Enhancing Neuromorphic Computing: A High-Speed, Low-Power Integrate-and-Fire Neuron Circuit Utilizing Nanoscale Side-Contacted Field Effect Diode Technology

Addressing power efficiency and performance in neuromorphic computing systems is crucial, particularly in minimizing the power usage of neuronal circuits without compromising their ability to rapidly generate spikes. To this end, we introduce the Nanoscale Side-contacted Field Effect Diode (S-FED), a promising solution for reducing power consumption and improving circuit speed, thereby facilitating the implementation of neuron circuits. This paper presents a novel integrate-and-fire (IF) neuron model that harnesses the capabilities of the S-FED. Our model lowers power usage to approximately 44 nW and reduces the energy per spike to 0.964 fJ, setting a new benchmark below current state-of-the-art neuron circuits. Additionally, our model's spiking frequency reaches 20 MHz, surpassing those of the existing neuronal circuits. Finally, we compare the performance of the proposed IF neuron circuit against the forefront of neuron circuit design.