From Synaptic to Neuronal Functionality Using Ferroelectric Tunnel Junctions

Parallel information processing, energy efficiency and unsupervised learning make the human brain a model computing system for unstructured data handling. Oxide memristors, from filamentary to metal-insulator transition devices have been shown to emulate synaptic and neuronal functionalities in artificial neuromorphic circuits. However, challenges like non-linear conductance update, cycle-to-cycle and device-to-device variability and leakage current related issues in a dense crossbar structure are still not resolved. For neurons, often a circuit with multiple active or passive components are required which can complicate circuit designs in a large-scale neural network. In our earlier work, we reported solution-processable ferroelectric tunnel junctions (FTJs) with P(VDF-TrFE) copolymer barriers on semiconducting bottom electrode can show analog memristive behavior with a broad range of accessible conductance states and low energy dissipation of 100 fJ for the onset of depression and 1 pJ for the onset of potentiation by resetting small tunneling currents on nanosecond timescales. Key synaptic functions like programmable synaptic weight, long and short-term potentiation and depression, paired-pulse facilitation and depression, and Hebbian and anti-Hebbian learning through spike shape and timing-dependent plasticity are demonstrated. 1 In the current work, we demonstrate that by manipulating the carrier concentration of the semiconducting bottom electrode in FTJs, it is possible to control the dynamics of ferroelectric domain rotations in a way so that from synaptic devices they can act similar to neuronal devices. 2 These results offer a promising outlook for the FTJ memristors on semiconducting bottom electrodes as both synapse and neuron devices in artificial neural networks by controlling carrier doping concentrations only.