Passive All-Optical Nonlinear Neuron Activation via PPLN Nanophotonic Waveguides

Abstract: AI is transforming modern life, yet the growing scale of AI applications places mounting demands on computational resources, raising sustainability concerns. Photonic integrated circuits (PICs) offer a promising alternative, enabling massive parallelism, low latency, and reduced electrical overhead, particularly excelling in high-throughput linear operations. However, passive and fully optical nonlinear activation functions with equally superb performance remain rare, posing a critical bottleneck in realizing all-optical neural networks on PICs. Here, we demonstrate a compact and readily integrated all-optical nonlinear activation function, experimentally realized through highly pump-depleted second-harmonic generation (SHG) in periodically poled lithium niobate (PPLN) nanophotonic waveguides, achieving 79% absolute conversion efficiency. This activation exhibits a sigmoid-like, wavelength-selective response with femtosecond-scale dynamics and light-speed processing, requiring no electrical control or auxiliary optical signals. We further validate its feasibility for neural inference by combining the measured SHG-based nonlinearity with linear operations implemented via a Mach-Zehnder interferometer system on a silicon PIC. Our demonstration achieves performance on par with digital implementations in real-world tasks, including airfoil regression and medical image classification. These results pave the way toward scalable, high-speed, and fully integrated all-optical neural networks for next-generation photonic AI hardware.