All‐Optical Nonlinear Neuron Based on Metallic Quantum Wells
Junru Niu, Hua Shao, Yiming Feng, Bingtao Gao, Yiyun Zhang, Ying Li, Hongsheng Chen, Haoliang Qian · Advanced Optical Materials · 2023
Abstract Optical nonlinear neuron, as an essential implementation method for optical interconnection, could greatly promote the development of optical neuron networks, which appears to be a promising alternative to electronic neural networks that are limited by computing speed and the end of Moore's law. However, restricted by weak nonlinearities, the modulation performance of optical nonlinear neuron networks is much lower compared to their electronic counterparts. Therefore, an effectual all‐optical nonlinear neuron is proposed by combining a double‐ring resonator with field‐enhancement and metallic quantum wells (MQWs) with a large Kerr nonlinear response. Based on the large and ultrafast Kerr nonlinearity provided by MQWs, the device can achieve a modulation extinction ratio of 18.78 dB and a transition rate of 13 GW cm−2 per 3 dB. The work may provide a new route for integrated, ultrafast, and high‐efficient optical nonlinear neurons for on‐chip neuron networks.