Fig. 1. Schematic of a two-WG inhomogeneous PT symmetric system in the x–y plane. When the PT phase changes very slowly, is it feasible to get a resonance in the exact PT region?

Fig. 2. Distribution of field amplitude E_z at the frequencies of (a) 227 GHz and (b) 215 GHz, respectively.

Fig. 3. (a) Distribution of field amplitude from the COMSOL simulation at 227 GHz. (b) The calculated real part of k_PT in the structure. At the central of the structure, the PT phase is exact, while across EPs, the PT phase is broken.

Fig. 4. (a) Distribution of energy flux S_x inside the two WGs at 227 GHz. (b) Distribution of the two-dimensional energy flux in the xy plane. Green arrows (with a number of M = 8) at the lower blank space represent the direction of energy flux in the coupled WGs.

Fig. 5. (a) Maximum field E_max and (b) the effective resonator length L_eff versus f.

Fig. 6. Source being placed outside the PT symmetric region can still excite the non-Hermitian resonances at (a) 227 GHz and (b) 215 GHz.