Optimization of Random Phase Approximation Calculations for Improved Energies of Molecules, Solids, and Surfaces

Area of Science:

• Computational Chemistry
• Materials Science
• Quantum Mechanics

Background:

• The Random Phase Approximation (RPA) is a method for calculating electronic correlation energy.
• Conventional RPA methods often require significant computational resources and can lack accuracy for certain systems.
• Developing accurate and efficient methods for electronic structure calculations is crucial for understanding chemical and material properties.

Purpose of the Study:

• To present an optimized Random Phase Approximation method (optRPA26) that improves accuracy without altering the core RPA functional.
• To establish optRPA26 as a reliable, general-purpose reference method for diverse chemical and material systems.
• To demonstrate the method's capability in accurately describing various bonding types.

Main Methods:

• The optRPA26 method uses an empirically constructed hybrid functional to generate Density Functional Theory (DFT) orbitals.
• RPA correlation energy is evaluated using these DFT orbitals and then scaled by a constant.
• The approach is designed to be compatible with standard RPA implementations.

Main Results:

• optRPA26 achieves high accuracy across molecules, solids, and surfaces.
• Mean absolute errors are reported for reaction energies (0.05 eV), cohesive energies (0.07 eV), and formation energies (0.09 eV).
• Accurate predictions are also shown for molecule adsorption on metals (0.11-0.12 eV) and oxides (0.06 eV), and phase stability is correctly captured.

Conclusions:

• optRPA26 offers a significant improvement over conventional RPA methods.
• The method accurately captures covalent, ionic, metallic, and van der Waals bonding.
• optRPA26 serves as a versatile, high-accuracy reference method for electronic structure calculations.