Trace correcting density matrix extrapolation in self-consistent geometry optimization.

Area of Science:

• Computational chemistry
• Materials science
• Quantum mechanics

Background:

• Geometry optimization is crucial for determining molecular and material structures.
• Achieving self-consistency in electronic structure calculations can be computationally intensive.
• Existing methods for accelerating convergence may have limitations in accuracy or scalability.

Purpose of the Study:

• To introduce a novel linear scaling trace correcting density matrix extrapolation method.
• To accelerate self-consistency convergence in geometry optimization.
• To improve the accuracy and efficiency of electronic structure calculations.

Main Methods:

• The proposed method utilizes nonorthogonal trace correcting purification and perturbation theory.
• It employs a linear scaling approach for density matrix extrapolation.
• The technique is applied to geometry optimization calculations.

Main Results:

• Extrapolated total energies show improved accuracy, often an order of magnitude closer to the self-consistent solution.
• The method exhibits low computational cost for insulators.
• The computational cost scales linearly with the size of the perturbed region (O(N(pert))).

Conclusions:

• The proposed method effectively accelerates self-consistency convergence in geometry optimization.
• For local perturbations, the computational cost is independent of system size (O(1)).
• This technique offers a promising advancement for efficient and accurate electronic structure calculations.