Orbital-optimized coupled-cluster theory does not reproduce the full configuration-interaction limit

Area of Science:

• Quantum Chemistry
• Computational Chemistry

Background:

• Coupled cluster (CC) methods are essential for accurate electronic structure calculations.
• Orbital optimization in CC methods aims to improve efficiency and accuracy.
• Previous methods like orbital-optimized coupled cluster (OCC) have shown limitations.

Purpose of the Study:

• To investigate the limitations of orbital-optimized coupled cluster (OCC) in reproducing the full configuration-interaction (full CI) limit.
• To identify the conditions under which OCC fails and propose a more accurate alternative.
• To evaluate the performance of OCC and Brueckner coupled cluster (BCC) for benchmark calculations.

Main Methods:

• Comparison of orbital-optimized coupled cluster (OCC) and Brueckner coupled cluster (BCC) methods.
• Numerical calculations performed up to the full CI limit for CH2 and an active-space model of ozone.
• Analysis of orbital gradients and their impact on the cluster operator.

Main Results:

• OCC fails to reach the full CI limit when the cluster operator is complete.
• Fulfillment of projected singles equations (Brueckner orbital gradient) is crucial for accurate behavior.
• Deviations of OCC from full CI are comparable to correlation errors from high-level excitations.

Conclusions:

• Standard OCC is not suitable for benchmark calculations due to its failure to converge to the full CI limit.
• Brueckner coupled cluster (BCC) provides a more reliable approach for achieving accurate results.
• A hybrid OCC approach is proposed for active-space models, combining Brueckner and variational orbital gradients.