Dresden 2026 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

CPP: Fachverband Chemische Physik und Polymerphysik

CPP 9: French-German Session: Simulation Methods and Modeling of Soft Matter I

CPP 9.7: Talk

Monday, March 9, 2026, 16:45–17:00, ZEU/LICH

NMR crystallography at finite temperatures — •Matthias Kellner and Michele Ceriotti — Laboratory of Computational Science and Modeling, Institut des Matériaux, École Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland

In this talk I will present our latest developments in machine-learning models for predicting NMR chemical shieldings in organic molecular solids. NMR shielding-driven structure determination protocols rely on comparing experimental solid-state NMR shifts with simulations whose accuracy is often limited by computational cost. We address these limitations by combining ensembles of transformer-based shielding predictors with molecular dynamics simulations using a novel, broadly applicable machine-learning interatomic potential. This framework systematically improves agreement with experiment across many benchmark systems, removes the need for empirical corrections and makes it applicable also to disordered and amorphous materials. I will conclude by showing applications of this framework to the structure determination of amorphous active pharmaceutical ingredients.

Keywords: Machine Learning; NMR; Molecular Dynamics