Construction of a Spin-Component Scaled Dual-Hybrid Random Phase Approximation

Area of Science:

• Quantum Chemistry
• Computational Chemistry
• Theoretical Chemistry

Background:

• The direct random phase approximation (dRPA75) method shows promise for reaction energies, barrier heights, and noncovalent interactions.
• However, dRPA75 exhibits systematic errors in atomization energy computations.
• Addressing these limitations is crucial for advancing theoretical chemistry methods.

Purpose of the Study:

• To develop a modified dRPA75 method that corrects systematic errors in atomization energies.
• To preserve the accuracy of the original dRPA75 method for spin-unpolarized systems.
• To enhance the performance of dRPA75 for open-shell systems and various chemical benchmarks.

Main Methods:

• Introduction of a constrained spin-component scaling formalism to the dRPA75 method, creating SCS-dRPA75.
• Utilizing the aug-cc-pVTZ basis set for calculations.
• Evaluation against established test sets like RC0-RC5 and the GMTKN30 database.

Main Results:

• SCS-dRPA75 achieves an average error below 1.5 kcal mol⁻¹ for hydrocarbon reactions (RC0-RC5).
• The method outperforms the PWRB95 method for open-shell main-group elements and is comparable to advanced double-hybrid methods.
• SCS-dRPA75 demonstrates well-balanced performance on barrier heights, rivaling state-of-the-art methods.

Conclusions:

• SCS-dRPA75 effectively overcomes the atomization energy errors of dRPA75 while maintaining accuracy.
• The new method shows improved performance on open-shell systems and comparable accuracy to established high-level methods.
• SCS-dRPA75 exhibits reduced self-interaction, delocalization, and static correlation errors.