Nanoionics-Based Three-Terminal Synaptic Device Using Zinc Oxide

Artificial synaptic thin film transistors (TFTs) capable of simultaneously manifesting signal transmission and self-learning are demonstrated using transparent zinc oxide (ZnO) in combination with high κ tantalum oxide as gate insulator. The devices exhibit pronounced memory retention with a memory window in excess of 4 V realized using an operating voltage less than 6 V. Gate polarity induced motion of oxygen vacancies in the gate insulator is proposed to play a vital role in emulating synaptic behavior, directly measured as the transmission of a signal between the source and drain (S/D) terminals, but with the added benefit of independent control of synaptic weight. Unlike in two terminal memristor/resistive switching devices, multistate memory levels are demonstrated using the gate terminal without hampering the signal transmission across the S/D electrodes. Synaptic functions in the devices can be emulated using a low programming voltage of 200 mV, an order of magnitude smaller than in conventional resistive random access memory and other field effect transistor based synaptic technologies. Robust synaptic properties demonstrated using fully transparent, ecofriendly inorganic materials chosen here show greater promise in realizing scalable synaptic devices compared to organic synaptic and other liquid electrolyte gated device technologies. Most importantly, the strong coupling between the in-plane gate and semiconductor channel through ionic charge in the gate insulator shown by these devices, can lead to an artificial neural network with multiple presynaptic terminals for complex synaptic learning processes. This provides opportunities to alleviate the extreme requirements of component and interconnect density in realizing brainlike systems.