Study On Coatings Of Lithium Metal Anodes By Using Quantum Mechanics Method And Machine Learning Force Field

Lithium(Li)metal is considered as a promising anode material for high energy lithium batteries due to its high theoretical capacity(3860 mAhg-1)and low redox potential(-3.04 V vs.the standard hydrogen electrode).However,the lithium dendritic growth during the cycle of lithium metal batteries may cause serious security risks,thus hinder the application of lithium metal anodes.Building suitable coatings on the surface of lithium metal anode is an effective solution to the issue with regard to dendrite growth.The coatings were found to guide the uniform deposition of lithium ions onto the anode surface,and therefore inhibit the growth of dendrite.However,there are still many difficulties in the rational designs of coating materials,which may because of the insufficient understanding of the underlying mechanisms.The research methods based on microscopic simulations are expected to provide an effective way to understand the mechanism of coating materials inhibiting the growth of lithium dendrite and guide the design of coatings.This thesis focuses on two typical anode coating materials:zinc oxide and lithium-silicon alloy(LixSi),using quantum mechanics(QM)method and machine learning force field(MLFF).The adsorption energies of Li on the zinc oxide surface and on the copper collector surface were calculated.The results show that the adsorption energy on the zinc oxide surface is higher than that on the copper collector surface,indicating that the zinc oxide coating may guide the deposition of lithium ions onto the anode surface uniformly and inhibit the growth of lithium dendrites,thus significantly improving the cycle life of metal lithium batteries.To develop an MLFF for molecular dynamics simulation of lithium depositing on the coating surface,this thesis explored the construction and optimization of the data set of lithiumsilicon(Li-Si)systems.The construction of data sets follows a strategy of "simple to complex,from bulk to interface".Three key points in the development of MLFF are explored as follows:(1)The construction of bulk data set.For the Li-Si bulk system,the melting-annealing simulation method is used to construct an MLFF suitable for simulating crystalline and amorphous Li-Si alloy bulk system.The accuracy of the calculated potential energy and dynamic properties are consistent with QM results.The MD simulation based on this force field can successfully describe the bulk density for the amorphous lithium-silicon system and is consistent with the experimental results.The method of collecting structural features of the binary system based on the phase diagram used in this process can also provide a useful reference for constructing MLFF of a similar system.(2)The simplification and optimization of bulk data set.Aiming at the problem that there may be a large number of redundant structures in the Li-Si bulk data set,the original data set is simplified using the structure similarity calculation and farthest point sampling strategy.The number of samples required for the lithium-silicon alloy bulk data set is reduced by 93.5%,while retaining high accuracy.Similar simplification method can be easily extended to other MLFF development process,which significantly improves the development efficiency of MLFF and reduces the development costs.(3)The construction and optimization of surface/interface systems data set.Based on the reduced bulk data set,more interface configurations are supplemented,including the surface configurations of the Li-Si alloy and the trajectories of the Li deposition process onto different silicon surfaces.The configuration distribution of different phases is optimized to construct a high-precision MLFF which is able to describe Li-Si bulk phase,surface and interface characteristics.The force field is used to simulate the deposition and diffusion process of lithium on the silicon surface,and the two-phase interface lithiation phenomenon and the density distribution of the lithium-silicon alloy layer are found to be consistent with the experimental results.This laid the foundation for the subsequent in-depth study of the influence of the Li-Si alloy coating on the lithium deposition process and the regulation mechanism.This thesis uses the QM method and the MLFF to investigate the properties of two coating materials,explains the regulation mechanism of zinc oxide on the lithium deposition process,and develops a MLFF that can be used to simulate the deposition and diffusion of lithium on the surface of the Li-Si coating.The effective combination of these two methods can provide a theoretical basis for the rational design of lithium metal anode coating materials.In addition,the MLFF construction research involved in this study also provides a useful reference for the development and optimization of MLFF for other similar surface/interface systems.