Design and optimization of photonic crystal based eight channel dense wavelength division multiplexing demultiplexer using conjugate radiant neural network

In this paper, two dimensional photonic crystal, based eight-channel demultiplexer is proposed and designed for DWDM applications. The performance parameters of the demultiplexer such as transmission efficiency, channel spacing, spectral line width, Q factor, and crosstalk have been evaluated. The proposed demultiplexer comprises of bus waveguide, drop waveguide and parellogram resonant cavity (PRC). The bus waveguide transmits light to the PRC and exits through respective drop waveguide. The PRC consists of a parellogram resonator with a nano ring cavity that is used for dropping eight specific wavelength for ITU-T G 694.1 standard with 50 GHz channel spacing. The circular ring resonator is placed above the PRC wherein a resonant air hole (Cr) is positioned for desired channel selection. The channel selection is done by altering the radius of the air hole. In addition, a conjugate radiant neural network is implemented for optimizing the radii of resonant air holes to select the required channel wavelength. The proposed device is very compact and it could be considered for implementing the photonic integrated circuits.