Track synthesis, materials and catalysis as it is published.
Search papers, labs, and topics across Lattice.
This paper introduces STRUCTURES-P, an all-electron, orbital-free machine-learned density functional for periodic CHNO systems, which analytically handles electrostatics and uses an equivariant network to learn kinetic and exchange-correlation energies from non-ground-state density gradients. The method achieves a density error of 0.4% of the electron count and an energy error of 0.25 mHa/atom on held-out crystals, outperforming a comparable interatomic potential. Notably, a single functional is trainable for both molecular and periodic systems, generalizes to larger unit cells, and preserves relative energy rankings, demonstrating the potential of ML-driven orbital-free DFT for variational density optimization in solids.
Orbital-free DFT density errors plummet from 6.7% to 0.4% with a single, generalizable ML functional for periodic systems.
We introduce STRUCTURES-P, a machine-learned all-electron orbital-free density functional for periodic CHNO systems. Electrostatics are evaluated analytically in an atom-centered Gaussian basis, while an equivariant network learns the kinetic and exchange-correlation energy from energies and gradients of non-ground-state densities. Minimizing the functional from a superposition of atomic densities recovers the Kohn-Sham ground state: on held-out crystals, the density error falls from 6.7% to 0.4% of the electron count and the energy error reaches 0.25 mHa per atom, half that of a same-architecture interatomic potential. A single functional can be trained for both isolated molecules and periodic systems, and the model generalizes to larger unit cells. It largely preserves relative-energy rankings and reproduces convex-hull vertices on the tested H-N composition slice. Together, these results demonstrate the viability of machine-learned all-electron orbital-free DFT for variational density optimization and energy evaluation in fixed-geometry solids.