Speaker
Yoshitaka Tateyama
National Institute for Materials Science,Japan
Date＆Time
2023.08.03（Thur.）AM 9:00
Location
吉林大学中心校区唐敖庆楼B区521报告厅
Reporter
Professor Yoshitaka Tateyama is the director of Research Center for Energy and Environmental Materials (GREEN) in National Institute for Materials Science (NIMS) in Japan. He received Ph.D from the University of Tokyo in 1998. 3. He was a visiting researcher hosted by Prof. Michiel Sprik in the department of Theoretical Chemistry, University of Cambridge during 2003-2004. He joined the NIMS as a researcher since 2001.4. He is also a visiting professor in Waseda University and Tokyo Institute of Technology. Prof. Tateyama is working on the theoretical/computational/data-driven AI researches on microscopic mechanisms of energy and environmental issues associated with battery, solar cell and catalyst.
Abstract
Development of solid electrolyte (SE) with high ionic conductivity has been a most important target for the large-scale application of solid-state battery. Although several pristine SEs with high conductivities have been found so far, the grain boundary (GB) effect is still an open question. As SE usually has high ion concentration, understanding of ion-ion correlation is also indispensable to improve the conductivity. However, the experimental observations of these microscopic phenomena are still difficult. In this work, using DFT-based molecular dynamics (DFT-MD) with a sufficient predictability, we have addressed the elucidation of microscopic ion behaviors in the SEs.
First targets are the electronic states and the ionic conductivities around the GBs of the garnet-type Li7La3Zr2O12 SE as well as the dopant effects (Al, Ga, Nb, Ta). The present DFT-MD simulations provided the interesting results: (1) Some GBs do not lower the ionic conductivities, (2) certain Li-Li correlation contributes to the conductivity, (3) the dopants segregated to the GB regions may have positive effect on the GB ionic conductivity. Besides, the electronic states calculations indicated the probable dendrite growth mechanism through the GBs.
In most MD studies including our previous works, self-diffusion has been treated with the Nernst-Einstein approximation. Meanwhile, the conductivity diffusion with ion-ion distinct correlation is more substantial in practice. However, the corresponding correlated ionic conductivity is difficult to be calculated because of the slow convergence of sampling. To mitigate this problem, we developed a new type of non-equilibrium MD method. The method generally makes the nine-times speed-up of the sampling, enabling the correlated ionic conductivity calculation with reasonable accuracy and cost. Applying to Li10GeP2S12, we successfully reproduced the conductivity as well as the Haven ratio, a measure of the ion correlation.
