Computational Modeling of Electrolyte/Cathode Interface for Li-Ion Batteries
Spinel dissolution and cathodic capacity losses in metal-oxide cells as a function of varying electrolyte solution setting, comprising different mixtures of solvents, Li salts and additives, is a major area of current research. It has been shown that spinel dissolution is induced by acids resulting from electrochemical oxidation of solvent molecules on composite cathodes. The widely utilized carbonate based solvents are considered to be relatively inert, whereas other related solvents such as various ethers are readily oxidized to produce acids. Experimental charge/discharge cycling of spinel-loaded composite cathodes found the acid concentration and the extent of spinel dissolution to be significantly higher in ether-containing electrolytes in comparison to their carbonate based counterparts. Further experimental evidence indicates that Li and transition-metal ion extraction is facilitated by an acid induced degradation of the cathode material, where continued spinel dissolution leads to oxygen loss from the Li-metal-oxide lattice. The nature of added Li salts largely influences solvent oxidation and spinel dissolution. Although solvent induced acid production for electrolytes containing fluorinated salts is not significant, the appreciable spinel dissolution in these cathode/electrolyte systems strongly suggests that acid is generated through various reaction pathways.
A thorough understanding of the reactions on the electrode surface of lithium batteries is central to the designing of new electrode interface material components to achieve efficiency and durability in high-voltage settings. In the present computational modeling work, we focus on the high-voltage lithium nickel manganese spinel (LiNi0.5Mn1.5O4) cathode material, where an experimental investigation of the voltage-dependent electrochemical reactions of a 1-M LiPF6/EC:DMC:DEC electrolyte on a LiNi0.5Mn1.5O4-based electrode, through characterization of the surface species by XPS and FTIR-ATR, showed that the increase in current flow coincides with changes to the surface chemistry of the cathodes and displayed a clear trend of increasing polyethylenecarbonate formation with increasing voltage. In order to provide predictive understanding in complement to experimental work, we performed high-level Quantum Chemical (QC) calculations and DFT-based Born-Oppenheimer Molecular Dynamics (BOMD) simulations on a series of non-aqueous, mainly carbonate-based electrolytes featuring a combination of solvents, salts, electrolyte additives, and cathode materials.
Initial BOMD calculations of molecule/surface systems (EC, DEC, FEC, DMC, PF6-, HFIB) served as a guide for subsequent unconstrained BOMD simulations of electrolyte mixtures on the delithiated LiNi0.5Mn1.5O4  and  surfaces to capture spontaneous electrochemical reactions. Nudged Elastic Band and Path Minimization calculations were employed to study the chemisorption characteristics of each electrolyte, such as the internal bond-breaking/bond-making evolution of the electrolyte and the electrolyte coordination to surface metal ions. Then, DFT potential-of-mean-force (PMF) simulations were conducted on explicit liquid electrolyte/electrode interfaces at finite temperature to investigate the previously determined pathways. From the PMF calculations, potential key reaction steps such as the transfer of protons and the lowering of the reaction barrier in the explicit liquid environment were determined. In particular, the two mechanisms of main interest in regard to interfacial electrolyte/cathode chemistry studied in the simulations is the retrieval of an oxygen ion from the cathode surface by an oxidized electrolyte molecular fragment and the proton transfer from the carbonate electrolyte to the metal-oxide. Our DFT based molecular dynamics simulations conducted at liquid EC/cathode interfaces are consistent with the view that reactions and electron transfer occur at the electrode interface.