Battery Design LLC. Overview of the. LLIBTA 2008 Tampa, Fl

Size: px

Start display at page:

Download "Battery Design LLC. Overview of the. LLIBTA 2008 Tampa, Fl"

Diana Simpson
6 years ago
Views:

1 Overview of the Life of Li-Ion Batteries Robert Spotnitz Tuesday May 13, 2008 Tampa, Fl

2 Overview The Importance of Battery Life Life Models Loss of Cyclable Lithium Li Loss through SEI Growth Site Loss Model Impedance Growth Conclusions Page 2

3 Battery Life Determined by the decrease in power/capacity p that occurs over time. Calendar Life loss on standing Cycle Life loss during charge/discharge cycles 60 o C 45 o C 20 o C Storage period (months) Page 3

4 Goal: Year Battery Life Begin of Life Disc charge Power (k kw) Discharge Power Goal Energy Goal Energy Goal+margin End of Life Regen wer (kw W) Re egen Po Net Energy Removed (at C/1 discharge rate) Page 4

5 Possible Types of Fading Curves in Cycle Life Testing Frank J. Wang, The 25 th Intl. Battery Seminar & Exhibit, March 17-20, 2008 Page 5

6 ATD Gen 2 Cell Behaviour I. Bloom et al. J. Power Src. 155 (2006) 415. Page 6

7 Life Modeling Want to account for impedance growth and capacity loss due to calendar time and cycling Current state of the art relies upon extrapolation of experimental results Simple functional forms are preferred Page 7

8 Key Aspects of Lithium-Ion Cells 1) Insertion compounds that provide sites for lithium 2) Amount of cyclable lithium Page 8

9 Approach Model loss of cyclable lithium and/or sites using simple differential equations that can be readily solved Life simulation is possible in reasonable time (minutes to hours) Arbitrary test conditions (rates, depth of discharge, etc.) Models available in Battery Design Studio for user-friendly interface Page 9

10 Solid Electrolyte Interface (SEI) Electron Tunneling Model SEI allows transport of lithium but inhibits electron transfer. Formation of SEI results in loss of cyclable clable lithium. Rate of SEI formation Li + dl dt = ke o fl L L50 40 A*ln(1+Bt) Graphite SEI Page time

11 Electron Tunneling Model e Lithiu um Cyclabl A:ko=1e-2, f=8675 B: ko=1e-4, f=6211 C: ko=1e-6, f=3737 dl D: ko=1e-7, f=2502 A B C D dt = ke o fl EquilFade-9-1 EquilFade-9-3 EquilFade-9-5 EquilFade-9-6 Page 11

12 Electron Tunneling Model Simulated Calendar Life Test hium Cyclable Lith Tunneling model does not predict an effect of voltage on capacity fade. Page 12

13 Schematic of SEI at Graphite Vinylene carbonate HH + 2e - + 2Li + Li 2 CO 3 + -(C=C-0)- Crystalline solid Polymeric film ion conductor Page 13

14 i Acid Generation at Positive LiMO MO + Li + e Li ( φ φ φ ρ ) F 1 2 eq i f L p = Fkp exp RT i ROH RO + H + e = Fk + ( φ φ φ ρ ) f p F i L eq, H exp H + H + RT Net Reaction + + Li + MO + ROH H + LiMO + 2 Page 14 2 RO LiMO 2

15 Acid attack of SEI CH CH 2 OLi + H ROH + Li H + ROH H + Li 2 CO 3 + H + LiHCO 3 + Li + Page 15

16 Electron Tunneling with Acid Generation Model Simulated Calendar Life Test ch. Neg. Li Stoic Acid generation model predicts lower voltages significantly reduce fade rate. EquilFade3-1-G Page 16

17 Electron Tunneling with Acid Generation Model Simulated Cycle Life Test 4.2 Volts Capacity 4.0 Volts Acid generation model predicts lower voltages significantly increase cycle life. EquilFade Page 17

18 Particle Expansion/Contraction Damages SEI R Expansion cracks SEI break = k I N sei I Q p x x x o is value at stoichiometry at last time current changed sign o Equation predicts capacity loss increases with charge/discharge current and magnitude of SOC swing Page 18

19 Simulated Cycle Life Acid Generation Models No acid generation Capacity Acid generation + SEI destruction due to SOC swings Acid generation Page 19

20 Possible Mechanisms for Site Loss Electrical isolation of particles due to Expansion of positive, recovery from compression or cycling Break-up of agglomerates Resistive film formation on surface Change in surface chemistry Phase change to insulating or material with low Li diffusivity Composition change, perhaps due to acidic impurities Active Binder C black Page 20

21 Site Loss Model 100 dq dt site k site Site Concn. 80 = 60 p site 40 Q site 20 0 k site =1.0 site k site =0.01 k site = time Expect site loss to accelerate as lose sites, since remaining sites are exercised more vigorously and have fewer neighbors to contact. Page 21

22 Site Loss + SEI Growth Models 100 dq sei = dt k Q sei sei Capacity time Capacity dq 60 site k = site p dt Q site 40 site Site Con ncn time time Page 22

23 100 Site Loss + SEI Growth Models Impedance Growth R = R + r N + cell o sei sei r Q site site 80 Imped dance time Page 23

24 Summary Defining feature of lithium-ion cells are the lithium sites and the amount of lithium available for cycling, which are determined d by electrode loadings, irreversible capacity loss on formation, and equilibrium curves. The most common shapes for capacity fade curves can be rationalized by simple models for lithium and site loss. Page 24

25 # Cy ycles Constant Ah Throughput Cell passes fixed amount of Ah (Ah life ) before reaching end of life 1.0E+06 10E E E+04 Basis: Cell Cap., Ah = 1.0 Ah-life = 10,000 N cycles = Ah Ah cell life DOD DOD, % Page 25

Capacity fade studies of Lithium Ion cells

Capacity fade studies of Lithium Ion cells by Branko N. Popov, P.Ramadass, Bala S. Haran, Ralph E. White Center for Electrochemical Engineering, Department of Chemical Engineering, University of South