Prediction of many-electron wavefunctions using atomic potentials: extended basis sets and molecular dissociation

Area of Science:

• Quantum Chemistry
• Computational Chemistry
• Theoretical Chemistry

Background:

• Accurate molecular wavefunctions are essential for predicting chemical properties.
• Existing methods often require significant computational resources or approximations.

Purpose of the Study:

• To develop a novel method for generating accurate molecular orbitals using one-electron potentials.
• To assess the accuracy of these orbitals for constructing wavefunctions and calculating energies.

Main Methods:

• Utilizing a one-electron Schrödinger equation with special atomic potentials.
• Employing the exact Hamiltonian for variational energy calculations and configuration interaction (CI) expansions.
• Extending previous work to larger basis sets and molecular dissociation.

Main Results:

• Generated molecular orbitals are accurate enough for direct use in wavefunction construction.
• Energies calculated using invariant atomic potentials deviate by <0.04 eV per bond/valence electron pair from CI energies.
• Errors reduce to 0.01 eV upon single Fock matrix diagonalization.
• Atomization energies and dissociation curves show good agreement with established CI methods.

Conclusions:

• The proposed one-electron potentials provide a computationally efficient route to accurate molecular orbitals.
• This method is effective for studying molecular geometries, bonding, and dissociation processes.
• The approach offers a reliable alternative for high-accuracy quantum chemical calculations.