Gain-Layer-on-Waveguides for All Optical Spiking Dynamics

Neuromorphic photonics is set to transform AI, enabling systems to interact and adapt to the ever growing computational demand with unprecedented speed, efficiency and accuracy. This need arises due to the fact that traditional electronic models currently face challenges in scalability and energy efficiency to meet demand. We introduce the innovative use of gain-layer-on-waveguides as bio-inspired, spiking synapses, leveraging flexible sol-gel fabrication for moderate on-chip amplification in a spiking neural network. We analyze the behavior of an erbium doped gain-layer-on-waveguide when activated by a pulsed pump-signal input, delving into the dynamic responses essential for neuromorphic computing. Our studies reveal how these waveguides handle non-linear pulse transformations of input signals, crucial for mimicking neuronal behavior and enhancing propagation through spiking neural network (SNN).