2025/12/15 by Jordan Dorrell, Graeme M. Day · 1 voice
Chemistry · Computer Science · Materials Science · #Computational Drug Discovery Methods #Crystallography and molecular interactions #Machine Learning in Materials Science
paper · doi:10.26434/chemrxiv-2025-l4ftw
openalex publication_date 2025/12/15 · openalex created_date 2025/12/16 · openalex updated_date 2026/07/14
Effective crystal structure prediction (CSP) relies on thorough exploration of potential energy surfaces (PES). For this reason, molecular CSP has historically focussed on simple crystals with a single molecule in the asymmetric unit. Increasing the number of molecules in the asymmetric unit increases the degrees of freedom of the crystal structure, and the burden on the crystal structure predictor. We mitigate this burden by modifying quasi-random structure searching (QRSS) with "sensible asymmetric units for crystal exploration" (SAUCE). Compared to traditional QRSS implementations, the crystal structures generated by SAUCE are denser, and in a more energetically favourable environment, leading to faster geometry optimisations and more frequent recovery of the low energy minima discovered by traditional QRSS. This will allow for structure predictors to approach more complex molecular materials and to reduce the computational cost of molecular CSP.