Improved cluster-in-molecule local correlation approach for electron correlation calculation of large systems

Area of Science:

• Computational chemistry
• Quantum chemistry
• Materials science

Background:

• Accurate electron correlation calculations are crucial for understanding molecular behavior.
• Calculating correlation energy for large systems is computationally demanding.
• Local correlation methods offer a computationally efficient alternative.

Purpose of the Study:

• To develop an improved cluster-in-molecule (CIM) local correlation approach.
• To enhance the accuracy and speed of electron correlation calculations for large molecular systems.
• To refine the construction of virtual localized molecular orbitals (LMOs) for various cluster types and basis sets.

Main Methods:

• Developed a refined strategy for constructing virtual LMOs within the CIM framework.
• Investigated the impact of the distance threshold (ξ) on recovering medium-range electron correlation.
• Applied the CIM approach to second-order Møller-Plesser perturbation theory (MP2) and coupled cluster singles and doubles (CCSD) calculations.

Main Results:

• The improved CIM method allows for more accurate and faster electron correlation calculations.
• A larger distance threshold (ξ) is necessary for high accuracy in large systems.
• The CIM-MP2 and CIM-CCSD schemes recover over 99.8% of correlation energies across various systems and basis sets.
• The CIM-MP2 scheme provides reliable relative energy differences for polypeptide secondary structures.

Conclusions:

• The enhanced CIM method significantly improves the efficiency and accuracy of electron correlation calculations for large systems.
• The choice of distance threshold (ξ) is critical for achieving quantitative accuracy.
• This method offers a robust and reliable tool for computational chemistry research, particularly for large molecules and biomolecular systems.