Assessing crystallisation behaviour in molecular crystals through particle rugosities.

Surface properties of molecular crystals play a central role in determining their nucleation, growth, and overall crystallisation behaviour. In particular, surface rugosity has been suggested as a meaningful descriptor of how readily a crystal may nucleate and grow, but the existing rugosity metrics have been limited by incomplete descriptions of surface topology. In this study, we present a new workflow for the computation of surface descriptors across crystal facets. Here the surface rugosity description for each facet is based on the Surface Area Ratio (SAR) definition proposed in the Cambridge Structural Database (CSD) tools. Given a crystal structure, our algorithm computes an average overall particle rugosity with three offset selection criteria that account for different crystallisation conditions. We then calculate particle rugosities and analyse data for a range of systems, including: polymorphic families, datasets from the CSD, and crystal structure prediction (CSP) landscapes. Our analysis shows that, in most cases, polymorphs with lower rugosities tend to nucleate and grow more readily, suggesting that this metric can distinguish experimentally accessible forms from more elusive ones. These findings demonstrate that particle rugosities can serve as a complementary tool for predicting and classifying the experimental feasibility of polymorphs generated computationally.