Frozen-Core Analytical Gradients within the Adiabatic Connection Random-Phase Approximation from an Extended...

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Theoretical Chemistry

Background:

• The random-phase approximation (RPA) is a powerful method for calculating electron correlation.
• Accurate RPA calculations can be computationally expensive, limiting their application to smaller systems.
• Developing efficient computational methods is crucial for advancing theoretical chemistry.

Purpose of the Study:

• To implement and evaluate a frozen-core option for analytic gradient calculations within the RPA framework.
• To assess the computational speedup and accuracy of the frozen-core RPA method.
• To extend the applicability of RPA calculations to larger and more complex molecular systems.

Main Methods:

• Implementation of a frozen-core option combined with the analytic gradient of RPA.
• Utilized density functional theory reference determinants and resolution-of-the-identity techniques.
• Employed an extended Lagrangian and Curtis-Clenshaw quadratures for correlation contributions.

Main Results:

• The frozen-core option significantly reduces computational cost by decreasing matrix dimensionality and grid size.
• Optimized geometries, vibrational frequencies, and dipole moments show only modest deviations from all-electron results.
• Achieved computational speedups of 35-55% for various molecular systems, including alkanes and metal complexes.

Conclusions:

• The combination of the frozen-core option and RPA provides an efficient and accurate method for calculating molecular properties.
• This approach broadens the scope of systems amenable to high-accuracy RPA calculations.
• The developed method offers a practical solution for computational chemists seeking to balance accuracy and efficiency.