Reduced-cost linear-response CC2 method based on natural orbitals and natural auxiliary functions

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Theoretical Spectroscopy

Background:

• Accurate calculation of molecular excitation energies is crucial for understanding photophysical processes.
• High-level correlated methods like coupled-cluster (CC2) are computationally expensive, limiting their application to small systems.
• Density fitting (DF) approximations can reduce computational cost but often require significant basis sets.

Purpose of the Study:

• To develop a reduced-cost density fitting (DF) linear-response second-order coupled-cluster (CC2) method.
• To enable efficient computation of excitation energies for medium-sized molecules.
• To significantly reduce computational expense while maintaining high accuracy.

Main Methods:

• Simultaneous truncation of molecular orbital (MO) and auxiliary basis sets for DF approximation.
• Construction of state-specific natural orbitals (NOs) from MP2 and CIS(D) density matrices.
• Truncation of the natural auxiliary function (NAF) basis.

Main Results:

• Achieved over an order of magnitude speedup in CC2 calculations.
• Reduced virtual MO basis by ~60% and fitting basis by a factor of five.
• Obtained a mean absolute error of only 0.02 eV compared to canonical CC2 results.

Conclusions:

• The developed reduced-cost DF-CC2 method is computationally efficient.
• Enables accurate calculation of excitation energies for molecules up to 100 atoms.
• Provides a practical approach for studying excited states of medium-sized molecules.