Mechanical\nand Electronic Properties of Bulk and Surface\nLi<sub>6</sub>PS<sub>5</sub>Cl Argyrodite: First-Principles Insights\non Li-Filament Resistance

Different Li-filament growth patterns\nhave been experimentally\nobserved in numerous solid electrolytes (SEs) with high ionic conductivity\nsuch as garnet Li<sub>7</sub>La<sub>3</sub>Zr<sub>2</sub>O<sub>12</sub> (LLZO) and argyrodite Li<sub>6</sub>PS<sub>5</sub>Cl (LPSC). Herein,\nwe probed the mechanical and electronic properties of LPSC, using\ndensity functional theory calculations, and compared with other SEs\nto determine the relevant descriptors for predicting Li-filament resistance.\nLPSC has a complicated structure that can incorporate S<sup>2–</sup>/Cl<sup>–</sup> inversion and has Li<sup>+</sup> distributed\namong two Wyckoff sites (24g and 48h). A representative bulk structure\nthat incorporates both phenomena was determined via systematic structure\nsampling. The lowest energy bulk structures had a majority of Li<sup>+</sup> in 48h sites after relaxation, agreeing with experimental\nstudies. The Young’s modulus and shear modulus of bulk LPSC\nare low, ∼10–30 GPa, and the fracture energy of cleaving\nalong the (100)-Li<sub>2</sub>S-deficient surface is also low, 0.20\nJ/m<sup>2</sup>, suggesting poor mechanical resistance to filament\ngrowth. The crack surfaces and pore surfaces in LPSC have a similar\nbandgap and excess electron distribution compared to bulk LPSC, suggesting\nthat these internal defects will not trap electrons to reduce Li<sup>+</sup> to Li-metal. Thus, LPSC is likely to experience “dry”\ncracks, with a mechanical crack opening up first, followed by a Li-filament\nfilling the crack. This is opposite to LLZO, which has a high fracture\nenergy and experiences electron localization at internal defects (e.g.,\ncrack surfaces, pore surfaces, and grain boundaries). LLZO has been\nexperimentally observed to suffer “wet” cracks.