K. Wang

With the development of neuromorphic electronics, much effort has been devoted to expand perception, memory, and computing integration capabilities. In this paper, an ionic-based graphene synaptic transistor with long-gate structure has been investigated to mimic memory, learning function and perceive humidity. By harnessing the tunable in-plane-field transport of charge carriers in graphene and ions motion in ion-gel, this transistor mimics various synaptic functionalities, including inhibitory postsynaptic current, excitatory postsynaptic current, paired-pulse facilitation, long-term depression, and long-term potentiation. Under short pules stimuli, the long-gate structure provides our transistor with an inertial assisted re-accumulation, generating two excitatory postsynaptic current peaks and enhanced paired-pule facilitation up to ∼265%. Furthermore, the presence of the long-gate structure enables our transistor to exhibit excellent learning and simulate Ebbinghaus' memory. In addition, physical mechanic about its humidity perception has been analyzed and discussed. This study provides a unique platform for designing high-performance carbon-based artificial synapses enabling integrated functions of sensing, storage, and computation for the neuromorphic system.