Computer vision is a core technology that enables emerging applications such as autonomous driving, intelligent robotics, and smart manufacturing. Neuromorphic vision computing dramatically breaks the arithmetic bottleneck by simplifying and optimizing data transfer between photodetectors and computation units. More importantly, neuromorphic vision sensors hold substantial potential for mimicking the environment adaptation and in-memory sensing capability of human vision system. However, the limited responsivity and data retention time of all-in-one neuromorphic sensors usually hinder their potential in multispectral machine vision, especially in the near-infrared (NIR) band which contains critical information for pattern recognition.
In a new paper published in Light: Science & Applications, a team of scientists, led by Professor Shaojuan Li from State Key Laboratory of Luminescence Science and Technology, Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences, China, and co-workers have demonstrated physisorption-assistant optoelectronic synaptic transistors based on Ta 2 NiSe 5 /SnS 2 heterojunction, which present tunable synaptic functionality in broadband (375-1310 nm). The strategy utilizes the localized state caused by the physical adsorption of gas molecules in the air to extend the carrier lifetime. The abundant intrinsic sulfur vacancies in SnS 2 can be used as gas adsorption sites and extend the carrier lifetime. Besides, these defects can effectively improve the performance of optoelectronic synapses by capturing and confining photogenerated carriers. As a ternary chalcogenide with great potential in infrared detection, Ta 2 NiSe 5 was used as NIR absorber to improve the adsorption for optoelectronic synaptic transistors and further enhance NIR light performance. Accordingly, the Ta 2 NiSe 5 /SnS 2 heterojunctions were explored for the application of photonic synapses from ultraviolet (UV) to NIR. The structure of the device and characterizations of the photodetection properties are shown in Figure 2.
Due to the synergistic effect of the energy band structure of Ta 2 NiSe 5 /SnS 2 heterojunction and the adsorption of gas molecules on the surface of SnS 2 (Figure 2d), the persistent photoconductivity effect of the device is enhanced, which can produce a remarkable photoresponse to light illumination in the range of 375-1310 nm (Figure 3a-c). Meanwhile, the photocurrent of the device exhibits excellent non-volatility, and this non-volatility can be effectively modulated by light pulses, which can mimic the neural behaviors produced by the retinal cells of the human eye when they receive light signals. Therefore, the device can be applied as an optoelectronic synapse in neuromorphic visual systems. As shown in Figure 3d-e, based on the significant difference in synaptic behaviors of the device for different wavelengths, the recognition and memory storage functions for RGB-light in the range of 2-13 μW light intensity were achieved. The broadening of the wavelength response range will help the extraction of image features and improve the accuracy of the recognition task.
Light Science & Applications
Physisorption-Assistant Optoelectronic Synaptic Transistors Based on Ta2NiSe5/SnS2 Heterojunction from Ultraviolet to Near-Infrared