Solution-Processed\nPolymer Memcapacitors with Stimulus-Controlled\nand Evolvable Synaptic Functionalities: From Short-Term Plasticity\nto Long-Term Plasticity to Metaplasticity

In the vanguard of neuromorphic engineering,\nwe develop a paradigm\nof biocompatible polymer memcapacitors using a seamless solution process,\nunleashing comprehensive synaptic capabilities depending on both the\nstimulation form and history. Like the human brain to learn and adapt,\nthe memcapacitors exhibit analogue-type and evolvable capacitance\nshifts that mirror the complex flexibility of synaptic strengthening\nand weakening. With increasing frequency and intensity of the stimulation,\nthe memcapacitors demonstrate an evolution from short-term plasticity\n(STP) to long-term plasticity (LTP), and even to metaplasticity (MP)\nat a higher level. A physical picture, featuring the stimulus-controlled\nspatiotemporal ion redistribution in the polymer, elaborates the origin\nof the memcapacitive prowess and resultant versatile synaptic plasticity.\nThe distinctive MP behavior endows the memcapacitors with a dynamic\nlearning rate (LR), which is utilized in an artificial neural network.\nThe superiority of implementing a dynamic LR compared with conventional\npractices of using constant LR shines light on the potential of the\nmemcapacitors to exploit organic neuromorphic computing hardware.