Machine-Learning Aided Screening of Organic-Inorganic Perovskites as Efficient Photoabsorbers
Abstract
This project is a multi-pronged computationally driven approach combining machine learning, molecular dynamics, and first-principles calculations to discover new two-dimensional hybrid organic–inorganic halide perovskites (2DHPs) for application as the active materials in optoelectronic devices. 2DHPs are gaining interest because of their superior chemical stability, structural diversity, and broadly tunable properties stemming from the possibility of combining an inorganic semiconductor comprised of a 2D connected network of an inorganic cation (B) with a counter anion (X, which is typically Cl, Br, or I for materials that absorb/emit in the UV or visible spectrum), separated spatially by targeted organic components. However, because of the immense space that typically needs to be explored, these materials have not yet been exhaustively charted, with only about 180 organic cations having been investigated to date. This research activity will create and leverage machine learning (ML) techniques for more effective computational predictions. Because of the presence of ions in the organic layer, a particular focus of this work will be the incorporation of long-range Coulombic interactions in ML-based force fields in order to predict reliably the solid-state structures. A key outcome of this work will be a comprehensive database of computationally predicted 2DHPs to facilitate the accelerated development of these materials by the wider community
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 21, 2022
- Source ID
- FA86552117010XX0
Entities
People
- Gábor Csányi
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- University of Cambridge