Energy Track: Material Needs in
Alternative & Renewable Energy for
the Automotive Industry
Lithium ion traction
batteries and the role of
ceramics
Mark Verbrugge and Mike Balogh
General Motors Research & Development
Ceramic Leadership Summit. August 1-3, 2011, Baltimore, MD
Outline
 Automotive context
 Short overview/primer on lithium
ion batteries
 Ceramics…now and in the future
of lithium ion battery technology
 Summary
Battery Technology Trajectory
Range
Energy
(Watt-hr / kg)
Maximum stored
energy per unit
of battery mass
Power (Watt / kg)
Acceleration
Maximum power per unit of battery mass
Future vehicles will use alternative energy sources
like bio-fuel, grid electricity, and hydrogen
COST!
Improved
Vehicle Fuel
Economy &
Emissions
Hydrogen
Fuel Cell
Displace
Petroleum
Battery Electric
Vehicles (E-Rev)
IC Engine and
Transmission
Improvements
Hybrid Electric
Vehicles (including
Plug-In HEV)
Time
Petroleum (Conventional & Alternative Sources)
Alternative Fuels (Ethanol, Bio-diesel, CNG, LPG)
Electricity (Conventional & Alternative Sources)
Hydrogen
Energy
Diversity
Really BIG questions
¶
Liquid fuels
–
Future price & availability of oil?
–
Efficacy of bio-derived fuels?
¶
What is the relative importance of zero on-vehicle
“regulated emissions” vs. fuel cost, CO2 emissions ,
& energy security?
¶
Fuel cell vision offers
1. Range
2. Short re-charge times
3. Zero emissions
¶
Another vision: EREV with bio-derived fuels
–
City: EV (~40 miles)…zero emissions
–
Between cities
 Liquid fuel: high Wh/kg
 Regulated emissions from ICE range
extender, but greatly reduced today and
low for highway driving
 Energy security, affordability, and reduced
unwanted emissions (including CO2)
Cost of gasoline to personaltransportation customers, $US/gal
4. Technical efficacy now
6
5
4
3
2
1
50
75
100
125
150
Cost of (crude) oil, $US/barrel
175
200
ENERGY CARRIER PROPERTIES: ONBOARD STORAGE
WHY IS PETROLEUM THE DOMINANT TRANSPORTATION FUEL?
Weight & Volume of Energy Storage System for 500-km Range
EREV: SMALL ZEV BATTERY + LIQUID FUEL FOR RANGE EXTENSION
GM Vehicle Electrification Strategy
Portfolio of solutions for full range of vehicles that provide customer choice
Petroleum and Biofuels (Conventional and Alternative Sources)
Electricity – ZEV Fuel
GM
Hybrid
2-Mode
2-Mode
PHEV
Voltec
EREV
Electrification
Battery
Electric
Fuel Cell
Typical Commute
Why Target 40 Miles?
40 Miles Is the Key
78%of customers
commute 40 miles
or less daily
Based on U.S. Department of Transportation 2003 Omnibus Household Survey
Electric Vehicle with RANGE-EXTENDER
Up to
40 miles
BATTERY
Electric Drive
HUNDREDS of miles
EXTENDED-RANGE
Driving
Variations on Electric Vehicles
Chevrolet Volt: The Electric Vehicle with Extended Range
http://www.youtube.com/watch?v=JZYN3TK3Fmo
Chevrolet Volt Wins 2011 Motor Trend Car of the Year!.flv
PHEV
Plug-in Hybrid
Electric Vehicle
(3 minute Motor Trend Video)
EV
with Extended Range
Electric Vehicle with
“Extended-Range”
Pure EV
Pure Electric Vehicle
• All-electric at low
speed/power
• All-electric for up to
40 miles
• All-electric for ~100
miles
• Blended electric/gas at
higher speed/power
• Gas generator for +300
miles extended driving
range
• Fuel is electricity
• Primary fuel is gasoline
supplemented with
electricity
(typical)
• Primary fuel is electricity
supplemented with
gasoline
(Volt)
(typical)
Electricity as Low-Cost Fuel (US costs)
7-13¢ per mile
1-2¢ per mile
Battery Technology Improvements
and the EREV Challenge
Range
Energy
(Watt-hr / kg)
Maximum stored
energy per unit
of battery mass
EV
PHEV
Power (Watt / kg)
EREV: High
specific power
(kW/kg) and
energy (kWh/kg)
EVre
(Volt: 8-year, 100k mile)
Acceleration
Maximum power per unit of battery mass
Charge
Port
Lithium-Ion
Battery
Electric
Drive Unit
Engine
Generator
North American Car of the Year for 2011
Motor Trend 2011 Car of the Year
Green Car Journal 2011 Green Car of the Year
Car and Driver 10 Best for 2011
Ward’s AutoWorld 10 Best Engines for 2011
AUTOMOBILE Magazine 2011 Automobile of the Year
2010 Breakthrough Technology, by Popular Mechanics
THE ALL NEW 2011 CRUZE
Introducing the Cruze Eco and amazing
highway fuel economy at 42 MPG that
doesn't sacrifice the sculpted exterior design
that sets Cruze apart from the competition.
Lithium ion battery challenges
 Cost
• Can we size pack closer to end-of-life requirements?
• Can we reduce materials & processes costs?
 Life
• How do electrodes fail?
• Can we develop an accelerated life test?
 Temperature tolerance
• Can we improve low temperature power?
• Why is battery life shorter at higher temperatures?
Charge mechanism (reverse of discharge)
V
e-
eLi+
e-
Li+
eLi+
PF6
Li+
Li+
e-
Li+
e-
Li+
Li+
Li+
Li+
PF6-
(+) Metal oxide,
phosphate, or silicate
Li+
-
PF6-
PF6-
Li+
PF6-
PF6-
Li+
PF6-
Li+
PF6-
Separator (Solvent + Salt),
often ceramic enhanced
Positive is full of lithium at the
end of discharge
Li+
PF6-
(-) Carbon,
titanate, etc.
By putting energy into the cell for charging,
lithium is forced out of the positive and into
the negative.
Electrode microstructure
V
•
e-
-
e
Li+
-
Li+
eLi+ Li+
e
Li+
e- eLi+ Li+
Li+
PF6-
Li+
Li+
Li+
PF 6-
PF 6-
Li+
PF 6-
PF6-
Li+
PF 6-
PF6-
PF6-
Li+
Li+
~25 mm
Negative charge reaction
1. At the particle surfaces:
Li+ + e-  Li0
~3 Å
2. Lithium diffuses rapidly into
host particle
•
Opposite reactions takes
place at positive particle
surfaces
•
Li plating must be
avoided
PF 6-
Porous electrodes (~100
mm thick) composed of
host particles (~1 to 5 mm
diameter) are used to
1. increase the surface
area for reaction
2. reduce lithium
diffusion resistance
~100 mm
Li
Li0
Li+
+
e-

Li0
~5 mm
Positive electrode materials
 Often ceramics
• LiMO2 (with M = Ni, Co, Mn, Al … or combinations thereof) is
the most used positive material (includes LiCoO2, NCM,
LNCA)
• LiMn2O4 (spinel) is low cost and provides high power density
along with good abuse tolerance
• LiMPO4 (with M = Fe, Mn, Mg, … or combinations thereof)
• Li2MnO3-Li(NixCoyMnz)O2 (with x + y + z = 1) is of strong
interest currently
• LiMSiO4 (with M = Fe, Mn, … or combinations thereof) is
showing promise as a low cost, high capacity positive
 The positive electrode material is a major cost driver in LiIon
batteries
 The potential for solvent oxidation at the positive electrode
leads to abuse tolerance concerns
Ceramic coatings to
suppress unwanted
side reactions
Particle cracking…back to the ceramics literature
Durability…terminologies, bathtub curves
 Chemical degradation
• Critical role of SEI (solid electrolyte interface) to impede deleterious
degradation reactions within lithium ion cells
• Calendar life determined by chemical degradation
 Mechanical degradation
• Cyclic expansion and contraction of insertion or alloy materials leads to
fatigue, cracking, and structural changes
• Cyclic life issues are affected by mechanical degradation and chemical
degradation
Number of
Failures
Early (infant
mortality)
failures
Wear out
failures
Useful life
(durability)
Time in service
Electroanalytical cell for characterization of graphite negative
 Substantially uniform current
distribution and cell pressure
• Li reference in the counterelectrode plane
• Electrochemical reaction on
the surfaces of the graphite
particles (WE):

-
Li  e  S
kc
ka
[Li - S- ]
• At the Li electrodes (CE, RE),
Li  e Li
Three Electrode Cell
Negative electrode…the solid electrolyte interface (SEI)
• Solvent reduction at ~0.8V vs Li
on first cycle
• Then ~100% Coulombic efficiency

B  k a kc
1- 
e
(1-  ) f (V -U  )
- e - f (V -U

)
 DRc
I
Formation of the SEI…solvent reduction on the negative
(ethylene carbonate)
CH2
H2C
2Li+ + 2e- +
Li2CO3
+ H2C=CH2
Inorganic
Layer (1st)
Gassing
(ethylene)
O
O
C
[Li(OCOO)CH2]2
Organic layer
Li+ + 2e- = Li
O
Vcell ~ mLi ~ ln(SOC)
(Calendar life influence)
+ H2C=CH2
Gassing
(ethylene)
LiCH2CH2(OCOO)Li
Organic layer
 Example reactions only…many others contribute to the formation of
the solid electrolyte layer
• For computed IR spectra of surface species in an EC electrolyte, see
S. Matsuta, T. Asada, and K. Kitaura. J. Electrochem. Soc.
147(2000)1695-1702…dimers found to be lowest energy
• Experimental FTIR data indicates predominance of [Li(OCOO)CH2]2 for
EC and EC+DEC systems with 1M LiPF6, see C. R. Yang, Y. Y.
Wang, C. C. Wan, J. Power Sources, 72(1998)66.
On the importance of
Coulombic efficiency I
Li+ + e-
1
+
2
→
Cycle
1
2
3
Capacity
(Ah0)I
[(Ah0) I ]I
[(Ah0) I I ]I
N
(Ah0)(I)N
1
LiCH2CH2OCO2Li
2
For N = 5000 cycles and a 12/16 or 75% capacity retention,
the current efficiency per cycle must be such that
[Ah0(I)N ]/Ah0 > 0.75, or I > (0.75)(1/5000) , hence I > 0.99994.
•
•
This is why very low rates of lithium-consuming reactions can lead to premature
cell failure. The rates can be so low that they are not measureable in terms of
seeing current maxima associated with solvent reduction.
Note: high capacity negatives (Si, Sn based)…large challenge!
Graphite|iron-phosphate cell…excellent power density, life, and
potential for low cost. Challenged on energy density.
4
The iron phosphate is
so reliable that we
deduce graphite
degradation within this
cell.
Graphite|FePO4 cell
45oC, C/2, 90% DOD
3.75
Cell potential, V
3.5
3.25
New
963 cycles
3
P0
Cell capacity loss
P5
2.75
2.5
2.25
2
0
Charged,
0.25
0.5
FePO4 & Li~0.8C6
0.75
1
1.25
1.5
Cell capacity, Ah
1.75
2
2.25
2.5
Discharged, LiFePO4 &
C6
Iron phosphate vs. Li
- Little voltage variation
Graphite vs. Li
- Voltage variations
Analysis of FePO4/
graphite cells
50% DOD, 6C,
45oC,1376 cycles
New
 Conventional differential
voltage spectroscopy, but here
on the full FePO4-graphite cell
 Peaks result from graphite
staging (next slide)
 Peak broadening indicating
reduction in crystallite size
Utility of dV/dQ vs Q, uniform shifting of
peaks for graphite/FePO4 cells
d(Cell potential, V)/d(Capacity, Ah)
2
Cell capacity loss
Graphite|FePO4 cell
45oC, C/2, 90% DOD
1.8
1.6
1.4
New
1.2
963 cycles
1
P0
0.8
P5
0.6
0.4
0.2
0
0
0.25
Discharged
0.5
0.75
1
1.25
1.5
Cell capacity, Ah
1.75
2
2.25
2.5
Charged
 Same as previous plot with the exception that origin now is at the fully
discharged state…clear that distance between graphite peaks is nearly
constant
 Conclusion: lithium consumption (at the negative electrode surface) is leading
to capacity decline
Chemical-mechanical degradation at the negative
electrode
SEI forms
on newly
exposed
surfaces
(cracks)
Expansion &
contraction
upon charge
& discharge,
respectively.
Cracks
via cycling
Electrode isolation and
loss of active material
when cracks join
• Consistent with
additional loss of
negative capacity
[Li - S]  R - H - O  SEI  gasses  ...
[Li - S] 
 S  LiCO 3   H 2 C  CH 2 
Increased disorder and cracking.
d002 peak-width at half max
amplitude increases with cycling.
Supported by Raman analyses.
• Consistent with loss of active
lithium.
Summary: role of surface layers on + and Newer lithium ion
Conventional
lithium ion
V Electrochemical reaction
1.35 NiOOH +H2O + e- = Ni(OH)2 + OH
~ 1.35 V
+
-
0.4 FePO4 + Li + e = LiFePO4
+
0 H + e = 0.5H2
~ 2.5 V
~ 1.2 V
Solvent oxidation on Pt ~2.1 V vs Li
~ 3.3 V
O2 + 2H+ + 2e- = 2H2O
+
-1.5 Li4Ti5O12 + 3Li + 3e = Li7Ti5O12
Solvent reduction on negative ~0.8 V vs Li
~4V
Lithium ion conventional stability window
+
1 Li1-xMO2 + xLi + xe = LiMO2 (M: Ni, Co, Mn)
1.3 V
+
-2.9 C6 + Li + e = LiC6
+
-3 Li + e = Li
 Underscores the importance of the SEI
• Disruption of the SEI (e.g., due to dilation, crack propagation, etc.) is deleterious to cell
life…even low reaction rates are a problem
– Loss of Li
– Gas generation
Separators and ceramics
V
e-
Li+
 Function
• “Zero” electronic conduction
– Requires mechanical integrity
e-
eLi+
PF6
eLi+ Li+
Li+
eLi+
eLi+
Li+
Li+
Li+
PF6-
(+) Metal oxide,
phosphate, or silicate
Li+
-
PF6-
PF6-
– High porosity is desired
– Wetted by conventional solvent+salt systems (e.g., LiPF6
in EC+DEC)
• Poly(propylene) and poly(ethylene)
 Relatively new development: ceramic enhancement
PF6-
Li+
PF6-
Separator (Solvent + Salt)
• Facile ionic conduction
 Current separator costs are significant
PF6-
PF6-
Li+
– Low porosity helps to mitigate dendrite shorting
 Strong element of cell abuse-tolerance strategy
Li+
Li+
PF6-
(-) Carbon,
titanate, etc.
Conventional
separator
PP or PP|PE|PP
Ceramic (alumina
and/or silica)
enhanced separator
• More on expansion and
contraction of active materials
due to diffusion induced stress
analogous to thermoelasticity
analyses of ceramics
Potential step, Q0  QR
Radius R

(1   ) K s
0
12
E
R
R
S1 
, S2  s
(1 - 2 ) 2 K
(1 - 2 ) 2 K s
1
1
E
R
E
R
Surface modulus Ks
r
Surface stress 0
1. Influence of both terms vanish as R  
2. Results now consistent with nano-particle
and thin film electrodes yielding enhanced
cycle life
3(1 - )
2
 r (t , r )  Q R ( S1 - 1)
3
ECS
 S
sin (nx) - nxcos (nx)  - n 2 2 3(1 - )
S2
- 4(Q R - Q 0 )  1 2  (-1) n

e
3
(

)
(

)
n
n
x
E
C

n 1 

S
3(1 - )
2
  (t , r )  Q R ( S1 - 1)
3
ECS

2 2
 2 S1
3(1 - )
n sin nx - nxcos nx 
n sin nx  
(
1
)
(
1
)

S2
- 2(Q R - Q 0 ) e - n   


2
3
(
n

)
(
n

x
)
n

x
E
C

n 1


S

For the stress functions, the transient terms
are proportional to DSOC (DSOC  stress)
Y-T Cheng and M. W. Verbrugge, “The Influence of Surface Mechanics on Diffusion
Induced Stresses within Spherical Nanoparticles,” J. Appl. Phys., 104(2008)83521.
Stress
DStress
Time
Steel
Stress
Amplitude,
DStress
Endurance or
fatigue limit
Aluminum,
Magnesium
log(Number of cycles)
SOC
DSOC
(~stress)
Time
1
EV
0.9
0.8
Plug-in HEV
DSOC
0.7
Supercapacitor
0.6
0.5
0.4
0.3
0.2
0.1
Dashed line:
DSOC = -0.320 log(N) + 1.87
R² = 0.999
HEV
0
1,000
10,000
100,000
1,000,000
Cycle life requirement
http://www.uscar.org/guest/article_view.php?articles_id=85
Direct analogy to
the lower cycle-life
fatigue…stress
amplitude is
replaced by DSOC
Model result:
maximum stress is
proportional to the
maximum
difference in SOC,
or DSOC
M. W. Verbrugge and Y-T. Cheng, “Stress
and Strain-Energy Distributions within
Diffusion-Controlled Insertion-Electrode
Particles Subjected to Periodic Potential
Excitations,” J. Electrochem. Soc.,
156(2009)A927.
Problem statement
Single Particle Model…particle
diffusion resistance dominates
 Phenomenological stress-strain
relations (cf. thermoelasticity)
 Phenomenological guest lithium
(intercalate) diffusion
1
 r - 2    1 
Cc
3
E
1
1




1 -    -  r  
 
Cc
3
E
r 
 Mechanical equilibrium
  2 c 2 c 
c

 D 2 
t
r r 
 r
c0, r   c0
cmax for t - nt cycle  t1
ct , R   
cmin for t - nt cycle  t1
Radius R
 -
d r
2 r
0
dr
r

r
 Boundary conditions
 Finite stresses at particle center
 Surface
r
•
R
2
 -  surf
r
2
 -  0  K s 
r

R
surface tension and surface
modulus
ct ,0  finite
 The equation system can be solved
analytically

 Simplified problem statement
R
 Decoupled concentration problem
 No variable physical properties
 Idealized (spherical geometry)
Stress profiles (stationary state)
 Define
(1   ) K
E
R
S1 
(1 - 2 ) 2 K s
1
E
R
1-
Stress
DStress
s
and S 2  -
Stress
Amplitude,
DStress
2
 Steel
0
Time
R
(1 - 2 ) 2 K s
1
E
R
Endurance or
fatigue limit
Aluminum,
Magnesium
log(Number of cycles)
 Solution. Simple in structure and similar to step concentration solution.
SOCof the cycle.
Analogous expressions result for the second portion
(~stress)
DSOC
Time
c R , max
3(1 -  )
3(1 -  )
2
 x ( , x;  T1 ) 
S 2  ( S1 - 1)
E(cmax - cmin )
E(cmax - cmin )
3
(cmax - cmin )
1
EV
0.9
0.8
Plug-in HEV
 e -  j (T  -T1 ) - e -  j
- 4 
-  jT
-1
e
j 1 


DSOC
0.7
Supercapacitor
0.6
0.5
  S1
j sin( jx ) - jx cos( jx ) 
 (-1)


2
3


(
j
)
(
j
x
)
 

0.4
0.3
0.2
0.1
Dashed line:
DSOC = -0.320 log(N) + 1.87
R² = 0.999
HEV
0
1,000
10,000
100,000
1,000,000
Cycle life requirement
http://www.uscar.org/guest/article_view.php?articles_id=85
c R , max
3(1 -  )
3(1 -  )
2
  ( , x;  T1 ) 
S 2  ( S1 - 1)
E(cmax - cmin )
E(cmax - cmin )
3
(cmax - cmin )
 e -  j (T  -T1 ) - e -  j
- 2 
-  jT
-1
e
j 1 


  2 S1
j sin( jx ) - jx cos( jx )
j sin( jx ) 
(
1
)

(
1
)

2
( jx) 3
jx 
  ( j )
Stresses are proportional to DSOC
Current and next generation
negative electrodes
LiC6
372 mAh/g (theoretical)
Dominique Larcher, Shane Beattie, Mathieu Morcrette, Kristina Edström,
Jean-Claude Jumas and Jean-Marie Tarascon, “Recent findings and
prospects in the field of pure metals as negative electrodes for Li-ion
batteries,” J. Mater. Chem., 2007, 17, 3759 – 3772
Si: clear
“theoretical winner”
Why is there such
interest in nanoscaled insertion &
alloy electrodes?
1. High power
capability…shorter
solute(intercalate)
diffusion distance and
higher surface area
for electrochemical
reaction
2. Absent surface-areaenhanced degradation
reactions, smaller
particles should be
more (mechanically)
robust
Summary

Automotive context

Short overview/primer on lithium
ion batteries

Ceramics…now and in the future
of lithium ion battery technology

Positives…most are ceramics

Negatives…SEI; LTO

Separators…ceramic enhanced

Dilation and fatigue…classical
ceramic investigations are
proving helpful
Acknowledgments
 HRL LLC
• Jocelyn Hicks-Garner, Ping Liu,
Elena Sherman, Souren
Soukiazian, John Wang
 GM
• Danny Baker, Brian Koch, Mike
Balogh, Mei Cai, Xiaosong Huang,
Hamid Kia, Anil Sachdev, Curt
Wong, Xingcheng Xiao
 University of Kentucky
• YT Cheng
 Brown University
• H. Haftbaradaran and H.Gao