A Linear-Scaling Integral-Direct Explicitly Correlated Second-Order Møller-Plesset Approach

Area of Science:

• Quantum Chemistry
• Computational Chemistry
• Method Development

Background:

• Accurate calculation of electron correlation is crucial for predicting molecular properties.
• Explicitly correlated methods, like MP2-F12, offer high accuracy but are computationally expensive.
• Previous local MP2 methods have limitations in accuracy and scalability.

Purpose of the Study:

• To develop a linear-scaling, iteration-free, local explicitly correlated second-order Møller-Plesset (MP2-F12) approach.
• To enable accurate and efficient calculations for large molecular systems.
• To extend previous local MP2 methods with enhanced accuracy and reduced computational cost.

Main Methods:

• Integral-direct, iteration-free, linear-scaling local MP2-F12 approach.
• Domain-based computation of electron pair correlation contributions.
• Multipole expansions for spatially distant electron pairs.
• Implementation of the 2B MP2-F12 ansatz for improved basis set convergence.

Main Results:

• Recovered at least 99.9% of the canonical MP2-F12 correlation energy.
• Achieved reaction energies with a mean error of less than 1 kJ/mol.
• Demonstrated excellent basis set convergence with double-ζ basis sets.
• Successfully computed correlation energy for a 644-atom protein, the largest system in explicitly correlated calculations.

Conclusions:

• The new local MP2-F12 approach offers a balance of high accuracy and computational efficiency.
• This method significantly reduces memory requirements and disk I/O.
• Enables the study of large, complex molecular systems previously inaccessible to explicitly correlated methods.