Pushing the limits: Efficient wavefunction methods for excited states in complex systems using frozen-density...

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Frozen density embedding (FDE) simplifies complex environment calculations by dividing systems into subsystems.
• Combining FDE with wavefunction methods is computationally intensive, limiting its use to smaller molecules.
• Steep computational scaling of wavefunction methods hinders their application in large, complex systems.

Purpose of the Study:

• To develop and validate a computationally efficient FDE approach combined with pair natural orbitals (PNOs).
• To enable accurate excitation energy calculations for complex molecular systems, such as molecular crystals.
• To mitigate the computational cost associated with traditional FDE-wavefunction method combinations.

Main Methods:

• Implementation of the uncoupled FDE (FDEu) approach for excitation energy calculations.
• Integration of FDE with pair natural orbitals (PNOs) within the TURBOMOLE software package.
• Utilizing efficient implementations of second-order correlation methods (ricc2 and pnoccsd programs).

Main Results:

• The combination of FDE and PNOs shows significantly smaller truncation errors than environment-induced shifts.
• PNO-based local approximations can be combined with FDE without losing significant digits.
• Demonstrated substantial computational savings compared to conventional supermolecular calculations.
• Excitation energies were found to converge rapidly with the size of the embedding environment.

Conclusions:

• The FDE-PNO combination offers a computationally efficient and accurate method for studying complex molecular systems.
• This approach overcomes the limitations of previous FDE-wavefunction method combinations.
• The findings suggest broad applicability for studying large molecular systems and materials.