Silicon Nanowire Polarization Rotator

The schematic shown in Fig. 1a illustrates the polarization rotator studied in this application note. It includes three parts: a polarization input section (set as TE in this case), a polarization rotation section, and a polarization output section (TM in this case). For the polarization rotator, the first two modes are sufficiently hybrid that their optical axes are rotated 45° with respect to the fixed coordinates of the device (x,y in Fig.1b), respectively. When linearly polarized TE light is launched into this longitudinally invariant waveguide, the two modes are excited with almost equal amplitude. At each half beat-length, full 90°-polarization rotation is achieved and TM output is achieved. The half beat-length Lπ is determined from the propagation constants β1, β2 of the two first modes as L_π = π/( β1 - β2).

To achieve a shorter half-beat length Lπ, a high-index contrast waveguide is preferred. To give a detailed analysis, we will use a FullWAVE simulation to compute the Polarization Conversion Efficiency (PCE), which is defined as:

Figure 4: The TE (top row) and TM (bottom row) modes calculated in the rotator section. These modes were calculated using the FEM-based FemSIM mode solver.

Note that the fields in the Fig 4 vector plots are tilted at approximately +/- 45 degrees, and that both modes are perpendicular to each other. When linearly polarized light is launched into this waveguide, both of these modes will be excited and, at each half beat-length, a full 90 degree polarization rotation will be achieved.