| （英） |
Metasurfaces can achieve functions that are difficult to be achieved with ordinary materials by arranging various
components smaller than the wavelength. In this study, in order to design high performance metasurfaces, we are developing
a topology optimization method that can automatically create optimum structures by just specifying the desired properties.
In our approach, the structure is represented by a multi-materials function expansion method, and the covariance matrix
adaptation evolution strategy (CMA-ES) is used to optimize structures. A full-vector three-dimensional finite element method
is used for numerical analysis, and automatic element discretization scheme suitable for the function expansion method is
also developed for accurate modeling of the given structure. To show the usefulness of our approach, a design example of
polarization-independent magneto-optic metasurface isolator is shown. The relationship between the degree of freedom of the
function expansion method and the obtained device performance is also investigated in detail. In the future, we are planning to
study a versatile optimal design method of metasurfaces with higher performance using various optical effects. |