Efficient Low-Scaling Calculation of THC-SOS-LR-CC2 and THC-SOS-ADC(2) Excitation Energies Through Density-Based...

Area of Science:

• Computational chemistry
• Quantum chemistry
• Theoretical chemistry

Background:

• Advances in computational hardware and algorithms allow for larger system calculations.
• Approximate coupled cluster (CC2) and algebraic diagrammatic construction (ADC(2)) methods are key for electronic structure calculations.

Purpose of the Study:

• To present efficient implementations of scaled opposite-spin (SOS) CC2 and ADC(2) methods.
• To achieve near-linear scaling for these quantum chemistry methods.

Main Methods:

• Implementation of least-squares tensor hypercontraction (THC) approximation.
• Novel density-based integral-direct reformulation for grid-projection.
• Screening using Cholesky-decomposed densities (CDD) and sparse linear algebra.

Main Results:

• Developed CDD-THC-SOS-LR-CC2/ADC(2) methods with effective O(N^2) scaling.
• Demonstrated capability to target excitation energies for systems up to ~1000 atoms.
• Achieved efficient calculations on a single compute node.

Conclusions:

• The new methods significantly enhance the scalability of CC2 and ADC(2) calculations.
• Enables accurate prediction of excitation energies for large molecular systems.
• Represents a substantial step forward in computational quantum chemistry for large-scale problems.