A density functional for core-valence correlation energy.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Theoretical Chemistry

Background:

• Accurate calculation of core-valence correlation energy is crucial for high-precision chemical predictions.
• Existing methods often involve significant computational cost or approximations.

Purpose of the Study:

• To develop a new density functional, εCV-DFT(ρc, ρv), for describing core-valence correlation energy.
• To assess the accuracy and efficiency of the developed functional against benchmark datasets and established methods.

Main Methods:

• Construction of εCV-DFT(ρc, ρv) as a linear combination of existing functionals and a nuclear charge-dependent term.
• Inclusion of 6 adjustable parameters optimized against a benchmark set of 194 chemical energy changes.
• Comparison of εCV-DFT(ρc, ρv) performance against coupled cluster (CCSD(T), CCSD) and Møller-Plesset perturbation theory (MP2) calculations.

Main Results:

• The εCV-DFT(ρc, ρv) functional achieved an rms error of ±0.27 kcal/mol for 194 chemical energy changes.
• This accuracy is comparable to CCSD/MTsmall (±0.39 kcal/mol) and superior to MP2/GTlargeXP (±0.65 kcal/mol), though slightly less accurate than CCSD(T)/MTsmall (±0.16 kcal/mol).
• The functional's evaluation is computationally inexpensive, requiring less time than a single self-consistent field (SCF) iteration.

Conclusions:

• The developed εCV-DFT(ρc, ρv) functional provides a computationally efficient and accurate method for core-valence correlation energy.
• Its accuracy is suitable for model chemistries aiming for CCSD(T) level of theory.
• This advancement offers a practical tool for various applications in computational chemistry.