Pareto multiobjective global optimization of plasmonic filter based on Q factor analysis

We propose an efficient multiobjective design approach for a plasmonic nanoslit array device based on extraordinary optical transmission using the Q factor analysis. This approach leads to a substantial reduction in computation time required for the analysis of resonance in slit array by computing the Q factor at a singular frequency, corresponding to the eigenfrequency of the dispersive wave equation, rather than by performing calculations for the overall transmission for evaluating transmission bandwidths. For rigorous numerical analysis, we investigate the optical properties of plasmonic nanoslit array devices such as transmission, return loss (RL), and Q factor using the finite element method. Furthermore, to investigate the signal distortion that the light may undergo, we calculate the group delay time and dispersion based on various structures of nanoslit array filters. Among them, the three objective functions are defined as transmission, RL, and Q factor, with slit height, slit width, and array period as design variables, respectively. Multiobjective optimization is carried out by using Pareto optimality and particle swarm optimizations. The proposed method allows us to evaluate the Q factor much faster than estimating the bandwidth from the full transmission spectrum of the periodic slit array, which significantly reduces the computational time. The Q factor analysis can be used in the multiobjective optimization process that would otherwise be impossible due to time constraints.