Multiple Ion-channel ISFET Neuron for Lab-on-chip applications

Abstract

Ion Sensitive Field Effect Transistor (ISFET) architectures for Lab-on-chip (LoC) applications have continued to evolve at a rapid pace over the past few decades. Multiple implementations have been presented towards making these devices low powered and robust. The field of neuromorphic electronics has the potential to be a great approach in improving ISFET architectures. Since both neurons and ISFETs respond to variation in ion concentrations, there is a great opportunity to combine neuromorphic circuits with potentiometric ISFET sensing. This work aims to present a novel approach to creating smarter LoC devices by emulating the electrochemical behaviour of neurons using a multiple ion channel ISFET architecture. The paper presents an ultra low power architecture with a power consumption per pixel of 191.4 nW to 1.04 µL W and a sensitivity of 47.8 KHz/dpH to 402 KHz/dpH. This is accompanied with a hardware implementation of a backpropogation algorithm for the learning of the ISFET neurons.