Artificial neural networks have shown effectiveness in the inverse design of nanophotonic structures; however, the numerical accuracy and algorithm efficiency are not analyzed adequately in previous reports. In this Letter, we demonstrate the convolutional neural network as an inverse design tool to achieve high numerical accuracy in plasmonic metasurfaces. A comparison of the convolutional neural networks and the fully connected neural networks show that convolutional neural networks have higher generalization capabilities. We share practical guidelines for optimizing the neural network and analyzed the hierarchy of accuracy in the multi-parameter inverse design of plasmonic metasurfaces. A high inverse design accuracy of ± 8 n m for the critical geometrical parameters is demonstrated.

A reverse ridge/slot hybrid chalcogenide glass (As2S3) waveguide with two vertical silicon dioxide slots is proposed in this work. The fundamental quasi-TE mode of the waveguide shows an ultra-flat dispersion with three zero-dispersion wavelengths. Its dispersion is confined between -26 and + 27 ps/nm/km over a bandwidth of1370 nm (from 1770 nm to 3140 nm). Two slots in the waveguide introduce the dispersion tailoring effect on the quasi-TE mode, which provides more flexibility for designing the waveguide dispersion. The nonlinear coefficient and the phase mismatching of the degenerate four-wave mixing (FWM) process in this waveguide are calculated showing that it can support broadband FWM processes in near-infrared and middle-infrared regions. Index Terms: Chalcogenide glass waveguides, reverse ridge/slot waveguide, waveguide dispersion, waveguide nonlinearity 1943-0655 (c)

We propose a fabrication method for chalcogenide glass (ChG) waveguides based on hot melt smoothing and micro-trench filling. ChGs has low melting points and good flowability. Experiments show that this method can realize high quality As2S7 glass waveguides with reverse ridge structures. The attenuations of the fundamental quasi-TE mode and quasi-TM mode are 0.1 and 0.9 dB/cm, respectively. This method avoids fabrication processes, such as photolithography, lift-off, and dry or wet etching that are directly applied to the ChG films. It provides a simple way to fabricate high quality ChG waveguides, which have great potential for applications in integrated nonlinear optical devices.

No abstract

The design steps of a millimeter waveguide filter will be presented. A sample filter at 35.75GHz will be designed as an example to show how to use the formulae. Tested results from the sample filter will be presented.