The Cation-anion Coupled Dynamics In The Superionic Conductors And High Performance Computing

Superionic state is an intermediate state between solid state and liquid state and has been widely researched in solid-state lithium batteries,fuel cells,sensors,astrophysics and geology.One of the key materials in next generation battery is solid state electrolyte which owns superionic state or high ionic conductivity during applicated temperature range.In order to meet the increasing requirements for higher energy density and safety in electric vehicle,consumer electronic and energy storage systems,the development di-rection of battery is to replace the liquid electrolyte by solid state electrolyte and make it operate with Li or Si anode in order to promote the battery performance.However,due to the lower ionic conductivity in solid than liquid electrolyte and the complexity of ionic transport and dynamics properties between solid electrolyte and electrode,one of the key tasks for next generate solid lithium battery is understanding the ionic dif-fusion mechanism and seeking the suitable fast ion conductors with ionic conductivity compared with liquid electrolyte.This thesis aims to explain the physical nature of cation-anion dynamics coupling mechanism.The author selected two materials which own the same atom groups while different dynamics properties to clarify the origin of two contradictory mechanisms and confirmed the conditions for different dynamics coupled mechanisms by high-throughput calculations.Based on it,the author developed a novel method assisted by natural language process to reveal a superionic material with cation-anion dynamics coupling mechanism.In addition,the author designed and optimized the calculate flow for search-ing fast ion conductors.In order to clarify the contradictory at first,the author selected two similar com-pounds with the same anion group（Li BF₄and Li₂BF₅）while they present different dy-namics difference due to the different configurations.The author analyzed the changes of BF₄^-accompanied with the diffusion of Li⁺by ab initio molecular dynamics and data mining and found that in Li₂BF₅with the shorter Li-Li distance the Li hops ac-companying with only little perturbation while in Li BF₄with longer Li-Li distance the Li hops with the help of the reorientation of BF₄^-.The author concluded three different modes for anion groups:local vibration,rotation and migration and presented harmonic oscillator model and potential energy surface model to explain the contradictory:low experiment accuracy and lack of the clear definition for rotation.After understanding the atomic mechanism in cation-anion dynamics coupling mechanism,it is possible to screen and search the fast ion conductors with same char-acter.Besides searching fast ion conductors in known structures from databases,the author tried to identify new structure assisted by natural language processing and an-alyzed the relevance between superionic and other properties.The author excavated a new superionic material from the synthesis literature.LiY（SO₄）₂was first reported as a nonlinear optics materials while its synthesis method and experiment results imply its superionic character.The author also analyzed the changes of SO₄^2-during the diffu-sion of Li⁺and found that the reorientation of SO₄^2-can reduce the nearest neighbour number of Li-O.In addition,the author also found Cu₄TiTe₄as a superionic material from thermoelectricity literature and analyzed two different local environments for Cu⁺.Based on above two examples,the author explained cation-anion dynamics coupling mechanism from the view of bond.These examples not only present a new method to find superionic materials but also deepen the understanding of ionic diffusion mecha-nism especially for cation-anion dynamics coupling mechanism.The author explained the ionic diffusion from the local environments and the changes of bond rather than restricted to rotation.Superionic materials or fast ion conductors mainly exist at high temperature and high pressure and the experiments cannot observe the diffusion mechanism in situ.It is necessary to develop high performance computing to research the ionic diffusion mech-anisms.The author analyzed the character of origin calculation especially grid sampling in BV method and finite element method and improved the calculation efficiency and identified the diffusion path of Li⁺using image recognition.The author also tested it with different software and hardware and promoted it one time faster.At last,the author discussed the methods to search fast ion conductors and the problems in calculating the properties of fast ion conductors.