Reducing the Cost of CCSD Basis Set Extrapolation in Ab Initio Computational Thermochemistry

Area of Science:

• Computational Chemistry
• Theoretical Chemistry
• Quantum Chemistry

Background:

• Coupled cluster singles and doubles (CCSD) calculations are crucial for accurate theoretical predictions.
• Approximating CCSD contributions is necessary to reduce computational cost.
• Existing methods like MP2 have limitations in accuracy and applicability.

Purpose of the Study:

• To develop and evaluate approximations to CCSD contributions for total atomization energies.
• To identify cost-effective basis-set intermediates for sub-chemical accuracy calculations.
• To propose new schemes for approximating large CCSD basis set limits.

Main Methods:

• Benchmarking approximations against the HEAT+CH4 test suite.
• Evaluating MP2 and SDQ-MP4 as basis-set intermediates.
• Developing a "CBS-like" scheme combining MP2 and SDQ-MP4.
• Assessing SDQ-MP4 for harmonic and fundamental frequency computations.

Main Results:

• MP2 is effective without empirical scaling for certain accuracy ranges.
• SDQ-MP4 offers significant improvement over MP2 as a basis-set intermediate at reduced cost.
• A "CBS-like" scheme provides a "cheap" three-part approximation of CCSD basis set limits.
• SDQ-MP4 shows promise for accurate zero-point corrections in frequency computations.

Conclusions:

• SDQ-MP4 is recommended as a de facto CCSD basis-set intermediate for sub-chemical accuracy when CCSD is unaffordable.
• The proposed "CBS-like" scheme enables affordable approximations of CCSD basis set limits.
• SDQ-MP4 facilitates accurate harmonic and fundamental frequency computations using composite methods.