Quantitative Estimation of Capacity Fade of Sony 18650 cells Cycled at Elevated Temperatures by Branko N. Popov, P.Ramadass and Bala S. Haran Center for.

Quantitative Estimation of Capacity Fade of Quantitative Estimation of Capacity Fade of

Sony 18650 cells Cycled at Elevated Sony 18650 cells Cycled at Elevated

TemperaturesTemperatures

by

Branko N. Popov, P.Ramadass and Bala S. HaranCenter for Electrochemical Engineering

Department of Chemical Engineering

University of South Carolina Columbia, SC 29208Center for Electrochemical Engineering

University of South CarolinaCenter for Electrochemical Engineering

University of South Carolina

ObjectivesObjectives

Develop a methodology to determine the cause of capacity fade in Li-ion cells:

Primary Active Material (Li+) loss

Secondary Active Material (LiCoO2/Carbon) losses

Rate Capability loss

Factors that control the capacity loss:

Charging protocol

Cycling Temperature

Charge and discharge rates

The depth of discharge (DOD)

Quantify capacity fade using experimental data.

Develop a capacity fade model that would predict cycle life of a Li-ion cell

Experimental – Cycling StudiesExperimental – Cycling Studies

Studies of 18650 Li-ion Cell:

Cells cycled using Constant Current-Constant Potential (CC-CV) protocol.

Charged at 1A current till potential reaches 4.2 V

Hold potential at 4.2V till current decays to 50 mA.

Cells were discharged at a constant current of 1 A.

Batteries were cycled at four temperatures: RT(25oC), 45oC, 50oC and 55oC.

Rate capability studies done after 150, 300 and 800 cycles

Cells charged at 1 A and discharged at different rates (C/9 to 1C).

EIS measurements were done for fresh and cycled cells. (100 kHz ~ 1 mHz ±5

mV)

Studies of fresh and cycled electrode materials were carried out using a T-Cell

assembly with Li metal being the counter and reference electrode.

Discharge Curves at Various CyclesDischarge Curves at Various Cycles

0.0 0.4 0.8 1.2 1.6 2.0

Capacity (Ah)

2.00

2.44

2.88

3.32

3.76

4.20

Volta

ge (V

)

55-Dicharge

1

150300

490

0.0 0.4 0.8 1.2 1.6 2.0

Capacity (Ah)

2.00

2.44

2.88

3.32

3.76

4.20

Vol

tage

(V)

RT-Dicharge

1

500150

300

800

0.0 0.4 0.8 1.2 1.6 2.0

Capacity (Ah)

2.00

2.44

2.88

3.32

3.76

4.20

Vol

tage

(V)

45 degree-Dicharge

1

150

300500

800

0.0 0.4 0.8 1.2 1.6 2.0

Capacity (Ah)

2.00

2.44

2.88

3.32

3.76

4.20

Vol

tage

(V)

50 degree-Discharge

1

150300

600

500

50 deg C 55 deg C

45 deg C25 deg C

Capacity Fade as a Function of Cycle LifeCapacity Fade as a Function of Cycle Life

Temperature% Capacity Fade (cycle number)

50 100 150 300 500 800

25oC 3.8 5.11 6.06 10.29 22.5 30.63

45oC 3.8 5.46 7 11.75 26.46 36.21

50oC 2.4 5.1 7.9 23.9 43.21 -

55oC 4.3 6.4 9.4 27 70.56 -

Change in Charging Times with CyclingChange in Charging Times with Cycling

Constant Current

Constant Voltage

RT 45 50 550.0

0.5

1.0

1.5

CC Time (h)

1

150

300

1

150

300

1

150

300800

800

300

1

150300

500

1

RT 45 50 550

1

2

3

CV Time (h)

1

150300

1150

300

1

150

300

800800

490

1

150

300

500

1

0.0 0.4 0.8 1.2 1.6 2.0

Applied Current (A)

0.75

1.00

1.25

1.50

1.75

2.00

Dis

char

ge C

apac

ity

(Ah)

Rate Capability comparison after 150 and 800 cycles

Fresh

800

300

150

0.0 0.4 0.8 1.2 1.6 2.0

Applied Current (A)

0.75

1.00

1.25

1.50

1.75

2.00

Dis

char

ge C

apac

ity

(Ah)

Rate Capability comparison after 150 and 800 cycles

Fresh

150

800

300

Rate Capability with CyclingRate Capability with Cycling

50 deg C 55 deg C

45 deg C25 deg C

0.0 0.4 0.8 1.2 1.6 2.0

Applied Current (A)

0.75

1.00

1.25

1.50

1.75

2.00

Dis

char

ge C

apac

ity

(Ah)

Rate Capability comparison after 150 and 800 cycles

Fresh

500

150

300

0.0 0.4 0.8 1.2 1.6 2.0

Applied Current (A)

0.75

1.00

1.25

1.50

1.75

2.00

Dis

char

ge C

apac

ity

(Ah)

Rate Capability comparison after 150 and 800 cycles

Fresh

150

300

Variation of Cell Impedance with CyclingVariation of Cell Impedance with Cycling

RT 45 50 550.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8O

vera

ll C

ell R

esis

tanc

e ()

1111 150150150150300300300300

800

800

600

1

Comparison of Electrode ResistancesComparison of Electrode Resistances

150 Cycles 300 Cycles

800 Cycles

RT 45 50 550

200

400

600

800

1000

1200

Ele

ctro

de R

esis

tanc

e (-cm2 )

Electrode resistance after 150 cycles for a fully charged cell

LiCoO2Carbon

1

RT 45 50 550

200

400

600

800

1000

1200

Ele

ctro

de R

esis

tanc

e (-cm2 )

Electrode resistance after 300 cycles foa a fully charged cell

LiCoO2Carbon

1

RT 45 500

200

400

600

800

1000

1200

Electrode Resistance (-cm2 )

Electrode resistance after 800 cycles for a fully charged cell

600 cyclesLiCoO2Carbon

800 cycles800 cycles

1

Specific Capacity of Positive and Negative Electrodes at Various Specific Capacity of Positive and Negative Electrodes at Various Cycles and TemperatureCycles and Temperature

TemperatureCycle

Number

Specific Capacity (mAh/g)

LiCoO2 Carbon

Fresh 148.132 339.896

25oC

150 145.61 334.03

300 141.07 325.04

800 122.14 271.10

45oC

150 143.74 332.84

300 139.26 325.71

800 118.43 264.56

50oC

150 143.28 331.20

300 133.56 293.91

600 94.05 202.53

55oC150 142.84 328.90

300 131.13 286.92

Capacity Fade BalanceCapacity Fade Balance

Q = QQ = Q11 + Q + Q22 + Q + Q33

Q: Total Capacity Loss

Q1: Capacity Fade due to rate capability loss Difference in capacity between C/9 and C/2 rate discharges.

Q2: Capacity Fade due to loss of secondary

active material (LiCoO2 and Carbon) Measurement done on T-cells

Q3: Capacity Fade due to loss of primary

active material (Li+) and other losses.

Capacity Fade - QuantifiedCapacity Fade - Quantified

150 300 8000

20

40

60

80

100

% o

f Tot

al C

apac

ity

Los

s

Electrode resistance after 150 cycles

Cycle Number

Q1Q2Q3

1

150 300 8000

20

40

60

80

100

% o

f Tot

al C

apac

ity

Los

s

Electrode resistance after 150 cycles

Cycle Number

Q1Q2Q3

1

2525ooCC 5050ooCC

Simplified Diffusion ModelSimplified Diffusion Model

Concentration variations in the solution phase can be

neglected for low to medium discharge currents.

Solid phase potential drop is negligible as compared to

kinetic and concentration over-potentials.

Eliminating the above two results in a simple diffusion

model which can be used to simulate the performance of

the Li-ion Battery.

Model considers Li+ reaction at carbon/LiCoO2 particle

interface and subsequent transport in these materials.

Governing Equations: Diffusion Model Governing Equations: Diffusion Model

Fickian diffusion in spherical coordinates in carbon and LiCoO2

rc

rr

Dt

c Li2

r

effLiLi2

Initial ConditionoLiLi c c 0,t

Boundary Conditions

Li

eff Lip Li Li

cr 0, 0

rc j

r R , D rate r̂ F

a1 c1P P

F FΦ Φ

RT RT

Li,in o,1j j e e

ref reff fP PU jR U jR

Electrode Reaction RatesElectrode Reaction Rates

Concentration dependent exchange current density:

Lithium intercalation/deintercalation reaction:

a2 c2N N

F FΦ Φ

RT RT

Li,de o,2j j e e

ref reff fN NU jR U jR

P NV

, ,max, , , , , 1, 2

a j c js so j j Li j Li j Li jj k c c c j

Rf refer to total resistance that includes ohmic R and polarization RP resistances. (Rf=R+RP)

UUrefref A Function of SOC A Function of SOC

( )ref

p PU fn SOC

ref

nU = ( )

nfn SOC

SOCp

SOCn

Ure

f p

Ure

f n

LiCoO2

Carbon

Parameters Considered for Diffusion ModelParameters Considered for Diffusion Model

State of Charge of the electrode limited by capacity

To account for capacity loss due to primary and

secondary active materials.

Solid Phase Diffusion Coefficient

To account for capacity loss due to rate capability.

Film Resistance

To account for the drop in cell voltage due to increase

in ohmic resistance and polarization losses.

Incorporation of QIncorporation of Q22 and Q and Q33 in Diffusion Model in Diffusion Model

SOC of the electrode material could be estimated

from active material losses.

Calculate SOC of the negative electrode based on

the capacity loss (Q2 + Q3).

Develop a correlation for variation of SOC of

negative electrode with cycle number.

Using this capacity fade due to active material losses

could be incorporated in the diffusion model.

Comparison of Diffusion Model and RT Comparison of Diffusion Model and RT Experimental Data Experimental Data

Parameter Property

SOCActive material

losses

Dns Rate Capability

losses

RfDrop in cell voltage

with cycling

Comparison of Diffusion Model and RT Comparison of Diffusion Model and RT Experimental Data for Utilization (Rate Capability)Experimental Data for Utilization (Rate Capability)

0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75

Current (A)

1.0

1.2

1.4

1.6

1.8

Cap

acit

y (A

h)

150 Cycles

Fresh

300 Cycles

800 Cycles

ConclusionsConclusions

A diffusion model was developed to simulate the discharge curves of Li-ion cell.

Empirical correlations have been developed for variation of SOC of Li-ion cell with continuous cycling.

Active material losses have been accounted through the variation of negative electrode SOC in the diffusion model.

Rate capability losses and Polarization resistance increase have been accounted through varying the diffusion coefficient and exchange current density respectively.

Inclusion of effect of charging and discharge rates and DOD into diffusion model is currently in progress.

AcknowledgementsAcknowledgements

This work was carried out under a contract with the

National Reconnaissance Office

for

Hybrid Advanced Power Sources

# NRO-00-C-1034.

Center for Electrochemical EngineeringUniversity of South Carolina

Center for Electrochemical EngineeringUniversity of South Carolina