# Improved method for generating exchange-correlation potentials from   electronic wave functions

**Authors:** Egor Ospadov, Ilya G. Ryabinkin, Viktor N. Staroverov

arXiv: 1701.08708 · 2017-02-27

## TL;DR

This paper introduces a modified iterative method to generate more accurate exchange-correlation potentials from electronic wave functions, ensuring robustness across various Gaussian basis sets and improving finite basis set approximations.

## Contribution

The authors derive a new modification to the RKS method that enhances its robustness and accuracy for all Gaussian basis sets, including small and medium sizes.

## Key findings

- Method becomes reliable for all Gaussian basis sets
- Improved exchange-correlation potentials closer to basis-set limit
- Enhanced accuracy over original RKS procedure

## Abstract

Ryabinkin, Kohut, and Staroverov (RKS) [Phys. Rev. Lett. 115, 083001 (2015)] devised an iterative method for reducing many-electron wave functions to Kohn-Sham exchange-correlation potentials, $v_\text{XC}(\mathbf{r})$. For a given type of wave function, the RKS method is exact (Kohn-Sham-compliant) in the basis-set limit; in a finite basis set, it produces an approximation to the corresponding basis-set-limit $v_\text{XC}(\mathbf{r})$. The original RKS procedure works very well for large basis sets but sometimes fails for commonly used (small and medium) sets. We derive a modification of the method's working equation that makes the RKS procedure robust for all Gaussian basis sets and increases the accuracy of the resulting exchange-correlation potentials with respect to the basis-set limit.

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/1701.08708/full.md

## Figures

6 figures with captions in the complete paper: https://tomesphere.com/paper/1701.08708/full.md

## References

62 references — full list in the complete paper: https://tomesphere.com/paper/1701.08708/full.md

---
Source: https://tomesphere.com/paper/1701.08708