Nitash Balsara, Lawrence Berkeley National Laboratory

Report
Characterization of Battery Failure
Nitash Balsara
University of California, Berkeley
Lawrence Berkeley National Laboratory
Intellectual support:
John Newman (Berkeley)
Support:
Soft Matter EM Program (BES), Lithium sulfur LDRD and
Program (BATT, EERE)
Ideal speciation spectroscopy
PS-b-PEO
NaPF6
PS
PEO
Signal
salt
PEO
Salt
200
500
PS
100
Energy
TEM of block copolymer electrolytes
PS-b-PEO
EFTEM SI datasets analyzed by
Principle Component Analysis
NaPF6
• Intense plasmon peak is
sensitive to the material’s
chemical structure.
• One big fat broad peak
regardless of chemical
details.
• Typically, we record a
series of EFTEM images
from 5-60eV in 1eV steps
using a 5eV energy slit.
Species identification
Principal component analysis (PCA)
enables speciation in spite of big fat
highly overlapping ugly peaks!
Frances Allen, Andy Minor, et al., Ultromicroscopy, 2012
Lithium Battery
-
current
collector
anode:
Li/graphite
alkyl
carbonates/
Li+X-
Li+
cathode:
FePO4
+
anode at 0.2 V (vs Li/Li+)
Li
Li+ + e-
cathode at 4 V
Li+ + e- + FePO4
LiFePO4
current
collector
conductive
carbon and
binder
Components of electrodes: active particles, electron conductor, ion conductor
Miracle of lithium battery electrodes
conductive carbon
inactive binder
reaction sites
e-
active
particle
Li+ + e- + FePO4
LiFePO4
Li+
Need equal electronic and ionic conductivity
electrolyte
Components of electrodes: active particles, electron conductor, ion conductor.
Every active particle needs to have access to electron and ion conducting
pathways at all times!
It is a miracle that any electrode works.
Time-honored Approach
 Synthesize a new cathode or anode material.
 Make a full cell (not half cell with excess
lithium).
 Make sure that the cathodes are thick enough
to be commercially viable.
 Count electrons coming in and out of the
battery.
 Measure voltage.
 Usual result: terrible
 Once in a while: result better than terrible base
line  Publish!
 Advantage: If it works you can go directly to the market.
 Disadvantage: You might be throwing out a perfectly good material.
Perhaps the binder failed.
Example of approach
Why did the miracle not happen?
What if the binder was not holding the
structure together?
What if the electrons were reacting
with the electrolyte?
What if ionic pathways were blocked?
What if the electronic pathways were
blocked?
What if the top half of the cell is
working fine but the bottom half is not?
e-
active
particle
Li+
PIs who go through the pains of
establishing a new synthesis route
may no be great cell builders.
JCESR desperately needs a facile lab to answer these questions.
We cannot afford to throw out good active materials.
The prototyping and characterization group has an opportunity to create such an
infrastructure.
Balsara’s Version of the Time-honored
Approach
 Synthesize a polymer binder.
 Make a full cell (not half cell with excess
lithium).
 Make sure that the cathodes are thick enough
to be commercially viable.
 Count electrons coming in and out of the
battery.
 Measure voltage.
 Usual result.
Joke: Balsara drops his keys in front of his home and the
area around the door is pitch-dark. An hour later a
neighbor drives by and sees Balsara searching for
something on the street near lamppost. “What’s up?” he
asks. “Looking for my keys”. “Why are you searching
here?” he asks. “Because there is light here and I can
see.”
Lithium-air saves the day
Perfect for fundamental studies because…
Easy to detect gasses!
Detect something other than
electrons
electrolyte:
carbonate/ether
McCloskey et al., JACS 2011
ether
Li-S needs miracles
Sulfur
Lithium
++- -+
+- -+
Quoting Venkat: Milk-to-cheese-to-milk-…
-+
Ideal speciation spectroscopy
Signal
Li2S4
Li2S6
200
500
Li2S2
100
Energy
Work in a pristine environment
(computer)
Simulation of
Li2Sx and S8
dissolution in
tetraglyme
(oligomeric PEO)
Tod Pascal
Tod Pascal and David Prendergast
Simulations show clustering
• Li2S6
These simulations show clustering that is
not limited by box
• Li2S8
Work in a real environment
XAS measurements of
Li2Sx in PEO or SEO
Kevin Wujcik
Wujcik, Velasco-Velez,
Cabana, Salmeron, Guo
XAS predictions Li2Sx in
tetraglyme (in progress…)
Bottom line: No clear distinction between
species…
Principal Component Analysis
Juan Velasco Velez
Eigenvalue
Indicator
72.1
7.2
1.29
0.49
8.58
2.78
1.15
0.993
• Truncate at 3
• Unconstrained PCA
analysis gives unphysical
underlying spectra.
Iterative transformation factor analysis (Marcus)
Theory vs Experiment
Ideal speciation spectroscopy
Signal
Li2S4
Li2S6
200
500
Li2S2
100
Waiting for this is not an option…..but developing such tools is!
Energy
Prototype Electrode Project
Tab
Electron and ion conducting binder
(electron side street)
contamination
Active particle
C nanotubes
(electron highway)
Ensure that electron
and ion are delivered
to particle.
Characterization Task: Make the
miracle happen.
 Respect the intuition of Stan Wittinghams
of the day. (You cannot convince me to
work on binder for Sn5Fe.)
 We must work under our lampposts.
Need to find lampposts near the key.
 Make sure we do not discard gem.
 Get your hands dirty and help the
synthetic folks succeed.
 Concerted deployment of tools: theory,
experiment, statistics.
e-
active
particle
Li+
Nobody can afford to throw out good active
materials.
JCESR prototyping and characterization groups
have an opportunity to create the necessary
infrastructure.
Credit: Frances Allen, Kevin Wujcik, Juan Velasco Velez, Tod Pascal
Andy Minor, Jinguha Guo, Miquel Salmeron, David Prendergast

similar documents