8/13/2019 3D Microstructrue Electrochemistry Model

1/1

THREE-DIMENSIONALLY RESOLVED SIMULATIONS OF ALICOO2 ELECTRODE STRUCTURE OBTAINED VIA FIB/SEM

CHRISTIANWALCHSHOFER1, TOBIASHUTZENLAUB2, SIMONT HIELE2,BORISKALUDERCIC1, ROBERTS POTNITZ3

1CD-adapco (christian.walchshofer@cd-adapco.com), 3 Battery Design LLC (rspotnitz@batdesign.com)2 Department of Microsystems Engineering - IMTEK, Laboratory for MEMS Applications (tobias.hutzenlaub@hsg-imit.de)

PREFACEEstablished simulation models for Li-ion batter-

ies are formulated in 0D or 1D (Dualfoil), and relyon assumptions, e.g. well defined particle shapes

and absence of diffusion between such particles.Fitted solid conductivities have to be applied tomatch experiments. The 3D Micro-Structural Elec-trochemistry model implemented in STAR-CCM+extends the well known Dualfoil model to 3D [1] toaccurately predict spatial phenomena.

SETUPA VARTA LIC 18650 WC lithium-ion battery

was segmented by FIB-SEM in [2] and recon-structed (right box). The reconstructed porous elec-trode block was contacted with a current collector,separator andnegativefoil (carbon) to form a carte-sian mesh simulation setup comprising 21 M cells.

Detailed model parametrization was conducted

by applying LiCoO2 equilibrium potentials from[3] and electrolyte parameters from [4].

LiPF6 in EC-EMC-DMC (1:1:1, vol.); 20C

The setup assumed an aluminum current col-lector, a perfectly porous (= 1) separator and acarbon anode foil. A 1C charge was applied.

CONCLUSIONSModeling electrically conductive pathways ac-

curately by introducing a third phase, which rep-resents conductive aid and binder, allows for us-ing physically measured conductivities for activematerials in simulations. The third modeled phaseleads to stronger gradients in electrolyte concentra-tion, affecting cell performance.

GEOMETRYGENERATION

The image stack provided by [2] was read into Scilab using the ScilabImage Processing toolbox (SIP), and coarsened (2x) by a majority wins

algorithm. Small regions of minor impactwere removed by applying asmoothing algorithm. A mesh was generated using pro-STAR and im-ported into STAR-CCM+. No transport was assumed between binderand electrolyte phases. A Java-macro was used therefore to removeelectrolyte regions purely contacting binder and vice-versa, thereby re-moving 0.390% of electrolyte and 0.015% of binder cells.

MODELF ORMULATIONTransport equations

Solid Liquid

J= J=+2RuT

F

1t0+

1 +

d(ln f)

d(ln c)

(ln c)

S

dS= 0 S

dS= S

d (ln c)dS

t

V

cdV =

S

D cdS

t

V

c dV =

S

D cdS

V

J t0+F

dV

D= D0

1

d (ln c0)

d (ln c)

Butler-Volmer kinetics

Jn,s= J0

e

a F

R Te

cF

R T

+C

t(sl) J0 = F k

cs

cs,max

11

cs

cs,max

2 clcl,ref

3

RESULTS

t = 8 (s) t = 360 (s) t = 3600 (s)

SOC = 0 SOC = 0.1 SOC = 1

REFERENCES

[1] Spotnitz et al. 2012 Geometry-resolved electro-chemistrymodel of li-ion batteries SAE Technical Paper2012-01-0663

[2] Hutzenlaub et al. 2012 Three-Dimensional Reconstructionofa LiCoO2 Li-Ion Battery Cathode Electrochemical and Solid-

State Letters, 15 (3), A33A36[3] Karthikeyan, Sikha, White, 2008 Thermodynamic modeldevelopment for lithium intercalation electrodes. Journal ofPower Sources185 (2), 1398 1407.

[4] Gering, K. L. 2006 Prediction of electrolyte viscosity foraqueous and non-aqueous systems: Results from a molec-ular model based on ion solvation and a chemical physicsframework.Electrochimica Acta51 (15), 3125 3138.

[5] Hutzenlaub et al. 2013 Electrochemical modelling in aFIB/SEM based three-phase reconstruction of a LiCoO2 Li-ion battery cathode to appear