Accelerating inverse Kohn-Sham calculations using reduced density matrices.

Area of Science:

• Quantum Chemistry
• Computational Physics
• Materials Science

Background:

• Accurate exchange-correlation (XC) potentials are crucial for electronic structure calculations.
• Existing methods like Ryabinkin-Kohut-Staroverov (RKS) and Kanungo-Zimmerman-Gavini (KZG) have complementary strengths in XC potential determination.
• RKS offers computational efficiency with finite basis sets, while KZG provides higher accuracy using complete basis sets.

Purpose of the Study:

• To develop a hybrid computational approach combining the RKS and KZG methods.
• To enhance the efficiency of XC potential calculations without sacrificing accuracy.
• To provide a direct comparison of XC potentials derived from RKS and KZG methods.

Main Methods:

• Implementing a hybrid approach where the RKS solution serves as an initial guess for the KZG method.
• Utilizing accurate configuration interaction calculations with a Slater orbital basis for ground state solutions.
• Comparing the accuracy of RKS and KZG derived XC potentials for various molecular systems.

Main Results:

• Achieved a significant 3-11× speedup in computational time by combining RKS and KZG methods.
• Demonstrated the feasibility of using RKS as an efficient initial guess for KZG.
• Presented a detailed comparison of XC potentials and their accuracy for weakly and strongly correlated molecules.

Conclusions:

• The hybrid RKS-KZG approach successfully balances computational efficiency and accuracy in XC potential determination.
• This combined method offers a practical advancement for electronic structure calculations.
• The findings provide valuable insights into the performance of different XC potential calculation methods.