Even-handed subsystem selection in projection-based embedding

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Projection-based embedding integrates correlated wavefunction methods with density functional theory.
• Current partitioning of subsystems relies on localized molecular orbitals, which can change unstably during geometry changes.
• Unstable partitioning leads to artifacts like cusps and discontinuities in potential energy surfaces.

Purpose of the Study:

• To develop a robust and consistent framework for subsystem partitioning in projection-based embedding.
• To address the limitations of existing methods that produce unphysical results during large geometry changes.
• To enable accurate electronic structure calculations for complex chemical systems and reactions.

Main Methods:

• Developed an even-handed framework for localized orbital partitioning.
• Ensured consistent subsystem definitions across a range of molecular geometries.
• Applied the method to model systems, including an SN2 reaction, a cobalt-based CO2 reduction catalyst, and CO binding to copper clusters.

Main Results:

• The even-handed partitioning framework successfully maintained consistent subsystems during geometry variations.
• Eliminated unphysical cusps and discontinuities in the potential energy surface.
• Achieved chemically accurate embedding results even with small embedded regions.

Conclusions:

• The proposed even-handed partitioning strategy provides a stable and accurate method for embedding correlated wavefunction methods in DFT.
• This approach overcomes limitations of previous methods, enabling reliable calculations for challenging chemical systems.
• Facilitates accurate computational studies of chemical reactions, catalysis, and molecular interactions.