Font Size: a A A

Research On Artificial Synaptic Devices Based On Memristors And Waveguide Structures

Posted on:2024-01-10Degree:MasterType:Thesis
Country:ChinaCandidate:J H YuFull Text:PDF
GTID:2568307136988889Subject:Microelectronics and Solid State Electronics
Abstract/Summary:PDF Full Text Request
With the advent of the big data era,traditional computer chips designed with CMOS logic gates and von Neumann architecture are struggling to handle the vast amount of information.In contrast,research on brain-inspired computing,which leverages the low-power,learning,memory,and parallel computing capabilities of neurons and synapses has become a hot topic.Scientists are delving into the development of hardware devices that mimic neurons and synapses to advance the realization of brain-inspired computing.Currently,various types of discrete synaptic devices have been developed,including transistor-based synaptic devices,memristor-based synaptic devices and waveguide-based synaptic devices.Of particular interest is the memristor-based synaptic device that exhibits photonic and electronic effects,showing tremendous potential in constructing artificial synapses that can directly capture external light signals.In this study,a novel two-terminal optoelectronic memristor based on the CuAlAgCr/Ti O2/W(CTW)structure.The synaptic properties under electrical and light stimulation were studied.Furthermore,designed and simulated a waveguide-based all-optical synaptic directional coupler,and the following main research work is carried out:(1)An investigation was conducted on a new alloy material,CuAlAgCr,and optoelectrically controlled memristor devices based on CuAlAgCr were fabricated using magnetron sputtering.The fabricated devices were subjected to annealing treatment,and their characteristics were analyzed through X-Ray Diffraction,Scanning Electron Microscopy,and Ultraviolet Visible Near Infrared Spectrophotometer.The impact of the annealing process on the device performance was examined.(2)Direct current electrical tests were conducted on the CTW optoelectronic memristor to explore their resistive switching characteristics.By adjusting the current limiting,the transition between volatile and non-volatile states was achieved.Furthermore,the CTW optoelectrically controlled memristor devices exhibited excellent retention time characteristics,low switching voltage,and good cycling durability.The resistance switching mechanism of the devices was investigated through theoretical calculations and experimental research.(3)Detailed testing was conducted on the optoelectronic characteristics of the annealed CTW optoelectrically controlled memristor devices.These devices exhibited responsiveness to visible light,with response currents reaching the microampere level.Additionally,they demonstrated a variety of artificial synaptic functionalities,including short-term memory,long-term memory,the transition from short-term to long-term memory and learning,forgetting,and relearning behaviors.Furthermore,the annealed CTW optoelectrically controlled memristor devices were explored for image perception,logic operations,and optoelectronic information demodulation.Finally,the conduction mechanism of the annealed CTW optoelectrically controlled memristor devices was investigated through theoretical calculations.(4)A study was conducted on a new phase-change material Ge2Sb2Se4Te1 and a waveguide-based all-optical synaptic directional coupler was designed based on this material.The all-optical synaptic directional coupler was simulated using COMSOL Multiphysics.By applying a 1200 nm wavelength laser to heat the Ge2Sb2Se4Te1 material,it switches between crystalline and amorphous states.The long-term potentiation and long-term depression functionalities of the optical synapse were achieved by varying the quantity of crystalline Ge2Sb2Se4Te1.Notably,the all-optical synaptic directional coupler exhibited excellent characteristics,including low power consumption and fast switching speed.
Keywords/Search Tags:Optoelectronic memristor, Waveguide, Artificial synapse, First Principles Computation
PDF Full Text Request
Related items