Reconfigurable Logic Circuits Based on Organic Electrochemical Neuromorphic Devices

Owing to the increasing demands of data-intensive applications, there exists an imperative need for next-generation electronic systems that are energy-efficient and possess low power consumption, facilitated by innovative architectures. In this study, we present the development of reconfigurable circuits built on organic neuromorphic devices. Our research focused on leveraging the learning capabilities of an organic electrochemical diode to execute multiple logic functions through high-performance reconfigurable logic circuits. A key feature of these devices is the tunability of their conductivity state, which can be modulated in response to the polarity of the drain-source voltage (<i>V</i>_DS), while the selective activation of various elements is governed by the manner in which the input voltage (<i>V</i>_logic) is utilized. Consequently, such manipulation facilitates the transition from an undefined state to selectively performing as either an AND or OR gate within the same hardware configuration. Additionally, it was demonstrated that the capacity for more sophisticated operations by integrating learnable AND and XOR circuits to construct a half-adder. We expect that our findings will provide an efficient approach toward novel digital circuitry through on-chip learning, thus surmounting the efficiency bottleneck that exists in conventional computing systems.