Induction of Synaptic Plasticity in Flexible Organic Synaptic Transistors with Cross‐Linked Polymer Dielectric

Somnath Bhattacharjee,Shree Prakash Tiwari
DOI: https://doi.org/10.1002/admt.202400552
IF: 6.8
2024-06-19
Advanced Materials Technologies
Abstract:Controlled cross‐linking of poly (4‐vinylphenol) with lower concentrations of methylated poly (melamine‐co‐formaldehyde) significantly enhances the memory and synaptic performance of flexible organic synaptic transistors (OSTs). Fabricated OSTs mimic biological synapses and exhibit low energy consumption and suitability for logic implementation for complex neuronal behaviors. These findings can revolutionize toward implementation of neuromorphic computing hardware. Controlled cross‐linking of polymer dielectric poly (4‐vinylphenol) (PVP) is demonstrated as an effective tool in enhancing the performance of flexible organic synaptic transistors (OSTs). Investigation of variation of concentration of the cross‐linking agent methylated poly (melamine‐co‐formaldehyde) (PMCF) in PVP in bilayer combination with high‐k hafnium oxide (HfO2) as gate dielectric in devices shows that the lower concentration of cross‐linking agent results in better memory performance. OSTs with 26% PMCF concentration in PVP (by mass) exhibit excellent memory performance with memory window > 4 V for VGS sweep of ±5 V, static retention of ≈104 s, dynamic retention for 500 cycles, and ≈125 continuous program/erase cycles. Pulse paired facilitation with relaxation time constants of 370 and 4670 ms respectively for slow and rapid phases with regulating modulation amplitude of ≈1 resemble a biological synapse. Through excitatory post synaptic current characteristics, spike timing dependant plasticity and spike voltage dependant plasticity are clearly observed, with low energy consumption per spike on the order of 10 pJ. Further, by leveraging the intricate interconnected data transfer and computation phenomenon, "AND" logic is effectively implemented using these OSTs. These exciting results may open up new directions toward the development of hardware for neuromorphic computing.
materials science, multidisciplinary
What problem does this paper attempt to address?