Stochastically accelerated perturbative triples correction in coupled cluster calculations

Area of Science:

• Quantum chemistry
• Computational physics
• Theoretical chemistry

Background:

• Coupled-cluster (CC) methods are highly accurate for electronic structure calculations.
• Perturbative triples corrections (CC(2,3)) are crucial for high accuracy but computationally expensive.
• Deterministic calculation of CC(2,3) is a bottleneck for large molecular systems.

Purpose of the Study:

• To develop a novel, computationally efficient algorithm for perturbative triples corrections in coupled-cluster theory.
• To balance accuracy and computational cost using stochastic sampling.
• To enable accurate electronic structure calculations for previously intractable molecular systems.

Main Methods:

• Introduction of a novel semi-stochastic algorithm.
• Leveraging stochastic sampling techniques within the coupled-cluster framework.
• Combining deterministic and random elements for computation.

Main Results:

• The algorithm allows calculations to be terminated at any point, yielding an unbiased estimate.
• Statistical error diminishes as the exact calculation is approached.
• Achieved 0.5 millihartree precision with only 10% of the computational cost of full methods.
• Demonstrated substantial computational savings compared to traditional deterministic approaches.

Conclusions:

• The developed semi-stochastic algorithm offers a significant reduction in computational cost for perturbative triples corrections.
• This method provides a pathway for accurate and efficient electronic structure calculations.
• Enables the study of complex molecular systems previously limited by computational resources.