Theoretical chemistry articles from across Nature Portfolio

Existing Moiré materials are mostly van der Waals heterostructures. Here the authors show that hydrogen-bond adaptability allows spontaneous formation of twisted bilayer ice at magic angles in 2D confinement, establishing a new class of Moiré materials.

Long-range interactions are challenging for machine learning interatomic potentials (MLIPs). Here, authors show that, by just learning from energies and forces, MLIPs can accurately capture electrostatics and predict atomic charges.

The scarcity of experimental training data limits machine learning applications in catalysis research. Here, the authors demonstrate that graph convolutional network models pretrained on a molecular topological index significantly enhance the prediction of catalytic activity, showcasing a promising transfer-learning strategy that leverages self-generated virtual molecular databases.

Catalytic descriptors are crucial to accelerating catalyst design. Here, the authors develop an equivariant graph neural network to enable robust structure representations and achieve accurate predictions of descriptors across complex catalytic systems.

Atomistic simulations are important for phase-change materials and devices. Here, the authors present fast and accurate machine-learned potentials, enabling full-cycle device-scale simulations and showcasing applications in studying memory and neuromorphic computing devices.

This issue of Nature Computational Science features a Focus that highlights both the promises and perils of large language models, their emerging applications across diverse scientific domains, and the opportunities to overcome the challenges that lie ahead.

Electron counting and oxidation states are fundamental concepts used for understanding a molecule’s electronic structure. A recent study of PuO₂ provides insights into how covalency can impact our assessment of the electronic structures of f-element materials.

Ahead of his 80th birthday, Michele Parrinello, Principal Investigator of Atomistic Simulations at the Italian Institute of Technology (IIT), discussed his career in science exploring chemical fundamentals and modelling molecular dynamics.

We reflect on ongoing topics and discussions at conferences in the field of computational chemistry.

A recent study provides intuition and guidelines for deciding whether to incorporate cheaper, lower-fidelity experiments into a closed-loop search for molecules and materials with desired properties.

Predicting the macroscopic properties of molecular liquids from first principles is a major challenge owing to the disordered nature of liquids and the weak link between microscopic forces and thermodynamic observables. A new workflow called BAMBOO produces accurate and transferable machine learning interatomic potential simulations of liquid electrolytes.