Properties of Sulfide Solid Electrolytes Studied by Electronic Structure Calculations

Rechargeable all-solid-state batteries (ASSBs) are traded as next-generation power sources for mobile applications, because they are believed to provide increased energy densities, higher power densities and improved cyclability compared to conventional Li-ion batteries (LIBs).[1] Moreover, the replacement of flammable liquid organic electrolytes, used in LIBs, with non-flammable solid electrolytes (SEs) might eliminate safety issues and enables new battery designs.[2] In this regard, sulfide SEs are promising candidates because they show ionic conductivities of up to ≈10 mS/cm at room temperature and convince with favorably soft mechanical properties that enable an easy integration into the battery.[3-5] Their disadvantage, however, is a lack of electrochemical stability against most electrode materials.[3,6,7] Despite the huge effort to study sulfide SEs, however, many of the related processes, such as exact diffusion mechanisms or interface degradation reactions, have not been understood in detail. Such an understanding could offer new optimization strategies, and we have therefore used atomistically resolved density functional theory (DFT) calculations and ab-initio molecular dynamics (AIMD) simulation over the past years to investigate selected sulfide SEs. A detailed introduction into the topic is given in Chapter 1 and a literature review for the materials of interest in Chapter 2 will lay out the specific research questions tackled in this work. The applied methods and theoretical background are explained in Chapter 3 and lay the foundation for the following chapters. In Chapter 4 we will discuss the litium thiophosphate (LiPS) system that comprises multiple crystalline phases such as Li3PS4, Li7P3S11 and Li4P2S6. Structurally, the situation is further complicated by the coexistence of glass phases exhibiting an amorphous structure.[8] Hence, most sulfide SEs are actually glass-ceramics whose properties are determined by the types and amounts of the underlying phases. We strongly focus on glass phases as their structure is difficult to analyze by experiments. To this end, we generated structure models for LiPS glasses at various compositions by applying a computational melt-quenching approach and compare the stability, structure and Li+ transport properties of crystalline and glassy phases. We find that all glasses are metastable and exhibit similar Li+ diffusion coefficients despite the fact that they are comprised of different basic structural units (PS4^3-, P2S7^4-, P2S6^4-). Furthermore, the occurrence of unusual structural units is observed and the association of structural units via cross-linking S-S bonds is derived as compensation mechanism in case of local Li deficiency. Finally, the interfacial stability against Li metal and internal interfaces are investigated. In this regard, the usage of defect formation energies as descriptors to judge the stability of interfaces is discussed. Next, the quaternary, argyrodite-type system Li6PS5Br is analyzed in Chapter 5. The key question concerns the experimentally observed Br-/S2- site-exchange among its 4a and 4d sites, that can be controlled via the synthesis procedure without altering the composition:[9,10] How does the Br-/S2- site-exchange influence the structure and properties of the material? We will show that the ordered structure is the most stable configuration and that the lattice constants show a minimum at 50% site-exchange. The main part discusses the Li+ transport properties and how the introduction of Br-/S2- site-exchange enables the transition from local to long-range Li+ diffusion. Moreover, we were able to identify the underlying diffusion mechanism and show that especially the Br- ions on S2- sites facilitate the generation of Li+ Frenkel pairs with mobile Li+ interstitials. Finally, we have a closer look on the Li+ substructure and analyze how the Br-/S2- site-exchange interacts with the Li+ transport properties of symmetrical tilt and twist grain boundaries. In Chapter 6 we will deal with the recently developed Li7SiPS8, which was found to crystallize in an orthorhombic phase (ortho-Li7SiPS8) with rather poor Li+ transport properties and a more promising tetragonal phase (tetra-Li7SiPS8).[11] As not much is known about the material we examined several of its properties, also in light of the Si/P disorder that is observed experimentally. We show that ortho-Li7SiPS8 is the more stable phase and experimental trend of poor transport properties is confirmed. Tetra-Li7SiPS8 is the much better conductor owing to its fast diffusion along the c axis. The Si/P distribution was found to have a negligible influence on the transport properties, and a compression of the material leads to decreased diffusion coefficients. Finally, the interfacial instability of tetra-Li7SiPS8 against Li metal was probed by means of explicit interface calculations. At long last, we will conclude this work in Chapter 7 and present open questions and promising directions for future studies