Kinetic Monte Carlo simulations applied to the study of LIB graphite anodes

GAVILÁN-ARRIAZU, E.M.; D.E. BARRACO; E.P.M. LEIVA

Antofagasta

Workshop; 7 °International Workshop on Lithium, Industrial (IWLIME); 2020

Universidad de Antofagasta-Center for Advanced Research on Lithium an Industrial Minerals-U. de Córdoba-U. Mayor de San Andrés,-U. Católica Boliviana, INIFTA (UNLP-CONICET)

Graphite is still the material mainly used as anode for Li-ion batteries (LIB), due to its considerablecapacity combined with a voltage close to Li+ / Li electrode, abundance and low cost as principalcharacteristics. Even so, there are phenomena that are not fully understood yet. This prevents theproper design of the electrodes to overcome certain technical difficulties, such as the impossibilityof operating at high current densities. Experimental and computational studies can shed light onthese problematics.The formation of structures called ?stages? upon Li-ion (de)intercalation into(from) graphite iswell known. These stages are periodic structure of Li ion planes repeating after a certain numberof graphite sheets. Two models have been proposed to explain the Li ion distribution on the stages:Ruddorf-Hoffman and Daumas-Hérold models. The (de)intercalation of Li ion into(from) graphite has demonstrated to be a very slow process. Aregular cyclic voltammetry experiment, for example, uses sweep rates in the order ofmicrovolts/second to observe well defined peak currents (several hours or even days to perform acomplete cycle). This is a problem when using simulation methods like molecular dynamics,because simulating times up to microseconds involves high computational efforts. Kinetic MonteCarlo (kMC) surmounts this problem neglecting the vibrations of ions / atoms, allowing to extendthe simulation time to the order of seconds.In this presentation we show the most relevant results that we have obtained simulating the(de)intercalating process of Li ion into graphite using kMC and standard Monte Carlo (MC), tocompare kinetic results with those at equilibrium. For example, we have discovered with kMC thatthe DH model has apparently a kinetic origin related with the slow exchange rate of Li ion at thegraphite/solution interphase. On the contrary, RH structures were obtained with MC (equilibrium)simulations. This suggest that DH structures are out of thermodynamic equilibrium. Furthermore,we could simulate the results obtained with electrochemical techniques of common use in the lab.This is the case of potentiostatic steps. With this kind of simulations, we could explain, in detail,the reason why intercalation is slower than deintercalation, as observed on literature [1]: Duringintercalation, lithium accumulates at the interface when a potential step is applied. Thisphenomenon is not observed during deintercalation (Figure 1). We could also predict too that thehigh exchange current density observed at occupation degrees next to 0.5 (stage II) [2] is due tothe formation of DH structures, because perfect RH structures would lead to low exchange currentdensity values.