Enhanced Multiwavelength Response of Flexible Synaptic Transistors for Human Sunburned Skin Simulation and Neuromorphic Computation

Abstract

In biological species, optogenetics and bioimaging work together to regulate the function of neurons. Similarly, the light‐controlled artificial synaptic system not only enhances computational speed but also simulates complex synaptic functions. However, reported synaptic properties are mainly limited to mimicking simple biological functions and single wavelength responses. Therefore, the development of flexible synaptic devices with multiwavelength optical signal response and multifunctional simulation remains a challenge. Here, flexible organic light‐stimulated synaptic transistors (LSSTs) enabled by AlOX with a simple fabrication process are reported. By embedding AlOX nanoparticles, the excitons separation efficiency is improved, allowing for multiple wavelength responses. Optimized LSSTs can respond to multiple optical and electrical signals in a highly synaptic manner. Multi‐wavelength optical synaptic plasticity, electrical synaptic plasticity, sunburned skin simulation, learning efficiency model controlled by photoelectric cooperative stimulation, neural network computing, “deer” picture learning and memory functions have been successfully proposed, which promote the development for future artificial intelligent systems. Furthermore, as prepared flexible transistors exhibit mechanical flexibility with bending radius down to 2.5 mm and improved photo‐synaptic plasticity, which facilitating development of neuromorphic computing and multi‐function integration systems at the device‐level.This article is protected by copyright. All rights reserved

Introducing Chiro-optical Activities in Photonic Synapses for Neuromorphic Computing and In-Memory Logic Operations

“…The effect of the first 40 cycles of optical pulses (LCP light with a duration of 1 s) followed by 40 successive electrical pulses (−1.0 V with a duration of 100 ms) shows optical potentiation succeeded by electrical inhibition in an R -device. Optical potentiation and electrical inhibition are generally utilized in pattern recognition and image processing …”

“…Since the synapses are triggered by optical pulses as an independent source of stimuli, they require low power consumption and offer high computing speed. , The development of synaptic devices with optical and electrical stimuli in tandem has also been envisaged . In the materials domain, a range of optically active materials are being considered. − Metal halide perovskites have turned out to be quite appropriate in designing photonic synaptic devices due to their appropriate optical and electrical properties, such as high extinction coefficient, small exciton binding energy, long carrier lifetime, ambipolarity, and efficient charge transport capabilities . The reports on all-optical photonic synaptic devices use multiple illumination sources with different wavelengths as inputs. − …”

“…17,18 The development of synaptic devices with optical and electrical stimuli in tandem has also been envisaged. 19 In the materials domain, a range of optically active materials are being considered. 17−19 Metal halide perovskites have turned out to be quite appropriate in designing photonic synaptic devices due to their appropriate optical and electrical properties, such as high extinction coefficient, small exciton binding energy, long carrier lifetime, ambipolarity, and efficient charge transport capabilities.…”

Vertically Integrated Monolithic Neuromorphic Nanowire Device for Physiological Information Processing

“…These characteristics hold promise for applications in neuromorphic computing. 41,42 ■ BIOMEDICAL APPLICATIONS Human physiological signals are crucial indicators of tissue status, physiological functions, and their dynamic variations. 43 These signals exhibit bidirectional fluctuations from their baseline conditions.…”

“…The p-type synaptic transistor displayed similar potentiation/depression characteristics when potentiation and depression pulses were sequentially applied at V bias = −1 V (Figure S11). These characteristics hold promise for applications in neuromorphic computing. , …”