Bootstrap Embedding For Large Molecular Systems.

Area of Science:

• Quantum embedding theories
• Computational chemistry
• Method development

Background:

• Quantum embedding theories offer efficient approaches for correlated calculations in large systems.
• Previous work established bootstrap embedding (BE) for correlated ab initio calculations.

Purpose of the Study:

• Extend bootstrap embedding (BE) to coupled cluster with singles and doubles (CCSD) level for large molecules.
• Introduce algorithmic developments to reduce computational cost and maintain accuracy.

Main Methods:

• Developed algorithmic improvements for bootstrap embedding (BE).
• Implemented BE for coupled cluster with singles and doubles (CCSD) calculations.
• Tested on conjugated molecules to assess accuracy and efficiency.

Main Results:

• Achieved O(N^3) scaling for integral transform and O(N) for CCSD calculation.
• Recovered >99.5% of total correlation energy with reasonably sized fragments.
• Computational resources favorably compare to domain localized pair natural orbital (DLPNO) methods.

Conclusions:

• The enhanced BE method enables accurate, cost-effective correlated ab initio calculations for large molecular systems.
• The approach scales efficiently and requires modest computational resources.
• Represents a significant advancement for applying BE to practical problems in computational chemistry.