Research On The Optoelectronic Properties Of Ferroelectric-gated MoS₂ Field Effect Transistors

With the advent of the Internet of Things and the era of big data,there is an increasing demand for more efficient and intelligent optoelectronic devices for information perception,processing,and storage.Compared with other heterostructure devices,ferroelectric-gated two-dimensional transistors have significant advantages in the multi-functional design due to its simple configuration,rich and flexible modulation,and excellent optoelectronic characteristics,which are expected to play an important role in innovative optoelectronic application.However,the research on different two-dimensional ferroelectric heterostructure devices for multifunctional integration in emerging optoelectronic applications is deficient recently.Considering these,this thesis has constructed ferroelectric-gated MoS₂ field effect transistors based on LiNbO₃/HfO₂/MoS₂ structure,and investigated their properties for diverse optoelectronic functions.The main results are as follows:1.Based on LiNbO₃/HfO₂/MoS₂ devices,photodetection and memory functions have been successfully implemented.Through photoelectric testing,the device showed a broadband response of 395 nm to 808 nm,while observing the phenomenon of photocurrent relaxation.Based on this,the photoelectric memory with 532 nm optical program and negative gate voltage erasure was realized in the thesis,achieving retention time more than 10⁴ s and switch cycle more than 40 times.Nonvolatile photocurrent and electrical erasure phenomena are attributed to carrier traps at the HfO₂/MoS₂ interface and the channel depletion modulated by ferroelectric polarization.The result lays the foundation for the subsequent multifunctional implementation of photoelectric modulation.2.Benefit from the controllable photoelectric induction mechanism,optoelectronic synaptic functions and neuromorphic applications have been successfully realized based on LiNbO₃/HfO₂/MoS₂ devices.By controlling external photoelectric pulse stimuli parameters,such as interval,intensity,duration,and numbers et al,various synaptic properties could be mimicked,including paired pulse facilitation,plasticity transition,and long-term plasticity et al.Furthermore,the linear conductance was obtained,which provided the basis for synaptic weight update to implement handwritten digital image recognition tasks in the artificial photoelectric neural network.After training,the image recognition accuracy achieved 87%.In addition,Image learning and memory behavior of human visual system was simulated successfully based on 3×3 synaptic array.These results indicate that ferroelectric-gated MoS₂ field effect transistors are available to simulate the visual synapses,which are expected to be applied to the next generation of intelligent artificial vision systems.3.Based on LiNbO₃/HfO₂/MoS₂ devices,optical logic operations have been implemented via regulation,which enabled different image processing functions to perform.The dynamic configuration of"OR"and"AND"logic operations were accomplished through positive and negative gate voltage modulation and different optical inputs.Then,two logical operations were used to demonstrate the completion of feature fusion and intersection processing operations for input images.The image processing hardware scheme has the advantages of multi-function integration and high area efficiency.