Integrated silicon photonic device design by attractor selection mechanism based on artificial neural networks: optical coupler and asymmetric light transmitter

Abstract

Recently, different nanophotonic computational design methods based on optimization algorithms have been proposed which revolutionized the conventional design techniques of photonic integrated devices. The intelligently designed photonic devices have small footprints and high operating performance along with their fabrication feasibility. In this study, we introduce a new approach based on attractor selection algorithm to design photonic integrated devices. In order to demonstrate the potential of the proposed approach, we designed two structures: an optical coupler and an asymmetric light transmitter. The designed photonic devices operate at telecom wavelengths and have compact dimensions. The designed optical coupler has a footprint of only 4 × 2 μm 2 and coupling efficiency of 87.5% at a design wavelength of 1550 nm with spatial beam width compression ratio of 10:1. Moreover, the designed optical coupler operates at a wide bandwidth of 6.45% where the transmission efficiency is above 80%. In addition, the designed asymmetric light transmitter with a size of 2 × 2 μm 2 has the forward and backward transmission efficiencies of 88.1% and 8.6%, respectively. The bandwidth of 3.47% was calculated for the designed asymmetric light transmitter where the forward transmission efficiency is higher than 80% and the backward efficiency transmission is under 10%. In order to evaluate the operating performance of the designed photonic devices, coupling losses are analyzed. The presented results show that the attractor selection algorithm, which is based on artificial neural networks, can bring a conceptual breakthrough for the design of efficient integrated nanophotonic devices.

Digitized subwavelength surface structure on silicon platform for wavelength-/polarization-/charge-diverse optical vortex generation

“…Thus we call it digitized subwavelength surface structure on silicon platform. This type of digitized nanophotonic structure has been increasing popular to design photonic integrated devices with arbitrary topologies used for multiple novel functionalities [68][69][70][71][72][73][74][75][76]. Unlike conventional design techniques with only a small amount of parameters optimized one by one, the digitized design based on optimization algorithms allows the full parameter space searching of possible structures even without user intervention, which could be seen as a promising technical innovation.…”

Deep learning: a new tool for photonic nanostructure design

“…The surrogate model of the SWG reported by the authors worked 1830 times faster than exact numerical simulations with 93.2% accuracy of the simulations. Bor and co-workers 77 introduced a new approach based on the attractor selection algorithm to design photonic integrated devices showing improved performance compared to traditional design techniques; specifically, an optical coupler and an asymmetric light transmitter were designed. Gabr and co-workers 78 considered the design of four common passive integrated devices (waveguides, bends, power splitters and couplers) with a forward DNN; they reported split-second evaluation speeds with errors less than 2%.…”

Tuning Topology of Photonic Systems with Transparent Conducting Oxides

“…A possible way of solving this issue is to use either more advanced high-efficiency modulator designs or to optimize TCO material composition that would provide the same order of tunability at a lower attenuation level. High-efficiency modulator designs can be obtained based on conventional optimization techniques or by using inverse design optimization methods, such as topology optimization and machine learning assisted inverse design optimization . The optical properties of TCOs can be efficiently adjusted via doping and tuning the fabrication parameters, , providing an additional degree of freedom for designing highly efficient, low-loss modulators.…”

“…High-efficiency modulator designs can be obtained based on conventional optimization techniques or by using inverse design optimization methods, such as topology optimization 53 and machine learning assisted inverse design optimization. 54 The optical properties of TCOs can be efficiently adjusted via doping and tuning the fabrication parameters, 55,56 providing an additional degree of freedom for designing highly efficient, lowloss modulators. For example, AZO permittivity is strongly dependent on oxygen pressures upon deposition.…”