While deep learning has demonstrated tremendous potential for photonic device design,it often demands a large amount of labeled data to train these deep neural network models.Preparing these data requires high-resolut...While deep learning has demonstrated tremendous potential for photonic device design,it often demands a large amount of labeled data to train these deep neural network models.Preparing these data requires high-resolution numerical simulations or experimental measurements and cost significant,if not prohibitive,time and resources.In this work,we present a highly efficient inverse design method that combines deep neural networks with a genetic algorithm to optimize the geometry of photonic devices in the polar coordinate system.The method requires significantly less training data compared with previous inverse design methods.We implement this method to design several ultra-compact silicon photonics devices with challenging properties including power splitters with uncommon splitting ratios,a TE mode converter,and a broadband power splitter.These devices are free of the features beyond the capability of photolithography and generally in compliance with silicon photonics fabrication design rules.展开更多
Integrated lithium niobate(LN)photonics is a promising platform for future chip-scale microwave photonics systems owing to its unique electro-optic properties,low optical loss,and excellent scalability.A key enabler f...Integrated lithium niobate(LN)photonics is a promising platform for future chip-scale microwave photonics systems owing to its unique electro-optic properties,low optical loss,and excellent scalability.A key enabler for such systems is a highly linear electro-optic modulator that could faithfully convert analog electrical signals into optical signals.In this work,we demonstrate a monolithic integrated LN modulator with an ultra-high spurious-free dynamic range(SFDR)of 120.04 dB·Hz^(4/5)at 1 GHz,using a ring-assisted Mach–Zehnder interferometer configuration.The excellent synergy between the intrinsically linear electro-optic response of LN and an optimized linearization strategy allows us to fully suppress the cubic terms of third-order intermodulation distortions(IMD3)without active feedback controls,leading to∼20 dB improvement over previous results in the thin-film LN platform.Our ultra-high-linearity LN modulators could become a core building block for future large-scale functional microwave photonic integrated circuits by further integration with other high-performance components like low-loss delay lines,tunable filters,and phase shifters available on the LN platform.展开更多
基金the Chinese Academy of Sciences(XDB24030600)National Natural Science Foundation of China(12004421,61635013,61675231)+3 种基金Youth Innovation Promotion Association of Chinese Academy of Sciences(2016535)West Light Foundation of the Chinese Academy of Sciences(XAB2017A09)Natural Science Basic Research Program of Shaanxi(2019JQ-447)Research Project of Xi’an Postdoctoral Innovation Base(201903).
文摘While deep learning has demonstrated tremendous potential for photonic device design,it often demands a large amount of labeled data to train these deep neural network models.Preparing these data requires high-resolution numerical simulations or experimental measurements and cost significant,if not prohibitive,time and resources.In this work,we present a highly efficient inverse design method that combines deep neural networks with a genetic algorithm to optimize the geometry of photonic devices in the polar coordinate system.The method requires significantly less training data compared with previous inverse design methods.We implement this method to design several ultra-compact silicon photonics devices with challenging properties including power splitters with uncommon splitting ratios,a TE mode converter,and a broadband power splitter.These devices are free of the features beyond the capability of photolithography and generally in compliance with silicon photonics fabrication design rules.
基金National Natural Science Foundation of China(61922092)Research Grants Council,University Grants Committee(CityU 11204820,CityU 21208219,N_CityU113/20)+1 种基金Croucher Foundation(9509005)City University of Hong Kong(9610402,9610455).
文摘Integrated lithium niobate(LN)photonics is a promising platform for future chip-scale microwave photonics systems owing to its unique electro-optic properties,low optical loss,and excellent scalability.A key enabler for such systems is a highly linear electro-optic modulator that could faithfully convert analog electrical signals into optical signals.In this work,we demonstrate a monolithic integrated LN modulator with an ultra-high spurious-free dynamic range(SFDR)of 120.04 dB·Hz^(4/5)at 1 GHz,using a ring-assisted Mach–Zehnder interferometer configuration.The excellent synergy between the intrinsically linear electro-optic response of LN and an optimized linearization strategy allows us to fully suppress the cubic terms of third-order intermodulation distortions(IMD3)without active feedback controls,leading to∼20 dB improvement over previous results in the thin-film LN platform.Our ultra-high-linearity LN modulators could become a core building block for future large-scale functional microwave photonic integrated circuits by further integration with other high-performance components like low-loss delay lines,tunable filters,and phase shifters available on the LN platform.