Effects of some organic additives on lithium deposition in propylene carbonate

Article abstract of DOI:10.1149/1.1516770

The effects of some film-forming organic additives, fluoroethylene carbonate (FEC), vinylene carbonate (VC), and ethylene sulfite (ES), on lithium deposition and dissolution were investigated in 1 M LiClO₄ dissolved in propylene carbonate (PC) as a base solution. When 5 wt % FEC was added, the cycling efficiency was improved. On the contrary, addition of 5 wt% VC or ES significantly lowered the cycling efficiency. The surface morphology of lithium deposited in each electrolyte solution was observed by in situ atomic force microscopy (AFM). In PC + FEC, the surface was covered with a uniform and closely packed layer of particle-like deposits of about 100-150 nm diam. The surface film seemed to be more solid in PC + VC, and inhomogeneous in PC + ES. From ac impedance measurements, it was revealed that the surface film formed in PC + FEC has a lower resistance than that in the additive-free solution, whereas that formed in PC + VC or PC + ES has a higher resistance. Large volume changes during lithium deposition and dissolution require that the surface film should be elastic (or soft) and be self-repairable when being damaged. In addition, a nonuniform current distribution is liable to cause dendrite formation, which requires that the surface film should be uniform and its resistance should be as low as possible. PC + FEC gave a surface film that satisfies all these requirements, and therefore only FEC was effective as an additive for deposition and dissolution of lithium metal.

Full text of DOI:10.1149/1.1516770

A1578
Journal of The Electrochemical Society, 149 ͑12͒ A1578-A1583 ͑2002͒
0013-4651/2002/149͑12͒/A1578/6/$7.00 © The Electrochemical Society, Inc.
Effects of Some Organic Additives on Lithium Deposition
in Propylene Carbonate
,c,z
Ryo Mogi,a,b Minoru Inaba,a,
Soon-Ki Jeong,a Yasutoshi Iriyama,a
*
Takeshi Abe,a, and Zempachi Ogumi
a,
*
*
aDepartment of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto University,
Sakyo-ku, Kyoto 606-8501, Japan
bKanto Denka Kogyo Company, Limited, Shibukawa Laboratory, Gunma 377-8513, Japan
The effects of some film-forming organic additives, fluoroethylene carbonate ͑FEC͒, vinylene carbonate ͑VC͒, and ethylene sulfite
͑ES͒, on lithium deposition and dissolution were investigated in 1 M LiClO4 dissolved in propylene carbonate ͑PC͒ as a base
solution. When 5 wt % FEC was added, the cycling efficiency was improved. On the contrary, addition of 5 wt % VC or ES
significantly lowered the cycling efficiency. The surface morphology of lithium deposited in each electrolyte solution was ob-
served by in situ atomic force microscopy ͑AFM͒. In PC ϩ FEC, the surface was covered with a uniform and closely packed layer
of particle-like deposits of about 100-150 nm diam. The surface film seemed to be more solid in PC ϩ VC, and inhomogeneous
in PC ϩ ES. From ac impedance measurements, it was revealed that the surface film formed in PC ϩ FEC has a lower resistance
than that in the additive-free solution, whereas that formed in PC ϩ VC or PC ϩ ES has a higher resistance. Large volume
changes during lithium deposition and dissolution require that the surface film should be elastic ͑or soft͒ and be self-repairable
when being damaged. In addition, a nonuniform current distribution is liable to cause dendrite formation, which requires that the
surface film should be uniform and its resistance should be as low as possible. PC ϩ FEC gave a surface film that satisfies all
these requirements, and therefore only FEC was effective as an additive for deposition and dissolution of lithium metal.
© 2002 The Electrochemical Society. ͓DOI: 10.1149/1.1516770͔ All rights reserved.
Manuscript submitted February 19, 2002; revised manuscript received May 25, 2002. Available electronically October 24, 2002.
Lithium metal is the most attractive material for use as a negative
lithium deposition and dissolution. The counter and reference elec-
trodes were lithium metal. The base electrolyte solution was com-
mercially available 1 mol dmϪ3 ͑M͒ lithium perchlorate (LiClO4)
dissolved in propylene carbonate ͑PC; Kishida Reagent Chemicals,
lithium-battery grade͒. FEC ͑Kanto Denka Kogyo͒, VC ͑Aldrich͒,
and ES ͑Aldrich͒ were added 5 wt % each to the base solution. Each
solution was dried over 4A molecular sieves for weeks and was used
for measurements after the water content dropped less than 30 ppm.
An electrochemical cell made of polytetrafluoroethylene ͑PTFE͒
was used for cycling tests. The geometric surface area of the work-
ing electrode was fixed at 0.8 cm2 using an O-ring. The current
density for lithium deposition and dissolution was 0.5 mA cmϪ2. In
each cycle, 0.3 C cmϪ2 of lithium was deposited and dissolved com-
pletely until the potential reached 1.5 V vs. Liϩ/Li.
electrode in rechargeable cells because of its high energy density.
However, dendritic deposition of lithium metal during repeated
charge/discharge cycles is a serious problem that is responsible for
low cycling efficiencies and safety issues.1 On lithium metal, a pro-
tective surface film, called the solid electrolyte interface ͑SEI͒, is
formed, which is known to have a great influence on the morphol-
ogy of deposited lithium.2,3 To modify and control the morphology,
and physical and chemical properties of the surface film, different
kinds of additives have been proposed. These include HF,4 CO2 ,5
AlI3 ,6 SO2 ,7 nitromethane,7 polyethylene glycol dimethyl ether,8
silicone/polypropylene oxide copolymer,9 cetyltrimethylammonium
chloride,10 ethyl trifluoroacetate,11 and aromatic compounds such
as benzene,12 toluene,12 2-methylfuran,12 2-methylthiophene,12
triazoles,13 dipyridyl derivatives,14 etc.
AFM observation of the surface morphology of deposited lithium
was carried out with a PicoSPM® system ͑Molecular Imaging͒
equipped with a PicoStat® potentiostat ͑Molecular Imaging͒. A
laboratory-made fluid cell made of PTFE was set on a sample stage.
The geometric surface area of the nickel substrate was fixed at
1.2 cm2 using an O-ring. AFM images were obtained in the contact
mode using a piezoelectric scanner with scan ranges of 7 ␮m in the
x- and y-directions. The microcantilever made of Si3N4 was
scanned at 21 ␮m sϪ1 to obtain AFM images. Lithium was depos-
ited at 0.5 mA cmϪ2. During deposition, the microcantilever was
moved out of the solution because the current distribution beneath
the tip would have been disturbed by its presence. After every
0.03 C cmϪ2 deposition, AFM images were obtained under open-
circuit conditions. In addition to the constant current deposition,
cyclic voltammetry ͑CV͒ was carried out at 5 mV sϪ1 between 2.5
and 0.05 V, and AFM images after five cycles of CV were obtained
in a similar manner.
AC impedance was measured with a frequency analyzer ͑SI
1255, Solartron͒ coupled with a potentiostat ͑model 273A, EG&G
PAR͒ and the cell used for the cycling tests. After lithium was de-
posited on the nickel substrate at Ϫ0.1 V for 300 s, the impedance
was measured under open-circuit conditions over the frequency
range of 100 kHz to 100 mHz. The perturbation amplitude for alter-
nating polarization was 5 mV.
The importance of the protective surface film is a common issue
to graphite negative electrodes. It has been recently reported that
some film-forming additives, fluoroethylene carbonate ͑FEC͒,15 vi-
nylene carbonate ͑VC͒,16 and ethylene sulfite ͑ES͒,17 are effective
for graphite negative electrodes. In a previous study using in situ
atomic force microscopy ͑AFM͒, we reported that all these additives
easily decompose and leave effective surface films on graphite nega-
tive electrodes at potentials more positive than 1 V before the main
solvent decomposes.18 These additives are expected to be effective
not only for graphite but also for lithium metal, though there has
been no report for the latter in the literature. In this report, we
studied the effects of these additives on lithium deposition and dis-
solution. The results of charge/discharge measurements were corre-
lated with the surface morphologies of deposited lithium observed
by in situ AFM, and with the resistances of the surface films esti-
mated by ac impedance spectroscopy.
Experimental
Nickel plates ͑Nilaco Co., 20 ϫ 20 ϫ 0.3 mm) were polished
with alumina powder to a mirror finish and used as substrates for
* Electrochemical Society Active Member.
c Present address: Department of Molecular Science and Technology, Faculty of
Engineering, Doshisha University, Kyotanabe, Kyoto 610-0321, Japan.
z E-mail: minaba@mail.doshisha.ac.jp
Downloaded on 2014-05-07 to IP 131.215.25.218 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, 149 ͑12͒ A1578-A1583 ͑2002͒
A1579
Figure 1. Cycling efficiencies for deposition and dissolution of lithium on
nickel substrates in 1 M LiClO4 /PC without additive and with 5 wt % FEC,
VC, and ES. The current density was 0.5 mA cmϪ2. The amount of lithium
deposition was 0.3 C cmϪ2 in each cycle. The cutoff voltage for dissolution
was 1.5 V.
All experiments were carried out at 30°C in an argon-filled glove
box ͑Miwa Industries͒ with a dew point below Ϫ60°C.
Results and Discussion
Figure 2. AFM images ͑7 ϫ 7 ␮m͒ of the surface morphology on a nickel
substrate obtained ͑a͒ before and after ͑b͒ 0.09, ͑c͒ 0.15, and ͑d͒ 0.3 C cmϪ2
deposition of lithium metal in 1 M LiClO4 /PC. Lithium was deposited at
Cycling efficiencies for lithium deposition and dissolu-
tion.—Lithium deposition and dissolution were carried out on nickel
substrates in 1 M LiClO4 /PC solutions with and without the addi-
tives. The variations of the cycling efficiencies for lithium deposi-
tion and dissolution with cycle number are shown in Fig. 1. In the
additive-free solution, the cycling efficiency was around 80% in the
first cycle. However, it decreased gradually and was leveled off
between 60 and 70% after the 10th cycle. The addition of 5 wt %
FEC obviously improved the efficiency, i.e., the efficiency was
higher than 80% up to the 30th cycle. The effects of VC and ES
addition were contrary to what we expected. The efficiency de-
creased rapidly after the 10th cycle and became lower than 10%. It
should be noted that all these additives are effective for graphite
negative electrodes as mentioned earlier. However, only the addition
of FEC improved the cycling efficiency for lithium deposition and
dissolution in the present study. Thus, the properties required for the
surface film on lithium seem to be different from that on graphite.
0.5 mA cmϪ2
.
reasonable to presume that these large agglomerates lead to the for-
mation of dendrites during repeated deposition/dissolution cycles,
which lowers the cycling efficiency.
AFM images obtained in 1 M LiClO4 /PC ϩ 5 wt % FEC are
shown in Fig. 3. Even after 0.09 C cmϪ2 of lithium was deposited,
the whole surface of the nickel substrate was completely covered
with fine deposits as shown in Fig. 3a. The surface was very smooth
and the original grooves on the substrate were invisible. Upon fur-
ther deposition, the particle size increased slightly; however, almost
no clear morphological changes were observed. Figure 3b shows an
image obtained after deposition of 0.3 C cmϪ2. The surface was still
covered with a thick, closely packed layer of fine particles with no
sign of large agglomerate formation. Figure 3c shows a magnified
image of the 1 ϫ 1 ␮m area indicated by a square in Fig. 3b. The
sizes of the particles were several tens of nanometers with a uniform
size distribution.
Morphologies of deposited lithium.—The surface morphologies
of lithium deposited on nickel substrates were observed by in situ
AFM to clarify the correlation between the surface morphology and
the cycling efficiency. AFM images obtained in the additive-free 1
M LiClO4 /PC are shown in Fig. 2. Figure 2a shows an image before
lithium deposition. Many grooves, which were made from polishing,
are seen on the nickel surface. After lithium deposition of
0.03 C cmϪ2, particle-like deposits of about 100 nm diam appeared
sparsely ͑not shown͒. The number of the deposits increased, and
their sizes grew up with the amount of deposition. After deposition
of 0.09 C cmϪ2, the sizes of particle-like deposits were not uniform
over the range of 100-200 nm diam ͑Fig. 2b͒. In addition, the
grooves on the nickel substrate became indistinct, which means that
the surface was covered with a thin film. Some noisy lines are seen
running horizontally in the image, which were made by the colli-
sions of the microcantilever with large bulges on the surface and
thus indicate that the surface was rough. In Fig. 2c, which shows an
image obtained after deposition of 0.15 C cmϪ2, a large deposit ap-
Surface morphologies after the following dissolution to 1.5 V
and after another 0.3 C cmϪ2 deposition are shown in Fig. 4. Sur-
prisingly, the morphology of the precipitate layer remained almost
unchanged during the dissolution and redeposition. Kanamura et al.
observed the morphology of lithium deposited in LiClO4 /PC con-
taining a trace amount of HF by scanning electron microscopy
4
͑SEM͒ and AFM.21 They reported that hemispherical deposits of
about 300 nm diam were formed on deposition. During the follow-
ing dissolution process, they observed the formation of holes and
defects on the deposits accompanied by shrinkage of the deposits
and concluded that each particle consisted of lithium metal covered
with a surface film. However, this does not seem to be the case for
the precipitates in Fig. 3 and 4, which showed no morphological
changes during deposition and dissolution. It is therefore reasonable
to consider that the particles in Fig. 3 and 4 were not lithium metal
but decomposition products of the electrolyte solution. This, in turn,
implies that metallic lithium was deposited uniformly underneath
the precipitate layer, which suppressed dendritic deposition of
peared with many noisy lines across the image. After 0.3 C cmϪ2
,
the surface became very rough, and many agglomerates appeared on
the surface ͑Fig. 2d͒. These morphology changes were similar
to those obtained in lithium bis͑perfluoroethylsulfonyl͒imide
͑LiBETI͒/PC at room temperature in our previous reports.19,20 It is
Downloaded on 2014-05-07 to IP 131.215.25.218 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A1580
Journal of The Electrochemical Society, 149 ͑12͒ A1578-A1583 ͑2002͒
Figure 3. AFM images of the surface morphology on a nickel substrate obtained after ͑a͒ 0.09 and ͑b, c͒ 0.3 C cmϪ2 deposition of lithium metal in 1 M
LiClO4 /PC containing 5 wt % FEC. Lithium was deposited at 0.5 mA cmϪ2. ͑a, b͒ 6.5 ϫ 6.5 ␮m; ͑c͒ 1 ϫ 1 ␮m.
lithium metal and resulted in the improvement in cycling efficiency
shown in Fig. 1.
decomposes on a graphite electrode at a higher potential ͑about 1.1
V͒ than PC is cointercalated within the graphite and decomposes
͑about 0.9 V͒.18 Thus, it seems that FEC accelerates the rate of the
electrolyte decomposition and provides a surface morphology that
can suppress dendrite formation.
The morphologies of the surface film in PC ϩ FEC in Fig. 4 and
5 are very similar to those obtained in LiBETI/PC at elevated tem-
peratures of 60 and 80°C in our previous studies,19,20 at which tem-
peratures the cycling efficiencies were greatly improved. In the pre-
vious studies, we concluded that the dense and uniform structure of
the surface film is given by rapid decomposition of the electrolyte
solution at the elevated temperatures, and such a uniform surface
morphology is one of the important factors for the suppression of
dendrite formation. In addition, we reported that FEC reductively
Figure 5 shows the morphological changes observed in 1 M
LiClO4 /PC ϩ 5 wt % VC. After deposition of 0.09 C cmϪ2 ͑Fig.
5a͒, a small amount of deposits appeared on the surface. The
grooves on the substrate became indistinct, and the whole surface
seemed to be covered with a very thin film of the decomposition
products. However, the image was very clear and sharp when com-
pared with that in the additive-free solution ͑e.g., Fig. 2b͒ so that the
film should be solid and hard. After deposition of 0.15 C cmϪ2 ͑Fig.
5b͒ the size of the deposits increased to 200-500 nm diam, and they
grew up to large agglomerates after deposition of 0.3 C cmϪ2 ͑Fig.
5c͒. These particles are most probably lithium metal and become
nuclei for lithium dendrites during repeated deposition/dissolution
cycles, as was the case in the additive-free solution.
AFM images obtained in 1 M LiClO4 /PC ϩ 5 wt % ES are
shown in Fig. 6. After deposition of 0.09 C cmϪ2 ͑Fig. 6a͒, the
whole surface was covered with a layer of precipitates. This surface
morphology was similar to that in PC ϩ FEC in Fig. 3a. However,
the precipitate layer became rough and inhomogeneous after depo-
sition of 0.3 C cmϪ2 ͑Fig. 6b͒. In some parts, large deposits, which
were probably lithium metal, were observed as shown in Fig. 6c.
The poor cycling efficiency in PC ϩ ES is thus attributable to den-
drite formation due to the rough and inhomogeneous structure of the
precipitate layer.
CV coupled with AFM observation.—The decomposition pro-
cesses of the three additives were investigated more closely by CV
coupled with in situ AFM observation. Figure 7 shows CVs of
nickel electrodes in 1 M LiClO4 /PC with and without the additives.
All voltammograms showed three peaks at about 1.1, 0.6, and 0.2 V
on the first cathodic sweep, which can be assigned to reductive
decomposition of electrolyte solutions to form surface films as dis-
cussed in a previous study using 1 M LiBETI/PC.22 In the second
cycle in each solution, the cathodic current decreased significantly,
and the peak at 1.1 V disappeared. The current further decreased
with cycling number, which indicates that the surface of the nickel
electrode was passivated by the surface film formed via the decom-
position of the electrolyte solution. We reported that the additives,
FEC, VC, and ES, reductively decompose at 1.1, 1.3, and 1.0 V,
Figure 4. AFM images of the surface morphology on a nickel substrate
obtained after ͑a, b͒ dissolution of lithium to 1.5 V and ͑c, d͒ the following
deposition of 0.3 C cmϪ2 in 1 M LiClO4 /PC containing 5 wt % FEC. The
current density was 0.5 mA cmϪ2. ͑a, c͒ 6.5 ϫ 6.5 ␮m; ͑b, d͒ 1 ϫ 1 ␮m.
Downloaded on 2014-05-07 to IP 131.215.25.218 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, 149 ͑12͒ A1578-A1583 ͑2002͒
A1581
Figure 5. AFM images (7 ϫ 7 ␮m͒ of the surface morphology on a nickel substrate obtained after ͑a͒ 0.09, ͑b͒ 0.15, and ͑c͒ 0.3 C cmϪ2 deposition of lithium
metal in 1 M LiClO4 /PC containing 5 wt % VC. Lithium was deposited at 0.5 mA cmϪ2
.
respectively, on a graphite negative electrode.18 However, these
peaks were not clearly observed on the nickel electrodes in the
present study.
Particle-like deposits are also seen in the image, but their shapes are
less distinct. This implies that the particles are also made of a soft
material. In contrast, a very clear and sharp image was obtained in
PC ϩ VC ͑Fig. 8c͒. The shapes of particle-like deposits were also
clearly observed. These facts mean that the surface film and the
particle-like deposits formed in PC ϩ VC are much more solid than
those formed in the other solutions. Particle-like deposits were
hardly seen in PC ϩ ES; however, the grooves were unclear and
hence a thin film was present on the surface. From these results it
can be concluded that the properties of surface films are rather dif-
ferent depending on the kind of additives, which affects the mor-
phology of deposited lithium.
After the fifth voltammogram in Fig. 7 was obtained, the surface
of the nickel electrode was observed by in situ AFM in each elec-
trolyte solution. The results are shown in Fig. 8. Figure 8a shows an
image obtained in the additive-free 1 M LiClO4 /PC. The image
became indistinct when compared with the original surface before
CV ͑e.g., Fig. 2a͒. This may be due to the presence of a thin and soft
film, which is slightly deformed by the microcantilever during scan-
ning. It is clear that the surface film was formed by reductive de-
composition of the electrolyte solution during the CV measure-
ments. The grooves made from polishing were still visible; hence,
the amount of deposition was not so large. In addition to the thin
layer, particle-like deposits were sparsely observed on the surface.
These particles were not lithium metal but decomposition products
of the electrolyte solution, because observation was carried out at
2.5 V.
AC impedance measurements of deposited lithium.—Lithium
metal was deposited on nickel substrates at a constant potential of
Ϫ0.1 V for 300 s in 1 M LiClO4 /PC with and without the additives,
and ac impedances were measured to compare the electrochemical
properties of respective surface films. The results are shown in Fig.
9. Figure 9a compares the impedances over the entire frequency
region. The resistances of the electrolyte solutions (ϳ60 ⍀) were
In PC ϩ FEC ͑Fig. 8b͒, the image obtained was unclear, and this
again indicates the presence of a thin and soft film on the surface.
Figure 6. AFM images (6.5 ϫ 6.5 ␮m͒ of the surface morphology on a nickel substrate obtained after ͑a͒ 0.09 and ͑b, c͒ 0.3 C cmϪ2 deposition of lithium
metal in 1 M LiClO4 /PC containing 5 wt % ES. Image ͑b͒ and ͑c͒ were of different regions. Lithium was deposited at 0.5 mA cmϪ2
.
Downloaded on 2014-05-07 to IP 131.215.25.218 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

A1582
Journal of The Electrochemical Society, 149 ͑12͒ A1578-A1583 ͑2002͒
Figure 7. CVs on nickel substrates in 1 M LiClO4 /PC ͑a͒ without additive,
and with 5 wt % ͑b͒ FEC, ͑c͒ VC, and ͑d͒ ES. The scan rate was 5 mV sϪ1
.
Figure 9. Cole-Cole plots for lithium deposited on nickel substrates in 1 M
LiClO4 /PC without additive, and with 5 wt % FEC, VC, and ES. Lithium
deposition was carried out at Ϫ0.1 V for 300 s.
subtracted for clarity. Two arcs are seen in each spectrum. The
smaller arc in the high-frequency region appeared at 1.0 V and has
been already assigned to the impedance of the surface film in a
previous study.22 The larger arc in the low-frequency region ap-
peared only after deposition of bulk lithium, and disappeared after
dissolution ͑not shown͒. The arc in the low-frequency region is thus
attributable to the charge-transfer reaction, though further investiga-
tion is necessary. In the present study, we compared only the arcs in
the high-frequency region to discuss the electrochemical properties
of the surface films. Figure 9b shows magnified spectra in the high-
frequency region. In PC ϩ FEC, the resistance of the surface film
was slightly lower than that in the additive-free solution. The char-
acteristic frequency of the arc was 3.98 kHz, which was equal to the
value in the additive-free solution, which implies that the surface
films formed in these two electrolyte solutions have similar electro-
chemical properties. In contrast, the resistances in the other two
solutions, in particular in PC ϩ VC, were higher than that in the
additive-free solution. The order in the conductivity of the surface
Figure 8. AFM images (7 ϫ 7 ␮m͒ of the surface morphologies on nickel
substrates obtained after five cycles of CV in 1 M LiClO4 /PC ͑a͒ without
additive, and with 5 wt % ͑b͒ FEC, ͑c͒ VC, and ͑d͒ ES.
Ͼ
Ͼ
Ͼ
PC ϩ ES PC ϩ VC. This or-
film is thus PC ϩ FEC
PC
Downloaded on 2014-05-07 to IP 131.215.25.218 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, 149 ͑12͒ A1578-A1583 ͑2002͒
A1583
der is consistent with the order for the cycling efficiency in Fig. 1;
that is, the lower the resistance of the surface film, the higher the
cycling efficiency obtained. A high resistance of the surface film is
liable to cause a nonuniform current distribution, which accelerates
the formation of dendritic lithium. Consequently, a low resistance is
one of the important properties required to obtain an effective sur-
face film on lithium metal.
fore VC and ES work as effective additives for graphite negative
electrodes.
Kyoto University assisted in meeting the publication costs of this article.
References
1. J. Yamaki and S. Tobishima, in Handbook of Battery Materials, J. O. Besenhard,
Editor, Part III-3, Wiley-VCH, Weinheim ͑1999͒.
2. E. Peled, J. Electrochem. Soc., 126, 2047 ͑1979͒.
Conclusions
3. J. G. Thevenin and R. H. Muller, J. Electrochem. Soc., 134, 273 ͑1987͒.
4. K. Kanamura, S. Shiraishi, and Z. Takehara, J. Electrochem. Soc., 143, 2187
͑1996͒.
The effects of FEC, VC, and ES as additives on the cycling
efficiency for lithium deposition and dissolution were investigated.
Although all the additives were reported to be effective for graphite
negative electrodes, only FEC improved the cycling efficiency for
lithium metal and the other two additives significantly lowered the
efficiency. These results suggested that the properties required for
the surface film on lithium are different from that on graphite.
In situ AFM observation revealed that deposited lithium was
covered with a uniform, closely packed layer of particle-like depos-
its of 100-150 nm diam in PC ϩ FEC. The surface film seemed to
be more solid in PC ϩ VC and inhomogeneous in PC ϩ ES. Large
volume changes are accompanied by lithium deposition and disso-
lution, which requires that the surface film should be elastic ͑or soft͒
and be self-repairable when being damaged. In addition, a nonuni-
form current distribution ͑local current concentration͒ easily results
in dendrite formation, which requires that the surface film should be
uniform and its resistance should be as low as possible. When VC
was added, the surface film was very thin and solid, and its resis-
tance was high. In PC ϩ ES, the surface film was inhomogeneous,
and its resistance was relatively high. PC ϩ FEC gave a surface
film that satisfies all the requirements, and therefore only FEC was
effective as an additive on lithium metal. In contrast to lithium
metal, graphite negative electrodes exhibit much smaller volume
changes during charging and discharging, and are free from dendrite
lithium deposition except when being overcharged. Softness and low
resistivity are not severely required for the surface film, and there-
5. T. Osaka, T. Momma, T. Tajima, and Y. Matsumoto, J. Electrochem. Soc., 142,
1057 ͑1995͒.
6. M. Ishikawa, M. Morita, and Y. Matsuda, J. Power Sources, 68, 501 ͑1997͒.
7. R. D. Rauh and S. B. Brummer, Electrochim. Acta, 22, 75 ͑1977͒.
8. K. Naoi, M. Mori, M. Inoue, T. Wakabayashi, and K. Yamauchi, J. Electrochem.
Soc., 147, 813 ͑2000͒.
9. M. Mori, Y. Naruoka, and K. Naoi, J. Electrochem. Soc., 145, 2340 ͑1998͒.
10. T. Hirai, I. Yoshimatsu, and J. Yamaki, J. Electrochem. Soc., 141, 2300 ͑1994͒.
11. T. Fujieda, Y. Xia, C. Wang, and T. Sakai, Extended Abstracts of the 41st Battery
Symposium in Japan, Abstract no. 1E02 ͑2000͒.
12. M. Morita, S. Aoki, and Y. Matsuda, Electrochim. Acta, 37, 119 ͑1992͒.
13. M. Ito, H. Saito, and K. Usami, Abstract 316, The Electrochemical Society and the
Electrochemical Society of Japan Meeting Abstracts, Vol. 99-2, Honolulu, HI, Oct
17-22, 1999.
14. C. Wang, H. Nakamura, H. Noguchi, and M. Yoshio, Electrochemistry (Tokyo,
Japan), 65, 941 ͑1997͒.
15. R. McMillan, H. Slegr, Z. X. Shu, and W. Wang, J. Power Sources, 81-82, 20
͑1999͒.
16. P. Biensan, J. M. Bodet, F. Perton, M. Broussely, C. Jehoulet, S. Barusseau, S.
Herreyre, and B. Simon, Abstract 286, The 10th International Meeting on Lithium
Batteries, Como, Italy, May 28-June 2, 2000.
17. G. H. Wrodnigg, J. O. Besenhard, and M. Winter, J. Electrochem. Soc., 146, 470
͑1999͒.
18. S.-K. Jeong, M. Inaba, R. Mogi, Y. Iriyama, T. Abe, and Z. Ogumi, Langmuir, 17,
8281 ͑2001͒.
19. R. Mogi, M. Inaba, T. Abe, and Z. Ogumi, J. Power Sources, 97-98, 265 ͑2001͒.
20. R. Mogi, M. Inaba, Y. Iriyama, T. Abe, and Z. Ogumi, J. Electrochem. Soc., 149,
A385 ͑2002͒.
21. S. Shiraishi and K. Kanamura, Langmuir, 14, 7082 ͑1998͒.
22. R. Mogi, M. Inaba, Y. Iriyama, T. Abe, and Z. Ogumi, J. Power Sources, 108, 163
͑2002͒.
Downloaded on 2014-05-07 to IP 131.215.25.218 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).