EXTREME FAST CHARGE CELL EVALUATION OF LITHIUM-ION BATTERIES (XCEL): OVERVIEW

BAT 386

EXTREME FAST CHARGE
CELL EVALUATION OF
LITHIUM-ION BATTERIES
(XCEL): OVERVIEW
VENKAT SRINIVASAN SAMUEL GILLARD
Argonne National Lab Department of Energy
vsrinivasan@anl.gov samuel.gillard@ee.doe.gov

 This presentation does not contain any proprietary,
 confidential, or otherwise restricted information

 AMR 2021

OVERVIEW
 Timeline Barriers
 § Start: October 1, 2017 § Cell degradation during fast charge
 § End: September 30, 2021 § Low energy density and high cost of
 § Percent Complete: 75% fast charge cells

 Budget Partners
 § Funding for FY20 – $5.6M § Argonne National Laboratory
 § Idaho National Laboratory
 § Lawrence Berkeley National Lab
 § National Renewable Energy Laboratory
 § SLAC National Accelerator Lab
 § Oak Ridge National Lab

RELEVANCE: FAST CHARGE REMAINS AN
ISSUE FOR WIDESPREAD ADOPTION OF EVS
 Energy Density
 (500 Wh/l)
 Tolerance to 100 Specific Energy
 Abuse 80 (235 Wh/kg)
 60

 Fast Charge 40
 Critical Materials
 (

RELEVANCE: WHAT LIMITS FAST CHARGE?
 Li plating Particle breakup

 Temperature rise Exponent Inc.
 38
 37
 4C
 36
 Temperature (°C)

 35
 3C
 34
 33
 2C
 32
 31
 30
 0 10 20 30 40
 Time (min)

 We need to understand how all three issues impact fast charge
 4

RELEVANCE: CHALLENGES AT MULTIPLE SCALES
 Anode Cathode

 Current Collector
 Current Collector

 Separator
 Li+

 e-
 Li2
 e- O LiR

 LiF
 Li+-e-
 Li+-e- Li+
 Graphite

 LEDC
 Li2
 Li+ O
 LiF

 Li2C Li
 O3 R

 5

COLLABORATION ACROSS LABS AND
UNIVERSITIES
 Cell and electrode design and building, performance characterization,
 post-test, cell and atomistic modeling, cost modeling

 Li detection, electrode architecture, diagnostics

 Performance characterization, failure analysis, electrolyte modeling and
 characterization, Li detection, charging protocols

 Thermal characterization, life modeling, micro and macro scale modeling,
 electrolyte modeling and characterization

 Li detection, novel separators, diagnostics

 Detailed Li plating kinetic models, SEI modeling

 6

CONTRIBUTORS AND ACKNOWLEDGEMENTS
Abhi Raj Eva Allen Michael Toney Victor Maroni
Alec Ho Fang Liu Molleigh Preefer Vikrant Karra
Alison Dunlop Francois Usseglio-Viretta Nancy Dietz Rago Vince Battaglia
Ankit Verma Guoying Chen Ning Gao Vivek Bharadwaj
Andy Jansen Hakim Iddir Nitash Balsara Volker Schmidt
Andrew Colclasure Hans-Georg Steinrück Orkun Fura Wei Tong
Antony Vamvakeros Hansen Wang Partha Paul Wenhan Ou
Anudeep Mallarapu Harry Charalambous Parameswara Chinnam Wenxiao Huang
Aron Saxon Ilya Shkrob Paul Shearing William Chueh
Bor-Rong Chen Ira Bloom Peter Weddle William Huang
Bryan McCloskey James W. Morrissette Pierre Yao Xin He
Bryant Polzin Ji Qian Quinton Meisner Yang Ren
Che-Ning Yeh Jiayu Wan Ravi Prasher Yanying Zhu
Chuanbo Yang Jeffery Allen Robert Kostecki Yi Cui
Chuntian Cao Johanna Nelson Weker Ryan Brow Yifen Tsai
Charles Dickerson Josh Major Sang Cheol Kim Yuqianz Zeng
Daniel Abraham John Okasinski Sangwook Kim Zachary Konz
Dave Kim Juan Garcia Sean Lubner Zhelong Jiang
David Brown Kae Fink Seoung-Bum Son Zhenzhen Yang
David Robertson Kandler Smith Shabbir Ahmed
David Wragg Kamila Wiaderek Shriram Santhanagopalan
Dean Wheeler Kevin Gering Srikanth Allu
Dennis Dees Maha Yusuf Steve Harris
Divya Chalise Manuael Schnabel Steve Trask
Donal Finegan Marca Doeff Susan Lopykinski
Elizabeth Allan-Cole Marco DiMichiel Swati Narasimhan
Eongyu Yi Marco Rodrigues Tanvir Tanim
Eric Dufek Matt Keyser Uta Ruett
Eric McShane Michael Evans Venkat Srinivasan

 Support for this work from the Vehicle Technologies Office,
 DOE-EERE: Samuel Gillard, Steven Boyd, and David Howell 7

APPROACH: UNDERSTAND THE PROBLEM TO
ENABLE SOLUTIONS
 Li plating Cathode cracking
 Link electrode- Quantify
 scale effects in inhomogeneities Quantify
 anode to Li and inter-
 Detect Li by conditions when
 plating (A. electrode effects
 readily-accessible different
 Jansen) (A. Colclasure)
 means verified by cathodes crack
 complex tools (T. Tanim)
 (J. Weker)
 Link charging
 protocols to long-
 term aging (E.
 Cathode
 Dufek)
 architectures to
 Quantify heat prevent cracking
New anode designs and generation rate and
 electrolytes to prevent causes (M. Keyser)
 Li plating
 Heat generation
 New protocols
 to extend life
 8

APPROACH COMBINES MULTIPLE EXPERTISE

 9

SUMMARY OF ACCOMPLISHMENTS
 2017 2021

• Plating free: 10 min charge at 1.7 mAh/cm2 • 10 min charge at 3 mAh/cm2*

• 532 cathode exhibits 22% capacity • 811 cathode exhibits 7% capacity
 fade (6C charge; 600 cycles)# fade (6C charge; 600 cycles)#
 • Developed electrolytes w/ enhanced transport

 • Developed multiple Li detection methods

 • Quantified heat generation during fast charge
*
 Hero cell with latest anode design, new and impact on life
electrolyte, and new separator
 • Developed model-informed charge protocols
#
 532 and 811 cathode cycled to 4.1V
 • Developed methods to detect SOC and Li
 10 plating heterogeneity across scales

WHAT DRIVES LI PLATING?
 6C charge; C/2 discharge full cells
 High loadings fade quickly and
 level out at ~ 1.3 mAh/cm²

 Round 1: 2 mAh/cm2 – baseline
 Round 2: 3 mAh/cm2 Future: 4 mAh/cm2

 XCEL developed a deep understanding of the underlying issues
 11

ALL THE ACTION HAPPENS CLOSE TO THE
SEPARATOR.
 4 mAh/cm2 (230 Wh/kg cell; 110 micron electrodes)

 Li plates near separator during charge § Next Gen Electrolyte = 2X ionic
 conductivity, 4X diffusivity, and
 transference number increased by 0.15

 Need a combination of solutions to enable fast charge
 12

See BAT456
“HERO” CELL SHOWS IMPACT OF SYSTEMATIC CHANGES
 Reduced CBD loading + Celgard separator 2500 + enhanced electrolyte
 6C CC Charging of Single Layer Pouch Cells 6C CCCV to 4.1 V (10 minutes cutoff)
 R2 BD 5 00 B 26
 rd C +2 0 + d R1
 Minimal lithium plating
 n da +4
 %
 B D, 5 0 r predicted for R2 loading with
 Sta 4 %C + 2
 n da electrode/separator/electrolyte
 + D Sta
 CB enhancements
 4%

 Dots represent measurements at INL
 Lines are NREL Macro-model Predictions
 Electrolyte properties from Kevin Gering’s AEM

 B26: EC-DMC-DEC-EP-PN (20:40:10:15:15,
 mass) plus LiPF6 with 3% VC and 3% FEC.

 3 mAh/cm2 behaves like 2 mAh/cm2
 cell. Model predictions match exp. data

IMPROVEMENTS TO ANODE DESIGN AND See BAT456

ELECTROLYTE CONTINUE

 = 25%

 = 45% Smaller particles
 e.g.: SLC1520P/SLC1506T
 = 45% = 45% Larger particles
 = 25% = 25%
 4.1 V
 Baseline rnd2 Anode only Both electrodes

 Particle size/graded porosity

Implementation of these changes in the pipeline Electrolyte 3 = Solvent C:EC:EMC,
 1.2M LiPF6 + 1% FEC + 1% VC

 14

See BAT461

 WE ARE EXAMINING OTHER HETEROGENEITIES
 Discharged Charged
 state state
 Saturation Distribution

 1.0 (red >
 orange >
 yellow >
 green >
 blue >
 ~3.5 mm indigo >
 violet) 0.9

 Neutron imaging confirms model observations
 that electrolyte wetting needs consideration
Phis-phie (V)

 Edge effects partially
 explain the data (~2 mm)
 15

See BAT461
MAPPING HETEROGENEITIES REMAINS AN AREA
OF IMPORTANCE
 Discharge Discharge Charge Mapping of ionic resistance
 Cycle 10 Cycle 452 Cycle 453
 C
 Cathode: c/a lattice ratio - increasing heterogeneity.
 Cathode SOC (red = less Li)
 5.002 5.085
 5.01
 5
 4.998 5 5.08
 4.996
 4.99 5.075
 4.994
 4.992 5.07
 4.98
 c/a = 4.9951±0.0033 c/a = 4.9962±0.0084 c/a = 5.0797±0.0042

 Anode: x in LixC6 - low heterogeneity A
 1 1 1

 0.8 0.8 0.8

 0.6 0.6 0.6

 0.4 0.4 0.4

 0.2 0.2 0.2

 0 0 0

 Quest continues to understand all the sources of inhomogeneity
 and to correlate it to Li plating
 16

CLASSIFICATION OF LI DETECTION See BAT457

TECHNIQUES

 Paul, Partha P., et al. "A Review of Existing and Emerging Methods for Lithium Detection and Characterization in
 Li-Ion and Li-Metal Batteries." Advanced Energy Materials 11.17 (2021): 2100372.

See BAT457
 MAP LI CONCENTRATION GRADIENT IN ANODE
 Stronger gradient precursor to Li plating
 300 nm

 insulation
 Separator
 (thickness)

 Cathode
 2 mm 100 µm

 Sensor
 Anode
 Cu CC
 Al CC
 Local relative Li concentration
 Ve
 ry
 ba
 d1C charging

 Less
 good 66% SOC

Points = sensor measurements 47% SOC
 Ok
Shading = agnostic predictions from
 28% SOC
electrochemical simulations

See BAT457
 PRESSURE EVOLUTION WITH CYCLING
 Pressure changes at discharged state correlate dead Li
30 psi 3rd cycle 6th cycle 8th cycle 10th cycle 10th cycle
 Graphite

 1 inch

0 psi Voltage vs. time
 Average pressure vs. time
 @ 6C

See BAT463
CATHODE CRACKING IDENTIFIED AS AN ISSUE

 Fresh 9C, 225 Cycles
 25 cyc
 225 cyc

• Li plating dominates 6C for cracks after cycling

 Program started with 532 cathode. Over the last year, transitioned to 811

 20

See BAT463

EARLY RESULTS ON NMC 811 PROMISING

 4C@425cyc
 532 cathode

 811 cathode

 While cracking remains an issue, does not appear to exacerbate on cycling.
 Now examining single crystal cathodes.
 21

See BAT459
HETEROGENEITY AT LARGE SCALES
 Voltage (Volts)

 25 Ah pouch cell simulation

• Primary reasons for heterogeneity: Temperature (°C)
 – Higher current density near the tabs.
 – Higher temperatures away from the
 cold plates leading to higher
 conductance of electrode.

See BAT459
 THERMAL RAMP AS A KNOB
 Investigating 6C CCCV charging at 50oC and various cooling rates
 back to 30oC during discharge
§ NMC811/1506T cells with Gen 2
 electrolyte and B26 electrolyte under
 test at 3 thermal ramp rates during C/2 1000 XFC cycles over 10 • 20% longer
 discharge (20, 40 and 60 minutes) years life
 1000 cycles over 3+ • 17% longer life
§ Tests now underway years
§ Modeling shows the importance of
 rapid cooling
 – Accounts for cathode cracking and
 SEI/CEI growth. Does not include
 Li plating

 Rapid cooldown important for extending calendar life
 23

RESPONSE TO PREVIOUS YEAR REVIEWER’S
 COMMENTS
§ This project was not reviewed last year

 24

REMAINING CHALLENGES AND BARRIERS

1. Can we ensure good cycle life on 3 Ah/cm2 Gr-811 cells charging at 6C?
2. What knobs can be push to get 4 mAh/cm2 cells charging at 6C?
3. Do single crystal NMC-811 cathodes allow cycling at 4.2 V without cracking?
4. Can we quantify impact of higher temperature charge on cycle life?
5. How much inhomogeneity can we tolerate without impacting cycle life?

 25

PROPOSED FUTURE WORK*

 1. Can we ensure good cycle life on 3 Ah/cm2 Gr-811 cells charging at 6C?
 2. What knobs can be push to get 4 mAh/cm2 cells charging at 6C?
 3. Do single crystal NMC-811 cathodes allow cycling at 4.2 V without cracking?
 4. Can we quantify impact of higher temperature charge on cycle life?
 5. How much inhomogeneity can we tolerate without impacting cycle life?

*Caveats related to funding apply

 26

SUMMARY
 0 26
 R2 CB
 D 50
 ard % , +2 0 + B R1
 tan
 d +4 BD 50 rd
 S 4 %C + 2 nda
 + B D S ta
 C
 4%
 Systematic approach has resulted
 Comprehensive effort required
 in improvements
 Dots represent measurements at INL
 Lines are NREL Macro-model Predictions
 to enable fast charging Electrolyte properties from Kevin Gering’s AEM

 B26: EC-DMC-DEC-EP-PN (20:40:10:15:15,
 mass) plus LiPF6 with 3% VC and 3% FEC.

 ! = 25%
 1000 XFC cycles over 10 • 20% longer
 ! = 45% years
 1000 cycles over 3+2
 life
 811 results promising to date; but
 We still need to drive to 4 mAh/cm • 17% longer life
 ! = 45% ! = 45% years
 more needed
 ! = 25% ! = 25%

Baseline rnd2 Anode only Both electrodes

 27