Abstract
Ruddlesden-Popper (RP) chalcogenides are an emerging class of layered semiconductors with tunable properties and chemical stability, making them promising candidates for a wide range of functional applications. Over the past four decades, the structural diversity of RP oxides has been exploited to realize advanced functionalities; however, similar strategies have not yet been applied to RP chalcogenides, whose structural behavior remains poorly understood. In this study, we develop a high-accuracy machine-learned interatomic potential to perform large-scale molecular dynamics simulations of the homologous RP series Ba_{n+1}Zr_{n}S_{3n+1}. We identify new polymorphs for each n value, predict the corresponding phase transition temperatures, and validate our approach through comparison with existing experimental data. We find that the n=1 phase exhibits in-plane negative thermal expansion, that the n=1 and n=3 phases undergo unusual ascending symmetry breaking, and that phases with n≥3 develop layer-dependent tilt patterns not previously observed in inorganic RP compounds. This distinctive behavior arises from the interplay between ZrS_{6} octahedral rotations and BaS rumpling at the perovskite-rocksalt interface, suggesting new strategies for realizing advanced functionalities and tuning properties in RP chalcogenides.
Funding Information
  • Chalmers Tekniska Högskola
  • Engineering and Physical Sciences Research Council
  • European Cooperation in Science and Technology (CA21148)
  • European Cooperation in Science and Technology (EP/X035859)
  • European Cooperation in Science and Technology (EP/T022213)
  • European Cooperation in Science and Technology (EP/W032260)
  • European Cooperation in Science and Technology (EP/P020194)
  • Vetenskapsrådet (2020-04935)
  • Vetenskapsrådet (2021-05072)
  • Vetenskapsrådet (2022-06725)
  • Knut och Alice Wallenbergs Stiftelse (2024.0042)
  • Northumbria University
  • ARCHER Service
  • Renewable Energy Northeast Universities (EP/S023836/1)
  • UK National Supercomputing Service
  • National Academic Infrastructure for Supercomputing in Sweden

This publication has 50 references indexed in Scilit: