Ministarstvo znanosti, obrazovanja i mladih, Hrvatska zaklada za znanostHybrid synaptic transistors for non-volatile visual perception and memory (Seal of excellence under the synergis program-DIGIT.2.2.01) | ||
| voditelj projekta | Christy Roshini Paul Inbaraj | |
| suradnici | Marko Kralj | |
| datum početka | 01.05.2026. | |
| datum završetka | 30.04.2028. | |
Sažetak
Artificial synaptic devices that emulate biological synapses to develop a parallel data processing system are becoming increasingly imperative in the present big-data era. Among the neuromorphic devices, visual synaptics is a must for vision-driven information processing, which is required in robotics and artificial intelligence. The advancement in such devices based on conventional semiconductors leads to a complex device structure, resulting in high power consumption. However, the properties of 2D semiconductors with strong light-matter interaction and tunable bandgap fit the requirements for synaptic transistors. In this project, we couple the highly photosensitive indium selenide (InSe) nanosheets with a ferroelectric polymer, poly(vinylidene fluoride trifluoroethylene) (P(VDF-TrFE)), to demonstrate an energy-efficient visual synaptic transistor with non-volatile memory. Here, the optical signals are directly converted to analog conductance by the photogenerated carriers in InSe, which can be tuned by the ferroelectric polarization, like the changes in synaptic weights. The stable remanent polarization of PVDF-TrFE enables multilevel conduction states and long-term memory retention. The research presents key features such as strong photoresponsivity, nonvolatile visual memory, and adjustable plasticity from short-term to long-term, while operating at low power. The proposed project aims to develop a single compact transistor that combines optical sensing with synaptic plasticity and non-volatile memory. It also involves investigating the synaptic functions and demonstrating visual pattern learning in device arrays. Moreover, the flexibility of InSe and the polymer facilitates their integration on flexible substrates for adaptable andwearable applications. This device can be advanced for in-sensing and in-memory computing systems, as it is scalable, energyefficient, and easily integrated, paving the way for next-generation bio-inspired electronics.

