Al current collectors for Li-Ion batteries made via a template-free electrodeposition process in ionic liquids

Article abstract of DOI:10.1149/1.3368667

Efficient electrical wiring of electrode is becoming a key issue in the design of Li-ion batteries utilizing insulating compounds. Presently, approaches acting at the material level are being pursued by combining techniques in decreasing the size of the particles and the use of carbon nanopainting. Here, we report the texturation of the Al current collector as an additional means to improve electrode kinetics. Aluminum electrodeposits were grown on aluminum and stainless steel substrates from ionic liquid electrolytic baths using pulse currents. The Al deposits, having a hilly morphology, nicely adhere to the substrate. Electrodes using these electrodeposits with host LiFePO₄ powders were shown to display excellent capacity retention and high rate capabilities.

Full text of DOI:10.1149/1.3368667

Journal of The Electrochemical Society, 157 ͑6͒ A641-A646 ͑2010͒
A641
0013-4651/2010/157͑6͒/A641/6/$28.00 © The Electrochemical Society
Al Current Collectors for Li-Ion Batteries Made via a
Template-Free Electrodeposition Process in Ionic Liquids
Cyrille Lecoeur,z Jean-Marie Tarascon, and Claude Guery
Laboratoire de Réactivité et Chimie des Solides, UFR des Sciences, ALISTORE–European Research
Institute, 80039 Amiens Cedex, France
Efficient electrical wiring of electrode is becoming a key issue in the design of Li-ion batteries utilizing insulating compounds.
Presently, approaches acting at the material level are being pursued by combining techniques in decreasing the size of the particles
and the use of carbon nanopainting. Here, we report the texturation of the Al current collector as an additional means to improve
electrode kinetics. Aluminum electrodeposits were grown on aluminum and stainless steel substrates from ionic liquid electrolytic
baths using pulse currents. The Al deposits, having a hilly morphology, nicely adhere to the substrate. Electrodes using these
electrodeposits with host LiFePO4 powders were shown to display excellent capacity retention and high rate capabilities.
© 2010 The Electrochemical Society. ͓DOI: 10.1149/1.3368667͔ All rights reserved.
Manuscript submitted January 4, 2010; revised manuscript received February 25, 2010. Published April 19, 2010.
Lithium-ion technology is the fastest growing and most promis-
ing battery technology, thanks to its high volume and specific energy
densities ͑ϳ3 times more than NiMH͒. To ensure this to be a sus-
tainable trend, new electroactive materials and better electrode de-
signs for batteries must be discovered to give a higher energy den-
sity and power rate capabilities. Moving from bulk to nanosized
materials has enabled the battery community to turn insulating com-
pounds ͑e.g., LiFePO4͒ into attractive electrode materials1 and to
trigger new reaction mechanisms such as conversion reactions,2
which have led to staggering capacity gains. Similarly, passing from
planar to nanostructured electrodes has been shown to greatly im-
prove kinetics, thanks to both the smaller diffusion length of Li+ in
the active material and a better electronic wiring of the electrode.3
Until now, obtaining a nanometer-sized material has revolved
around physical or chemical techniques, most often relying on the
help of templates as pioneered by Martin.4,5 Research involving
templates has recently found great resonance among researchers
with the successful assembly of high power rate negative electrodes
consisting of nanostructured copper current collectors coated with
Fe3O4 as an active material. Two different approaches have been
utilized to successfully prepare these electrodes: ͑i͒ a template-
assisted electrodeposition6 and ͑ii͒ a bottom-up approach pioneered
by our group,7 which only requires electrochemical and chemical
steps. Owing to the success of this technique, it was tempting to
prepare a nanoarchitectured Al current collector to host positive
electrode materials. Perre et al.8 showed that electrode capacity can
be enhanced as the surface contact between the current collector and
the LiCoO2 active material is increased.
selectively obtain a structured columnar deposit using a pulsed po-
tential electrodeposition method and an alumina membrane
template.8
Here, we report a template-free method to electrodeposit struc-
tured aluminum onto aluminum substrates via the use of four
different imidazolium-based ionic liquids: 1-ethyl-3-methyl-
imidazolium chloride ͓͑EMIm͔Cl͒, 1-ethyl-3-methylimidazolium
bis͑trifluoromethanesulfonyl͒imide
͓͑EMIm͔TFSI͒,
1-octyl-3-
methylimidazolium bis͑trifluoromethanesulfonyl͒imide ͓͑C8MIm͔
TFSI͒, and 1-hexadecyl-3-methylimidazolium bis͑trifluoromethane-
sulfonyl͒imide ͓͑C16MIm͔TFSI͒. Our pulsed electrodeposition
method utilizes an electrolytic bath composed of the above-
mentioned ionic liquids and AlCl3 as the source of aluminum. After
deposition, the electrochemical performances of these newly pro-
duced Al current collectors were evaluated using LiFePO4 as the
active material.
Experimental
All chemicals, manipulation, and electrochemical measurements
were conducted under inert atmosphere in an argon filled glove box
͑O2, H2O Ͻ 0.1 ppm͒. ͓EMIm͔TFSI, ͑99.9%, Solvionic͒ and
͓C8MIm͔TFSI ͑99.9%, Solvionic͒ were dehydrated under primary
vacuum for 1.5 h at room temperature, 2.5 h at 40°C, and 15 h at
70°C. This protocol decreases the amount of water from 20 to 4
ppm, as deduced by Karl–Fisher measurements. As ͓EMIm͔Cl
͑98%, Solvionic͒ and ͓C16MIm͔TSFI ͑99.9%, Solvionic͒ are solid at
room temperature, thermogravimetric studies were performed to de-
termine if the amount of water could interfere in our experiments. In
both cases, no loss of mass was observed. We can conclude that less
than 0.01% of water was present. We can assume that this amount is
less than 50 ppm because aluminum was deposited, which agrees
with the reported required condition.22 These ionic liquids were so
used as received as anhydrous AlCl3 ͑99%, Alfa Aesar͒. The elec-
trochemical baths were made under continuous stirring by a slow
addition of AlCl3 into the ionic liquids at room temperature. Tem-
peratures of 70°C were required for electrolytic bath based on
͓C16MIm͔TFSI, whose melting point is 67°C.
Electrochemical depositions were performed using a three-
electrode cell configuration connected to a potentiostat/galvanostat
BioLogic VMP3. The working electrode was a 1.7 cm2 aluminum
disk electrode ͑99%, Goodfellow͒, the counter electrode was a
2.5 cm2 aluminum strand ͑99%, Goodfellow͒, and the reference
electrode was an aluminum wire ͑99%, Goodfellow͒. All the experi-
ments were conducted without any stirring of the solution and in
total absence of external disturbances. Besides, for comparative pur-
poses, all measurements were strictly performed in the same cell
geometry ͑electrodes placed horizontally with a distance between
working and counter electrodes of 5 mm͒. Finally, all the potentials,
The electrodeposition of Al is not an easy task as it cannot be
done in aqueous media due to its high reactivity ͑E° = −1.67 V vs
normal hydrogen electrode͒. To solve problems arising from non-
aqueous solvent systems while preserving the feasibility of growing
aluminum deposits on a variety of substrates ͑e.g., Au,9,10 W,11-13
8,12,13
Mg alloy,14,15 Li–Al alloy,16 stainless steel,17-21 or Al
͒, many
groups have switched to the use of ionic liquids ͑ILs͒ as an electro-
lytic bath. All depositions were performed with a potentiostatic or
intentiostatic method that resulted in dense layers regardless of the
substrate. Additionally, Zein El Abedin et al.9 demonstrated that it is
possible to attain different sized Al particles on a Au͑111͒ substrate
by using different ionic liquids with the feasibility to obtain
Ϸ1.5 ␮m Al particles in the presence of ͓EMIm͔TFSI/AlCl3 as
opposed to nanometric Al growth with ͓BMP͔TFSI/AlCl3 at room
temperature. Moving to a less noble substrate ͑Al͒, Jiang et al.12
showed the feasibility to grow large ͑10 ␮m͒ and nicely faceted
aluminum particles in an ͓EMIm͔Cl/AlCl3 bath by electrochemical
deposition in a potentiostatic mode. Finally, Perre et al. were able to
z E-mail: cyrille.lecoeur@u-picardie.fr
Downloaded on 2014-05-13 to IP 193.146.32.73 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A642
Journal of The Electrochemical Society, 157 ͑6͒ A641-A646 ͑2010͒
Table I. Different ionic liquids used in this work.
Ionic liquid
Name
Cation
Anion
Cl−
l-Ethyl-3-methylimidazolium
chloride
N+
N
N
N
͓EMIm͔Cl
F
O
O
F
l-Ethyl-3-methylimidazolium
bis͑trifluoromethylsulfonyl͒imide
F
F
F
S
N
N
N
S
F
F
F
N+
͓EMIm͔TFSI
F
O
O
F
F
O
S
O
S
F
N+
l-Octyl-3-methylimidazolium
bis͑trifluoromethylsulfonyl͒imide
F
O
O
F
͓C8MIm͔TFSI
͓C16MIm͔TFSI
7
F
O
S
O
S
F
N+
N
l-Hexadecyl-3-methylimidazolium
bis͑trifluoromethylsulfonyl͒imide
F
O
O
F
15
if not otherwise specified, would be reported vs AlIII/Al. Lastly,
cyclic voltammograms ͑CVs͒ were collected, if not otherwise speci-
threshold concentration varies depending upon the nature of the sub-
17
strate ͑5 M for gold9 and 6 M for less noble substrates such as
fied, at 10 mV s−1
.
aluminum and stainless steel͒.
Before being used, substrates were mechanically polished,
washed in concentrated potassium hydroxide for 2 min, and subse-
quently washed in distilled water, ethanol, and acetone for 15 min
under ultrasonication. After deposition, the obtained deposits were
washed in acetone.
All deposits were observed by a scanning electron microscope
͑SEM, Philips XL30 FEG͒. SEM micrographs were then analyzed to
determine particle size and spacing, with Mesurim Pro software to
check the morphology of the deposits.
Homemade LiFePO4 powders, consisting of 150 nm carbon-
coated particles, were made via a previously described solvothermal
process.23 Acetone suspensions containing 80% LiFePO4/C, 10%
carbon, and 10% polyvinyldifluoride-hexafluoropropene ͑PVDF-
HFP͒ were prepared and placed dropwise separately on both planar
and aluminum structured current collector for comparative studies.
To obtain a meaningful comparison between the different elec-
trochemical baths, AlCl3 concentrations must be expressed in terms
of molar ratios ͑IL:AlCl3͒ because ͓C16MIm͔TFSI is solid at room
temperature. In addition to ͓EMIm͔TFSI, the other ͓XMIm͔TFSI
ionic liquids under investigation ͑where X corresponds to the alkyl
chain length͒ show phase segregation. Moreover, the ͑IL:AlCl3͒ mo-
lar ratio required to preserve electroactivity in the upper phase must
increase with the increasing alkyl chain length ͑1:1.5, 1:1.6, and 1:2
for ͓EMIm͔TFSI, ͓C8MIm͔TFSI, and ͓C16MIm͔TFSI, respectively͒.
In contrast, ͓EMIm͔Cl, which is solid at room temperature melts
with gas release in the presence of AlCl3, leading to a monophasic
bath regardless of the ͑IL:AlCl3͒ concentration. Therefore, we de-
cided to operate with a high molar ratio of AlCl3 ͑1:2͒ so as to
facilitate the formation of the well-known Al2Cl−7 electroactive an-
ion, first described in 1981 by Welch.24
Mindful of these early results on the phase stability of the vari-
ous IL-AlCl3 electrolytic baths, CVs ͑Fig. 1͒ were collected at a rate
of 10 mV s−1 as follows. For ͓EMIm͔Cl, the potential was cycled
from the open-circuit voltage ͑Ϸ+0.075 V vs AlIII/Al͒ to a lower
potential limit of Ϫ0.6 V in the negative direction and back to the
upper limits of +1 V along an anodic current. In contrast, wider
scanning voltage ranges of Ϫ1.4 to +2 V, Ϫ2 to +1 V, and Ϫ4 to
Electrodes,
having
the
same
quantity
of
materials
͑Ϸ0.24 mg cm−2͒͒, were dried overnight before assembling into a
battery.
Electrochemical tests vs Li were conducted in coin-cell-type
hardware with the cells connected to a potentiostat/galvanostat Bio-
Logic VMP2. The positive electrode was composed of aluminum
electrodeposits as a current collector and LiFePO4 as the active ma-
terial. A lithium metal foil, separated from the positive electrode by
a Whattman glass fiber soaked in a 1 M LiPF6 ethylene carbonate
dimethyl carbonate electrolyte solution, was used as the counter
electrode.
3.5
V were needed for ͓EMIm͔TFSI-, ͓C8MIm͔TFSI-, and
͓C16MIm͔TFSI-based electrolytic baths, respectively, to observe the
full AlIII reduction response. The CV profiles show that bulk depo-
sitions begin at potentials of Ϫ0.18, Ϫ0.26, Ϫ0.35, and Ϫ0.39 V for
͓EMIm͔Cl, ͓EMIm͔TFSI, ͓C8MIm͔TFSI, and ͓C16MIm͔TFSI, re-
spectively. More specifically, a reduction phenomenon, without a
defined peak, can be observed followed by two oxidation phenom-
ena at +0.04 and +0.41 V for ͓EMIm͔Cl. In contrast, well-defined
reduction peaks can be observed at Ϫ0.86, Ϫ1.38, and Ϫ2.9 V
͑bottom of peaks͒ together with their reversible oxidation peaks at
+0.31, +0.52, and +1.96 V ͑top of peaks͒ for ͓EMIm͔TFSI,
͓C8MIm͔TFSI, and ͓C16MIm͔TFSI, respectively. Current loops ob-
servable at Ϫ0.16, Ϫ0.12, Ϫ0.16, and Ϫ0.19 V for ͓EMIm͔Cl,
͓EMIm͔TFSI, ͓C8MIm͔TFSI, and ͓C16MIm͔TFSI, respectively, are
attributed to the aluminum nucleation. Overall, there is a decrease in
the deposition potential inducing an increase in ⌬E ͑the difference
between the bottom of the reduction peak and the top of the oxida-
tion peak͒, reminiscent of a sluggish kinetic, with increasing alkyl
chain length and as we move from smaller to bigger anions ͑Fig.
1e͒. Steric hindrance is most likely the reason for this increase in
⌬E.
Results and Discussion
The key to any successful electrodeposition resides in the com-
position of the electrolytic bath, and ionic liquids are no exception.
An important aspect to consider when selecting the proper ionic
liquid ͑in addition to solvation properties͒ is electrochemical stabil-
ity. Therefore, in light of such considerations and previously con-
ducted studies on Al electrodeposits, we focused our attention on
imidazolium-based ionic liquids ͑Table I͒, which offer electrochemi-
cal stability up to 5 V.
Mixtures of ionic liquids and AlCl3 were prepared and handled
in an argon dry box owing to the extreme moisture sensitivity of the
Al-based precursor. For ͓EMIm͔TFSI, a strong phase miscibility de-
pendence is obtained depending upon the ionic liquid/AlCl3 concen-
tration with a phase separation occurring when the AlCl3 concentra-
tion approaches 2 M. The upper brown phase became darker with an
increasing aluminum chloride concentration. The lower phase of
higher viscosity is translucent and crystallizes within 1 week. Elec-
troactivity was found only in the upper phase when the AlCl3 con-
centration reached 6 M. These results did not come as a total sur-
prise, as they were previously observed in ͓EMIm͔TFSI by Zein El
Abedin et al. Additionally, they reported that the electroactive
In light of the previous findings, regarding the obtention of dense
deposits via potentiostatic and intentiostatic methods, we decided, to
achieve structured Al deposits without the help of a template, to
move to a pulsed current electrodeposition method. Through fine-
tuning of the electrodeposition parameters ͑total deposition time,
Downloaded on 2014-05-13 to IP 193.146.32.73 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, 157 ͑6͒ A641-A646 ͑2010͒
A643
E (V vs. AlIII/Al)
-0.5
0.0
0.5
1.0
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5
15
30
20
10
0
a) [EMIm]Cl/AlCl3
b) [EMIm]TFSI/AlCl3
10
5
0
-10
-20
-5
-30
4
-10
20
c) [C8MIm]TFSI/AlCl3
d) [C16MIm]TFSI/AlCl3
2
0
15
10
5
-2
-4
-6
-8
0
Figure 1. CVs at
a
scan rate of
10 mV s−1 of ͑a͒ ͓EMIm͔Cl/AlCl3 ͑1:2͒,
-5
Tamb
Tamb
,
,
͑b͒ ͓EMIm͔TFSI/AlCl3 ͑1:1.5͒,
͑c͒ ͓C8MIm͔TSFI/AlCl3 ͑1:1.6͒,
-10
-15
-20
-25
Tamb, and ͑d͒ ͓C16MIm͔TSFI/AlCl3 ͑1:2͒,
70°C. ͑e͒ Potential values of bottom of
reduction peaks, current loops, bulk depo-
sitions, and top of oxidation peaks de-
duced from the CV curves for each elec-
trochemical bath.
-2.0 -1.5 -1.0 -0.5
0.0
0.5
1.0 -5 -4 -3 -2 -1
0
1
2
3
4
E (V vs. AlIII/Al)
bottom of reduction peak
current loop
bulk deposition
top of oxidation peak
2
1
e)
0
-1
-2
-3
[EMIm]Cl [EMIm]TFSI [C8MIm]TFSI [C16MIm]TFSI
pulse intensity, duration, relaxation time, and temperature͒, we were
able to reach optimum conditions. Deposition times ranging from 52
s to 1 h, pulse times from 100 to 300 ms, relaxation times from 1 to
11 s, and temperatures from 23 to 100°C were tried. The optimal
parameters to obtain Al deposits covering the whole Al substrate at
room temperature for ͓EMIm͔Cl, ͓EMIm͔TFSI, and ͓C8MIm͔TFSI
and at 70°C for ͓C16MIm͔TFSI were a pulse duration of 200 ms and
a relaxation time of 5 s.
The deposits formed from the ͓EMIm͔TFSI/AlCl3 bath ͑Fig. 2͒
highlights the influence of pulse intensity on the morphology of the
Al. While maintaining a constant pulse duration and deposition time,
pulse intensities were increased from Ϫ0.6 to −6.21 mA cm−2. All
deposits consisted of aluminum balls that varied in size and dis-
tances between deposits, leading to various surfaces roughness as
deduced by SEM micrographs ͑Fig. 2͒. More specifically, there is an
evolution from the isolated aluminum particles in ͑a͒ for low cur-
rents to agglomerated particles forming three-dimensional ͑3D͒ en-
tities in ͑d͒ for higher currents.
evenly spaced.25 To best quantify these two parameters, SEM mi-
crographs were exploited with Mesurim Pro software26 treating iso-
lated particles and agglomerates as unique entities. Histograms ͑Fig.
2͒ were drawn, and every reported entry is an average of 200 mea-
surements. A low current deposition results in good homogeneity of
particle sizes with diameters between 100 nm and 2 ␮m ͑a ͒. In-
Ј
creasing the current intensity enlarges this repartition ͑b ͒ and
Ј
shows two distinguishable parts due to the presence of the isolated
͑0–2 ␮m range͒ and agglomerated ͑Ͼ2 ␮m range͒ particles at the
same time. Operating at −4.52 mA cm−2
͑cЈ͒ results in agglomer-
ates with a sharpening of their size distribution ͑cЈ͒ compared to
either lower ͑bЈ͒ or higher ͑d ͒ currents. For reasons not yet deter-
Ј
mined, the spreading of the distances between the particles some-
what tracks down their size distribution with the best distance uni-
formity for low currents ͑a ͒. In contrast, at higher currents, the
Љ
distances between the agglomerates, regardless of two-dimensional
͑2D͒ or 3D growth, are highly dispersed ͓͑b ͒, ͑c ͒, and ͑d ͔͒. Over-
Љ Љ
Љ
all, we can conclude that the highly structured deposits consisting of
single particles are obtained at low currents. However, for the appli-
cations, aggregates are preferred because they offer a better grating
For an efficient current distribution, an architectured current col-
lector should have agglomerates that are homogeneous in size and
Downloaded on 2014-05-13 to IP 193.146.32.73 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A644
Journal of The Electrochemical Society, 157 ͑6͒ A641-A646 ͑2010͒
a)
b)
c)
d)
3 µm
4 µm
4 µm
4 µm
100 µm
100 µm
100 µm
100 µm
%
d’)
a’)
b’)
c’)
Figure 2. SEM micrographs of aluminum
deposits on aluminum substrates obtained
in ͓EMIm͔TFSI/AlCl3 ͑1:1.5% mol͒ with
corresponding histograms of agglomerate
size and distance between agglomerates
15
10
5
after 10 min at intensities of ͓͑a͒, ͑aЈ͒, and
͑a ͔͒ j = −0.6 mA cm−2
, ͓͑b͒, ͑bЈ͒, and
0
Љ
, ͓͑c͒, ͑cЈ͒, and
0
2
4
6
8
0
2
4
6
8
0
2
4
6
8
0
2
4
6
8
͑b ͔͒ j = −1.07 mA cm−2
Љ
͑c ͔͒ j = −4.52 mA cm−2
Particles diameter (µm)
%
Љ
, and ͓͑d͒, ͑dЈ͒,
and ͑d ͔͒ j = −6.21 mA cm−2
.
d’’)
a’’)
b’’)
c’’)
20
15
10
5
Љ
0
0
5
10 15 20 25 30
0
5
10 15 20 25 30
0
5
10 15 20 25 30
0
5
10 15 20 25 30
Distance between particles (µm)
of the surface. Within this context, the deposits prepared at
−4.52 mA cm−2 ͑Fig. 2c͒, showing an acceptable size distribution
but poorly spaced agglomerates, were worth further investigations
with respect to the effect of the nature of the ionic liquid
͓͑EMIm͔Cl, ͓EMIm͔TFSI, ͓C8MIm͔TFSI, and ͓C16MIm͔TFSI͒
used on the quality of the Al deposit structuration.
Figure 3 compiles the results of such a study. All the deposits
were performed with a total deposition time of 10 min and an inten-
sity of −4.52 mA cm−2. SEM micrographs were collected for all
the deposits and exploited as above to draw histograms. The SEM
micrographs for the deposits made from the ͓EMIm͔Cl- and
͓EMIm͔TFSI-based electrochemical baths ͑Fig. 3a and b͒ display
the same morphological characteristics, with the presence of homo-
geneously distributed 2D aluminum agglomerates having mean di-
ameters and mean spacing distances of 2.3 and 7.5 ␮m and 1.95
and 7.9 ␮m, respectively. The nature of the anion does not seem to
play a major role in the deposition processes or in the structuration
of the aluminum deposits. In contrast, more apparent differences in
the aluminum deposits were observed by changing the nature of the
cations. Deposits from ͓C8MIm͔TFSI and ͓C16MIm͔TFSI ͑3c and
3d͒ offer isolated and smooth spherical agglomerates when com-
pared with that of ͓EMIm͔TFSI with, additionally in both cases, a
narrowing of the spreading in size and distance. The deposits ob-
tained from the ͓C8MIm͔TFSI-based electrolytic baths show
0.8 ␮m agglomerates separated by 2.1 ␮m gaps, while those ob-
tained from an ionic liquid having
a longer alkyl chain
͓C16MIm͔TFSI have a mean diameter of 2.3 ␮m and are separated
by 7.5 ␮m. These morphological characteristics, which approach
those of ͓EMIm͔TFSI, are most likely due to the minimal growing
temperature ͑70°C͒, which allows a better diffusion of electroactive
species. Overall, it results from this study that the use of cations
with long alkyl chains favors deposits having smaller or isolated
entities.
At this juncture, thanks to both the use of a pulsed electrodepo-
sition method combined with the proper ionic liquid formulation,
current collectors with surfaces made of homogeneously sized and
spaced Al agglomerates were achieved. The most elegant way to
determine the positive attributes of ͑quasi-͒structured current collec-
tors in improving electrode kinetics is to test them in combination
with electrode materials showing poor kinetics. For this reason,
LiFePO4, whose poor electronic and ionic conductivities have been
shown to be overcome by a combination of downsizing particle size
and through the use of conducting nanocoatings,27 appears as an
ideal candidate for such a study. Figure 4a shows evidence that the
active material follows the surface landscape of the newly produced
current collectors.
a)
b)
c)
d)
2 µm
2 µm
2 µm
2 µm
100 µm
100 µm
100 µm
100 µm
8
7
6
5
4
3
2
1
0
a) [EMIm]Cl
b) [EMIm]TFSI
c) [C8MIm]TFSI d) [C16MIm]TFSI
Figure 4b shows the voltage composition profile for a Li/half-cell
using the LiFePO4 self-supported on our structured Al current col-
lector as the positive electrode together with its capacity retention.
The cell shows excellent capacity retention up to 45 cycles; the
electrode still maintains 95% of its initial capacity ͑inset͒. Similar
capacity retention was obtained for the flat aluminum substrate. The
Figure 3. Mean diameter and mean distance between agglomerates and cor-
responding SEM micrographs for deposits obtained on aluminum substrates
for 10 min at a current density of −4.52 mA cm−2 in ͑a͒ ͓EMIm͔Cl/AlCl3
͑1:2͒, ͑b͒ ͓EMIm͔TFSI/AlCl3 ͑1:1.5͒, ͑c͒ ͓C8MIm͔TSFI/AlCl3 ͑1:1.6͒, and
͑d͒ ͓C16MIm͔TSFI/AlCl3 ͑1:2͒.
Downloaded on 2014-05-13 to IP 193.146.32.73 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, 157 ͑6͒ A641-A646 ͑2010͒
A645
a)
100
80
60
500 nm
2 µm
40
20
0
40 µm
0.01
0.1
1
10
5 µm
Rate (C)
b)
4.0
3.5
3.0
2.5
2.0
Figure 5. Comparative normalized high rate capabilities of LiFePO4 on
stainless steel substrates: ͑ᮀ͒ Planar substrate and ͑᭿͒ aluminum deposit
with an inset showing its corresponding SEM micrograph ͑deposit obtained
in ͓EMIm͔TFSI/AlCl3 ͑1:1.5͒ for an intensity of j = −6.15 mA cm−2 for 10
min͒.
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
C/10
using either ͓EMIm͔TFSI or ͓C8MIm͔TFSI ionic liquids show the
same behavior as on the flat substrate ͑see curve ࡗ͒. In that case,
even if the Al deposited balls are well structured ͑Ϸ1.0 ␮m diam-
eter and Ϸ2.0 ␮m distance͒, they are too small so that they were
totally covered by the active material, hence behaving like a planar
surface. By comparison, the composite electrodes built on the Al
deposits obtained from ͓EMIm͔Cl, ͓EMIm͔TFSI ͑agglomerates of
particles͒, and ͓C16MIm͔TFSI ͑isolated particles͒ ionic liquids and
having the following characteristics ͑Al “balls” of Ϸ2.0 ␮m in di-
ameter separated by Ϸ7.5 ␮m distance͒ can still deliver 78% of its
initial capacity at a rate of 20C ͑see curve b͒. The high rate capa-
bility of these confectioned electrodes is displayed by plotting the
cell voltage profile for various current drains ͑inset of Fig. 4c͒. The
composite electrodes made on the Al deposits having bigger par-
ticles ͑Ϸ7.0 ␮m diameter͒ but less distanced ͑Ϸ5.5 ␮m͒ obtained
by modifying the deposit conditions ͑duration time and intensity;
−4.52 mA cm−2 for 1 h͒ still show ͑see curve ᭿͒ rate capability
improvement with respect to planar Al substrates ͑48% of the initial
capacity at a rate of 20C͒. Although we acknowledge that the
makeup of our LiFePO4 loaded PVDF-HFP film placed on top of
our rough Al deposited surfaces does not allow taking full benefit of
the whole aluminum deposit surface, it still has the merit to reflect
the importance of the current collector texturation.
0
10
20
30
40
50
Cycle
0.0
0.2
0.4
0.6
0.8
1.0
x in LixFePO4
c)
100
80
60
40
20
0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
C/10
20C
C/50
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
x in LixFePO4
0.01
0.1
1
10
C rate
Another substrate, which has a great interest as a current collec-
tor, is stainless steel. The inset of Fig. 5 shows a homogeneous
deposit obtained for an intensity of −6.15 mA cm−2 for a total
deposition time of 10 min in the ͓EMIm͔TFSI/AlCl3 electrochemi-
cal bath. The pulsed intensity electrodeposition methods allowed the
same texturation as the one we observed on aluminum substrates,
i.e., aluminum aggregates forming spheres. By a potentiostatic
method, Liu et al.19,20 and Zein El Abedin et al.17 observed some
problems of adherence of their deposits. They bypassed these diffi-
culties by preoxidizing the substrate to obtained dense aluminum
coatings free of structuration. For not yet understood reasons, using
the pulsed current deposition method, we experienced the formation
of adherent and structured deposits without any need of pretreat-
ments. A newly prepared LiFePO4/aluminum deposit on a stainless
steel electrode can still deliver 80% of its initial capacity at a rate of
10C ͑e.g., full discharge in 6 min͒, as shown in Fig. 5. The rate
capability of such electrodes is benchmarked to a corresponding flat
one, whose capacity starts decreasing at a rate of C/10 and which
only delivers 65% of its capacity at 10C ͑curve ᮀ in Fig. 5͒. This
finding further proves the positive attributes of architectured current
collectors over planar ones, with the difference between stainless
steel and planar Al substrates resulting from the quality of the inter-
Figure 4. ͑a͒ SEM micrographs of homemade LiFePO4 ͑inset͒ deposed on
an aluminum deposit from ͓EMIm͔TFSI/AlCl3 bath at an intensity of
−4.52 mA cm−2 for 10 min, ͑b͒ voltage composition profile and capacity
retention ͑inset͒ of Li/half cell using LiFePO4 self-supported on structured Al
deposit, ͑c͒ comparative normalized high rate capabilities of active material
on aluminum substrates: ͑᭺͒ Planar substrate, ͑ࡗ͒ small aluminum balls
͑Ϸ1.0 ␮m diameter and Ϸ2.0 ␮m distance͒, ͑b͒ medium aluminum balls
͑Ϸ2.0 ␮m diameter and Ϸ7.5 ␮m distance͒ with inset showing its corre-
sponding high rate capability response, ͑᭿͒ big aluminum balls ͑Ϸ7.0 ␮m
diameter and Ϸ5.5 ␮m distance͒.
high rate capability obtained for deposits made on aluminum sub-
strates is shown in Fig. 4c. Batteries are charged at a rate ͑C/10͒ and
then successively discharged at decreasing rates ͑20C, 10C, 5C, 2C,
1C, C/2, C/5, C/10, C/20, and C/50͒ with a relaxing time of 30 min
between each step. The electrode made by placing the active mate-
rial on a flat aluminum substrate ͑see curve ᭺͒ keeps its full capac-
ity till 1C but loses it by increasing the discharge rate to 20C ͑e.g.,
full discharge in
3
min͒. The electrodes consisting of
LiFePO4/C/PVDF-HFP composites deposited onto nanotextured Al
substrates obtained at low currents from the electrochemical baths
Downloaded on 2014-05-13 to IP 193.146.32.73 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A646
Journal of The Electrochemical Society, 157 ͑6͒ A641-A646 ͑2010͒
4. C. R. Martin, Science, 266, 1961 ͑1994͒.
5. C. R. Martin, Chem. Mater., 8, 1739 ͑1996͒.
6. P. L. Taberna, S. Mitra, P. Poizot, P. Simon, and J. M. Tarascon, Nature Mater., 5,
567 ͑2006͒.
face ͑e.g., adhesion͒, which appears to be better for Al substrates.
This does not come as a total surprise when electrodepositing on the
same metal vs on a different metal.
7. A. Finke, P. Poizot, C. Guery, D. Mazouzi, and J. M. Tarascon, Adv. Funct. Mater.,
18, 3598 ͑2008͒.
Conclusions
8. E. Perre, L. Nylhom, T. Gustafsson, P. L. Taberna, P. Simon, and K. Edström,
Electrochem. Commun., 10, 1467 ͑2008͒.
9. S. Zein El Abedin, E. M. Moustafa, R. Hempelmann, H. Natter, and F. Endres,
Phys. Chem. Chem. Phys., 7, 1535 ͑2006͒.
10. S. Zein El Abedin, E. M. Moustafa, R. Hempelmann, H. Natter, and F. Endres,
We show that it is possible to obtain a structured aluminum de-
posit by a pulsed intensity electroreduction method without the help
of a template. Aluminum is electrodeposited by the reduction of AlIII
species in imidazolium-based ionic liquids where aluminum chlo-
ride was added. Intensity steps were applied periodically to obtain a
homogeneous deposit. These deposits were used as current collec-
tors in Li batteries using LiFePO4 as an active material. Electro-
chemical tests confirmed the well-known capacity retention for
LiFePO4 electrodes and reveal the feasibility to achieve high rate
capability, thanks to the specifically designed current collector ar-
chitecture. In light of such a finding, the direct electrochemical
growth of carbon-free LiFePO4 in ionic liquid media on textured Al
current collectors is presently being studied.
Electrochem. Commun., 7, 1111 ͑2005͒.
11. Y. Zhao and T. J. Vandernoot, Electrochim. Acta, 42, 1639 ͑1997͒.
12. T. Jiang, M. J. Chollier Brym, G. Dubé, A. Lasia, and G. M. Brisard, Surf. Coat.
Technol., 201, 1 ͑2006͒.
13. T. Jiang, M. J. Chollier Brym, G. Dubé, A. Lasia, and G. M. Brisard, Surf. Coat.
Technol., 201, 10 ͑2006͒.
14. J. K. Chang, S. Y. Chen, W. T. Tsai, M. J. Deng, and I. W. Sun, Electrochem.
Commun., 9, 1602 ͑2007͒.
15. J. K. Chang, S. Y. Chen, W. T. Tsai, M. J. Deng, and I. W. Sun, J. Electrochem.
Soc., 155, C112 ͑2008͒.
16. U. Bardi, S. Caporali, M. Craig, A. Giorgetti, I. Perissi, and J. R. Nicholls, Surf.
Coat. Technol., 203, 1373 ͑2009͒.
17. S. Zein El Abedin, Z. Phys. Chem., 220, 1293 ͑2006͒.
18. S. Caporali, A. Fossati, A. Lavacchi, I. Perissi, A. Tolstogouzov, and U. Bardi,
Corros. Sci., 50, 534 ͑2008͒.
19. Q. X. Liu, S. Zein El Abedin, and F. Endres, Surf. Coat. Technol., 201, 1352
͑2006͒.
Acknowledgments
The authors thank the European Commission for the financial
support in the frame of SUPERLION program no. 214832-2 ͑FP7
program͒.
20. Q. X. Liu, S. Zein El Abedin, and F. Endres, J. Electrochem. Soc., 155, D357
͑2008͒.
Université de Picardie Jules Verne assisted in meeting the publication
costs of this article.
21. G. Yue, X. Lu, Y. Zhu, X. Zhang, and S. Zhang, Chem. Eng. J., 147, 79 ͑2009͒.
22. A. P. Abbott, I. Dalrymple, F. Endres, and D. R. MacFarlane, Electrodeposition
from Ionic Liquids, Chap. 1, Wiley-VCH Verlag, Weinheim ͑2008͒.
23. C. Delacourt, P. Poizot, and C. Masquelier, World Pat. WO 2007/000251 ͑2007͒.
24. B. J. Welch and R. A. Osteryoung, J. Electroanal. Chem. Interfacial Electrochem.,
118, 455 ͑1981͒.
References
1. C. Delacourt, P. Poizot, S. Levasseur, and C. Masquelier, Electrochem. Solid-State
Lett., 9, A352 ͑2006͒.
2. J. M. Tarascon, S. Grugeon, M. Morcrette, S. Laruelle, P. Rozier, and P. Poizot, C.
25. N. Li and C. R. Martin, J. Electrochem. Soc., 148, A164 ͑2001͒.
26. http://pedagogie.ac-amiens.fr/svt/info/logiciels/Mesurim2/Telecharge.htm, 2008.
27. J. M. Tarascon and M. Armand, Nature Mater., 414, 359 ͑2001͒.
R. Chim., 8, 9 ͑2005͒.
3. A. S. Aricò, P. Bruce, B. Scrosati, J. M. Tarascon, and W. van Schalkwijk, Nature
Mater., 4, 366 ͑2005͒.
Downloaded on 2014-05-13 to IP 193.146.32.73 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).