Organic Thin-Film Memristors And Memtransistors

In recent years,the fields of artificial intelligence and machine learning have witnessed rapid development.However,most of the achievements are realized based on software algorithms,which rely on the traditional von Neumann system.In the current computer architecture,computation and storage need to be separated into two independent units,and data need to be continuously transferred between the two units when processing instructions,which restricts the running speed of the computer and generates large energy consumption.In contrast,the human brain can carry out computation and storage simultaneously.Because there are trillions of neurons in the brain and their connections are synapses.The weight of each synapse is a basic unit of memory,and these memory units are scattered in the whole brain,so the brain can efficiently process tasks in parallel.Inspired by human brain,new neuromorphic networks are developing rapidly,in which synaptic devices are the most basic units.In order to simulate the plasticity of synapses,synaptic devices need to be capable of analog resistive switching and non-volatile storage.Therefore,memristor and memtransistor with these two capabilities are the best choice to simulate the synaptic behaviors.Currently,most memristors and memtransistors are based on inorganic materials.Comparing to inorganic materials,organic materials have many advantages such as biological compatibility,mechanical flexibility and ductility,low cost and low temperature manufacturing process,and modifiability of molecular groups.Therefore,the main goal of this thesis is to fabricate synaptic devices based on organic materials.The details are as follows:Model and device construction of organic thin film memristor.A novel memristor model based on charge trapping and detrapping in electrets is proposed:under external electric field,charge exchange will occur at the interface between semiconductor channel and electret,and the residual charge in electret will change the total resistance of semiconductor channel.According to the model,an organic thin film memristor with asymmetric top electrode and floating gate electrode is proposed,which is based on the equivalent topological transformation of a standard organic thin film transistor storage device.The realized memristor has the properties of analog resistive switching and non-volatile storage,which can simulate the synaptic plasticity very well.Application of organic thin film memristor in synaptic simulation.Based on the constructed organic thin film memristor,a variety of important synaptic behaviors have been emulated including:repetitive synaptic potentiation and depression behaviors are simulated through analog resistive switching properties;The paired-pulse facilitation behavior,which represents the short-term plasticity,is simulated by the bias stress effect inherent in organic devices;The Hebbrian learning rule,which represents the long-term plasticity,is simulated by the non-volatile storage properties.Model and device construction of organic thin film memtransistor.A novel memtransistor model based on UV controlled charge trapping and detrapping in electrets is proposed:the charge trapping/detrapping rate between semiconductor channel and electret interface depends on both vertical field intensity and UV intensity which can control the speed of resistive switching.According to this model,an organic thin film memtransistor with UV gate is proposed.Because of the UV-controllable properties,organic thin film memtransistors can simulate more advanced synaptic plasticity.Application of organic thin film memtransistors in simulating the plasticity of synapses at different growth stages.Based on the constructed organic thin film memtransistor,the behaviors of synapses at different stages have been emulated including:The synaptic potentiation and depression behavior of synapses at different stages through UV controlled analog resistive switching properties;the paired-pulse facilitation behavior of synapses at different stages was simulated by UV induced current decay;the long-term memory behavior of synapses is simulated by the non-volatile storage properties in the dark environment.In this thesis,organic thin film memristor and memtransistor based on charge trapping and detrapping in polymer electret are proposed,and a variety of important synaptic behaviors were successfully simulated with the prepared devices,including:synaptic potentiation and depression,paired-pulse facilitation,spike-time-dependent plasticity,and the plasticity in different synaptic growth stages.