Documente Academic
Documente Profesional
Documente Cultură
BATTERIES
Science and Technology
Gholam-Abbas Nazri
Gianfranco Pistoia
[Eds.)
LiNi02 Structure
LITHIUM BATTERIES
Science and Technology
LITHIUM BATTERIES
Science and Technology
Edited by
Gholam-Abbas Nazri
General Motors Research & Development Center
Warren, Michigan, U.S.A.
Gianfranco Pistoia
Consultant Rome,
Italy
Springer
Editors
Gholam-Abbas Nazri Gianfranco Pistoia
General Motors R&D and Planning Center Via G. Scalia, 10
MC 480-102-RCEL 00136 Rome
30500 Mound Road Italy
Warren, MI 48090-9055
USA
springer.com
ents
Contributors XV
xvii
damentals
ode Materials
athode Materials
nds in Cathode Materials for Rechargeable Batteries
Pasquali, S. Passerini and G. Pistoia
11.1. Introduction 315
11.2. Methods of Synthesis 317
11.3. Effect of Particle Size and Morphology on Cathode Behavior 321
11.4. Natural vs. Artificial Cathodic Films 324
11.5. Specific Cathode Materials 327
11.5.1. LiNi02 and LiCo02 327
11.5.2. Manganese Spinels 335
11.5.3. Layered LixMn02 and Similar Cathodes 341
11.5.4. 3 V Cathodes 344
11.5.5. A Special Case: LiFeP04 347
11.5.6. Sulphur and Organic or Inorganic Sulphides 350
11.5.7. High Voltage Materials 351
11.6. Concluding Remarks 352
References 353
lectrolytes
uid Electrolytes: Some Theoretical and Practical Aspects
Nazri 509
17.1. Introduction to Solvents 511
17.2. Coulombic Forces Between Ions and Dipolar Solvents 511
17.2.1. Ion-Dipole Forces 513
17.2.2. Dipole-Dipole Interactions 513
17.2.3. Dipole-Induced Dipole Interactions 514
17.2.4. Instantaneous Dipole-Induced Dipole Interaction Forces 515
17.2.5. Hydrogen Bonding
17.3. Electron Pair Donor-Electron Pair Acceptor Interactions 516
(EPD/EPA Interactions) 517
17.4. Solvation 518
17.4.1. Solvation Number 518
17.4.2. Selective or Preferred Solvation 519
17.5. The Basic Requirements of Electrolytes for Lithium Batteries 520
17.5.1. Electrolyte Conductivity 525
17.5.2. The Electrochemical Stability of Organic Electrolytes 527
17.6. Conclusions 527
References
rent Issues and Market Trends of Li-Ion Batteries for Consumer Applications
acArthur
23. I. Introduction 70 I
23.2. Lithium-Ion batteries 702
23.2.1. The Market 702
23.2.2. Different Technologies 704
23.2.3. Targets: Small and Large Batteries 705
References 706
Index 707
f Contributors
Amatucci
mundsen
rbach
Besenhard
oussely
der
osnier
Dudney
ugeon
Huggins
ikawa
err
Kim
cher
uelle
Leising
cArthur
nthiram
Marschilok
squelier
Moeller
orcrette
ce:
INTRODUCTION
50
Smaller
00
50 1-o
..Q...).
Lithium-ion
00
C)
oO
50
Ni/Cd Ni/MH
Lead-acid
0
0 100 200 300 400
Volumetric Energy Density (Wh/L)
s
Ti ............................
............................
s
Li 00
s
............................ TI
1.2. Illustration of lithium insertion/extraction into/from the layered TiS 2 during the
ge/charge process.
V = f.lu(c) - f.lu(a)
(1.1)
oc F
(1.2)
Chapter 1
LUMO JA.a(Li)
Jlc(Li)
eVoc
Cathode Anode
Electrolyte/separator
1.3. Schematic energy diagram of a lithium cell at open circuit. HOMO and LUMO
spectively, to the highest occupied molecular orbital and lowest unoccupied molecular
n the electrolyte.
ertion/extraction in order to provide a good cycle life for the cell. This
plies that the insertion compound LixMyXz should have good structural
bility without breaking any M-X bonds.
e insertion compound should support mixed conduction. It should have
d electronic conductivity cre and lithium-ion conductivity cru to
imize polarization losses during the discharge/charge process and
reby to support a high current density and power density. This
ends on the crystal structure, arrangement of the MXn polyhedra,
metry and interconnection of the lithium sites, nature and electronic
figuration of the Mn+ ion, and the relative positions of the Mn+ and xn
rgies.
2 13
since the Co + + redox couple lies within the S2-:3p band as seen in
1.4. Accordingly, several transition metal oxide hosts crystallizing in
ty of structures (2-dimensional layered and 3-dimensional framework
res) have been pursued during the past two decades and Figure 1.5
res the electrochemical potential ranges of some selected lithium
on compounds versus metallic lithium. Cell voltages of as high as 5 V
metallic lithium have been achieved with oxide insertion compounds
n in Fifiure 1.5. The readers may refer to some of the review articles
6
ooks s- in the field to get a more detailed account of the various
insertion compounds that have been investigated over the past three
s.
E E
Co3+/2+
Co3+12+
Co4+/3+
1.4. Relative energies of metal:d (for example, Co:3d) and nonmetal:p in (a) a sulfide
an oxide.
>
-s
g-
Q)
C)
ctS
LixGriaphite ILixCoke
0- - Li Metal
-LiAl
Insertion Compound
Cathode
Discharge
u+
Charge
Electrolyte
und LixMyXz should also satisfy several important criteria like the
e insertion host: a large degree x of lithium insertion/extraction
capacity), high electronic and ionic conductivity, good chemical and
ural stabilities (reversibility), and affordable cost. The main difference
en the anode and cathode insertion hosts is that the former should have
lithium chemical potential (f.!u(a)) in order to maximize the cell
e, which implies that the Mn+ ion in the LixMyXz anode should have
xidation state.
Apart from the various intrinsic characteristics of the insertion
ode materials discussed, several other criteria are important in
ing a good lithium-ion cell that can offer high performance with long
life. The electrolyte should have high lithium-ion conductivity, but
be an electronic insulator in order to avoid internal short-circuiting.
h ionic conductivity in the electrolyte is essential to minimize IR drop
mic polarization and achieve good rate capability. With a given
lyte, the IR drop due to electrolyte resistance can be reduced and the
pability can be improved by having a higher electrode interfacial area
in separators. The electrolyte should also have good chemical stability
ould not undergo any direct reaction with the electrodes. Additionally,
gineering involved in cell design and fabrication plays a critical role in
erall cell performance. 12 For example, high electronic conductivity
hium-ion diffusion rate in the electrodes are essential to minimize cell
zations and the electronic conductivity of the electrodes can be
ved by adding electronically conducting additives such as carbon.
ver, the amount of additive should be minimized to avoid any undue
ce in gravimetric or volumetric capacity. Finally, cell safety,
nmental factors, and raw material and fabrication costs are additional
tant considerations in both materials selection and cell design.
o Li
•co
c/2 Go
Layered LiCo02
c a
b-
3
0 50 100 150 200
Capacity (mAh/g)
1.8. Typical discharge curves of (a) layered LiCo02, (b) layered LiNi0.85Co0.850z, and
el LiMn204.
23
m). However, it has been found recently that the cyclability of the
o02 cathodes could be improved significantly down to (l-x)::::0.3,
g reversible capacities of close to 200 mAh/g, 24by29 modifying its
e with other oxides such as Ah03, Zr02, and Ti02. - For example,
1.9 compares the cyclability of Ah03-modified LiCo02 cathode
29 with
unmodified LiCo02 cathode to various cut-off voltages. As can be
he surface modified samples exhibit good capacity retention to higher
f voltages with practical capacities of close to 200 mAh/g. Cho et
attributed the improved cyclability from ex-situ X-ray diffraction data
uppression of both the changes in the c axis with (1-x) and the
onal to monoclinic phase transition arising from an ordering of the
m ions around (l-x)=0.5. However, in-situ X-ray diffraction data
te that the surface modified samples also exhibit changes in c axis
1-x) and the hexagonal to monoclinic 30 phase transition around (1-
just like the uncoated Li 1_xCo02. Also, Transmission electron
copic (TEM) studies show that the coating oxides such as29Ah03 30 and
re present as loose nanoparticles on the surface of LiCo02. ' These
gs suggest that the capacity fade commonly encountered by the
2 cathodes (uncoated) on cycling to (l-x)<0.5 may not be simply due
hexagonal to monoclinic transition at (l-x)=0.5 and some other factors
ay a role.
250
- 200
(f)
<E
-u
(e)
150 (d)
(b)
(a)
Bu 100
l
J
50
0
0 20 40 60 80 100
Cycle number
The lack of a clear understanding of the factors that limit the practical
ty of the LiCo02 (uncoated) cathode to 140 mAhlg is partly due to the
22 31 34
hat most of the investigations in the literature • - have focused
on the structural characterization (both in-situ and ex-situ) of the
chemically charged Lh-xCo02 samples. No chemical characterization
uch as the variations of the oxidation state of Co and the oxygen
t with lithium content were available since it is difficult to chemically
terize the electrochemically charged samples as they are contaminated
arbon, binder, and electrolyte. To overcome these difficulties and to
p a better understanding of the factors that control the practical
ties, recent investigation by our group35 -38 has focused on both the
cal and structural characterizations of the Li 1_xCo02 samples that were
ed by chemical delithiation. The chemical delithiation reactions were
d out by stirring the LiCo02 powder under argon atmosphere with an
itrile solution of an oxidizing agent such as N02PF6 (or N02BF4) that
39
reduction potential of around 5 V versus metallic lithium and is
enough to extract all the lithium from LiCo02:
10 20 30 40 50 60 70 80
Cu Ka 2e (degree)
14.4
'$
CJ
14.8
14.0
• 000
--.
0
.. • •
0 0
13.6 0
IT1I] 3
....... •
2.84 0 0 0 oo
OJ
2.82 •
ca
2.80 •
0.0 0.2 0.4 0.6 0.8 1.0
Lithium content, (1-x)
•u
•
0
OCo
c c eo
0•
a a
03 P3
.12. Crystal structures of 03-type LiCo02, P3-type Co02, and 01-type Co02 viewed
e (100) plane.
A A
c BA AC B
c
. BA
AC
ce c
c B
ce
B A
A
B
A :!c
P3 01
.13. Schematics of the sliding of some Co0 2 sheets leading to the transformation of
tructure to P3 and 01 structures. The letters refer to oxygen stacking sequence along
s.
Chapter I
1.98
-
c ••• o•
0 o • •o
Q)
§ • (a)
u 1.92 • 0
ffi 1.86 00
0>
1.80 0
0 0
1.74
0
3.8 A (b)
AAA.a
Q)
1.14. Variations of the (a) oxidation state of Co and (Ni 0_ 85Co0.15) and (b) oxygen
with lithium content (1-x) in Li 1.xCo02 and in Li1.xNio.8sCoo. 150z.
Figure 1.14 shows the variations of the oxidation state of Co and the
n content with lithium content for the chemically delithiated Li 1.xCo02
38
es. The data in Figure 1.14 were obtained by a wet-chemical analysis
40
ing iodometric titration and the application of charge neutrality
ple. The oxidation state of cobalt increases linearly with (1-x) and the
n content remains around 2 for 0.5s(1-x)sl.O in Lh.xCo02.s- For Os(l
, the oxidation state of cobalt remains nearly constant and the charge
nsation is achieved by a loss of oxygen from the lattice with a 8=0.28
e end-member Co02. 8. The loss of oxygen from the lattice signals
a cal instability for (l-x)<0.5 in Li 1.xCo02.0, which may be the
factor
g its reversible capacity. The increased reversible capacity for the
e modified samples in Figure 1.9 is due to a more stable electrode
lyte interface and the consequent improvement in the chemical
ty. Interestingly, the lithium content (1-x) at which the oxygen loss
to occur coincides with the lithium content at which the new phase
gins to form (Figures 1.10 and 1.11). The results indicate that the P3
is oxygen deficient with a shorter 0-0 distance across the van der
gap between the Co02 sheets, signaling an increased 0-0 interaction.
l Aspects: An Overview 19
. Layered LiNi02
•• •m 0 Layer
M Layer
0 Layer
OJ 0
-
Li Layer
oo
b_L
cP
•• •
1.15. Schematic representation of the paths for the migration of M 3+ ions in layered
2 (M = Mn, Co and Ni). The open squares and the dotted squares refer, respectively,
m ion vacancies and tetrahedral sites. T 1 and T2 refer to tetrahedral sites at (0, 0,
nd (0, 0, 0.375) respectively.
1.1. Crystal field stabilization energies (CFSE) and octahedral site stabilization
s (OSSE) of some 3d transition metal ions.
3 0
t ge
2 g
(HS)
t3 el
2g g
(HS)
3 2
t2 geg
(LS)
t6 g
2g
eo (LS)
6 I
t2 geg
10 20 30 40 50 60 70 80
Cu Ka 29 (degree)
14.8
.--,14.4 A A&
·::su 14.0 AA
Ae,.
A A
13.6 6. 4.
A 6. [I;] '
A
2.86 A
· 2.84 6.
ro 6. &A
2.82 J!ii.. A
6.
sidering the qualitative band diagrams for the Li 1_xCo02 and Li 1_xNi02
s (Figure 1.18). With a low-spin Co3+:3d6 configuration, the t2g band
pletely filled and the eg band is empty (t;
e ) in LiCo02• As
8
is extracted from LiCo02, the Co + ions are oxidized to Co4+ by a
3
E E
M n3+1 +:e8
Co3+14+:c8
Mn +14+:tz8
3
Ni 3+14+:e
Co3+14+:t z8
8
Ni 3+14+:tz8
o2-:2p o2-:2p o2-:2p
N(E) N(E) N(E)
1.18. Comparison of the energy diagrams ofLi 05Co02, Li0_ 5Ni02, and Li0_5Mn 204•
The qualitative band diagrams shown in Figure 1.18 and the observed
n loss behavior are consistent with (i) the spectroscopic data that holes
roduced into the 02":2p band rather than into the Co:3d band in Li1.
, but into the Ni:eg band in Li1-xNiOz and Lii-xNio.ssCoo.IsOz during
57 59
chemical charging, - and (ii) the theoretical calculations that the
e charge on oxygen decreases with decreasing lithium content in Li1 _
60 61 Furthermore, the introduction of holes into the 0 2-:2p band
. '
ates the 0-0 interaction across the van der Waals gap between the
heets, which could in tum be the driving force for the formation of the
03' phases with smaller c parameters than the 03 phase (Figures 1.11
17). In fact, the P3 phase is generally believed to be inaccessible since
ide ions lie one above the other along the c axis (Figure 1.12), and the
ion of the P3 phase only in the case of Co02.8 could very well be due
2
presence of significant amount of holes in the 0 -:2p band and
erable amount of oxygen vacancies.
63 65
nge reactions of NaM02• - Also, a partial substitution of Cr orAl
n has been found to access the 03-type LiMn02 by direct ceramic
53 66 68
ures. • - Unfortunately, the 03-type LiMn02 transforms to spinel
hases during cycling due to the low OSSE of Mn 3+ (Table 1.1) and the
3
quent migration of the Mn + ions from the Mn plane to the Li plane.
rly, the orthorhombic LiMn02 that is obtained by conventional
69
ures also transforms to spinel-like phase, leading to poor cyclability.
3-type LiFe02 also exhibits poor electrochemical cycling properties
3 5
structural instabilities since the high-spin Fe +:3d with an OSSE
of zero (Table 1.1) can readily migrate from the octahedral to
dral sites.
0 Li (Sa site)
0 Mn (16d site)
• 0 (32e site)
m can also be inserted in to the empty 16c octahedral sites of the spinel
work to give the lithiated spinel Li2 [M2]04. However, an electrostatic
ion between the u+ ions in the 8a tetrahedral and 16c octahedral sites,
share common faces, causes a displacement of the tetrahedral Lt ions
e neighboring empty 16c sites to give an ordered rock salt structure
a cation distribution of {Lizh 6c[Mzh 6ct04. Thus theoretically two
m ions per LiM204 formula unit could be reversibly inserted/extracted.
the edge shared M06 octahedral arrangement with direct M-M
ction, as in the layered LiM02 oxides, provides good electrical
ctivity O'e, the interconnected interstitial (lithium) sites in the three
sional spinel framework provide good lithium-ion conductivity cru. As
lt, the spinel LiM204 oxides have also become attractive candidates
ey are discussed below.
2
s the top of the 0 -:2p band (Figure 1.18).
Another strategy that has been pursued to improve the cyclability of
Mn204 spinel is the modification or coating of its surface with other
77
such as LiCo02, V205, Alz03, or Mg0 '95 -97 with an aim to suppress
solution of manganese in contact with the electrolyte. The modified
es exhibit improved cyclability at both ambient and elevated
ratures. However, firing at higher temperatures (-800°C) after coating
to much better capacity retention compared to firin§. at lower
9
ratures (-400°C) in contrast to the coated LiCo02 samples discussed
that requires firing only at low temperatures (-300°C). This suggests
a diffusion of the coating or modifying oxides MyOz into LiMn 204 is
al to achieve good cyclability and (ii) the improved cyclability need
simply due to the protection of the cathode surface from the
l Aspects: An Overview 29
120 •
·
0)
\
-
110
..c 100
•\ (b)
--
<(
E 90
(c)
(d)
'(3 >. 80
<tl
c.. 70
<tl
()
60
(a)
50
0 10 20 30 40 50
Number of Cycles
4.4
-
C/5
C/10
Q)
C)
ctS 3.6
0
>
4.4
4.0
3.6
0 20 40 60 80 100 120
Capacity (mAh/g)
1.22. Discharge profiles of LiMn 2.2yLiyNiy04 at various Crates (C/10, C/5, C/2, IC,
4C).
0.08 •
•
0.06 • •
o<(
'-"
0.04
co
<l
•
•
0.02
0.00 ....
••r=--.---,.--.---,-----,.---r---.--.--
0 10 20 30 40
Capacity Loss (%)
1.23. Relationship between % capacity loss and lattice parameter difference (Lla)
the two cubic phases in Li 1.xMn2.yMy04 and Li 1.xMn2•2yLiyMy04 (M = Li, Ti, Co and
102 103
ould also be extracted from the 8a tetrahedral sites of LiV204. •
er, LiV204 suffers from a migration of the vanadium ions during
rocesses, which leads to poor capacity retention.
Both LiCr204 and LiFe204 are not known. LiCo204 could not be
y conventional high temperature methods, but it could be accessed as
al spinel (Li)8a[Co2h6ct04 by chemically extracting 50% of lithium
20 21 104
queous acid or N02PF6 in acetonitrile medium ' from the low
17 19
ature form of LiCo02 that has a lithiated spinel structure. ' LiCo204
is metallic due to the direct Co-Co interaction and the partially filled
3 14 104
d associated with the low-spin Co + +. While the extraction of
from the 8a sites occurs around 3.9 V, the insertion of an additional
21
into the 16c sites occurs around 3.5 V. However, the system
from a huge polarization loss as indicated by a large separation
n the discharge and charge profiles, which could possibly be related
low-temperature synthesis. Attempts to make LiNiz0 4 spinel by
ally extracting 50% of lithium from LiNi02 followed by heating at
°C result in a spinel-like cubic phase, but with the Ni3+14+ ions in both
and 16d sites. Heating at T>200°C results in a disproportionation to
iNi02 and Ni0.53
Chapter 1
1.24. Crystal structure of olivine (a) LiFeP04 and (b) FeP04 consisting of Fe06
ra and P04 tetrahedra (Ref. 62). The circles refer to lithium ions.
dra share comers with the S04 tetrahedra with a Fe-0-S-0-Fe linkage
thium ions could be inserted into the interstitial voids of the
2 13
work. The higher voltage of 3.6 V observed with the Fe + + couple
2(S04)3 compared to that found in simple iron oxide is due to the
6
ive effect caused by the counter cation S +. A stronger S-0 covalent
g weakens the 7t-bond Fe-0 covalence through inductive effect,
2 13
results in a lowering of the Fe + + redox couple and an increase in the
ltage. However, a poor electronic conductivity associated with the Fe-
-Fe (X= S or P) linkages leads to poor rate capability.
Following this initial concept of using poly anions, several phosphates
been investigated in recent years. 107- 109 Among them, LiFeP04
lizing in the olivine structure with Fe0 6 octahedra and P04 tetrahedra
l Aspects: An Overview 33
20 121 122
· and LiV308 having layered structures have been shown
to t high capacity. The vanadium oxides generally exhibit good
chemical ty with better safety characteristics without releasing oxygen
from the
3 14
compared to the cobaIt or m.e keI ox1'des as th e V +ts+ and y + +
4
2
s lie well above the top of the 0 -:2p band.
1.25. Crystal structure of V 6013 consisting of single strings (I) and double ribbons (II).
ANODE HOSTS
Carbon has become the material of choice for anode in the present
tion of lithium-ion cells. 137' 138 The lightweight and low
chemical potential lying close to that of metallic lithium (Figure 1.5)
made carbon an attractive anode. It has a theoretical capacity of 372
, which corresponds to an insertion of one lithium per 6 carbon atoms
in LixC6). One of the drawbacks with the carbon anodes is the
ence of significant amount of irreversible capacity during the first
rge-charge cycle due to unwanted, irreversible side reactions with
lyte. Also, with electrolytes consisting of propylene carbonate (PC),
l graphite cannot be charged as it leads to an evolution of gas at
1 V. However, with electrolytes consisting of other solvents such as
ne carbonate (EC) and diethyl carbonate (DEC), the side reactions are
ssed and it can be cycled without much difficulty.
Carbon materials can be classified into soft carbon (graphitic carbon)
rd carbon (glassy carbon). The hard carbons are typically obtained by
mal decomposition of phenolic and epoxy resins and products from
um pitch. They have significant amount of open rnicropores, which
become closed on heating at higher temperatures, and some of them
t of single graphene sheets. Generally, hard carbons show higher
ties than graphite, which is thought to be due to the adsorption of
on both sides of the single graphene sheets, accommodation of extra
into nanometer size cavities, and storage of additional lithium at the
and surfaces. Also, hard carbons can be used with PC-based
lytes unlike graphite. However, the hard carbons show a sloping
rge profile between 0 and 1 V unlike graphite, which shows a nearly
scharge profile between 0 and 0.3 V. With an aim to improve the
Chapter 1
CONCLUSIONS
After providing a brief introduction to the basic concepts of
geable lithium batteries, this chapter presented an overview of the
s cathode and anode materials. In view of maximizing the cell voltage
nergy density, transition metal oxides offering around 4 V versus
have emerged as the choice for cathodes while the light-weight
with an electrochemical potential close to that of metallic lithium has
e the anode. With carbon anode that does not contain lithium, the
d LiM02 (M = Co and Ni), spinel LiMn204, and olivine LiFeP04
ning lithium have emerged as the leading cathode candidates.
er, only 40 to 65% of the theoretical capacity of the layered LiM02
inel LiMn204 could be practically utilized due to the chemical and
ral instabilities encountered by these cathodes at deep lithium
ion. Although recent studies indicate that 100% of the theoretical
l Aspects: An Overview 37
ty of LiFeP04 can be utilized, the added weight of the P04 units and
oor electronic conductivity lower the energy and power densities
cantly.
The future challenge is to develop simple oxide cathodes without
elements such as P in which at least one lithium ion per transition
ion could be reversibly extracted/inserted to give close to 300 mAhlg
keeping the materials cost and toxicity low. Such cathodes can double
ergy density compared to the present level. There are also possibilities
rease the capacity of anodes perhaps by focusing on amorphous
als and metal nitrides, pnictides, borides, and carbides with significant
nt character. An alternative approach is to develop cells with lithium
ning anodes and lithium-free cathodes. This strategy will allow the use
e of the already known high capacity cathodes such as the vanadium
and nanocrystalline manganese oxides that offer higher capacity with
chemical stability and safety characteristics compared to the currently
iCo02 cathode. From a safety, and cycle and shelf life points of view,
es with a lower voltage (3 to 4 V), but with an increased capacity are
ble for future applications. Such cathodes would also be attractive for
er batteries from a stability point of view.
NOWLEDGMENT
ERENCES
ium Battery Energy Storage (LIBES) Publication, Technological Research
ociation, Tokyo (1994).
. Whittingham, Science 192 (1976) 1126.
K. Aydinol, G. J. Ceder, J. Electrochem. Soc. 144 (1997) 3832.
S. Whittingham, A. J. Jacobson, Intercalation Chemistry, Academic Press, New York
2).
Mizushima, P.C. Jones, P.J. Wiseman, J.B. Goodenough, Mat. Res. Bull. 15 (1980)
.
Goodenough, K. Mizushima, T. Takeda, Jap. J. Appl. Phys. 19 (1983) 305.
. Thackeray, W.I.F. David, P.G. Bruce, J.B. Goodenough, Mat. Res. Bull. 18 (1983)
INTRODUCTION
ition metal oxide host. The valence electron of each lithium ion is
rally donated to the host where it can either shift the valence state of the
ition metal ion and/or alter the nature of the bonds between the transition
l and the oxygen ions. Simultaneously, lithium removal from the host
structurally destabilize the metal oxide structure or may lead to order
der phase transitions between lithium and vacancies once a critical
ncy concentration is reached. These phenomena often affect the voltage
cteristic or the lattice parameters of the compound in important ways.
Within the last decade, much attention has been devoted to
1
rstanding the properties of LixCo0 2, LixNi02, LixMn02, LixMn204 and
ePO/, currently among the most important candidate cathode materials
echargeable lithium batteries. Other compounds receiving attention are
ally doped variants of the Co, Ni and Mn compounds. While many
m transition metal oxides have similar crystal structures, either a layered
or one derived from the spinel structure (LixFeP04 has the olivine crystal
ture), they are often characterized by very different electrochemical
rties owing to the unique electronic structure of each transition metal.
o02, for example undergoes a concentration driven metal-insulator phase
ormation whereby the electrons over much of the lithium concentration
(x<0.75) are delocalized and exhibit metallic properties? In contrast, the
n204 compound in important lithium concentration regions is
cterized by more localized electrons with large magnetic moments
46
ng to phenomena such as charge ordering and Jahn-Teller distortions. -
anisms and rates31' ' can be predicted. The first principles study of
m transition metal oxides has also received an important boost with
opments in alloy theory during the last twenty years, most notably the
opment of the cluster expansion27' 44.4 5 which enables the statistical
anical treatment of solids with significant configurational disorder. In
alation compounds this capability is critical due to the many degrees of
om of distributing lithium ions and vacancies over the interstitial sites.
arly, the disorder in mixed compounds with multiple transition metal
28
ents can only be described properly with cluster expansion techniques.
combination of accurate DFf calculations with sophisticated statistical
anics techniques have enabled accurate calculations of phase diagrams
'
39
tercalation compounds, 18 28' voltage intercalation curves 12' 15' 17 and
m diffusion coefficients.31•36
In this chapter, we focus on a description of important electronic
rties of LixCoOz, LixNi02 and LixMn02 • Most of our information will be
ed from first principles calculations. The chapter starts with an overview
ole of Electronic Properties ... 45
2.1. Hydrogenic d-orbitals within an octahedral oxygen environment. (a) The d3z2-,2
) the dx2-y2 orbitals have lobes pointing toward the oxygen ions. (c) The dxy orbital,
is symmetrically equivalent to the dxz and dyz orbitals, has lobes pointing between
n ions.
Chapter 2
it: \
\
4p \
\
\ \ .... ---..;a;:;_g...., \
\
\
4s ---...--';-- e*
\
\
\
\
\ \
g \ \
.
\ \ \ \
\ \ \ \
\ \ ' \ \ \
\ \ ,' \
\
\
\
\
\
\
>' \ \ \ \
-:.' g
\ _,,\ =
\ \\
3d \\\
\ ,,
,,,
,,
_ \ \ \
\
\
\
\
\
\
b
,,
\ \ \
\ \ ':==l:i§u= ---,",',
\ \ \ 2p
:-
\ \
\ \
'
,','
\ \
\ \ b
alg '
''
\ \
\ ''--"'-- '
''
\
\
''
\
' Oxygen
Metal '·===='
2.2. Schematic illustration of the bonding and antibonding levels that arise for a
ion metal ion in an oxygen octahedral environment due to the hybridization between the
e electronic states of the transition metal ion with those of the oxygen ions.
imental input involves solving the many body Schrodinger equation for
50 • 51
olid. The time-independent Schrodinger equation (in this section, all
ions are expressed in atomic units) is an eigenvalue equation
(2.1)
e H is the Hamiltonian operator for the solid, '¥ is the many body wave
ion describing Ne electrons and E is the total energy of the solid. Within
Born-Oppenheimer approximation, the electrons are assumed to
ntaneously adjust their state to any change in the positions of the nuclei.
coordinates of the nuclei Rn then serve only as parameters in the
dinger equation and the Hamiltonian can be written as
H = T +Vee+ L v(;:;)+ L L
-ZnZn:_ I
j n m<n
Rm -Rn
(2.3)
(2.4)
the rj refer to the positions of the electrons. The sums in Eqs. (2.2),
Chapter 2
and (2.4) are over all the electrons in the solid. The third term in Eq.
describes the coulomb interaction between the electrons and the nuclei
e solid and the last term is the Ewald energy arising from the coulomb
action between the nuclei having charge Zm· The Ewald energy is simply
ditive term and does not affect the electronic wave function \f. It can be
ped when solving the eigenvalue equation for the electronic ground state
nly needs to be added to the electronic energy to obtain the total energy
e crystal.
For solids of practical interest, solving the many-body Schrodinger
ion is intractable. One approach to solving Eq. (2.1) is with density
ional theory (DFT).41-43 As was shown by Hohenberg and Kohn (and
41 52
ded by Levy), ' the ground state properties of a crystal are uniquely
mined by the electron density p(r) (which is determined by the many-
wave function \f). The fundamental theorems of density functional
y state that the ground state energy of a solid is a functional of the
on density41
(2.6)
versal functional and v(r) the coulomb potential due to the nuclei of the
The lJ' in Eq. (2.6) is the electronic wave function that minimizes
+ Vee I' ') subject to the constraint that lJ' produces the electron density
ccording to DFT, if the functional F[p] were known, the ground state
y of the solid with external potential v(r) is obtained by variationally
mizing the functional E[p] with respect to p41•42 The minimization
eds over the set of all densities p that can be obtained with
mmetric wave functions.5 2
ole of Electronic Properties ... 51
(2.7)
(2.10)
cludes the difference between the kinetic energy of the actual system of
cting electrons with density p and the kinetic energy of a system of
endent electrons with the same density. This difference, though, is
42
ally expected to be small and Exc primarily accounts for a correction to
arising from the correlations between electrons. Since the electron
ty corresponds to a probability distribution and not a charge density in
Chapter 2
vj(r)dr
1
e '¥0 and the If/ 1 are such that they produce the electron density p.
ole of Electronic Properties ... 53
e it is possible that more than one 'I'n could produce the same density, the
cular 'I'n appearing in Eq. (2.11) is the one that minimizes ('PviTI'Pv)
xed p.
With the 'I'n and p expressed in terms of the independent orbitals
Kohn and Sham53 variationally minimized the energy functional
(2.12)
(2.13)
Veff
(- )r -V
Jipr (_rr')'l d-'
+
Vxc .
(-) (2.14)
( -)
r
+ T T
An equation of the form of Eq. (2.13), exists for each electron in the
The Kohn-Sham procedure replaces the many body eigenvalue equation,
(2.1) and (2.2), with a set of independent-electron-like eigenvalue
ions. Despite their appearance, however, the Kohn-Sham equations are
dependent due to the dependence of Veff on the density p (Eq. (2.14))
Chapter 2
h itself is determined by all the orbitals 1f1 j. This means that the Kohn
equations are to be solved self-consistently.
(2.16)
ally guaranteed and for some solids, properties predicted within LDA
more with experiment than those predicted within GGA. Typically,
predicts lattice parameters and volumes that are slightly larger than
observed experimentally.
. Spin Polarization
ixCo02 system has shown that the pseudopotential method and LAPW
62
lations agree to within 10 meV per LixCo02 formula unit. Modem
dopotentials are determined from all electron calculations of atoms,
63 68
ng the pseudopotential method a first principles approach. -
-> 5
>.
0
...... l
O
(])
c
UJ
-5
z r F L r
(a)
> 0
-
>.
Ol
Cii
c
UJ
-5
p
r--....
-- " =--
"
--------- - - -:::::::
-10
z r F L r
(b)
2.3. The partial band structures of (a) Co0 2 and (b) LiCo02 as calculated within the
ensity approximation. The dashed line shows the Fermi level.
2.4. Charge difference plot between Li 114Co02 and Co02• Dark shade signifies regions
rge accumulation and light areas correspond to regions of charge depletion.
--c o cobalt
•Lithiu
--B
Qoxygen
--A
c
--c
--B
--A
2.5. The 03 crystal of LiCo02. The crystallographic plane on which the electronic
density difference is viewed in Figure 2.4 corresponds to the shaded plane in this figure.
Chapter 2
cter. This causes the cr bond between Co and oxygen to become less
ent and more ionic with increasing x. The result is a significant increase
e electron density at the oxygen sites immediately surrounding the
m ions. The net effect is that the electron transfer from lithium to the
s very local and occurs predominantly to the oxygen ionsY
The conventional picture of lithium transition metal oxides is that the
on donated to the host by lithium is transferred to the transition metal
7 8 17 18
e it shifts its valence state. First principles work within LDA ' ' • show
he electronic changes of the host with x are in fact more subtle. The
ion of lithium to the Co02 host causes a shift in the nature of the bond
een oxygen and cobalt from a covalent character at low lithium
entration to a more ionic character at high x. While the electron donated
hium is added to the t2g band, which is predominantly of Co d character,
ffect of this addition is to polarize the Co-O bond such that the charge
d the oxygen ions actually increases upon lithium insertion (it is
rtant to realize that oxygen is not in its fully ionized state of -2 in Co02).
e, while it is reasonable to talk about formal valence changes of Co upon
ertion/removal, one has to keep in mind that the rehybridization around
obalt ions occurs in a manner that ensures that the local net charge d
17
the cobalt ions remains fixed as the lithium concentration changes.
18 27
on-hole states in the semi-conducting state. ·
0- ------------:0
o::·:::::.....>:o-.-.-::::: : ------ :0
0 ...- - ·· · ··o
···o:::·:.-_·_·_·_· . ·_--:::o·········
- --- - . . .. - . ...... -.-- --
The tetrahedral site along the lithium migration path shares a face with
jacent octahedral site containing positively charged cobalt. The resulting
rostatic repulsion energetically penalizes the tetrahedral site relative to
ctahedral site. This electrostatic repulsion is screened to some extent by
egative electronic charge on the oxygen ions. Nevertheless, the screening
nishes with decreasing lithium concentration because of the
mpanying reduction of the charge on oxygen. As the Li concentration is
ed, the accompanying reduction of screening, along with the contraction
e c axis, which results in a reduction of volume of the tetrahedral site,
to a dramatic increase of the activation barrier for lithium migration.
Q Oxygen
e Lithium
• Nickel
2.7. In LixNi0 2 a long range attraction between a pair of lithium ions of different lithium
is mediated by a Jahn-Teller distortion of a Ni+ 3 octahedron.
2.8. Attraction between four lithium ions over long distances in Li,Ni0 2 is mediated by
tive Jahn-Teller distortion of a Nt 3 octahedron (legend as in Figure 2.7).
-> Q)
E........
100
0
...
>.
0 >
.... . -100
Q)
••
c
Q) -200 ••••
Q) ••••
Q )
.... .
LL
-300 ••••
••••••••••
-400
2.9. Free energy of layered 03 (solid line) and spinel (dashed line) LixCo02 as a
on of lithium concentration.
Li layer
Mn layer
lli, 0 Oxygen
'
' o Lithium s:
: : : : : : e Manganese
'
'
2.10. The layered 03 host and spinel have the same oxygen skeletal structure.
ormation from layered to spinel can occur by a redistribution of the transition metal ions
this structure. The first step of such a transformation involves migration of 114 of the
ons to tetrahedral sites (thick dashed lines) in the lithium layer.
2.11. Projection of a metal layer (transition metals are the filled circles) and the oxygen
oxygen ions are the empty circles) above it in the 03 crystal structure. Transformation to
involves migration of 114 of the metal ions to the lithium layer. This leaves a vacant
site (empty squares) within the metal layer. The migrated metal ions must migrate to a
site above a triangle of metal ions (shaded circle). In order for this rearrangement to
o spinel, the arrangement of vacancies within each metal layer must be ordered as
ated in the above figure.
+3 +2 =4
2 Mnoct --7 Mntet + Mnoct
)
m layer and sufficient Mn+3 ions are available in the metal layer to feed
harge-disproportionation reaction of Eq. (2.17).
The availability of vacancies in the lithium layer and Mn+3 ions in the
layer are requirements that have an opposite dependence on lithium
entration. Decreasing the lithium concentration increases the number of
m vacancies but it decreases the number of Mn+3 ions. A maximum in
ctahedral to tetrahedral migration rate can therefore be expected at
mediate lithium concentration, the concentration which also happens to
ide with the maximal thermodynamic driving force for the layered to
l transformation.
90 93
e with tetrahedral Mn has been observed experimentally. - Further
formation of the splayered phase to spinel occurs by thermally activated
hops within the lithium layers and therefore proceeds slower than the
ed to splayered transformation, which has no barrier.
In a battery, charging rarely stops at .x=l/2 but continues to lower
entrations. Further lithium removal from the spontaneously formed
2
ered phase would require oxidation of Mn+ ions, destabilizing the Mn
in the tetrahedral sites. Lithium removal below .x=l/2 therefore leads to
migration from tetrahedral sites back to octahedral sites. Nevertheless, a
on in a tetrahedral site can migrate to four neighboring octahedral sites,
which resides in the metal layer and three which reside in the lithium
. The higher availability of octahedral sites in the lithium layer leads to a
igration of tetrahedral Mn ions to octahedral sites in the lithium layer g
further deintercalation. Repeated cycling will enhance the migration of
ons to the lithium layers and this is observed experimentally where
ng of the material leads to an increased spinel formation.
CONCLUSIONS
KNOWLEDGMENTS
ERENCES
8.
reedan, N.P. Raju, A.S. Will, C. Morin, S.M. Shaw, J.N. Reimers, Chern. Mater. 10
98) 3058.
0.
1999) 1.
eder, M.K. Aydinol, A.F. Kohan, Camp. Mat. Sci. 8 (1997) 161.
7) 1354.
Deiss, A. Wokaun, J.L. Barras, C. Daul, P. Dufek, J. Electrochem. Soc. 144 (1997)
877.
Ceder, A.F. Kohan, M.K. Aydinol, P.D. Tepesch, A. Vander Ven, J. Amer. Cer. Soc. 81
ole of Electronic Properties ... 81
1998) 517.
Ceder, Y.M. Chiang, D.R. Sadoway, M.K. Aydinol, Y.l. Jang, B. Huang, Nature 392
1998) 694.
. Vander Yen, M.K. Aydinol, G Ceder, J. Electrochern. Soc. 145 (1998) 2149.
. Vander Yen, M.K. Aydinol, G Ceder, G Kresse, J. Hafner, Phys. Rev. B 58 (1998)
2975.
Benco, J.L. Barras, M. Atanasov, C.A. Daul, E. Deiss, Solid State lonics 112 (1998) 225.
Benco, J. L. Barras, M. Atanasov, C. Daul, E. Deiss, J. Solid State Chern. 145 (1999)
503.
S. Braithwaite, C. R. A. Catlow, J.D. Gale, J. H. Harding, Chern. Mat. 11 (1999) 1990.
1999) 4335.
0 (2000) 239.
2000) 3841.
Vander Yen, C. Marianetti, D. Morgan, G Ceder, Solid State lonics 135 (2000) 21.
2001) 144107.
M. Dunn, D. S. McClure, R. G Pearson, Crystal Field Theory, Harper and Row (1965).
W. Ashcroft, N.D. Merrnin, Solid State Physics, Saunders College Publishing (1976).
J. Singh, Planewaves, Pseudopotentials and the IAPW Method, Kluwer Academic Pub.
1994).
C. Payne, M.P. Teter, D. C. Allan, T. A. Arias, J.D. Joannopoulos, Rev. Mod. Phys. 64
1992) 1045.
2002) 1286.
.-S. Yoon, K.-B. Kim, M.-G Kim, M.-K. Lee, H.-J. Shin, J.-M. Lee, J.-S. Lee, C.-H. Yo,
esearch Society (MRS) Fall meeting, Boston (MA), Dec. 2-6, 2002.
998) 194.
2002) 41.
pter 3
M. S. Whittingham
INTRODUCTION
Synthesis is critical not only to the generation of new materials for
ation of new structures and fundamental properties, but also for the
ion of materials with the optimum electrochemical behavior for
ercial devices. The technique used is quite often very different for the
two cases. The initial synthesis of a material should provide it in a
enough state that its intrinsic behavior and properties can be
ined; it should for example be possible to accurately determine
er its structure is cubic spinel, layered or some other form. There is
tly much confusion in the literature because the synthesized material
sufficiently well-characterized. Once well characterized, the material
to be synthesized in a form suitable to be used in a battery, for
le it may be doped or coated to enhance the conductivity, its particle
nd morphology will be optimized for maximum reactivity but
um corrosivity and side-reactions.
Many battery materials are metastable phases, and therefore non
nal synthesis methods must be devised to allow kinetics to over-ride
dynamics. Hence, many soft chemistry techniques have come to the
uch as hydrothermal, ion-exchange, intercalation, etc. In addition, to
ze the formation of the desired material frequently the reactants are
mixed prior to final reaction by for example sol-gel formation or co
tation as hydroxides.
A number of critical synthetic parameters will be discussed in this
r. The temperature of formation of the electrode material determines
ect structure, for example in all the layered structures strict ordering
cations is desired. Thus, in TiS2 no titanium must reside in the lithium r
diffusion is limited. This restricts the final firing temperature to a
Chapter 3
he gas phase reaction of TiC14 with H2S. This produced a sulfide with
phology with many plates growing in three dimensions from a single l
point, which shows excellent electrochemical behavior.
The stoichiometry and ordering of the titanium is critical to the
chemical behavior of TiS2• Stoichiometric and ordered TiS2 has been
3
to exist if the temperature is kept below 600°C, and to have metallic
4
ctivity. The titanium disorder can be readily measured by attempting
ercalate weakly bonding species such as ammonia or pyridine. In
e a slight excess, :::;1%, is beneficial in that it reduces the
iveness of the sulfur without significantly impacting the cell potential
lithium diffusion coefficient. It is preferable to add this extra metal to
tial reaction medium.
ILMENITE
(+rutile)
CHLORINATION
850-950°C
PETROLEUM COKE
I
2
ENT ILE) H.,S 500-600°C
I
KROLL PROCESS
Mg metal
I
TITANIUM
SPONGE
s, l;jQQ-l;jl;jQOQ TiS2
24 Hours
I
ARC FURNACE
I
TITANIUM s gQQOQ TiS2
INGOT 1 Week
Li
...
n-Butyllithium
Electrointercalation
0
Ti Li
•
.2. 110 projections of the structural changes occurring during the lithiation of disulfide
in the presence and absence of a co-intercalating solvent such as ammonia.9
is of Battery Materials 89
Oxidation-MX 2 - Reduction
In NH3 i 00 H g -. --.-....--'
BD in Y
t
Polar solvent
z
A, , Y 11 _z., M X 2
72
2 and NaxCr02. The complete phase diagrams of these systems are
ell understood. Much less was known about the corresponding lithium
unds, and indeed the corresponding LiMn02 compound cannot be
sized at elevated temperatures.
Although ion-exchange methods had been used extensively in the
74 76
sis of solid electrolytes, in particular of beta-alumina - and beta" a
much less effort had been made to prepare lithium based cathodes
on-exchange until the 1990s. As noted above, layered LiMn02 cannot
nthesized at elevated temperatures because of its thermodynamic
lity. Thus, a number of studies have been made on ion-exchanging the
77 78
m analogs to the lithium form. Groups at Bordeaux • and St.
79
ws showed simultaneously that layered NaMn02 could be readily
ged to the layered form of LiMn02• In these experiments, the sodium
und is immersed in hot lithium salts dissolved in water-alcohol
es for periods of hours to days. Care has to be taken that the
ature is not too high, otherwise the spinel structure can be formed.
doped material has also been synthesized by the same approach of
80
reparing the sodium material followed by ion-exchange; • 81 these
81
containing materials show much enhanced cycling even though they
ally convert to the spinel phase even for LiMn0.5Co0.502•
The above layered LiMn02 phases have the same rhombohedral
re as the lithium, and other alkali phases formed by the hydrothermal
38
position of LiMn04, and other permanganates.58 However, these
hermal compounds have a lower alkali content, for example Li0.5Mn02
their initially being hydrates. A comparison for the potassium
nese oxides is given in Figure 3.4, which also distinguishes between o
high temperature regimes. Higher alkali contents can be obtained by g
out the hydrothermal reactions under very high alkali hydroxide
trations. The range of synthesis reactions and redox/intercalation
ns for the layered manganates are shown schematically in Figure 3.5.
2 85
- ion exchanged the Na 0.44Mn02 tunnel structure 86 giving
inantly the lithium analog, which shows excellent cycling with no
of structure unlike the layered manganese dioxides. These results
d convincingly that a stable manganese dioxide can be formed. The
can be improved by replacing part of the manganese by titanium, but
e capability is negatively impacted.87• 88 Pure Li0.44Mn02 was prepared
exchange,89 and also showed excellent cycling.
Chapter3
Lithium iron oxides, such as the layered LiFe02, have also been
90
sized by ion exchange of the sodium compounds. The sodium
und a-NaFe02 was formed by hydrothermal treatment of FeOOH
large excess of NaOH (Na:Fe = 250) at 220°C; this sodium compound
hen partially converted to the lithium compound by ion exchange
rothermal
Mn02 KxMn02 •nH20
3 block
C
Mn02 Ko.2sMn02 KxMn02-8 Ko.7Mn02 KMn0 2
• •
tunnel 3 block
• •
C
Mn02 Ko.2sMn0 2 KxMn02-8 Ko.7Mn02 KMn0 2
• •
tunnel 3 block 2 block • •
3.4. Composition of the potassium manganese oxides, both hydrotherma158 and high
69
ture forms.
90
hydrothermal conditions, with 80% conversion in four hours at
. Reaction between NaFe02 and molten lithium halides results in
complete ion-exchange results in the formation of LiFe02, but with
ic lattice constant of 4.162 A indicating that the lithium and iron have
91
e disordered giving a NaCl lattice. Reducing the temperature by
a eutectic mixture of LiCl and LiN03 resulted in the formation of a
phase. An extensive study of the formation of lithium-iron-oxygen
92
failed to synthesize the layered LiFe02 compound. However,
93
e et al. reported the formation of "layered" LiMn0 2 which consisted
layered a-NaFe02 phase and a spinel phase; they synthesized this by
change of a-NaFe02 in a LiCl/KCl molten salt at 400°C.
The electrochemically active vanadium oxide nanotubes are also
sized by ion exchange from the initially formed alkylammonium
55 56
giving the sodium or manganese compound.
is of Battery Materials 97
Heat Aqueous
phous Solid Solution at 170°C
AxMn1 2 •y"H20
wherey" >1
.5. Schematic showing the synthesis, redox and intercalation reactions for Mn02 .
Chapter 3
ng. There are two distinct ways of doing this: geometric and electronic
zation of Mn02 layers. In the former, the structure is prevented from
to ccp, for example by placing pillars between the Mn02 sheets, or by
a none ccp structure, such as a two-block structure. Here, one is trying
feat thermodynamics by making the transformation difficult -
ally controlled reaction. In the latter, the manganese is made to be
cobalt-like by doping the lattice with a transition metal to the right of
such as nickel. Whichever way is chosen presents a synthetic
nge, as in many cases a metastable structure must be formed.
In the geometric approach, cations larger than lithium must be inserted
en the sheets. The potassium manganese oxides, described above,
the potassium ions are in trigonal prismatic sites thus breaking-up the
n ccp are readily formed at elevated temperatures or by hydrothermal
ques (see Figure 3.4). Though potassium enhances the cyclability,
ions eventually diffuse out of the lattice and it reverts to a spinel
re. Thus, immobile ions must be inserted into the lattice, between the
. Vanadium oxide pillars, probably V-0-V groups, were successfully
d using a combination of ion-exchange resins to form the starting
94
al followed by hydrothermal reaction; the structure is believed to be r
to that of the pillared Zn3(0H)2V207•2H20.95 This vanadium pillared
re was stable and did not convert to spinel even at 5 volts; however,
e characteristics were poor relative to LiMn02 itself. Leroux et al.96
sized a similar Ranceite based compound, which also was stable but
had poor rate capabilities.
The second geometric approach again aims to prevent the formation of
97
lattice, by forming a LiMn0 2 in which the oxygen layer stacking is
ger the three-block IABICAIBCIABI structure, but for example is the
ock IABICBIABI structure or even the hexagonal close-packing found
iS2 or in Co02• As none of these structures are stable for lithium at
emperatures, the sodium or potassium compounds must first be
sized by standard high temperature techniques; these may then be ion
ged to give the corresponding lithium compound as discussed above.
itial composition of the sodium or potassium compound may be varied
a wide range of stacking opportunities (see Figure 3.4); moreover, in
potassium compounds and in some of the sodium compounds the
cation is in trigonal prismatic coordination; as the lithium ion must be
hedral coordination, there will be sliding of the Mn02 layers leading
further stacking opportunities. All these structures are inherently
dynamically metastable, as the Mn02 sheets can slide relative to one
r and the manganese ions are quite mobile.
In the electronic stabilization approach, the goal is to make the
nese look electronically more like cobalt where the layer structure is
rmodynamically stable phase, for example by doping with nickel. 98
is of Battery Materials 99
5
4
3
2
1
0
,-.....
.>._.., 4
3
00
ro
....... 2
........
0
> 1
0
4
3
2
1
0
0 50 100 150 200
Capacity (mAh/g)
.6. Effect of synthesis firing temperature on the electrochemical capacity of
102
n0.4Co0.202.
s of Battery Materials 101
n 600 and 900°C forming LiFeP04. Such material has been well
erized,108 and has lattice parameters of a=10.329 A, b=6.007 A,
1 A and volume= 291.0 N. LiFeP04 can also 109 110been synthesized by a
of other techniques, including hydrothermal ' and microwave.
In the hydrothermal process, lithium acetate, ferrous oxalate and
horic acid dissolved in water are heated at around l20°C for up to 5
110 The reaction was found to be complete in much less than an hour. It
so be carried out under microwave heating. However, X-ray analysis
d that there is some iron on 111the lithium sites which restricts the rate of
insertion and removal, making the hydrothermal approach
able for LiFeP04 synthesis. Such disordering can be recognized from
tice parameters, a=10.403 A, b=6.023 A, c=4.726 A and volume =
N, which are significantly larger than the ordered material. A high
ature anneal appears to be necessary to cause the ordering of 112 the
and iron atoms. A recent report on the microwave processing of
04 would seem to have the same problem, as indicated by the rather
ectrochemical capacity even at elevated temperatures.
The original publication on the LiFeP04 electrode suggested that the
resistivity of the material was the cause of the rather low
chemical capacity, around 0.6 Li!Fe. Several efforts were directed at
ing the effective electronic conductivity, including modifying the
sis process so that a conductive coating is formed at the time of
sis - in-situ coating. Such approaches have been used in the case of
113 114
um oxides. Ravet et al. • proposed firing the initial reactants with
aceous materials such as sugar, so as to form a carbon coating thus
ing the overall conductivity of the composite so formed. These
sites allowed almost complete utilization of the lithium. Even more
sive results were obtained when a carbon gel was fired with the
leading to complete utilization of the LiFeP04 at 5 mg/cm2
115
ts,
116
gs. Masquelier showed the effectiveness of ball-milling carbon with
04 in giving complete capacity at elevated temperatures. A
rative study of the effect of synthesis/coating procedure, including in-
r bon ge I, sugar and aqueous ge1 versus post synthes1.s coatm. g
117
bed above for the layered manganese oxides. ° Chiang et al. have
6 120
the lattice with for example 1 wt% niobium and showed that the
9
nce is reduced from around 10 ohm-em in very pure LiFeP04 to
2
10 ohm-em. The very low doping levels have a surprisingly large
t; the material is black, but there are apparently some carbonaceous
2
es present. The electrochemical capacity is around 80% at 0.4 mA/cm
2
2.5 mg/cm loading, significantly lower than that from the carbon gel
115
gs. The resistance of regular battery grade LiFeP04 formed from
5 6 118
t grade chemicals is around 10 -10 ohm-em, not significantly
60
nt than that of undoped manganese oxide. The impact of synthesis
ure on the lithium ion conductivity is still not known and this might
as critical if not more critical than the electronic conductivity.
n even if only as a solvating specie. This can be equally true for the
e as for the anode. Thus, when propylene carbonate was used as the
lyte solvent in the LiffiS2 cell, if a trace of water was present then
ene carbonate was co-intercalated with the lithium expanding the
hree fold. The ether dioxolane was found to be a particularly ve
electrolyte solvent for lithium batteries, not being co-intercalated eing
128 129
very good for lithium electroplating. " Such solvents were
ed for use with the lithium-graphite anode. °13
Carbon based anodes
131
w used in essentially all commercially lithium-ion batteries, and
l to their success is choice of solvent with ethylene carbonate being
successful. Hard carbons rather than pure graphite are preferred for
132
cycle life and capacity. The batteries are constructed in the
rged state, and the lithium is electro-intercalated into the graphite on
st charge. Thus, the cathode material provides all the lithium for the
which means that excess anode cells are not constructed. Thus, any
side-reactions have an immediate negative impact on cell capacity.
133
et al. showed that the graphite surface is coated with a protective
n the first synthesis of C6Li which irreversibly consumes some of the
. This film is stable and so need not be reconstructed on each
chemical cycle but it does permanently reduce the capacity of the
. Hence, there is much interest in synthesizing other potential anode
als that have a higher capacity, and safer operation than graphite based
unds. The outstanding stability and rechargeability of graphite after
t cycle will be hard to surpass.
OJ)
600
-400
u
200
10
Cycle
.7. The electrochemical capacity of several tin based anodes, synthesized by different
es.
is of Battery Materials 105
144
The intermetallic InSb has also been prepared by ballmilling. In this
dium and antimony metals in a 1:1 ratio were ballmilled in air at room
ature for 20 hours; after which the resulting powder was annealed
flowing argon at 400°C for 12 hours. This material had a capacity of
220 mAh/g for the 2nd to 20th cycle.
Mechanical alloying is particularly amenable to producing
145 146
etallic alloys or mixtures of them. Mao et al. ' prepared
sites of SnFe2 and SnFe3C. Powders of tin, iron and carbon were
in hardened steel vials together with hardened steel balls; these were
in an argon environment, and then shaken for 20 hours. X-ray
tion showed that the resulting grain size was about 10 nm. In this case
e SnFe2 phase is electrochemically active, the carbide phase being the
matrix; the cell capacity which exceeds 1500 mAh/cc (around 200
) for 24% SnFe2 can be enhanced by increasing the ratio of active
nent but it is at the expense of capacity retention. The same
147
nical alloying process has also been used to reduce the particle size
e2 formed by high temperature processes. A recent review of tin-based
148
etallics details some of the other synthetic processes used.
NOWLEDGEMENTS
ERENCES
Gaines, R. W. Francis, G. H. Newman, B. M. L. Rao, "Ambient Temperature Electric
le Batteries Based on Lithium and Titanium Disulfide", in lith lntersociety Energy
ersion Conference (1976) 418. Stateline, Nevada.
ines, R. Cuenca, "Costs of Lithium-Ion Batteries for Vehicles", in Argonne National
atory Report ANUESD-42. 2000, Argonne National Laboratory, p. 57.
Thompson, F. R. Gamble, C. R. Symon, Mater. Res. Bull. 10 (1975) 915.
Thompson, Phys. Rev. Lett. 35 (1975) 1786.
not, A. L. Blanc, J. Rouxel, Bull. Soc. Chirn. France (1969) 2670.
xel, M. Danot, J. Bichon, Bull. Soc. Chirn. France (1971) 3930.
dorff, Chirnia 19 (1965) 489.
not, J. Rouxel, Cornptes Rendus Acad. Sc. Paris C276 (1973) 1283.
Whittingham, Prog. Solid State Chern. 12 (1978) 41.
. Whittingham, U.S. Patent 4040017 and U.K. Patent 1468244 (1973)
. Whittingham, J. Chern. Soc. Chern. Cornrn. (1974) 328.
. Dines, Mater. Res. Bull. 10 (1975) 287.
. Whittingham, M. B. Dines, J. Electrochern. Soc. 124 (1977) 1387.
. Murphy, F. J. DiSalvo, G. W. Hull, J. V. Waszczak, lnorg. Chern. 15 (1976) 17.
chollhom, H. Meyer, Mater. Res. Bull. 9 (1974) 1237.
Chapter 3
2.
Y. Zavalij, M.S. Whittingham, Acta Cryst. B55 (1999) 627.
Chen, P. Zavalij, M.S. Whittingham, Chern Mater 8 (1996) 1275.
Chen, T. Chirayil, M.S. Whittingham, Proceedings of the lOth International
posium on Solid State lonics, Singapore, December 1995. Solid State Jonics 86-88
96) 1.
harma, G. Moore, F. Zhang, P. Y. Zavalij, M. S. Whittingham, Electrochem. Solid-
te Lett. 2 (1999) 494.
Zhang, M.S. Whittingham, Electrochem. Solid-State Letters 3 (2000) 309.
Feng, H. Kanoh, Y. Miyai, K. Ooi, Chern. Mater. 7 (1995) 1226.
Ching, J. A. Landrigan, M. L. Jorgenson, N. Duan, S. L. Suib, C.-L. O'Young, Chern.
er. 7 (1995) 1604.
Bach, M. Henry, N. Baffler, J. Livage, J. Solid State Chern. 88 (1990) 325.
ranger, S. Bach, J.P. Pereira-Ramos, N. Baffler, J. Electrochem. Soc. 147 (2000)
6.
eroux, D. Guyomard, Y. Piffard, Solid State lonics 80 (1995) 299.
Tsuji, S. Komarneni, Y. Tamaura, M. Abe, Mat. Res. Bull. 27 (1992) 741.
Omomo, T. Sasaki, M. Watanabe, Solid State lonics 151 (2002) 243.
Delmas, C. Fouassier, Z. Anorg. Allg. Chern. 420 (1976) 184.
Mendiboure, C. Delmas, P. Hagenmuller, J. Solid State Chemistry 57 (1985) 323.
Delmas, C. Fouassier, P. Hagenmuller, J. Solid State Chemistry 13 (1975) 165.
Delmas, J.-J. Braconnier, C. Fouassier, P. Hagenmuller, Solid State lonics 3/4 (1981)
.
Delmas, M. Devalette, C. Fouassier, P. Hagenmuller, Mat. Res. Bull. 10 (1975) 393.
Y. Yao, J. T. Kummer, J. lnorg. Nucl. Chern. 29 (1967) 2453.
S. Whittingham, R. A. Huggins, J. Electrochem. Soc. 118 (1971) 1.
S. Whittingham, R. A. Huggins, NBS Special Publications 364 (1972) 139.
Delmas, F. Capitaine, 8th IMLB 8 (1996) 470.
apitaine, P. Gravereau, C. Delmas, Solid State lonics 89 (1996) 197.
R. Armstrong, P. G. Bruce, Nature 381 (1996) 499.
D. Robertson, A. R. Armstrong, A. J. Fowles, P. G. Bruce, J. Mater. Chern. 11 (2001)
.
R. Armstrong, R. Gitzendanner, A. D. Robertson, P. G. Bruce, Chern. Commun. (1998)
3.
M. Doeff, M. Y. Peng, Y. Ma, L. C. DeJonghe, J. Electrochem. Soc. 141 (1994) L145.
M. Doeff, T. J. Richardson, L. Kepley, J. Electrochem. Soc. 143 (1996) 2507.
M. Doeff, M. Y. Peng, Y. Ma, S. J. Visco, L. C. DeJonghe, US Patent 5,558,961
96)
M. Doeff, K.-T. Hwang, T. J. Richardson, L. C. DeJongue, Electrochem. Soc. Proc.
2 (1999) 184.
G. Mumme, Acta Cryst. B24 (1968) 1114.
M. Doeff, T. J. Richardson, K. T. Hwang, A. Anapolsky, Electrochem. Soc. Proc. 99-
2000) 48.
M. Doeff, T. J. Richardson, K.-T. Hwang, A. Anapolsky, ITE Battery Lett. 2 (2001) B.
R. Armstrong, H. Huang, R. A. Jennings, P. G. Bruce, J. Mater. Chern. 8 (1998) 255.
Tabuchi, C. Masquelier, T. Takeuchi, K. Ado, I. Matsubara, T. Shirane, R. Kanno, S.
u, H. Sakaebe, H. Okuyama, 0. Nakamura, Abstracts of the lOth Int. Conf on Solid
e Ionics, Singapore, December 3-8 (1995) 197.
uchs, S. Kemmlersack, Solid State lonics 68 (1994) 279.
Tabuchi, K. Ado, H. Sakaebe, C. Masquelier, H. Kageyama, 0. Nakamura, Solid State
cs 79 (1995) 220.
hirane, R. Kanno, Y. Kawamoto, Y. Takeda, M. Takano, T. Kamiyama, F. Izumi,
d State lonics 79 (1995) 227.
Chapter 3
. Nishi, in: M. Wakihara and 0. Yamamoto (Eds.), Lithium Ion Batteries, Kodansha,
kyo, 1998,p. 181.
. Fong, U. v. Sacken, J. R. Dahn, J. Electrochem. Soc. 137 (1990) 2009.
. V. Boldyrev, Experimental Methods in Mechanochemistry of Inorganic Solids, 1986,
ovosibirsk. Nauka (in Russian).
. V. Boldyrev, Mechanochemistry and Mechanical Activation of Solids, 2004, World
ientific.
. Senna, Solid State Chemistry 63-65 3.
Gutman, Mechanochemistry of Materials, 1997, Cambridge International Science
b.
. V. Kosova, E. T. Devyatkina, Electrochem. Society Abstracts, 2003-1 (2003) 1009.
. D. Kepler, J. T. Vaughey, M. M. Thackeray, Electrochem. Solid-State Lett. 2 (1999)
7.
T. Vaughey, K. D. Kepler, C. S. Johnson, T. Sarakonsri, R. Benedek, J. O'Hara, S.
ackney, M. M. Thackeray, Electrochem. Soc. Proc. 99-2 (1999) 280.
. M. Thackeray, J. T. Vaughey, A. J. Kahaian, K. D. Kepler, R. Benedek, Electrochem.
omm.1(1999)111.
. Tamura, R. Ohshita, M. Fujimoto, S. Fujitani, M. Kamino, I. Yonezu, J. Power
urces 107 (2002) 48.
Yang, P. Y. Zavalij, M. S. Whittingham, Mater. Res. Soc. Proc. 756 (2003) 295.
T. Vaughey, J. O'Hara, M. M. Thackeray, Electrochem. Solid-State Lett. 3 (2000) 13.
Mao, R. L. Turner, I. A. Courtney, B. D. Fredericksen, M. I. Buckett, L. J. Krause, J.
Dahn, Electrochem. Solid-State Lett. 2 (1999) 3.
Mao, J. R. Dahn, J. Electrochem. Soc. 146 (1999) 414.
Mao, R. A. Dunlap, I. A. Courtney, J. R. Dahn, J. Electrochem. Soc. 145 (1998) 4195.
. Winter, J. 0. Besenhard, Electrochim. Acta 45 (1999) 31.
pter 4
8 9 0 11 12
hree components of the cell. ' ,t , ' These have resulted in significant
vements in the performance (increase in energy density and rate
lity by a factor of two) and cost (more than a factor of two). Aspects of
sitive electrode are discussed in Part III and the electrolyte in Part IV of
ook. From the anode point of view, these directions have concentrated e
areas: improvement of carbonaceous materials; utilization of metallic m
in combination with stable electrolytes; and a search for new als.
The most suitable direction for new anode materials is not
ndent of the other components in the cell - for example, successful
mentation of a 5V cathode and electrolyte system would allow for
s that have working potentials above 1 V to be utilized. Similarly,
ons at improving the stability of metallic lithium anodes are directed
ds implementation of lower voltage cathodes that have the advantage of
lyte stability.
Carbonaceous Materials
19 21
g surface bonds. ' Capacities well above the theoretical limit of 372
are observed, but are inevitably associated with a large hysteresis.
pecific new forms of carbon which have replaced petroleum coke and
rise to high capacity/rate capability anodes include MCMB
arbon microbeads or spherical graphite), MCF (microcarbon fiber),
ith natural and synthetic graphite flakes. MCMB's in particular, are
utilized. They exhibit large capacities after heat treatments at low
22 23 24
tures, as well as after graphitization at higher temperatures. • •
materials are derived from petroleum industry byproducts or coal
s, but are expensive owing to low yields. They, along with other
ceous and graphitic anodes are discussed in detail in Chapters Sand 6.
material c AI Si Sn Bi
ical specific
372 2235 4010 959 385
y (Ah/kg)
ical volumetric
y (Ah/1) 833 6035 9340 7000 3773
1. Tin compounds
.2. Antimonides
.2. CoP 3 skutterudite structure with comer-joined CoP6 octahedra (left) and FeP2
structure with edge-joined FeP6 octahedra (right).
imilar to CoSb3, the first reaction step (which takes place at 0.6 V and
vs. Li+/Li for CrSb2 and FeSb2 respectively), leads to a complete
ization of the structure and the formation of LhSb alloys as shown by
tterns (Figure 4.3).
3500
3000
2500
2000
c::1"' 1500
0
(.)
1000
500
0
50 60 70
10 20 30 40
20( 0
)
.3. XRD patterns of FeSbz (a): reference, (b): pnstme material synthesized by
g, (c): after 5 cycles stopped at the end of discharge and (d): after 5 cycles stopped at
charge. (*) and ( )represent Li3Sb and Sb peaks respectively.
nd Composite Anodes: An Overview 123
ore stable cycling behavior was noted for FeSb2 than in CoSb3,
r as a capacity of 300 mAhlg can still be obtained after 30 cycles at a
e.56
eversing (increasing) the M:Sb ratio provides similar outcomes,
ified by Li uptake in Mn 2Sb which also gives simple dissociation of
59
ting material into Mn and Sb to form LhSb alloys (equation 4.1).
n be contrasted with the behavior of copper antimonides, such as the
tural Cu2Sb (see below), where the apparent limited region of Li-Sn
Cu solubility initially leads to ternary phases.
b.
4.4. Phase transformation of'YJ-Cu 6Sn5 during lithiation (a) to form "Li2CuSn" (b).
66
has been recently reviewed, and details are given elsewhere.
nd the second step involves reaction of the extruded tin with Lito form
71 119 7
alloys. A recent study using Sn Mossbauer combined with Li MAS
spectroscopy confirmed the dissociation of the starting material during
rst discharge and also demonstrated that a partial recovery of the
ure occurs upon charge for SnSb single phase electrodes.The good
lity of this system can be attributed to the fact that 1) both components
equentially: LhSb formation at 700 mV vs. Li+!Li is followed by LixSn
41
ormation at lower voltages as shown in the first report on this system.
fore at each step of the reaction, the lithiated phase is embedded in an e
72
matrix; 2) the extruded tin and subsequent LixSn alloys are more ble
for reaction than in the case of InSb, since bulk tin crystallites are ntly
not formed on the surface. However, low voltage reaction ing
formation of tin alloys result in further electrode expansion and loss
erparticle contact between LhSb and Sn, causing irreversibility and
ty fading.
The best results were obtained for a Sn/SnSb mixture with a reported
73
-limited specific capacity of 360 mAh/g during 200 cycles. However,
materials exhibit a large irreversibility during the first cycle due to the
ion of a SEI layer and also because of residual impurities inherent in
nthesis process.74 '75
mercial batteries.
. Aluminides
Silicides
everal efforts have also been undertaken on silicon-based electrodes.
76 77 79
on has been given to Mg2Si, ' 78 ' ' but despite much effort its
ies have proven disappointing. More promising are silicon-based
ites prepared/designed by a variety of techniques.
MgzSi
Si-Based Composites
METAL OXIDES
etal oxide anode materials can be divided into two classes. One is the
ertion compounds: metal oxides capable of intercalation chemistry at
ential. These materials are characterized by relatively high potentials
between 1.4-1.8 V), low lithium capacities and often highly
le insertion reactions. The other class of metal oxides display Li
es derived from reaction with the lithium to form another material,
s usually the reduced metal oxide encapsulated in a matrix of LhO,
to the case of the tin oxides described above. These materials also
a relatively high average charge potential, but also a large hysteresis
nt with the necessity of oxygen transport within the bulk during
n. Critical to the functioning of these materials, therefore, is the
mixture at the nanoscale regime, of metal sub-oxide (or pure metal,
ng on the degree of Li uptake) and oxygen from the lithium oxide
hat facilitates the oxygen recovery process.
Chapter4
Nitrides
121
-ion cells. In recent work, the material has also been coupled with
122
ntaining anode materials to compensate for irreversibility in the latter.
The electrochemical behavior of main group nitrides capable of acting
123
i "sink" have been recently described by the Telcordia group. They
that uptake of Li by Zn 3N 2 results in formation of LiZn in a matrix of
. On charge, complex reactions result in the oxidation of LiZn to
ic Zn, and reaction of -LhN with LiZn to form LiZnN which is the end
er phase for all subsequent cycles. This first-reported reversible LhN
rsion gives a reversible capacity of 555 mAh/g but the cycle life was
d.
Phosphides
Oxide
Density of states Density of states Density of states
Where a binary MP exists but no stable LiMP ternary phase of the same
oichiometry, two possibilities for Li uptake can be predicted. The
n can either yield a metastable ternary phase, such has been proposed
hium insertion in FeP2; or decomposition to yield Li 3P and metal can
akin to the reaction in antimonides. The latter is exhibited by CoP3• In
ases, Li uptake (and subsequent extraction) results in an amorphous
al, making the nature of the Li uptake process difficult to discern.
nd Composite Anodes: An Overview 137
1. FeP2
2.5
+ 2.0
_J
rn
> 1.5
Q)
Cl
.l!! 1.0
g 0.5
2.2. CoP3
CONCLUSIONS
The task of finding alternate materials for anodes in lithium-ion cells to
nd Composite Anodes: An Overview 139
RENCES
996) 216.
Alcantara, J. M. Jimenez-Mateos, J. L. Tirado, J. ElectrochemSoc. 149 (2002) 201.
R. Dahn, W. Xing, Y. Gao, Carbon 35 (1997) 825
Tatsumi, N. !washita, H. Sakaebe, H. Shioyama, S. Higuchi, A. Mabuchi, H. Fujimoto,
Electrochem. Soc. 142 (1995) 716.
lnaba, H. Yoshida, Z. Ogumi, J. Electrochem. Soc. 143 (1996) 2572.
Alcantara, F. J. Femandez-Madrigal, P. Lavela, J. L. Tirado, J. M. Jimenez-Mateos, C.
mez De Salazar, R. Stoyanova, E. Zhecheva, Carbon 38 (2000) 1031.
lnaba, Z. Siroma, A. Funabiki, Z. Ogumi, T. Abe, Y. Mizutani, M. Asano, Langmuir
(1996) 1535; M. Inaba, Z. Siroma, Y. Kawatate, A. Funabiki, Z. Ogumi, J. Power
urces 68 (1997) 221; S. K. Jeong, M. Inaba, T. Abe, Z. Ogumi, J. Electrochem. Soc.
8 (2001)A989.
J. Wen, B. A. Boukamp, R. A. Huggins, W. Weppner, J. Electrochem. Soc. 126 (1979)
58; Y. Hamon, T. Brousse, F. Jousse, P. Topart, P. Buvat, D. M. Schleich, J. Power
urces 97 (2001) 185.
.S. Foster, C. E. Crouthamel, S. E. Wood, J. Phys. Chern. 70 (1966) 3042.
J. Wen, R. A. Huggins, J. Electrochem. Soc. 128 (1981)1181.
. Courtney, J.S. Tse, 0. Mao, J. Hafner, J. R. Dahn, Phys. Rev. B 58 (1998) 15583.
Chouvin, J. Olivier-Fourcade, J.C. Jumas, B. Simon, 0. Godiveau, Chern. Phys. Lett.
8 (1999) 413; R. A. Dunlap, D. A. Small, D. D. MacNeil, M. N. Obrovac, J. R. Dahn,
lloy Compd. 289 (1999) 135.
J. Weydanz, M. Wohlfahrt-Mehrens, R. A. Huggins, J. Power Sources 81-82, (1999)
7.
R. Goward, N.J. Taylor, D. C. S. Souza, L. F. Nazar J. Alloy Compd 329 (2001) 82.
Iodata, T. Kubota, A. Matsufuji, Y. Maekawa, T. Miyasaka, Science 276 (1997) 1395.
A. Courtney, J. R. Dahn, J. Electrochem. Soc. 144 (1997) 2045.
A. Courtney, J. R. Dahn, J. Electrochem. Soc. 144 (1997) 2943.
R. Goward, F. Leroux, W. P. Power, G. Ouvrard, T. Egami, W. Dmowski, L. F. Nazar,
ctrochem. Solid-State Lett. 2 (1999) 367; G. R. Goward, L. F. Nazar, W. P. Power, J. ter.
Chern. 10 (2000) 1.
Chouvin, C. Branci, J. Sarradin, J. Olivier-Fourcade, J. C. Jumas, B. Simon, Ph.
nsan, J. Power Sources 81 (1999) 277; J. Chouvin, J. Olivier-Fourcade, J. C. Jumas, B.
mon, 0. Godiveau, Chern. Phys. Lett. 308 (1999) 413; A. Hightower, P. Delcroix, G. Le
er, C-KK. Huang, B. V. Ratnakumar, C. C. Ahn, B. Fultz, J. Electrochem. Soc. 147
00) 1.
A. Courtney, J. S. Tse, 0. Mao, J. Hafner, J. R. Dahn, Phys Rev. B. 58 (1998) 15583.
. Corbett, E. Garcia, A. M. Guloy,W-M. Humg, Y-U. Kwon E. A. Leon-Escarnilla,
ern. Mater. 10 (1998) 2824.
. Tirado, Mat. Sci. Eng. R40 (2003) 103.
Yang, M. Winter, J. 0. Besenhard, Solid State lonics 90 (1996) 281.
A. Huggins, B. A. Boukamp, 1984 U.S. Patent 4,436,796.
Mao, R. A. Dunlap, J. R. Dahn, J. Electrochem. Soc. 146 (1999) 405; J. R. Dahn, ibid
6 (1999) 414; 0. Mao, J. R. Dahn, ibid., 146 (1999) 423
Larcher, L. Y. Beaulieu, 0. Mao, A. E. George, J. R. Dahn, J. Electrochem. Soc. 147
00) 1703.
Crosnier, T. Brousse, X. Devaux, P. Fragnaud, D. M. Schleich, J.Power Sources
(2001) 169.
M. Ehrlich, C. Durand, X. Chen, T. A. Hugener, F. Spiess, S. L. Suib, J. Electrochem.
. 2000 147 (2000) 886.
Y. Beaulieu, J. R. Dahn, J. Electrochem. Soc. 147 (2000)3237; L. Beaulieu, D.
rcher, R. A. Dunlap, J. R. Dahn, J. Alloys Compd. 297 (2000) 122.
. Dahn, R. L. Turner, 0. Mao, R. A. Dunlap, A. E. George, M.M. Buckett, D. J.
Clure, L. J. Krause, Thin Solid Films 408 (2002) 111.
nd Composite Anodes: An Overview 141
Kim, Y-J. Kim, D.G. Kim, H-J. Sohn, T. Kang, Solid State Ionics 144 (2001) 41
Sakaguchi, H. Honda, T. Esaka, J. Power Sources 81 (1999) 229
M. Wilson, B. M. Way, J. R. Dahn, T. van Burren, J. Appl. Phys. 77 (1995) 2363.
Kim, P. N. Kumta, G. E. Blomgren, Electrochem. Solid-State Lett. 3 (2000) 493.
S. Wang, G. T. Wu, X. B. Zhang, Z. F. Qi, W. Z. Li, J. Electrochem. Soc. 145 (1998)
51.
Li, X. Huang, L. Chen, Z. Wu, Y. Liang, Electrochem. Solid-StateLett. 2 (1999) 547.
Wang, M. Yoshio, A. Takeshi, Z.Ogumi, Electrochem. Solid-State Lett. 4 (2002) A499.
Yoshio, H. Wang, K. Fukuda, T. Umeno, N. Dimov, Z. Ogumi, J. Electrochem. Soc.
9 (2002) A1598.
L. Turner, World Intellectual Property Organization PatentApplication 2000
00/03444.
Y. Beaulieu, K. C. Hewitt, R. L. Turner, A. Bonakdarpour, A.A. Abdo, L. Christensen,
W. Eberman, L. J. Krause, J. R. Dahn, J. Electrochem. Soc. 150 (2003) A149.
Ohzuku, A. Ueda, N. Yamamoto, J. Electrochem. Soc.142 (1995)1431.
Kavan, M. Gratzel, Electrochem. Solid State Lett. 5 (2002) A39.
Ohzuku, K. Tatsumi, N. Matoba, K. Sawai, J. Electrochem. Soc. 147 (2000) 3592.
Huang, L. Kavan, I. Exnar, M. Gratzel, J. Electrochem. Soc.l42 (1995) L142.
O'Regan, M. Gratzel, Nature 533 (1991) 737
P. Feist, P. K. Davies, J. Solid State Chern. 101 (1992) 275; G. Nuspl, K. Yoshizawa,
Yamabe, J. Mater. Chern. 7 (1997) 2529.
Wagemaker, A. P.M. Kentgenst, F. M. Mulder, Nature 418 (2002) 397.
M. Thackeray, W. I. F. David, J. B. Goodenough, J. Solid State Chern. 55 (1984) 280;
A. Picciotto, M. M. Thackeray, Mat. Res. Bull. 21 (1986) 583.
Sigala, D. Guyomard, Y. Piffard, M. Tournoux, C.R. Acad.Sci.Paris. 320 (1995) 523;
Guyomard, C. Sigala, A. LeGal LaSalle, Y. Piffard, J. Power Sources 68 (1997) 692.
Kumagai, Jap. Patent08-241101 (1996).
Denis, E. Baudrin, M. Touboul, J-M. Tarascon, J. Electrochem. Soc. 144 (1997) 4099.
Leroux, G. Piffard, Ouvrard, J-L. Mansot, D. M. Guyomard, Chern. Mater. II (1999)
48.
Leroux, G. R. Goward, W. P. Power, L. F. Nazar, Electrochem. Solid-State Lett. 1
998) 201; F. Leroux, L. F. Nazar, Solid State Ionics 133 (2000) 37.
Leroux, L. F. Nazar, Mat. Res. Soc. Symp. Proc, New Materials for Batteries and Fuel
lls, II, 575 (1999) 173.
Nazar, G. Goward, F. Leroux, M. Duncan, H. Huang, T. Kerr, J. Gaubicher, J. Int. J.
rg. Mat. 3 (2001) 191.
Denis, E. Baudrin, F. Orsini, G. Ouvrard, M, Touboul. J-M. Tarascon, J. Power Sources
(1999) 79; S. Denis, E. Baudrin, F.Orsini, J-M. Tarascon, Chem.Mater. 12 (2000)
33.
Rowsell, J. Gaubicher, L. F. Nazar, J. Power Sources in press.
Poizot, S. Laruelle, S. Grugeon, L. Dupont, J. M. Tarascon, Nature 496 (2000) 407.
Duncan, L. F. Nazar, Mat. Res. Soc. Symp. Proc, Solid State Ionics 548 (1999) 71; M.
ncan, G. Ouvrard, L. F. Nazar, J. Gaubicher, Abstract# 235, The Electrochemical
ciety Meeting Abstracts, Hawaii,l999.
H. Gregory, J. Chern. Soc., Dalton Trans. (1999) 259.
H. Elder, L. H. Doerrer, F. J. DiSalvo, J. B. Parise, D. Guyomard, J. M. Tarascon,
rn. Mater. 4 (1992) 928.
Juza, K. Langer, K. von Benda, Angew. Chern. Int. Ed. Engl. 7 (1968) 360.
Nishijima, N. Tadokoro, Y. Takeda, N. Imanishi, 0. Yamamoto, J. Electrochem. Soc.
1 (1994) 2966; S. Suzuki, T. Shodai, Solid State Ionics 116 (1999) 1.
Nishijima, Y. Takeda, N. Imanishi, 0. Yamamoto, J. Solid State Chem.l13 (1994)
5.
Shodai, S. Okada, S. Tobishima, J. Yamaki, Solid State Ionics 86-88 (1996) 785
nd Composite Anodes: An Overview 143
Gudat, R. Kniep, A. Rabenau, Angew. Chern. Int. Ed. Engl. 30 (1991) 199.
latyk, R. Z. Kniep, Kristallogr. NCS 214 (1999) 447.
. C. Rowsell, V. Pralong, L. F. Nazar, J. Am. Chern. Soc. 123 (2001) 8598.
. Kim, M.G. Kim, J. M. Lee, T. Kang, H-J. Sohn, Electrochem. Solid-State Lett. 5
2) A103.
hodai, Y. Sakurai, T. Suzuki, Solid State Ionics 122 (1999) 85.
akeda, M. Nishijima, M. Yamahata, K. Takeda, N. Imanishi, 0. Yamamoto, Solid e
Ionics 130 (2000) 61
ang, Y. Takeda, N. Imanishi, 0. Yamamoto, J. Electrochem. Soc. 147 (2000) 1671.
ereira, L.C. Klein, G. G. Amatucci, J. Electrochem. Soc., 149 (2002) A262
onconduit, M. Tillard-Charbonnel, C. Belin, J. Solid State Chern. 156 (2001) 37.
ouza, V. Pralong, A. J. Jacobson, L. F. Nazar, Science 296 (2002) 2012.
ouza, V. Pralong, A. J. Jacobson, L. F. Nazar, 11th IMLB, Abstract #60, Monterey
fornia 2002.
L. Doublet, F. Lemoigno, F. Gillot, L. Monconduit, Chern. Mater. 14 (2002) 4126.
azar, D. Souza, V. Pralong, Patent Cooperative Treaty International Application 2002
/CA02/00016
ralong, D. C. Souza, L. F. Nazar, Electrochem. Comm. 4 (2002) 516
lcantara, J. L. Tirado, J. C. Jumas, L. Monconduit, J. Olivier-Fourcade, J. Power
ces 109 (2002) 308.
apter 5
hould be noted here, that cation and anion intercalation can also occur into carbon
ials used as conductive electrode additives. At the anode, this cation intercalation is a r
reaction occurring in parallel to intercalation in the active carbonaceous anode ial.
At the cathode, anion intercalation can happen during overcharge of the electrode.
onaceous and Graphitic Anodes 145
E ctrolyte
c
2
0 0 0
0 0 0
Electrolyte
0 0 0
Dts.charg•
tr
0 0
0 0
- 0
H 0
0
0
0
0 0
000 0
Discharge
lnMrtion Ho•t
( Or l!lph ile Anodto)
ln1ortlon Ho•l
(Cothodo)
c,.,x c,,x
oo
n
000
000
000
o g Elec trolyte
ooo
0 0 0 + . 980
000
0 0 92 2
M C,
0 "'"""..
c,x c.
+---
Charge g oo
c,x
c, c,
Charge
e 5.1. Rechargeable cells with active graphite intercalation electrodes: (A) graphite as
m cation intercalation host anode (negative electrode) in lithium ion cells, (B) graphite as
(accepting cations M) and cathode (positive electrode, accepting anions X) in dual ite
cells, and (C) hydrogen sulfate ion transfer cell based on graphite anode and cathode
accepting anions X=HS04.), CnX denotes a stage-1, C2nX a stage-11
gensulfate/graphite intercalation compound (cf. Section 5.3.2.2).
argeable batteries have been investigated since decades. In these so d
"dual-graphite" cells, 18 22 - the charge/discharge reaction is the
calation/release of electrolyte cations and anions into graphite (Figure
). As the ion concentration of the electrolyte changes during cycling,
gh electrolyte salt (cations +anions X) must be incorporated in the cell
uarantee the electrolyte conductivity, and enough solvent must be
Chapter 5
outline of
unit cell
e 5.2. Crystal structure of hexagonal graphite showing the AB layer stacking sequence,
it cell, and the subdivision in basal plane and prismatic surfaces.
Chapter 5
charge
Chapter 5
r:
37 0
·l!::;;:
$
ccupation of nearest neighbor sites is avoided (Figure 5.3B).
A Li layer B C
A
-1§ ij@@ . 0 0
cl
"""'
0· "<?'t:>
0.430 nm grap hene Iayer I} "<'fao
e 5.3. AA layer stacking sequence and the aa interlayer ordering of the intercalated
m in LiC6 (A). In-plane structures of LiC6 (B) and LiC2 (C).
A higher lithium in-plane density by occupation of nearest neighbor
is obtained in the phases LiC2 - LiC4, i.e., x in LixC6 = 2-3, which are
ared chemically from graphitic carbon under high pressure (-60 kbar)
89 173
high temperature (-300°C) conditions. • 170 - The close Li-Li distance
C2 (Figure 5.3C) results in a higher chemical activity 174
of lithium than that
thium metal (Li-Li bond length (20°C) = 0.304 nm ). Under ambient
itions LiC2 decomposes slowly via various metastable intermediate Li/C
89 172
es to LiC6 and metallic lithium. ' A study of the electrochemical
onaceous and Graphitic Anodes 155
vior of LiC 2 can be found in Ref. 89. For more details on the chemical
hesis of LixC6, see Refs. 26,88,89,175,176.
A
-
- -" Li
0.3
B
·m
3
S
== tag e II
la yer
Is> IV (1' Li)
1
IV]
stage I
I
I
0.1
IV II L
+ +
I
+ & (Reduction)
Ill :II L: II
I I
II+ I
t
0 0 '-'------+
t (i = const.)
I
-0.2 0.34 0.5 1 xin Li,C.
-30 18 12 6 n in LiC,
Ill II L II I Stage
e 5.4. Stage formation during electrochemical lithiation of graphite. A: Schematic
nt current charge curve. B: Schematic voltarnmetric curve.
Chapter 5
The first reasoning is only to a minor part responsible for the excess
ge. It is believed that reduction of impurities like H20 or 02 present on
arbon surface, reduction of surface groups at the prismatic surfaces, and
ersible lithium incorporation into the carbon matrix ("formation of
205 207
ue compounds" - ), e.g., by reduction of heteroatom-groups located
rbon layer edges in polycrystalline carbons, take place.
The main part of the irreversible charge loss, however, is due to
tions with the electrolyte, which typically consists of a mixture of
nic solvents and a highly soluble lithium salt. The strong reducing
er of graphite (and other carbon anodes), at potentials where lithium
53 208 211
calation proceeds, leads to reductive electrolyte decomposition. • -
suitable nonaqueous electrolytes, however, "passivating" films of
2.00
0
8 A
:::. 1.60
c
6 _J
4 .; 1.20
Discharge > Discharge
2
>
0 0.80
"iii
8 'E
Q)
6 0.40
0
4 a..
2 0.00
0
Charge 0 200 400 600 800
200 400 Capacity I Ah kg·'
Capacity I Ah kg·' 2.00
8
0 ·- +
::J
:::. D
6
B _J 1.60
ui
.4- > 1.20
2 >
0 Iii 0.80
8 .. 'E
6 ..
4 ..
a
a.. 0.40
2
0 0.00
100 200 0 200 400 600 800
Capacity I Ah kg·' Capacity I Ah kg·'
e 5.5. Typical schematic I" cycle charge/discharge curve of (A) graphite, (B) soft
n (coke), (C) soft carbon (low temperature, hydrogen containing), (D) hard carbon.
ductive additives like carbon blacks, which reversibly take up lithium, can not be
ered as inactive.
Chapter 5
rominent example for the latter effect are fluorinated binders like
(tetrafluoro ethylene) (PTFE; "teflon", (-CF2-CF2-)n) and several
239
(vinylidene fluoride) (PVdF, (CHrCFr)n)· Kavan has disclosed that
rinated polymers can be electrochemically reduced ("electrochemical
onization") at the conditions a graphite anode experiences in contact
an organic electrolyte in a lithium ion cell during charge. In general,
alogenated polymers react with lithium very easily. The dehalogenation
ds LiHal (LiF, LiCl...) and elemental carbon. The reduction of PTFE by
um is well investigated as PTFE was suggested by Gabano and
240
eagues as cathode for lithium primary cells. PTFE can be completely
241
erted into carbon by electrochemical reduction, therefore increases
42
irreversible capacitl and consequently should be avoided as binder
carbon anodes. The reaction of Li with halogenated polymers
aining vicinal halogen atoms is facilitated due to the formation of
getically favored products, i.e., conjugated C=C double bonds and LiF.
ially halogenated polymers which contain no vicinal halogen atoms, such
VdF and poly(vinyl chloride) (PVC) usually do not react, because the C
ond is more resistive against reaction with Li, which blocks the
239
ation of a conjugated carbon-carbon bond system. In contrast,
2431
san et al. [ have observed a highly exothermal reaction of PVdF with
ated graphite at elevated temperatures. It is likely that the presence of
cts or irregularities in the monomer sequence, which locally lead to
al halogen atoms in the partially halogenated polymer, may alleviate the
rochemical reduction by lithium.
°C reveal that the size of the single layers as well as the size of the
Chapter 5
Despite the fact that the high specific charge carbons show a multiple
e specific charge of graphite, there are still some problems to solve.
In many cases extremely high irreversible specific charges were
rved,157• 253• 265'277- 279• 284• 313- 315 occasionally also at higher cycle
bers.265•278•313 •314•316 •317 The' irreversible capacities can be correlated with
formation of the SEI. 157 278279 '
313 317
' ' However, an "irreversible lithium
265 317 320
rporation" into the carbon is discussed, too. • - Apart from an
ersible trapping of lithium in the pore volume of disordered carbons, 320
arge fraction of (internal and external) heteroatorns in the non-graphitic
ons, such as hydrogen and oxygen, can irreversibly bind a considerable
unt of lithium during reduction of the carbon.Z65• 284·317 -319 In addition, the
amount of polar surface groups adsorbs large amounts of impurities
air, which can irreversibly react with lithium and thus have to be
321
ved.93•312• All of the above charge losses have to be compensated by
xcess of cathode material as lithium ion cells are assembled in the
arged state. As, in contrast to hydrogen-containing carbons, hard carbon
electrochemical properties suitable for application in lithium ion
143
ries, there are attempts to reduce the irreversible capacity of hard
312 315 318
ons by changing the surface properties.284 • • •
3221
With one exception r , carbons exhibiting hysteresis show a low
iency and poor cycling performance. Furthermore, they can be
251
arged only over a broad potential region of about 1-2 V.87• 157 158 -
• •
7-279'305'308'313 As a result the practical energy efficiency of a lithium ion
s reduced.
The end of charge potential of non-graphitic carbons must be chosen
close to 0 V vs. LilLi+ in order to obtain the available capacity. The
w "safety gap" separating from the potential where metallic lithium is
sited on the carbon surface might give rise to some safety problems of
carbons. In particular, if fast charging is applied, there is a high risk of
m plating at the end of charge. In some cases 278 •279305
•
308
• the hard carbon
rode was indeed charged below 0 V vs. Li/Lt and dendritic323 lithium
sition on such carbons was observed. In contrast, there is a difference of
Chapter 5
Though
• the SEI concept has been basically adapted for carbonaceous
08 210 212 213
es/ • ' there are significant differences between film formation
esses on metallic lithium and lithiated carbon. Film formation on
llic Li takes place upon contact with the electrolyte. The electrolyte,
h consists of various electrolyte components (solvent(s), electrolyte
ives, electrolyte salt, impurities) decomposes spontaneously with low
14
tivity, and some of the decomposition products form the film. On the
ary, film formation on the surface of carbonaceous hosts takes place as
arge consuming side reaction in the first few Lt intercalation/de
calation cycles, especially during the first reduction of the carbon host
rial, leading to the so-called "irreversible capacity". In this case, the
rolyte components which are least stable towards reduction selectively
first. The chemical composition of the SEI on lithiated carbon varies
14
gh the depth of the film. In addition, the film composition and
hology is also laterally non-homogeneously distributed over the carbon e
79
surface? As a consequence of the difference in the reduction
vior to metallic lithium, dissimilar intermediate reduction products and
rent decomposition mechanisms of the electrolyte may occur in the
nce of lithiated carbon.
Finally, the behavior of metallic lithium and of lithiated carbon,
ctively, during battery charge and its influence on the SEI has to be
ssed. Upon recharge of the Li anode, lithium plating occurs
ltaneously to lithium corrosion and "passivation" (formation of SEI).
, lithium deposits as highly disperse, highly reactive metal particles.
e dendrites are covered with SEI films, which in some cases
ronically insulate them from the rest of the Li. Thus, the dendrites are
onaceous and Graphitic Anodes 169
hite. In the "best" case reduced charge storage capabilities, in the worst
complete electrode destruction are the results of this reaction.
As long as the content of lithium in the graphitic carbon is low (
in LixC6), the ternary lithium-graphite intercalation compounds are
modynamically favored over the unsolvated binary compounds, LixC 6
ure 5.7A). Hence, the potentials of their electrochemical formation are
e positive than those for the formation of the corresponding compound
6. At this stage of Lt intercalation the coulombic interaction between
lithium guest layer (Ln and the balancing negative charge distributed
the graphene layers (Cn) is weak, and space to accommodate large
12 14 59 60 34 435
ent molecules is still available. • • • .4 •
The solvated GICs are thermodynamically unstable with respect to the
12 14 59 60
ction of the co-intercalated solvent molecules. The kinetically
• · •
rolled reduction depends on the type of co-intercalated solvent. It is slow
e.g., dimethyl sulfoxide, where even staging of solvated GICs can be
.4 6.4 46
where the
' -
11 3 37 43 438 441
rved, but very much faster for, e.g., PC, '
trochemical intercalation followed by fast decomposition of the
calated Lt(solv)y can be misunderstood as simple electrolyte
mposition. The reduction of Li\solv)y inside the graphite is associated
extra irreversible capacity.
Both effects resulting from solvent co-intercalation, the mechanical
ruction and the higher irreversible capacities, seriously complicate the
ation of graphitic anode materials. Due to the fact that ternary lithiated
ons are thermodynamically favored at low lithium concentrations in the
hite, i.e., at the beginning of intercalation, kinetic measures have to be
ied to diminish or even to completely suppress solvent co-intercalation.
Binary (unsolvated)
A lithium-graphite
intercalation compound
00000000
00000000
Ternary (solvated)
B lithium-graphite
intercalation compound
-graphene layer o
lithium
0 solvent
e 5.7. Schematic drawing (A) of binary and (B) of ternary lithium-graphite intercalation
ounds, as well as (C) of the SEI formed on/in graphite via electrochemically formed
ted lithium/graphite intercalation compounds.
onaceous and Graphitic Anodes 173
hite matrix, where they are reduced. Again gas is produced, which leads
rther cracking and so on. This reaction mechanism involves no PC co
calation step. This behavior seems to be only typical for PC electrolytes,
or electrochemical reduction or oxidation of other nonaqueous and
ous electrolytes, which results into strong gas evolution; a similar and
rapid mechanical destruction mechanism has not been reported so far.
465
Takamura and colleagues investigated highly graphitized carbon
rs in PC based electrolytes. They found a dependence of the extent of
liation on the number of used carbon fibers and their relative orientation
ach other. With a single fiber no exfoliation and very low irreversible
city could be observed. With a bundle of five fibers in which the fibers
been irregularly oriented to each other, strong exfoliation takes place;
irreversible capacity is high. However, if the fibers in the bundle are
nged in parallel to each other, the extent of the exfoliation reaction is
much reduced. They attributed the exfoliation to an inhomogeneous
ent distribution in the bundle, which does not occur with a single carbon
r. One might, however, also speculate that the irreversible capacity in
case is also reduced, when the current density is high and furthermore is
orm for the total surface area of the anode material.
466
The authors have investigated the reactions of graphite vs. PC by in
electrochemical dilatometry. Strong macroscopic expansions of graphite
r, which can be partially reversible under certain conditions. The results
est that PC co-intercalation into graphite is the initial step, which opens
ssures and pores in the graphite. The second step is physical penetration
C electrolyte into these openings of graphite, then gas evolution inside
hite, which causes strong mechanical stress and graphite destruction.
462 63 67
subsequent reaction process is also supported by Chung et al. .4 .4
=/X
+e· .. =/X
=--'X
•e
Many of the above film forming compounds take advantage from the
that they are usually reduced at quite positive graphite electrode
ntials, i.e., at the early stages of the first reduction of the graphite
trode, before detrimental side reactions such as solvent co-intercalation
eed. The formed films have SEI properties. Solvent co-intercalation n
as a side reaction) is suppressed. This sensitiveness towards reduction
be explained (i) by the electron-withdrawing effect of chlorine- or
rine-groups introducing partial positive charge on the molecule
orinated and fluorinated solvents) or (ii) by the presence of atoms or
ps in the molecule which have a high electron affinity, e.g., sulfur in the
ation state +IV (sulfites) or double bonds (vinylene compounds). The
ction behavior and the reduction potentials of film forming components
461
be also calculated, '486 though the dependence on the chosen
rimental conditions (type of electrode and co-electrolyte components,
ge regime, kinetics, etc.) will inevitably make a practical test necessary.
In many cases, electrolyte additives not only improve typical lithium
battery electrolytes, but also allow the use of PC-based electrolytes and
r electrolytes not compatible with graphitic anodes. Especially, the use
C instead of EC in a lithium ion battery electrolyte is highly desired, as
as a large liquid range from -49°C to +240°C, whereas EC is a solid at
temperature. Thus, PC-based electrolytes usually show a low
erature performance desired for use in lithium ion cells. The low
osity of the PC at lower temperatures is counteracted by the still high
uctivity of the SEI film formed in the presence of PC.530
There is no universally applicable electrolyte additive or electrolyte
ent available and therefore additive/solvent mixtures are used at the
ent. However, the cross-interaction of the various electrolyte
ponents contributing to SEI formation is an issue of concern. For
ple, in electrolytes designed for good low temperature performance, not
still comparatively high electrolyte conductivity, but the low ionic
uctivity • of lithium within the SEI formed in these
rolytes,324 506' 531' 532and especially the low conductivity within the carbon
24
riae •506• 533 at these temperatures, limit the low temperature
ormance. Similar observations were made with flame retardant or other
y-relevant electrolyte additives which may deteriorate,534• 535 but also
improve536 '537 the SEI formation process. Finally, electrolyte additives
esired, which are in parallel beneficial for the interphase formation at
e and cathode.
Disordered soft and hard carbons are much less sensitive to solvent co-
onaceous and Graphitic Anodes 177
A B
SEI Products
SEI Products Nuc
Substrate Substrate
(Carbon Surface) (Carbon Surface)
e 5.8. Deposition (plating") of SEI film products. Composition and morphology of the
eposit are controlled by the "plating conditions", i.e., by (i) the electrolyte composition, e
pretreatment (coating or modification) of the carbon surface ("substrate"), and (iii) the g
conditions (constant current or pulse plating, etc.): A) Plating conditions favoring
ation result in a compact and dense SEI. B) Plating conditions which facilitate the crystal
h of the initially formed nuclei induce the formation of a rough SEI deposit, which
es more SEI products to get pin-hole-free, i.e., consumes more irreversible capacity.
considerations can be applied to other anode and cathode materials, too.
Carbons have been and still are considered to be the anode material for
rgeable high energy density lithium batteries, due to a synergetic
ination of properties beneficial for performance and production. These
erties include: high lithium storage capacities, negative redox potentials
he cathode, long term cycling performance, low costs and large
dance. Any other anode material discussed in this book may possess
al, but not all of these properties. Nevertheless, as (surface)
fications or the use in composites seem to be the perspective for future
n anode materials, the same might be true for other anode materials,
e in a composite together with carbon.
Two "wonders" are associated with the story of success of carbon
es. The first wonder is strongly related with the properties of metallic
m. Apart from having the highest capacity, among the light metals M =
a, and Mg, solely metallic Li shows a chemical and electrochemical
vior which favors its use in high energy density batteries, because
ce films with the unique SEI properties are formed. In contrast,
cts of electrolyte decomposition of alkali metals other than Li are
Chapter 5
KNOWLEDGEMENT
FERENCES
M. B. Armand, in: Fast Ion Transport in Solids, W. van Gool, Ed., Elsevier, Amsterdam,
1973,665.
B. C. H. Steele, in: Fast Ion Transport in Solids, W. van Gool, Ed., Elsevier, Amsterdam,
1973, 103.
M.S. Whittingham, Science 1976, 192, 1226.
M.S. Dresselhaus (Ed.), Intercalation in Layered Materials, NATO ASI Series, Plenum,
New York, 1986, Vol. 148.
A. P. Legrand, S. Flandrois (Eds.), Chemical Physics of Intercalation, NATO ASI Series,
Plenum, New York, 1987, Vol. 172.
M.S. Whittingham, A. J. Jacobsen (Eds.) Intercalation Chemistry, Academic Press, New
York, 1982.
W. Miiller-Warmuth, R. SchOllhom (Eds.), Progress in Intercalation Research, Kluwer,
Dordrecht, 1994.
S. S. Prasad, in: Handbook of Solid State Batteries and Capacitors, M. Z. A. Munshi,
Ed., World Scientific, Singapore, 1995,467.
onaceous and Graphitic Anodes 181
RAPHITE-ELECTROLYTE INTERFACE IN
LITHIUM-ION BATTERIES
1 2 2
M.Nazri , B.Yebka and G.-A. Nazri
INTRODUCTION
The recent commercialization of advanced lithium batteries is mainly
the breakthrough in the stabilization of the anode-electrolyte interface.
gh metallic lithium has an energy density (3860 mAhlg) higher than
other alternative anodes, its poor performance and safety issues
to the low melting point of lithium (180 OC), dendritic growth during
deposition (charge), and high reactivity toward the electrolytes have
ed the commercialization of rechargeable lithium-anode batteries.
There have been several approaches to solve the problem of lithium
The development of lithium alloys, particularly the binary and ternary
has received considerable attention (see Chapter 9 of this book).
er, the performance of these alloys is unsatisfactory mainly due to the
olume change (100-200%) during lithiation and delithiation processes.
xpansion and contraction processes may cause the alloy particles to
nd lose contact with the electrode substrate. In addition, the lithium
with high concentration of lithium are very reactive toward the
lytes and cause decomposition. Therefore a problem similar to that of
tallic lithium also exists for Li-alloy anodes. Recently, some success
en made using intermetallic alloys such as Cu 6Sn5 that insert lithium
tically over a wide composition range LixCu 6Sn5 (0<x<l3).1 Despite
d volumetric energy density of these alloys, their gravimetric energy
is poor and there is a significant capacity loss during multiple
.
The early transition metal oxide anodes have also been proposed for
tion in rechargeable lithium batteries -in particular, the use of lithium
lated titanium oxide.2 -7 The lithium titanium oxide anode is described
Chapter6
Graphite has an ideal layered structure with weak van der Waals
g between the graphene layers. Thus, Li can be inserted between the
6 7
of graphite. According to recent Li- and Li- NMR studies, the nature
um in graphite is partly ionic. There is partial charge transfer from
to the n-orbitals of the graphene planes. The lithium in the graphite
so a high mobility, which is ideal for applications in high-power
s. In addition, the graphite has very high electronic conductivity,
electrode resistance is very low. Natural graphite is abundant, and
tic graphite also can be made from a variety of precursors. Heat
nt of various organic polymers, particularly those with conjugated
e rings, goes through a carbonization process around 1200-1500 OC
aphitization around 2300-3000 OC. During graphitization, the rate of
in a and c directions is strongly temperature dependent. Therefore,
e flakes with extended a or spherical with extended c directions can
de for different applications. The ideal graphite structure consists of
Graphite Types
Several natural and synthetic graphite samples commercially available
4, BG-35, CPC, CN-39, CN-39A, SFG-15) with different morphology,
.2. SEM morphology of natural graphites BG34, BG35, KS44, and synthetic graphite
CPC, CN-39, CN39A, and SFG-15.
-Electrolyte Interface in Lithium-Ion Batteries 199
size, and surface area have been examined. Figure 6.2 shows the
ology of these graphite samples.
raphite 2
ace area (m )
The surface area of the graphites is listed in Table 6.1. These samples
ow surface area and, so, may be suitable for battery applications.
e of the high reactivity of lithiated graphite, lower surface area
s are preferred.
The X-ray diffractions of the graphite samples listed above are shown
re 6.3. The high and narrow 002 peak is an indication of the good
nity of the samples.
/ 0
0
N
0
......
......
MCMB.10
l 1L ± A
BG·34 A
/ SFG-15 Jl /
L KS·44 A /
L N·39A A /
/ CN·39 A /
/ CPC J\. /
/ BG·3S A /
/
10 20 30 40 so 60 70 80
2 Teta (degree)
rbon
EPDM Sam pie
ack binder Carbon
powder
.1 .
Mixture
Micronizer
..
. 4.
Coa t ing
Electrode
.4. Coating process to make electrodes for lithium cells. Anode is metallic lithium, the
is the graphite electrode, and reference electrode is lithium wire
in the cell between two layers of porous separator.
e-Electrolyte Interface in Lithium-Ion Batteries 201
From the coated foil, electrodes are cut to make the lithium cells. Two
2
of cells can be constructed. A lOx10 cm electrode can be used to
the surface reactions and electrolyte decomposition using a GC-MS
. Figure 6.5 shows the structure of the cell used for measuring gaseous
s formed during electrolyte decomposition. The graphite electrode is
ted from the metallic lithium electrode by a porous polyethylene
opylene separator soaked with electrolyte. The cell is housed in a
hylene coated aluminum pouch.
2
A 5 cm disc electrode can be used for evaluation of capacity and cycle
sts. Figure 6.6 shows the assembly of this cell. The electrolyte
position occurs on the surface of graphite during initial lithium
lation. In order to investigate the nature of the evolving gas, a special
sembly is constructed to allow the generated gas to escape from the
d be sampled by in-situ GC-MS analysis. Before sealing the cell
in Figure 6.5, a flexible 2 mm diameter tube is inserted in the cell. The
s connected to an on-line GC-MS system. Figure 6.7 shows the
ned GC-MS equipment used for in-situ analysis of the gaseous
ts.
Chapter6
Anodc(CE)
Cathode (WE)
CurTcnt collector
.6. Schematic of lithium cell for evaluation of graphite anodes. WE= working
e, CE =counter electrode andRE= reference electrode.
PLOTTER
.7. Schematic of on-line GC-MS for analysis of gases generated during lithium
tion in graphite anodes due to electrolyte decomposition.
e-Electrolyte Interface in Lithium-Ion Batteries 203
-0.5
·1.11
· .. :.:.:.:.... . . .
3.0 ... :k .-··
... -1.5
r\ ·.....,
.".c 1.5 r=
·0 10m0 ,._, 0.6 0.7 0.8 0,9 1.0
1.1 1.2
=
u"
=I Cell voltage (volts)
.9. The voltage profile of graphitic electrodes (BG-35, CN-339, CPC, BG-34, CN-
d KS44).
3.5 S2/CPC
3.0
2.5
V) EC-DEC-LiPF6
2.0
PC-LiPF6
1.5
EC-PC-LiPF6
1.0
0.5
0.0
0 100 200 300 400 500 600
Q (mAh/g)
apacities close to theoretical one (372 mAhlg) are achievable . The rate
ch the electrode can be charged depends also on the amount of active
als loaded on the substrate. For the design of a high rate (power)
0
0 10 20 30 40 50
Time (hours)
, the coating is thinner that for the electrode designed for high
applications. The usual loading for high power battery is about 10-20
2
and for the high-energy cells the loading may exceed 100 mg/cm2.
Disordered carbons have also been used as anodes. However, their
ty is almost half that of the graphitic anodes, and their voltage profile
lithium intercalation does not show the staging processes often
ed for the graphitic anodes. Some special carbon anodes with extended
layer graphene sheets, store lithium to a higher level than the
tical capacity. In this case, the lithium is being adsorbed on both sides
graphene sheets and in micropores. However, the cycle life of this type
bon is poor and there is significant capacity loss during multiple
-discharge cycles. In addition, there is a considerable amount of
e hysterises between charge and discharge. The sloping charge
rge profile also suggests that there are various sites for lithium with
nt affinities.
e-Electrolyte Interface in Lithium-Ion Batteries 207
Graphite Electrolytes
BG-35
CPC
CN-39
KS-44
SFG-15
BG-34
BG-34
; &: DMC-LiPF•
Wavenumber (cm"1)
.12. IR spectrum of the film formed on the surface of graphite anode during
intercalation in (DMC + 1M LiPF6) electrolyte.
1.0 .------.:---------------------:J
BG-34
EC-DMC-LiPF6
0.8
0.6
0.4
0.2
o
2000 1800 1600 1400 1200 1000 800 600 400
Wavenumber (cm-1)
6.13. IR spectrum of the film formed on lithiated graphite during the first lithium
ation process in EC-DMC (1: I mol) + 1MLiPF6 electrolyte.
Assignment
observed (cm'1) v assigned bands Compounds
1958vw
1866w 1870-1750 Organic Carbonate
1870 Ethylene Carbonate
1780sp 1870-1750 Organic Carbonate
1760 Dimethyl Carbonate
1737sh 1870-1750 Organic Carbonate
1638sh 1629 LiPF/
1566sh 1580 0-C-0 stretch RCOOLi
1483sp CH2 bend
1414sp 1441 LhCOJ
1420 LhCOJ
1369sh 1380-1360 CHbend LiOCH3
1198br 1170-1150 CH stretch LiOCH3
1083sp 1089 CO stretch Li,C03
1085 CO stretch RC03Li
1074 Ethylene Carbonate
1083 LiPF.'
977m 974 Ethylene Carbonate
909s
844br 846 C03 bending Li,C03
858 Dimethyl Carbonate
850 LiP
832 LiPF.'
780sp 745-789 LiO
774 Ethylene Carbonate
The peak assignments for SEI layer formed on the surface of graphite
the first lithium intercalation are listed in Table 6.5. The peaks
pond to the formation of residual lithium carbonate with peaks around
1
cm- , formation of various organa-lithium compounds as well as peaks
C=O stretching vibrations at 1650-1800 cm-1 •
The IR spectrum of the film formed on the surface of graphite anode
lithium intercalation in PC-LiPF6 solution is shown in Figure 6.15,
e frequencies and products in Table 6.6. This spectrum is rich in by
cts of the PC decomposition. The IR signals of LizC0 3, around 850 and
1
cm- have also been observed. As already mentioned, the film formed in
se is much thicker than the film formed by EC decomposition, this
ting further electrolyte decomposition upon cycling.
Chapter 6
EC-DEC-LiPF6
3.0
25
05
Wavenumber cm-1
.14. IR spectrum of the film formed on BG-34 in EC-DEC + IMLiPF6 during first
n.
(PC-LiPFti)
-O.Il
-O.Il5
-O.'l
-0.'15
Wavawmher ao. _,
Assignment
Observed vassigned bands compounds
1966s
1870 Eth 1ene Carbonate
1809s
1771sE 1750-1870 Orsanic Carbonate
1746m 1750 Diethi:1 Carbonate
1699w
1685w
1652m 1650 C02 stretch RC03Li
1555sp 1580 RCOOLi
1580 LiOH
1505sh 1500 (CHzOLi)z
1479w Eth lene Carbonate
1457w 1450-1480 CHbend LiOCH3
1445 CH, CH3 asy. Ethyl-C03Li
bend
1390w Ethylene Carbonate
1360-1380 CHbend LiOCH3
1374w 1351 COz s m. stretch Eth l-C03Li
1296sh
1267sE 1262 Dieth,rl Carbonate
1195sh CO stretch
1161sp 1162 Ethylene Carbonate
1152 CO stretch Eth,rl-C03Li
1072sp 1074 Ethylene Carbonate
1085 CO stretch Ethyl-C03Li
1100 CO stretch (CHzOLi)z
1017sp 1021 Diethyl Carbonate
1010 CO stretch Eth l-C03Li
973m 974 Eth lene Carbonate
903sh 890 (CHzOLi)z
844sp 858 Diethyl Carbonate
826 co3 bend Ethyi-C03Li
850 LiOH
774sE 774 Eth lene Carbonate
730sh 738
717m 720 C02 as m. bend Eth I-C03Li
666w 660-550 LiO stretch LiOCH3
556sE 500-600 LiO stretch Eth I-C03Li
529w 400-500 LiOH
(w: weak, m: medium, s: strong, sh: shoulder, br: broad, v: very, sp: sharp)
Assignment
observed (cm'1) v assigned bands Compounds
1862sh
1817s
1785sh 1760 Dimethyl Carbonate
1732w
1715br
1630w 1629 LiPF6
1588sh 1580 RCOOLi
1500br
1430br 1441 LhC03
1360br COz sym.
stretch
1338br
1289br 1280 Dimethyl Carbonate
1250w 1263 LiPF6'
1233w
1134m 1122 LiPF6
1088m 1089 CO stretch LhCOJ
1083 LiPF6'
1065w 1060 LhO
960w 969-914 Dimethyl Carbonate
865sp 866 C03 bend LizC03
832 LiPF6'
870 LhO
866 LiO
791sp 793 Dimethyl Carbonate
789-745 LiO
734br 738 COz asym. bend LhCOJ
69lsh 730-689 LiO
562sp 557 LiPF6
650-550 LiO stretch LiOCH3
(w: weak, m: medium, s: strong , sh: shoulder, br: broad, v: very, sp: sharp)
e-Electrolyte Interface in Lithium-/on Batteries 215
CPC
PC-DMC-LiPF
I
0
":!':
:;::
:!:
l::i:!:
...
6.16. IR spectrum of the film formed on the surface of graphite anode during lithium
ation in PC-DMC (l:lmol) +1M LiPF6.
3.0 LiC6
X
X=O.S
2.5
BG-34 X=0.6
2.0
X=0.4
1.5
X=0.2
1.0
0.5
0.0
Before cycling
-0.5
.17. Differential scanning calorimetry of the surface film formed on graphite anode at
states of charge (x=0.2-0.8 in LixC6).
The sample has an endothermic peak near the boiling point of DMC,
n the exothermic process starts around 150 OC. A major exotherm was
ed at 220-240 OC. The intensity of the exotherm peak correlates well
e amount of lithium remaining in the graphite. It is very important to
the safety aspects of the Li-ion cells when they are exposed to
atures above 150 oc. Further electrolyte decomposition with massive
neration may occur above 150°C, and cell temperature may rise to
e-Electrolyte Interface in Lithium-Ion Batteries 217
t'iO
50
i .w
; 3:0
20
10
•-
0 2
" Surface area (m 'fg)
6 ll 10 12 14
.18. The correlation between the volumes of gas evolved during lithium intercalation
te anode and the surface area of the graphite. Upper line: PC-DMC; middle: EC
wer: EC-DEC (all with LiPF6).
CONCLUSIONS
he following general conclusions can be made from this study.
Organic carbonates are not electrochemically stable over a lithiated
raphite electrode. However, the decomposition by-products formed on
he surface can protect the electrode, and the electrolyte decomposition
s eliminated in subsequent charge-discharge cycles.
The nature of the film formed on the surface of lithiated graphite
epends on the composition of electrolyte, and in the case of
Chapter 6
ERENCES
M. Thackeray, J.T. Vaughey, A.J. Kahaian, K.D. Kepler, R. Benedek, Electrochem.
mm. 1( 1999) 111.
Akimoto, Y. Gotoh, Y. Oosawa, J. Akimoto, Y. Gotoh, Y. Osawa, J. Solid State
emistry 7 (1997) 129.
. Kartha, D.P. Tunstall, J.T.S. Irvine, J. Solid State Chemistry 152 (2000) 397.
E. Arroyo de Dompablo, A. Varez, F. Garcia-Alvarado, J. Solid State Chemistry 153
00) 132.
van de Krol, Thesis, 2000, Universal Press, Veeneddaal, Netherland.
Sodergren, H. Siegbahn, H. Rensmo, H. Lindstrom, A. Hagfeldt, S.E. Lindquist, J.
ys. Chern. B 101 (1997) 3087.
W. Murphy, R.J. Cava, S.M. Zahurak, A. Santoro, Solid State Ionics 9 (1983) 413.
Ohzuku, A. Ueda, N. Yamamoto, J. Electrochem. Soc. 142 (1995) 1431.
Nishijima, Y. Takeda, N. Imanishi, 0. Yamamoto, J. Electrochem. Soc. 141 (1994)
66.
Shodai, S. Okada, S. Tobishima, J. Yamaki, Solid State Ionics 86 (1996) 785.
Suzuki, T. Shodai, Solid State lonics 116 (1999) 1.
.C. Rowsell, V. Pralong, L.F. Nazar, J. Am. Chern. Soc. 123 (2001) 8598.
Pereira, L.C. Klein, G.G. Amatucci, J. Electrochem. Soc. A262 (2002) 149.
Souza, V. Pralong, A.J. Jacobson, L.F. Nazar, Science 296 (2002) 2012.
Pralong, D.C. Souza, L.F. Nazar, Electrochem. Com. 4 (2002) 516.
Nagaura, K. Tozawa, Prog. Battery and Solar Cells 9 (1990) 209.
. Holmes, A.R. Landgrebe, The Electrochemical Society Proceeding 91 (1997) 18.
Ohzuku, A. Ueda, J. Electrochem. Soc. 144 (1997) 2780.
te-Electrolyte Interface in Lithium-Ion Batteries 219
boratoire de Reactivite et Chimie des Solides, Universite de Picardie Jules Verne and
CNRS (UMR-6007), 33 rue Saint Leu, 80039, Amiens, France
INTRODUCTION
In response to the needs of today's mobile society and the emergence
cological concerns such as global warming, one of the major
nological challenges in this new century is undoubtedly energy
ration and storage. Ninety percent oftoday's electrical power generation
comes from fossil fuels, and we are constantly struggling to reduce the
on dioxide emissions per unit of electric power so as to help curtail
al warming. It is now mandatory that new and environmentally friendly
gy/storage sources be found. Hence, the fast developing research in that
involving, among others, fuel cells, primary and rechargeable batteries,
supercapacitors. As a result of this worldwide ecological priority,
ical concerns have come into play, and science has suffered from
itisation based on both industrial pressure and media reports, rather than
he clear and rigorous scientific identification of technological stoppers
ent in each storage system. Needless to say, this applies to battery
ms as well.
In the past two decades, intensive efforts have given birth to the
argeable Li-ion battery technology that has dominated the market place,
an be regarded as one of the great successes in modern electrochemistry
te. But these Li-based systems still suffer from the lack of suitable
rode and electrolyte materials, which they require if they are ever to
mmodate the increasing user's demands. Aware of this limitation,
ists have been acting at several levels to incrementally improve the Li
erformance. They have followed a dual approach, dealing with either
ive or negative electrode materials, with efforts centered around: 1) the
fication of existing materials through cationic/anionic substitution,
re modification and surface treatments, 2) the making of composite
Key Role ofNanoparticles in the Reactivity of 3d Metal Oxides Towards Li 221
EXPERIMENTAL
The Co-based metal oxide powders used within this study were
mercial products (CoO), or were synthesized in-house (Co304, Mn304,
203). The other metal oxides (M= Ni, Fe, Mn, Cu) were used as
ved from commercial sources, unless otherwise specified. LiCo02 was
provided by the market.
Powder purity and crystallinity were examined by X-ray diffraction
D), operating in Bragg-Brentano geometry with a CuKa radiation. The
ders morphology and composition were investigated by scanning
ron microscopy (SEM) with a field emission gun (FEG), coupled to an
ument for energy-dispersive X-ray spectroscopy (EDS). High-
Chapter 7
3.5
C/2
--
3
_
2.5
-
j
V)
> 2
>
1.5
0
1
c
+- 1
0
> 0.5
0 3.5
0 0.5 1 1.5 2 2.5 3
x in "LiXCoO"
e 7.1. Voltage-composition traces for a CoO!Li cell cycled between 0.01 V and 3 a
rate of C/2 (e.g., 1 lithium per formula unit in 2 hours) shown over several s.
Capacity retention for the same cell shown in the inset.
3.5
3
0 Q
5'2.5
r:
-0.05
2 -0.1
0.5
-0.15
0 0.5 1 1.5 2 2.5 3 3.5
0 x in "Li CoO"
X
e 7.2. PITT measurements performed on a CoO!Li cell. The voltage steps are of
V with a current limit Cimin) corresponding to C/300 (e.g., 1 Li in 300 hours).
Key Role of Nanoparticles in the Reactivity of 3d Metal Oxides Towards Li 225
e 7.3. TEM observations realized on CoO electrodes recovered from CoO!Li cells that were
ed at various discharged and charged states, as indicated by the full circles on the
ge/composition curves. When necessary, the SAED patterns are joined to the micrograph as
ans to indicate the electrode composition. The nanometric character of the composite
her with the growth/disappearance of a polymeric layer are nicely depicted.
Whatever the a or forms, all the above results convey the message
the reduction of CoO through the first discharge plateau is reversible
rding to the following reaction.
(7.1)
=
hermodynamic data - given in the literature for massive compounds,
an deduce for CoO a .:l,G -347 kJ/mol that should correspond to a
au located at Eeq = 1.798V vs Li+/Li • Similar calculations can be
0
36 37
nded to most of the 3d metal oxides, - , with the main result being that
tabulated Gibbs free energy values are negative. This result
mbiguously proves that such oxides are, under these conditions,
aneously reducible at Eeq values, as reported in Figure 7.4.
Key Role of Nanoparticles in the Reactivity of 3d Metal Oxides Towards Li 227
o Bard data
3.5
.---.. • Handbook data ·························· ······················i····················· 3
.
s
3 • Kubaschewski data
+.....
.
2.5 ------------------·--r--·-------------------r ---------------------!·----------------------r·----- 2.5
r.n
> 2 ·····················i····················+······················ -- --······ -- ----····................ 2
>'-" i aj •••f • :•oo
Q.)
1.5
.. .
.. . .... ..
0 1
> ··••••••••••••••m ••••• •• •••··••r•:••••••••••••••••r•••••••••••••••• :•••••••••••• ··· 0.5
:·'
.
- --- ---··•··-i·-·····················i······················"·······················f·····················
0.5 '
e 7.4. The equilibrium voltage values of MO < -- > M conversion reaction are
ted for various oxides. The values were obtained by first determining the reaction
nergy (dG) and then by applying the Nernst law.
3.5 0.1
Eeq(Mn0/Mn°)=1,03V
+' 3 Eplateau"'0.2V/Li 0.05
2.5
CZl
> 2
0
-0.05
§g
>
1.5
.......
-0.1
..... 1 -0.15 '--'
0.5 -0.2
e 7.5. The voltage composition curves as deduced from PITT measurements for
Li cells cycled between 0.02 V and 3V. The voltage steps were of 10 mY with a t
limit corresponding to C/300.
0 0.5 1 1.5 2
Scan potential (Volts)
e 7.6. Cyclic-voltammetry data as a function of the scan rate for a CoO!Li cell that
ycled between 0.02 and 1.8 V. One inset shows the linear variation of the current as
ction of the scan rate. The others, a, b, and c are micrographs of CoO electrodes
ered from CoO!Li cells that were fully discharged and fully recharged 10 times
en 0.02 and 1.8 V, and fully recharged to 3 V, respectively.
2000
1800
1600
1400
-'
;
1200
d 1000
. 800
600
;
) 400
..
) 200
0
0 50 100 150 200
Cycle number
e 7.7. Variation of the capacity retention of CoO/Li cells, both as a function of
mperature and of the electrode content (e.g., with or without carbon). Effect of
rature sweep between 20°C and 75°C on the capacity values shown as an inset.
Key Role of Nanoparticles in the Reactivity of 3d Metal Oxides Towards Li 231
First, common to all the cells once the first ten cycles were achieved,
ote a continuous increase in the cell capacity that was more pronounced
increasing temperature. At 75°C, a doubling of the initial capacity was
hed after 200 cycles. Such a capacity was found to be reversibly
perature dependent, since a decrease in temperature resulted in a capacity
ease, which is fully recovered as the temperature return to its original
e39 • Worth mentioning is that more than 500 cycles were achieved
out any evidence of short-circuit, reminiscent of dendrite formations.
hermore, on cells opened after 500 cycles, we noted that the polymeric
r has expanded through all the separator, suggesting at first that the
meric layer could not be a good electronic conductor, without which the
would have a short. The feasibility of cycling CoO/Li cells at 75°C at a
ate for more than 400 cycles is somewhat amazing, bearing in mind
these are the best conditions to favour dendrite formation. This can be
ained either by: 1) the fact that the polymer film formed on the CoO
rode prevents dendrite penetration, and by the same token, cell short or
y a modification of the Li surface morphology, due to the deposit of
rolyte degradation products, to the point that it cancelled dendrite
ation. Needless to say these facts could help in engineering better
ectrolyte interfaces so that rechargeable Li batteries could finally
me a commercial reality.
Finally, by conducting cycling measurements with different discharge
ff voltages (Figure 7.7, inset), we reached the conclusion that this
erature-enhanced capacity was the result of electrochemical processes
g place at voltages close to lithium (e.g, over the 0.02-0.15 V).
iously, this result could simply be explained by assuming a temperature
nced electrolyte degradation that would lead to an electrochemically
e polymer. Therefore the amount of this polymer would increase, the
city would increase, and should remain constant whatever the cycling
erature once the film is formed. This is not as simple, since we
rved that the capacity decreases once the temperature is lowered. The
that the temperature-driven extra capacity is located near 0 V vs. Li+/Li 0,
ates that one needs to decrease the cell polarization by rising the cell
erature to tap into this low potential extra capacity. Thus, the capacity
g observed with changing the temperature is the result of polarization
ts, most likely associated to a limited ionic or electronic conductivity of
rown polymeric layer.
To further explain the cycling data at 70°C, we plotted the derivative
es (Figure 7.8) as a function of the cycle number. We noted the
arance of a 1.2V peak corresponding to the Co07Co conversion
ess, which disappeared upon cycling at the expense of a 0.8 V peak.
implies that, upon long cycling, the Co07Co conversion process
lly corresponding to the main part of the overall cell capacity vanished.
Chapter 7
0.15V-3V
2.5 3
0 0.5 1 1.5 2
U(Vvs LVLi)
7.8. Derivative dx/dV plots as a function of the cycle number for CoO/Li cells.
ells were cycled at 75°C at a C rate. Two scanning ranges were used, 0.02 to 3 V
.15 to 3 V. Note the appearance of a 0.8 V to the expense of the 1.2 V peak
ponding to the MO < -- > M conversion process, and of a large extra capacity
d around 0.1 V.
140 30 discharges
120
0
U100
4-1
0 80 30 charges
60
::E 40
20
0 ---- ----
4
----
4
---- 4----4 -- 4
0 110 2 10 3 10 4 4 10 5 10 6 10
FIELD (Gauss)
e 7.9. Variation of the room temperature remnant magnetization for fully charged
ischarged CoO electrodes recovered from CoO/Li cells that were cycled 30 and 100
, respectively, at a C rate over a 0.02 to 3 V voltage range.
14
O, FeO, MnO, CuO), such as 3d metal sesquioxides (Co304), Li-based
27
ary oxides (LiCo02, LiNi02, ...), 3d metal sulphides (CoS) , or nitrides
4041
3N), were found to behave identically, leading to the
ation/decomposition upon cycling of Li2S and LhN, respectively,
ther with the formation of a polymeric gel-like layer. Similarly, there
recent reports describing the potential reactivity of 3d metal-based
tes or phosphates in similar terms, thus suggesting the universality of
a proposed mechanism.
(b)
4 ,. I
Output Cell '
Voltage , 1
---"--,, I
0
CZl
........
3 ----
,.,I
>
"-"'
2
(!.)
OJ)
C\$ 1
........
0
> 0
0 2 4 6 8 10
x in LixCo 304
0.2
\ Co3b4 /LiCo02
1 I I I I
. 0.15
:_l
<. 0.1
c::
"-"'
;>..
......
u
C\$
0.05 Rate: C/5
u Mr=8.5 T = 25°C
C\$ I I I I I I
0
0 10 20 30 40 50 60 70 80
Cycle number
re 7.10. The first cycle of a LiCo02/Co304 cell cycled between 4.1 and l.IV, as
ured by means of three-electrode configuration, is shown in (a). The cell output
ge is of about 2 V and the weight ratio (M,) of positive to negative equals 8.5. The
city retention for such Li-ion cell is shown in (b).
ains, and in the same way the domain wall boundaries that act as a
er layer, is crucial to the performance of this electrode, since, upon
sequent cycling, the so called "Ni(OHh formation cycle" remains a
able know-how specific to each battery manufacturers. This situation
Chapter 7
100
90
50 750
150 250 350 450 550 650
Temperature COC)
Co304 : 250°C
2000
Co304 : 800°C
BeO A
BeO + BeO coJ04
Co3 Co304
1750 BeO
1500
1250
1000
ea_n
(.)
750
500
250
re 7.12. In situ X-ray data for CoO/Li cells discharged at very low current drains
2)/ close to equilibrium at a C rate. The cells contained either Co0250 or Co0800
rodes .
4.2. a-Fez03
- Ill 300
-
c.
0
200
100
Chapter 7
>
:!:::
0
-
Ill
I::
Q)
1200
I:: 800
400
30 40 50 60 70 80
Scattering Angle (deg.) CoKo:
e203• While structurally similar, these two samples strongly differ in the
h of their Bragg reflections. From the Scherrer's equation, we estimated
mean crystallite size at about 200 A for n-Fe203, and at least 1000 A for
e203• By TEM, the lozenge-shaped n-Fe203 particles were found to be
olithic and very small (200 A), in total agreement with the crystallite
deduced from XRD analysis. M-Fe203 particles are much larger (0.5
in mean diameter), as shown by SEM images. The measured BET
2 2
ific surface areas are 60 m /g and 2 m /g for n-Fe203 and M-Fe203•
The first electrochemical discharge curves collected at C/5 rate for the
hematite samples previously described are presented in Figure 7.14.
n discharged down to 0 Volt, 7.2 Li per M-Fe2 03 and 8.3Li per n-Fe203
be reacted, corresponding to capacities of 1200 and 1400 mAhlg. Both
harge curves show a nicely developed and long voltage plateau lying
w 1.0 Volt, until an x value of about 6 (x in "LixFe203"), in agreement
3 0
the theoretical capacity for the Fe + 7 Fe reaction, is reached. The
r excess in capacity measured for n-Fe203 is consistent with its much
r specific surface area, when compared to that of M-Fe203•
Key Role of Nanoparticles in the Reactivity of 3d Metal Oxides Towards Li 241
400Capacity (mAh/g)
-·
200 600 800 1000 ' 1200 1400
30
s. 2.5
r:l.l
2
c 1.5
s
1
Oll
0 0.5
0
0 1 2 7 8
e 7.14. First discharge for a-FezO/ Li cells, using bulk or nano a-Fez0 3 as positive
ode, together with, as insets, the open circuit voltage measurements for similar cells.
However, the most relevant part of the curve, with respect to the
ortance of nano particles, lies at low x values. Indeed, note that the onset
he 1 V plateau is very quickly reached for M-Fe203 (x = 0.1), while it
rs at x=2 for n-Fe203, suggesting a different Li reactivity mechanism
een the two samples. To better isolate such differences, PITT
riments were conducted on n-Fe203 and M-Fe203 up to x=2. Within this
position domain, the reduction of n-Fe203 clearly proceeds in two steps.
, a monophasic process characterized by a continuous decrease in the
i-equilibrium OCV is observed up to x close to 1, and then a biphasic
tion characterized by a constant equilibrium voltage close to 1.85 Volts.
The Li reactivity difference was well diagnosed by in situ X-ray
surements (Figure 7.15), with a continuous shift of the hematite Bragg
s towards low angles for the n-F 03 sample, as compared to a
ressive vanishing of the hematite Bragg peaks for M-Fe203 reflections,
he expense of a new set of reflections. These extra peaks, that were
xed within a cubic symmetry having an axis of 8.422 A, were found to
espond to a Li-Fe-0 rock-salt compound described by Thackeray et al.50
not to a FeO rock-salt phase (as confirmed by EXAFS and Mossbauer),
should have an a axis of 4.31 A (e.g., 2a = 8.62 A).
Chapter 7
12000
00oo ::o--- "-'
10000:;
/1.\"---v
o:- - :::.
In
8000
-n
0ooob
6000 .!!
37 38 39 40 41 42 43 37 38 39 40 41 42
Scattering Angle (deg.) CoKu Scattering Angle (deg.) CoKu
e 7.15. in situ X-ray data for a-F 03/ Li cells, using either bulk or nano a-F 03
ositive electrode. For reasons of clarity, only part of the pattern (33°
< 43°) depicting the evolution of the 104 and 110 Bragg peaks as a function of the
nt of reacted Li is shown. Note a disappearance of these peaks for the bulk sample,
mpared to a shift for the nano-materiaL
. _-·r-0('1
- - "o .
0 0.
1om
8 --r--------r
45
c::::::
....- 4
, 3.5 -
-
e" il o.
0.
eo L
40.:;,
o..o. --
.a d
OJ)
r JJ -
>
0.1 201;!_
3_ u oL- -- - ---'---' €
> 2.5 -
'""-'
8 12
Cycle number
18 28
2-
1.5-
1
0
>o.5 0 0.5 1
x in LixFe203
,...-._ 5
•
+.-<
5.....:l
4.5 n-Fe203
4
rJJ
> 3.5
>
'""-'
3
OJ) 2.5
ro
,.-.+..-.>.; 2
0
> 1.5 0 0.5 1
x in LixFe203
re 7-16. Voltage-composition profile for o:-F O/ Li cells, having bulk (upper) or nano
er) o:-F 03 as the positive electrode. Their capacity retention upon cycling is shown by
nsets.
es, nitrides, fluorides and sulphides, and another family of materials like
rmetallics. For instance, our group recently reported that the reactivity of
b3 towards Li is reversible. Although Co-Sb bonds are broken during the
harging process, leading to a full disintegration of the CoSb3 structural
ework into Co and LhSb nano particles, we show that upon recharge
e destroyed bonds, due to the high reactivity of the divided metal nano
icles, reformed and arranged to convert back to the initial CoSb3
51 2
cture Such an example, also observed with Cu6Snl , further contrasts
previous beliefs that bond-breakings were prohibited for reversible Li
tivity.
CONCLUSIONS
KNOWLEDGMENTS
The authors would like to give special thanks to L. Dupont, M. Dolle,
ayen, F. Badway, and I. Piitz for their contribution to some of the
ented data.
ey Role of Nanoparticles in the Reactivity of 3d Metal Oxides Towards Li 245
FERENCES
M.B. Armand, New Electrode Materials, in Fast Ion Transport in Solids (Van Goo!, W.,
Ed., North Holland, Amsterdam) (1973) 665.
M.S. Whittingham, Science 192 (1976) 1226.
D. W. Murphy, P.A. Christian, Science 205 (1979) 651.
A. R. Armstrong, P.G. Bruce, Nature 381 (1996) 499.
B. Ammundsen, J. Desilvestro, T. Groutso, D. Hassel, J.B. Metson, E. Regan, R. Steiner,
P.J. Pichering, ECS Fall Meeting, Hawai, Abstract N"138 (1999).
A.K. Padhi, K.S. Nanjundaswamy, C. Masquelier, J.B. Goodenough, J. Electrochem.
oc. 144 (1997) 1609.
Y. Idota, T. Kubota, A. Matsufuji, Y. Maekawa, T. Miyasaka, Science 276 (1997) 1395.
.A. Courtney, J.R. Dahn, J. Electrochem. Soc. 144 (1997) 2045.
0. Mao, R.A. Dunlap, J.R. Dahn, J. Electrochem. Soc. 146 (1999) 405.
K.D. Kepler, J.T. Vaughey, M.M. Thackeray, Electrochem. Solid-State Lett. 7 (1999)
07.
N. Pereira, L.C. Klein, G.G.Amatucci, J. Electrochem. Soc. 149 (2002) A262.
R. Alcantara, F. J. Fernandez-Madrigal, P. Lavela, J. L. Tirado, J-C. Jumas, J. Olivier-
ourcade, J. Mater. Chern. 9 (1999) 2517.
V. Pralong, D.C.S. Souza, K.T. Leung, L.F. Nazar, Electrochem. Comm. 4 (2002) 516.
. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.-M. Tarascon., Nature 407 (2000) 496.
G.G. Amatucci, J.M. Tarascon, L.C. Klein, J. Electrochem. Soc. 143 (1996) 1114.
. Bates, Y. Jumel, Lithium batteries J. P. Gabano Ed., Academic Press, London (1983).
Y. Matsuda, K. Teraji, Y. Takasu, Denki Kagaku 44 (1976) 363.
. Novak, Electrochim. Acta 30 (1985) 1687.
. Sigala, D. Guyomard, Y. Piffard, M. Tournoux, C.R. Acad. Sci. Paris Il320 (1995)
23.
. Poizot, E. Baudrin, S. Laruelle, L. Dupont, M. Touboul, J-M. Tarascon, Solid state
onics 138 (2000) 31.
. Baudrin, S. Denis, F. Orsini, L. Seguin, M. Touboul, J-M. Tarascon, J. Mater. Chern. 9
1999) 101.
. Denis, E. Baudrin, F. Orsini, G. Ouvrard, M. Touboul, J-M. Tarascon, J. Power
ources 81-82 (1999) 79.
. Baudrin, S. Denis, S. Laruelle, M. Touboul, J-M. Tarascon, Solid State Ionics 123
1999) 139.
. Laruelle, P. Poizot, E. Baudrin, V. Briois, M. Touboul, J-M.Tarascon, J. Power
ources 97-98 (2001) 251.
. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J.-M.Tarascon., Nature 407 (2000) 496.
. Poizot, S. Laruelle, S. Grugeon, L. Dupont, B. Beaudoin, J-M. Tarascon, C.R. Acad.
ci. Paris II, 3 (2000) 681.
. Poizot, S. Laruelle, S. Grugeon, L. Dupont, J-M. Tarascon, Ionics 6 (2000) 321.
. Grugeon, S. Laruelle, R. Herrera-Urbina, L. Dupont, P. Poizot, J-M. Tarascon, J.
lectrochem. Soc. 148 (2001) A285.
. Dehart, L. Dupont, P. Poizot, J-M. Tarascon, J. Electrochem. Soc. 148 (2001) A1266.
. Poizot, S. Laruelle, E. Baudrin, S. Denis, M. Touboul, J.-M. Tarascon, J. Power
ources 97-98 (2001) 235.
N.N. Obrovac, R.A. Dunlap, R.J. Sanderson, J.R. Dahn, J. Electrochem. Soc. 148 (2001)
576.
M.M. Thackeray, W.I.F. David, J.B. Goodenough, Mat. Res. Bull. 17 (1982) 785.
.M. Thackeray, W.I.F. David, J.B. Goodenough, J. Solid-State Chern. 55 (1984) 280.
Chapter 7
Standard Potentials in Aqueous Solution, A.J. Bard, R. Parsons, J. Jordan, Eds., Marcel
Dekker (1985).
0. Kubaschewski, C. B. Alcok, Metallurgical Thermochemistry, 5th Edition, Pergamon
Press (1987)
S. Laruelle, S. Grugeon, P. Poizot, M. Dolle, L. Dupont, J-M. Tarascon, J. Electrochem.
Soc. 149 (2002) 627.
P. Poizot, S. Laruelle, S. Grugeon, J-M. Tarascon, J. Electrochem. Soc. 149 (2002) 627.
M. Dolle, P. Poizot, L. Dupont, J-M. Tarascon, Electrochem. Solid-State Lett. 5 (2002)
A18.
S. Grugeon, S. Laruelle, J.M. Tarascon, J. Electrochem. Soc. submitted (2002).
N. Pereira, L.C. Klein, G.G. Amatucci, ECS and ISE Joint International Meeting, San
Francisco, CA, Sept. 2001, paper 203.
N.Pereira, L. Dupont, J.-M. Tarascon, L. Klein, G.G Amatucci, J. Electrochem. Soc. (in
press).
F. Badway, I. Piitz, S. Grugeon, S. Laruelle, M. Dolle, A.S. Gozdz, J-M. Tarascon,
Electrochem. Solid-State Lett. 5 (2002) A115.
A. Delahaye-Vidal, B. Beaudoin, M. Figlarz, Reactivity Solids 2 (1986) 223.
W.I.F. David, J.B. Goodenough, M.M. Thackeray, M.G.S.R. Thomas, Rev. Chim. Miner.
20 (1983) 636.
M.M. Thackeray, W.I.F. David, J.B. Goodenough, Mat. Res. Bull. 17 (1982) 785.
M.M. Thackeray, W.I.F. David, P.O. Bruce, J.B. Goodenough, Mat. Res. Bull. 18
(1983) 461.
M.M. Thackeray, S.D. Backer, K.T. Adendorff, Solid State Ionics 17 (1985) 175.
D. Larcher, G. Sudant, J-B. Leriche, Y. Chabre, J-M. Tarascon, J. Electrochem. Soc. 149
2002) A234.
D. Larcher, C. Masquelier, D. Bonnin, Y. Chabre, V. Masson, J-B. Leriche, J-M.
Tarascon, J. Electrochem. Soc. submitted (2002).
M.M. Thackeray, J. Coetzer, Mat. Res. Bull. 16 (1981) 591.
.-M.Tarascon, M. Morcrette, L. Dupont, Y. Chabre, C. Payen, D. Larcher, V. Pralong, J.
Electrochem. Soc. in press (2003).
D. Larcher, L.Y. Beaulieu, 0. Mao, A.E. George, J.R. Dahn, J. Electrochem. Soc. 147
2000) 1703.
apter 8
Telcordia Technologies, 331 Newman Springs Road, Red Bank, NJ 07701, U.S.A.
INTRODUCTION
The search for negative electrode materials which exhibit improved
rochemical properties relative to the widely used graphite intercalation
rials has lead to a renaissance of research on inorganic negative
rodes.
The intercalation of Li+ into graphite follows the basic topotactic
tion process:
(8.1)
Beyond basic binary lithium alloys, a large focus has been on the
lopment of conversion oxides. The conversion oxides MexOy, as
ussed in a previous chapter, consist of metals (Me) which can alloy with
um. Upon the first lithiation, an electrochemically induced conversion
ess occurs creating a nanostructure consisting of nanodomains of Me in
trix of nonstructured Li2 0 (8.2):
(8.4)
The quest for new inorganic anodes has recently proceeded into areas
ch have been largely ignored, namely, nontraditional MexXy materials,
re the anion X is N or P. The number of papers on these type of
erials is relatively small, but nonetheless the electrochemistry is found to
urprisingly rich and somewhat exotic. In this chapter, we take a close
at the work that has been done to this point on binary and ternary metal
des along with a look at the often overlooked metal silicides.
. THE NITRIDES
Metal nitrides have been investigated for their properties related to fast
um ion conduction for the past two decades. The most famous of these
des is the highly conducting Li 3N 1- 7 which exhibits ionic conductivity on
2
order of 1'10- S/cm. A number of papers have been published on this
rial; however, because of its poor resistance to anodic decomposition at
potentials, it has found little use as a solid state ionic conductor in a
tical cell. In spite of this drawback, the initial work spurred interest in
nvestigation of alternative nitrides that could also support appreciable
conductivity as a result of the low polarizability of the nitrogen anion.
y of these structures are related to Li2x_ 1MNx which is based upon the
luorite structure. Containing pathways for facile ion movement, these
rials exhibited appreciable Li+ conductivity. Materials such as Li 3BN2, 8
iN3,9 •10 and Li 3AlN2 all share this structure type and have well
lished ionic activity. Due to their electronically insulative properties,
materials were investigated solely for use as solid state ionic
uctors.
The facile movement of the lithium ions within the antifluorite
ture gave rise to the question whether such structures could be utilized
radaic insertion intercalation electrodes for lithium batteries. This issue
addressed by the use of transition metals as a substitute for M in Lizx-
x. The resulting Lizx_ 1MNx structure gained appreciable electronic
uctivity to parallel the existing ionic conductivity. The materials, due to
ubstitution, were mixed conductors, opening the possibility for their use
e and Silicide Negative Electrodes 249
e 8.1. Crystal structure of Li 7 MnN4 from the [110] viewpoint. Dark spheres represent
2.0
Capacity (mAh/g)
..,..,...........,."'"-.:LizN Layer
(Li\M+(l·x)) Layer
01(1
rted is related to Li vacancy population but the specific capacity does not
ed 100 mAh/g. Practically speaking, Liz.6Co0.4N can be treated as being
he lithiated state; therefore, lithium must be deintercalated from the
cture in an initial anodic oxidation. During the first charge, Li 2.6Co0.4N
les the removal of approximately 800 mAh/g of charge at an extended
au at 1.1 V corresponding to a delithiated composition of approximately
Co0.4N. Subsequent discharge and charge results in the reversible
rtion of 800 mAh/g from 1.5 to 0.1 V (Figure 8.4). After the initial
hiation cycle, subsequent delithiations remove capacity in a
2.0,..----------------,
e 8.4. Voltage profile of Li2.6Cao.4N vs. Li metal counter electrode. Cycle numbers are
ated (from Ref. 15).
e 8.5. N-K EELS spectra of lithiated and delithiated Li 2.6Cao.4N (from Ref. 15).
ility, on par with that of Liz,6Co0.4N. The copper compound revealed 650
h/g in the 0-1.3 V range. The Ni system, in contrast, was found to show
versible specific capacity of only 180 mAh/g in the range of 0.0 to 1.4 V.
Most recently, a layered lithium nitride has been investigated in the
17
uit of a less toxic and more cost effective negative electrode material.
eN2 is a layered nitride with hexagonal LhN layers analogous to the Lh_ N
materials described above. A similar material Li2(Li0.7Fe0.3)N
7Fea.3N) was investigated for its electrochemical performance. The
rial exhibited 550 mAh/g at an average potential (between 0.05 and 1.2
f 0.6 V. Analogous to Li 2.6Co04N, the structure decomposes to a quasi
rphous state upon delithiation.
().1.5V- 23mAig
-. SO% Irreversible Loss
1.5
c
y
0 0.5
;;.
e 8.6. Voltage profile of crystalline Ge 3N 4 cycled vs. Li metal between 0 and 1.5 Vat a
ant current of 23 rnA/g.
1 800
E
1 E 600
c.
...
u
..
0 1t>.O 400
;
-.l/l.=.. 200
Q LW LW LL LLU
0 10 20 30 40 50 60 70 80
Cycle#
e 8.7. Cycle life of crystalline Ge 3 N 4 cycled vs. Li metal between 0 and 0.7 V at a
ant current of 23 rnA/g.
Chapter 8
Residual Ge3N 4
Amorphous
•Li. Ge +Li3N
tion of Zn 3N2 into LiZn and a matrix of LhN, the high-pressure analog
i3N:
(8.9)
2.5
2
>
';' 1.5
0.0
1 SEI
0.5
0 2 3
#LiperC N
e 8.9. First cycle voltage profile of a Cu 3N electrode cycle vs. Li metal between 0 and 3
constant current of 22.4 mA/g (equivalent to a C/6 rate).
(8.13)
3. Summary - Nitrides
700
c:
.. . : 600
.§,
c
· ;:;
500
400
"-
u"'
2 "?
300
.."c::' 200
\.
-
100
i5
0 200
0 50 100 150
Cycle#
e 8.10. Cycle life of Cu 3N cycled vs. Li metal between 0 and 3 V at a rate equivalent to
c.
SILICIDES
1. Introduction
c:.
ii 1.0
:;::;
l
r:::
.!
0
0.5
11.
0.0 0 200 400 600 800 1 000 1200 1400
Capacity (mAh/g)
e 8.11. First three cycles voltage profile of an Mg 2Si electrode cycled vs. Li metal
en 0 and 2 Vat a constant current of 10 mA/g (from Ref. 33).
+ ..L
§ 1•• step
I
2•d step
L:r
.;
>
=----
Ill
Cl
0
> 0.1
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
x In Li.Mg2Si
e 8.12. First discharge voltage profile of an Mg 2Si electrode cycled vs. Li metal at a
2
ant current of 0.5 mA/cm (from Ref. 34).
described above. The first discharge voltage profile in Figure 8.12 shows
voltage remained quite flat during the insertion of the first two lithium
e it gradually decreased during further lithiation. The first step was
ved to involve the extraction of Mg from Mg 2Si concomitant to the
ation of a ternary LhMgSi phase (equation 8.16). The second step was
ciated to the alloying reaction of lithium with the extracted Mg metal
ation 8.17).
Si
Li
(A) (B)
a
Tetrahedron
occupied by Mg
/ ,L. /
;>
r
--
--....
./ ' '{3)
12.83A <"
/
--
1 / (A) ' \:J /,
/
.2. CrSh
36
Lithium was found to intercalate into the CrSiz structure. Because the
ChapterS
8.2. Reconstitution reactions as found from X-ray diffraction analyses after chemical
ion of CrSi2 (from Ref. 36).
eded the ones obtained with the binary system. This benefit in capacity
be related to the Cr-Si phases. The latter remain mostly
rochemically inactive after the first discharge and may act as stabilizing
uctive matrix.
.3. NiSi
rmined from the width of Bragg peaks using the Scherrer equation. The
harge voltage profile (Figure 8.14), which is smooth and approximately
ows a half parabolic curve down to 0 V, shows high initial capacity, 1180
h/g. Even though 80% of the initial discharge capacity was reversible,
capacity fade during cycling is significant as NiSi electrodes retained
65% of the initial discharge capacity after 25 cycles. Based on previous
rts of Sn-Sb materials, NiSi was proposed to decompose into Li-Si
ys and Ni during the first lithiation (equation 8.18). In subsequent charge
discharge, the Li-Si alloys reversibly transform into Si metal, which
titutes the active center of the electrode, while Ni metal remains
trochemically inactive (equation 8.19).
4. Summary - Silicides
2.5
2.0
..J 1.5
C:.
Gl 1.0
g 0.5
0.0
e 8.14. Discharge voltage profile of a nanosize NiSi electrode cycled vs. Li metal at a
ant current of 0.08 mNcm2 (from Ref. 37).
FERENCES
R. A. Huggins
INTRODUCTION
Early work on the commercial development of rechargeable lithium
es to operate at or near ambient temperatures involved the use of
ntal lithium as the negative electrode reactant. Binary phases, generally
ing a solid solution of lithium in one of the forms of carbon, are
tly employed on the negative side of lithium cells.
There is considerable interest in finding alternative materials that might
re attractive than the lithium-carbons. Improvements might involve the
to operate safely at higher current densities, less first cycle irreversible
ty loss, better cycling behavior, reduced specific volume, and lower
7
te at room temperatures when using a polymeric electrolyte. Prior
ments with liquid electrolytes were unsuccessful due to co-intercalation
cies from the organic electrolytes that were used at that time. This
m has been subsequently solved by the use of other electrolytes.
There has been a large amount of work on the development of graphites
lated carbon-containing materials for use as negative electrode materials
ium batteries in recent years. This is due in large part to the successful
pment by Sony of commercial rechargeable batteries containing
ve electrodes based upon materials of this family.
Lithium-carbon materials are, in principle, no different from other
-containing alloys. However, since this topic is treated in more detail
ere, only a few points will be briefly discussed here.
One is that the behavior of these materials is very dependent upon the
of both the nanostructure and the microstructure. Therefore, the
sition and the thermal and mechanical treatment all play important
in determining the resulting thermodynamic and kinetic properties.
als with a more graphitic structure have more negative potentials,
as those with less well organized structures typically operate over much
potential ranges, resulting in a cell voltage that is both lower and more
f-charge dependent.
Another important consideration in the use of carbonaceous materials
gative electrodes in lithium cells is the common observation of a
erable loss of capacity during the first charge-discharge cycle due to
sible lithium absorption into the structure. This has the distinct
antage that it requires that an additional amount of lithium be initially t
in the cell. If this irreversible lithium is supplied from the positive
de, an extra amount of the positive electrode reactant material must be
to the cell during its fabrication. As the positive electrode reactant
als often have relatively low specific capacities, e.g., around 140
, this irreversible capacity in the negative electrode leads to a
ment for an appreciable amount of extra reactant material weight and
e in the total cell.
There are some other matters that should be considered when
ring normal metallic lithium alloys with the lithium-carbons. The
c volume of some of the metallic alloys can be considerably lower than
the carbonaceous materials.
a+ Liquid
Liquid
a l
--------'-----------------
/ U+ p \
---------------------- '-•
I
I
--
p+y \
I
II
I
p \
Pure A Pure B
9.1. Schematic binary phase diagram with an intermediate phase 13, and solid solubility
inal phases a andy.
composition.
A schematic diagram for a hypothetical binary alloy system A-B is
in Figure 9.1. In this case there are four 1-phase regions. The
hases are designated as phases a, and y. In addition, there is a liquid
at higher temperatures. It can be seen in the figure that the single phases
separated by 2-phase regions as the composition moves horizontally
rmally) across the diagram.
According to the Gibbs phase rule all intensive properties, including the
cal potential, vary continually with the composition within single phase
s in a binary system. Correspondingly, the intensive properties are
sition-independent when two phases are present in a binary system.
the equilibrium electrical potential of such an electrode, E, in an
chemical cell is determined by the lithium activity, it also varies with
sition within single phase regions, and is composition-independent
there are two phases present under the equilibrium conditions that are
ed here.
The variation of the electrical potential difference V (voltage between
ctrode and some reference potential) with the overall composition in
pothetical system at temperature T 1 is shown in Figure 9.2. It is seen
alternately goes through composition regions in which it is constant
tial plateaus) and those in which it varies. If B atoms are added to pure t
A the overall composition is initially in the solid solution phase a and
ctrical potential varies with composition. When the solubility limit is
d lntermetallic Anodes 275
a
<I t><l t>
a+ +1
Pure A PureB
2. Variation of electrical potential with composition across the binary phase diagram
Figure 9.1.
nt reactions. The first involves the formation of the phase LhSb, and
written as:
2 Li + Sb = LhSb (9.1)
Upon the addition of more lithium a second reaction will occur that
in the formation of the second intermediate phase from the first. This
1600
1400 1320
Liquid
1200
.a
(/)
.P
....1
1000
"'"
795
BOO
630.7
600 89 99
.a .a
(I)
(/)
.!)
....1 .$:1
400 I!
....1
200 Sb-
10 20
00 30 40 50 60 70 80 90 100
Li Atomic Percent Sb Sb
written as:
(9.2)
Li +- conducting Li b
Li electrolyte
4. Schematic drawing of the electrochemical cell used to study the Li-Sb system.
F = C-P+2 (9.3)
on (9.1).
There is also a very narrow composition range in which only one phase,
is present and the potential varies. Upon the addition of further Li the
YS.
mv)
1oocP
L.
I I I
1000
I I ·
-Asec/o Sb
I I
2000
I 2!K)()
Li1SII
600 E
vs. AI."LiA(
800 Li 1SD 400 (mv)
600
200
400r- -
r- 0
200r- T • 360-c
1- -200
I I
0
0 2 3
y in Li.ySb
9.5. Results of coulometric titration experiment on Li-Sb system.
l composition moves into the region of the phase diagram in which two
are again present, in this case LhSb and LhSb, and the potential
s along a second plateau related to Equation (9.2).
(9.4)
(9.5)
his case:
(9.6)
ult is that the potential of this plateau is 871 mV vs. pure Li.
his is the same result that would be found if it were assumed that the
diate phase, LhSb, did not form, and that there is only one voltage
If the electrochemical titration curve were calculated from this value y,
it would have only a single plateau, at a voltage that is the weighted of
the voltages of the two reactions. This is false, due to the lack of tion
of the existence of the intermediate phase. Thus one has to be to be
aware of all of the stable phases when making predictions from
ynamic data.
his discussion has assumed that thermodynamic data are available,
one wants to predict the electrochemical properties. The opposite can
done, of course; experimental measurements can provide information
he relevant phase diagram and the thermodynamic properties of the
phases.
imilar experiments have been made on a number of other binary
systems under conditions such that equilibrium can be reached. This
ment is fulfilled within reasonable times much more easily at elevated
tures. In some cases, however, equilibrium data can also be obtained
ent temperatures, albeit with a bit more patience. Elevated temperature
a number of phases in the Li-Al, Li-Bi, Li-Cd, Li-Ga, Li-In, Li-Pb,
i-Si, and Li-Sn binary lithium alloy systems, made using a LiCl-KCl
salt electrolyte, are listed in Table 9.1.
Li + Bi = LiBi (9.7)
d lntermetallic Anodes 281
3
<
.•<
r
...
600
.6. Temperature dependence of the potentials of the two-phase plateaus in the Li-Sb i
systems.
in slope at the eutectic melting point (243°C), and the data for the two
s converge at about 420°C, which corresponds to the fact that the LiBi
s no longer stable above that temperature. These features can be seen
phase diagram for that system, shown in Figure 9.7. At higher
tures there is only a single reaction:
3 Li + Bi = LhBi (9.9)
n addition, the potentials of the second reaction fall along two straight
ments, depending upon the temperature range. There is a significant
in slope at about 210°C, resulting in a negligible temperature
ence of the potential at low temperatures, due to the melting of
.
s before, all the potentials are related to the standard Gibbs free
changes G,o relating to the relevant reactions.
Chapter 9
12001 ----------------------------------------------,
1000
Liquid
800
415
400 400
..,
:::;
iii
200
o;nrrrrrrrrmmmmmmnnnrrrrrrmmmmmmnnrrrrrrrrmmmmmm rrrrmmmm
0 10 20 30 40 50 60 70 80 90 100
Bi Atomic Percent Li Li
The temperature dependence of the values of Gr0 can be seen from the
g equation:
(9.10)
Mlr 0 is the change in the standard enthalpy and Sr0 is the change
in ndard entropy resulting from the corresponding reaction. Thus it can
be at:
(9.11)
From these data, one can obtain values of the standard molar entropy
es involved in these several reactions. They are shown in Table 9.2.
llows the prediction of the potentials at any temperature within this
An example that was of great interest some years ago involved the so
sodium-sulfur battery that operates at about 300°C. In this case, both of
ctrodes are liquids, and the electrolyte is a solid sodium ion conductor,
only called sodium beta alumina. This configuration can be described as
S/L system. It is the inverse of conventional systems with solid
des and a liquid electrolyte, SILlS systems. The negative electrode is
sodium, and the positive electrode is the product of the reaction of
m with liquid sulfur. Thus the basic reaction can be simply written as:
(9.12)
900
BOO
u 700
•
''
'
\
\
90 ICO
SULPHUR ,
Percont
2.0
I I I
Na 2S15 Na2S4 Na 2s3
1.0
DEPTH OF DISCHARGE,%
e 50 H WO%
0.7
0.5
c .1 ,
0.1
T
yin
10. Composition dependence of the potential in the Li-Sn and Li-Si systems.
Chapter 9
0.39
40mA.tml
D.38
0.01
- 0.36
D.35
l !) O.JJ
-
>
- l ,.sl plateau IOCVI
::-0.01 0.33-
ci -0.02
rnAicrn2 -
!i!
l-0.03
-0.04
-!!.05 charging •
0.28
-0.06 - I I I I t t
0 10
charge stored as% o1 CCij)CIC:ily -
9.11. Charge and discharge curves of the Li-Si alloy in the matrix of the
hemically inert mixed-conducting Li-Sn alloy at different current densities.
ore, lithium can react with Si to form the phase Lit.71Si (Li12Si7) inside
solid composite electrode containing the Lh_ 6Sn phase, which acts as a
-transporting, but electrochemically inert matrix.
Figure 9.11 shows the relatively small polarization that is observed
the charge and discharge of this electrode, even at relatively high
t densities. It is seen that there is a potential overshoot due to the free
involved in the nucleation of a new second phase if the reaction goes
pletion in each direction. On the other hand, if the composition is not
quite so far, this nucleation-related potential overshoot does not
, as seen in Figure 9.12.
This concept has also been demonstrated at ambient temperature in the
of the Li-Sn-Cd system.26•27 The composition-dependence of the
ials in the two binary systems at ambient temperature is shown in
9.13, and the calculated phase stability diagram for this ternary system
wn in Figure 9.14. It was shown that the phase Li4.4Sn, which has fast
cal diffusion for lithium, is stable at the potentials of two of the Li-Cd
titution reaction plateaus, and therefore can be used as a matrix phase.
d lntermetallic Anodes 289
Q.l9
0.38
0.37
I• 40 mA/cm'
0.36
0.35 r
r--+-------S-i-U-171Si plateau IOCVI------------ 0.34
0.33
0.32
:g
. o.J'll
Q.30
l
0.29
12. Charge and discharge curves of the Li-Si, Li-Sn composite if the capacity is
o that the reaction does not go to completion in either direction. There is no large
n overshoot in this case.
1000 r-------------------------------------
..........
aoo-
sao- \
.. .. .
• U·Sn
• • • • •• • •
400
•• ... •
•••
•
. ........
•
200
• •
••
....
LI·Cd
0
AMA ... A
•••
0 2 3 4 5
.13. Potential versus composition for Li-Sn and Li-Cd systems at ambient
re.
Sn
9.14. Calculated phase stability diagram for the Li-Cd-Sn system at ambient
ture. Numbers are voltages vs. Li.
DECREPITATION
A phenomenon called "decrepitation" can occur in materials that
o significant volume changes upon the insertion of guest species. It is
ometimes called "crumbling". These dimensional changes cause
nical strain in the microstructure, often resulting in the fracture of
es in an electrode into smaller pieces.
This can be a striking, and sometimes disastrous, phenomenon, for it is
ecifically related to fine particles, or even to electrochemical systems.
example, it has been shown that some bulk solid metals can be caused
ture, and can even be converted into powders by repeated exposure to
en gas if they form metal hydrides under the particular thermodynamic
ons present. This is, of course, different from the hydrogen
tlement problem in metals with body-centered cubic crystal structures,
d lntermetallic Anodes 291
100
5"
> ..
g01 60
IIi
·····························
01
80
>
.§.
Q)
•.•···•··•·· .-..=:Ch;a.rg.e.. EQUifibtium
• •
0 •
01
4
..... ...............D.is.
J: +
(...). 0 ••
2
15. Charge-discharge curve of the Li-Cd system with a fast mixed-conducting phase
ium-tin system at a current density of 0.1 mA/cm2 (ambient temperature).
50
45
40
35 -·-5
. a!
s 30
2 25
--o--10
-•-15
20 ----<>-- 20
.tj
._ 15
!'12
·c
u
10
0
0 2 3
Dilation strain parameter
9.16. Variation of the critical particle size as a function of the dilation strain and the
toughness of the phase in tension.
d Intermetallic Anodes 293
SUMMARY
Because of safety and cycle life problems with the use of elemental
m, essentially all rechargeable lithium batteries use lithium alloys as
ve electrode reactants today. The theoretical limits of the potentials and
ies of such electrodes can be determined from a combination of
odynamic data and phase diagrams. This has been demonstrated for
l examples of binary systems.
There are two general types of reactions that can take place: homophase
ns, in which guest atoms are inserted into an existing phase, often
ctically, and reconstitution reactions in which phases nucleate and grow
terophase microstructures. The potential varies with the overall
sition of an electrode in the insertion reaction homophase case, but is
sition-independent when reconstitution reactions take place in
phase microstructures.
The electrochemical titration method can be used to investigate the
nt parameters experimentally. Experimental results are now available
number of systems, both at elevated temperatures and at ambient
ratures.
Under equilibrium, or near-equilibrium, conditions the potentials are
y related to the standard Gibbs free energies of formation of the phases
ed. Thus thermodynamic data can predict experimental results.
ise, experiments can provide thermodynamic data. As an example, the
rature dependence of potential plateaus can be used to determine the
rd entropy changes in the relevant reaction. These experimental data
orrelate with the stability of phases in the phase diagram. Furthermore,
aximum theoretical specific energy of an electrochemical system can
e determined from the equilibrium electrochemical titration curve and
ated thermodynamic data.
These principles are applicable to liquid, as well as solid, binary
als. An example is the Na-S system, which is the basis for a high
rature LIS/L battery system that uses sodium beta alumina as a solid
lyte.
Composite multiphase microstructures can be useful in some cases, if
ermodynamic and kinetic properties of one binary system are
tible with those of another binary system. The critical question is the
ial range in which the desired reaction takes place. It must be within
d Intermetallic Anodes 295
ility range of the other phase with fast kinetics. Experiments have
trated that this mixed-conductor matrix concept can be used at both d
temperatures and ambient temperatures.
ne of the practical problems with a number of alloy systems in which
lues of capacity are achieved by the use of large composition changes
d to the corresponding volume changes. This can lead to decrepitation
ctronic disconnection of reactant particles. The general features of this
r, and the importance of the particle size and the mechanical
es of the respective phases are now understood in principle.
n interesting recent development relating to alloys has been the
nce of experimental results showing that electrochemical lithiation
se the appearance of amorphous phases with quite different properties
ose of the equilibrium phases. Especially interesting, and potentially
nt, is the observation that displacement reactions can take place during
n of some materials containing silicon that result in the formation of
ous silicon as a product. This amorphous silicon can then react
ly with lithium at attractive potentials and with a high capacity.
RENCES
ao, L.A. Heredy and R.C. Saunders, J. Electrochem. Soc. 118 (1971) 1039 .
ay, D.R. Vissers, F.J. Martino and K.E. Anderson, J. Electrochem. Soc. 123 (1976)
Discharge
Li < > Lt + e-,
Charge
Anode-Calhode
loy-liS,
r grapllite-
ous V:Os
oy-Polyaniline
oy-LIOII/ MnO,
loy-Vi),
,composite oxide
e,
,
,
Applicalion
iah·temperBtun:
3
battety)
eralUtC operation)'
IEPI (Central Research Institute ofEieclric Power Industry, Japan)j 'NEooiDENSO; 3USABC/Hydro-Quebec/3M; 4 USABC/ANU
T-America; 'Malsushita 8altcry Industrial; 'Lithium Technology; PolyPius
Chapter 10
-up coin-type cells, research activity for such alloy electrodes is now
g diminished. In Europe, nonetheless, Besenhard and co-workers
inue basic studies on alloy electrodes,6 and in the USA mainly USABC
veloping rechargeable batteries for high-temperature applications with
Al anode for vehicles.7
Electrolyte
General PC systems
PC/LiPF6
PC/LiCl04
General EC systems
S021LiAlC14
THF/LiAsF6
2MeTHF/LiAsF6
DME/LiAsF6
General BL systems
oup
HF
2Me-furan,
derivatives
Cetyltrimethylammonium
ganic
chloride
Nonionic
Crown ethers
Benzene
C02
as
N20,CO
Current Issues of Metallic Lithium Anodes 303
.2. Effect of Additives
0 60
i:E
d)
. 40
E
0
:::I
0 20
u
0
0 5 10 15 20
Cycle number
A)
@
@@
Reacti @ --..
High resistance film
Limetal
100 e
80
. ,'
---..--
70
60
50
0 5 10 15 20
Cycle number
e 10.4. Addition effects of Lil, Snl2, AII3 , and AII3 + 2MeF on coulombic
2
ency of Li in PC with LiCI0 4 (1 moVI); current density: 2.0 mA/cm ; charged
2
ricity: 0.2 C/cm2 ; a: PC, b: PC+ Lil (r 300 ppm), c: PC+ Snl
2 (Sn + 100 ppm),
d: AII3 (ArJ+ 100 ppm), e: PC +AII3 (Al3+ 100 ppm)+ 2MeF (0.5 vol %).
SCSI Board
r--· '--------,
........... lll!llfl!lllll-.. - lIl
Actuator 1I
eootro er II
II
Galvanoatat
Pot.ntlosblt 1-----*1
--Signal Lile
NlSubstr11t. --- RS-2320
···-GPIB
-1
"D
JS
t3 0.5
J
01)
....
c 2$
e
t o.s
c:l ! l ..
00
Sni2: 100 ppm
.....
c
m
J Q.S
J 35
FERENCES
E. Peled, J. Electrochem. Soc. 126 (1979) 2047.
T.B. Reddy, S. Hossain, "Rechargeable Lithium Batteries (Ambient Temperature)" in D.
Linden, T.B. Reddy Eds., Handbook of Batteries, 3'd Edition, McGraw-Hill, New York,
2002, Chapter 34, pp. 34.1-34.62.
J.O. Besenhard, J. Electrochem. Soc. 94 (1978) 77.
R.A. Huggins, "Lithium Alloy Anodes" in J.O. Besenhard Ed., Handbook of Battery
Materials, Wiley-VCH, Weinheim, 1999, pp. 359-381.
Y. Toyoguchi, J. Yamaura, T. Matsui, T. Iijima, Proc. Electrochem. Soc. 88-6 (1986)
659.
J.O. Besenhard, J. Yang, M. Winter, J. Power Sources 68 (1997) 87.
GL. Henriksen, A.N. Jansen, "Lithium/Iron Sulfide Batteries" in D. Linden, T.B. Reddy
Eds., Handbook of Batteries, 3'd Edition, McGraw-Hill, New York, 2002, Chapter 41, pp.
41.1-41.23.
P. C. Symons, P. C. Butler, "Advanced Batteries for Electric Vehicles and Emerging
Applications-Introduction" in D. Linden, T.B. Reddy Eds., Handbook of Batteries, 3rd
Edition, McGraw-Hill, New York, 2002, Chapter 37, pp. 37.1-37.25.
K. Kanamura, H. Tamura, Z. Takehara, J. Electroanal. Chern. 333 (1992) 127.
K. Kanamura, H. Tamura, S. Shiraishi, Z. Takehara, Electrochim. Acta 40 (1995) 913.
K. Kanamura, S. Shiraishi, Z. Takehara, J. Electrochem. Soc. 141 (1994) 2379.
Y. Malik, D. Aurbach, P. Dan, A. Meitav, J. Electroanal. Chern. 282 (1990) 73.
D. Aurbach, 0. Chusid, J. Electrochem. Soc. 140 (1993) L155.
T. Osaka, T. Momma, T. Tajima, Y. Matsumoto, Denki Kagaku 62 (1994) 451.
T. Osaka, T. Momma, T. Tajima, Y. Matsumoto, J. Electrochem. Soc. 142 (1995) 1057.
M. Ishikawa, K. Otani, M. Morita, Y. Matsuda, Electrochim. Acta 41 (1996) 1253.
T. Hirai, I. Yoshimitsu, J. Yamaki, J. Electrochem. Soc. 141 (1994) 611.
M. Ishikawa, Y. Takaki, M. Morita, Y. Matsuda, J. Electrochem. Soc. 144 (1997) L90.
M. Ishikawa, M. Kanemoto, M. Morita, J. Power Sources 81-82 (1999) 217.
Chapter 10
1 2 3
M. Pasquali , S. Passerini and G. Pistoia
1
Dipartimento ICMMPM, Universita di Roma La Sapienza, Rome, Italy
2
Enea Casaccia, Rome, Italy
3
Consultant, Rome, Italy (E-mail: pistoia@libero.it)
. INTRODUCTION
of Li+ content.
4. The number of sites for Lt has to be high and the hosting molecule
has to feature low weight and high density (high gravimetric and
volumetric capacity).
5. The coupled diffusion of electrons and Lt in the host, as a function
of concentration gradient, must be fast enough to grant a good rate
capability (high power).
6. The Lt intercalation/insertion has to be reversible, i.e., with no or
reversible structural changes, to allow cyclability.
7. The cathode must be stable towards the electrolyte over the entire
operating voltage range.
To these criteria, one has to add the prerequisites of:
8. Low cost.
9. Lack of toxicity.
10. Ease of synthesis.
However, it is becoming evident that the above are necessary but non
cient criteria. One has also to add:
11. A favourable interface formed in contact with the electrolyte (a
cathode may be stable in a given electrolyte, but affected by a
highly resistive interface).
12. The synthesis has to be not only simple, but also reproducible and
must allow the production of a material with the desired particle
size (the latter feature being of the utmost importance).
13. It should be possible to prepare a material encapsulated by another
one to protect it from unwanted reactions with the electrolyte.
14. The material has to be easily processed into a practical electrode.
Indeed, some materials may be difficult to fabricate as sheets (e.g.,
a LiNi02 slurry tends to be thixotropic).
Furthermore, the choice of a cathode is oriented by the applications.
instance, in portable devices, a material can be tolerated which is
ively expensive and moderately performing at high rates; conversely for
power applications, such as in EV, cheapness and rate capability are
datory.
In Table 11.1, materials which have particularly attracted the attention
esearchers in the last 10-15 years are reported together with their
ating voltage ranges. More details on them will be given in the
opriate sections. Before starting the description of selected cathode
rials, we think it useful reviewing the principal methods of synthesis.
section will be followed by two other sections dealing with the effect of
cle size on cathode performance and with the properties of the cathode
ion interfaces. These aspects are of a general character and can apply to
athode materials.
s in Cathode Materials for Rechargeable Batteries 317
5 LiMnz_xMx04 (x .5)
4 LiNiOz, LiCoOz, LiCol-x-yNixMy0z,
LiMnz04, Lil+yMnz-xMx04 (low x andy)
4-3 LiMnOz, LixMnl-yMyOz, Li[LixMyMnl-x-y]Oz
3.5 LiFeP04
3 Mn Spinels, LixMnOz, LixVyOz
2 Sand Polysulfides
1.5 FeSz
. METHODS OF SYNTHESIS
For many years, the only way of synthesizing cathode materials was
hermal treatment at high temperatures of mixtures of reactancts.
sion limitations in the solid state impose both long reaction times and
nding between subsequent firings. Even so, the possibility of having less
pure final products remains high. This can be partially alleviated if a
ant has a low melting point, so that the melt can impregnate the solid
impregnation technique). However, especially if the solid reactant has
particles and limited porosity, the core of the solid may not be reached
e viscous melt, again leading to incomplete reactions.
Since the 90's, alternative methods of synthesis have been proposed.
neral, they tend to form a better mixture of the precursors and to reduce
the time and temperature of the reaction. The most popular of these
sses is sol-gel synthesis. Originally, this term was used for the
lysis of an alkoxide and subsequent heating of the hydrolysis product
ve an oxide. At present, we call solution-sol-gel (or sol-gel, for short)
process in which a solution gives, through a chemical reaction, a sol
h converts into a gel. The latter is typically converted by annealing or
nt-exchange into the desired ceramic (normally an oxide) .
The sol-gel process allows a high degree of control over the physico
ical characteristics of the product obtained with high purity and
geneity. This is made possible by the initial solution step, as the
mixing of the reactants may reach the molecular level. Then, lower
eratures and shorter firing times are necessary to achieve materials with
Chapter 11
Li(CH3C02) • 2H20
Mn(CH3C02)2 • 6H20 Adipic acid
aqueous solution aqueous solution
I I
I Mixing
l Drying at 80-
90°C
Adipic acid/metal sol
I Drying at 90-1oooc
Gel precursors
Polycrystalline LiMn204
powders
Capacity (mAh/g)
0 50 100 150 200
2
O.lmA/cm
3 --------------
/t'\
,
lOth 5th lst
3
[\. .............-···-·-·-·-·-·-. ..
0.5mA/cm
2
/t""
40th 20th IOth 1st
2-
1 --.---.--.---.--,---.--.
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
x in Li 4+Mn sO 12
odimensional tunnels 0.9 Li per vanadium 16 (290 mAhlg) vs. 0.5 Li per
dium for the thermally prepared analog.
A peculiar synthesis of LiNi02 (and LiNi1_xCox02) is carried out by
izing, in a LiOH solution, Ni(OH)2 to NiOOH with Na2S208, followed
17
he H+;u+ exchange in the same solution. This form of LiNi02 shows a
4 -7 2
electron redox reaction (Ni + Ni +) to which a specific capacity of 300
2
/g corresponds in the potential range 4.2-1.5 V. 30 cycles at 0.9 mA/cm
eported, with a final capacity of 260 mAh/g.
Ion-exchange has been frequently exploited in the last years,
cially to obtain LiMn02 or similar materials. Orthorhombic LiMn02 was
prepared by Dahn et al. by ion exchange in an acetonitrile solution from
18
OH and LiOH. Bruce et al. have substituted Li for Na in NaMn02 and
4Mn02 19by21 ion exchange in alcoholic solutions at moderate
eratures. - The products so obtained show high capacities and long
e lives. They will be more extensively treated in another section of this
ter.
An interesting solution procedure, sometimes utilised to tailor
rials 22with
25 structures otherwise difficult to achieve, is hydrothermal
hesis. - By controlling such parameters as the pH of the medium,
erature, solvent, time and, optionally, size of the templating cation (if a
ial architecture is sought), modified oxide cathodes may be obtained. An
ple is the layered MxMn1_yCoy02 (M=Na or K) which has an electronic
uctivity two orders of magnitude higher than pure KxMn02 .Z Also, a
4
hermic peak for the 5-!.lm material, while the heat evolved by the two
r materials is rather limited. The conclusion is that the large surface area
2
m /g) of the 5-f..lm cathode and the subsequent reaction with the
rolyte prevent its use in practical batteries.
A similar solution technique ("controlled crystallization") was used by
et al. to obtain spherical LiNi 0.8Co0.202.44 These authors could obtain
cles with a very high tap density (3.24 g/cm\ while a commercial
le not based on spherical particles has a density of 2.27 g/cm3• The
er material could maintain a stable capacity of 170 mAh/g when cycled
2
5 mA/cm for 50 cycles.
A study of commercial Mn spinels has again stressed the importance
e morphology factor in determining the electrochemical performance.
ng the different samples, those having uniform spherical particles
ting from the agglomeration of spherical crystallites showed better
45
bility.
The issue of capacity fading for the Mn spinel upon storing and
ng, especially at high temperatures, has received a wealth of attention.
we wish to point out the reported implication of the spinel particle size
is phenomenon. Tarascon et al. have investigated in detail this aspect
ound a correlation between surface area and capacity loss on storing or
ng at 55°C.46 They prepared samples with different surface areas and
d that as this parameter increases the irreversible self discharge also
ases, due to the enhancement of both electrolyte oxidation and Mn
lution. A low surface area of 0.5-1 m2/g is considered optimal by these
ors.
The issue of spinel surface area vs. capacity fading has also been
essed by Cho through measurements of the surface area by both a
entional BET method and methylene blue (MB) method.47 With the
the area is measured through N2 adsorption, while the second is based
e adsorption of MB molecules. The difference between the two lies on
act that N2 can access the nanopores of the particles, while the larger
molecules, having the same size of the electrolyte molecules, -15 A,
ot. When evaluating the electrochemical characteristics of 5 samples
particles in the 5 to 50 f..lm range, a precise correlation was found
een irreversible first cycle capacity (and capacity fading) vs. MB
ce area. As the latter increased, higher values for the irreversible
city and capacity fading on cycling were measured. Interestingly, BET
MB surface areas did not correlate and even had opposite trends for the
es in the 13-25 f..lm range.
There are some discrepancies in the literature about the influence of
cle surface area on storing and cycling the spinel at high temperatures,
me authors have noticed a better performance with higher surface areas.
It seems that these disagreements may be partly reconciled if one takes
ccount the fact that the BET measurements do not provide a value of
Chapter II
real area accessible by the electrolyte. On this basis, one may perhaps
nalise the results of Yoshio et al. on cycling at 55°C Li-rich spinels of
2 50
ace areas 1.2 and 6.2 m /g. In 100 cycles, the former had a 5% capacity
while the latter had a loss of 8%. With a 511 ratio in surface areas, one
ht have expected a larger difference. It may be conceivable that the two
ples, prepared by different techniques, have a different pore structure in
s of number and size.
All of the above remarks emphasise that any synthesis or post
hesis process should produce particles with an appropriate size and
phology if they are to be used successfully in rechargeable batteries.
niques leading to compounds with exceedingly high surface area or
d particle size distribution are unlikely to be of practical interest.
This view was rejected by Chen and Dahn, who found, for
crystalline ZrOz-coated LiCo02, the same variation of the c axis
66
ured for the uncoated sample. According to these authors, the
llent electrochemical performance of this material has to be possibly
ected with a reduction of side reactions involving oxygen loss from
o02 to the electrolyte.
A reconsideration of the effect of coating LiCo02 with oxides has led
s group to attribute the capacity stabilization to a reduced Co
lution at high potentials.63b Even after charging to 4.8 V, Zr02-coated
o02 showed a much lower Co dissolution with respect to the uncoated
LiM02 A.-Mn02
Symmetry:Hexagonal Symmetry: Cubic
Space Group: R 3 m Space Group: Fd3m
c d
o-LiMn02 LiFeP04
Symmetry: Orthorhombic Symmetry: Orthorhombic
Space Group: Pmnm Space group: Pnma
e 11.3. a) shaded octahedra: M0 6 , white octahedra: Li0 6; b) note two of the pathways
Li+ transport; c) hatched octahedra: Mn06, circles: Li+; d) tetrahedra: P04 ,
edra:Fe06, circles: Li+ (in octahedral sites). a-c) from Ref. 70b; d) from Ref. 70c.
For LiCo02, the second monoclinic phase at low Li+ content (Figure
1
b) is not observed by Me Breen et al./ who found instead
her hexagonal phase. The phase composition above -4.3 V has recently
reconsidered by other authors. Chen and Dahn 77 " have found that the
gonal 03 structure·becomes a stage-2 compound of formula Li 0.12Co02,
hexagonal as previously proposed by Ceder et az.77b At full delithiation,
2 with an 01 structure is formed. This latter finding is in agreement with
esults of Yang et al.71c and Tarascon et alY" The latter authors, in a
equent paper, assign a mixture of two 01 phases to Co02.78 However,
15.5 Hl H3
HI +
•
6.0
Cl
II 4.1S
!!
0
> 4.0 (a)
3.5
3.0
o.o o.1 0.2 o.3 o.4 o.5 o.e o.7 o.8 o.g 1.0
x In L1 1_"NI02
---- - ;:!/
(b)
----
Cl
! 4.5
0 4.0
)
3.5
3 0
" o.o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 O.SI 1.0
X In Ll 1_wCo02
.. \ ••• ••••••
•••••
- 13.5
u
14
;;:
o -Monoclinic
"""""
(c)
13
•• •+CoO2
12.5 ....._ ......._......_......._.......... _._. _._............
-0.2 0 0.2 0.4 0.6 0.8
X in LixCo02
14.5
14.5 oo
co 0
14.4 CtJ 0 OofJOO
0 0
14.4 cP
-< 14.3 OJD
(.)
14.3 cx:fX)
13.6 (d)
()
14.2 13.5
14.2
11.4. (a) Phase diagram of Li 1.xNi02 (from ref.71); (b) Phase diagram of Li 1_xCo02
Ref. 71); (c) Variation of the c axis during delithiation of LiCo02 (negative x values:
olyte decomposition) (from Ref. 67a); (d) Variation of the c axis during delithiation of
2 (from Ref. 72).
s in Cathode Materials for Rechargeable Batteries 331
79
rding to Venkatraman and Manthiram, Co02 has the P3 structure. e
authors observe formation of a P3 phase for x=0.45 in LixCo02, h
coexists with the 03 phase in the range 0.2<x<0.45 and is
acterized by a faster decrease of the c axis as x decreases. The P3 phase
oxygen more rapidly than the 03 phase (P3-like Co02 should be
en as CoOu) and this can be connected with the practically usable
city of LiCo02.
Examples of the cyclability of pure LiNi02 and LiCo02 are given in
e of the references of the next sub-sections as well as in the references
ng with coated LiCo02 in Section 11.4.
We have already noted that LiCo02 is the cathode used in most of the
ently commercialised Li-ion batteries. LiMn204-based cells have a
inal market. Batteries with LiCo02 have now reached energy densities
cess of 150 Wh/kg and 350 Wh/1, cycle lives in excess of 1000 cycles
low self-discharges (<3%/month). The industrial processes are well
lished and high levels of mass production of portable batteries have
achieved through huge investments. On this basis, it is very difficult to
ider the substitution of LiCo02 with another cathode, unless the
native is highly attractive in terms of performance and price. One has
to consider that the price of cobalt has been drastically decreasing in
t years and that in small portable batteries the relative cost of active
rials is lower.
have found that both ions are simultaneously oxidized. These latter
have also examined the relationship between irreversible capacity
d depth of first charge. They found that most of the capacity is lost at
ginning of the first charge and is not influenced by the phase
ons occurring during further charging/discharging. The presence of
uces the irreversible capacity, which is due to the oxidation of Ni2+ in
layers and preferential removal of u+ around the Ni ions, this
a local structural collapse.82
Co not only improves the layered characteristics of LiNi02, but also
es its thermal stability in the charged state. This is because the Co-O
s stronger than the Ni-0 bond (binding energies: 1067 KJ/mol for
nd 1029 KJ/mol for Ni02).83 As a consequence, a thermogravimetric
s has shown progressively reduced weight losses to 1000°C as Co
s Ni in LiNi 1_xCox02• 83 The stronger Co-O bonds also contribute to
e the structure during Lt intercalation/deintercalation. In addition to
excess helps to reduce the amount of Ni2+ in the Lt layers, as these
rated by Li+, and thus a limited irreversible capacity in the first cycle
ood cyclability are observed. 83
An investigation on the thermal stability of delithiated Li 1_yNi 1_xCox02
.3), has produced surprising results. All samples have shown
al damages on storage, demonstrated by a decrease of the cia ratio
temperatures as low as 50°C. The rate of change of the cia ratio
es with time (several days), temperature and oxidation degree, and
es with the Co content: a Ni-free sample does not show any
n.84 The thermal stabilization of oxidized Li 1_yNi 1_xCox02 at higher Co
s is confirmed by DSC experiments. In samples charged to 4.2-4.3 V,
e exothermic peak at about 195°C of pure Li 1_yNi02 is transformed
85
small peak centred at 220°C for Li 1_yNi Cyclic
0.7Co0.302.
metry tests of this latter compound have shown the disappearance of
agonal-monoclinic phase transitions and delithiation seems to occur
85
gle phase up to 4.4 V. Several authors have attributed the im roved
6 87
hemical stability of the Co-containing cathodes to this feature. •
ccording to Ref. 79, the original 03 phase coexists, for x<0.25 in
5Co0.1502, with another hexagonal phase, 03'. At full delithiation,
he new 03' phase exists. Oxygen loss would occur only after
on of this phase. This behaviour is different from that of LiCo02 n
loss at x<0.45, see Section 11.5.1) and would explain why
5Coo.1502 can reach a practical capacity of 180 mAh/g vs. 140 mAh/g
o02•
he references of this sub-section give details on the best syntheses to
e efficient materials. Here, we only wish to point out an unusual low
ature synthesis based on the reaction at 400°C (instead of the usual
86
0°C) between f3-Ni1_xCoxOOH and LiN03• Since Ni is already in the
t state in the starting material, a lower temperature may be used. A
s in Cathode Materials for Rechargeable Batteries 333
j
80
= +------- ---------+---------- --------------
30
20 +------- ---------+---------- ------------
10
0
0 100 200 300 400 500 800 700 BOO 900 1000 1100 1200
Cycle number
ence of Mg in the interslab sites from the relative heights of the (003)
(104) peaks in the diffractograms: a lower (003)/(104) ratio for x=0.05
0.075 was taken as a proof of the distribution of Mg between Li and
91
o layers.
Conversely, Cho et a/.89 and Chang et a/.90 have refuted the presence
g in the 3b sites. According to the former authors, as the Mg content
ases from 0 to 0.06 atoms/mol, the Mg and Ni occupancy of the Li sites
mes negligible (Rietveld analysis). Chang et al. also found a very low
concentration in the Li sites, through Rietveld, chemical and
mogravimetric analyses. Despite the disagreement regarding Mg
pancy, all of the authors agree that these materials have excellent
abilities with high and stable capacities (e.g. 160-170 mAhlg 89).
1
Mn, Ti and AI substitutions also give excellent cyclability.92 -95 LiNi _x
Aiy02 is now receiving special attention at the technological level as it is
ved to be a candidate for a new generation of Li-ion batteries. 95 Indeed,
n be produced on a large scale and is commercially available. A 10 Ah
n cell was cycled at AEA for 270 cycles (C/5) with a 90% capacity
tion. The specific energy of this cell was 150 Wh/kg in a "soft
aged" version. 95 "A similar performance (up to 145 Wh/kg) was reported
AFI' for a 40 Ah cell.95b,c This cell can be long cycled with a rather
tant energy output- see Figure 11.5.
Dahn et al. have reported the structure and electrochemical behaviour
2
i[NixCo1_2xMnx]02 (x=l/4 and 3/8), prepared by substituting Ni + and
92
+ for Co3 in LiCo02. " This new compound, which maintains the a
+
e02 structure of LiCo02, can deliver a stable capacity of 160 mAhlg
een 2.5 and 4.4 V, and proves more thermally stable than LiCo02•
The double cationic and anionic substitution in LiNi02 has also been
96
osed, e.g., in: Li .075Ni 0.755Co0.1701.9F0.1 • This material has an initial
city of 182 mAh/g and a capacity fading of 2.6% in 100 cycles at 1
cm2 .
7 98
' Ohzuku et al. have found that Al in Li1-xAlu4Nh,402 produces a
e-phase behaviour and allows charging up to 4.5 V with excellent
97
sibility of the 160 rnAh/g initial capacity. The presence of non
3
izable Ae+ limits the degree of Ni + oxidation, thus avoiding the
iderable shrinking of the c axis at high x values and consequent loss of
city. In contrast, Al substitution in LiCo02 does not have a positive
t and fast capacity fading is observed. Al substitution increases the
ge in both LiNi02 (by -0.1 V) and LiCo02 (by -0.2 V).
2
Instead of counterbalancing the tendency of Ne+ to form Ni + in
4
i02 with di- or tri-valent ions, in one report Ti + was used to compensate
99
he electric charge deficiency and structural instability due to N?+. The
bility of LiNi1_xTix02 (0.025:SX:::;0.2) was rather good with initial
cities ranging from 170 to 240 rnAh/g as a function of x.
A double Mg-Ti substitution has been reported to have a dramatic
t on the thermal stability of LiNi02. Indeed, compounds of formula
1_xTix12M&1202 show a progressive disappearance of the large
hermic peak of pure, fully oxidized LiNi02 as x increases. For .x=0.3, no
76
herm was observed. " These materials, however, have large irreversible
cities. Including Co among the dopants, the materials become highly
hiometric (99% Li+ in the Lt layers for LiNio.7Co0.2Ti0.05Mg0.0502) and
an excellent cycle performance in addition to thermal stability.76 b This
eature has been explained by arguing that the highly ionized Ti4+ and
(plus unremovable Lt) increase the ionic character of these
pounds and prevent depletion of electrons from the oxygen sites during
ge.76c
6 100 101
There are reports of 0 substitution with p9 • or S. The latter seems
e particularly effective in stabilizing the capacity of a compound2 with
ula Li1.QJNi01.98S0.02 at -140 rnAh/g in 100 cycles at 0.4 rnAJcm .
95
90
.g 85
80
z:. 75
-o-
· 8 70 ... -&- 45'C ............·-·-·--- --·-··-··-··"·-·-··
65 -o- 60'C ..----··-·-······-·-·····---·--··········.......
600 1.000 2.000 3.000
Cycle (Charge 4.1V, Discharge 2.5V)
number
e 11.6. Cycling of a NEC battery with a Mn spinel cathode. Rate: 1 C (from Ref. 4a).
x<0.5. Li-rich spinels (and spinels doped with another ion) are single
e throughout the entire x range, and, finally, oxygen-defective spinels
Mn204-S) show two biphasic plateaux separated by a voltage knee at
5. The two-phase regions show the fastest capacity decay on
cycling.
Much debate is also ongoing regarding the failure mechanisms upon
ge or cycling at high temperature. The hypotheses are: 1) Mn2+
olution; 2) J-T effect with formation of the tetragonal spinel LizMn204;
ructural instability and 4) decomposition reaction of the electrolyte at
voltage. It is quite possible that such effects are interrelated. For
nce, points 1-3 can occur if the following sequence is considered: in a
harge down to -3 V under non-equilibrium conditions, LizMn 204 is
ed; this can then disproportionate according to the reaction:
(11.1)
120
city fading on storage at high temperatures. The cathode of the above
ioned NEC battery is formed by blending the Mn spinel with 10-15%
121
0 .sCoo.zOz?' The latter material acts as a scavenger for HF, thus
ressing Mn dissolution.
We have already mentioned the effect of coating the spinel particles
such materials as Co304 or LiCo02 prepared by sol-gel synthesis. 5 9 The
layer formed on the spinel, possibly with the LiMn 2.xCox04 composition,
sistant to HF attack and prevents Mn dissolution. As a result, a steady
city of 105 mAh/g was obtained on cycling at 60°C.
3
Substituting part of Mn + with other ions was suggested by the
ence of a J-T effect in the 4 V range. The spinel was initially doped with
ent ions as Mg2+ or Zn 2+. 122' 123 The capacity was indeed stabilized
3
ugh a penalty in its absolute value was paid, as the amount of Mn +
h determines the capacity was decreased, according to the general
3 4
ula: Li[Mn\Mn + 1+(n-4)xMn + 1_(n-J)x]04. The numerous reports on this
have shown that the doping ions also have other beneficial effects: they
124
ce the Mn solubility in the electrolyte; often give higher energies for
125
M-0 bond; limit the extraction of u+, and give rise to a single-phase
126
ge-discharge process. All the above effects contribute to the structure
lization, which is ultimately the basis for a good cyclability.
Several elements have been used as dopants. Substitution with Li is
ost obvious choice: it is generally agreed that Li+ replaces Mn 3 + in the
octahedral sites, from which it cannot be extracted. However, Howard et
y observing that Li+-doped spinels have an experimental capacity larger
the theoretical one, have suggested, also on the basis of neutron
127
action data, that Li+ may also reside in tetrahedral 8b sites. In another
128
rt, occupancy of octahedral 16c sites is suggested. Lii+xMn2_x04
ls, even though showing a better cyclability than the undoped one,
problems on storage at high temperature, with an enhanced formation
107
i2Mn03/LizMn 4 09, and do not compare favourably with other doped
129
ls in terms of capacity retention.
Divalent metals, as Zn and Ni, or trivalent as Co, Cr, Ga and Al show
est performance. Ni and Cr are particularly effective in reducing the Mn
124
lution and, so, the spinel instability. Capacity losses as low as 0.02-
mAhfg·cycle at room temperature and 50°C have been measured during
129 130 132 133
cycling. ' ' ' Reports indicate that AI, Zn and Ga may reside in
131 132
8a tetrahedral sites. ' In principle, they could create diffusion
ems for u+ along the 8a-16c-8a paths, but, at least for the limited
itutions considered, this does not occur and, instead, both show a well
129 132 133 131
lized capacity. ' ' According to Thackeray et al. substitution in
hedral sites may retard the Mn diffusion to these sites, a phenomenon
mpanying its dissolution.
Excellent data have often been reported for double doping. Not only
Chapter 11
57
different cations be used to replace Mn, e.g., Ga-Cr, butS or F may be
to partially replace 0. Tarascon et al. have proposed LiA10.2Mn1.804_sFz
0.06
0.04
0.03
PLiON™Li-ion
LAMOFvs. Li 4Ti5 012
O.ot 55°C C/7
approx. 6 months
e 11.7. Cycling of aLi-ion cell with LiAio. 2 Mnl.804 _sFz as a cathode (from Ref. 134).
134
h z;::;(}.5). The improvement of cyclability of this material at 55°C over
-doped spinel is dramatic as shown in Figure 11.7, where cycling of a
mer Li-ion cell at 55°C is reported.
Sun et al. have substituted 0 with a limited amount of S:
135
0.24Mn1.7603.98S0.0z. AI stabilizes cycling in the 4 V range, whereas S
is tive in stabilizing cycling in the 3 V range. Overall,
capacities eding 200 mAh/g could be obtained. Leaving aside for the
moment the
ical interest of a cathode showing a sharp voltage drop of 1 V, the
lity in the 3 V range is highly surprising. Furthermore, the authors have
n that, even upon long cycling down to 2.3 V, the cubic structure of the
l spinel is preserved, i.e., there is no J-T effect causing the tetragonal
rtion, perhaps due to a more flexible structure of the oxysulphide. A
136
lar result has been found by using Cr instead of Al. Kang and
denough have commented on these results. According to them, sulphur
(r=l.70 A vs. 1.26 A for oxygen) would only reside on the spinel
ce and the J-T distortion should be present as the 3 V plateau does
137
r.
Yoshio et al. have shown that heavily doping the spinel with Co or Ni,
in LiM0.5Mn 1504, also provides a good stabilization of the 3 V
138
city.
ds in Cathode Materials for Rechargeable Batteries 341
C>
c: 250
<
E
-
200
>-
(3 150
C tl
c.
Ctl
.) 100
Q) 50
C>
...
Ctl
c:
.)
0
0 0 20 40 60 80 100
Cycle number
e 11.8. Cycling of a cell with a Li,Mn 0.975Li0.02502 cathode (from Ref. 21c).
The solid solution between LizMn03 and LiCr02 has also received
h attention and, in particular, Li[Li0.2Cr0.4Mn0.4]02 has proved
139
ctive. Just like for the other solid solutions, the doping transition metal
3 6
e electrochemically active species (Cr + 7Cr + in this case), while Li
Mn form an inert matrix.
5.4. 3 V Cathodes
ors report capacities of -250 mAh/g while cycling at 0.4 rnNcm2 in the
.0 V range. 164
Recently, Li-rich spinels (e.g., Li 4Mn50u) and 0-rich spinels have
reconsidered as 3 V cathodes. The former has been examined by
thiram (see Section 11.2),u while the latter, with a typical composition
n 204.54, has been shown by Kilroy et al. to cycle with capacities above
165
mAh/g.
Surprisingly stable capacities have been reported for the Mn spinel in
V range by Goodenough et al. These authors have prepared a nearly
hiometric spinel by a sol-gel synthesis with glycolic acid, producing a
137 2
rial with an average particle size of 0.8 f-Lm. Cycling at 1 rnNcm ,
cities of -110 mAh/g have been obtained in spite of the evident two e
behaviour testifying to a J-T effect. The authors, although unable to
ide a sound explanation of this unusual behaviour, tended to attribute it
e small particle size. Indeed, the same authors made experiments on
milled spinel samples obtained by both sol-gel and thermal syntheses
found even higher and stable 3 V capacities. In this work, also
ioned in Section 11.3, the conclusion was reached that the J-T distortion
d be accommodated by strained, nanosized particles.37 Excellent cycling
vior was also reported at 60°C and 80°C.
Vanadium oxides, e.g., V205, V6013 and LiV308, have long been
tigated as cathodes for 3 V batteries with aLi anode.
V205 is very amenable to sol-gel processing by different routes. Using
ell known ion-exchange process that yields a precursor HV03 solution
self-assembles into the V205 hydrogel, a variety of materials of different
sity have been synthesized in form of films, powders and free-standing
branes. In the gel stage, the V205 solid network is bicontinuous with
connected pores filled with water. Depending on the way the liquid
nt is removed (drying process) any of the following three types of solid,
phous vanadium pentoxide can be produced:
1. Xerogel (XRG): nominal surface area= 1-10m2/g.
2. Aerogel (ARG): nominal surface area= 350-450 m2/g.
3. Aerogel-like (ARG-like): nominal surface area= 100-300 m2/g. 166
All materials are reported to have high specific capacities (see Table
) in near-equilibrium or very slow insertion/release tests.167 However, as
ate increased the capacity of xerogel decreased to 380 mAh/g (at C/20),
mAh/g (at C/5), and 200 mAh/g (at 1C). The dependence of the capacity
e rate indicates that diffusion limitations are key factors determining
the
rmance of the xerogel material.
Chapter II
VzOsXRG 560mAh/g
Cu-doped V z05 XRG 300 mAh/g
VzOsARG 560 - 780" mAh/g
Vz05 ARG-Like 560 mAh/g
c-V205 148 mAh/g
a. Obtained by chemicallithiation
a)
.-.
:::1 3.8
..;
;o.
3.4
:..!..
c 3.0
!2.6
0 0.2 0.4 0.6 0.8 1.0
xLi/UFeP04
b) 160
0.8
-
140 'eG
0...
0.6
r - 5C
- 120
.... 100 ..e....
lo<
....J 80 'Cl
0.4 •
60
•c.
t,)
0.2 40
20
0 0
0 200 400 600 800
Cycle number
rly confirm that the reduction of the material particle size and the
ovement of the electronic conductivity lead to a better electrochemical
ormance of LiFeP04.
202
hesis procedures. These compounds have been tested in prototype
ries under conditions simulating those typical of cellular phones. A
(phenylenedithiazol) has also been reported, whose electrochemical203
ity is based on an S-S bond and an unsaturated dithiazolium ring. Its
llent first-discharge capacity (420 mAh/g at low rate) was only partly
rsible, perhaps due to solubility problems.
Peled's group has investigated in detail the Li/FeS2 couple in a
204 205
posite polymeric electrolyte based on PE0. ' Thin-cathode batteries,
harging at the mean potential of 1.7 V (at 135°C), have a reversible
city of 625 mAh/g (2.8 Li/FeS2) and a projected specific energy of 170
kg in a bipolar configuration. 500 cycles with 100% DOD have been
rted. The Li/FeS2 couple has been proposed as a good candidate for EV
cations. The same couple has also been cycled at room temperature in a
206
meric electrolyte based on PVdF.
. CONCLUDING REMARKS
The research on rechargeable cathode materials has been very fruitful
e the introduction of the first commercial Li-ion batteries. Here, we shall
o summarize the most important advances.
LiCo02, the cathode still used in mass production, has been optimized
41
s preparation, especially as far as particle size is concerned (-20 -tm).
esting reports on the coating of this cathode to improve the thermal
lity and to allow the use at higher voltages (with higher capacities) have
ared.59• 63 As mentioned in Section 11.5, the price of this material has
decreasing, consolidating its predominant position in batteries for
ble electronics. In 2002, the price of LiCo02 was about twice that of
1_xCox02 and about four times that of LiMn 204, the two major
idates for future developments. However, if the price is related to the
gy (Wh) of a given battery, a LiCoOz-based battery is only 30% more
nsive than batteries with cathodes containing Ni or Mn.95b This can be
ated in small cells, but not in the large ones intended for EV and
strial applications. Indeed, in view of the fact that for such large
ries mass production processes are not as well established as for the
l ones, it can be foreseen that only Ni- and/or Mn-based cathodes will
sed.
s in Cathode Materials for Rechargeable Batteries 353
FERENCES
. Nagaura, K. Tozawa, Prog. Batt. Solar Cells 9 (1990) 209.
. Numata, C. Amerniya, J. Iriyama, T. Miura, M. Shirakata, NEC R§D 41 (2000) 8.
M. Tsunoda, Y. Oshima, M. Yoshinaga, T. Shirasu, NEC R§D 41 (2000) 13.
a) C. Amerniya, J. Kurihara, M. Yonezawa, NEC R§D 42 (2001) 241; (b) S. Narukawa,
. Nakane, N. lrnachi, S. Fukuoka, M. Yamasaki, II'" IMLB, Monterey, CA, USA, June
Chapter 11
2002, Abs. 2.
Y.-S. Lee, Y-K. Sun, K.-S. Nahm, Solid State lonics 109 (1998) 285.
H.-B. Park, J. Kim, C.-W. Lee, J. Power Sources 92 (2001) 124.
C. Julien, M.A. Camacho-Lopez, T. Mohan, S. Chitra, P. Kalyani, S. Gopukumar, Solid
State 1onics 135 (2000) 241.
S. Chitra, P. Kalyani, T. Mohan, R. Gangadharan, B. Yebka, S. Castro-Garcia, M.
Massot, C. Julien, M. Eddrief, J. Electroceram. 3:4 (1999) 433.
W. Yang, G. Zhang, J. Xie. L. Yang, Q. Liu, J. Power Sources 81-82 (1999) 412.
Y.-S. Han, H.G. Kim, J. Power Sources 88 (2000) 161.
A. Manthiram, J. Kim, Chern. Mater. 10 (1998) 2895.
S. Bach, J.P. Pereira-Ramos, N. Baffier, Solid State lonics 80 (1995) 151.
Q. Feng, H. Kanoh, K. Ooi, M. Tani, Y. Nakacho, J. Electrochem. Soc. 141 (1994)
L135.
J. Kim, A. Manthiram, J. Electrochem. Soc. 145 (1998) L53.
A. Manthiram, A. Dananjay, Y. Zhu, Chern. Mater. 6 (1994) 1601.
C. Tsang, A. Manthiram, J. Electrochem. Soc. 144 (1997) 520.
. Maruta, H. Yasuda, M. Yamachi, J. Power Sources 90 (2000) 89.
.N. Reimers, E.W. Fuller, E. Rossen, J.R. Dahn, J. Electrochem. Soc. 140 (1993) 3396.
A.R. Armstrong, H. Huang, R.A. Jennings, P. G. Bruce, J. Mater. Chern. 8 (1998) 255.
A.D. Robertson, A.R. Armstrong, P.G. Bruce, Chern. Mater. 13 (2001) 2380.
a) A.R. Armstrong , P. G. Bruce, Nature 381 (1996) 499; (b) A.D. Robertson, A.R.
Armstrong, A.J. Peterson, M.J. Duncan, P.G. Bruce, 2001 ECS and ISE Joint
nternational Meeting, San Francisco, CA, USA, September 2001; (c) A.D. Robertson,
A.R. Armstrong, A.J. Peterson, M.J. Duncan, P.G. Bruce, /11h IMLB, Monterey, CA,
USA, June 2002, Abs. 74.
M.S. Whittingham, J.-D. Guo, R. Chen, T. Chirayil, G. Janauer, P. Zavalij, Solid State
onics 75 (1995) 257.
M.S. Whittingham, S. Yang, K. Ngala, P. Zavalij, 2001 ECS and ISE Joint International
Meeting, San Francisco, CA, USA, September 2001.
.K. Sharma, G.J. Moore, F. Zhang, P. Zavalij, M.S. Whittingham, Electrochem. Solid-
tate Lett. 2 (1999) 494.
R. Chen, M.S. Whittingham, J. Electrochem. Soc. 144 (1997) L64.
M.S. Whittingham, Solid State Ionics 134 (2000) 169.
a)Y. Li, C. Wan, Y. Wu, C. Jiang, Y. Zhu, J. Power Sources 85 (2000) 294; (b) K.
1
Konstantinov, G.X. Wang, J. Yao, H.K. Liu, S.X. Dou, ll h IMLB, Monterey, CA, USA,
une 2002, Abs. 100.
a) S.-T. Myung, H.-T. Chung, J. Power Sources 84 (1999) 32; (b) S.-T. Myung, S.
Komaba, N. Hirosaki, N. Kumagai, K. Arai, T. Kodama, Y. Terada, I. Nakai, ll 1 h 1MLB,
Monterey, CA, USA, June 2002, Abs. 90.
Y.-M. Chiang, Y.-1. Jang, H. Wang, B. Huang, D. Sadoway, P. Ye, J. Electrochem. Soc.
145 (1998) 887.
H. Yan, X. Huang, H. Li, L. Chen, Solid State lonics 113 (1998) 11.
H. Yan, X. Huang, L. Chen, J. Power Sources 81-82 (1999) 647.
.S. Whitfield and I.J. Davidson, J. Electrochem. Soc. 147 (2000) 4476.
M.N. Obrovac, 0. Mao, J.R. Dahn, Solid State 1onics 112 (1998) 9.
N.V. Kosova, N.F. Uvarov, E.T. Devyatkina, E.G. Avvakumov, Solid State /onics 135
2000) 107.
a) S. Soiron, A. Rougier, L. Aymard, J.-M. Tarascon, J. Power Sources 97-98 (2001)
02; (b) W.T. Jeong, J.H. Joo, K.S. Lee, 1 J'h IMLB, Monterey, CA, USA, June 2002,
bs. 68.
. Singhal, G. Skandan, G. Amatucci and N. Pereira, Electrochemical Society
roceedings Volume 2000-36, A.R. Landgrebe and R.J. Klinger, Eds., p. 244, 2001.
.-H. Kang, J.B. Goodenough, L.K. Rabenberg, Chern. Mater. 13 (2001) 1758.
s in Cathode Materials for Rechargeable Batteries 355
S.-H. Kang, J.B. Goodenough, L.K. Rabenberg, Electrochern. Solid-State Lett. 4 (2001)
A49.
Y.-I. Jang, B. Huang, H. Wang, D. R. Sadoway, Y.-M. Chiang, J. Electrochern. Soc. 146
1999) 3217.
Y. S. Lee, M. Yoshio, Electrochern. Solid-State Lett. 4 (2001) A166.
Y. Nishi, J. Power Sources 100 (2001) 101.
. Cho, G. Kim, H.S. Lim, J. Electrochern. Soc. 146 (1999) 3571.
. Cho, B. Park, J. Power Sources 92 (2001) 35.
. Ying, C. Wan, C. Jiang, Y. Li, J. Power Sources 99 (2001) 78.
H. Huang, C.H. Chen, R.C. Perego, E.M. Kelder, L. Chen, J. Schoonman, W.J.
Weydanz, D.W. Nielsen, Solid State Ionics 127 (2000) 31.
G. Amatucci, C.N. Schmutz, A. Blyr, C. Sigala, A.S. Gozdz, D. Larcher, J.-M. Tarascon,
. Power Sources 69 (1997) 11.
. Cho, Solid State Ionics 138 (2001) 267.
V. Manev, T. Faulkner, J. Engel, in: Proceedings of the 1' 1 Hawaii Battery Conference,
anuary 1998.
H. Huang, C.A. Vincent, P.O. Bruce, J. Electrochern. Soc. 146 (1999) 481.
Y. Xia, N. Kumada, M. Yoshio, J. Power Sources 90 (2000) 135.
D. Aurbach, J. Power Sources 89 (2000) 206.
D. Aurbach, K. Gamolsky, B. Markovsky, G. Salitra, Y. Gofer, U. Heider, R. Oesten, M.
Schmidt, J. Electrochern. Soc. 147 (2000) 1322.
G. Amatucci, A. Du Pasquier, A. Blyr, T. Zheng, J.-M. Tarascon, Electrochirn. Acta 45
1999) 255.
T. Aoshima, K. Okahara, C. Kiyohara, K. Shizuka, J. Power Sources 97-98 (2001) 377. A.
Du Pasquier, A. Blyr, A. Cressent, C. Lenain, G. Amatucci, J.-M. Tarascon, J. Power
ources 81-82 (1999) 54.
C.H. Chen, J. Liu, K. Amine, J. Power Sources 96 (2001) 321.
A. Antonini, C. Bellitto, M. Pasquali, G. Pistoia, J. Electrochern. Soc. 145 (1998) 2726. A.
Du Pasquier, F. Orsini, A.S. Gozdz, J.-M. Tarascon, J. Power Sources 81-82 (1999)
07.
a) J. Cho, G.B. Kim, H.S. Lim, C.-S. Kim, S.-1. Yoo, Electrochern. Solid-State Lett. 2
1999) 607; (b) J. Cho, T.-J. Kim, Y.J. Kim, B. Park, Chern. Cornrn. 2001, 1074.
.-C. Park, Y.-M. Kim, S.-C. Han, S. Ahn, C.-H. Ku, J.-Y. Lee, J. Power Sources 107
2002) 42.
. Endo, T. Yasuda, A. Kita, K. Yamaura, K. Sekai, J. Electrochern. Soc. 147 (2000)
291.
. Endo, T. Yasuda, K. Yamaura, A. Kita, K. Sekai, J. Power Sources 93 (2001) 87.
a) J. Cho, Y. J. Kim, T.-J. Kim, B. Park, Angew. Chern. Int. Ed. 40 (2001) 3367; (b) Y.
. Park, personal communication.
a) J. Cho, Y. J. Kim, B. Park, Chern. Mater. 12 (2000) 3788; b) J. Cho, Y. J. Kim, B.
ark, J. Electrochern. Soc. 148 (2001) A1110.
. Cho, C.-S. Kim, S.-I. Yoo, Electrochern. Solid-State Lett. 3 (2000) 362.
. Chen, J. R. Dahn, Electrochern. Solid-State Lett., 5 (2002) A213.
a) G. Amatucci, J.-M. Tarascon, L.C. Klein, J. Electrochern. Soc. 143 (1996) 1114; (b)
.H. Lee, H.S. Kim, K.T. Lee, S.M. Oh, 1 I' 11 IMLB, Monterey, CA, USA, June 2002,
bs. 102; (c) B. Markovs1y, A. Rodkin, Y. S. Cohen, E. Levi, D. Aurbach, H.-J. Kim, M.
chmidt, U. Heider, 1 J'h IMLB, Monterey, CA, USA, June 2002, Abs. 327.
. Cho, Chern. Mater. 13 (2001) 4537.
. Cho, Y. J. Kim, T.-J. Kim, B. Park, Chern. Mater. 13 (2001) 18.
a) K. Mizushima, P.C. Jones, P.J. Wiseman, J.B. Goodenough, Mater. Res. Bull. 15
1980) 783; (b) M.M. Thackeray, "The Structural Stability of Transition Metal Oxide
nsertion Electrodes for Lithium Batteries", in Handbook of Battery Materials, J.O.
esenhard Ed., Wiley-VCH, 1999; (c) A.S. Andersson, J.O. Thomas, J. Power Sources
Chapter 11
1. INTRODUCTION
Secondary lithium batteries have long been studied because they exhibit
ighest specific energy among all rechargeable batteries. Nevertheless, the
ed safety and rechargeability associated with the use of metallic lithium
prevented their widespread acceptance in the market. Unsatisfactory
s with lithium metal have directed the research towards the development e
so called "rocking-chair" or "lithium-ion" batteries (LIB), based on a
13
ition metal oxide cathode and a carbon anode. - In 1990, Sony Energytec
3
commercialized this type of batteries. Today, lithium ion batteries are
in cellular phones, notebook-size personal computers, video cameras,
to a limited extent, in electric vehicle (EV) applications. Most commonly,
egative electrode acts as a "lithium sink" and the positive LiAzBy
rode acts as a "lithium source". Layered lithium transition metal oxides
pinel lithium manganese oxide have been selected as preferred cathode
rials for lithium-ion batteries. Table 12.1 summarizes the cathode
rials that have been, or possibly will be, used in LIB. LiCo02 is the
de of choice for small-size batteries. In view of the economic and
onmental advantages, spinel-structured manganese oxides have
nstrated to be the most promising positive electrode materials for
-size batteries for EV and industrial applications, as they are cheaper, less
, and show higher safety.
In this chapter, we will describe the fundamental aspects of spinel
Chapter12
erial
o02
i02
n204
o.2Nio.s02
no.sNio.s02
P04
--- [Lil-xMxls.[Mn2]16d[04h2e
(Ga3+, Zn 2+)
(3V, 4V)
[Li]s.[ Mn2-xll6d[04]32e
(Li+, Mg 2+, Ga 3+, AP+,etc)
(3V, 4V)
[Li]s.[MxMn2-xll6d[04]32e
(Ni2+, Cu 2+, Co 3+, cr-3+, Fe 3+, etc)
(4V, 5V)
[Li]s.[Mn2]16d[Ny04-y]32e
[Li]s.[Mn2]16d[04h2e (N = F, S. 3V, 4V)
(32e) • Mn (16d) e Li
(8a)
4.8 ..........................:::::::::::::::···········--"
4.4
l€ 3.6 I
I
;> 3.2
' ,--------------
2.8 -;. - - - - - - - - - - - - - -
, u,l\fn.012
'
''
2.4 t
Otpacity (mAh/g)
150
125
100 Cl
75
O"t"ftTl"l·n·
· ;50
-= (;a}
I: 25
§. 100
.....
1"-' rc::Ll:n:t: .:a:l:
so
2.5
00
(bl
10 20
l 0
t:l
4Vte:nl
3 Vtest
Cycle number
Figure 12.3. Cycle tests of LiAI 0_ 1MnJ.904 in the 4 V region (3.0-4.3V) and 3 V region
(2.2-3.6V) at 1/3 Crate: (a) using Tosoh Mn 30 4; (b) using Chuo Mn 3 04 •
150
'@I "'(a)
:=
g
125
_e.
·c too
§.
u
75
50L--- _L L _j
0 10 20 30 40 50
Cycle number
12.4. Specific capacity vs. cycle number for the Li/lM LiPF -EC/DMC/Zn0-coated
6
Mnl.p4 cell at 113 Crate [cathode obtained by sol-gel method] ((a) at 55°C) and for
LiPF -EC/ DMC/LiNi Mnl. 0 [cathode obtained by the composite carbonate process]
6 05 5 4
°C and (c) 25°C).
Chapter12
160
0
140
e iMn,o,
f-
LiMill Mn19 0" -------
120
100
> --• 0
80
•
Li+ 0
/ AI"' eos•
'
er>+
'
!it O Ni>+,Mg2'
u 60 LiM02 nuO"
40
20
l 2 3 4
12.5. Relationship between capacity and doping level and oxidation state of doping
ions.
The lines parallel to the line yz show changes in the average Mn valence
3.5 to 4.0, and the lines parallel to the line xz variations of Li/Mn molar
Chapter12
from 0.5 to 0.8. It is evident that the stoichiometric spinel LiMn204 and
nonstoichiometric spinels, Li 2Mn409 and Li4Mn5012, are on the three
28
ces of the equilateral triangle. If the Li/Mn molar ratio is represented as
d the average Mn valence as m:
(i) the composition of each compound on the straight lines xy, xz and yz
be expressed as:
and (ii) any compound lying within the xyz triangle can be written as:
Mn: 3.5-4.0
Mn: 3.5-4.0
Lithium-rich
12.6. Semiquantitative Li-Mn-0 phase diagram for optimization of the capacity and
bility.
hanisms will also to some extent occur in metal-ion doped spinels, with a
rence in their insertion/extraction reactions (two-phase for the former vs.
phase for the latter). Capacity fading has to be referred primarily to the
el structure. Lithium-ion intercalation in the stoichiometric LiMn204
ework leads to a one-phase reaction (cubic a= 8.163 to 8.247 A) in the
e O<.x<0.55 in Li 1.xMn204 (4.0 V plateau), and a two-phase reaction
ic a= 8.154 and 8.072 A) in the range 0.55<.x<0.9 (4.2 V plateau), while
the lithium-rich compound Li 1+xMn2 04+y• with a high capacity
lization, a one-phase reaction (cubic a= 8.105-8.223 A) was observed
the entire intercalation range 0.25<Li<l.04.33.4 6 Moreover, the
eability of LiMn204 critically depends on the operating temperature.
39
n 204 loses less capacity during cycling at 0°C. In situ XRD experiments
irmed that split peaks were still observed after 50 cycles at 0°C
esponding to two phases in the high voltage region), indicating that the
phase structure is more stable for lithium ion insertion/extraction at this
erature. On the other hand, during cycling at 50°C, split peaks were
ly detected and each diffraction line shifted almost continuously to
er diffraction angles, indicating a continuous contraction of the unit
Figure 12.7 shows the cell lattice parameter changes as a function of
• • •
•
• •
='" 8.20 After cycling
i
s
.
•• •
•
8.16
!a0
1.1
Q,j
:=
8.12
8.08
• •
••
0
1a 8.04 0
.1
8.00
0.0 0.2 0.4 0.6 0.8 1.0
X inLi,Mn2 04
e 12.7. Variation of the cubic lattice parameter as a function of oxidation depth of spinel in
st charge and after 50 cycles at 50°C (113 C rate).
Spinel Cathode Materials for Lithium-Ion Batteries 371
ation depth (in the first charge and after cycling at 50°C for 50 cycles).
hermore, the lattice parameter, a, decrease from 8.246 A of original
n204 to 8.235 A after cycling at 50°C. Both results are consistent with a
ge to a defect structure, in which more Mn vacancies are present in 16d
, e.g., "lithium-rich" or "vacancy-rich" spinels are formed. Mn
lution and structural changes are facilitated at high temperature.47• 48
arently, all above proposed mechanisms are mainly associated with the
al structure and morphology. Accordingly, many researches have been
ed to these aspects. The most common way to stabilize the structure is by
ng with metal ions, as previously discussed. A second way is to reduce
the issolution by minimizing the LiMn20Jelectrolyte interface through
51
the
49
hesis of spinels with small surface area; (a,b),so coating with Lh0.B203,
27
or LiCo02 52; "'b using a salt containing less HF, e.g. LiBF4 " or Li
53
53
alatoborate (LiBOB) b instead of LiPF6; or using (CH)3SiNHSi(CH)354
h is able to react with water traces to act as a scavenger for HF, thus
ressing Mn dissolution.
140
I
!;>
·
8
100
0 50
Cycle number
12.8. Effect of the surface area on the cycle profile of spinel electrode materials at 50°C
2 2
rate). (a) LiMnz04 (3.0 m /g), (b) LiuMnz04.19 (6.3 m /g) and (c) LiuMn2 04.17 (1.2
. .,
'O l
0
0 oxygen defect spinel
8.. 100 •
e o·o·o·o·o·oooo
· o·
·o
o•o•o••••
§""' 50
u
QL---._-- -- -- -- ----L---
0 10 15 20
Cyclfl number
4.4
A
4.0
l.1
4.4
B
4.0
.u
l.1
u
10 40 0 80 100 110 10
Capacity (mAlt/g)
It has been demonstrated that the capacity loss for stoichiometric spinel
n 204 is mainly due to the unstable two-phase region of the 4.2 V plateau.
hown above, this two-phase characteristic, and the capacity fading, can be
3
ressed by making the sample lithium-rich (one-phase model)? Moreover,
57
ree-phase model was also proposed in the literature. Recent research d
on systematic in situ XRD studies has shown that the phase transition
cally depends on the oxygen stoichiometry, this affecting the cycling
55
ile. When a lithium-rich compound is heat-treated to yield an
en-deficient compound, it shows clear evidence of two biphasic regions in
the 4.0 V and 4.2 V plateaus, and significant capacity fading was observed
oth plateaus. In the compound with no oxygen deficiency, the two-phase ns
were suppressed, and the capacity fading was also reduced. These results ide
strong evidence relating the capacity fading during cycling to the itions
among three cubic phases.
58
is in agreement with Yamada's data. However, after an
en-deficiency was created by heat treatment, Li1.05Mn 2 03.97 (sample D)
s clear exothermic and endothermic peaks. The peak amplitude increases
increasing degree of oxygen deficiency. Therefore, the disappearance of
hase transition observed in the lithium-rich samples was actually due to
limination of the oxygen-deficiency. As soon as the oxygen-deficiency
created, the phase transition reappeared. Therefore,
the
en-deficiency is the primary condition and the content of lithium is only a
ndary factor.
l U
---
100
so c oo lin g
0 A
-SO
-100
-ISO
H - -eati ng B
--
200
D cooli ne
'-"
100
.. .. .. ., .
0:1
OJ
-100
-200
_, . . . ,. ._.._ ..L... -'
-
U ..ati
- :ng
-300 ._ ....
-30 -20
_.__
-10 0 10 20 30 40 so
Temperature (C)
e 12.11. DSC traces between -30°C and 50°C for four kinds of Li 1±yMn204 ±s spinels:
le A (Li0.99Mnz04.oz). Sample B (Lio.9sMnz03.97), Sample C (Lii.osMnz04.IO), and Sample
.osMnz03.92)-
Spinel Cathode Materials for Lithium-Ion Batteries 377
. SUMMARY
FERENCES
M. Armand, in Materials for Advanced Batteries, D.W. Murphy, J. Broadhead and C.H.
Steele, Eds., Plenum Press, New York, 1980, p.145.
T. Nagura, K. Tozawa, Prog. Batteries Solar Cells 9 (1990) 209.
K. Brandt, 1hiMLB, Boston, MA, USA, May 1994, Abstract FRI-30.
J. Hunter, J. Solid State Chern. 39 (1981) 142.
M. M. Thackeray, A. De Kock, M. H. Rossouw, D. C. Liles, D. Hoge, R. Bittihn, J.
Chapter12
Y.-K. Sun, Y.-S. Lee, M. Yoshio, K. Amine, Electrochem. Solid-State Lett. 5 (2002) A99.
Y. Xia, T. Sakai, T. Fujieda, M. Wada, H. Yoshinaga, J. Electrochem. Soc. 4 (200l)A9.
T. Ohzuku, K.Ariyoshi, S.Yamamoto, Y. Makimura, Chern. Lett. 1270 (2001).
M. Yoshio, H. Noguchi, Y. Todorov, Y. Hideshima, Denki Kagaku 66 (1998) 1198.
H. Noguchi, H. Nakamura, M. Yoshio, Electrochemical Society Proceedings Volume
2000-21, M. Doyle, E. Takeuchi, K.M. Abraham, Eds., p. 18,2001.
Y. Xia, M. Okada, M. Nagano, M. Yoshio, Batteries for Portable Applications and
Electric Vehicles, C. F. Holmes and A. R. Landgrebe, Eds., The Electrochem. Soc. Inc.,
Pennington, NJ, 1997, PV97-18, p. 494.
Y. Xia, M. Yoshio, J. Electrochem. Soc. 143 (1996) 825.
Y. Xia, M. Yoshio, J. Power Sources 57 (1995) 125.
R.I. Gummow, A. de Kock, M.M. Thackeray, Solid State lonics 69 (1994) 59.
Y. Xia, M. Yoshio, J. Electrochem. Soc. 144 (1997) 4186.
M.M. Thackeray, L.A. de Picciotto, A.de Kock, P. Johnson, V. Nicholas, K. Adendorff,
Mater. Res. Bull. 21 (1987) I.
Y. Gao, J.R. Dahn, J. Electrochem. Soc. 143 (1996) 100.
Y. Xia, Y. Zhou, M. Yoshio, J. Electrochem. Soc. 144 (1997) 2593.
D. H. Jang, Y. J. Shin, S.M. Oh, J. Electrochem. Soc. 143 (1996) 2204.
D. Guyomard, J. M. Tarascon, Solid State lonics 69 (1994) 222.
G Pistoia, A. Antonini, R. Rosati, D. Zane, Electrochim. Acta 41 (1996) 2683.
S. J. Wen, T. J. Richardson, L. Ma, K. A. Striebel, P. N. Ross, Jr., E. J. Cairns, J.
Electrochem. Soc. 143 (1996) L136.
M. M. Thackeray, Y. Shao-Horn, A. J. Kahaian, K. D. Kepler, E. Skinner, J. T. Vaughney,
and S. A. Hackney, Electrochem. Solid-State Lett. I (1998) 7.
J. M. Tarascon, W. R. McKinnon, F. Coowar, T. N. Bowmer, G Amatucci, and D.
Guyomard, J. Electrochem. Soc. 141 (1994) 1421.
Y. Xia, H. Noguchi, M. Yoshio, J. Solid State Chern. 119 (1995) 216.
X. Wang, Y. Yagi, Y-S Lee, M. Yoshio, Y. Xia, T. Sakai, J. Power Sources 97-98 (2001)
27.
T. Inoue, M. Sana, J. Electrochem. Soc. 145 (1998) 3704.
Y. Xia, N. Kumada, N. Yoshio, J. Power Sources 90 (2000) 135.
M.Yoshio, X. Wang, H. Nakamura, 203rd ECS Meeting, to be presented (2003).
G G Amatucci, C.N. Schmutz, A. Blyr, C. Sigala, A.S. Gozdz, D. Larcher, J. M.
Tarascon, J. Power Sources 69 (1997) II.
G G Amatucci, A. Blyr, C. Sigala, P. Alfone, J. M. Tarascon, Solid State lonics 104
(1997) 13.
.Cho, GB.Kim, H.S.Lim, C.-S.Kim, S.-In.Yoo, Electrochem. Solid-State Lett. 2
(1999) 607.
S.-C.Park, Y.-S.Han, Y.-Skang, P.S.Lee, S.Ahn, H.-M.Lee, J.-Y.Lee, J. Electrochem.
Chapter12
B. Ammundsen
Pacific Lithium New Zealand Limited, PO Box 90725, Auckland, New Zealand
. INTRODUCTION
The first commercial lithium-ion batteries used LiCo02 as the cathode
e material, and this material continues to be used in most lithium-ion
ries manufactured despite the high cost and safety hazards associated
cobalt. Apart from LiCo02, only the isostructural nickelate LiNi02, and
particularly the Co-substituted nickelate LiCoxNi 1_x02, have been
idered to have sufficient energy density and cycling stability to be of
mercial interest. However the nickelates present safety and toxicity
erns which are still greater than LiCo02• Manganese oxides offer lower
and toxicity than cobalt or nickel, and have been demonstrated to be
on overcharge. A lithium manganese oxide based cathode should
fore, at least in principle, provide significant technological advantages
ithium-ion system over LiCo02, LiNi02, or LiCoxNi 1_x02•
One of the reasons that LiCo02 functions so well as a lithium-ion
de material is that it has a well-ordered and stable layered crystal
ture which is easily prepared and handled in air, and enables a fast and
sible lithium intercalation. Nickelates also have a layered crystal
ture, although the synthesis conditions are more difficult, requiring
nation and sintering under a controlled oxygen-rich atmosphere. The
readily prepared lithium manganese oxide is LiMn 204, which does not
a layered crystal structure but a spinel structure (refer previous chapters
is volume). Although the spinel crystal structure of LiMn 204 permits
intercalation of lithium ions, its stoichiometry means that it has a lower
city than LiCo02• Furthermore, the stoichiometric spinel LiMn 204
s a large capacity fade with cycling. Lithium-rich spinel compositions
xMn2_x]04 or materials substituted with cations such as Co, Cr, Al,
Chapter 13
e 13.1. Crystal structures of layered LiCo02 (left) and o-LiMn02 (right). Small white
es represent Li ions, small black spheres represent transition metal ions, and large gray
es represent oxygen.
Chapter 13
e 13.2. Crystal structure of spinel LiMn 204 (refer Figure 13.1 for key).
8 11
el-like structure over subsequent cycles. -
Figure 13.3(A) in Section 13.3.3 shows the consequence of the phase
formation toward the spinel structure on the discharge voltage profile of
pical LiMn02 material (here shown for o-LiMn02 prepared by high
erature processing). The transformation toward the spinel structure is
mpanied by evolution of a voltage profile characteristic of spinel
ganese oxide, which drops sharply from approximately 4 V vs. Li/Li+ to
below 3 V near the Li 0.5Mn02 stoichiometry, i.e. at the halfway point of
harge. The voltage step in the spinel structure is associated with a change
ntercalation mechanism. For O<x<l in spinel LixMn 204, Li ions are
ted directly into tetrahedral sites with high binding energy for lithium, a
ess which takes place around 4 V vs. LilLi+. For l<x<2 in spinel
n204, Li ions are inserted into octahedral sites to form a tetragonally
rted rock salt structure of stoichiometry Li2Mn204 (=LiMn02) in a two
e process which occurs at a potential of just below 3 V vs. LilLi+. Such a
drop in voltage midway through discharge poses problems for power
agement in actual systems. Furthermore, while a crystalline spinel
ture provides fast diffusion for lithium into tetrahedral sites at around
conversion of LiMn02 materials can lead to a less ordered spinel
ture with lower 4 V capacity and poorer rate capability. Additionally,
wo-phase process that occurs at around 3 V typically has much poorer
tics than the insertion process at 4 V, manifesting as polarization of the
ge and incomplete discharge as the current rate is increased. This also
s problems for practical constant-power applications that require higher
ent rates as the voltage lowers toward the end of discharge.
On the other hand, spinel-type cathodes formed in situ by cycling
n02 materials often show better cycling stability over a large lithiation
e (up to nearly LiMn02 or LizMn204) than spinel materials prepared ex
by conventional methods. The retention of high capacity- in many cases
e than 200 cycles have been reported - is surprising considering that
entionally prepared LiMn2 04 spinel materials almost always show large
ersible capacity loss over only a few cycles when cycled over both the 4
3 V regions. Data from high resolution electron microscopy have shown
the cycling stability may largely result from the formation of antiphase
domains (at a length scale of 20-50 nm) in cathode materials starting
12 13
LiMn02 phases. • The antiphase nanodomains arise from nucleation
he spinel structure at different points in the original crystal structure,
wed by growth of the spinel domains from the point of nucleation to
at phase boundaries where there is incoherence of the spinel crystal
ture between neighboring domains. The formation of this nanodomain
ture, in which there is a high density of phase boundaries, provides an
icity which can better accommodate transformation strains and
otropic changes in volume arising from cycling the material over such a
voltage range. Some cation inversion (Mn ions in tetrahedral Li sites)
ed Manganese Oxides as Cathodes 387
also been found in the spinel structure obtained from LiMn0 2 phases,
13
h is also considered to contribute to the cycling stability, but would
an adverse effect on rate capability.
It has been pointed out that substitution of LiMn02 with species
ng strong octahedral site stabilization energies could hinder the second
e of the layered-to-spinel transformation, i.e. the stage that requires
7
erative re-arrangement of transition metal ions in the structure. This
oach has been shown to be partially successful for materials made by
soft chemistry routes and high temperature processing (see below).
rally show better cycling stability over a wide voltage window than the
rials prepared at low temperature. Gummow et al. prepared o-LiMn02
ucts by solid state reaction of y-Mn02 with LiOH at 600- 620°C in an
n atmosphere in the presence of carbon black as a reducing agent, and
25
ned capacities of up to 160 mAhlg over 4.45- 2.5 V. Davidson et al.
ned capacities up to 140 mAhlg over 4.4 - 2.0 V for HT o-LiMnOz
26 prepared by reaction of Mn02 and Li2C03 at 800- 1000°C under
ucts
n. Croguennec et al. examined in detail the effect of crystallinity on
lectrochemical performance of HT o-LiMn02 phases, using solid state
ions between Mn203 and LiOH.H20 at 700°C designed to produce
7 28
rials with very small crystallites (<0.5 )..tm).Z • Capacities up to 200
29
/g were obtained for materials with crystallite sizes of around 0.3 )..tm.
small-crystallite-size o-LiMn02 phases were characterized by cationic
der identified as stacking faults in the crystal structure, which were
30
idered to be important for the electrochemical activity. However Jang
. later reported high capacities for o-LiMn02 materials prepared by a
olled high temperature synthesis process giving small crystallite size
low content of stacking faults, indicating that crystallite size is the most
al parameter in determining the electrochemical performance of o
31
n02 materials.
The problem for application of o-LiMn02 as a cathode material
ins its instability in transforming to a disordered spinel structure during
ng. To the best of our knowledge the stabilization of the o-LiMn02
ture against transformation to a spinel-type structure has never been
ted. The speed of transformation to spinel, and the rate capability
lem on the 3 V plateau, can both be controlled to some extent by
fying the microstructure of the o-LiMn02 material. However the sharp
ge drop in the middle of discharge remains a problem for applications.
ns should be able to perform this function, but only AI, Cr and Ga have
shown to work in practice. This is at least partly because the high
erature conditions using inert atmosphere are not suitable for the use of
tituting elements like Co and Ni, which tend to reduce to divalent oxides
g the attempted synthesis. AI, Cr and Ga have in common that the
lent oxidation state is stable under the synthesis conditions.
The stabilization of the layered structure over the orthorhombic
ture in a high temperature process was first reported by Davidson et al.,
38 3
Cr substitution. Substitution of as little as 5% of the Mn by Cr + in
n02 resulted in crystallization of a layered monoclinic phase by a simple
state reaction at 1000°C under argon. The original report was unclear in
ifying the crystal structure of the material, but subsequent work
rmed that it was indeed layered with the same monoclinic C2/m
ture reported for LiMn02 obtained by ion-exchange.39 It was reported
the Cr-substituted material seemed to convert to a spinel-type structure
cycling, but extended cycling data was not shown.
Later, Jang et al. reported that monoclinic layer-structure LiMn02
also be crystallized in a stable form at high temperature (ca. 950°C) by
ituting 5-10% of the Mn with AI.40 The materials were synthesized
a precursor powder prepared by atomizing and freeze-drying a
ension of Al and Mn hydroxide, coprecipitated from nitrate precursors,
OH solution. During cycling, the Al-substituted materials were again d
to convert to a spinel-type structure. Despite the transformation of the
ture and consequent evolution of the discharge voltage profile, however,
Al-substituted cathodes could be cycled over the voltage range 4.4 -
with good retention of capacity, and showed unusually good cycling
lity at 55°C.41
The formation of monoclinic layered oxides by high temperature
42 3
ssing was investigated in further detail by Ammundsen et al. .4
red monoclinic LiMn02 was successfully prepared by substituting
een 3 and 10% Al 3+, Ga3+, or Cr3+ in simple low-cost solid state44
ods of preparation which could be implemented on an industrial scale.
monoclinic unit cell parameters for the substituted materials showed a
matic decrease in unit cell volume and in the extent of the monoclinic
rtion by comparison with non-substituted LiMn02 prepared by ion
ange.
AI- and Cr-substituted materials were evaluated as cathodes to
mine whether stabilization of the layered crystal structure in a well
ed highly crystalline material through chemical substitution could
late to increased stability of the layered structure against transformation
42
inel during cycling. '43 Figure 13.3 shows the discharge voltage curves
monoclinic layered materials substituted with 5% AI and 3% Cr,
ared with the o-LiMn02 material prepared under the same high
erature processing conditions in the absence of a substituting element.
Chapter 13
o-LiMn0
2
(B)
4.0
AJ-doped man. LiMn02
>
-;; 3.5
g
C)
.l!l -----1·····
:
3.0
'iii
0 2.5 (C)
2.0 Cr-doped man. LiMn0
2
4.0
Q) 3.5
C)
3.0
'iii
0 2.5
13.3. Discharge voltage profiles for LiMn0 2 phases prepared by high temperature
sing: (A) o-LiMnOz, (B) monoclinic layered LiMn02 with 5% AI substitution,
onoclinic layered LiMn02 with 3% Cr substitution. Data recorded in lithium cells
between 4.4 and 2.0 V at 55°C, 30 mA/g current rate. Cycle numbers are indicated.
ed Manganese Oxides as Cathodes 395
ting from the ion-exchange process. During Na-Li exchange, the Mol>
ts glide with respect to one another, creating the stacking faults.
However substitution of a fraction of the Mn with Co or Ni improves
the ion exchange kinetics and the electrochemical properties. A layered
rial of composition Li2,3 [Mn0.85Co0.15]0z showed a smoothly sloping
ge profile and good cycling performance without transforming to
14
el. Substitution with Ni allowed a successful ion exchange at room
erature resulting in a crystalline layered phase with a novel T2 type
16
ture. T2 Lizi3[Ni 113Mnz,3]0z shows Ni 113Mnz,3 honeycomb type
ring in the transition metal planes and unusual occupation of tetrahedral
47
layer sites by Li ions. On first charge the material converts to the 02
ture. The material can then be cycled in the 02 structure over hundreds
cles without collapse of the layer structure, with capacities of up to 180
lg if cycled between 2.5 and 4.6 V. However, unlike the Co-substituted
rial, only one-third Li per formula unit Li213 [Ni113Mn213 ]02 (i.e. half the
able Li) can be removed from the lattice on charge. In order to obtain a
city of 180 mAhlg, an additional 1/3 Li needs to be inserted into the
e on discharge (requiring an additional source of Li not normally
able in lithium-ion cells). Furthermore, although the 1/3 Li removed on
charge is reversibly re-intercalated at around 3.8 V, the additional 113 Li
ercalated at around 2.8 V. Although the material does not convert to a
l, the voltage profile is almost the same as that of 03 phases that do
ert to spinel. In the case of Li 213 [Ni113Mn 213]02 the two steps with a 1 V
in the middle are due to different redox couples being involved. The 3.8
ercalation is related to reduction of nickel, while the 2.8 V intercalation
sponds to reduction of Mn4+ to Mn3+.
Substitution of Co for Ni in Li213 [Nil/3Mnz,3]0z has been found to
t in formation of a stacking faulted 02 structure with incomplete Li-Na
ange.48.49 However substitution of Co for both Ni and Mn allows ion
ange to go to completion to give materials of composition Li213 [CoxNi113.
213.x12]02, and allows increased amounts of Li to be extracted on first
e, enlarging the reversible capacity available in a sloping discharge
17
een 4.5 and 3.8 V.
The 02 layered cathode materials are still far from commercialization
o the complicated methods of preparation, and lower capacity at higher
nt rates in their current state of development. Although the substitution i
and/or Co improves the rate performance, the resulting T2 and 02
rials need to be cycled over a wide voltage range and require an
ional source of Li to capture their full capacity, which is delivered in
teps with a large voltage drop mid-discharge.
The structural studies carried out on the 02 and T2 layered oxides do,
ver, provide interesting data on the behavior of Mn, Ni and Co ions
sharing the same oxide lattice. In particular, the data show that best
tural and electrochemical results are achieved when Mn is fully
ed Manganese Oxides as Cathodes 397
lithium lithium
stoichiometric saturated
phases phases
LiMn02
e 13.4. The Li2Mn03 - LiMn02 - LiM02 solid solution system showing the
fication of phases according to their lithium stoichiometry.
All compounds within the phase triangle shown in Figure 13.4 obey
ules described by Dahn et al. for these layered rock salt solid solution
51
es. Firstly, the sum of the cation occupations on the sites in the
ition metal layers (the 3b sites in the space group R 3m) is equal to one.
ndly, the sum of the cation oxidation state times the cation occupations
e transition metal layer sites equals three.
It should be noted that the results of different investigations into solid
ions over the past decade have shown considerable variation, depending
gly on the conditions used to prepare the materials. These variations can
ed Manganese Oxides as Cathodes 399
4.5
Q)
Cl 3.5
ctl
15
> 3
2.5
Capacity (mAh/g)
e 13.5. First and second cycle charge-discharge voltage curves for a material of
osition Li[Li0.2Cr0.4Mno.4]02 cycled between 4.4 and 2.5 V (3 mA/g current rate, 25°C).
4.5
Q)
Ol 3.5
.!9
0
> 3
2.5
e 13.6. First and second cycle charge-discharge voltage curves for a material of
osition Li[Lio.05Co0.15Mno.4Ni0.4 ]02 cycled between 4.4 and 2.5 V (30 mA/g current rate,
.
FERENCES
. Capitaine, P. Gravereau, C. Delmas, Solid State lonics 89 (1996) 197.
. R. Armstrong, P. G. Bruce, Nature 381 (1996) 499.
M. Tabuchi, K. Ado, H. Kobayashi, H. Kageyama, C. Masquelier, A. Kondo, R. Kanno,
. Electrochern. Soc. 145 (1998) lA9.
. Ceder, S.K. Mishra, Electrochern. Solid-State Lett. 2 (1999) 550.
. Ceder, A. Vander Ven, Electrochirn. Acta. 45 (1999) 131.
. Wolverton, A. Zunger, J. Electrochern. Soc. 145 (1998) 2424.
. Reed, G. Ceder, A. VanDer Ven, Electrochern. Solid-State Lett. 4 (2001) A78.
. Vitins, K. West, J. Electrochern. Soc. 144 (1997) 2587.
. M. Koetschau, J. R. Dahn, J. Electrochern. Soc. 145 (1998) 2672.
Y. Shao-Horn, S. A. Hackney, A. R. Armstrong, P. G. Bruce, R. Gitzendanner, C. S.
ohnson, M. M. Thackeray, J. Electrochern. Soc. 146 (1999) 2404.
.-J. Hwang, H.-S. Park, J.-H. Choy, G. Campet, Chern. Mater. 12 (2000) 1818.
Y.-I. Jang, B. Huang, H. Wang, D. R. Sadoway, Y.-M. Chiang, J. Electrochern. Soc. 146
1999) 3217.
H. Wang, Y.-I. Jang, Y.-M. Chiang, Electrochern. Solid State Lett. 2 (1999) 490.
. M. Paulsen, C. L. Thomas, J. R. Dahn. J. Electrochern. Soc. 146 (1999) 3560.
. Delmas, J. Braconnier, A. Maazaz, P. Hagenmueller, Rev. Chirn. Miner. 19 (1982)
43.
. M. Paulsen, C. L. Thomas, J. R. Dahn, J. Electrochern. Soc. 147 (2000) 861.
. Lu, R. A. Donaberger, C. L. Thomas, J. R. Dahn, J. Electrochern. Soc. 149 (2002)
l083.
. Chen, M.S. Whittingham, J. Electrochern. Soc. 144 (1997) L64.
M. H. Rossouw, M. M. Thackeray, Mater. Res. Bull. 26 (1991) 463.
. S. Johnson, S. D. Korte, J. T. Vaughey, M. M. Thackeray, T. E. Bofinger, Y. Shao
orn, S. A. Hackney, J. Power Sources 81-82 (1999) 491.
. Paik, C. P. Grey, C. S. Johnson, J.-S. Kim, M. M. Thackeray, Chern. Mater. (2002) in
ress.
. Ohzuku, A. Ueda, T. Hirai, Chern. Express 7 (1992) 193.
. N. Reimers, E. W. Fuller, E. Rossen, J. R. Dahn, J. Electrochern. Soc. 140 (1993) 3396.
Koetschau, M. N. Richard, J. R. Dahn, J. B. Soupart, J.-C. Rousche, J. Electrochern.
oc. 142 (1995) 2906.
. J. Gummow, D. C. Liles, M. M. Thackeray, Mater. Res. Bull. 28 (1993) 1249.
J. Davidson, R. S. McMillan, J. J. Murray, J. E. Greedan, J. Power Sources 54 (1995)
32.
. Croguennec, P. Deniard, R. Brec, A. Lecerf, J. Mater. Chern. 5 (1995) 1919.
. Croguennec, P. Deniard, R. Brec, P. Biensan, M. Broussely, Solid State lonics 89
1996) 127.
. Croguennec, P. Deniard, R. Brec, J. Electrochern. Soc. 144 (1997) 3323.
Chapter 13
1 2
Byungwoo Park\ Yang Jeong Kim , and Jaephil Cho
1
School of Materials Science and Engineering, and Research Center for Energy
Conversion and Storage, Seoul National University, Seoul, Korea
epartment of Applied Chemistry, Kumoh National Institute of Technology, Gumi, Korea
.1.1. Introduction
Cathode Material
LiCo02
LiAlo.25Coo.1s02
LiMgo.osCoo.9s02
e Li+
c Co +
3
0 oz-
B
A
c
--A
(a)
2.82 I
-$ 2.81
1::1 II
14.3
-<
';:;' 14.2
14.1
I
14.0
(c)
60
040 ci
II I+II I
II s:::
izo
0
re 14.2. Lattice constants a (a) and c (b) as a function of the lithium concentration, x,
,.Co02• (c) Phase diagram for Li..Co02 (Ref. 8).
4 Chapter 14
(a)
ure 14.3. Severely damaged LiCo02 particles from a cycled cathode showing
rofractures (arrows) and a strong diffraction contrast from the internal strain and
ended defects (Ref. 1).
0.5 •
0.4
. "s'
- ";
,Q
u
Q
0.3
ure 14.4. Cobalt loss (percentage) as a function of cycling voltage (Ref. 20).
6 Chapter 14
20 30 40 50 60 70
Scattering Angle 29 (degree)
4.0
c
,-...
=n
s0 3.5 LiCoo2
> Crate= 0.1 C (= 14 rnA/g)
u
3.0
2.5 L-,J...._J.._..L.,_..J...._...J........I........L......J........L.---L.---L.---L---L----I.......JL........L-..L.-.L..-.J
o w rn
Capacity (mAh/g)
ure 14.6. Voltage profiles of the LiCo0 2 cathode at a rate of 0.1 C (4.4- 2.75 V).
180
r
170 Charge cut-off voltage
160
+ 4.4V
! A 4.2V
e 4.1V
150
140
-
=ii' 130
LiCo02
120
u
cu 110
Oil
-"='
100
s"' 90
80
Crate= 0.5 C
70
0 10 20 30 40 50
Number of Cycle
ure 14.7. Cycle-life performance of the LiCo02 cathode with a charge cutoff voltage
4.4 V, 4.2 V, and 4.1 Vat the rate of 0.5 C. The discharge cutoff voltage is 2.75 V.
C/LiCo02
LiCo02
Charge Rate: 0.5 C
Discharge Rate: 1 C
ure 14.8. Plots of the capacity retention for C/LiCo02 cells as a function of cycle
ber. The charge and discharge rates are 0.5 and I C, respectively.
thodes Based on LiCo02 and LiNi02
419
C/LiCo02
0.2C
- -3 : : : :: - - - _
_- - -- - _- _- : :
0.5C
'
..
IC ' ''
... ' \
.... \ 1\
... ... 2C ', \ \I
... \
1
,u
II
JC ', 1\1
,u
'
'
2.5 J--L J--L J--L
0 20 40 60 80 100 120 140 160
Capacity (mAh/g)
ure 14.9. Rate capabilities of LiCo02 at various discharge rates of 0.2, 0.5, 1, 2, and 3
The charge rate was fixed at 0.5 C.
1500
1000
_, 500
d
0
="' ""'-500
u
1-1000
3.2 3.4 3.6 3.8 4.0 4.2 4.4
Voltage (V)
ure 14.10. Cyclic voltammograms of a LiCo02 cathode. Scan rate: 0.02 mY/sec. The
bols Hand M denote the hexagonal and monoclinic phase, respectively.
0 Chapter 14
4.4 &.............. A
-·········· y=O.S
o----Q o---o- ·-.......A..... ,.......
E =--o. ··A···AA
4.0 y = o.ff·-oo-o.--o-o. <e
A
y=O •0
3.6
u
! 3.2
0.4 0.6 0.8 1.0
ure 14.12. Open-circuit voltage (OCV) as a function of the Li content in pure LiCo0 2
in Al-doped materials (Ref. 3).
--Charge at 55°C
200 -o- Discharge at 55°C
.......... Charge at RT
-o-Discharge at RT
Cycle Number
ure 14.13. Specific capacity vs. cycle number for LiA10.25Co0.7502, tested against aLi
2
al anode at 0.4 mA/cm between 2.0 and 4.5 V at room temperature and 55°C (Ref. 3).
2 Chapter 14
.2.1. Introduction
Cathode Material
LiNi02
iNio.zsCoo.soMno.2s02
"' Z'
'§ LiNi02
.cl
1;
'-'
B
j ;:::;
=0
15 20 25 30 35 40 45 50 55 60 65 70
Scattering Angle 28 (degree)
4.5
4.0
"'
i >
3.5
=
u"'
3.0 Crate= 0.1 C (= 18 rnA/g)
200
Io:l
150
100
o:l
u
r"';
-5 50
Crate= 0.2 C (= 36 rnA/g)
i5"'
0
0 10 20 30 40 50 60 70
Cycle Number
ure 14.15. (a) Initial voltage profiles of LiNi0 2 between 4.3 and 2.75 V at a rate of 0.1
8 rnA/g), and (b) cycle life ofLiNi02 at a rate of 0.2 C (36 rnA/g).
2000
1000
=' .
' - 0
.
8 -1000
-2000
3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4
Voltage (V)
ure 14.16. Cyclic voltammograms ofLiNi0 2• The scan rate was 0.02 mY/sec.
14.50
0
"' 14.45
j
:r:l 14.40
u
..
"'
: 0
14.35
.'l.U.. 14.30
14.20
2.90
2.89
2.88
0
2.87
<:!
=s0 2.86
" '
u=
..
"'
: 0
2.85
2.84
.'.lU..
2.83
2.82
M M M U U
X in Lil·XNi02
.3.1. Introduction
4.S
4.0
3.0
u
2.S
2.0
0 so 100 1SO 200 2SO 300 3SO 400
Capacity (mAh/g)
180
!
t>
-
=u - 140
"'., 130
0 11
.. .
120
s"' 110
10 20 30 40 so 60 70
Cycle Number
ure 14.18. (a) Initial charge and discharge curves for various metal-oxide coated
o02, and comparison with uncoated LiCo02 (first cycle). (b) Cycle-life performances
ZrOz-, Al203-, TiOz-, B 203-, Si02-coated, and uncoated LiCo02. The cells were
ed at a rate of 0.1 C for the first two cycles, followed by a 0.5 Crate between 4.4 and
Vat 21oc.
, -..
o<
,="'
'-'
Q
14.25
u
Col
14.20
';I
14.15
....:l
14.10
14.05
14.00
2.818
2.816
, -..
o<
,= <:::1
' '
... 2.814
2.812
Q
u
';I
C!l
...
Col
2.810
....:l
2.808
600
.
.
200 .••.••••.•.••...•..••.••.••••.••-<] Zr0
2
....................--
ZrOrcoated LiCo02
u
0%
deintercalation
hexaQ9(1al
UCo02
.4.1. Introduction
50
2
0.2mA/cm
10
40 0 -- --
0 20 40 60 80 100
30
c ---o--- Discharge Capacity
a
·20 -o- Charge Capacity
10
60
(b) 30 nm Coating 50
50 40
30
20
to
0
0 20 40 60 80 100
Cycle Number
ure 14.22. Cycle-life performance (charge and discharge capacities vs. cycle number)
(a) bare, and (b) 30 nm-thick Al203 -coated LiCo02 thin films between 4.4 and 2.75 V
2
0.2 mA/cm (= 6 C). The inset shows the same cycle-life performance, including the
a capacity from the constant-voltage mode.
0... 3 2
80
,-.._ ··o···················O····················o···················O
=
'-'
0
;I 60
.=
fj o Bare LiCo02
0 -D
40
_e. 0 0
·c'II
§' 20
u
0
50
(b)
_e. Q-···················9··········· 0
·c
a
30 o Bare LiCo0
'II 2
1,.:.1.1 20
' II
.. c :
(,J
Q"' 10
:...!.
·..a.. 0
200 400 600
2
800
Charge/Discharge Rate (jLA/cm )
ure 14.23. (a) Capacity retention vs. current rate after 100 cycles, and (b) initial
charge capacity vs. current rate of uncoated and Ah03-coated LiCo02 thin films.
2
0.2 mA/cm • In contrast, the change in the voltage profiles of the
zOrcoated LiCo02 thin films during 80 cycles is more gradual.
wever, compared with the cycling profiles of bare LiCo02 thin films, a
ht increase in cell potential is observed in the initial cycle of A}z03-
ted cathodes. This is attributed to the insulating effect of the
minum-oxide layer. From the 2nd cycle on, the polarization in the
rge and discharge processes of coated sample is not significant, as the
Al-0 coating layer on LiCo0 2 acts as a solid electrolyte (details in
tion 14.4.2.2).
4 Chapter 14
E 4.0
-a;c:J '.
=
3.6 \
'·
I '.
U I '
3.2 \ ·,
·,·,
' ·,
•. ·,, •• I
2.8 '··-··--- ·-·-·-·-·-· ....... 0 '
4.4
E 4.0
-a;c:J
= 2
£ 3.6 0.2mA/an
'
=
u --1st Cycle
\
\
- - - 2nd Cycle
3.2 ·····20th Cycle
I
i
-·-··40th Cycle
I
-··-··80th Cycle
2.8
0 10 20 30 40
Capacity (jLAh/cm1
ure 14.24. Voltage profiles of (a) uncoated, and (b) 30 run-thick Al 20rcoated LiCo02
2
s between 4.4 and 2.75 V at a rate of 0.2 mA/cm • At each charge/discharge cutoff
, the cell voltage was potentiostated until the current decreased to 0.02 mA/cm2 .
the uncoated LiCo02 thin film is only -25% after 100 cycles, as shown
Figure 14.25(b). This is possibly because the A}z03 coating layer acts
a solid electrolyte with a low electronic conductivity and a reasonably
h Li-ion conductivity. For example, the Li-ion conductivity of a
7 68
Liz0-0.3A}z03 solid electrolyte is -10- S/cm at room temperature.
50
"s: (a) Initial Discharge Capacity
l
a
· o
40
45
< II
c . 35
till
30
- .=...
s 25
1 ...... 20
-o---
100
(b) Capacity Retention
...,.........o-0
---o---!:c-
-6i
---a-- --------Q
0 40 80 120 300
Alp3 Thickness (nm)
re 14.25. The effect of the Al20rcoating thickness on (a) initial discharge capacity
(b), capacity retention after 100 cycles, all excluding the extra capacity from the
tant-voltage mode. The cells were cycled between 4.4 and 2.75 V at 0.2 and 0.4
2
cm •
6 Chapter 14
1200
After floating for 12 days
D
I
1000
_ I
I
. I
-
800 I
I
I
I
. 600 I
I
I
I
I
I
,0
,,
l
400 , ,
. , ' o·
,
200 Al 0 -coated LiCo0
-----o·-
--- ---
0
o--o
o-- ---
0
0 2 3 2
gure 14.26. The amount of cobalt dissolution in the electrolyte, from bare and Ah03-
ated LiCo02, after an initial charge and being immersed for 12 days at 25°C.
hodes Based on LiCo02 and LiNi02 437
50
=t: 0 ··-··-··-..........
u=
-50
--During 1st Cycle
······After 40 Cycles
-100 a' -··-·· After 80 Cycles
100
50
=5I . . .
< 1 , )
0
u
-50
-100 a'
ure 14.27. Cyclic voltammograms of (a) bare, and (b) 30 nm-thick Al203 -coated
o02 thin films, during the 1st cycle, and after the 40th and 80th cycle. The sweep rate
0.1 mV/sec between 3.0 and 4.4 V. The symbols Hand M denote the hexagonal and
oclinic phases, respectively.
14.5
14.4
,-...
·s - -········ ·········
.... 14.3
B ··•·· LiCo02 Powder
-o- Bare LiCo02 Film
"
u='
14.2 t' - -o-· Al 0 -coated LiCo0 Film
0
2 3 2
c.J
E 14.1
C'iS !
= --
14.0
atOCV
4.1 4.2 4.3 4.4
2.7 3.9 4.0
ure 14.28. Changes in the lattice constants c as a function of the cell potential for bare
Al203-coated LiCo02 thin films during the first charge, with a comparison with
der samples. For thin films, each cell was charged at a 0.1 mA/cm 2 rate to the
determined voltages, then potentiostated until the current density decreased to
2
tA/cm •
ure 14.29. Lithium diffusivities as a function of cell potential for (a) bare, and (b) 30
thick AI20rcoated LiCo02 thin films, determined by GITT during Li deintercalation
rging).
11 2
-lxl0- cm /sec after 80 cycles, while the deterioration of the Li
usivities in the coated ones remains at -3xl0-11 cm2/sec (Figure 14.30).
se results suggest that an A}z03 coating on the LiCo02 thin films can
rate the electrochemical cycling with a high cutoff voltage of 4.4 V,
nks to the suppression of cobalt dissolution by A}z03 coating.
thodes Based on LiCo02 and LiNi02 441
During Li Intercalation
gure 14.30. Lithium diffusivities as a function of cell potential for (a) bare, and (b) 30
-thick Al20rcoated LiCo02 thin films, as determined by GITT during Li intercalation
scharging).
KNOWLEDGEMENT
FERENCES
S.-T. Myung, N. Kumagai, S. Komaba, H.-T. Chung, Solid State Ionics 139 (2001) 47.
K. Kubo, M. Fujiwara, S. Yamada, S. Arai, M. Kanda, J. Power Sources 68 (1997)
553.
C.-C. Chang, J. Y. Kim, P. N. Kumta, J. Electrochem. Soc. 149 (2002) A331.
J. Cho, G. Kim, H. S. Lim, J. Electrochem. Soc. 146 (1999) 3571.
Z. Lu, D. D. MacNeil, J. R. Dahn, Electrochem. Solid-State Lett. 4 (2001) A200.
M. M. Thackeray, J. Electrochem. Soc. 142 (1995) 2558.
P. G. Bruce, Chem. Commun., 1817 (1997).
P. Koksbang, J. Barker, H. Shi, M. Y. Sakdi, Solid State lonics 84 (1996) 1.
H. Arai, S. Okada, H. Ohtsuka, M. lchimura, J. Yamaki, Solid State lonics 80 (1995)
261.
T. Ohzuku, A. Ueda, M. Nagayama, J. Electrochem. Soc. 140 (1993) 1862. A.
Rougier, P. Gravereau, C. Delmas, J. Electrochem. Soc. 143 (1996) 1168. J.
Barker, R. Koksbang, M. Y. Saidi, Solid State lonics 89 (1996) 25.
R. V. Moshtev, P. Zlatilova, V. Manev, A. Sato, J. Power Sources 54 (1995) 329.
H. Arai, S. Okada, Y. Sakurai, J.l. Yamaki, Solid State lonics 95 (1997) 275.
J.P. Peres, PhD thesis, Univ. of Bordeaux I, France, (1996).
W. Li, J. N. Reimers, J. R. Dahn, Solid State lonics 67 (1993) 123.
T. Ohzuku, A. Ueda, M. Kouguchi, J. Electrochem. Soc. 142 (1995) 4033.
J. Cho, H. Jung, Y. Park, G. Kim, H. Lim, J. Electrochem. Soc. 147 (2000) 10.
Z. Wang, C. Wu, L. Liu, F. Wu, L. Chen, X. Huang, J. Electrochem. Soc. 149 (2000)
A466.
Z. Chen, J. R. Dahn, Electrochem. Solid-State Lett. 5 (2002) A213.
Z. Wang, L. Liu, L. Chen, X. Huang, Solid State lonics 148 (2002) 335.
J. Cho, T.-J. Kim, Y. J. Kim, B. Park, Electrochem. Solid-State Lett. 4 (2001) A159.
D. Rahner, W. Plieth, M. Kloss, "Modified electrode material and its use," U.S. Patent
No. 6348259, March 30, 1999.
T. Miyasaka, "Non-aqueous lithium ion secondary battery," U.S. Patent No. 6037095,
March 30, 1998.
H.-J. Kweon, G.-B. Kim, D.-G. Park, "Positive active material for rechargeable lithium
battery and method of preparing same," U.S. Patent No. 6183911, October 29, 1999.
D. Aurbach, K. Gamosky, B. Markovsky, G. Salitra, Y. Gofer, U. Heider, R. Oesten,
M. Schmidt, J. Electrochem. Soc. 147 (2000) 1322.
W. D. Jr. Callister, Materials Science and Engineering: An Introduction, John Wiley &
Sons: New York, 4th ed., Appendix C, p. 787 (1997).
M. C. Frede!, A. R. Boccaccini, J. Mater. Sci. 31 (1996) 4375.
J. Sehgal, S. Ito, J. Non-Cryst. Solids 253 (1999) 126.
P. Birke, W. F. Chu, W. Weppner, Solid State lonics 93 (1997) 1.
J. B. Bates, G. R. Gruzalski, N.J. Dudney, C. F. Luck, X. Yu, Solid State lonics 70-71
(1994) 619.
4 Chapter 14
C. Julien, G.-A. Nazri, Solid State Batteries: Materials Design and Optimization,
Kluwer Academic, Boston, 1994.
C. Wolverton, A. Zunger, J. Electrochem. Soc. 145 (1998) 2424.
D. G. Fauteux, A. Massucco, J. Shi, C. Lampe-Onnerud, J. Appl. Electrochem. 27
(1997) 543.
B. Wang, J. B. Bates, F. X. Hart, B. C. Sales, R. A. Zuhr, J. D. Robertson, J.
Electrochem. Soc. 143 (1996) 3203.
J. B. Bates, N. J. Dudney, B. J. Neudecker, F. X. Hart, H. P. Jun, S. A. Hackney, J.
Electrochem. Soc. 147 (2000) 59.
M. Antaya, J. R. Dahn, J. S. Preston, E. Rossen, J. N. Reimers, J. Electrochem. Soc.
140 (1993) 575.
Y.-1. Jang, B. J. Neudecker, N. J. Dudney, Electrochem. Solid-State Lett. 4 (2001)
A74.
J.-K. Lee, S.-J. Lee, H.-K. Baik, H.-Y. Lee, S.-W. Jang, S.-M. Lee, Electrochem. Solid
State Lett. 2 (1999) 512.
K. A. Striebel, C. Z. Deng, S. J. Wen, E. J. Cairns, J. Electrochem. Soc. 143 (1996)
1821.
Y. Iriyama, M. Inaba, T. Abe, Z. Ogumi, J. Power Sources 94 (2001) 175.
J. D. Perkins, C. S. Bahn, J. M. McGraw, P. A. Parilla, D. S. Ginley, J. Electrochem.
Soc. 148 (2001) Al302.
P. J. Bouwman, B. A. Boukamp, H. J. M. Bouwmeester, P. H. L. Notten, J.
Electrochem. Soc. 149 (2002) A699.
. D. Perkins, C. S. Bahn, P. A. Parilla, J. M. McGraw, M. L. Fu, M. Duncan, H. Yu, D.
S. Ginley, J. Power Sources 81-82 (1999) 675.
Y. J. Kim, T.-J. Kim, J. W. Shin, B. Park, J. Cho, J. Electrochem. Soc. 149 (2002)
1337.
A.M. Glass, K. Nassau, J. Appl. Phys. 51 (1980) 3756.
J. M. McGraw, C. S. Bahn, P. A. Parilla, J. D. Perkins, D. W. Readey, D. S. Ginley,
Electrochim. Acta 45 (1999) 187.
J.-S. Hong, J. R. Selman, J. Electrochem. Soc. 147 (2000) 3190.
M.D. Levi, K. Gamolsky, D. Aurbach, U. Heider, R. Oesten, J. Electroanal. Chem.
477 (1999) 32.
D. Aurbach, D. L. Levi, E. Levi, H. Teller, B. Markovsky, G. Salitra, J. Electrochem.
Soc. 145 (1998) 3024.
ter 15
POLYANION-BASED POSITIVE
ELECTRODE MATERIALS
oire de Reactivite et Chimie des Solides, UMR CNRS 6007, Universite Picardie Jules
Verne, 33, rue Saint-Leu, 80000 Amiens, France
[001]
_f_
j-*_
j_
J-.\11
>1: l\.'12 M3
E
a)
G
Li•/Li
Na•/Na c;;;;;,;;;iiiiiiiiiiiijii;l
.2. Illustration of the cell voltages measured for a) lithium or sodium insertion into
4 h and b) lithium insertion into Fe2 (Mo04)3 and Fe2 (S04)3.
31
/extraction. The major advances were reported by the groups of
Quartoii and Goodenough and are summarized below.
4
V.,7 V.,
1 §fj (J 0
lg
31
1 8-l\.00
1 gg
M'
* )I J\.
ql )... - [A]
'
[U]
::;. :,.; .
·
:,..,.
..,.;_
"A
36 38 40
IAI
24 26 28 30 32 34
0
29 ( )- CuKa
15.5. In situ X-ray diffraction patterns recorded during a full discharge and charge cycle
h(P04)3 I Li cell.
4
3.6
zGIS I
3.2
-
2.8
fli
GIS 2.4
..=.
1
2
1.6
..e 1.2
1.5 2 2.5
2
x in
Na
Ti ..
l
(PO
) J
l
4.5 s
';
-;::
:u
c c::u
. ..c. l
1.8
0
15.7. Voltage- composition profiles recorded vs. Na+/Na during sodium
n/extraction into/from NASICON Na2+xTiM(P04h (M =Fe, Cr).
35
(P04)3 as a result of the complete filling of both M(l) and M(2) sites.
5 4 3
ent was the ftrst to locate the Nb +/Nb + and Nb4+/Nb + redox couples in
SICON structure, at 1.90 V and 1.20 V vs. Na, respectively.
As shown in Figure 15.6, the NASICON composition Na3V2(P04)336 may
either oxidation of V3+ -7 V4+ through sodium extraction or to partial
2
ion of V3+ -7 V + through sodium insertion. The difference between
two couples is of 1.8 V (3.4 V and 1.6 V vs. Na+/Na, respectively).
in situ X-ray diffraction during cell operation, Patoux showed that the
tion of sodium from Na3V2(P04)3 to NaV2(P04)3 is a two
phase n, the latter phase being isotypical with NaTh(P04)3.
A similar approach was followed by Goodenough's group to evaluate
ative positions of these same transition elements couples when lithium
rted or extracted into/from the NASICON framework. To this end,
M'(P04)3 compositions were prepared through Na+- u+ ion exchange
the stable37 39sodium analogs 37immersed
39 in molten
27 29 40LiN03 at 300°C.
b(P04)3, " Lh,FeTi(P04)3, ' Li3Fe2(P04h, ' " and LhV2CP04)
3
4143
ong the most significant examRles. in situ X-ray diffraction
Recent neutron diffraction4 <a.b),44 and
45
ments demonstrated that the the composition LhV2(P04)3 is
ctural with LhFe2(P04)3 and LhTh(P04)3. Lithium extraction leads to
P04)3 isostructural with LiTh(P04)3 (Figure 15.8). This occurs at an
ting potential of 3.75 V vs. Li+/Li, among the highest for polyanionic
res. The overall data relative to lithium insertion or extraction is
tly consistent with the one obtained for sodium
n-Based Positive Electrode Materials 455
LiV,(I :(4 V)
-
Li 3V1(P04)3 (3.4 V)
24 28 32 36 40
29 (")- CuKa
5.8. X-ray diffraction patterns collected in situ for NASICON LiM 2(P04h and
4)J (M = Ti, V)
Vs Li • I Li
1.5 1.7 2.2 2.5 2.8 3.7
V4•t3•
Vs. Na • INa
Nb4•/3• Nb5•t4• Ti4•t3• fe3+12•
0 2 3 4 5
x(Li)
V 2 (504)
3 .... .-... ¥. Li2 V2 (504 )3
Li Fe (50 )
3.6 v
Fe (50 )
..............
2 4 3 2 2 4 3
2 5 et 2 . 2 et 1 . 5 V
TiNb(P0 4h .,.. ... .... ... .. . ... ... ..3. .... ... 4Li3 TiNb(P0 )
2
LiTi4
(P0 ) ..... :. ... Li Ti (P0 )
3 3 2 4 3
2.8 et 2.5 v
Li2 FeTi(P04)3 .,......::::......... Li4 FeTi(P04 h
.....2....8....v
.....
......3........v... .....1.........v......
M/P =1
M/P =1
M/P = 2/3
M/P = 2/3
M/P = 1/2
Rhomboedrique (R-3)
Monoclinique (PZ/n) Echange d'ions
Vo1e ceramique
A B
,,,
;;;
,,,
,,,
8
Manthiram was the first to report on the insertion of lithium into 56
linic Fe2(S04) 3 at an attractive potential of 3.6 V vs. Li+/Li. Okada
d that NASICON and monoclinic Fe2(S04)3 had similar electrochemical
ures. Insertion of lithium into monoclinic Fe2(S04)3 occurs through a
hase reaction, the end-member LhFe2(S04)3 being of orthorhombic
51
etry (Pcan).
In the same study, Okada investigated3 for 2the first time the behavior of
linic LhFe2(P04)3 and located the Fe +/Fe + couple at 2.8 V vs. Li+/Li.
point, Li3Fe2(P04)3 was envisaged as an interesting buffer material to
ed with Fe2(S04) 3 for protection against overdischarge. More detailed
27 29
gations, ' revealed though that two hi-phasic reactions take place at
and 2.73 V vs. Lt/Li with the existence of a definite composition
56
P0 4)3 in between. In situ X-ray diffraction during a discharge process
ed that the framework is very well-maintained and that the intermediate
involved are extremely similar to the pristine one. From
iodynamic Cycling 29 with Galvanostatic Acceleration technique,
ette determined that these phase transitions are kinetically limited
by electron transport within each phase rather than from phase front
ion (Li diffusion) (Figure 15.12).
-Based Positive Electrode Materials 459
0 3.5 . . . . . . . . ,. . . . . . . . . f.:.. . . . . :. . . . . . . .
- . Ml-+-+.-+-f-4-..JJ
-513 ....
-10 2.5
; ,----
!····..···············....1·························t····················.. ···t··"
3.5 4 4.5 5
10 20 x in A-Li Fe (PO )
Timelh • 2 4 3
.12. Discharge in PCGA mode (left) and galvanostatic cycling at C/10 (right) of a
c Li3F (P04)3 I Li cel1.29
same redox potential values (4.0 V vs. Li+/Li) for the V 5+N4+ couple in
P04 and LixVOAs04.
5
;::3
4.5
4
ta
'+=I
5 3.5
p
0..
3
0 0.5 1 1.5 2 2.5 3 -20 0 20
x in LiX V2 (PO4 )3 -dxldV(V 1)
64 65 3
ire and Masquelier showed that partial chemical reduction of Fe +
2
e + could occur through adding of extra sodium ions in the interstitial
according to the structural formula :
ce, by analogy with the sodium analog, to the reduction ofFem(1) into
The second process corresponds to reduction of Fe on the Fe(2) sites
versible structural changes occur.
.14. Partial representation of the crystal structures of (a) LiMP 207 (M =Fe, V) and
s207.
7
adhi showed firse that the electrochemical insertion of lithium into
07 occurred at 2.9 V vs. Li+/Li i.e. at a slightly higher operating voltage
r LhFe2(P04h. However, at this early stage of investigations on
onic structures, only part of the theoretical capacity (-60%) was
Chapter 15
68
ed due to the poor electronic conductivity of such material. Wurm
d that the electrochemical activity of these materials is drastically
ced by an intimate mixing of finely divided particles of pristine
als with conductive carbon, through ball milling.
3.6
Fe, P
il
;.::3
.;
>
03
· . p 2.2
.@ 2.0
1.8
1.6
1.0 1.1
1.2 1.3 1.4
xdans LiX MX2 07
15.15. Potential-composition curves recorded in GITT mode during the first discharge
MX207 (M =Fe, V ; X= P, As) I Li cells.
on, and VPz07 crystallizes in space group P2J. Compared to LiVP 207,
a very lirrge cell distortion due to a modification of the shape of the
that contained lithium ions along with an oxidation of the vanadium
iV that induces shorter V-0 bond lengths.
10
5
-"? - 0
X
------:::::::::--;::--..:::=!
-5
3.2
c
3
.,;
..
;.. 2.8
>. .
2.6
-;:
c
....
0
. 2.4
2.2
2
1.8 0.2 0.4 0.6 0.8 1 1.2
0
15.17. Potential-composition curve ofTiP 20, recorded in a GIIT mode vs. u+/Li.
The ot draws the incremental capacity voltammogram.42
29 76
As stated by Morcrette and Masquelier, the poor electronic
ction of Fern phosphates may be partly compensated by selecting highly
d powders intimately mixed, or "coated" (through ball-milling) to
ctive carbon. The beneficial role of constitutional water, that may act as
moter for faster ionic diffusion has been demonstrated. Within the
ing voltage window used (2V-4V vs. Li) these water molecules look
cted as very good electrochemical stability was demonstrated for
ive cycling. This stability suggests that, as for the crystalline form of
·2H20, where the 6-fold coordination of Fe is ensured by 4 oxygen
ing to P04 groups + 2 oxygen belonging to H20 groups, a similar local
ement occurs for the amorphous FeP04 • nH20 powders.
Despite a whole panel of kinetic responses that depend strongly on
ic parameters such as powder crystallinity, morphology, texture, and
de preparation, all the materials investigated in Ref. 76 lead to a very
r operating voltage vs. Li, i.e. comprised mainly between 3.4 V and 2.6
for crystalline LiFeP0 4 and Li3Fe2(P04)3, the position of the
Fe3+/Fe2+ couple is thus strongly modified by the inductive effect of
the P04 that diminishes the strength of the Fe-0 bonds, even for
amorphous sitions.
n-Based Positive Electrode Materials 465
Capacity (mAh/g)
40 20 40 60 80 100 120 140
Amorphous
::3 FeP04.2H20 _......
>>"'
'-' 3.5 Galvano, C/S
c;; 3
·.::::
11)
0
tl.. 2.5
2 _L
Capacity (mAh/g)
4 0 20 40 60 80 100 120 140
Amorphous
;::1 Fe4(P207),.4H20
">' 3.5
_......
GITT
e:,
3 3
'i::
B
0
tl..
2.5 Eeq. = 3.2 V
2
0 0.5 1.5 2 2.5 3 3.5 4
x in Li Fe (P 0 ) .4H 0
X 4 2 73 2
oth profiles of Figures 15.18 and 15.19 show S-shape voltage profiles
any well-defined intercalation plateau over the whole composition
3 2
The position of the Fe +/Fe + redox couples for both materials were
termined, after a first full cycle in GITT mode, by measuring the
ium potential after relaxing the battery for 2 days at half state of
ge. The obtained values are 3.0 V and 3.2 V for amorphous FeP04·
and Fe4(P207)3-4H20, respectively. This difference may arise from
t local environments around Fe atoms. In this regard, it is interesting
that the higher position of the Fe 3+/Fe2+ redox couple in FeiP207 h-
Fe/P = 2/3) may be related to the fact that Fe is statistically surrounded
Chapter 15
61 62 79 80
A very complete series of studies was initiated by Gaubicher • • •
llowed by Dupre 81'82and Kerr83 on various LixVOX04 structures84' 85
or the reversible insertion of lithium. As for the NASICON structure,
uctive effect of xo4n· polyanion generates attractive operating voltages
e V5+N4+ and V4+N3+ redox couples in lithium batteries. However, as
discuss extensively, each M06 octahedron is connected to 4 X04
dra only, instead of 6 for the NASICON structure. This generates
r global inductive effect in VOX04-type structures than in AxM2(X04)3.
We will focus here on the relative positions of these couples as a
on of the crystal structure and of the chemical nature of the polyanion.
(010)
(010) (100)
4
Nb5+/Nb + Couple in LixNbP05
a b
5.22. Schematic representation of the crystal structures of a-NbOP04 (left) and 13-
(right).
On the other hand, two kinds of more regular Nb06 octahedra are
ntered in r3-NbP05 that crystallises in the monoclinic space group P2/c
8) with a = 13.028(1) A, b = 5.278(1) A, c = 13.225(1) A and r3 =
88 89
7(1)0 • • This second high temperature form, is characterised as the
er m=2 of the "MonoPhosphate Tungsten Bronze with pentagonal
s" (MPTBp) family of general formula Ax(Nb03)2m(P02)4; (A= Li, Na,
The structure is described as Re03-type chains of Nb06 octahedra,
ted by slabs of P04 tetrahedra delimiting pentagonal tunnels (Figure
. In this structure, double strings of Nb06 octahedra parallel to the b
e isolated, and linked to others by tetrahedra only.
Both structures present distinct behavious with regard to lithium
on (Figure 15.23). r3-LixNbP05 rapidly and reversibly uptakes lithium
g to a sustained capacity of 90-120 mAhlg (i.e. operating to the
4
Nb + couple) for more than 100 cycles at - 2 V vs. Li+/Li. This
bility was confirmed by in situ X-ray diffraction in the range 0x0.7
dditionally revealed a series of single- and two-phase insertion
nisms. On the other hand, a-LixNbOP04 displays an insertion plateau at
V vs. Li+/Li indicative of a two-phase process that was found to be
reversible. Both materials, when discharged at voltages lower than 1
sent irreversible reactions with complete amorphization and reduction
allic Niobium.
-Based Positive Electrode Materials 469
0 U OA G.6 0.1 1
x in cx-Li,.NbOP04
'r-----'' [Te0c2]layer #1
FeOr. isolated }
from each other
within each
dense layer
108
.25. SEM image of optimized :UFeP04 particles (-l -tm).
160
140
'i ii
:c 120 ...
<E
. _,
100 ...
,e. 80
!:! •
'D
c . 60 ;::.
f l ...
u
fl
40
20
OL---
.... 1.1 ... ...
.
x In Ll FePO
. .. --
0 1 u 2 3
Rate (number of Li atoms exchanged per h)
Optimised Standard
,.....
4.S
...t
;: 4 -·:
I
I
3.5 3.S
-
3 i!:.3
2.S
:
: ! 2.S
2
1.5
0
>
2
l.S
0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8
x lnLI FePO
• • •
x Ia Li FePO
•
1>8 llruker, CDK.
15.27. X-ray diffraction patterns ofLiFePOIC pristine electrode and at the end of
or two different electrodes. 110
CONCLUSIONS
1 A-
LI1Fe2(P04)1 B-LI F8z(P04)1 LIFeP04
UFeP10r
A-
LIVP10r B-L11V1(P04h FeP04.2HzO
LI1V2(P04h
ERENCES
Guyomard, in "New Trends in Electrochemical Technology: Energy Storage Systems
Electronics", T. Osaka and M. Datta (Editors), Gordon & Breach Science Publishers,
apter 9 p. 253 (2000).
B. Nalbandyan, I. L. Shukaev, Russ. J. lnorg. Chern. 32 (1987) 3.
B. Goodenough, M. M. Thackeray, W. I. F. David, P. G. Bruce, Rev. Chim. Miner. 21
984) 435.
. M. Thackeray, W. I. F. David, J. B. Goodenough, Mater. Res. Bull. 17 (1983) 785.
M. Thackeray, W. I. F. David, P. G. Bruce, J. B. Goodenough, Mater. Res. Bull. 18
983) 461.
M. Thackeray, P. J. Johnson, L.A. de Picciotto, P. G. Bruce, J. B. Goodenough,
ter. Res. Bull. 19 (1984) 179.
Manthiram, J. B. Goodenough, J. Solid State Chern. 71 (1987) 349.
Manthiram, J. B. Goodenough, J. Power Sources 26 (1989) 403.
Delmas, F. Cherkaoui, A. Nadiri, P. Hagenmuller, Mat. Res. Bull. 22 (1987) 631.
Delmas, A. Nadiri, Mat. Res. Bull. 23 (1988) 65.
Collin, J.P. Boilot, in "Superionic Solids and Solid Elecrolytes ", Academic Press,
89) p 227 .
Y-P. Hong, Mat. Res. Bull. 11 (1976) 173; L. Hagman, P. Kierkegaard, Acta Chern.
and. 22 (1968) 1822.
Hagman, P. Kierkegaard, Acta Chern. Scand. 22 (1968) 1822.
G. Sizova, A. A. Vronkov, N. G. Shumyatskaga, V. V. Pynkhin, N. V. Belov, Dokl,
ad. Nauk SSSR, Ser. 205, Issue 1 (1972) 90.
Tillement, These de doctorat de l'Universite Paris VI, Mars 1992.
B. Goodenough, H. Y-P. Hong, J. A. Kafalas, Mat. Res. Bull. 11 (1976) 203.
Tillement, J. Angenault, J.C. Couturier, M. Quartan, Solid State Ionics 44 (1991) 299.
Tillement, J.C. Couturier, J. Angenault, M. Quartan, Solid State Ionics 48 (1991) 249.
Tillement, J. Angenault, J.C. Couturier, M. Quartan, Solid State lonics 53-56 (1992)
1.
Nadiri, C. Delmas, R. Salmon, P. Hagenmuller, Rev. Chim. Miner. 21 (1984) 537.
M. Reiff, J.H. Zhang, C.C. Torardi, J. Solid State Chern. 62 (1986) 231.
C. Torardi, E. Prince, Mat. Res. Bull. 21 (1986) 719.
G. Bruce, G. Miln, J. Solid State Chern. 89 (1990) 162.
d'Yvoire, M. Pintard Screpel, E. Bretey, M. de Ia Rochere, Solid State Ionics 9&10
83) 851.
Pintard-Screpel, F. d'Yvoire et Francis Remy, C. R. Acad. Sc. Paris 286 (1978) 381.
Delmas, R. Olazcuaga, F. Cherkaoui, R. Brochu, G. Le Flem, C. R. Acad. Sc. Paris
7 (1978) 169.
Masquelier, A. K. Padhi, K. S. Nanjundaswamy, J. B. Goodenough, J. Solid State
ern. 135 (1998) 228.
. Andersson, B. Kalska, P. Eyob, D. Aemout, L. Haggstrom, J.O. Thomas, Solid State
ics 140 (2001) 63.
Morcrette, C. Wurm, C. Masquelier, Solid State Sciences 4 (2002) 239.
K. Padhi, V. Manivannan, J. B. Goodenough, J. Electrochem. Soc. 145 (1998) 1518.
Rangan, J. Gopalakrishnan, Inorg. Chern. 34 (1995) 1969.
n-Based Positive Electrode Materials 475
Masquelier, F. d'Yvoire, G. Collin, Journal of Solid State Chemistry 118 (1995) 33.
Gabelica-Robert, P. Tarte, Solid State Chern., Proceedings of the 2nd European Conf
9 June 1982, p.475.
Belkouch, L. Monceaux, E. Bordes, P. Courtine, Mat. Res. Bull. 30 (1995) 149.
. Wurrn, M. Morcrette, L. Dupont, G. Rousse, C. Masquelier, Chern. Mater. 14(7)
002) 2701.
Riou, N. Nguyen, R. Benloucif, B. Raveau, Mat. Res. Bull. 25 (1990) pp 1363.
. H. Lii, Y. P. Wang, Y. B. Chen, S. L. Wang, J. Solid State Chern. 86 (1990) 143.
Wang, C. Wu, S. Liu, J. Solid State Chern. 113 (I) (1994)37.
. Rousse, C. Wurm, J. Rodriguez Carvajal, M. Morcrette, C. Masquelier, Int. J. lnorg.
ater. 3 (2001) 881.
V. Pechkovskii, E. D. Dzyuba, G. I. Salonets, V. N. Yaglov, A. I. Volkov, Rus. J.
rg. Chern. 20(3) (1975) 329.
Sanz, J.E. Iglesias, J. Soria, E. R. Losilla, M.A. G. Aranda, S. Bruque, Chern. Mater. 9
997) 996.
Song, S. Yang, P.Y. Zavalij, M.S. Whittingham, Mater. Res. Bull. 37 (2002) 1249.
Masquelier, P. Reale, M. Morcrette, C. Wurm, L. Dupont, D. Larcher, J. Electrochem.
c. 149(8) (2002) Al037.
.S. Hong, K.S. Ryu, Y.J. Park, M.G. Kim, J.M. Lee, S.H. Chang, J. Mater. Chern. 12
002) 1870.
. Prosini, M. Lisi, S. Scaccia, M. Carewska, F. Cardellini, M. Pasquali, J.
ctrochem. Soc. 149(3) (2002) A297.
Gaubicher, Y. Chabre, J. Angenault, A. Lautie, M. Quarton, J. of Alloys and
mpounds 262-263 (1997) 34-38.
Gaubicher, These de Doctorat, Universite Pierre & Marie Curie, Paris (1998).
Dupre, J. Gaubicher, J. Angenault, G. Wallez, M. Quarton, J. Power Sources 97-98
01) 532.
Dupre, J. Gaubicher, T. Le Mercier, G. Wallez, J. Angenault, M. Quarton, Solid State
ics 140 (2001) 209.
A. Kerr, J. Gaubicher, L.F. Nazar, Electrochem. Solid State Lett. 3(10) (2000) 460.
V. Lavrov, V. P. Nikolaev, G. G. Sadikov, M. A. Porai-Koshits, Sov. Phys. Dokl.
(9) (1982) 680 0
H. Lii, C.H. Lii, C.Y. Cheng, S. L. Wang, J. Solid State Che., 95 (1991) 352.
M. Longo, P. Kierkegaard, Acta Chern. Scand. 20 (1966) 72.
G. Amos, A. Yokochi, A. W. Sleight, J. Solid State Chern. 14 (1998) 303.
G. Amos, A. W. Sleight, J. Solid State Chern. 160 (2001) 230.
Leclaire, H. Chahboun, D. Groult, B. Raveau, Z. for Kristallographie 177 (1986).
Chahboun, D. Groult, M. Hervieu, B. Raveau, J. Solid State Chern. 65 (1986) 331.
K. Padhi, K. S. Nanjundaswamy, C. Masquelier, S. Okada, J.B. Goodenough, J.
ctrochem. Soc. 144(5) (1997) 1609.
S. Andersson, B. Kalska, L. Haggstrom, J.O. Thomas, Solid State Jonics 130 (2000)
Y. Shao-Hom
INTRODUCTION
h2 2
lf/ + (E + V)lf/ = (1)
V
8.1Z' moe
2 0
Where eE is the total energy of the electron and -eV is the potential
Within the kinematical approximation, the amplitude of the wave
d from an atom at a point P (at r from the origin) can be expressed by:
(
exp 2m"K'r
2 1 1
h r atom
r Z- fx) (3)
de for electrons:
(
x is the atomic scattering amplitude for X-rays and Z is the atomic
The interaction between the atom and electrons consists of scattering
oth nucleus and the electron cloud. This is different from X-ray
on, which interacts with atomic electrons and the atomic scattering r
X-rays can be expressed by:
Chapter 16
e2 (4)
fx-rays ((}) = --2fx
m0 c
o.s -----;:::============::-!
--e- x-ray diffraction
- Electron Diffraction
" 0.4
::;;
:::;
..
0 0.3
N0 o.2
.".,
l!!
!
..
rl 0.1
...
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Sine/A
16.1. The ratios between the X-ray and electron scattering factors of lithium to
ese atoms lfLI!Mn) as a function of SinS//.,. Note that the ratio, fd!Mn, for electron
on is considerably larger than that of X-ray diffraction.
(1 ((})is the atomic electron scattering factor of site i in the crystal. The
te intensity of the scattered wave can be obtained by taking the product
wave function with its complex conjugate.
Electron diffraction patterns of a powder sample can be obtained by
d area electron diffraction from regions as small as 0.5 J..lm or by
dy of Battery Materials 483
. TEM Imaging
Diffraction contrast in the TEM images results from the variation in
racted intensity of the material. Bright-field images (the transmitted
tensity) and dark-field images (the diffracted beam intensity) provide
tion about microstructure features of the materials. Imperfections such
cations, stacking faults, domain and phase boundaries within the
displace the unit cell and thus introduce a phase change in the scattered
which leads to intensity variation in the dark-field and bright field
Dark-field imaging has been particularly useful to show and study
structures within lithium transition metal oxide crystals induced by
intercalation or de-intercalation.
In addition to diffraction contrast, imaging of lattice fringes has been
nly used in studying phase transformations in lithium battery materials
nanometer scale. The intensity distribution of the incident beam and
racted beam gives a modulation with the spacing of the plane being
ed for a given crystal orientation and thickness, which reveals the
f the lithium electrode material in form of sine fringes as shown as
2 2 (6)
I= 1+ R - :-
2Rsin( S0 )
;] •
r-:-·--·::-::
.
c::.a·:--i
I ··•• ••••• I
1
INtO.:.
...
:t .t · -········..-·········...,--...··-,-·-·...- -,.-· ·-·-,X
0.0 O.l 0.. O.J O.J t .t
16.2. The voltage profile of Li,Ni0 2 vs. lithium as a function of lithium content x from
et al.56 . Note that the electron diffraction patterns obtained from various points in the
profile show different superstructures, which is not visible in the X-ray powder
on patterns of Li,Ni02 samples.
. &s
3.8
3.6
.3. Typical voltage profiles of Li/LixCo0 2 cells and the inflexion point is correlated
-disorder of lithium and vacancy in Li,Co02 (x ""0.5)3 •37 .
Chapter 16
37
Experimental electron diffraction analysis of LixCo0 2 with x z 0.5
ectron diffraction simulation of disordered Li0.5Co02 with space group
and ordered Li0.5Co02 with space group P2/m have confirmed the
ed in-plane lithium ordering. The in-plane ordering long the [110]hex.
uivalent directions allows the formation of mazed crystals and electron
tion patterns with appearance that resembles those of the spinel
re can be fully explained by lithium and vacancy ordering and
37
tence of two or three variants, as shown in Figure 16.4. • Therefore, it
37
eved that electron diffraction data do not suggest the layered to spinel
rmation in the LixCo02, and the stability of the layered LixCo02 upon
removal is in agreement with the excellent cycling performance of
Co02 batteries. In addition to in-plane lithium ordering, evidence of
ng of lithium and vacancy arrangements perpendicular to the c axis of
xagonal parent cell is also found, which is consistent with first
les predication 78 •
We now compare and discuss the difference in lithium ordering
ements between LixNi02 and LixCo02 systems. The number of ordered
arrangements found in LixNi02 obtained from electron diffraction and
rinciples is significantly more than LixCo02. Lithium ordering occurs at
02 with x = 0.5 and x = 113 where lithium ordering in LixNi02 has been
ed at x = 0.25, 0.4, 113, 0.5, 0.6 and 0.75 ' 63• In theory, in-plane
42
ng of lithium and vacancy should lead to deviation in the alb ratio of the
ohedral layered structure as the electrostatic interactions of lithium ions
the a and b directions are different. The deviation in the alb ratio from
al value (1.732) decreases the crystal symmetry from rhombohedral to
linic, where the crystal structure can be modified by varying the alb
nd/or ·It is interesting to note that the structural changes associated
ithium ordering at the Li0.5Co02 composition and in the LixNi02
; x:::; 0.75) system are different. Changes in the alb ratio would
erably modify the cobalt-oxygen bonds and distort the Co06 octahedra,
is energetically unfavorable as both Co3 + and Co4+ ions prefer to
ideal or nearly ideal octahedra. Therefore, the Li0.5Co02 layered
re adopts a negligible change in the alb ratio (1.733) but a significantly
nt (107.89°) value from the rhombohedral symmetry upon lithium
cancy ordering. In contrast, the LixNi02 (0.50:::; x:::; 0.75) structure has
l change in (109.75°) but a large increase in the alb ratio (1.764) as
3
eller distorted Ni + ions can accommodate the distortion in the a and b
on upon lithium and vacancy ordering.
y of Battery Materials 489
6.4. Left, three possible orientations for lithium and vacancy ordering in the lithium
d the monoclinic unit cell of Li05Co02 and right, an experimental diffraction pattern
[103lmon. zone axis of the monoclinic unit cell or the [OOllhex. zone axis of the
l unit cell, which revealed the reflections associated with lithium and vacancy
37
rom all three variants •
Layered 02 LiCo0 19 2
with a unique AB CB AB oxygen packing, first
by Delmas et al. , is metastable to layered 03 LiCo02 with 03-type
BC oxygen packing. Recently 02 LiM02 (M = Ni, Mn, Co, etc.)
nds have been reinvestigated for lithium rechargeable battery
ions80 -89 as their structural derivatives upon lithium de-intercalation
ercalation are stable with respect to the spinel-related framework.
removal from 02-LiCo02 leads to the formation of various novel
as shown by X-ray powder diffraction, T#2, T#2' and 06, which are
82 86 90
related to the 02 crystal structure ' ' , as shown in Figure 16.5.. Of
ar interest is the T#2-LixCo02 phase that is stable over a large
itional range (0.52 < x 0.72) and the stability of this structure is not
understood. The r2 structure has an orthorhombic unit cell and
roup Cmca, for instance, 'Lio.61Co0z' with aorth. = 2.8097 A,
corth. = 9.9082 82 . . 87
0
ium ions are located on the 8e type tetrahedral sites in the T#2
e, where two 8e sites are available per cobalt ions and fewer than half
e sites are occupied by lithium ions. As simultaneous occupancy of
8e sites is unlikely due to small site-to-site distances (1.4 A and 2.4
ium ordering on the 8e sites is expected in the T#2 structure.
eless, it should be noted that the voltage-capacity profile for the entire
positional range is relatively smooth (Figure 16.5.). The absence of
Chapter 16
Be Variant II
Ntl
16.6. The [51 2 lorth. and the [02 T lorth. diffraction patterns collected from the
02' sample reveal the presence of the 2 aorth. x 2 borth. x 2 corth. superstructure, in
ly fundamental reflections are indexed. Lithium and vacancy ordering configurations
o describe the doubling of the aorth. and borth. cell parameters are presented. Note that
ants are allowed as two possible orientations exist for the 8e Li04 tetrahedra. The
of the corth. cell parameter results from the alternation of these variants in the
38
space along the corth. axis .
16.7. The [0 ll]onh. and [OOl]onh. electron diffraction patterns collected from the
o02 ' samples, where the first-order, incommensurate reflections are marked by
8
. Two variants of the incommensurately ordered structure are highlighted in the
. electron diffraction pattern.
(11)
Chapter 16
equations indicate that the total scattering intensity from the (003)
ions in the layered structure equals that from the { 111} reflections of
hiated-spinel structure.
Now we consider intermediate cation distributions that deviate from
eal lithiated-spinel and layered configurations, which can be described
following formula: { (Lit6-4x)J6c[Li4x] t6cthayerl{ (Mt6-4x)J6ct[M4xh6c }Jayer2032
nel notation. When x = 0, the structure has the ideal lithiated-spinel
uration, (Lit6)t6c[Mt6h6ct032, or (Liz)t6c[M2h6ct04. When x = 1, it has the
d configuration, { (Lit2)t6c[Li4h6d}tayert { (Mdt6d[M4h6c}tayer2032
atively, in the layered notation, (Li3)33[M3hb06). The magnitude of the
re factor and the scattering intensity for the set of the { 111} planes
ncreasing transition metal density, e.g., the (1 T 1) reflection can be
bed as a function of x:
F(lll) = (8fM +8xfM -8fu -8xfu) (12)
agnitudes of the structure factors and the scattering intensities for each
other three sets of the { 111} planes with decreasing transition metal
ies, can also be described as a function of x:
F ( lll) = (8 f M - 8 xf M - 8 f Li + 8 xf Li ) (13)
thiated-spinel o.•
o.a
0.7
o.e
0.5
o.•
0.3
02
0.1
6.9. The scattering intensity ratio of the (111) and (1 T 1) planes, m, decreases as the
istribution varies from the ideal lithiated-spinel to the ideal layered configuration31 •
dy of Battery Materials 497
' U 03 Ol U
Uhlllled-SpFui'
00 0.7 Oj DO I
la,...-.:
16.10. A high-resolution TEM image (top left) of LT-LiCo02 crystals showing the
crystal size on the order of 20 nm, and two single-crystal electron diffraction patterns
similar scattering intensity ratios of the (111) and (Ill) planes 31• The number of
analyzed in the LT-LiCo02 sample is plotted as a function of x (bottom right).
dy of Battery Materials 499
16.11. Left: A typical high-resolution TEM micrograph and the corresponding electron
on pattern {[1 T 0]1ayered zone axis}of layered LiMn02 crystals and Right: A high
on TEM image and the corresponding electron diffraction pattern { [0 1 !]spinel zone
a randomly selected crystal in the Li05Mn02 electrode after one charge and one partial
e21 •
PERSPECTIVES
. .
,.•,. .• .•
• •• • •
, ,•
•
,,•
•
.• .•
t • • ••
..
6.12. Left: Simulated Images of LiCo0 2 in the [110] zone axis under thin-crystal
hase-object") conditions for increasing resolution (marked) show that atoms of cobalt e
visible at 2 A resolution, oxygen at 1.4 A, and lithium at 1 A. Right: Experimental
of Li columns confirmed by simulation a) LiCo02 ESW-phase image, reconstructed
experimental images, shows lithium ions in rows (arrowed) between 0-Co-0 groups.
phase simulation for 48 A thickness and 0.9 A resolution •
39
Chapter 16
wledgments
NCES
den & T. Reddy (eds.) Handbook of Batteries (McGraw-Hill, New York, 2002).
ushima, P. C. Jones, P. J. Wiseman & J. B. Goodenough, Materials Research
5,783-789 (1980).
eimers & J. R. Dahn, J. Electrochem. Soc. 139,2091-2097 (1992).
uku & A. Veda, J. Electrochem. Soc. 141,2972-2977 (1994).
ales, C. Perez-Vicente & J. L. Tirado, Materials Research Bulletin 25, 623-630
roguennec, P. Deniard, R. Brec, M. T. Caldes Ricos & L. Brohan, Mol. Cryst. Liq.
1, 101-108 (1998).
aetz, C. C. Ahn, R. Yazami & B. Fultz, Journal of Physical Chemistry B 107, 2887-
03).
abrisch, R. Yazami & B. Fultz, Electrochemical and Solid State Letters 5, A111-A114
Dompablo, A. Vander Yen & G. Ceder, Physical Review B 66, art. no.-064112
. Aydinol, A. F. Kohan & G. Ceder, Journal of Power Sources 68, 664-668 (1997).
ander Yen, M. K. Aydinol & G. Ceder, J. Electrochem. Soc. 145, 2149-2155 (1998).
ander Yen, M. K. Aydinol , G. Ceder, G. Kresse & J. Hafner, Physical Review B 58,
7 (1998).
elmas, J. J. Braconnier & P. Hagenmuller, Materials Research Bulletin 17, 117
M. Nazri
INTRODUCTION TO SOLVENTS
The basic requirements of a suitable electrolyte for electrochemical
s are high ionic conductivity, low melting and high boiling points,
cal and electrochemical stability, and safety. Electrolyte conductivity
ectrochemical stability are key parameters in selecting an electrolyte
odem electrochemical devices such as advanced batteries, fuel cells,
capacitors, sensors, and electrochromic displays. These parameters,
ctivity and electrochemical stability, will receive particular attention in
apter. Although progress has been made in enhancing the conductivity d
electrolytes, particularly the polymeric ones, liquid electrolytes are sed
in most electrochemical systems. The solvent properties, and ics of
ion solvent interactions, must be understood in designing new lytes. In
this chapter, a short but general introduction to properties of ts and ion-
solvent dynamics is discussed.
The history of electrolyte development goes as far back as the work of
philosophers in search for a universal solvent, the so-called
hest". In search of Alkahest, many solvents and chemical rules were
ered such as "like dissolves like" (similia similibus solvuntur) as
in Table 17.1. Later, the theory of osmotic pressure by van't Hoff
1911), and the theory of electrolyte dissociation by Arrhenius (1859-
were discovered. Many speculations about the nature of solute-solvent
tions and the influence of solvent media on the rate of chemical
n were proposed in the early eighteen-century. The role of solvents on
cal equilibrium, on tautomerism (i.e. keto-enol tautomerism), and the
menon of solvatochromism (shift of UVNis absorption bands due to
nges of the index of refraction) were discovered. 1' 2 Scheibe et al. have
Chapter 17
Carboxylic Esters
drocarbons
Alcohols, 1,4-Dioxane
des Acetic Acid
s Alcohols, Water
ylic Acids
c Acids
hydrophilic Water
than the time scale of NMR, and solvent molecules in the solvation
annot be distinguished from those in bulk solution. The complexity of
tions in liquid state makes the theoretical treatment of electrolyte
uid Electrolytes:Some Theoretical and Practical Aspects 511
re a difficult task, and still most reliable data are those obtained
mentally, using NMR, vibrational spectroscopy, and fast spectroscopy
ues.
When dealing with mixed solvents, the solubility parameter relates to
4
lar enthalpy of vaporization for a gas of zero pressure by
. Ion-Dipole Forces
olvents with unsymmetrical charge distribution (dipolar solvents)
a permanent dipole moment, - The magnitude of the dipole moment
s on the value of charges on the opposite sides of the molecule and
eparation distance as defined by = q·r, where q is the charge
ing equal value of charge on each side of the molecule), and r is the
5
ion distance between charges. Most organic solvents have permanent
3
°
moment in a wide range from 0 to 18xl0. Coulomb-meters (Cm). In
lar, cyclic and linear organic carbonates, used in lithium batteries and
larly dealt with here, have a large 3difference in permanent dipole
t, (i.e.
10"
30 °
llpropylene carbonate = 16.7 X 10" Cm, and J.l.dietbyl carbonate= 3 X
he dipole moment, relative dielectric constant, melting point, and
point of several common organic solvents are listed in Table 17.2.
ter molecule is also listed as a reference for comparison.
Chapter 17
6 8
7.2. Common dipolar organic solvents and their physical properties. -
ide
ol
acid
exanol
ne Carbonate
ethane
trile
yl sulfoxide
ne
e
enzene
nzene
ethylphophoric
amide
henone
e
acetate
xanone
ne
drofuran
carbonate
yl carbonate
propyl carbonate
isopropyl carbonate
opyl carbonate
opropyl carbonate
e carbonate
xane
oethene
ether
e
disulfide
oromethane
ne
lting point, tbp = boiling point, E = dielectric constant, 1.1 = dipole moment.
uid Electrolytes:Some Theoretical and Practical Aspects 513
ze is the charge of the ion, e is the dipole angle relative the line joining
to the center of dipole, and Eo is the vacuum permittivity. For a
endate ion-solvent configuration, the solvent molecule and the ion are
on the same line, and e = 0. The ion-dipole forces are very important
relative stability of the solvation shells. It also plays a major role on
ed solvation when multi-component solvents are used.
. Dipole-Dipole Interactions
olvent molecules with permanent dipole moments interact with each
ia directional electrostatic interactions. The dipole-dipole orientation
gly temperature dependent, and thermal energy prevents optimum
tion of solvents toward each other. The dipole orientation is
ally favored, and at higher temperatures all other dipole orientations
more populated. The dipolar interaction is mainly responsible for the
10 11
tion of dipolar solvents. ' The dipole-dipole interaction is given by 12
(17.3)
(17.7)
Udispersion = - (3a2I)/4 6
r (17.8)
6
The dispersion energies are very short range as expressed in 1/r . This
important for molecules with polarizable 1t electrons. The molecular
ation of polyaromatic molecules are mainly related to the dispersion
among polyaromatic molecules with conjugated 1t electrons. From the
ons above, it is clear that the dispersion energy is also important for the
on of molecules with high polarizability values. For example, the
sion interaction energy of two molecules with a = 3x1o- m , I =
30 3
<:1-<:IIz-<:liz-<:1
EPD + Sb<:ls ====================EPD+- -sb<:ls
At room temperature
The donor number is a very useful concept and correlates well with
ctrochemical half-wave redox potentials, kinetic rates, and the NMR
20
al shift signals for various EDP solvents. - 26 The donor numbers of
ommon organic solvents are listed in Table 18.3 of <:hapter 18.
A quantitative electron accepting ability (accepting number, AN) for
n pair acceptor solvents has also been developed to determine their
philic properties. This concept is developed by Gutmann and
31
ers based on P-NMR chemical shift of triethyphosphine oxide
27
0) in presence of an EPA-solvents at room temperature. - 29 The
n used to determine the acceptor number AN is
· Ucorr· · ppm
8corr(Et 3 PO- SbCl 5 )- 8(n- C6 H 14
)
(17.11)
31
N is obtained from the relative P-NMR chemical shift values
to the 1:1-adduct Et3P=0 - Sb<:l 5 dissolved in 1,2-dichloroethane.
emical shift for n-hexane has been used as reference, ANn-hexane= 0.
N values of a number of common organic solvents are also listed in
18.3 of chapter 18. The acceptor numbers are dimensionless number
to those of Sb<:Is.
SOLVATION
he interaction between ions and solvent molecules is important in
ytic media. The shell formed around the ions by solvent molecule is
solvation (hydration in aqueous electrolytes). The energy of solvation
hange in Gibbs energy when an ion is transferred from vacuum into a
: Gsolvatiom is the solvation ability of the solvent. In order to dissolve a
a solvent, the interaction energies between solvent molecules, g
the all-electrostatic, polarization, and dispersion energies, as well
attice energy of the salt, must be overcome.
he electrostatic field that solvent molecule experience in close
ty of the ion in a solvation shell is in the order of 106-107 V/cm. This
Chapter 17
1 Solvation Number
The solvation number is the number of solvent molecules in the
y shell attached to the ion in a way that their translational and
nal motions exhibit changes from the bulk solution. In some cases, the
t molecules in the second sphere around the ion are also considered as
ion number. The residency time of a solvent molecule in the solvation
is short, so that in most cases they can be monitored in the time scale
17
R spectroscopy. The 0-NMR spectroscopy of Ae+ and alkali ions in
4
us solution has indicated a slower than 10 /sec solvent exchange for
4
nd faster than 10 /sec exchange for the alkali ion hydration. 31• 32 NMR
oscopy has proved to be a very useful tool for studying the dynamic of t
molecule in solvation shells of various ions.
The solvation number is usually measured by electrolytic transport
ds. The solvation number is higher for the harder ions (small size), as
3 34
red to the softer ions (large size), with similar charge? • The data
electrolytic transport measurement and from conductance
rements indicate that solvation numbers increase according to: Li+>
+>Rb+>Cs+; the alkaline earth ions are more solvated than the alkali
nd their order is: Mg++>Ca++>Sr++>Ba++. Anion solvation numbers for
35
n ions follow the order: F>CI">Br->L
on, and when the cation is solvated preferentially by one solvent and
ion by other solvent molecules, it is called heteroselective
36 37
on. '
everal methods have been developed to study selective solvation in
mponent electrolytes. The IR and NMR spectroscopy have been used
y spectroscopists and the conductance method and Hittorf transference
6
ements by electrochemists?
1. Electrolyte Conductivity
We wish to stress in the final part of this chapter points 1 and 3 of the
list. To this end, some data on the anodic stability of several
ates and their conductivity in a broad temperature range will be
ted here.
mong various solvents presented in Table 17.2, the organic
ates, particularly the cyclic carbonates, are the most suitable solvents
plication in a battery electrolyte. However, the viscosity of the cyclic
ates is high and needs to be modified to improve their low temperature
mances. Addition of a linear carbonate with high fluidity enhances the
conductivity of multi-solvent blend electrolytes. The conductivity of
l organic carbonates at various salt concentrations and at different
ratures has been measured. The salt concentration was chosen from
to 1M LiPF6• Normally, below 0.4M salt concentration a marked
se in conductivity of the electrolyte is observed and considered not to
actical. Above the 1M concentration, there is significant salt
itation at low temperatures, t<-2o·c, that is detrimental to most
chemical devices. The temperature range was selected from -40 to
in our experiments. This is a practical temperature range for most
chemical devices. It has been shown that there is no electrolyte with a
solvent that can provide all the requirements of a practical electrolyte.
ore multi component electrolytes containing multi-blend solvents are
Several binary and ternary organic carbonate solvents are used to
ce the performance of the electrolytes, particularly their low
rature conductivity. In fluid dynamics, the mixing rules for the
38
ity and dielectric constant have been defined
(17.12)
(17.13)
(17.14)
UPF•
: LiPF•
iPF•
-50 0 50
Temperature ( C)
17.1. Conductivity of ethyl methyl carbonate (EMC) containing, ( +) 1M, (.&) 0.8M,
M, and (o) 0.4M LiPF6.
"' "
§
""
.
....
en....
20,00
s 15,00
. e 10,oo
' - "
>
•; :J
.§ 5,00
s:::::
0 0,00 -50
u 0 50 100
Temperature ( C)
"'s" 15,00
"" 10,00
en
'-"
s
·- ..>..
..
>
......
u
.g
s:::::
0 0,00
u
-50 -30 -10 10 30 50
Temperature ( C)
The most striking feature of the EC-EMC mixed electrolyte is that its
ctivity does not change significantly for 1M, 0.8M, and 0.6M LiPF6
oncentrations, even at high temperatures. Using this remarkable
lyte, the cost of salt can be reduced by 40%. The other remarkable
ty of this electrolyte is that the EC with melting point of 39°C does not
separate at temperatures significantly below its melting point.
-..
] 20,00
Cl)
5 15,00
>..
·-.§ 10,00
......
5,00
§ 0,00
u 100
-50 0 50
Temperature ( °C)
s 15,00
Cfl
5 10,00
.......
:.>...
.........
5,00
§
u 0,00
-50 -30 -10 10 30 50
Temperature ( °C)
Oxygen reduction
OH reduction
Oxide formation ..,
Solvent oxidation .
5
--r.ithium deposition
Li-M alloy formation (UPD)
Solvent Reduction and formation of SEI layer
-----------H 20 reduction
Electrolyte solvent
Propylene carbonate
Ethylene carbonate
Dimethyl carbonate
Diethyl carbonate
Methyl ethyl carbonate
Gamma-butyrolactone
Diethoxyethane
Tetrahydrofuran
2-methyltetrahydrofuran
I ,3 dioxolane
Acetonitrile
CONCLUSIONS
RENCES
1. INTRODUCTION
The major families of this class are ethers, esters, and alkyl carbonates,
h are highly important for the field of high energy density batteries.
e solvents and their formulas are summarized in Figure 18.1. Other
ortant solvents are acetonitrile (AN), dimethylsulfoxide (DMSO), N-N
thyl formamide (DMF), methylene chloride and nitromethane. The latter
solvents are mainly important for electroanalytical use and organic
hesis. As discussed later in this chapter, the type of salt plays a major
in determining the electrochemical window of a polar aprotic system.
monly used anions, in conjunction with Li+, are PF6-, Cl04- and BF4-. All
r aprotic solutions may contain atmospheric contaminants such as 02,
and C02, which are reactive and may play an important role in
rmining the electrochemical behavior.
MOyXz type (M=S, P, N, and even metals such as V and Cr; X=Cr, Bf,
hat were mentioned in the literature in connection with possible battery
ications. However, SOC}z and S02 seem to be the most important
1
ents in this group for electrochemical systems.
ETIIERS
Diethyl ether (DEE) CH,CH,OCH,CH,
Tetrahydrofuran (TifF)
6
2-Methyl-THF (2Me ·
T HF) q
Diethoxyethane CH,CH,OCH,CH,OCH,CH,
1-3 Dioxolane
0
ALKYL CARBONATES
Ethylene carbonate (EC)
)a
\._/
Methylacetate 9
CH3COCH3
(a)
18.1. The most important organic polar aprotic solvents for Li batteries.
Chapter 18
5
.1.4. Molten Salts
N-Methyl-2- ii:
pyrrolidinone ;,:
5·
I'yridine "'
Sulfolane
Acetone
Acetonitrile
Benzonitrile
y-Butyrolactone
Diethyl carbonate
Diethyl sulfite
Dimethyl carbonate
Ethyl acetate
Ethylene carbonate
Ethylene glycol sulfite
Methyl acetate
Meth_yl formate
Ul VJ VJ
Nitromethane
ylene carbonate
Trimethyl phosphate
Diethyl ether
Diglyme
1,2-Dimethoxyethane
1,4-Dioxane
1,3-Dioxolane
2-Methyl-tetrahydrofuran
Tetrahydrofuran
Tetrahydropyran
Triglyme
Benzene
Carbon tetrachloride
Chloroform
1,1-Dichloroethane
Methylene chloride
Toluene
Trichloroethylene
Cyclohexane
n-Heptane
n-Hexane
Sulfur dioxide
Sulfuryl chloride
-
-
'1:1
Thionyl chloride
.Oo
dvanced Liquid Electrolyte Solutions 535
e 18.2. Typical approaches for defining solvent polarity. Taken from Reichardt<8l.
scale, appears at a lower wave number. The scale itself is defined by the
wing equation:
The values range between zero and about 60 kcal/mole. Polar solvents
as nitromethane are very weak donors. Solvents such as acetonitrile and
ylenecarbonate are moderately strong donors. Ketone, ester, and ether
ents show medium donor properties. Amides and sulfones are strongly
solvents with donor numbers of 25 to 30, and aliphatic amines are very
g donors.
The characterization of the acceptor properties of solvents is a more
cult problem. The definition of an analogous thermochemical quantity is
possible because many solvents contain atoms with lone pair electrons,
hus may undergo adduct formation with strong reference acceptors.
9 12
This problem was solved by Mayer • by the use of triethylphosphine
5 5
oxide (Et3P + --- 0 - ) as a basic reference donor, in a solvent acting as an
ptor.
Triethylphosphine oxide contains a highly basic oxygen atom, which is
y accessible to solvent electrophilic attack. This causes a polarization of
= 0 bond and a downfield shift of the 31P NMR signal. The observed
ical shifts (8) referred to the reference solvent n-hexane, and
polated to infinite dilution, may be taken as a measure of the acceptor
erties of the solvents. The acceptor number (AN) is defined as follows:
AN = _8_(E_t..::...•P_O_in_so_l_ve_n_t_S_)_-_8_(E_t..:;_.P_O_in_h_ex_a_n_e) (18.3)
8(SbC15 ·Et 3 POinC 2 H 4 Cl 2
)
2
However, when plotting Ev. values of a redox couple such as Cd/Cd +
E of the solvents, the results provide a very scattered picture. In contrast,
s been clearly shown by Mayer et al. that redox potentials of metals (e.g.,
n2+, Cd/Cd 2+, Eu/Eu 2+) can be nicely correlated to the donor numbers of
solvents, and that the Ev. values measured behave linearly vs. the
ents' DN.9
Table 18.3, taken from Refs. 8 and 9, lists some solvent evaluation
meters, in light of the above discussions and the definitions in Table
, for a number of solvents of interest.
In conclusion, extensive work on solvent properties has revealed that
le physical properties, such as dielectric constant or dipole moment, are
dequate measures for solvent polarity (which can correlate well with the
ence of solvents on thermodynamic and kinetic reaction parameters in
). Better solvent parameters, which correlate well with the impact of the
en solvent on electrochemical and chemical reactions, are donor and
ptor numbers or parameters based on solvatochromic effects, because
reflect not only pure electrostatic effects, but rather the entire electronic
erties of a solvent. We should also mention the efforts to develop tools
18.3. Er30 values, donor numbers (DN), and acceptor numbers (AN) of various
nts. Taken from Refs. 8 and 9.
Solvent Er30 DN AN
% 2Me-THF (vol)
0 20 40 60 80 100
10
e 8 •
'
Q
-=e LiAsF6(DN)
e 6 IMLiA5F6
in 2Me THF/PC
'\.
•
2
u LiAsF6(PC)
L iC 104(DN)
0
0 0.5 I 1.5 2
M/Li SALT.
ion interfaces. Hence, some esters that can have a positive impact on the
uctivity may interfere badly with the passivation of Li or Li-C anodes.
, the choice of solvent mixtures for low temperature applications
ires a rigorous study of the electrodes' surface chemistry and stability in
mixtures at ambient and high temperatures. Figure 18.2 shows typical
ples of the behavior of polar aprotic solutions showing dependence of
onductivity on the salt concentration, and on the solvent ratio, in binary
ions containing low dielectric constant-low viscosity and high
26 27
ctric constant-high viscosity solvents (e.g., 2Me-THF and PC). '
28
Table 18.4 shows typical parameters related to the conductivity of
ted polar aprotic solutions and demonstrates the effects of salts,
ents, solvent mixtures, and temperature. There are many publications
ng with the conductivity of polar aprotic solutions. Typical examples are
. 29-32.
. ELECTROCHEMICAL WINDOWS OF
NONAQUEOUS ELECTROLYTE SOLUTIONS
Reactive electrodes refer mostly to metals from the alkaline (e.g., Li,
and the alkaline earth (e.g., Ca, Mg) groups. These metals may react
taneously with most of the nonaqueous polar solvents, salt anions
ining elements in a high oxidation state (e.g., Cl04-, AsF6·, PF6-,
F3") and atmospheric components (02, C02, H20, N2). Note that all the
Salt Solvents
liPF6 ECJPC
2Me-TIIF/ECJPC
liAsF, ECIDME
PCIDME
2Me-TIIF/ECJP
liCF,SO, ECJPC
DMEIPC DMEIP
2Me-TIIF/ECJP
liN(CF1SOJ2 ECJPC
ECIDME PCID
2Me-1HF/ECJP
liBF. ECJPC
2Me-1HF/ECIP
ECIDMC ECID
'1:1
"....
01)
uao.
a.
LiPF,
s:
! :!.
a
LiAsF,
c;·
I i!
LiCF,SO,
LiN(CF,S
LiBF4
VI
" ""
U l
Chapter /8
solvents have groups that may contain C-0, C-S, C-N, C-CI, C-F, S-0,
, etc. These bonds can be attacked by active metals to form ionic species,
hus the electrode-solution reactions may produce reduction products that
more thermodynamically stable than the mother solution components and
h less soluble. Consequently, active metals in nonaqueous systems are
34
ys covered by surface films.
When introduced to the solutions, active metals are usually already
red by 'native' films (formed by reactions with atmospheric species),
then these initial layers are replaced by surface species formed by the
5
ction of solution components? In most of these cases, the open circuit
ntials of these metals reflect the potential of the MIMX/Mz+ half-cell,
e MX refers to the metal salts/oxide/hydroxide/carbonates which
prise the surface films. The potential of these half-cells may be close to
of the M!Mz+ couple.
At potentials positive to the OCV, the electrochemical window may be
limited because dissolution of the active metal takes place either via
ation of Mz+ ions through the surface films or by the breakdown of the
ce films, exposing the active metal to the solution and enabling a 'direct'
6
lution process? At potentials lower than the OCV, the situation is more
plicated. Since there is no possibility of a cathodic breakdown
anism of the surface films, the possibility of obtaining active metal
sition depends either on the feasibility of migration of the M'+ ion
gh the surface films, or on electron tunneling through them to form
l deposits outside the surface layer. Consequently, in the case of alkaline
metals (Ca, Mg, etc.) whose ions are bivalent, and whose migration
gh thin layers of their salts is difficult (if not impossible), there may be
37
ossibility of metal deposition. '38 In such cases, the cathodic limit of the
rochemical window may be determined by reduction reactions of
ions components. In the case of alkaline metals such as lithium, the
rochemical window is very narrow because Li dissolution and
sition take place above and below the OCV at very low overpotentials,
39
ions migrate through the surface films that cover Li electrodes.
40
cules or salt anions containing central atoms of high oxidation states.
ever, they are not reactive due to passivation phenomena. Hence, the
rochemical window of nonaqueous solutions with electrodes such as
inum or stainless steel may be determined by the electrochemical
41
ion of the solution components.
A typical example is aluminum, which is used as a current collector for
odes in lithium batteries. The stability of aluminum in many Li salt
ions at potentials as high as 4.5 V vs. Li/Lt is due to the formation of
ly insoluble AI halides on the aluminum surface, which remain stable
42
hus protect this active metal from corrosion.
Intercalation electrodes
Red-ox electrodes
I 30JLAICiT?
Li + SaII Solution
o.v
o+ ,.....
---, ,.----.
I
\ I
I
/ 1/
I
,...-'\ I
0. / vi DME or THF
O'(F;t;;z;;;::::===P=C::=::==:;::: :::
I ,f o+\1AVr:..::::==BL=. =::::=::::::::::::=:: _
I '
t
11 /' TBA Solt
1 I
I I I
t, l
I
ov
,i 2V 3V
4l/
18.3. Steady state cyclic voltammograms obtained with four electrolyte solutions. The
lower ones were obtained from 0.2 M TBAP solutions with THF, BL and PC, as
ted. The upper one was obtained from 0.2 M LiAsF6 or LiC104 solutions in BL. The
ial scale refers to the LilLi+ 0.1 M reference electrode. The potential sweep rate was 20
The current scale is common to all CVs. Solid lines: dry solutions (-15-30 ppm of
dashed line: water contaminated solutions (0.001 M). The potential scan in all cases
from open circuit potential (2.7-3 V in these systems). Ref. 46.
ced Liquid Electrolyte Solutions 547
onates or esters, which are more reactive than ethers, the major cathodic
46
ions in TBA salt solutions relate to solvent reduction.
When the cations belong to the alkaline metals (Li, Na, K), reduction
ace oxygen, water, C02 solvents, and salt anions form insoluble oxides,
oxides, carbonates, halides, and organic salts, which passivate the
rodes and form surface films that block massive reduction of solution
47
ies. -53 Hence, a first polarization of a non-active electrode in a polar
tic Li (or Na) salt solution from open circuit potentials (usually around 3
. Li/Li+) to low potentials, is characterized by irreversible cathodic peaks
nd 2 V and 1.5 V (corresponding to traces of 0254 and H2055 reduction)
rimposed on a broad cathodic wave related to reduction of solvent
46 56
cules and salt anions, which form surface films. -
The steady state voltammetric response of Li (and Na) salt solutions
non-active electrodes includes mostly underpotential deposition and
ping of the active metal on/from the noble metal electrode, possible alloy
ation and decomposition (active metal + non-active metal), and bulk
sition-dissolution of the active metal. As discussed in detail and well
lished in the literature, UPD processes and alloy formation depend on
46 51 52
nature of both the host (substrate) and the deposited metals. ' •
mes 18.1-3 show reaction paths of alkyl carbonates, commonly used salt
ns and contaminants, respectively, in the presence of Li-ions with Li
l, lithiated carbon and/or non-active metals and carbons at low
ntials, as well as the product distribution, based on Refs. 26, 27, 46-53,
7.
Finally, we mention the fundamental basic cathodic reactions of the
most important polar aprotic inorganic solvents, thionyl chloride and
(due to their application for Li batteries\ as listed below:
2 SOC}z +4e- S0
SOz + 2e- Szol-
2.1. Introduction
not supported
?
l..i.zO + EC ---r...iOCH 2CH2 0C02Li.
_ _, CH 2 -0, 1
_.+ .
RO --.../" I .--C=O ------.. ROCH2-CH2-0C02Li pooriiDcleophile
CH2 -0
secood
RO-yH2
nucleC1'hilic
.,ock &
II nucleophile
CH2 -0COz Li ROCH 2CH20R + 1
C,
Q,r\Q
Ro-.J 8" 8"
Li'"
LizC03.. ionpairing
competition
IIDCleophilic
yHz-OCOzlit attack
CH2 -0C02li
(Relevant to PC as well)
---
(In a similar manner other acvclic alkvl carbonates: DRC.. RMC.. PMC. reacts).
. 2ROC0
2 Ii+ H20
-
HF
R- Li.
R·+ Li.0 ---
vanced Liquid Electrolyte Solutions 549
heme 18.2: Surface reactions of commonly used Li salts (Refs. 26, 49, 50, 53,
, 58, 61-67)
LiAsF6 + 2Li+ + 2e" 3LiF J, + AsF3 (sol)
heme 18.3: Reaction patterns of common contaminants (Refs. 26, 47, 48, 50,
53-55, 61, 62, 64-66)
COz + e- + u+ COzLi
COzLi + COz0 = C-0-COzLi
O=C-0-COzLi + e- + Li+ CO i + LizC03 J,
N2 + 6e·+ 6Li+2Li 3N
02 + e·+ u+Li02
Li02 + e- + Li+Li202
Li 202 + 2e·+ 2Li+2Li 20
HF + e·+Li+LiF + Y2 H2
18.5. Oxidation potentials (V vs. Li/Lt) for several electrolyte solutions. Taken from
128 and Refs. therein.
(a)
------,;;--- .
----. . ....
• ,., .
..
" ' .,, .
.. :
.t
,f ' !
' · I'
-'·:"' ,
· ..
117
2.1. EC-PC/LiAsF6, MOLl Energy Inc. Li-MoS2 Cells
This solution was also found to be very good for reversible, high
ng efficiency of Li anodes. While relatively high cycling efficiency
Chapter 18
surface films
77
.2.3. 1-3 Dioxolane/LiAsFJTributylamine
10 150 u
..
> 8 120 :!'
c
>
g
!! c.
4
90 t E
60
1 30
0 llO 0
0 60 Time, min 180
b.
hetic graphite flakes seem to be the best probe to test suitable solutions
Li-ion batteries, because what works well with these electrodes should
even better with other types of graphite and with disordered carbons. 135
e intensive studies converged into the development of standard
rolyte solutions which are commonly used throughout the world in the
production of commercial Li-ion batteries. These solutions include
6 as the electrolyte and alkyl carbonate solvents. EC is an obvious
ponent, 139 together with cosolvents that are acyclic alkyl carbonates from
40 141
DEC/ DMC, EMC,59 and PMC60 list. The alkyl carbonate solvents
he most suitable because their anodic stability is higher than that of other
lies of solvents such as ethers and esters. Also, at low potentials both Li
Li-C compounds reach a very good passivation in a variety of alkyl
onate solutions. It should be noted that neither ethers nor esters promote
passivation of Li-C electrodes. 132- 138 In the case of ethers, these solvents
not sufficiently reactive to form passivation layers at high enough
ntials. Thus, ether molecules can easily cointercalate into graphite with
ns, which leads to the destruction of the graphite's structure
oliation). 133 The reduction products of esters are not cohesive and
sive enough to form sufficiently passivating surface films on carbon
rodes.
Highly interesting, and deserving of discussion herein, is the failure of
hite electrodes in PC solutions. No graphite carbons can insert Li
sibly in pure PC solutions with Li salts. 134 It appears that some
uctive processes take place at relatively high potentials (in which PC is
ced, 1.5-0.8 vs. Li/Li+), and the electrodes cannot reach all of the Li
132
calation stages. Reversible behavior of Li-graphite electrodes is
ned in mixtures of ethers and PC, PC solutions containing crown ethers
142
, 12 Cr 4), and in EC-PC mixtures 143• In the latter solutions, the
ersible capacity is very strongly dependent on the morphology and
ee of disorder of the graphite particles. 135 The fact that in solutions based
C, which is similar in structure to PC, the behavior of Li-graphite
rodes is so reversible, is striking. PC is reduced on graphite below 1.5 V
species such as CH3CH(OC02Li)CH20C02Li and CH3CH=CH2
pylene gas), while EC is reduced to species such as (CH2 0C02 Lih and
46
ene gas. •50 The EC reduction products, when formed, precipitate
kly on the graphite surface to form cohesive and adhesive surface films
passivate the electrode and block electron transfer, thus preventing any r
surface reactions except for Li insertion (which include Li-ion
ation through the surface films). The fast precipitation of the surface
in this case prevents detrimental side processes such as cointercalation
lvent molecules.
In the case of PC, its reduction product contains the methyl group,
h interferes badly with the good cohesion and adhesion of the surface
ced Liquid Electrolyte Solutions 563
unavoidable contamination with HF. 145 LiPF6 decomposes into LiF and
in an equilibrium reaction. 146 The latter species readily hydrolyzes
protic substances to form POxFy, P(OR)xFy species and HF, which reacts
ROLi or ROC02Li surface species formed by alkyl carbonate reduction,
replaces them with highly resistive LiF surface films. HF also reacts
the LixMOy cathode materials. The impact of the HF reactions on the
trodes' passivation is negative and the electrodes' impedance increases
F surface reactions are more intensive.
In conclusion, the standard electrolyte solutions for Li-ion batteries
LiPF6 solutions in mixtures of alkyl carbonates that always contain EC as
ghly polar component and a precursor for very good passivating agents
61 62
, (CH20C02Li)2) ' and solvents from the DMC, EMC, DEC
39 141
- It should be emphasized that ternary mixtures such as EC-DMC
enable a very good low temperature conductivity of LiPF6 solutions to
eached.25 The choice of these systems is mostly due to an acceptable
ivation and stabilization of both anodes and cathodes for Li-ion batteries
a wide electrochemical window (i.e., high anodic stability > 4.5 V vs.
t with lithiated transition metal oxide cathodes).41 Major problems with
use of these solutions in practical batteries relate mostly to HF
amination, which interferes badly with the electrodes' passivation. In
tion, at elevated temperatures there are accelerated surface reactions on
anodes and cathodes, which increases their impedance and lead to
rent capacity-fading of Li-ion batteries.147' 148 There is, therefore, a
g driving force for R&D of improved electrolyte solutions for Li-ion
ries, as summarized in the next section.
I( ., ,
s;r- !- (CH _ O+C02 Li)2
2ne- +2nEC+2nLit----+"'?i=.r _..,.,;; ::i: -'2
Current collector CCH
2
CH3
I
CH - OC02Li
I
C -OC02Li
The methyl group in
the RC02Li species interferes
--
wi th a +C H-
good
cohesion
CH=C
c==:)
Worse
passivation
Gas bubbles are
Formed inside the
particles
aphite particles
e
litted ¢::=::J
e More
reactive
s
essure su rface
D
exposed,
further surface reactions
6. MOLTEN SALTS
6.1. Introduction
Molten salts may be useful non-aqueous electrolyte systems for a
ty of applications, including batteries. 169' 170 Ionic liquids of anhydrous
, namely molten salts, usually have high conductivity and wide stability
large potential ranges. 171
Especially important are the room temperature molten salts. The most
ly studied are the chloroaluminate melts with organic cations such as
yl-ethylimidazolium, N-butyl pyridinium, pentamethylimidazolium, and
dimethyl-1,2,4 triazolium.172
Regarding the conductivity of molten salts, unlike conventional
queous solutions, where the medium is almost unaffected by the
rical field, in molten salts the properties of a medium which consists
of ions are affected by the applied electric field. It is often found that
quivalent conductivity increases as the electrical field applied is higher,
to a certain point where a plateau is reached. 173 Regarding the
uctivity of molten salts, several conduction mechanisms have been
osed to explain their high electrical conductivity. 174' 175 With organic RT
en salts, a maximum conductivity is usually reached with neutral
ms.176
Chapter 18
The potential windows of molten salts are very important for their
rochemical applications. The limiting potentials depend on the
position of the melts and the temperature, which affect the standard
ntial of reduction and oxidation of their components.
In the case of room temperature molten salts, the anion is typically a
ure of aluminum halide complexes in equilibrium. These anions can
rgo several oxidation reactions, depending on the composition of the
. The cathodic reactions may involve reductions of the organic
177
ns.
It was possible to obtain reversible Li and AI dissolution and
172
sition in some RT organic molten salts. Hence, these molten salts
be considered as possible electrolyte systems for rechargeable
queous lithium and aluminum batteries.
Recent work on the behavior of cathodes for Li batteries (e.g.,LixMOy,
178 179
Mn, Co) • clearly demonstrates that RT molten salts may be
ising electrolytes for secondary Li-ion batteries as well, comprising
ns such as ethyl methyl imidazolium and anions such as cr, BF4-, and
4-.
Finally, there are reports on the use of these systems in super (EDL)
180
citors.
FERENCES
POLYMERIC ELECTROLYTES:
AN OVERVIEW
J. B. Kerr
INTRODUCTION
Ionically conducting polymers have been the focus of much
mental and applied research for many years. Polyelectrolyte
ranes have found significant technological use in the production of
1
ne and caustic soda, as separators in fuel cells2'3 and in
45
dialysis · for example. The discovery of ionic conductivity in
67
hylene oxide solutions of alkali metal salts ' led the way for the
uction of polymer electrolyte in devices such as lithium batteries and
89 0
chromic windows. ' Since those early days many books/ -13 book
8 9 14 19
rs and reviews ' - have been published on these materials and the
is referred to these for more detailed information.
15
In a recent article Scrosati and Vincent have listed the desired
ties of a polymer electrolyte for use in lithium batteries. These are
ate conductivity, high cation mobility, good mechanical properties,
interfacial contact with electrodes, wide electrochemical stability,
cal and thermal stability, safety and ease of processing. This set of
ties is understandably difficult to obtain in a single material and a
r of different classes of polymer electrolyte materials have been
4
d which possess most but not all of the desired properties. Wrighe has
elaborated on this classification and defined them as:
Solvent-containing gel and hybrid systems.
Solvent-free, ion-coupled systems.
"Single-ion" systems and systems with reduced anion mobility.
Decoupled systems.
These classifications are useful for distinguishing different
nisms of ion transport in the bulk separator. However, the complete
ic Electrolytes: An Overview 575
Load
""
l nter raccstSEI)
- "' / x ummmmu uu H i
:-.Li -t---- Li+ ----+!.! ! !m;<; mlllllllllllllllllllll
Electrolyte HH vxoy
Li+ ----E m:;-LL=='=() l ,,== =
X \ jjjUjjjjHjjijjjjillllllllll
\
"Dry" Polymers, Solvent swollen gels.
Binary salt systems, single-ion conductors
ncy. Impressive cycle lifetimes were obtained for the lithium electrode
certain conditions (LiFoM > 100). Parameters such as the purity of the
lyte and the lithium metal, prior history (exposure to heat or light),
ce of surface-active additives or impurities and cell stack pressure
ll found to affect the results. With lithium metal electrodes, loss of y
was found to be due to chemical reactions and morphological
s that convert the lithium into a form that is not easily recoverable.
an be insoluble lithium salts or isolation of lithium metal through
e growth that produces "mossy" lithium. The suppression of lithium e
growth was and still remains a critical safety and reliability issue uld
not be easily solved with liquid electrolytes. The reaction of and
flammable organic solvents with finely divided lithium powder und to
be an insurmountable safety problem. Polymer electrolytes, er, are
not volatile and although they do react with lithium metal and ng
28
metal oxides the slow delivery of fuel to the reaction site limits of
reaction and prevents runaway reactions. This safety feature plus a ion
that polymer electrolytes inhibit dendrite growth have lead to the
interest in solvent-free materials over the last twenty years.
EO :- $4-8/lb
MO :- M.Wt. -150k
ymethylene-linked PEG400
-PEM0-20
-3 .......................,...
, -<>- PVES-20
; ··•··PEPE5-20
j ---PEO-LITFSI [30:1]
················ • -HQ data-30 TFSI
PEO = Polyethylene Oxide
PEA= Polyacrylate ether
-5 .........................,... PVE = Polyvinylether
PEPE= Polyepoxide ether
-6 ..................... .
PEMO = oxymethylene
-7 '
[
'
i
linked PEG400
-a crystalline PEO ' HQ =EO/PO copolymer
0.004 0.0045
0.0025 0.003 0.0035
1fT
4
ations requires a conductivity of between 10- and 10-3S/cm. Thus PEO
ms adequately at 80°C but no system provides acceptable transport
ties at ambient temperatures. In order to provide adequate performance
temperatures it is important to elucidate the factors that control the
dence of transport properties on temperature.
0
o ·-, 4-:o
l'·-\
,#'\,.
·'-A h .
.r' ' rr emus
Controlled by Control
Segmental Motion -related to
- related to Tg solvation
--><-· PEPE5{20:1]
1MO units appear to lead to
-4 • • '• • PCE3E(20o1)
higher rmbility
• ,..._ PP3A(20:1]
·5 _,_PEPP3(20:1)
1P1 PEPP3
·10
0.0025 0.003 O.OOJS
1fT
..
d TMO units and the temperature dependence of conductivity, the
ransition temperature as a function of salt concentration are shown in
'
F U ··t Jg ·g·r
•
28.2
(!Ui)
1 -·A , 1 .
6
I U "-t (wtcr)
7
'
G U ·-tgt "
E U ·., JJl • .t:.t (ccnlcr)
9.8. Calculated reaction pathway for migration of the lithium cation along a
lene oxide (PEO) chain. The PEO is modeled by diglyme and the cation moves by
and breaking Li-0 bonds. The energies are in kcal/mol from HF/6-310*
67
ons.63 •
19.9. The polymers in this case are the polyepoxide ether polymers
ifferent proportions of EO and TMO units in the side chains which
67
om all EO to all TMO units. ' 68 It can be clearly seen that as the
ion of TMO in the polymer increases the low temperature
tivity increases and the dependence of tg upon the salt concentration
ses. However, it is very striking that the temperature dependence of the
tivity above 40°C is unaffected indicating that the motion along the
s calculated is not rate determining.
There are a number of significant differences between these comb
rs (presence of PO units in the backbone, chain ends with OH groups)
linear polymer chain used for the calculations but it appears that the
nism used in the calculations does not dominate the rate.
ements of the linear PTMO-LiTFSI complexes show a slightly
Chapter 19
wer slope but no overall increase in conductivity over the linear PEO
s which indicates that the differences in the comb structure are not
cant. The decreased dependence of the tg on salt concentration
ed for both comb and linear TMO polymers indicates that the ionic
linking between chains is reduced with TMO. Although this does not
the problem of how to obtain high conductivity at room temperature, it
important result as the TMO structures may be used to mitigate the
of salt concentration gradients that occur in lithium-polymer batteries,
larly in the presence of surfaces such as electrodes or ceramic filler
es.
-20
T
g
of Polyepoxide Ethers
........ --fPE(TM6Eo>2
Me
-40
I• u• ncrea·.· se .,· · "•e·PETMd,!E03 ..
-
-70
-- L- ....
026 0.003 0.0032 0.0034 0.0036 0.0036 0.004 -80 20 40 60 80 100 120
0
1FT O:Ll
The results of these combined synthetic and theoretical efforts have yet
blish the physical phenomena that limit the mobility of the ions.
er, the possibilities have barely been explored as there are many
res such as carbonates, polyethylene imide (PEl) and polyalkylsulfide
1
that may be synthesized to test the limitations of these electrolyte
s. To date, the barrier to ion transport remains undefined in the simple
s. Efforts have been made to investigate the influence of different
such as the use of block copolymers 69-74 that provide different
ic Electrolytes: An Overview 587
5
nments for the solvated ions, rigid polymer systems such as glassi or
76
line electrolytes that contain defects as well as "polymer-in-salt"
77
s. These approaches show promise to provide information that will
te the nature of the limiting process in the ionic motion although they
oduce further complexity into the system. The presence and effect of
nt phases on the behavior of the polymer electrolyte is of great interest
y for the effect on ionic mobility but also for the understanding of the
cial behavior at electrode surfaces.
as 100°C above the first one (see Figure 19.10). The second tg was
ed using dynamic mechanical analysis but the behavior has been
med by recent neutron scattering measurements which showed the
89
nce of two distinct relaxation processes in the presence of the fillers.
ehavior is dependent upon the nature of the polymer-surface
tion, the polymer molecular weight and any cross-linking, be it
nt or ionic as in the case of a polymer electrolyte. It has been reported
or high M.Wt. polymers the second tg is not observed but the tg
ses due to the restriction of the polymer chain motion by multiple
90
es. As has been noted by Eisenberg these effects are not confined to
olymer systems but also play a role in polymer gel systems and hence
ely to influence transport properties in composite electrolytes and
des in lithium ion and lithium-polymer gel batteries as well as solid
polymer electrolytes.
>50%, no Tg
© © 0 ooo
o0oo
0 = filler particle
= immobilized
© © <10% polymer
=restricted
1 oneTg r polymer
©©© oooo
oo 0 >20%, smaller
©© -10%,
two Tg's upperTg
+ Ultra-pure
<> Lab-pur ied
b
• Hydrophobic Si02 filled
Hydrophilic Si02 filled
E
.!::!
<> +
+ (a)
• •
<>
'-"
• ••
(I) <>
+
<>
<> +
>- 10-4 ... +
... ...
=ti:::; ...
"0
c
... ...
u ...
+ _.
0
0)
0 10"5 b
...J ...
...
--- •
... <>
10-6 I I
E'
(b)
0
10 Hz
19.11: (a) Conductivity; (b) compression elastic modulus of low molecular weight
P(EO)z0LiTFSI at 80°C. Amount of fumed silica added is 10%.
Chapter 19
his case the polymer binder in the electrodes is identical to the polymer
separator and the co-polymer formulation is chosen to reduce the
linity and optimize the uptake of solvent into the separator and
des. These gels often contain fumed silica particles to provide
114 116
ved mechanical properties. - In fact, gel electrolytes with excellent
nical properties and liquid-like transport properties have been prepared
nly the presence of fumed silica particles which form a network
re that contains a polyethylene glycol dimethylether (PEGDME)
n of LiTFSI salts. ll7-!2o These systems show great promise as they
ne excellent mechanical properties with high rate capability and the
ial for good chemical and electrochemical stability with the electrodes
providing enhanced safety, lower cost and more flexibility in
ing than a conventional lithium ion cell.
The operation of the gel polymers is dependent upon the presence of
vent molecules that travel with the ions just as occurs in a liquid. Thus,
nductivity and salt diffusion coefficients approach those of liquid
111 121 122
lytes (>10-3S/cm, 10-7cm% at room temperature). ' ' The
ers and ceramic networks provide the mechanical strength by forming
ks through which the liquid is able to migrate. This network formation
nction of the properties of the polymer or particle surface and how it
ts with the solvent cluster as it moves through the matrix. The best
system for this is the presence of clusters and multiplets that form
ar pathways in polyelectrolyte gels as is well known for Nafion® and
3 123
polyelectrolyte membranes. ' As such these solvent-network
tions and gel-electrode interactions are simply extensions of the issues
ered above for the effect of surfaces on dry polymers and consequently
of the considerations are similar. The ability of the network system to
etely contain the liquid and prevent leakage is of great importance as is
tential for the gel to retard reactions, prevent thermal runaway and
ce safety. From a manufacturing point of view polymer gel systems are
rest as the gel precursors can be introduced into the cell as a liquid and
in place, thereby greatly simplifying the manufacturing process. There
wever, a number of drawbacks with this procedure that will be more
nt after discussion of the design needs of practical lithium batteries in
t section.
ular models of transport and reactivity to help guide the selection and
of the appropriate materials that can provide the desired improved
68
mance.
.:-'._:·:"!-".
...
I
I
.•
:;_;;:;f \
I
I
I
ium
) !1
: t?e; p,,---
- '·- ·- >
./:\;:::
9.13. Effects of transference number and salt diffusion coefficient on salt
ation profiles upon Li-polymer cell discharge.
de processes in the cell. In most lithium ion battery systems only the
ion undergoes any reaction at the electrode while the anion does not
o a reaction. Thus, when the lithium ion is removed from solution by
Chapter 19
90
'
•••••
10·11
•20 • i ,a
• 4
D D •
<>
IJ
D
C]
•D D /85
D /00
10·12 <)
<> D /40
<>
0 D
<>
"
[J
f t
10·11 0 0.5 15 2 2.5 3 3.5
Conc./M
0.5
0.4
• 1+85 c
.l I+ SOC
[J
{) <> l+40C
0.3
••
0.2
" [J
<}
lijl
j • • •
f
. [J
<}
+...
[J
0.1 [J
" D
1:1
(>
<> <>
Conc.IM
9.14. Salt diffusion coefficient (D,) and transference number (t+ 0) for LiTFSI
s a function of salt concentration and temperature (40, 80 and 85°C).
Chapter 19
;
TT
:····:·:;
!•
r
:· ······,·· j 51"/.PE0301-12TFSibypulseSieppolarlzatlon
:
. .. . '
..
2.7
22
·
2.1
400
• 1imp
1imp:before first charge
'
D
1Z'1
2Z1
0 3Z1 "'su
4Z1
]
X
+ 5Z1
0
N
0 100 150 200 250 300 350 0 50 100 150 200 250 300 350
z Ohmcrn' Z Cllun.cm2
(a) (b)
i
; 100
0
11.
.II:
ca
Gl
11.
50
0 0.002 0.004 0.006 0.008 0.01 0.012
152
9.18. Dependence of peak power on anode film (SEI) resistance.
Chapter 19
<±>
i
I
LowM.Wt.
fragment,
/..........._/ /'--.._ -',,' analyze by
I I I I GC-M:
( ( (\/
Radical cleava o
(electrochemica
here +
) ) \/\;
0 0
(chemical - lithium \
(oVo\ 0
metal)here (
\/\f% ) ',
;
'
)
structure affects Li ion transport
through bulk polymer and across interfaces
(a)
:z
-s:IJIJ IJ s:m t til" t5 Hi' :z to" :Z5 to" ::l11l" ::!5 ui'
11m' I
I Side-Chain !li
II""
fragments
li!
l:o
I
I
lllo<l
'll:il
IW
1l:il
t:l
!ill
:SOOQ
4:11
.ll:llQ "Ill
IQ Tnr <{ra!)
1
(b)
2
9.2J.(a) Cycling of Li/PE(TMOh-LiTFSI(lO:1)/Li at 85°C and 0.4 mA/cm . Cell
ss= 256 Jlm; (b) Gas chromatogram of extracted membrane from (a).
rane had a thickness of 256J.Lm and the current density was set at 0.4
2
• The cell polarization behavior shows a transition time that indicates
rrent is operating above the limiting current. The delivery of lithium
the electrode is insufficient to support the current and hence a second
n occurs at a different voltage. One may observe that the cell rapidly
ue to a dendritic short circuit. Extraction of the cell components for
is by gas chromatography shows the formation of alkoxy alcohols
ponding to cleavage of the side chains as shown in Figure 19.20 with
M0)3 unit predominating. However, a higher molecular fragment is
en that corresponds to cleavage of the backbone and indicates more
ex behavior than our analysis has so far encompassed.
ic Electrolytes: An Overview 609
2
Prior to the cycling at 0.4 rnNcm shown in Figure 19.21(a) the cell
2
en cycled at 0.2 rnNcm for two days to pass a total of 134 coulombs
mAh) with little visible change in the impedance (<5% increase in the
cial impedance). The cell had been cycled during this time at a rate
as below the limiting current where the transport properties were
ent to ensure an adequate supply of lithium ions. The limiting current
135
y be calculated according to the following equation :
. Dendrite Growth
ution at the anode will be non-uniform and this will encourage dendrite
. Thus, poorly controlled chemistry such as UV and thermal radical
d cross-linking is likely to lead to non-uniformity in the polymer
ane and cause problems during cycling particularly at high rates. This
reason why in situ curing of polymer gel systems may not be wise.
pment of alternate cross-linking chemistry that is more
35 197
lable • is critical to the production of quality lithium polymer
es.
The previous discussion on the effects of surfaces on the polymer
ics is pertinent to recall at this time. One may recall that the glass
on temperature of polymer electrolytes rises with increasing salt
tration (see Figure 19.9). During discharge of a lithium metal cell with
ry salt system, the concentration gradients at the lithium metal
de rise to levels that induce phase changes (glass formation,
tation) and leads to changes in interfacial impedance (Figure 19.17).
changes will be enhanced by the presence of filler particles and cross n
the polymer and inhomogeneities in the membrane will induce non m
current distribution which may encourage dendrite initiation at the te
electrode in a symmetrical lithium/lithium cell. On charging of the
metal electrode in a lithium /polymer/metal oxide cell, the depletion
concentration at the plating electrode will lead to a lower tg, higher
y at the electrode surface and the ability to sustain a higher current
. During the charging cycle strong mechanical properties are desired
33 68
plating electrode surface to restrain dendrite growth. • However, in
tal oxide electrode, salt concentration rises, leading to a higher tg.
sed mobility and high interfacial impedance. The presence of strongly
ting particles such as carbon black or cross-linking in the polymer
lyte, binder or polymers formed from electrolyte solvents during the
ar or cycle life will lead to decreased transport properties in the
site electrode. This will be the case for dry polymer, gels and liquid s
and will result in high impedance and low rate capability. The
uence of a non-uniform composite electrode composition will be a
neven current distribution which will encourage the initiation of
e growth at the lithium metal counter electrode.
There is a well-developed literature on morphological instability
electrodeposition of a variety of metals, which indicates that even a
199 203
perturbation on the surface will initiate dendrites. - The case of
is complicated by the reactivity of the metal and the need for a
ve film on the surface, commonly known as the SEI. A physical e
of a perturbation would be non-uniformity in the SEI that can give
instability and dendrite growth. It is obvious from our discussion on
r dynamics at surfaces how perturbations may arise from the changing
ies of the polymer electrolyte during cycling and that any design of
Chapter 19
r9t
T PEPE3_Silane·X_Tg_lhird
1 o'
10°
66 6
... ..
6 lo !' o
", ZrTg • '
-,".,'
../r I
)
'
,.
66 D D
0 c 0 0 6
'
:& 6A66 6AA6
0 0 0
00
0 I .o r> 0 o o0
""
0
0 0
o oacaooooP c 0
o
10. 1
·t OO. O
c.\spmdata\06040002.hdf
9.22. Polymer membranes prepared by cross-linking the co-polymer EPE 3 and AGE
35
sing hydrosilylation; (a) DMTA in compression mode at 10Hz. The elastic (E') and
(E") moduli and the tan I> response are shown as a function of temperature. (b) AFM
ct mode.
ic Electrolytes: An Overview 613
. Mechanical Properties
The foregoing discussion on dendrite growth highlights the need for
understanding and control of the mechanical properties of the
lytes. The inevitable desire to push the battery to the limits of its
mance will impose great stress on the polymer electrolytes that will
expose any weaknesses. Since the mechanical properties of the
er electrolytes are closely coupled with the transport properties, the
of the polymer systems needs to take close account of whether it will
d next to a lithium metal electrode, in the bulk of the separator or in a
site electrode. Even the seemingly simple process of cross-linking the
r membrane can introduce undesirable properties. For example,
19.22(a) shows the dynamic mechanical thermal analysis (DMTA) of
-linked polymer membrane of co-polymer EPE3 and AGE (10%) with
(20: 1) that shows the presence of a second tg indicating non
204
mity in the material. Examination of the film by contact mode
force microscopy (AFM) is shown in Figure 19.22(b), which shows
sence of non-uniform areas. Phase imaging AFM in tapping mode has
hown to be even more useful for the study of membrane morphology
205
l cell membranes and clearly these are very useful techniques for
control of membranes.
n this case the non-uniformity of the membrane is due to the non
nature of the co-polymer. During the co-polymerization it was noted
e AGE and EPE3 monomers disappeared at different rates indicating
e polymer was a block co-polymer and hence the cross-linking was
trated in the AGE blocks. The membrane showed rapid formation of
es when it was cycled in a symmetrical lithium/lithium cell and this
ascribed to the non-uniformity of the membrane and the fact that the
ane was very thick (400f.Lm) so that the limiting current was low.
unately the requirements for quantitative mechanical measurements
hickness) are difficult to reconcile with the need for thin membranes m)
for lithium batteries. Nevertheless, useful results can still be d if
the membrane can be separated from the electrodes after cycling.
19.23 shows the frequency sweep (a) and dynamic mechanical l
analysis (b) of a composite membrane consisting of the cross-linked
mer of EPE3 and AGE (X-PEPE3), LiTFSI (20:1) and A200 Aerosil
silica before and after cycling at 85°C in a lithium/lithium cell at 0.2
2 204
• The membrane in this case was about 300 !lm thick and the
2
ted limiting current was about 1 mA/cm based on the transport
ies measured in the absence of cross-linking and filler particles. The
cy sweep experiments show a decrease in the moduli after cycling
he DMTA shows an increase in the area under the tan 8 peak,
ng an increase in the proportion of polymer undergoing the transition,
Chapter 19
7
10
o'e
:-
w
f-
N
wE: 1o' - ·
ooooooooo
· o o·
o ooo o 000 o ooooooo
•a• • • •o • • • • • • •• • •• • •••• • • • •
10
5
1-- ----------LH-I'--------------10-' -------------J
10"
F eq (radls]
(a)
·
!I= 0.8
111
!! -!- . -
(b)
CONCLUSIONS
The design of polymer electrolytes for lithium batteries is a complex
m that needs to be approached with a combination of sophisticated
ng, diagnostic techniques and significantly greater synthetic effort s
been applied in the past. Molecular modeling is needed to predict
olecular features desired in an electrolyte to attain the transport
ies required to support the desired rate capability and to provide the
ry mechanical strength to act as a separator, inhibit dendrites and to
gether the particles of the composite electrodes. These molecular
s must also be stabilized towards reaction at the electrodes by the
on of an SEI layer with almost magical properties. System modeling
ed to set the performance goals of the materials that need to be d
and rigorous comparison of the model predictions with experiment
ssary. Failure analysis must be accomplished by a combination of
icated diagnostic techniques together with reaction modeling and
is of model materials designed to test the modes of failure. Only then
rates of side reactions be related to lifetime issues.
Chapter 19
It is clear that this task is daunting for dry polymer systems and is
to prove difficult for polymer gels if demands are put upon them
d acting as an adhesive to hold lithium ion electrodes together as is
practiced today. Our survey of polymer electrolyte properties has
strated that the desired properties depend critically upon the
nent of the cell where the electrolyte has to function. The electrolyte o
lithium metal has to have sufficient mechanical strength to inhibit
te growth but the means of attaining this should not reduce the
ort properties. The polymer electrolyte in the composite electrodes
have much more flexibility as the surface effects of the electrodes
o major difficulties with ion transport. This is true in dry polymer, gel
quid systems where polymers are expected to bind the particles
er. These polymer systems also have to resist reactions with these
des over the life of the battery so that these desirable properties do not
e and degrade the performance of the battery.
Given all of these rigorous requirements, it is hardly surprising that the
uction of the lithium metal polymer battery has not yet been successful.
t, it is more surprising that progress has been as successful as it has
given how primitive have been the polymer systems used so far in
rison to the severity of the demands placed on them by battery
ers. This chapter has attempted to illustrate how modeling, diagnostics
nthesis can be used to design better materials for the application. Many
e methods have only become available in recent years so there is much
g work to be done in the corning years. The lithium metal system is
larly difficult due to the dendrite issue but holds the most promise for
high energy density, particularly if ways can be found to increase the
oltage to 4 volts or more. This will require the development of multi
ystems that have a layer resistant to lithium metal on the lithium with
echanical strength and a polymer in the cathode that is stable to four
Since no single material can combine these chemical properties
nt materials will be necessary in different parts of the cell. As we have
it is necessary for the electrolyte to possess different mechanical
ties in different parts of the cell anyway. The nature of these
ayers next to electrodes will be the subject of much research and it is
aging to note that already attempts are being made to design and
ate protective layers on both anodes and cathodes?06" 208 Such efforts are
only the beginning of an exciting new chapter in the development of
batteries which will lead to more realistic demands on the battery
als and better designed materials for the applications.
ic Electrolytes: An Overview 617
NOWLEDGEMENT
ERENCES
T. Leah, N. P. Brandon, V. Vesovic, G. H. Kelsall, J. Electrochem. Soc. 147 (2000)
3.
. Kerres, J. Membrane Science 185 (2001) 3.
D. Kreuer, J. Membrane Science 185 (2001) 29.
Mazrou, H. Kerdjoudj, A. T. Cherif, A. Elmidaoui, J. Molenat, New J. Chern. 22
98) 355.
Bazinet, F. Lamarche, D. lppersiel, Trends Food Sci. Tech. 9 (1998}107
E. Fenton, J. M. Parker, P. V. Wright, Polymer 14 (1973) 589.
V. Wright, Br. Polymer]. 7 (1975) 319.
Armand, J. Y. Sanchez, M. Gauthier, Y. Choquette, in Electrochemistry of Novel
terials (J. Lipkowski and P.N. Ross, eds.), VCH, New York, NY, (1994) p. 65.
Armand, Solid State Ionics 69 (1994}309.
A. Ratner, (J. R. MacCallum, C. A. Vincent, eds.), Elsevier Applied Science, London
87).
. MacCallum, C. A. Vincent, Polymer electrolyte reviews--2, Elsevier Applied
ence, London; New York, (1989).
M. Gray, Solid polymer electrolytes :fundamentals and technological applications,
H, New York (1991).
M. Gray, Royal Society of Chemistry (Great Britain), Polymer electrolytes, Royal
iety of Chemistry, Cambridge, 1997.
V. Wright, MRS Bulletin 27 (2002) 597.
Scrosati, C. A. Vincent, MRS Bulletin 25 (2000) 28.
Fauteux, A. Massucco, M. McLin, M. Vanburen, J. Shi, Electrochim. Acta 40 (1995)
5.
A. Ratner, P. Johansson, D. F. Shriver, MRS Bulletin 25 (2000) 31.
A. Ratner, D. F. Shriver, MRS Bulletin 39 (1989).
Hooper, M. Gauthier, A. Belanger, Electrochem. Sci. Techno[. Polym. 2 (1990) 237.
Aurbach, in Advances in lithium-ion batteries (W. A. v. Schalkwijk and B. Scrosati,
.), Kluwer Academic/Plenum Publishers, New York, NY, (2002) p. 7.
. Sloop, J. K. Pugh, S. Wang, J. B. Kerr, K. Kinoshita, Electrochem. Solid-State Lett.
001}A42.
Vogdanis, B. Martens, H. Uchtmann, F. Hensel, W. Heitz, Makromol. Chern. 191
02._465 .
. Goldman, L.A. Dominey, V. R. Koch, J. Power Sources 26 (1989) 519.
Gofer, M. Benzion, D. Aurbach, J. Power Sources 39 (1992}163.
M. Abraham, Electrochim. Acta 38 (1993) 1233.
R. Koch, J. Electrochem. Soc. 128 (1981}1293.
M. Abraham, J. S. Foos, J. L. Goldman, J. Electrochem. Soc. 131(1984) 2197.
Chapter 19
J. Wen, J. Kerr, M. Rubin, J. Slack, K. von Rottkay, Solar Energy Materials & Solar
lls. 56 (1999). 299.
Oradd, L. Edman, A. Ferry, Solid State lonics 152 (2002) 131.
G. Andreev, P. G. Bruce, Electrochim. Acta 45 (2000) 1417.
G. Andreev, P. G. Bruce, J. Physics-Condensed Matter 13 (2001) 8245.
J. Wilson, C. V. Nicholas, R. H. Mobbs, C. Booth, Br. Polymer J. 22 (1990) 129.
Vallee, M. Duval, F. Brochu, M. Kono, E. Hayashi, T. Sada,US Patent 5,755,985,
98
R. Allcock, Chern. Mater. 6 (1994). 1476.
R. Allcock, D. E. Smith, Y. B. Kim, J. J. Fitzgerald, Macromolecules 27 (1994). 5206.
M.G. Cowie, V. M. C. Reid, I. J. McEwen, Br. Polymer J. 23 (1990). 353.
M.G. Cowie, A. C. S. Martin, Polymer 32 (1991). 2411.
Buriez, Y. B. Han, J. Hou, J. B. Kerr, J. Qiao, S. E. Sloop, M. M. Tian, S. G. Wang, J.
wer Sources 89 (2000). 149.
Pantaloni, S. Passerini, F. Croce, B. Scrosati, F. Roggero, M. Andrei, Electrochim.
a 34 (1989). 635.
Marchese, M. Andrei, A. Roggero, S. Passerini, P. Prosperi, B. Scrosati, Electrochim.
a 37 (1992). 1559.
Andrei, L. Marchese, S. Passerini, A. Roggero, B. Scrosati,US Patent, 5,162,174,
92).
Andrei, L. Marchese, P. Prosperi, A. Roggero,US Patent 5,173,205, (1992).
Andrei, L. Marchese, A. Roggero, P. Prosperi, Solid State lonics 72 (1994). 140.
M.G. Cowie, K. Sadaghianizadeh, Solid State Ionics 42 (1990) 243.
M. Blonsky, D. F. Shriver, P. Austin, H. R. Allcock, J. Am. Chern. Soc. 106 (1984)
54.
R. Allcock, S. J. M. Oconnor, D. L. Olmeijer, M. E. Napierala, C. G. Cameron,
cromolecules 29 (1996) 7544.
R. Allcock, D. L. Olmeijer, S. J. M. Oconnor, Macromolecules 31 (1998) 753.
C. Zhang, S. B. Fang, Electrochim. Acta 45 (2000) 2131.
Hooper, L. J. Lyons, D. A. Moline, R. West, Organometallics 18 (1999) 3249.
Nishimoto, M. Watanabe, Y. Ikeda, S. Kohjiya, Electrochim. Acta 43 (1998) 1177.
Kono, E. Hayashi, M. Watanabe, J. Electrochem. Soc. 145 (1998) 1521.
Watanabe, T. Endo, A. Nishimoto, K. Miura, M. Yanagida, J. Power Sources 82
99) 786.
Nishimoto, K. Agehara, N. Furuya, T. Watanabe, M. Watanabe, Macromolecules 32
99) 154.
M. Mao, R. F. Perea, W. S. Howells, D. L. Price, M. L. Saboungi, Nature 405 (2000)
3
M. Mao, M. L. Saboungi, D. L. Price, M. B. Armand, W. S. Howells, Phys. Rev. Lett.
(2000) 5536.
L. Saboungi, D. L. Price, G. M. Mao, R. Fernandez-Perea, 0. Borodin, G. D. Smith,
Armand, W. S. Howells, Solid State lonics 147 (2002). 225.
W. Halley, Y. H. Duan, J. Power Sources 89 (2000) 139.
Borodin, G. D. Smith, R. L. Jaffe, J. Comput. Chern. 22 (2001) 641. Aabloo, M.
Klintenberg, J. 0. Thomas, Electrochim. Acta 45 (2000) 1425. Kasemagi, M.
Klintenberg, A. Aabloo, J. 0. Thomas, Solid State lonics 147 (2002)
.
ohansson, Polymer 42 (2001) 4367 .
J. Kearley, P. Johansson, R. G. Delaplane, J. Lindgren, Solid State lonics 147 (2002)
.
C. Redfern, L. A. Curtiss, J. Power Sources 110 (2002) 401.
ic Electrolytes: An Overview 619
N.J. Dudney
Solid State Division, Oak Ridge National Laboratory, P. 0. Box 2008, Oak Ridge,
Tennessee 37831-6030, U.S.A.
1. INTRODUCTION
In recent years there have been important advances in the stability,
ty, and performance of lithium and lithium-ion batteries. Many of the
trolyte materials being examined are based on organic liquids or
mers, although solid inorganic electrolytes still have an important role a
variety of applications. There are numerous reports of new
positions, advanced synthesis routes, and novel cell architecture for the s
and ceramic electrolytes.
For any rechargeable lithium battery, the electrolyte material must
it the repeated and rapid transfer of Lt ions between the anode and
ode over the expected range of operating conditions (voltage,
perature, and current), without significant deterioration. Additionally,
ideal electrolyte material would be an electronic insulator, ultra-thin,
tweight, free of hazards and inexpensive. The inorganic solid
trolytes offer both advantages and disadvantages over liquid and organic
mer electrolytes. For the required rapid transport of Lt across the
trolyte, the product of the resistivity and electrolyte thickness must be
mized. Typical room temperature conductivities are 10-1 S/cm for
ds, 10-2 S/cm for superionic conductors such as -alumina, 10-3 to 10-6
for various gel and solvent-free (dry) polymers, and 10-4 to 10-8 S/cm
ypical glass and ceramic solid electrolytes. If formed as a very thin film
ess than about 1 f1m, even a rather resistive material may compete
rably when compared to a much thicker cast polymer or liquid-filled
us separator membrane. An ultra-thin electrolyte also provides a
iderable savings in terms of volume and mass for the battery, if not
et by the need for a thick inactive support material. It comes as a
Chapter 20
rise to many that 1-10 )..tm thick glass and ceramic sheets are quite
ible.
One of the principle advantages of the inorganic solid electrolytes is
these materials are generally single ion conductors, meaning that only
Lt ions have an appreciable mobility, while the anions and other cations
a rigid framework. This has many beneficial consequences.
inating the anionic concentration gradient across the electrolyte may
suppress undesirable side reactions or decomposition of the electrolyte.
single ion conductor expands the choice of electrode materials and
its operation at higher voltages. Few polymer electrolytes mimic this
1
le ion behavior and they are usually associated with a higher resistivity.
wise, a negligible electron transport across the electrolyte is necessary to
ent self-discharge of the battery. For a shelf life of months or years, the
trolyte must be an excellent insulator with an electronic conductivity <
2
S/cm. Conductive electron paths may be formed along lithium
rites for liquid or polymer electrolytes, while grain boundaries, pores, or
ks may become electronic paths for glass and ceramic electrolytes.
ificant concentrations of multivalent ions in inorganic electrolytes may
contribute to electronic transport. Ideally, the lithium ion transference
transport numbers should be as close as possible to identically 1.0.
The resistance of the solid electrolyte-electrode interface is just as
ortant and often as large as the bulk resistivity of the electrolyte.
ause of this, reports of an electrolyte's properties are insufficient without
valuation in contact with reversible electrodes comparable to those for
Li-ion battery. One might assume that ion exchange across the solid
rolyte-electrode interface would be more resistive than that with a liquid
oft polymer electrolyte. This is not necessarily the case, due to the
ation of a "solid electrolyte interphase" (SEI) reaction layer, which adds
e interface resistivity for most liquid or polymer electrolytes. The solid
rode-electrolyte interface has the added advantage of providing a dense,
surface that maintains its integrity and inhibits roughening of the
ostructure that leads to mossy or dendritic deposits. Without good
sion however, stresses may cause a rigid interface to form permanent
ks or voids.
The nature of the highly active lithium battery chemistry places severe
rochemical requirements on the electrolyte. Many of the glass and
mic electrolytes proposed in the literature decompose under high
ges or react with lithium, but some are indeed believed to be
modynamically stable. The organic electrolytes, on the other hand,
riably rely on the formation of a protective SEI layer to limit the reaction
composition rate.
As there are hundreds of different electrolyte compositions in the
ture, this chapter will cover work published since 1995 or 1997,
ding newly investigated electrolytes, as well as known electrolytes that
sy and Ceramic Electrolytes for Lithium and Lithium-Ion Batteries 625
. GLASS ELECTROLYTES
With many glassy lithium conductors already in the literature, the
rity of recent work is tending away from the traditional glass processing
xplore twin roller, vapor deposition, high-energy mechanochemical
ng, or even explosive glass forming processes. These processes expand
normal glass-forming region. A number of studies reveal that
positions with a higher concentration of the Li modifier, e.g. LizO and
give a higher conductivity and lower activation energy for u+
port. For many compositions, the glass is almost fully depolymerized.
Commonly perceived advantages of glassy electrolytes over
alline ceramic materials include: isotropic conductivity, ease of
cation as dense thin films, and lack of grain boundaries. Grain
daries in polycrystalline or granular electrolytes have an inherently r
density, higher defect concentration, space charge polarization, and
gated impurities. These may form resistive barriers for Li+ but fast
for other species, increasing the likelihood of an internal short circuit,
ecipitated Li metal dendrite, or breakdown related to electrochemical
s. The nature and effect of boundaries in glass electrolytes formed as
ed compacts of crushed glass powder has not been determined.
Chapter 20
rly small, but the precise effect of the transition metals on the electronic
uctivity has not been assessed.
All solid state batteries, In/LiCo02 and Li/LiCo02, have been
ared using the lithium silicon oxysulfide electrolyte compositions shown
30
able 20.1. '31 '32 The cells were formed by compacting the ground
trolyte powder together with the cathode mixture and In or Li foil in a 1-
diameter insulating tube. After several initial cycles, the cycling
eeded with a 100% coulomb efficiency for 100 cycles indicating good
trochemical stability. Although the electrolyte thickness was not given,
bserved cell resistance (0.7-2 kohm) was higher than anticipated for an
mated 0.5 mm thick layer of the highly conductive electrolyte. This high
nal resistance was attributed to the electrode interfaces, an important
e discussed briefly below. The reversible capacity was only 60 mAh per
LiCo02 when using the In anode, while a higher capacity of 80-90
/g is reported for the Li cell. A variety of cathode materials including
S2 and LizFeC14 were evaluated with this same tubular cell construction
ompacted powders.33 These cathode materials, which had been deemed
ly reversible in liquid cell experiments, were shown to cycle with good
city and efficiency in the cells using the glassy solid electrolyte.
In recent years, little additional work has been reported for Li halide
d glass compositions, although numerous compositions were reviewed
er.' The most conductive Li electrolyte reported to date, >2 mS/cm, is a
6
c transport? The mechanism of ion transport in the glassy structure
inues to be a topic of experimental and theoretical research.
. CONCLUSION
KNOWLEDGEMENT
RENCES
. Lee, K. H. Joo, J. H. Kim, J. H. Woo, H.-J. Sohn, T. Kang, Y. Park, J. Y. Oh,
Stare Ionic.\ 149 (2002) 59.
. Rao, M. Ganguli, "Lithium Ion Conducting Glasses", in Handbook of Solid State
eries and Capacitors, M. Z. A. Munshi, Ed., Singapore, World Scientific, 1995.
. Jones, J. R. Akridge, "Lithium Glassy Electrolyte Batteries", in Handbook of Solid
Batteries and Capacitors, M. Z. A. Munshi, Ed. Singapore, World Scientific, 1995.
. Robertson, A. R. West, A. G. Ritchie, Solid Stare Ionic.\· 104 (1997) 1.
unke, Prog. Solid State Chem. 22 ( 1993) 111.
manaka, T. Shimizu, G. Adachi, Solid State Ionic 62 (1993) 167.
Bates, N.J. Dudney, G. R. Gruzalski, R. A. Zuhr, A. Choudhury, C. F. Luck, J.D.
ertson, 1. Pmn!r Sources 43-44 (1993) 103.
u, J. B. Bates, G. E. Jellison Jr., R. X. Hart, J. Electrochem. Soc. 144 (1997) 524.
. Bates, N. J. Dudney, B. J. Neudecker, A. Ueda, C. D. Evans, Solid State Ionic. ·
(2000) 33.
ereda, R. B. Goldner, T. E. Haas, P. Zerigian, Electrochem. Solid-State Lett. 5
2) A239.
. Park, S.-H. Lee, B.-I. Lee, S-K. Joo, E/ectrochem. Solid- State Lett. 2 (1999) 58.
Jeon, Y. W. Shin, S. C. Nam, W. I. Cho, Y. S. Yoon, J. Electrochem. Soc. 148
1) A318.
ee, S. A. Akbar, C. 0. Park, Sensor Actuat. B-Chem. 80 (2001) 234.
. Yoon, W. L Cho, J. H. Lim, D. J. Choi, 1. Power Sources 101 (2001) 126.
Dudney, B. J. Neudecker, Curr. Opin Solid State Mater. Sci. 4 ( 1999) 479.
ereda, N. Clay, A. Gerouki, R. B. Goldner, T. Haas, P. Zerigian, 1. Power Sources
2000) 20 l.
hao, Z. Fu, Q. Qin, Thin Solid Films 415 (2002) 108.
. Choi, W. I. Cho, B. W. Cho, H. S. Kim, Y. S. Yoon, Y. S. Tak, Electrochem. d-
State Lett. 5 (2002) A 14.
. Roh, S.-D. Lee, H.-S. Kwon, Scripta Mater. 42 (2000) 43.
ang, B.S. Kwak, B. C. Sales, J. B. Bates, J. Non-Cry.1·t. Solids 183 (1995) 297.
Dudney, in preparation.
e Sauze, E. Gueguen, R. Marchand, J. Non-Cry.1"t. Solids 217 (1997) 83.
irkc, W. F. Chu, W. Weppner, Solid State Ionic.\ 93 (1996) l.
. Kennedy, Z. Zhang, J. Electrochem. Soc. 136 ( 1989) 2441.
hang, J. H. Kennedy, Solid State lonics 38 (1990) 217.
Morimoto, H. Yamashita, M. Tatsumisago, T. Minami, J. Am. Ceram. Soc. 82 (1999)
2.
ayashi, S. Hama, H. Morimoto, M. Tatsumisago, T. Minami, 1. Amer. Ceram. Soc.
2001) 477.
inami, A. Hayashi, M. Tatsumisago, Solid State Ionics 136-137 (2000) 1015.
. Bartholomew, D. M. Young, A. J. G. Ellison, J. Non-Cry.1t. Sol. 257 (1999) 242.
Hayashi, H.Yamashita, M. Tatsumisago, T. Minami, Solid State lonics 148 (2002)
Li.POyNz "
(x=2.9, y=3.3, z=0.46, typical) (0.24 < z < 1 .2)
"' 'o>
Li. POy Nz ...
t""'
§:
-
Li,BOy Nz
§
:::...
t""'
§:
Li 2S•P2S5 ii:'
(0.75, 0.25) :!
::s
1
Ea, activation energy for Li conduction from simple Arrhenius relation near room temperature. Those shown in ( )
were estimated
from the published graphs.
·
2
t., transference number= transport number for these single ion conductors, generally by Wagner de polarization .
3 ll.V, low/high voltage range of the electrochemical window with respect to Li+/Li.
0\
w
\0
Li2S•SiS2
(0.6, 0.4)
LizS•SiS2•LixM04 M=Si,P,Ge
(0.57, 0.38, 0.05)
LizS•SiS2•LixM04 M=Si,P, Ge
(0.57, 0.38, 0.05)
Li3N•SiS 2
(0.4, 0.6)
9
-€;;:
....
N
0
{mole (kJ/mol) {V)
ratio)
'':"''"
l..a213-xLiJzTi03 solid state reaction & 1 X 10'' 0.7 28±2 1.51>4
s.
Q
cyclic voltamrnetry [37,
0,03 Sx :s; 0.167 pressed x=0.13 10"
& titration 41]
l..a113-xLi3zTa03 solid state reaction & 9x 10"' 2 !'
[38] ;
0.025SxSO.I67 pressed x=0.06 to"' ·
[48]
pressed (1.8 X 10'3)0
-
-
-
-
-
-
-
-
·
-
Lil·2.t(Sc1.,M.J:z(PO,J, M = Zr, Ti, .x=O.l,0.2
Li3Sc ,5Feo.s(P04)]
Li6.1Yo.61Sio.J90s.l6
1.4..rGeJ•.Pa.S4 (.r-:0.25)
L4.rGeJ.J> ..s.
(.x=0.6S)
Li9Bi2Cl3
Li31nBr6
Li3P+3LiCI
composite
LiCF3S •NbF, (0.33, 0.37)
Li,SiAl01
Li_.BPO..,xn [62)
x=0.07
- -
.,
1'\
apter 21
M. Broussely
1. INTRODUCTION
So called "industrial batteries" are generally sized in the range from
to hundreds of kWh, addressing a market roughly divided in two
ments: transportation and stationary applications. This market is mainly
essed by lead-acid and NiCd systems. Large special batteries for space
defense generally use more advanced and sophisticated chemistries, such
ickel-hydrogen, silver-zinc, silver-aluminum oxide, etc. One of the most
ting questions in this battery business today is: will the "Li-ion
lution" which occurred in portable batteries extend to larger battery
ems? Obviously, that would only occur if two main conditions were
fied: better answer to a real need, and in a cost-effective way. This
ter will review the most important applications in which Li-ion might
g benefits, and will describe the technical status today.
Going back a little over the last 5 to 10 years, the possible use of Li
in large batteries was first considered in electric vehicles, for which high
ific energy is crucial to grant a sufficient range. In parallel with the
lopment of portable Li-ion, very important R&D programs started about
ears ago aiming at demonstrating the feasibility of such batteries in
tric vehicles. These programs were supported by a strong financial effort
over the world: European Commission and french government
nizations in Europe, DOE I USABC in USA, MITI I NEDO I LIBES in
n (addressing first the stationary batteries for home individual power
ces). Saft was one of the few companies involved from the beginning.
n, with the fast developments of REV's, Li-ion also had the opportunity
emonstrate its very high potential in terms of specific power and power
ity.
Chapter 21
This is certainly the critical part of aLi-ion cell that is responsible for
differences in cell properties. The use of lithiated transition metal es,
e.g. LiNi02 and LiCo02, as possible positive materials for
1
rochemical power sources was discovered more than 20 years ago.
e then, these compounds were widely investigated as positive materials
2
rechargeable lithium cells. •3 When LiCo02 became the undisputed
4
ode of the Li-ion, it was available in very small quantities, and LiNi02
ly did not exist. Studying the possible preparation processes for
i02,5 the difficulty of preparing a "good" electrochemically active
rial was pointed out. It was later found that the tendency for Ni2+ to
py sites in the u+ layers introduces a structure distortion, which
ically affects its ability to de-insert/insert Li+ ions.6 Therefore, in spite of
Lithium-Ion Batteries for HEVIEV and Other Industrial Applications 647
9
Substitution by cobalt (ideally 20%) was early identified as a solution
2
void Ni + formation, allowing easy manufacturing of the appropriate
tural form. In the frame of the LIBES program, the mixed Ni/Co
10
rial LiNi0.7Co0.302 was studied by Sanyo to design a 76 Ah cylindrical
dedicated to home load leveling systems.
For some years, many "doping" metallic ions were tested by
11
erous laboratories, showing interesting specific properties.
12
minum was found to produce a significant stabilization of the
hiated form, inducing a better thermal stability, as shown by differential
13
ning calorimetry (DSC) experiments.
That discovery focused the research work, made jointly with material
ufacturers, towards the definition of a mixed Ni/Co/Al material, which
ars to exhibit excellent electrochemical properties, now recognized by
14
rallaboratories. The exact composition used by Saft is still proprietary,
he ratio of nickel to total metals (except lithium) is greater than 0.75,
wing the expectation of a reasonable cost of the raw material.
Unlike LiNi02, LiNi 1_x-yCoxAly02 does not exhibit monoclinic/
gonal phase transformations during cycling. This may at least partly
ain the excellent stability on cycling, as described later. The crystalline
ture is very similar to that of LiCo02, however the charge discharge
es are different, as shown in Figure 21.1.
Figure 21.2 describes a particular feature of nickel based compounds.
2
to a small remaining quantity of Ni +, the first charge/discharge cycle is
Chapter 21
-- - 1-
-- t--...
1\
-LINIO
-- ueoo,
PC/EC/3DMC+LIPF (1M)
e 21.1. Compared voltage profiles Figure 21.2. Deep DOD voltage profile
tal oxides at low rate (C/10, 25°C). of a LiNil-x-yCoxAlyOz/graphite cell at low
rate (C/10). Ambient temperature.
rably to pure LiNi02 or LiCo02, and is a big advantage over LiMn 204,
which interaction with electrolyte is still a major drawback.
.1.2. LiMnz04
There are two main reasons put forwards to promote the use of the
n204 spinel as the positive for a large Li-ion battery. The first usually
tioned is the cost, expected to be the lowest. However, the cost of Mn02
in primary dry cells is often given as example and goal, ignoring the
ssary lithium salt and added processing costs. In addition, the pure
erial alone fails to provide the sufficient chemical stability to ensure the
ired long life time, and additional processes such as coating are
ssary. Finally, as long as costs of other battery components are not
tically reduced, the final cost/Wh is strongly impaired by the lower
gy density achieved, compared to nickel based materials. In high power
, where the cost/W is more relevant, the lower relative amount of active
erials reduces their influence on final cost. Finally, it is quite difficult
y to predict which one of Mn or Ni based products will lead to the more
effective system, but it can be anticipated that the difference would be
ginal.
Safety is the second mentioned advantage, which is of a particular
ortance for large batteries. This aspect is described in more detail in
her chapter of this book, and is only briefly discussed here. The totally
hiated phase, A.-Mn02, which is the regular full state of charge, is
ally thermodynamically stable. This is not the case of Ni02 or Co02,
h are ultimately formed when LiNiOz or LiCoOz cells are fully
charged. These non-stable forms may decompose via exothermic
tions with the electrolyte and can result in cell burning. This generally
rs when the cell starts to overheat because of electrolyte electrolysis,
the positive material is totally oxidized. The combustion is all the more
tive since lithium metal will have been deposited on carbon electrode,
to the excess of lithium extracted from the positive. Although there are
rent means provided for preventing all Li-ion batteries to go into
charge, the better stability of LiMn2 04 provides an additional safety
l.
That does not mean however that the cell is intrinsically safe and
d be abused in any manner. The oxidized form, A.-Mn02, is still a strong
izing agent and can react violently with organic solvents when
heated to about 300°C. This is usually studied using differential
ning calorimetry (DSC), and can be observed when the experiment
lves a mixture of active material and electrolyte as in a real cell. Figure
describes the DSC heat output from 100% charged materials, in contact
electrolyte in proportions similar to a real cell. It must be pointed out
Chapter 21
the heat power (W) is not related, as is usual, to the weight of the
ple, but to the quantity necessary to bring one Wh to the cell, which is
er to the real case.
The following comments can be made: 1) pure LiNi02 exhibits a low
t temperature for thermal runaway (temperature sharp peak), as low as
t 200°C. 2) Although heating delithiated LixMn2 04 alone would not
uce any thermal effect, the presence of electrolyte leads, as expected, to
ong exothermal reaction at about 300°C. 3) Aluminum doping in the
ed Ni-based material brings a very significant improvement, ranking the
rial to an even better level than LiMn 204 • Therefore, LiMn204-based
e batteries must be protected against overcharging, just like the other
ems. Indeed, a high rate overcharge will produce heat as soon as
trolyte is oxidized, i.e. from the beginning of overcharge, and may reach
point where the positive material reacts with the flammable electrolyte. t
transfer in cell design is also one of the key factors in improving ry
safety, regardless of the electrochemical system.
Beside the advantages of a better safety, a LiMn204-based system still
ers of poor calendar life even at moderately high temperature, which has
attributed to the tendency of Mn 3+ ions to disproportionate into Mn4+
Mn2+. The problem comes from the solubility of Mn 2 + ions in the
rolyte, which reacts on the negative side, and destroys the stability of
passivation layer on the carbon, which is the key point of Li-ion system
ility. Although this can be tolerated for consumer batteries, whose life
ctation is not too high and working temperature normally low, this is a
drawback for industrial batteries. Although improved solutions have
studied in the past years, the problem is not yet satisfactorily solved.
0.03
0.025
esc
(W/Wh) 0.02
0.015
0.01
0.005
e 21.3. DSC of different positive materials at full state of charge in the presence of
olyte. Stainless steel crimped sample vessels.
Lithium-Ion Batteries for HEVIEV and Other Industrial Applications 651
. EV APPLICATIONS
.1.1. Weight
e 21.1. Comparison of different technologies of EV batteries for a typical size of 250 kg.
hnology
icle curb weight
tery weight allocation (typ.)
tery structure cooling etc.
dule weight allocation
rgy density (module)
board energy
c. range @ 120 Whlton/km
chnology
ttery volume allocation (typ.)
ttery structure, cooling etc.
dule volume allocation
lumetric Energy (module)
board energy
nge @ 120 Wh/tonlkm
.1.2. Volume
.1.3. Power
e 21.3. Power characteristics comparison for a 250 kg battery using different chemistries.
hnology
icle curb weight
ery weight allocation (typ.)
ery structure cooling etc.
ule weight allocation
er density (module)
ery Power
The general cell design is described in Section 21.3 All battery makers
selected the wound coil technology that is the optimum industrial
ion to achieve a low process cost. These coils are then inserted into
drical or round-edge prismatic metallic cans. The can material may be
el-plated steel or aluminum depending on which of the coil polarities is
ected to the cell structure. A pressure release valve
Chapter 21
ls
minal Voltage
d Capacity at C/3
ical Power at 80% DOD, 30s pulse
ensions (0 I h)
ical Weight
cific Energy
cific Power at 80% DOD
rgy Density
er Density at 80% DOD
Capacity
Nominal voltage
Whlkg
Wh/1
W/kg at 85% DOD
21.6. Main characteristics of JSB high-energy cells for EV (from JSB catalogue).
acity Capacity
ight Energy
ensions Energy density
tage (V) Specific energy
19
NGK Insulators designed a 25 Ah cylindrical cell using substituted
ganese spinel, described in Table 21.8. The material and appropriate cell
gn allow to get a an improved safety behavior in severe abuse
ditions, such as nail penetration and overcharge. High power cells are
developed (see 21.5.3.3).
e 21.5. Open view of the Saft Li-ion battery for EPIC Daimler Chrysler van.
9 : -Howe< 50,0
ti SO -+-Energy f
0,0
ti
J....._J....._J...._..L..,_.,!_..J_..J j j .J...._J
0
600 BOO 1000 1200
0 5 10 15 20 25 :D 35 40 45 50 0 200 400
nergy remains remarkably stable during more than 1000 cycles, just like
power ability, measured with a 300 A pulse at 80% DOD. These results
trate the good behavior of the positive material, as well as the stability of
passivating layer on the carbon, preventing lithium loss. As a matter of
Figure 21.2 describing the lithium excess has been recorded after 1000
es, which demonstrates the excellent stability of the layer, which is
20
y due to the patented VC additive.
::::.-:::!:
-
·--- ·-
.3.4.3. Safety
As discussed in Sections 21.2 and 21.3.1, the main safety concern for
-ion battery is the fact that many of the components are flammable,
ing with the electrolyte. Cell overheating may ultimately lead to battery
ion in the worst case, through active material decomposition. Therefore,
r intrinsic cell stability would come from the discovery of a non- or
flammable electrolyte, which is presently one of the most important
arch tasks for Li-ion throughout the world.
There are, however, means to prevent abusive situations at all levels,
cell to complete battery assembly. They usually include shutdown
rators internal to the cell, fuses to avoid heavy short circuits, and voltage
temperature monitoring to shut down power in case of overvoltage or
heating. Ultimately, no explosion should ever occur, thanks to the case
safety vent design. Considering the worst case, should an individual cell
e, the fire must be contained and/or extinguished, and not propagate to
whole battery. These are the considerations that are taken into account
attery design on a case-by-case basis.
Complete EV batteries have been subjected to abuse tests by
akers in the frame of joint development programs. For example, a
21
re crash test was applied to the Saft battery for EPIC, simulating an
dent where the battery bumps against a tree or a telegraph pole. The
ct speed of the battery was 52 kmlh. Sparks during crash resulted in a
as the electrolyte was burning. It was allowed to burn for about 15 min,
was extinguished without any problem.
The observations were as follows: there was no explosion, the fire was
dramatic and could have been easily extinguished in a short time, there
a very slow transmission to other modules. At the end of the experiment,
the modules of battery in the crash zone were destroyed. It must be
hasized that the crash with the battery system only was really a worst
. It is much worse than for a real crash with the same battery integrated
erly in a vehicle. From this behavior the carmaker considered that the
is acceptable and allows carrying on with the development of this
nology for EV application.
Chapter 21
At the time this chapter is being written, the Li-ion for EV is not a
mercially available product, and real life experience with EV Li-ion
ries is limited to fleet operations. More than one hundred vehicles using
on batteries have now been put on the road by various carmakers. These
ries are corning from small production lines set up by the battery
ers. Nissan initially experimented with Li-ion batteries using Sony
alt positive) batteries. After Sony stopped this battery activity, Shin
e (manganese positive) started supplying the batteries. Most european
manufacturers are experimenting with the Saft batteries. Some of these
cles have achieved more than 40000 km and the batteries are still in
operation. A 200-km range on a single charge has been demonstrated on
a small Peugeot 106 and a large Chrysler Voyager minivan. Among he
vehicles running with these Li-ion batteries, not a single problem
ed to the cells has been reported.
.3.6. Cost
• For the same onboard energy, a vehicle fitted with Li-ion will have a
longer range thanks to the lighter weight of the battery.
Comparing two identical vehicles, one fitted with aLi-ion battery, and
other with a lead acid battery of the same energy, the lead acid powered
cle consumes from the charging network twice the energy needed to the
n vehicle to achieve the same driving range. Assuming an acceptable
hasing price, and taking into account zero maintenance costs and long
ating life, Li-ion can be a very efficient technology based on life cycle
.
5. HEV APPLICATIONS
The hybrid car concept is based on combining two or more different
er sources in a vehicle, which would complement each other. The goal
nd this approach is making cars more environmentally friendly by
cing emissions as well as improving gas mileage.
Usually the primary power source in the vehicle provides main
ulsion and the second power source is only used when an additional
er boost is needed, for example during vehicle acceleration. The primary
ce has long been considered to be a smaller, more economical internal
bustion engine (ICE). In some more recent scenarios it is foreseen that
ICE could be replaced by a fuel cell. The second power source is a
argeable battery capable of delivering the high power pulses needed for
cle acceleration. In the early days of HEY, other technologies were
idered as this second power source. Supercapacitors are capable of
atedly delivering hundreds of thousands of high rate pulses even at low
eratures. However, the very low energy density of these devices,
lly in the range of 3 to 5 Wh/kg, renders them unsuitable for this
ication. Another energy storage device considered for supplying the
er for the acceleration burst was the flywheel. Although this technology
nteresting potentialities, carmakers were not convinced of its ability to
h maturity due to its complexity, and the issue of containment. As a
lt, the rechargeable battery remains the only viable option that can meet
tringent requirements of the application.
haracteristics
ulse discharge power
eak regenerative pulse power
otal available energy
Minimum roundtrip efficiency
old cranking power at -30°C
ycle life at 25Wh cycle energy
alendar life
Maximum weight/volume
perating voltage limits
Maximum allowable self-discharge
emperature range peration
urvival
The competing battery systems for future hybrid electric vehicles are
-acid, Ni-MH and Li-ion technologies. Other technologies such as Ni-Zn
proven deficient in one or more critical performance aspects.
Lead-acid batteries have well established roots and traditions in the
motive industries. The main advantages are very low cost and proven
power capability. Drawbacks include low energy density, short life and
ations in charge acceptance at lower states-of-charge making this
em poorly suited for traditional HEV.
Ni-MH batteries became popular as a replacement for Ni-Cd. They
ide better energy density and their life, longer than that of Lead-acid
ries, is apparently sufficient to meet some of the life requirements in
application. The main disadvantages are their poor performance at
me temperatures and high self-discharge rates.
A number of different types of hybrid vehicles are being considered
he automotive industry. Several japanese carmakers have already
duced power-assist hybrid vehicles. These vehicles presently use NiMH
ries due to the more mature production status of this technology
pared to Li-ion.
From the very beginning Li-ion has promised and achieved great
ntage in energy density and specific energy over existing technologies.
ever, due to the use of organic electrolytes with fairly low ionic
Lithium-Ion Batteries for HEV/EV and Other Industrial Applications 665
ductivity, its possible use in high power applications was not originally
idered. Only in the last several years, R&D conducted for hybrid electric
cles has proven that Li-ion can successfully compete in the high power
argeable battery arena. The success of the development work has
mpted interest from other industries which could use a light but powerful
ce of energy. Why is this possible? The main power limiting factors in
on cells are indeed the low mobility of u+ in the organic electrolyte and
diffusion rate inside the active materials. The most practical way to
ease the power is therefore to increase the electrodes surface area, by
ing them thinner, which also reduces the ionic pathway in the electrolyte
arator and electrode porosity).
Li-ion technology utilizes electrodes that are already much thinner
electrodes used in other rechargeable chemistries. Commercial Li-ion
typically have negative and positive electrodes with thicknesses in the
e of 200 microns. This means that coating, calendering, slitting and
ding processes of thinner electrodes for high power cells must be
rolled much more tightly than traditional mass-produced Li-ion cells.
Another technical challenge is associated with the need to carry much
er currents. Therefore, engineering of the high power cells must address
nal electrode bussing as well as outside connections.
In order to meet the low temperature requirements, which commonly
e by default with most high power applications, electrolyte properties
t be specially tailored. High ionic conductivity, low freezing point and
viscosity enable low temperature discharge. This is accomplished
ugh selection of appropriate solvents and optimizing salt molar content.
ever, selecting solvents that would improve low temperature
ormance may sometimes result in compromised stability at elevated
perature and reduced calendar and cycle life.
24
.3.1. Perl'ormance of Saft Li-Ion HEV Battery Technology
.3.1.1. Power
A key parameter for a high power battery system is the point where
battery can deliver as much discharge power as it can accept charge.
point is called the "sweet spot". Usually this point is around 50% SOC is
determined by conducting a series of discharge and charge pulse at rent
states-of-charge. Figure 21.10 shows the sweet spot for one of
s high power cells.
A different way to characterize the power capability of a system is to
ect it to a complete discharge at various rates. In addition to the voltage
it is important to make sure that the cell does not overheat during the
harge because overheating will severely reduce its life. Figure 21.11 s
a set of discharge curves under different discharge loads. Figure 21.12
s the cell temperature rise under 10C (6 min) discharge.
Cells
Nominal Voltage
Rated Capacity at C/3
Power 50% DOD, ISs pulse
Dimensions (0 I h)
Typical Weight
Specific Energy
Specific Power
Energy Density
Power Density
--
4,--------------------------,
0
0
l ,/'2sa: regen
>
3,6
3,8
3,4
1i: 3,2
f 3
0
/ .......... 3 ---- -z
0
/ "18se:: clscharge g
'iji 2,8 ----------------------'k\--\-1
l r
0 u 2,6 -----------------------\\--1-i
0 2.4 -f-.-:ce'"lNl "'o=m>=ina"c'l -=a=-=pac:;;citycc-.•7."'5A'"h--------------l
2,2 -----------------·
0
2 -r y
0
0% 20% 40% 60% 60'% 100% 0 2 4 6 8
Depthofrlscharge Capacity, Ah
-- t
80 4,5 ,............................................................................................................................................,
CURRENT
70
35
'
50 !.: .p.....i\1------.>....J. ------'...........;,
80 -- ' f I f I i
40 291Wikg 293W/1<g c2=98.:..:W.:..c/1<::!.g-i
2
30 1,5 t---------------i
TEWERATURE
20 1i-- PNGV Cold Cranking Teat --;
10 0,5 _Three 40Apulo ondo al·40oC
Cal Nominal Copocly.7.5Ah
0 o+-
0 2 4 8 0 2 4 8 8 10 12 14 18 18 20 22 24 26 28
capacily, Ah Time,aec
Cycle life of Li-ion like any other rechargeable system varies with
h of discharge, cycle rate and temperature. Figure 21.14 shows that Li s
capable of maintaining power and energy over 1 million shallow high
cycles, representing the HEY utilization profile. Figure 21.15 displays
eep cycles recorded at C/1 rate. 5000 cycles are projected at an average
ient temperature of 20°C. Except for NiH2, which has much lower
gy and power density (see Section 21.7.1.), this performance is
atched by any other battery system. An evaluation of cycle life
ndence on temperature still remains to be performed.
18
--RoomTeffi!>iratiire
--
16
14 ;!. W'k ........... .
12 ln'/o :"""
E kr------ --
10 ln'/o
= --
8 00'/o
6
·-«1'kT---------------
4
i
2 10'k ·----
0 O'k
0 250 000 500000 750 000 1 000 000
o &XJ 1000 1&XJ am 2m 3lD
Cycles
(),des
e 21.14. Deep cycling of a high
cell under HEY load profile Figure 21.15. Cycle life of a high power cells
dic 100% DOD diagnostic). (Crate, 100% DOD).
Chapter 21
Last but not least is the calendar life of the system. It has been
ected that Saft HEY cells could maintain the required peak power and
gy necessary for the HEY application over a period of 15 years. These
ates are based on actual data of cells stored at various temperatures and
projections based on the Arrhenius law. It has to be mentioned that the
re of the application, which requires the battery to be at around 50%
-of-charge, is very beneficial for achieving long calendar life in Li-ion.
re 21.16 shows actual capacity and power fading from cells on storage
ree different temperatures.
%
I . ._.
120%,--------------,
A 2s•c
% 2 •c t -- - !:'! ·c
-.o;,c; 1f!.100%
t . ""' "- /-t>---.-...---..,,_._
% eo% t------;;:;=:;=Y.::;:=::::;::=:::::::: ::; :; ;--oo_·cl
]Average Gapaclty flleaau&d at 2s•c1 ]Avaraga Paak Power Maasurod at 2s•q
_J Cella stored at 50%SOC l ': 60% --l Cella stored at 50%SOC _
e 21.16. Capacity and power evolution of cells stored at 50% SOC at various
ratures.
Test descriPtion
rnal short VL8P
VL16P
h VLSP
VL16P
ct Droo Test
rate overcharge
A)
discharge (reversal)
25
er discharge capability, as shown in Figure 21.18. A peak power of
W/kg can be obtained at 50% SOC, 20°C.26•27
Japan Storage Battery has also developed a high power design
atteries using LiMn 204. The 3 Ah cell exhibits 1500 W/kg on
Chapter 21
e 21.17. Example of 50 kW high power battery for power assist HEY (2.8 kWh, 8 Ah at
).
ge and discharge at the 40% state of charge "sweet spot". Good deep
D cycle life is obtained at lC rate at 25°C, while some capacity loss is
ved at 60°C (Figure 21.19), as is typical of the Mn-based cells.
Using a substituted manganese spinel, NGK insulators 19 developed
ntly a high power 4 Ah cylindrical cell of enhanced safety, exhibiting a
ific power of 2500 W/kg at ambient temperature and 800 W/kg at -25°C.
edium size 12 Ah cell has been also designed for a 42 V battery.
45'1::
:--.
- 25'1::
60'1::
5 10 15 0
0 200 400 600 800 1000 1200
discharge energy (Wh) Number of Cycles
6. STATIONARY APPLICATIONS
Stationary battery applications can be categorized into broad subsets
d on the way the battery is charged and discharged.
Standby (float) operation. The battery is maintained in a fully
ged condition by a so-called 'float' or constant potential charger. The
ery and the system loads are connected in parallel with the charger
inals. In this mode of operation, the charger supports continuous loads
the battery accepts whatever current is necessary to maintain it at a set
age. If the charger output is lost, the battery immediately assumes the
em loads.
Intermittent charging (cycling) operation. Batteries in these
ications are charged only when the primary energy source is available.
primary energy may come from a solar (photovoltaic, or PV) panel, a
d turbine, an engine-driven generator, the utility grid, or some other
ce. When a load is connected and the primary source is unavailable or is
equate to support the load, the battery is discharged.
Buffer operation. Halfway between the other application groups, the
ary energy source in these applications is generally available but is
uently insufficient to supply the imposed loads. This may be because the
s exceed the rated output of the energy source, or because the rate of
ge in the loads exceeds the capability of the energy source to respond.
n the load level falls below the rated output of the primary source the
ry is recharged.
Each of these application sets imposes its own particular requirements
Li-ion batteries. The following sections describe the applications and
requirements in greater detail, and discuss the advantages and
dvantages of Li-ion batteries in meeting those requirements.
The battery duty in these applications can be quite varied. Some duties
ire the battery to support a relatively constant load for a duration of
ral hours, such as telecommunications network switching or railway
aling. Others demand a short, high-rate discharge, such as in UPS
ems or generator starting. A few applications feature a mixture of low l
continuous loads and short high-rate loads, such as utility switching
ems.
Batteries may be located in controlled temperature environments, or
be subjected to extreme temperatures in outdoor cabinets. Battery
gns that minimize high temperature aging and maximize low
perature performance are obviously beneficial in the latter case. The high
metric energy density of Li-ion can be an important factor in some
ications, particularly in cabinetized systems, but may be unimportant in
rs.
0 Although Li-ion batteries are more
D
=
......
... energy efficient than most other
b.
types, with better round-trip
·-...-...
D
m
charge/discharge efficiency and
+3/N 4I'C
-&3/NIIfC lower power consumption on float,
D
+3/Nli!J'C this feature is relatively unimportant
+3/N41'C
D -&-3/NfDC in standby operation. More important
m is that the battery operates reliably
0 :m 40 !IX) 8ll 1<DO over a long lifetime. A general good
FkD1irrn,d¥;
behavior of Li-ion batteries on
e 21.20. Residual capacity evolution of continuous float charging has been
high-energy Li-ion cells under float demonstrated.28 '29
e at various temperatures and voltages.
It is worth noting that this type of buffer operation may also require a
ry to absorb power as generator output is ramped down after a load has
abruptly terminated. It is likely that the most efficient way to
mmodate this requirement will be to operate the battery in a less than
charged state. This can be easily achieved with Li-ion batteries, with
relationship between open circuit voltage and state of charge.
Capacity
Dimensions
Weight
Voltage
. SPACE APPLICATIONS
Since the start of the space era, rechargeable batteries were used to
ide power to satellites during eclipse periods. In fact, solar panels are
to power the satellite payload (and recharge the batteries) when the sun
inates them. Batteries take over from the solar panels
Lithium-Ion Batteries for HEV/EV and Other Industrial Applications
675
e 21.2I. Prototype of a Li-ion Figure 21.22. High power lithium ion
ry for Telecoms. battery assembly.
n the satellite enters into the earth's shadow. NiCd in the 60's, NiH 2 in
90's and Li-ion now are the only three technologies for satellite use.
The telecommunication satellites that are presently in geosynchronous t
(GEO), use nickel-hydrogen batteries. The size of telecommunication
llites, and, as a consequence, the electrical power demand during solar
pses, is continuously increasing. In the 80's satellites weighing 0.5 ton,
ered with nickel-cadmium batteries providing 0.5 kW were the standard.
e 90's, the prevailing design was 1 to 3 tons and 2 to 8 kW provided by
el-hydrogen batteries. Now, satellites may weigh 5 tons with power
and in the range of 15 to 30 kW. With the current nickel-hydrogen
nology, 30 kW of power leads to a battery weight of at least 800 kg
esponding to 15-20% of the total satellite weight! That means that the
ry, which was already one of the heaviest pieces of equipment on board
atellite, is now reaching a critical size. Combined with the current price
0-100 k$ to place one kg in orbit, battery weight has become one of the
n issues facing competitive telecommunication satellites.
Thanks to the Stentor project, started in '96, Saft has qualified Li-ion
nology for the first time as the main power source of a satellite. Stentor
e first GEO satellite that will fly this promising new battery technology
pace.
e 21.13. Performance comparison of NiCd, NiH2 and Li-ion batteries for satellites.
Energy gauge
Charge management
Modularity
Started at Saft in '95, the development of Li-ion cells for space took
dvantage of the manufacturing line originally built for electric vehicles.
VES140 (140 Wh and 40 Ah) space cell has the same electrochemical
k (electrode assembly) as for the EV cell. The vital parts of the VES 140
are (Figure 21.23): the terminal, which must insure an extremely high
of hermeticity for the mission duration, and the support of the stack in
ontainer, to withstand the high vibration levels during the launch phase.
Chapter 21
>
8.
r--
>
·a 1000 1!m
Numerous life tests have been performed by ESA, CNES and satellite
ufacturers to measure fading and calendar losses on Saft cells. For GEO,
e than 25 life tests (accelerated or real time) have been conducted (some
still ongoing) with various DOD (from 50 to 90%), various end of
ge voltages (from 3.9 to 4.1V), and with or without electric propulsion
es.3 1.32
More than 56 GEO seasons (corresponding to more than 26 years)
been completed on a 2P3S battery. The test conditions were:
Chapter 21
bility. Figure 21.28 shows half of the 151 Wh battery qualified for the
S Rosetta Lander project.
CydtNumba
ourtesy of Yardney).
cycling of 35 Ah cells (60% DOD max) (courtesy of
Yardney).
um-Ion Batteries for HEVIEV and Other Industrial Applications 683
7.5. Launchers
Launchers also need batteries, mainly based today on NiCd or
/Zn technology. Here again, energy and power density of Li-ion are
ng incentives. A new 1.26 kWh battery has been recently developed by
and is shown in Figure 21.31. A particular feature of this design is to
stand very drastic vibration conditions.
Weight= 18 kg
Capacity =35 Ah
Voltage = 36 V
Operating temperature: -20 to +30°C
Discharge profile is 600 Amp 100 ms +
340 Amp for 5 minutes
8. CONCLUSIONS
Li-ion electrochemistry has now demonstrated its ability to be used in
e batteries for a very wide range of applications. The main features of
e power sources are high energy or high power, extended cycling ability,
calendar life, and high reliability. Continuous technical improvements
expected, particularly in the field of safety, to enlarge the domain of
zation. Their superiority has already been proven in various applications
ugh extended prototype testing such as EVs. The space industry, where
n brings very substantial benefits, has already adopted and qualified this
nology, which will progressively replace the existing NiH2 batteries. e
is no doubt that this system will very soon find its place in the
strial battery market. Opening the road to wide market segments, such as
motive where HEVs or new 42 V batteries are the biggest stakes, will
tly depend on cost reduction. Material costs and associated electronics
esent the major part of the present costs, and should significantly
ease with increasing production level. Ultimately, Li-ion large batteries
expected to be competitive with the today's common systems on the
of life cycle cost.
KNOWLEDGEMENT
The data on products presented here cannot be considered as a warranty from the
facturers.
FERENCES
Mizushima, P.C.Jones, P.J. Wiseman and J.B. Goodenough, Mater. Res. Bull. 17 (1980)
.
lichta, M.Salomon, S.Slane and M.Uchiyama, J. Power Sources 21 (1987) 25.
Broussely, F.Perton, J.Labat, R.J.Staniewicz and A.Romero, J. Power Sources 43-44
993) 65.
Yoshino, Asahi Kasei, JP 1989 293 (priority 1985), US Patent 4,668,595 (1987).
ecerf, M.Broussely, and J.P.Gabano, EU Patent 0345 707, US Patent 4,980,080 (1989).
Delmas, J.P. Peres, A. Rougier, A. Demourgues, F. Weill, A. Chadwick, M. Broussely,
erton, Ph. Biensan, and P. Willmann, J. Power Sources 54 (1995) 329.
arre, M.Broussely and R.J.Staniewicz, Proceedings of EVS12, Anaheim, USA (1994).
Lithium-Ion Batteries for HEV/EV and Other Industrial Applications 685
1 1
E. S. Takeuchi\ R. A. Leising , D. M. Spillman ,
2 2
R. Rubino\ H. Gan\ K. J. Takeuchi and A. C. Marschilok
INTRODUCTION
PRIMARY BATTERIES
ne
nyl chloride
860 2322
o 315 1521
02 308 1540
o02 131 668
ric capacity based on the true density of the active cathode materials.
Chapter 22
4
ment techniques of microcalorimeter/ and AC impedance have been
ensively to study the discharge process in Li/I2 cells.
ue to the solid electrolyte of the system, Li/I2 cells typically supply
ents, in the J..tA range. This current output has been able to meet the
ent for implantable pacemakers over the last several decades.
nally, the formation of a self-healing separator system, as well as high
nd reliability has made the Li/I2 cell a good choice as a pacemaker
ource. Li/I2 cells have demonstrated very long life under usage
ns, as illustrated by the real time discharge data displayed in Figure a
pacemaker battery on test for -10 years. Other larger implantable
attery designs have demonstrated longevity under similar test
ns for over 14 years. The Li/I2 battery from Figure 22.1 was
5
ted as previously described, using a complex mixture of iodine and
inylpyridine as the cathode material. The cell was tested under
kohm constant resistance load at 37°C. At the beginning of the
e test, this load corresponds to -28 J..tA current drain on the battery.
2.500
2.000
1.500
1.000
500
the thionyl chloride acts as both the electrolyte (in combination with a
ch as LiA1Cl4) as well as the active cathode material. In the physical
uction of the cell a carbon black cathode is used to catalyze the
chemical reaction of SOCh. The theoretical capacity of thionyl
de as a cathode material is high, as given in Table 22.1, and practical
etric energy densities as high as 1.2 Wh/cc have been reported for small
6
al LiffC batteries. A typical discharge curve for an implantable LiffC
is presented in Figure 22.2.
4.000
3.500
3.000
> 2.500
.§.
8, 2.000
! 1.500
;
! 1.000
g 500
0
0 500 1000 1500 2000 2500 3000
Capacity (mAh)
2.2. Discharge of a lithiurnlthionyl chloride battery under a 49.9 kohm load at 37°C.
LiffC cells can display voltage delay after prolonged storage periods
the voltage of the cell is temporarily depressed, due to a passivation n
the anode. Techniques have been employed to minimize voltage
including careful attention to the purity of the electrolyte, the
7
tration and type of the electrolyte salt, and the use of additives.
As noted in the introduction, an important aspect of battery systems for
table medical devices is a state-of-charge, or end-of-life (EOL)
ion. In practice this means that it is important for the battery to give
g significantly before all of the useable capacity is depleted. This is an
ant limitation of the LiffC chemistry, since cells provide a flat
rge curve with a rapid loss of voltage at the end of the discharge curve,
trated in Figure 22.2. Several methods have been described to provide
dication, including using S02 discharge voltage as an indicator, as well
use of lithium alloys.6
um Batteries for Medical Applications 691
3500r--------------------------------------------,
3000
2500
2000
1500
1000
500 32kohm
22.3. Discharge of a lithium!CFx battery under constant resistance loads of: 1, 2.7, 7.5,
100 kohm at 37°C.
sulted in the loss of crystallinity of the material over the range O<x<
LixAg2V4011 with a concurrent reduction of Ag+ to Ag0. The
n of metallic silver greatly increased the conductivity of the cathode
, which plays a role in the high current carrying capability of the
5 4
system. At x values >2.4 the reduction of V + to V + and V3+ was
d, and at x>3.8 mixed valent materials were found, containing V 3+,
5
V + in the same sample. The presence of several different oxidation
f vanadium, as well as silver reduction, in the discharge reaction of
cells results in a stepped discharge curve. This step-wise change in
is predictable and provides state-of-charge indication for the battery
ed in the implantable device.
3500 .,----------------------..,
Prepulse Voltage
r------
3000
/
2500
2000
1500 /
Pulse Minirrum Voltage
1000
500
0 --- - ----------- -
0 2 3 4 5 6 7 8 9 10
Tlrre (years)
.4. Discharge of a lithium/silver vanadium oxide battery under long-life test pulse
at 37°C. Four 10 sec, 2 Amp pulses are applied every 30 days, with a background
0 kohm applied continuously.
SECONDARY BATTERIES
in that the anodes are not lithium metal, which allows for safe
ng of this chemistry under controlled conditions. Lithium ion
also enjoy several distinct advantages over other secondary battery
gies such as nickel-cadmium (NiCd) or nickel metal hydride
, namely: 1) lithium ion cells possess a much higher discharge voltage
which is about three times the running voltage of NiCd and NiMH
ries, 2) lithium ion cells display a much lower self-discharge rate, and
memory effects of NiCd and NiMH systems are not present with
27
ion cells. '30
thium ion cells for implantable medical applications have mainly used
28
cathodes with graphite anodes. The gravimetric energy density of is
listed in Table 22.1, and while the value of 131 mAh/g is the lowest
ble, the rechargeability of the material over several hundreds of cycles
n a much larger useable capacity over the life of the battery. The
fe of a small (40 mAh) lithium ion cell designed for medical
able application is displayed in Figure 22.5. Here the cell was
120%
100%
-....
80%
60%
40%
20%
0%
0 100 200 300 400 500 600
Cycle Number
.5. Discharge capacity for a small (40 mAh) lithium ion cell cycled between voltage
.10 V and 2.75 V using a charge and discharge current of 7 rnA at 37°C.
apacity retention. The higher currents used in this test simulate faster
ng (2 h) and high current drain that are a requirement for heart-assist
ations, as outlined below.
Due to the rechargeable nature of lithium ion cells, it has become
ant to develop end of life (EOL) indicators for lithium ion batteries.
ample, impedance spectroscopy has been tested as an EOL indicator
120%
100% .
-·
.
1
80% -
60%
40%
20%
2.6. Discharge capacity for a 1.1 Ah lithium ion cell cycled between voltage limits of
and 2.75 V using a charge current of 550 rnA and a discharge current of 1100 rnA at
29
hium ion batteries. Sealed prismatic Sanyo UF653467 cells with a
al capacity of 930 mAh were cycled at a 1C charge-discharge rate.
286 cycles, the battery capacity dropped from 1005 to 700 mAh.
ance measurements of the fully discharged batteries were made and the
st plots as a function of cycle number showed an increase in the size of
w frequency semicircle with the number of cycles, due to the increasing
cial resistance of the cathode and anode. It was proposed therefore that
ance spectroscopy could be a good cycle life predictor for a sealed
ion battery.
The significant difference between lithium metal anodes and graphite
in secondary lithium ion cells is that lithium metal anodes suffer from
dendrite formation, while under controlled cycling conditions,
te anodes do not. Lithium dendrite formation poses significant safety
ns, which ultimately results in a non-viable battery for medical
ations. In addition to lithium dendrite concerns, safety considerations
significant role in the design and testing of lithium-ion cells employed
dical applications2730
' Abuse testing for small lithium ion cells have
um Batteries for Medical Applications 697
120 12
11
100 10
u
...
.
E 80
!. 8
7
"..'
i
u
60
'5
"' :)_To
i Voltage 6 >
<
40
4
20 3
0 20 40 60 80 100 120
%of Charge
22.7. Temperature and voltage curves for a 1.5 Ah lithium ion cell under overcharge
onditions. The cell was charged at a constant current of 1.5 A (C-rate) at RT where
f charge equaled complete removal of lithium from the LixCo02 cathode material.
Small button-sized lithium ion cells have been developed for potential
external medical devices such as common hearing aids (HAD),
ermal electrical nerve stimulation devices (TENS), and devices for the
39
ermal application of other drugs (TAD). It has been proposed that
mall lithium ion secondary cells could mitigate the cost, inconvenience,
nvironmental hazard associated with the continual replacement and
al of small primary cells for external medical devices. The 0.45 g, 0.18
iniature cells successfully completed 2700 cycles at 60% DOD and
um Batteries for Medical Applications 699
e rates of 1-3 rnA. The miniature cells had average energy densities
Wh/1 and specific energies of 75 Whlkg, offering approximately 80%
nergy density and 70% of the specific energy of large, commercial
on batteries.
SUMMARY
ENCES
lmes, J. Power Sources 97-98 (2001) 739.
tbatch, The Making of the Pacemaker (Prometheus Books, Amherst, NY, 2000) p.
more information regarding TAH devices, see the Jewish Hospital, University of
isville, Health Sciences Center. Website: www.heartpioneers.com
. Leising, M.J. Palazzo, E.S. Takeuchi, K.J. Takeuchi, J. Electrochern. Soc. 148
1) A838.
asserini, B.B. Owens, J. Power Sources 97-98 (2001) 750.
Schmidt, P.M.Skarstad, J. Power Sources 97-98 (2001) 742.
51
Takeuchi, R.A. Leising, 201 Electrochemical Society Meeting, Philadelphia, PA,
12-17,2002, Abstract No. 260.
ter 23
D. MacArthur
INTRODUCTION
The focus of this chapter is Li-Ion batteries in consumer applications.
is a long history and a large market of non Li-Ion batteries--many of
rimary lithium batteries--but this other field is too extensive for
on here except to mention the recent apparent success to
rcialize, after a long gestation period, the lithium metal-polymer
lyte technology in the telecommunications backup application. The
should also note that Li-Ion batteries are now finding their way into
non-consumer applications. This, too, is outside the scope of the
chapter.
The subject begins with the lithium metal/molybdenum disulfide
introduced by Moli Energy in the late 1970, early 1980 period. Other
eable lithium systems were introduced for consumer applications
his same time but the Moli cell was the first to reach wide spread use.
ery well known, the product encountered a safety issue. Was it just
construction or was the lithium metal-liquid electrolyte system so e
as to be unacceptable for consumer applications? (And does the
polymer electrolyte solve this problem?) Whatever the technical
ation, the technology gained a perceived image as unsafe and there
need to move on. Sony researchers (and others) saw the opportunity
veloped a high capacity carbon based anode. (Research into lithium
lation in graphite has a much longer history but the capacities needed
ctical use had generally not been observed.) This anode, along with a
oltage metal oxide cathode was key to what came to be called the Li
ttery, first officially described in 1991. The name was intended to
ate the new system from the "unsafe" lithium battery. It is, of course,
imply a "rocking chair" type battery (or the physicist's battery as
Chapter 23
LITHIUM-ION BATTERIES
1. The Market
he growth of the Li-Ion market has been huge, a matter which refutes
mage of battery technology as slow-moving, out-of-date, and
onsive. During the period of 1991 to 1999 growth of not just the Li
arket but the so-called advanced battery market closely paralleled that
semiconductor industry, doubling and redoubling in periods of just a
ars or less. By year 2000, however, growth had greatly diminished
mostly because of the faltering world economy but perhaps also the of
reduction in number and significance of technical advances. The
rowth and following slowdown is shown in the following Table 23.1.
rent Issues and Market Trends of Li-Ion Batteries for Consumer Applications 703
(millions of cells)
illions of year 2000 rs)
ductiOn m 1992 was entirely Sony captive. Value m the table 1s based on estimated factory cost.
cted from the Powers Reports.
.2. Growth in world Li-Ion manufacturing capacity (millions of cells per year at
of the year).
olymer
a from the Powers Reports. Almost all Ll-Ion polymer are of the m-sltu polymenzed type of a
esign intended for wireless telephones.
REGION
Japan
China (including Taiwan)
Korea
North America and Europe
2. Different Technologies
Early constructions were 18650 (18 mm diameter, 650 mm long) jelly
lls which, over the years, have grown from about 1 Ah capacity in
to 2 Ah capacity in 2000. (The Molicell of 1998 which was "AA" size
r to the 18650 size had about 500 mAh capacity.) Much of the increase
acity is the result of higher density carbon materials for the anode and
packing. By about 1995 there was a shift to 17650 cells of better fit to
nner portable computers. About 300 million cells go to the wireless
one market and about 200 million to the portable computer market.
50% of Li-Ion production is now prismatic. In Table 23.4, the gain in
content over the years is reported.
By 1998, there was growing interest in "flat packs" with superior
ing flexibility for wireless telephones. The introduction of polymer
lytes follows a long history beginning even before the appearance of
batteries. Several technologies have been developed: the "dry" er,
the "solid state" in-situ cross-linked polymer, the gelled in-situ er,
the extracted-activated polymer of the Bellcore technology. Early
vision of the polymer technology was to make batteries like paper is
The expectation of lower cost has not been fulfilled and the critical
has become the form factor as the packages can be thin and shaped to
hin the contours of the telephone. By 2000 Li-Ion chemistry dominated
reless telephone and portable computer markets but with the saturation
ting markets, if Li-Ion production is to increase there is a need to find
pplications. Although the proliferation of portable electronic devices
ation assistants, communicators, tools, and toys-- is growing it does
pear to predict a doubling in battery demand. A new market is needed.
is interest, in smaller sizes such as credit card sizes but Li-Ion
logy is not well suited and many uses are served by primary
logy. Much of the focus is on markets for larger sizes. This includes
backup, power tools and electric vehicles. The prospect of a huge
vehicle market is very enticing. The challenges are cost and safety.
teristics of small and large batteries are reported in Table 23.5
ent issues and Market Trends of Li-lon Batteries for Consumer Applications 705
Density (Whll)
ic Energy (Whlkg)
numbers are for commercial cells. The change m specific energy, year 2000, 1s related to
n of the plastic envelope package. The change in energy density over the 1992-98 period was f
changes in capacity of the anode and in packaging.
TTRIBUTE
E(Whlkg)
D (Whll)
D (W/1) peak/sustained
ife
emperature Range ( C)0
osts ($/Wh)
afety/Robust
ecycle
Chapter 23
ERENCES
mamura, T. Inoue, H. Yoshida, M. Mizutani, M. Goto, Power Sources Conference,
ry Hill, NJ, June 10-13, 2002, Paper 26-5.
Dowall, A. Brenier, M. Broussely, P. Lavaur, INTELEC 2002 Conference, Montreal
ec, Sept. 30-0ct 3, 2002, Paper 22.2.
ject Index
LiFeP207 cathode 471
ves for lithium battery 303, 174
materials LiMX04 445
Aluminides Li3Mz(X04h cathodes 457
Antimonides 116 121, 275 LiNi02 cathode 19
Cu-Sn 123 LiVO.33, 345
Graphite anodes 144 Polyanion cathodes 32
Intermetallic anodes 12, 125, 270 Role of bond covalency 64
Lithiated carbons 146 Spinel oxide cathode 25, 30, 362
Metallic lithium anode 297 Spinel manganese oxide 26
Metal oxide anodes 129 Sulfides cathode 350
Mg2Si 127 Synthesis of cathode materials
Nitrides 132, 247 86,317
LiMNx248 Titanium disulfide 5, 90
InN 254 V6013 34
Sn3N4 254 Composite anodes 112
G N4254 Covalent bonding in Li.Co02
Cu3N 257 Crystal field stabilization energy 21
Phospides 134 Electronic properties of LiNi02 , 410
Silicides 127, 247, 259
Li.Si 260 Decrepitation of alloy anodes 290
Li-Si-Mg 261
CrSi2 265 Electrode fabrication 200
NiSi 266 Electron diffraction in crystals 481
Silicon based composite 127 Electrolytes related
Tin anode 116 Additives 174, 303
Tin oxide anode 117 Alkyl carbonates 509, 531
Transition metal oxide 130 Binary electrolytes 522
synthesis 102 Cathodic reactions 545
Ceramic electrolytes 623
Dipole-dipole interactions 513
y buffer operation 673
Electrochemical stability 525
y cost 662
Electron-pair donor (EPD) 516
y design 657
y market 702 Electron-pair acceptor (EPA)
y medical applications 686 Glass electrolytes 625
y safety 661 Gel electrolytes 591
y satellite applications 675 Hydrogen bonding 515
y stationary applications 671 516
y standby operation 671 Ion-dipole forces 511
y space applications 674 Ion transport 575
Liquid electrolytes 509
Mixed electrolyte 509
naceous anodes 114
de materials, 315 Molten salts 567
Cr oxide cathode 345 Physical properties 532
LiNi 1.•Co.02 334, 410 Polar solvents 530
First principle calculations 42 Polyelectrolytes 602
General electronic structure 45 Polymer electrolytes 575
Layered oxide cathodes 11, 24 Selective solvation 518
LiCo02 12, 410, 414 Solvation 517
LiCo.Nit-xOz 333,410 Stability 604
LiFeP04 99, 347, 445 Ternary solvent blends 524
Subject Index
lasses 626
ide Glassy electrolytes 626
ide glassy electrolytes 628
·- 5,00