Analytic Nuclear Gradients for Complete Active Space Linearized Pair-Density Functional Theory.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Photochemistry

Background:

• Modeling photochemical reactions is challenging due to conical intersections and multiconfigurational excited states.
• Accurate potential energy surface calculations require multistate methods with state interaction.
• Traditional multireference methods are computationally expensive.

Purpose of the Study:

• To derive and implement analytic gradients for linearized pair-density functional theory (L-PDFT).
• To demonstrate the utility of L-PDFT gradients for predicting molecular properties.
• To provide a cost-effective alternative for accurate excited-state calculations.

Main Methods:

• Derivation of analytic gradients for L-PDFT.
• Implementation of L-PDFT gradients in the PySCF-forge software.
• Application to formaldehyde, s-trans-butadiene, phenol, and cytosine.

Main Results:

• Successful derivation and implementation of L-PDFT analytic gradients.
• Accurate prediction of ground- and excited-state equilibrium geometries.
• Reliable calculation of adiabatic excitation energies.

Conclusions:

• L-PDFT with analytic gradients is a powerful and cost-effective tool for studying photochemical reactions.
• The method accurately models potential energy surfaces with strong nuclear-electronic coupling.
• This advancement facilitates more accurate predictions of molecular behavior in excited states.