Bidirectional synapse based on anti-resonance modulation of photonic microring

The synaptic device is fundamental for memory and learning functions of neural networks. In this work, we demonstrated a GST(Ge2Sb2Te5)-based microring synapse with non-volatile reconfigurable characteristics. The device shows bidirectional weight modulation during unidirectional crystallization or amorphization of GST, simulating long-term depression (LTD) and long-term potentiation (LTP). In addition, we adopted an anti-resonance pump scheme to reduce the fluctuations in programming pump energy to 12%. This scheme significantly enhances the multi-weight programming capability of the microring synapse, achieving continuous programming of 64 highly linear states. This work holds promise for laying the foundation for novel photonic computing architectures and provides possibilities for the implementation of vision and adaptive optical neural networks that rely on bidirectional plasticity.