Putting a Temperature on Electron Correlation
π arXiv:2610.22134 Β· π₯ PDF Β· 2026-10-15 Β· quant-ph
Authors: Paul G. Graf [arXiv Β· scholar] , Florian Matz [arXiv Β· scholar] , Markus Penz [arXiv Β· scholar] , Christian Schilling [arXiv Β· scholar]
π Abstract
It is common wisdom that a better understanding of how to quantify electron correlation should lead to more efficient treatments of the correlation problem. Here, we not only substantiate this view but turn it into a concrete, systematic framework for ground states. Since prominent correlation measures such as the nonfreeness depend only on the one-particle reduced density matrix, modeling the correlation energy as a function of such a measure naturally places the framework within reduced density matrix functional theory, while keeping the cost at mean-field level. Strikingly, the variable conjugate to the measure emerges as a ``correlation temperature'', thereby connecting correlation measures, functional theory, and effective temperatures. The framework rationalizes methods such as i-DMFT, and we illustrate its key concepts analytically for the Hubbard dimer and numerically for small molecular systems. This brings a fundamental question within reach: how much of the correlation problem can be solved at mean-field cost?