Organic-Based Devices for Neuromorphic Computing

Polymer-based artificial synapses have shown outstanding performance in terms of switching speed, switching energy and endurance. The working principle of these devices leverages the dynamics of ion diffusion in polymers. The same dynamics can be used to fabricate organic circuits that mimic spiking neurons that exhibit adaptive behavior. These circuits comprise p- and n-type devices which are fabricated separately on chips and subsequently connected. Furthermore, the same materials set is used to fabricate multi-gate devices that reproduce the features of dendrites, specifically spatial and temporal pulse sequences. Device and circuit design allows to tune the temporal response of organic synapses, neurons and dendrites. Semiconducting polymers are thus attractive to fabricate all components suitable for neuromorphic computing.