Local Natural Orbital Coupled Cluster Method Benchmarks for Thermochemistry

A 408.5 KB Excel dataset presents benchmark results for the accuracy of local natural orbital approximations in coupled cluster quantum chemistry methods. The data, created by Vladimir Fishman and uploaded to figshare in April 2026, assesses contributions from high-order methods like CCSDTQ for predicting thermochemistry and noncovalent interactions.

Use Cases

• Benchmarking the CCSDT–CCSD(T) and CCSDT(Q)–CCSD(T) energy contributions against default and tightened LNO threshold settings.
• Assessing basis set completeness and LNO approximation accuracy for thermochemistry predictions like atomization energies and barrier heights.
• Evaluating the performance of LNO-CCSDTQ methods for calculating noncovalent interaction energies in molecular systems.
• Analyzing the relative and absolute accuracy (e.g., 0.05–0.1 kcal/mol) of post-CCSD(T) corrections across different molecular complexities.

Strengths

• Benchmarks include results for systems with over 100 orbitals, demonstrating applicability beyond small molecules.
• Data supports methodological claims with specific accuracy metrics, such as 85–95% relative accuracy for certain contributions.

Limitations

• The dataset is small (408.5 KB), indicating limited scope and likely a constrained number of benchmark systems or data points.
• Specific row counts and detailed column structures for the Excel file are not provided, limiting reproducibility analysis.

Provenance

Source

Vladimir Fishman, uploaded to the figshare repository.

Collection Method

Results from computational development and benchmarking of local natural orbital coupled cluster methods.

Freshness

Last updated in April 2026, indicating recent methodological assessment.