1. A unified quantum framework for electrons and ions: The self-consistent harmonic approximation on a neural network curved manifold
- Author
-
Monacelli, Lorenzo, Siciliano, Antonio, and Marzari, Nicola
- Subjects
Condensed Matter - Materials Science ,Condensed Matter - Strongly Correlated Electrons - Abstract
The numerical solution of the many-body problem of interacting electrons and ions beyond the adiabatic approximation is a key challenge in condensed matter physics, chemistry, and materials science. Traditional methods to solve the multi-component quantum Hamiltonian tend to be specialized electrons or ions and can suffer from a methodological gap when applied to both electrons and ions simultaneously. In addition, ionic techniques often limited as $T\to$ 0K, whereas electronic methods are designed for 0K. Thus, efficient strategies that simultaneously address the thermal fluctuations of ions at ambient temperature without struggling to describe the electronic quantum state from first principles are missing. This work extends the self-consistent harmonic approximation for the ions to include also the electrons. The approach minimizes the total free energy by optimizing an \emph{ansatz} density matrix, solving a fermionic self-consistent harmonic Hamiltonian on a curved manifold, which is parametrized through a neural network. We demonstrate that this approach, designed initially for a flat Cartesian space to treat quantum nuclei at finite temperatures, can efficiently tackle both the ground and excited state properties of electronic systems, thus paving the way to a unified quantum description for electrons and atomic nuclei. Importantly, this approach preserves an analytical expression for entropy, enabling the direct computation of free energies and phase diagrams of materials. We benchmark the numerical implementation in several prototypical cases, proving that it captures quantum tunneling, electron-ion cusps, and static electronic correlations in the dissociation of H$_2$, where other mean-field approaches fail.
- Published
- 2024