Molecular Dynamics Analysis of Lithium-Ion Transport Properties in All-Solid-State Lithium-Ion Battery

Abstract

Lithium-ion batteries have been widely used in electronic products since their commercialization, but the use of liquid electrolytes carries safety risks of electrolyte volatilization, leakage, and corrosion. Meanwhile, solid-state electrolytes have the advantage of high safety performance compared with liquid electrolytes. Therefore, all-solid-state Lithium-ion batteries has become a new research topic. LiNi x Co y Mn z O2 (NCM, x+y+z =1) material have been used as a representative ternary compound cathode material in Lithium-ion batteries. However, understanding the influencing conditions during the transport of Lithium ions in NCM materials is incomplete and difficult to observe experimentally. Therefore, our goal is to understand the factors influencing Lithium-ion transport properties. In this work, molecular dynamics (MD) simulations of NCM materials were used to analyze the transport properties of Lithium-ion.According to Wei's study, we created three patterns for transition metal alignment for LiNi0. 33Co0. 33Mn0. 33O2(NCM333). Then, MD simulations were performed for the three types of NCM333. The Lithium-ion transport properties are considered in different Lithium-ion content rate (100%, 88%, 77%, 66%, 55%, 44%, 33%, and 22%) which represent different state of charge (SOC). The occupied positions of Lithium-ion in the initial NCM structure were randomly constructed for different percentages of Lithium-ion content. According to Zhu's research, we used the Morse potential energy in the LAMMPS as the basis and added the corresponding Coulomb force and long-range repulsive interaction force. In the end, the potential model based on Morse-type interaction, Coulomb interaction and a long-range repulsive interaction were used. The size of the simulation box was 34.74 Å×29.72 Å×28.54 Å. The total number of atoms was 3456. The temperature of the system during the MD simulation was set at 300 K, the timestep was set at 1.0 fs, and the sampling interval for trajectories was 500 steps. To relax the system, a 0.5 ns NPT process was first applied. After 10 ns of NVT process was performed, the size of the simulation box at the final equilibrated state was compared with that of 100% Lithium-ion content. The results showed that the simulation box volume variation was less than 2% in all calculated Lithium-ion content. Radial distribution function (RDF) analysis was also performed using the same conditions, and the results were consistent with Lee's study. These two analyses confirmed the integrity of the structural NCM in the simulation.The mean-squared displacement (MSD) analysis was performed on the NCM333 material throughout the trajectories in the 10 ns NVT process. From the MSD data, all three patterns of NCM333 have very minimal Lithium-ion mean square displacement at 100% Lithium-ion content, and the values of Lithium-ion mean square displacement for each pattern were very close to each other, which was not only consistent with the structural properties of NCM333, but also with the Lee's study. In the simulation of Pattern 1 with different Lithium-ion contents, it can be seen that there was a more significant migration of lithium ions compared to 100% Lithium-ion content, which was also reflected in the mean square displacement data. The mean square displacement data at 88% Lithium-ion content was significantly smaller than those at 55% Lithium-ion content, but all rang of the mean square displacement data were much larger than those at 100% Lithium-ion content. Because at 100% Lithium-ion content, there was no vacancy around the Lithium-ion, so only vibrations occured. When the Lithium-ion content is less than 100%, vacancies appear, so the Lithium-ion occur in vibration and diffusion.The Lithium-ion diffusion coefficient was obtained by linear fitting the MSD data in the region of 2.5ns-5ns. The figure shows the Lithium-ion diffusion coefficients (DLi) of the NCM333 material for different Lithium-ion contents. As shown in the figure, the experimental value is 3.3x10-11cm2/s and the theoretical value of the Wei's study is 5x10-9cm2/s when the Lithium-ion content of NCM333 was 33%. Also the calculated value was approximately one hundredth times larger than the experimental, consistent with the Wei's study. The diffusion coefficient maximum peak exists around 50% Lithium-ion content, which was inconsistent with our prediction. Our predictions suggest that as the Lithium-ion content decreases, more vacancies will occur and therefore the DLi should increase. Further refinement and improvements will be considered to confirm these phenomena. Figure 1