Microscopic insights into glassy solid electrolytes

最近，科学界在SOLBAT项目的重大贡献下，在理解固体电解质中金属锂“枝晶”形成和传播的起源方面取得了重大进展[Ning等人，2021]。所有证据都指向Li-SSE电解质界面[McDowell等人，2019]和结晶固体中晶界[Sakamoto等人，2017]的反应性中间相的有害作用。玻（非结晶）固态电解质（SE）由于其相对于结晶对应物的优点而吸引了很多关注：各向同性离子传导、大的可用组成尺度、热历史操纵、对传导率的弱电子贡献、易于制造成膜以及最重要的是，没有晶界和相关电阻。一些最有前途的SE具有玻璃或玻璃陶瓷微结构，包括LiPON [Dudney et. 2018]和硫代磷酸盐[Wang et al. 2021]。尽管如此，Li的原子结构、微观结构、离子和电子导电性以及电化学机械性能之间的相互作用仍然存在。|SE间期仍然难以捉摸。因此，科学界仍在努力设计具有高锂离子电导率的SE，能够在商业相关的电流密度下防止枝晶形成，并可大规模加工成薄膜。在这个项目中，学生将结合联合收割机先进的实验和机器学习驱动的模拟，探索固体玻璃电解质的合成和表征特性（结构，物理化学，电化学和机械）。卤氧化物玻璃将用作模型系统，因为其具有良好的离子导电性和成分可调性[Goodenough等人，2016年; Lunz等人，2019年]。在第一部分中，结构和机制的离子传导在散装电解质将探索实验（X射线PDF，固态NMR和阻抗谱）和计算，使用ML为基础的方法。在第二部分中，利用XPS，AFM和先进的电子显微镜表征（包括冷冻FIB横截面和低剂量TEM）的现有专业知识，将探测Li-SE界面。将这些实验与ML驱动的10纳米长度尺度的模拟相结合，该项目有望对组成和形态对机械性能和离子传输的影响有新的认识。Pasta小组最近研究了晶体结构和微观结构对Li 2 OHX离子导电性的影响（X=Cl，Br）固体电解质，并在熔融-固化过程中开发了独特的专业知识来控制晶粒尺寸[Pasta et. 2021]。熔融盐的快速淬火将产生玻璃，然后将在结构上（与钻石光源的光束线科学家Maria Diaz-Lopez博士合作的X射线PDF），机械和电化学上进行表征。Deringer团队专注于开发和应用基于机器学习的力场，用于模拟结构复杂的材料[Deringer et al. 2021]。实验-计算相结合的方法将首先在定义良好的模型系统上进行验证，即卤化锂反钙钛矿（对于其熔化淬火可以用建立的模拟协议直接描述）-但是更雄心勃勃的长期-术语目标是计算以更一般地支持和加速材料选择和处理条件工作流，这是一个为期4年的法拉第研究所学生项目，（部分课程费用由牛津材料基金支付）

Recently the scientific community, with significant contribution from the SOLBAT project, has made great progress in understanding the origins of metallic lithium "dendrite" formation and propagation in solid electrolytes [Ning et al. 2021]. All evidence points toward the detrimental role of reactive interphases at the Li-SSE electrolyte interface [McDowell et al. 2019] and the grain boundaries in crystalline solids [Sakamoto et al. 2017]. Glassy (non-crystalline) solid-state electrolytes (SE) have attracted much attention due to advantages over their crystalline counterparts: isotropic ionic conduction, large available composition scales, thermal history manipulations, weak electronic contribution to the conductivity, ease of fabrication into films and, most importantly, absence of grain-boundaries and associated resistance.Some of the most promising SEs have glass or glass-ceramic microstructures, including LiPON [Dudney et. al. 2018] and thiophosphates [Wang et al. 2021]. Nevertheless, the interplay between atomic structure, microstructure, ionic and electronic conductivity, and electro-chemo-mechanical properties of the Li|SE interphase remains elusive. Therefore, the scientific community is still struggling to design SEs with high Li-ion conductivity, able to prevent dendrite formation at commercially relevant current densities and processable as thin film at scale.In this project, the student will combine advanced experiments and machine-learning-driven simulations to explore the synthesis and characterization properties (structural, physico-chemical, electrochemical and mechanical) of solid glass electrolytes. Oxy-halide glasses will be used as a model system because of their promising ionic conductivity and compositional tunability [Goodenough et al. 2016, Lunz et al. 2019]. In the first part, structure and mechanism of ion conduction in the bulk electrolyte will be explored experimentally (X-ray PDF, solid-state NMR and impedance spectroscopy) and computationally, using ML-based methodology. In the second part, leveraging existing expertise in XPS, AFM and advanced electron microscopy characterisation (including cryo-FIB cross sectioning and low-dose TEM), the Li-SE interphase will be probed. Combining these experiments with ML-driven simulations on the ten-nanometre length scale, the project promises new insight into the effect of composition and morphology on mechanical properties and ion transport.The Pasta group has recently investigated the effect of crystal structure and microstructure on ionic conduction in Li2OHX (X=Cl, Br) solid electrolytes and has developed a unique expertise in melting-solidification processes to control grain size [Pasta et. al. 2021]. Rapid quenching of the molten salt will produce glasses that will then be characterized structurally (X-ray PDF in collaboration with Dr. Maria Diaz-Lopez, beamline scientist at Diamond Light Source), mechanically, and electrochemically. The Deringer group specialises in the development and application of machine-learning-based force fields for simulations of structurally complex materials [Deringer et al. 2021].The combined experimental-computational approach will be first validated on a well-defined model system, viz. lithium halide antiperovskites (for which the melt-quenching can be directly described with established simulation protocols) - but the more ambitious long-term aim is for computation to support and accelerate the material selection and processing conditions workflows more generally, ultimately leading to a combined approach for the discovery of glassy solid-state electrolytes.This is a 4-year Faraday Institution Studentship (part of the course fee paid from Oxford Materials funds)