Unconventional Materials-Based Flexible Synaptic Transistors for Sustainable Neuromorphic Systems: Advancements, Challenges, and Future Trends

Click to copy article linkArticle link copied!

Synaptic transistors have lately emerged as a viable method for the implementation of hardware-based neuromorphic systems, emulating a biological synapse, the junction between two adjacent neurons. These artificial synaptic devices operate on minimal power (in the femtojoule range) while possessing the capability for parallel computation, hence offering the possibility to surmount the von Neumann bottleneck. The research articles published in the past decade have mostly focused on conventional transistors to emulate artificial synaptic function, which are neither sustainable nor flexible. The need for a sustainable environment has necessitated the adoption of benign and eco-friendly materials. Unconventional materials, including natural polymers, biodegradable substances, and waste-derived materials, have lately garnered attention for the fabrication of these devices. The primary objective is to attain neuromorphic functions with particular emphasis on device designs. This review mainly examines sustainable materials, including gelatin, chitosan, and cellulose-based polymers and biodegradable substrate-based devices, which have recently been reported to emulate synaptic behavior and attain neuromorphic functionalities. Moreover, this review briefly addresses details of synaptic activities such as short-term plasticity, long-term plasticity, and pulse paired facilitation, among others. Advanced AI applications, including pattern recognition and on-device learning, have been examined alongside circuit-level applications, such as various logic gates. The environmental impact of these synaptic transistors has also been discussed at the end of their lifespan. Through a detailed discussion of these topics, a comprehensive perspective has been provided to researchers across several domains, and a future pathway toward development of flexible and low-cost neuromorphic systems has been addressed.