Two-dimensional Semiconductor Transistors Based On Charge-trapping Mechanism For Prototype Applications

Layered two-dimensional（2D）materials have become an essential candidate for fabricating small-size,low-power electronics,due to the advantages such as atomic-level thickness,adjustable band gaps,and low effective mass.2D-semiconductor transistors can not only alleviate the failure trend of Moore’s law,but also break through the bottleneck of von Neumann architecture in memory and brain-like devices and other applications.The charge trapping/de-trapping phenomenon is the inherent effect of 2D semiconductor transistors.And it can achieve abundant electrical and photoelectric characteristics through accurate modulation of optical and electrical means,which is of great significance for the realization of multifunctional neuromorphic brain-like devices.Therefore,in this paper,2D semiconductor transistors were studied to achieve abundant electrical and photoelectric characteristics,through electrical,optical,and mechanical modulation of charge trapping/de-trapping effect.At the same time,the low-frequency noise technique was used to analyze and verify the relevant mechanism.The prototype applications in multimodal-driven memory and neuromorphic devices are further explored.Firstly,field-effect transistors（FETs）were prepared based on two-dimensional hafnium sulfide（Hf S₂）.With a method of modulating charge trapping through electric and optical fields cooperate,the effective defect states on the dielectric surface can be dynamically regulated to control the carrier trapping and de-trapping process in the channel under the optical field.Based on this,the optical response characteristics strongly dependent on the gate electric field were explored.The low-frequency noise measurement technique was used to analyze the effective interfacial defect state concentration regulated by the gate electric field,verifying the photoelectric responses were modulated by charge trapping/de-trapping events.The short-term plasticity,long-term plasticity,and paired-pulse facilitation for biological synapses had been simulated successfully by gate-dependent photocurrents.And the positive and negative correlations of the synaptic weight index A_n/A₁ with the stimuli frequency and number were achieved under the"same"light stimulation condition,which successfully simulated the dose-modulated drug neuroplasticity.Furthermore,this work also studied high-performance indium selenide（In Se）transistors and further explored the application of In Se floating-gate transistors（FGFETs）for memory.The charge trapping/de-trapping mechanism based on the interface effect of the semiconductor channel and dielectric layer in Chapter 3 has some limitations in achieving the memory characteristics,and most reported electronics to achieve various storage states by adjusting electrical or optical signals.Here,this work proposed an innovative touch-modulated memory,based on the coupling of indium selenide/hexagonal boron nitride/graphene（In Se/h-BN/Gr）van der Waals heterostructures coupling with triboelectric nanogenerator（TENG）.The devices were fabricated using a simple metal mask method with high mobility(829 cm² V^-1 s^-1).The floating gate effect of In Se FGFET was verified by low frequency noise analysis.A nonvolatile memory with mechanical-writing ability,a retention time of 10³ s,repeated erase and write operations of more than 100 cycles,and multi-valued storage through mechanical modulation,was further implemented.Finally,tactile neural artificial synapses were constructed to realize multifunctional neuromorphic devices based on the mechanically modulated nonvolatile memory above.A variety of excitatory and inhibitory synaptic plasticity behaviors were successfully simulated by controlling the connection direction of the In Se FGFET and the distance between the two friction layers of TENG.Due to the high mobility of In Se and the self-powered effect of TENG operation,a power consumption of 165 a J was achieved in the In Se synapse in this work.Finally,utilizing the long-term excitatory and inhibitory plasticity behaviors of the In Se synaptic device,digital image recognition is further simulated,and the recognition accuracy is 92%.The electronics with mechanical-writing ability have great potential in simulating multifunctional low-power neural morphological electronic devices.