Biomimetic Sensors Based On Ontoelectronic Behaviors Of Cerium Oxide Thin Film

Intelligent visual perception systems play an important role in a wide variety of fields such as autonomous vehicles,robotics,and industrial manufacturing.However,due to the physically separated architecture of sensing and processing units,machine vision systems based on traditional electronic devices face great limitations such as high latency and power consumption during data transmission,which are not conducive to efficient processing and real-time decision-making.On the contrary,human retina can simultaneously sense and preprocess visual information,which effectively reduces the transmission load of redundant data and greatly improves the processing efficiency of the human brain.More importantly,human retina has visual adaptation and photonic nociceptive perception capabilities.On the one hand,visual adaptation can not only prevent the visual system from saturating under strong light,but also discard a large number of irrelevant background signals to focus attention on the vital visual information for human beings,which is very crucial for efficiently processing a large amount of visual data.On the other hand,the photonic nociceptive perception ability plays a critical role in protecting the retina from light damage,which makes human eyes have self-protection functionality and can work stably in various service environments.Constructing new optoelectronic devices to realize the above-mentioned information perception and processing functions at the hardware level is of great significance for the development of low-power,high-efficient intelligent neuromorphic bionic vision systems.However,the realization of such biomimetic devices is a severe challenge at present.The content of this thesis focuses on the above problems.First,the regulation rule of light and electric field on the space charge region of the metalsemiconductor interface is studied,and the visual adaptation function is simulated by using a photoelectric synergy approach.Further,the dynamic-trapping effect of buried defects within the dielectric layer is used to emulate the function of the photonic nociceptor in human eyes.The specific work is as follows:(1)Research on the construction and photoelectric properties of CeO2-x-based photodetector.The typical photosensitive semiconductor CeO2-x with abundant defects is preferred as the dielectric layer,and a photovoltaic photodetector based on CeO2-x/Pt Schottky junction is constructed.By adjusting the charge state of oxygen vacancies in the space charge region of the interface using illumination,the dynamic modulation of the built-in electric field intensity of the device is realized,and the photocurrent attenuation behavior similar to the visual adaptation process is obtained.The result provides a material and device basis for developing advanced intelligent optoelectronic information processing system that integrates sensing and processing functions.(2)Research on the biomimetic adaptive sensor based on the photoelectric effect of ITO/CeO2-x/Pt interface.Photovoltaic and photoconductive effects are simultaneously introduced at the device interface,and the visual adaptation process,including transient photo-excitation,dynamic adaptation and off response,is realistically simulated in a single device unit by using photo-excitation induced by photovoltaic effect and suppression dominated by interface charge detrapping under illumination.By changing the parameters such as the interval time and the onset speed of light pulses,the visual adaptation habituation and dynamic speed recording characteristics are simulated.The functional demonstration,including environmental background signal filtering and selective detection of moving objects,is further explored.The result provides a new strategy for constructing a highly brain-like neuromorphic bionic vision system.(3)Research on the biomimetic photonic nociceptor based on the dynamictrapping effect of buried defects within the CeO2-x dielectric layer.Inspired by the working principle of the photonic nociceptor in human eyes,a biomimetic photonic nociceptor based on the dynamic-trapping effect of buried defects within the CeO2-x dielectric layer is designed and demonstrated.Key characteristics of the photonic nociceptor,including threshold,relaxation and sensitization,are simulated in a single device unit by changing the intensity,frequency and duration of the external light pulse.In particular,the threshold light intensity(TG=1.2 pW/μm2)and the color characteristics of the threshold response(TR/TB=36.25)of the device are highly consistent with those of the photonic nociceptor in human eyes.Furthermore,the threshold light intensity of the device can be readily manipulated using an external voltage,resembling the ambient luminance-dependent tunability of nociceptive threshold of the human visual system.The biomimetic photonic nociceptor has broad application prospects in the fields of visual prosthesis,artificial eyeballs and humanoid robots in the future.