A decoupled phase engineering strategy for Si-GSST active metasurfaces

Designing multifunctional metasurfaces employing phase change or other kinds of optically active materials is challenging due to the coupling of the designing phases, meaning that designing the metasurface for a specific function also impacts the other functionalities. Hence, the phase and amplitude requirements of all these functionalities must be met at once. In this article, a new strategy of designing active silicon-Ge 2 Sb 2 Se 4 Te 1 (Si-GSST) metasurfaces is introduced. Close refractive index of GSST in its amorphous state to that of Si makes it possible to separate the design process into two completely independent phases. In fact the key enabling factor of the strategy to design a decoupled metasurface is the specific material property of GSST ( n amorphous = n Si ). A pure Si metasurface is firstly designed which also resembles the Si-GSST metasurface functioning in the amorphous state of GSST. Afterwards, design process continues to the crystalline phase of the GSST. Here, geometrical parameters of the GSST layer, do not impact the metasurface response in its amorphous state, whereas, these factors can be engineered to meet the requirements of the metasurface for the crystalline phase. As an example, in this article, a bifocal waveguide coupled metasurface is presented based on Si-GSST meta-atoms. Firstly, a monofocal lens made entirely of Si is presented that effectively focuses light to a focal point of f 1 = 4000 nm above the metasurface. This structure is composed of Si nano-pillars carefully engineered to meet the required phase and amplitude map of the corresponding monofocal lens. Afterwards, a portion of each nano-pillar is replaced with GSST with the thickness and relative position of this inner GSST layer as the controlling factors. The metasurface is designed as a lens focusing to a focal point of f 2 = 7000 nm in this phase. In this article, the operating wavelength is considered to be \(\lambda\) = 1570 nm and all simulations are performed with a full vectorial 3D finite-difference time-domain method. Focusing behavior is 2D with spot sizes of approximately 780 nm and 1850 nm along the X and Y directions for f 1 = 4000 nm and 1002 nm and 2520 nm along the X and Y directions for f 2 = 7000 nm, respectively. Moreover, in this article, the feasibility of the proposed structure and the robustness of its response to inevitable fabrication tolerances are also discussed.