Skip to main content
QUICK REVIEW

[Paper Review] Alchemical perturbation density functional theory (APDFT)

von Rudorff, Guido Falk, von Lilienfeld, O. Anatole|arXiv (Cornell University)|Sep 5, 2018
Advanced Chemical Physics Studies32 citations
TL;DR

This paper introduces Alchemical Perturbation Density Functional Theory (APDFT), an orbital-free method that uses Taylor-series expansions of electron density derivatives under alchemical perturbations (e.g., nuclear charge changes) to predict electronic properties of iso-electronic molecules. It achieves accuracy comparable to high-level methods like PBE0 at negligible cost, using only a single reference density and its perturbations.

ABSTRACT

We introduce an orbital free electron density functional approximation based on alchemical perturbation theory. Given convergent perturbations of a suitable reference system, the accuracy of popular self-consistent Kohn-Sham density functional estimates of properties of new molecules can be systematically surpassed---at negligible cost. The associated energy functional is an approximation to the integrated energy derivative, requiring only perturbed reference electron densities: No self-consistent field equations are necessary to estimate energies and electron densities. Electronic ground state properties considered include covalent bonding potentials, atomic forces, as well as dipole and quadropole moments.

Motivation & Objective

  • To develop a computationally efficient method for predicting electronic properties of iso-electronic molecules.
  • To surpass standard DFT accuracy using only reference electron density and its derivatives.
  • To eliminate the need for self-consistent field iterations in property prediction.
  • To enable systematic accuracy improvement via higher-order perturbation terms.
  • To allow large-scale screening of chemical space using one high-accuracy reference calculation.

Proposed method

  • Uses alchemical perturbation theory to model changes in nuclear charge as a continuous parameter λ.
  • Expands the energy and electron density in a Taylor series around the reference system (λ = 0).
  • Applies the Hellmann-Feynman theorem to express energy derivatives as integrals over the external potential difference and density derivatives.
  • Constructs a lambda-averaged electron density ˜ρ via summation of density derivatives, enabling energy prediction via ∫∆v·˜ρ dr.
  • Derives density and energy derivatives using finite differences or analytical derivatives of the reference density.
  • Truncates the series after few terms, achieving convergence for small ΔZ (e.g., ΔZ = 1) systems with fixed geometry.

Experimental results

Research questions

  • RQ1Can alchemical perturbation theory be used to predict electronic properties without self-consistent field iterations?
  • RQ2Can the accuracy of DFT-like properties be systematically improved using only reference density derivatives?
  • RQ3Does the lambda-averaged density ˜ρ converge and yield accurate energies and properties for iso-electronic systems?
  • RQ4Can this method predict forces, dipole, and quadrupole moments with high fidelity?
  • RQ5Can one high-accuracy reference calculation enable reliable predictions across a broad chemical space?

Key findings

  • APDFT achieves energy, force, and electrostatic property predictions comparable to PBE0 at perturbation orders 3–4 for CO and BF relative to N2 reference.
  • For He, energy errors drop below 1 mH with n = 4 expansion order using HF or DFT functionals.
  • The method predicts dipole and quadrupole moments with accuracy matching the reference level.
  • Convergence is observed for hydrogenic atoms and free atoms, with numerical evidence supporting convergence for small ΔZ systems.
  • The approach enables geometry optimization and chemical composition scanning via alchemical gradients, with higher-order terms extending applicability to larger chemical space regions.
  • Only the reference electron density and its derivatives are needed, allowing application to any reference method that outputs densities.

Better researchstarts right now

From reading papers to final review, dramatically reduce your research time.

No credit card · Free plan available

This review was created by AI and reviewed by human editors.