This proposal puts forth a novel strategy to tackle large quantum systems. A variety of highly sophisticated methods such as quantum Monte Carlo, configuration interaction, coupled cluster, tensor networks, Feynman diagrams, dynamical mean-field theory, density functional theory, and semi-classical techniques have been developed to deal with the enormous complexity of the many-particle Schrödinger equation. The goal of our proposal is not to add another method to these standard techniques but, instead, we develop a systematic way of combining them. The essential ingredient is a novel way of decomposing the wave function without approximation into factors that describe subsystems of the full quantum system. This so-called exact factorization is asymmetric. In the case of two subsystems, one factor is a wave function satisfying a regular Schrödinger equation, while the other factor is a conditional probability amplitude satisfying a more complicated Schrödinger-like equation with a non-local, non-linear and non-Hermitian “Hamiltonian”. Since each subsystem is necessarily smaller than the full system, the above standard techniques can be applied more efficiently and, most importantly, different standard techniques can be applied to different subsystems. The power of the exact factorization lies in its versatility. Here we apply the technique to five different scenarios: The first two deal with non-adiabatic effects in (i) molecules and (ii) solids. Here the natural subsystems are electrons and nuclei. The third scenario deals with nuclear motion in (iii) molecules attached to semi-infinite metallic leads, requiring three subsystems: the electrons, the nuclei in the leads which ultimately reduce to a phonon bath, and the molecular nuclei which may perform large-amplitude movements, such as current-induced isomerization, (iv) purely electronic correlations, and (v) the interaction of matter with the quantized electromagnetic field, i.e. electrons, nuclei and photons.
Fields of science
Funding SchemeERC-ADG - Advanced Grant
See on map