A Parallel Block Preconditioner-Based VIE-FFT Algorithm for Modeling the Electromagnetic Response From Nanostructures

TL;DR

This paper introduces a parallel block preconditioner-based VIE-FFT algorithm that efficiently models electromagnetic responses of nanostructures, significantly reducing computational time and iterations in nano-electromagnetic simulations.

Contribution

It presents a novel parallel VIE-FFT method with a block preconditioner and OpenMP parallelization for improved nanostructure electromagnetic modeling.

Findings

Reduces iteration count in simulations.

Decreases computational time significantly.

Effective for dielectric metasurfaces and plasmonic solar cells.

Abstract

The superior ability of nanostructures to manipulate light has propelled extensive applications in nano-electromagnetic components and devices. Computational electromagnetics plays a critical role in characterizing and optimizing the nanostructures. In this work, a parallel block preconditioner based volume integral equation (VIE)-fast Fourier transform (FFT) algorithm is proposed to model the electromagnetic response from representative nanostructures. The VIE using uniform Cartesian grids is first built, and then the entire volumetric domain is partitioned into geometric subdomains based on the regularity and topology of the nanostructure. The block diagonal matrix is thus established, whose inverse matrix serves as a preconditioner for the original matrix equation. The resulting linear system is solved by the bi-conjugate gradient stabilized (BiCGSTAB) method with different residual error tolerances in the inner and outer iteration processes; and the FFT algorithm is used to accelerate the matrix-vector product (MVM) operations throughout. Furthermore, because of the independence between the inner processes of solving block matrix equations, the OpenMP framework is empolyed to execute the parallel operations. Numerical experiments indicate that the proposed method is effective and reduces both the iteration number and the computational time significantly for the representative nano-electromagnetic problems like the dielectric focusing metasurfaces and the plasmonic solar cells.