Sulfur - from hell to powering your car M. Marinescu EFL Energy Seminar June 7th , 2016 Motivation We need better batteries Emission reduction policies ⇒ Decrease reliance on carbon fuels • De-carbonise transport ⇒ xEV • De-carbonise electrical power grids Better = • cheap & long-lasting • energy & power dense • light, small • safe • versatile Contestabile, Offer, North, Electric vehicles: a synthesis of the current literature with a focus on economic and environmental viability, 2012 Battery requirements > 95% Li-ion manufacture for consumer electronics Standard EV Cell capacity (Ah) 4 40 Energy density (Wh/kg) 150 150 Life expect (yrs) 3 15 Cell cost ($/kWh) <250 400 United States Advanced Battery Consortium energy storage system goals, http://www.uscar.org, 2016 From cell to pack ? 35Wh −→ 30 kWh • structural support − mechanical vibrations, crash test • thermal management − large cells + tight pack of many • balanced electrically - many cells Cost of battery pack = double the cost of cells ($800/kWh, $21000/30kWh) Attributes for high energy density 1. Large specific capacity of electrodes Amount of charge they store 2. High Electropotential Potential difference between anode and cathode 3. Large stoichiometric coefficient # e- per mole of reactants 4. Low molar mass of active species Avenues • Reversibility! • Mass and volume of non-active materials! Oxis Energy Ltd Avenues P. Bruce, Nat Mat, 2012 Conversion vs intercalation batteries Intercalation - limited by: • active material intercalates reversibly • Li: 1e- , 4V cell (th): 300mAh/g=1200 Wh/kg Conversion • active material changes phase reversibly • many eType Example Vcell Intercalation (now) LMO 3.8 Intercalation (th) Wh/kg (th) Wh/l (th) 550 Wh/kg (max prac) 110-165 1200 240-300 S LiS 2.2 2567 2200 300-800 M-air LiO 3.0 3505 3430 700-1000 Reproduced from Cluzel, Douglas, Cost and performance of EV batteries, Element Energy report for The comittee on climate change, 2012 Why LiS • gravimetric energy density 3 • safety (*) 3 • cheap 3 • low temperature performance 3 http://www.shorpy.com/ • power density (discharge rate) 7 • cycle life 7 • volumetric energy density 7 * Hunt, Patel, Szczygielski, Kabacik, Offer, Journal of Energy Storage 2, 2015 LiS in use www.airbusgroup.com/int/en/story-overview/ http://www.oxisenergy.com/applications/ • lightweight • safe • low operating T https://protonex.com/blog/what-do-soldiers-carry-and-whats- LiS operation Discharge LiS operation Charge LiS operation Effect of current LiS operation Effect of current LiS challenges • improve rate capacity • improve cycle life - capture poly-S in cathode - ensure homogeneous Li plating on anode - design cathode microstructure • improve SOC estimation Manthiram et al, Acc Chem Res 46, 2012 Improving performance and use of LiS through modelling REVB: Revolutionary Electric Vehicle Battery • Target: 400Wh/kg LiS cells • Physics-based models & reduced-order models • Battery pack design, drive-cycle testing Funded by EPSRC and Innovate UK 0D model formulation S8 0 + 4 e− ←−→ 2 S4 2− S4 2− + 4 e− ←−→ 2 S2− ↓ + S2 2− 0D model formulation I = iH + iL IL V = (E + η)cathode −(E + η)anode − | {z } | {z } κA Vcathode Vanode • species formed/destroyed by e-chemical reactions • shuttle = S80 → S42− without e- • precipitation model with nucleation • Nernst equations: equilibrium potential of reactions • Butler Volmer approximation: reaction currents dSn dt Species evolution during operation • Can track charge status of battery • Qualitative charge/discharge voltage curves retrieved without need of differing mechanism Precipitation during discharge 2.4 EL = EL0 + RT 4F ln S 2− cL (S 2−4)2 S 2− 2 Potential (V) 2.35 [S 2(g), k * 2.3 p (s -1 )] = no ppt 0.0001, 100 0.0050, 100 0.0001, 500 0.0050, 500 2.25 2.2 2.15 2.1 0 0.5 1 1.5 2 2.5 Discharge capacity (Ah) Flat voltage plateau during discharge is caused by precipitation. Zhang, Marinescu, O’Neill, Wild, Offer, Phys. Chem. Chem, Phys. 17, 2015 Marinescu, Zhang, Offer, Phys. Chem. Chem. Phys 18, 2016 3 3.5 Charging rate Limited by shuttle and precipitation • fast charge ⇒ ’slow’ dissolution • slow charge ⇒ shuttle Limitations of 0D model No mass transport ⇒ mass transport in 1D model 1D model formulation • PDE’s species conc • flux of species mass • diffusion and migration Kumaresan, Mikhaylik, White, J. Electrochem. Soc. 155, 2008 Ghaznavi & Chen, J. Power Sources 257, 2014 1D model predictions Discharge rate retrieved for slow Li+ diffusion T. Zhang, M. Marinescu, S. Walus, G. J. Offer, submitted, 2016 1D model predictions Capacity recovery retrieved for slow Li+ diffusion T. Zhang, M. Marinescu, S. Walus, G. J. Offer, submitted, 2016 What we learnt from modelling Models show • precipitation/dissolution limits charge rate • transport through separator limits discharge rate • ’recent history’ effect: • accumulated precipitate • concentration gradients in electrolyte • crucial for SOC estimation Current work: • degradation: irreversible vs reversible loss • thermal coupling Predictions Time to market Chemistry 1st paper/patent 1st commercial cell 1st use series car LCO 1979 1991 2008 (Tesla) LMO 1983 1996 2009 (iMieV) LFP 1994 2006 2007 (MODEC van) Ni-MH 1967 1990 1997 (Prius) Reproduced from Cluzel, Douglas, Cost and performance of EV batteries, Element Energy report for The comittee on climate change, 2012 Conclusions Chemistry in car in 10 yrs? • Companies scaling up now • Only incremental increase When step change increase? • Current status = Nature paper • Step change → 20 yrs Application targeted chemistry? • high energy vs high power • numbers are driven by behaviour Extra slides 0D model formulation Species evolution dS80 dt = −nS8 aiH − ks S80 dS42− dt = nS8 aiH + (1 − fl dS22− dt = nS2 aiL 1 kp Sp S 2− − S∗2− = 2nS aiL − vρS 1 = kp Sp S 2− − S∗2− vρS MS8 = ks S80 , a=− ne F dS 2− dt dSp dt dSs dt Ss )ks S80 − nS4 aiL mS 0D model formulation Reaction equilibrium and kinetics S8 0 + 4 e− ←−→ 2 S4 2− S4 2− + 4 e− ←−→ 2 S2− ↓ + S2 2− Nernst equation for equilibrium potential ! RT S80 0 EH = EH + ln cH 2− 4F (S4 )2 ! RT S42− 0 EL = EL + ln cL 2− 2 2− 4F (S ) S2 Butler-Volmer approximation for reaction current ne F ηH , 2RT ne F ηL iL = 2iL,0 ar sinh . 2RT iH = 2iH,0 ar sinh ηH = V − EH ηL = V − EL I = iH + iL V = (E + η)cathode −(E + η)anode − | {z } | {z } Vcathode Vanode IL κA Definitions from 0D model n2S4 MS8 v , nS8 2 v2 n2 nS2 MS8 cL = S , nS4 cH = (1a) (1b)