Direct energy functional minimization under orthogonality constraints.

Area of Science:

• Computational Chemistry
• Materials Science
• Quantum Mechanics

Background:

• Direct energy functional minimization in electronic structure theory requires optimizing single-particle orbitals under orthogonality constraints.
• Existing methods can be computationally intensive and complex to implement.

Purpose of the Study:

• To develop an efficient and numerically stable method for direct energy minimization in electronic structure theory.
• To address the challenge of orbital orthogonality constraints in computational chemistry.

Main Methods:

• An orbital transformation based on the inverse factorization of the overlap matrix is presented.
• This transformation maps constrained to approximate unconstrained energy functionals.
• A conjugate gradient scheme is employed to minimize the transformed functionals.

Main Results:

• The presented technique offers an efficient, robust, and numerically stable approach for total energy minimization.
• For sparse problems, the computational effort scales linearly with the number of basis functions (N).
• A large-scale density functional theory calculation of a hydrated DNA decamer was successfully performed.

Conclusions:

• The developed orbital transformation method significantly enhances the efficiency and stability of electronic structure calculations.
• This approach is applicable to various first-principles methods like tight-binding, Hartree-Fock, and density functional theory.
• The method's scalability and robustness were demonstrated through a complex biomolecular system calculation.