Ion Conducting Glasses for Use in Batteries 2:
Glassy Solid Electrolytes
Steve W. Martin
Department of Materials Science & Engineering
Iowa State University of Science & Technology
[email protected]
Glass in Energy
Spring 2012
Anode and Cathode Combinations Determine the Voltage
and Energy Density of Lithium Batteries
Cathodes
Anodes
Where we are…
Where we want to go…
J.M. Tarascon, M. Armand, Nature, 414, 15 (2001) 359
[email protected]
Glass in Energy
Spring 2012
2
Anode and Cathode Combinations Determine the Voltage
and Energy Density of Lithium Batteries
Cathodes
Anodes
Where we are…
Where we want to go…
J.M. Tarascon, M. Armand, Nature, 414, 15 (2001) 359
[email protected]
Glass in Energy
Spring 2012
3
Just for comparison…
Where we are…
Where we can go…
10 fold increase
in energy density!
[email protected]
Glass in Energy
Spring 2012
4
Anode and Cathode Combinations Determine the Voltage
and Energy Density of Lithium Batteries
Cathodes
Anodes
Where we are…
Where we want to go…
J.M. Tarascon, M. Armand, Nature, 414, 15 (2001) 359
[email protected]
Glass in Energy
Spring 2012
5
So… Why Li-Air Batteries?
•
•
•
•
•
•
•
•
•
Because today’s Li batteries use Li1C6 as the Anode
And the Cathode, Li1CoO2, is equally less energy dense
Comparison….
Gasoline: C8H18 + 12 ½ O2  8CO2 + 9H2O + Energy
44,400 J/g C8H18
Li: ~4,000 mAhr/g x 3,600 sec/hr x ~4 V = 57,600 J/g Li
Gasoline and Li are comparably energy dense
Compare…
Graphite: ~ 400 mAhr/g x 3,600 sec/hr x ~ 4 V = 5,760 J/g
[email protected]
Glass in Energy
Spring 2012
6
Safety….Lithium Dendrites in Li ion Batteries
Li metal
Non-epitaxial deposition of lithium after each cycle leads to the
growth of uneven “fingers” or dendrites of lithium
Internal dendrites can cause short circuits of the battery
M. Dolle et al. Electrochemical and Solid-State Letters, 5(12) (2002)A286
[email protected]
Glass in Energy
Spring 2012
7
Applications of Ion Conducting Glasses
Solid electrolytes in Lithium Batteries
Solid state Li-cells
–
–
–
–
–

Anode
No liquids
Lightweight
High voltage
Wide T
Thin film form
Li+
Li+
Solid state electrolyte
High d.c.
– >10-3(cm)-1
– Stability to Li
– Wide voltage window
–
[email protected]
Cathode
FIC Glass
Electrolyte

eLoad
Glass in Energy
e-
Spring 2012
8
Glassy Solid Electrolytes
• Using highly conducting glasses in solid state batteries
– Increases safety?
• By mitigating lithium dendrite formation
– Increases energy density?
• By enabling lithium metal (or similar high activity) anodes
– Reduces cost?
• By simplifying design and using lower cost materials
[email protected]
Glass in Energy
Spring 2012
9
Fast Ion (Li+) Conducting Sulfide Glasses
• Typical glass compositions
– Lithium salt +
Lithium modifier
Mobile cations
+ glass former + additives
Glass structure
Chemical/mechanical/
electrochemical durability
– LiI + Li2S + SiS2, B2S3, GeS2 …
– LiI + Li2S + GeS2+ Ga2S3, La2S3, ZrS2…
– LiI + Li2S + GeS2 + GeO2 +….
[email protected]
Glass in Energy
10
Spring 2012
10
Ionic Conduction in Glass
xNa2S + (1-x)SiS2
Glass in 2-D
+
+
|E|
+
+
MD Simulations
+
+
y
+
+
-
Energy
NBS
BS
+1/rn
-
NBS
BS
Eact s Ec
Es = Strain Energy
Ec = Coulomb Energy
ES
EC
r
[email protected]
+
+
x
-e2/r
r
Glass in Energy
Spring 2012
11
Ionic Conduction in Glass
•
o
200
Temperature ( C)
0
600
1600
1400
1200
1000
800
– N is the number density
– eZ is the charge, +1 most of the
time
–  is the mobility
Now what are common values of?
– n ~ # M+/cm3 ?
400
•  = neZ
0
10
- N a
Glassy
Crystalline
-AgI
Tg
Tm
Al O
11
17
-1
-1
30
SO
Li 2
-1
4
.8 A
gO
-4 0
Li
1 0 Ge
PS
2
.4 (
12
Ag
l) 2-
12
9(
4.
2
S
3
2.5
3
O3
B2
O3
1B 2
-1
iO 2
-6
l) 2
B 2O
-5
.5S
Li C
-75
S iO 2
bO 3
O
Li 2
LiN
I)
28.6
Ag
2 - 29
Ag
.8 P
2 O-4
2 S
2
-5
.8(A
40
2O
Ge
gI)
Li
5
I+
S2 28.
6Mo
45
2
O
7L
Ge
3
i
S
2 S
2
+
27
G
eS
2 +
6G
a
LiC
OLi 2
25
- 50
O3
Al 2
.9
0(
26
-2 5
dc (-cm)
50
OLi 2
iO 4
Li 2O
-4
10
35
lS
25
Y 2O 3
-3
10
29
11
2
Li A
10
S
3
- 9%
Z rO 2
-2
Li P
7
Cl
O 12
B7
Li 4
10
RbAg4I5
10
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
-1
1000 / T (K )
[email protected]
Glass in Energy
Spring 2012
12
Ionic Conduction in Glass
10
200
0
400
600
o
Temperature ( C)
0
Glassy
Crystalline
-AgI
-Na
Tg
Tm
Al O
11
17
50
OLi 2
Ag
2
30
iO 4
-1
29
11
35
lS
-1
0(
LiC
l) 2-
OLi 2
12
9(
gO
4.
2.5
12
gI)
2 -29
Ag
.8P
2O
40 -42.8(A
2O
gI)
Li
5
eS
I+
2 -2 8 .
-4
6M o
27
5G
O
Li
3
eS
2 S
2
+
27
G
eS
2 +
6G
a
2
S
3
O3
B2
2.0
PS
2
.4(
A
2 .5
O3
1B 2
3
1.5
Li
10 Ge
-4 0
-1
iO 2
-6
l) 2
B 2O
1.0
.5 S
Li C
SiO 2
-75
-5
0.5
5G
4
.9
-50
O3
Al 2
10
.8A
28.6
S-
SO
Li 2
26
25
bO 3
O
Li 2
L iN
-4
25
10
OLi 2
-3
S
3
2
25
Y 2O 3
10
7
LiA
-2
RbAg4I5
Li P
-9 %
Zr O 2
10
-1
Cl
O 12
B7
Li 4
10
dc (-cm)
•
– N is the number density
– eZ is the charge, +1 most of
the time
–  is the mobility
Now what are common values of?
– n ~ 1022 M+/cm3
–  ~ # (cm)-1 ?
1600
1400
1200
1000
800
•  = neZ
3.0
3.5
4.0
4.5
-1
1000 / T (K )
[email protected]
Glass in Energy
Spring 2012
13
Ionic Conduction in Glass
•  = neZ
•
– N is the number density
– eZ is the charge, +1 most of
the time
–  is the mobility
Common values
– n ~ 1022 M+/cm3
–  ~ 10-6 (cm)-1
• (T) = oexp(-Ea/RT)
• The d.c. conductivity is
“Arrhenius”
• What does this give for
–  ~ # cm2/V-sec
[email protected]
Glass in Energy
Spring 2012
14
Ionic Conduction in Glass
•  = neZ
•

– N is the number density
– eZ is the charge, +1 most of
the time
–  is the mobility
Take
– n ~ 1022 M+/cm3
–  ~ 10-3 (cm)-1
–  ~ 10-6 cm2/V-sec
Compare Si
•  ~ 10+3 cm2/V-sec
• e- are 109 times more mobile
[email protected]
• (T) = oexp(-Ea/RT)
• The d.c. conductivity is
“Arrhenius”
Glass in Energy
Spring 2012
15
log10(a.c.)
AC versus DC ionic conductivity

+
Energy
0 2 4 6 8 10
103K/T
+  > 1
2 4 6 8 10 12
log10(f/Hz)
y
x
+
r
D.C. Conductivity
A.C. Conductivity
Charles - Polarization/Diffusion
Anderson/Stuart - Coulomb & Strain Energies
Moynihan/Macedo - Debeye &
Faulkenhagen Theory
Ravaine/Souquet - Weak Electrolyte
Malugani- AgI Micro domains
Ingram - Cluster Pathways
Elliott - Local Structure/Diffusion
Controlled Relaxation
Jonscher - Universal Response
Ngai - Coupling Theory
Moynihan - Modulus
Dyre - Power Law
Funke - Jump Relaxation
Dieterich/Bunde Coulomb Interaction
Elliott - Diffusion-Pathways
[email protected]
Glass in Energy
Spring 2012
16
10
o
200
Temperature ( C)
0
400
600
1600
1400
1200
1000
800
Arrhenius Ionic Conductivity…
0
-Na
Glassy
Crystalline
-AgI
Tg
Tm
Al O
11
17
SO
Li 2
-1
4
0(
-1
PS
2
.4(
A
12
4.
S
3
O3
B2
2.5
2
2.5
O3
1B 2
-1
iO 2
-6
l) 2
.5 S
L iC
2.0
- 40
Ge
12
9(
1.5
10
l) 2-
OLi 2
LiC
.9
bO 3
1.0
Li
gI)
2 - 29
Ag
.8P
2 O- 4
2 S
2. 8(
-5
4
2O
0L
A gI
Ge
5
) -2
iI
S2
8
+
.6M
45
2
oO
7L
Ge
3
i2 S
S
2
+
27
G
eS
2 +
6G
a
-5
0.5
.8 A
gO
28.6
Ag
30
O4
dc (-cm)
50
OLi 2
i
lS
-2 5
25
L iN
10
29
11
35
LiA
-4
26
S iO 2
3
-5 0
B 2O
O3
-7 5
Al 2
O
Li 2
10
Li 2O
-3
S
3
2
25
10
Y 2O 3
-2
7
-9 %
ZrO 2
10
RbAg4I5
Li P
Cl
O 12
B7
Li 4
10
-1
3.0
3.5
4.0
4.5
-1
1000 / T (K )
[email protected]
Glass in Energy
Spring 2012
Ion Conduction in Glass
-1
10
-3
10
-5
10
-7
10
-9
dc (-cm)
0
200
400
o
Temperature ( C)
Glassy
Tg
Lithium battery operation
35L
i2 O
-30
Li S
2 O
-1
10
600
1
10
1600
1400
1200
1000
800
Ion motion (conduction) is typically very limited in oxide glasses
4
25
Li
-10
(LiC
l) -1
2
2.5
SiO
O25
Al
25
2O
Li
3 -50
O
2
SiO
-7
5B
2
2 O
2
-11
10
2
-12
.5B
2O
3
3
-13
10
-15
10
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
-1
1000 / T (K )
[email protected]
Glass in Energy
Spring 2012
18
10
o
200
Temperature ( C)
0
400
600
1600
1400
1200
1000
800
Arrhenius Ionic Conductivity…
0
 -N a
Glassy
Crystalline
-AgI
Tg
Tm
Al O
11
17
SO
Li 2
-1
4
0(
-1
PS
2
.4(
A
12
4.
S
3
O3
B2
2.5
2
2.5
O3
1B 2
-1
iO 2
-6
l) 2
.5 S
L iC
2.0
-40
Ge
12
9(
1.5
10
l) 2-
OLi 2
LiC
.9
bO 3
1.0
Li
gI)
2 -29
Ag
.8P
2 O-4
2 S
2. 8(
-5
4
2O
0L
A gI
Ge
5
) -2
iI
S2
8
+
.6M
45
2
oO
7L
Ge
3
i2 S
S
2
+
27
G
eS
2 +
6G
a
-5
0.5
.8 A
gO
28.6
Ag
30
O4
dc (-cm)
50
OLi 2
i
lS
-2 5
25
L iN
10
29
11
35
LiA
-4
26
S iO 2
3
-5 0
B 2O
O3
-7 5
Al 2
O
Li 2
10
Li 2O
-3
S
3
2
25
10
Y 2O 3
-2
7
-9 %
ZrO 2
10
RbAg4I5
Li P
Cl
O 12
B7
Li 4
10
-1
3.0
3.5
4.0
4.5
-1
1000 / T (K )
[email protected]
Glass in Energy
Spring 2012
Examples of Ion Conduction in Glass
Salt doped lithium phosphate and thiophosphate glasses
dc
[email protected]
act
Glass in Energy
Spring 2012
20
AC ionic conductivity in glass
•
Connection to Far-IR vibrational modes, Angell ‘83
[email protected]
Glass in Energy
Spring 2012
21
DC ion conductivity in glass
•
•
•
•
•
Arrhenius temperature
dependence
xLi2O + (1-x)P2O5
Creation of non-bridging
oxygens
“Mobile” lithium ions
The higher the
concentration of Li2O, the
higher the conductivity
– Lower resistivity
•
Activation energy
decreases with Li2O
content
S. Martin, C.A. Angell JNCS ’83
22
[email protected]
Glass in Energy
Spring 2012
Mobility and Number Dependence of the Conductivity
0
  Eact 
 (T )  n(T )eZ c  (T )  exp

T
 RT 
  E c 
n(T )  no exp

 RT 
 (T ) 
 (T ) 
0
  Es 
exp

T
 RT 
Z c en0  0
  Ec  Es  
exp

T
RT


Question: What are the magnitudes of ES(M) and EC ?
[email protected]
Glass in Energy
Spring 2012
23
Ion Conduction in Glass: Coulombically or Structurally
Constrained?
•
•
•
•
Oxide glasses, Eact ~100 kcal/mole
Sulfide glasses, Eact ~10 kcal/mole
Eact Es Ec
Are alkali cations coulombically, Ec , constrained?
– Weak Electrolytes like HOAc, kA ~ 1 x 10-5 ?
– Cations are only weakly dissociated
• Are alkali cations structurally, Es, constrained?
– Strong electrolytes like NaCl?
– Completely dissociated, Na+ Cl- ? K ~ 106
[email protected]
Glass in Energy
24
Spring 2012
Models of the Activation Energy
•
•
Both activation energies appear to be non-zero and contribute to the total
activation energy
Anderson-Stuart1 model calculation
C struct . Z c Z a e 2  1
2
EC 
 


 ( rc  ra )  
•
•
 E S  G ( rc  rd ) 2  / 2
x Na2O + (1-x)SiO2 Es (calc)
kcal/mole
Ec (calc)
kcal/mole
Eact(calc)
kcal/mole
Eact2
kcal/mole
11.8
11.7
66.9
78.6
68.1
19.2
10.9
62.3
73.2
63.7
29.7
10.0
56.1
66.1
59.7
Calculation shows that the Ec term is the larger of the two energy barriers.
Coulombically constrained?
1
Anderson, Stuart, J. Amer. Cer. Soc., 1954
SciGlass 5.5, Average of many glasses
2
[email protected]
Glass in Energy
Spring 2012
25
Alkali Radii Dependence of Strain and Coulomb
Activation Energies
Es
Ec ~ 1/rc
Es ~ rc
2
0.0
+
H (?)
[email protected]
Es dominated
Ec dominated
E s , E c (A.U.)
Ec
0.5
Li
1.0
+
Na
1.5
+
K
2.0
+
Glass in Energy
rcation ()
2.5
3.0
Cs
+
Spring 2012
26
Short Range Order models
•
•
•
•
Anderson-Stuart Model
Assignment of Coulombic and Strain energy terms, EC + Es
“Creation” or Concentration versus Migration energy terms, EC + Em
Coulomb energy term, EC attractive force between cation and anion
Cstruct.   Zc Zae2  Zc Zae2  Cstruct.Zc Za e2  1
2


 



   / 2

(rc  ra ) 
 (rc  ra )  
Cstruct.Zc Za e2
 const.
Lim Eact 
 
  (rc  ra )
[email protected]
Glass in Energy
Spring 2012
27
Short Range Order models
 Es(kcal/mole)
• Strain energy term - Es
• “Work” required to “dilate the network so large cations can
migrate
E S   G ( rc  rd ) 2  / 2
G
Shear modulus
rc
Cation radius
Cation size affect on Strain Energy
rd
Interstitial site radius
50

Jump distance
40
30
20
10
0
0
0.04
0.08
0.12
0.16
0.2
Cation Radius (nm)
[email protected]
Glass in Energy
Spring 2012
28
Thermodynamic Models
•
•
•
•
•
•
Glass is considered as a solvent into which salt is dissolved
If dissolved salt dissociates strongly, then glass is considered a strong
electrolyte
If dissolved salt dissociate weakly, then glass is considered a weak
electrolyte
Coulomb energy term calculations suggest that the salts are only weakly
dissociated, largest of the two energy terms
Migration energy term is taken to be minor and weak function of
composition
Dissociation constant then determines the number of mobile cations
available for conduction, dissociation limited conduction
[email protected]
Glass in Energy
Spring 2012
29
Weak Electrolyte model, Ravaine & Souquet
M2O
 
Undissociated)
(Dissociated)
Kdiss = aM+ aOM- / aM2O
~ [M+][OM-]/aM2O = [M+]2/ aM2O
[M+]
~ Kdiss1/2aM2O1/2 = n

= zen zeKdiss1/2aM2O1/2 ~ C aM2O1/2
log Kdiss ~ -Ne2RT/4 r+ + r-)
As r+, r- increase, Kdiss increases
As  increases, Kdiss increases
[email protected]
Glass in Energy
Spring 2012
30
Strong and Weak Electrolyte models
•
•
•
•
“Strong electrolyte” model suggests all
cations are equally available for
conduction.
– Each cation experiences an energy
barrier which governs the rate at
which it hops
“Weak electrolyte” model suggests only
those dissociated cations are available for
conduction
– Dissociation creates mobile carriers
available for conduction
SE models suggests that EC + Es both
contribute, one could be larger or smaller
than the other
WE model suggests that Ec is the
dominant term
31
[email protected]
Glass in Energy
Spring 2012
Intermediate Range Order models
• Models recognize that ion conductivity requires ion motion
over relatively long length scales
• Ions must be able to move from one side of the electrolyte to
the other
• Long range connectivity of the SRO structures favorable to
conduction must exist
• Deep “traps” along the way must be infrequent and not severe
• Rather, low energy conduction “pathways” are thought to exist
which maximize connectivity and minimize energy barriers
and traps
• Cluster pathway model of Greeves ‘85, for example
[email protected]
Glass in Energy
Spring 2012
32
Intermediate Range Order models
•
Cluster pathway model, Greeves et
al ‘85
33
[email protected]
Glass in Energy
Spring 2012
Intermediate Range Order models
• Proposed Percolation Models - Johari et al. ‘87
– At low dopant concentrations
• Cations are far separated
• Mobile species are diluted in a non-conducting host glass
– At intermediate concentrations
• Cations begin to approach proximity
• Preferential conduction paths form
• Sites percolate
– At high concentrations
• Cations are fully connected
• Conduction pathways are fully developed
– Percolation Behavior is NOT observed
[email protected]
Glass in Energy
Spring 2012
34
Failure of Conductivity percolation
[email protected]
Glass in Energy
Spring 2012
35
Intermediate Range Order Models
• Microdomain models of conductivity
• Dopant salts such as AgI to oxide glasses, especially AgPO3,
are added to increase conductivity
• AgI is itself a FIC crystal above 150oC
• Extrapolations of  to xAgI = 1 give ~ AgI(298K)
• The question then is: Does the AgI create “microdomains” of AgI giving rise to the high conductivity?
[email protected]
Glass in Energy
Spring 2012
36
AgI Microdomain model
• Most well known of all glasses is xAgI + (1-x)AgPO3
• AgPO3 is a long chain structure of -O-P(O)(OAg)-O repeat
units
• Intermediate range structure is for these long chains to
intertwine and as such frustrate crystallization
• Added AgI dissolves into this liquid without disrupting the
structure of the phosphate chains
• Microdomain model then suggests that this dissolved AgI
creates increasingly large clusters of -AgI between the
phosphate chains
[email protected]
Glass in Energy
Spring 2012
37
AgI Microdomain model
[email protected]
Glass in Energy
Spring 2012
38
Composition Dependence of the Conductivity
•
•
•
Binary lithium phosphate glasses,
Li2O + P2O5, are relative poor ion
conductors
Binary lithium borate glasses, Li2O
+ B2O3, are slightly better
conductors
Binary lithium silicate glasses,
Li2O + SiO2 are slightly better
conductors yet.
Li2O+ SiO2
Li2O+ B2O3
Li2O+ P2O5
Li2O:SiO2
Li2O:P2O5
Li2O:B2O3
39
[email protected]
Glass in Energy
Spring 2012
Salt doped phosphate glasses
•
Halide doping strongly increases
the conductivity
40
[email protected]
Glass in Energy
Spring 2012
Mixed GlassFormer Cation Glasses
•
•
•
•
•
•
•
•
Na2O + B2O3 + P2O5 Glasses
Non-Additive (+’ive) behavior in
physical properties
Increased ionic conductivity
Increased Tg
Improved glass forming ability
Improved mechanical properties
Improved chemical properties
Overall improvement in nearly all
properties required for Lithium battery
applications
Zielniok, Cramer, Eckert . Chem. Mater. 19 (2007) 3162
[email protected]
Glass in Energy
Spring 2012
41
Mixed Glass Former Na B P O Glasses
A Maximum in the Ionic Conductivity…
0.35Na2O + 0.65[(1-x)B2O3 + xP2O5]
-5
d.c./ (cm)
-1
10
Na2O
B2O3
P2O5
443 K
423 K
80
403 K
75
383 K
-7
10
363 K
70
65
-8
10
0.0
0.2
0.35Na2O + 0.65B2O3
[email protected]
85
-6
10
90
Glass in Energy
0.4
0.6
x P2O5
60
0.8
1.0
0.35Na2O + 0.65P2O5
Spring 2012
42
Mixed Glass Former Na B Si O Glasses
A Minimum in the Ionic Conductivity…
10
-6
0.20Na2O + 0.8[(1-x)BO1.5 + xSiO2]
75
Na2O
BO1.5
10
-7
73
T = 423 K
10
72
383
71
363
70
-8
SiO2
10
403
Eact(kJ/mol)
log(d.c./(cm)
-1
74
-9
0.0
0.2
0.4
0.20Na2O + 0.8BO1.5
0.6
x SiO2
0.8
1.0
0.20Na2O + 0.8SiO2
R. Christensen et al., to be published
[email protected]
Glass in Energy
Spring 2012
43
Mixed Glass Former Effect
•
•
Phosphate additions cause a positive MGFE in Borate glasses
Silica additions cause a negative MGFE in Borate glasses
Na B P O
Na B Si O
d.c./d.c(linear)
10
1
0.0 0.2 0.4
0.2Na2O + 0.8BO1.5
[email protected]
0.6 0.8 1.0
x 0.2Na2O + 0.8PO2.5
0.2Na2O + 0.8SiO2
Glass in Energy
Spring 2012
44
Mixed Glass Former Cation Glasses
•
Sulfide glasses
A Pradel, N Kuwata and M Ribes J. Phys.: Condens. Matter 15 (2003) S1561–S1571
[email protected]
Glass in Energy
Spring 2012
45
Mixed GlassFormer Cation Glasses
•
Sulfide glasses
Annie Pradel, et al, Chem. Mater. 1998, 10, 2162-2166
[email protected]
Glass in Energy
Spring 2012
46
Mixed Glass Former Glasses
•
AgI-Ag2MoO4-Ag3PO4 Glasses
Nobuya Machida, et al. Solid State Ionics 79 (1995) 273-278
[email protected]
Glass in Energy
Spring 2012
47
Salt doped phosphate glasses
•
LiI doped LiPO3 show highest conductivity and lowest activation
energy among the halides
T = 298 K
48
[email protected]
Glass in Energy
Spring 2012
Mixed glassformer systems
•
Phosphate and borate mixed
glasses show non-linear “Mixed
Glassformer” effect
49
[email protected]
Glass in Energy
Spring 2012
Lithium Thiophosphate Glasses
[email protected]
Glass in Energy
Spring 2012
50
Silver Phosphate Glasses
[email protected]
Glass in Energy
Spring 2012
51
Other Silver Sulfide Silver Halide doped glasses
[email protected]
Glass in Energy
Spring 2012
52
Summary and Conclusions
•
•
•
•
•
•
•
•
Fast Li+ ion conducting glasses can be compositionally tailored
Over wide ranges of composition and Li+ ion conductivity
Advantages of solid structure reduce lithium dendrites
Improves safety of the batteries
Simplifies battery construction
May enable the use of metallic lithium anodes
Dramatically and safely increasing the energy density of the batteries
Future work will focus on cycle life time
[email protected]
Glass in Energy
Spring 2012
53