Study On Oxygen Ion Behavior Control And Performance Optimization Of ZnTPP-Based Memristors

With the high-level information storage and processing capabilities of synaptic activity control,the human brain can perform a variety of complex learning,memory,cognition and recognition functions whose efficiency is incomparable with traditional information storage and processing equipment.At present,the artificial intelligence system based on neuromorphic computing has become one of the most promising technologies for the efficient storage and big data processing in the post-Moore era by breaking through the "Von Neumann bottlenecks".To date,many artificial architectures such as traditional complementary metal oxide semiconductor(CMOS)devices,nano cells,quantum dots and molecular devices have been used to simulate human brain function.However,neural networks based on such devices are highly dependent on software programming without inherent hardware learning capabilities,therefore these devices can't perform complex neural functions efficiently.In recent years,the emergence and development of memristors provide a simple and effective way to construct artificial neural networks(ANNs)that don't rely on software programming.In this paper,an ion-/electron-transporting memristor with coordination-assisted ion assisted ion traps is proposed in this dissertation.The configuration of organic memristor structure is ITO/tetraphenylporphyrin zinc(ZnTPP)/Al2O3-x/Al.The optimization of the device structure,the solution to destructive bubble problem,and the application of neural function simulation are studied which provide a technical alternative for the theoretical development and practical application of the organic memristor.(1)Based on the ZnTPP memristor,adjusting the Al2O3-x layer thickness improves the device performance.With the characterizations of the basic electrical properties of the device,the device performance from the normal diode to the memristor and eventually changing to the insulator was studied.Based on this,the optimized devices were simulated for functional neuromorphology.Therefore,this chapter not only validates the role of Al2O3-x layer in the device structure,but also studies the effect of thickness on device performance for further optimizing the device structure thickness.(2)Under the stimulation of voltage pulses,a large amount of bubbles are generated in the ZnTPP memristor and the bubbles may even break after suffering the excessive stimuli.These destructive bubbles severely affect the memristors' simulation of synaptic function.Because the bubbles are generated between the anode of the device(i.e.the ITO electrode and the active layer),we add a modified layer between the two layers to prevent the appearance of such destructive bubbles.(3)Based on the oxygen ion transport model with coordination-assisted mechanism of ZnTPP diode memristor and the dynamic analysis of resistance change behavior in previous researches,this article builds a diode based on different organic substances(such as P5,C60,PS and PS&C60).The paradigm verifies that this diode-based device structure based on oxygen ion transport could be expanding.