L-Shaped Double Gate Bipolar Impact Ionization MOSFET Based Energy Efficient Leaky Integrate and Fire Neuron for Spiking Neural Network

In this article, an L-shaped double gate bipolar impact ionization MOSFET (L-DG BIMOS) is proposed and demonstrated as an ultra-low energy artificial leaky integrate and fire (LIF) neuron for spiking neural network. The L-shaped double gate structure offers a higher rate of impact ionization (II) essential for triggering LIF action and hence can efficiently be applicable to mimic a biological neuron. Using a well-calibrated 2D TCAD tool, the functioning of the proposed device as an LIF neuron is manifested by exhibiting a low threshold voltage of 0.2 V to fire a spike, reduced breakdown voltage of 0.68 V and 0.1050 pJ of energy per spike, making the device more energy efficient as compared to conventional neurons. In addition, the spiking frequency of L-DG BIMOS is calculated to be in the gigahertz (Gz) range when the drain is biased at 1.5 V, which is much higher than biological neurons.