On the Kinetic Energy Density Functional: The Limit of the Density Derivative Order

Abstract

Within ``orbital-free'' density functional theory, it is essential to develop general kinetic energy density (KED), denoted as $t(\mathbf{r})$. This is usually done by empirical corrections and enhancements, gradient expansions, machine learning, or axiomatic approaches to find forms that satisfy physical necessities. In all cases, it is crucial to determine the largest spatial density derivative order, $m$ in, $t(\mathbf{r})$. There have been many efforts to do so, but none have proven general or conclusive and there is no clear guide on how to set $m$. In this work, we found that, by imposing KED finitude, $m=D+1$ for systems of dimension $D$. This is consistent with observations and provides a needed guide for systematically developing more accurate KEDs.

References in corpus (4)

• 'Hartree-exchange' in ensemble density functional theory: Avoiding the non-uniqueness disaster
• Toward Orbital-Free Density Functional Theory with Small Data Sets and Deep Learning
• Order-$N$ orbital-free density-functional calculations with machine learning of functional derivatives for semiconductors and metals
• Kinetic Energy Density Functionals by Axiomatic Approach