Due to the excellent mechanical,electrical and optical properties,twodimensional(2D)materials have become promising materials for the development of multi-functional and wearable nano-devices.In order to improve the properties of 2D material and the performance of corresponding nano-devices and computing system,the modification methods of 2D material have aroused intensive attention in the field of electronics.Compared to other modification methods,such as chemical doping,strain engineering and plasma treatment,optical irradiation can not only thin and pattern 2D materials,but also induce the selective modification of materials,including oxidation,phase transition,deformation and formation of defects and nanostructure.This technology is a critical method to fabricate novel 2D materials and nano-devices.Moreover,the memristors based on 2D materials have demonstrated excellent performance and stability,Hence,optical-induced modification of 2D materials is also a promising way to daevelop various memristor and corresponding efficient computing and storage system.In this paper,to develop fantastic memristor and corresponding computing system,the theoretical and experimental investigations about the optical-induced modification of 2D indium selenide(InSe)and tellurium(Te)are carried out.Volatile and non-volatile memristors with high performance,low power consumption and high stability are fabricated.Besides,the applications of these memristors in computing system and artificial neural network are also explored.Based on the density functional theory,the atom structures are calculated to investigate the exfoliation of bulk crystal and the formation of 2D InSe and Te.Then,with the "Energy-Strain" method,acoustic phonon limited method and other calculated methods,the intrinsic properties of 2D InSe and Te are calculated to predict the enormous potential of these materials in the development of nano-devices with high performance.Besides,the calculation results are also theoretical foundations for modifying properties of 2D InSe and Te by optical irradiation.Furthermore,the mechanisms of tuning mechanical and electrical properties of 2D InSe and Te by optical-induced modification,including changes of thickness and size,deformation and formation of defects,are provided by calculating the atom structures,Young’s modulus,Poisson’s ratio,formation energy of defects,band structures,diffusion barrier of metal atoms and so on.These investigations provide theoretical guidance for the experimental investigation of optical modification 2D InSe and Te.In the experimental investigations,high quality 2D InSe and Te flakes are firstly fabricated to explore the modification of materials induced by laser irradiation and ultraviolet-ozone(UVO)treatment.Then InSe and Te flakes are exposed to laser with different power and irradiation time in air condition to investigate the evolution of the surface morphology,atom structure and the element composition.Using suitable laser parameters,defects and oxides are selectively induced in InSe and Te flakes for developing memristors.In addition,the modification of 2D InSe induced by UVO treatment is completed.Moreover,the evolutions of atom structure and element composition in InSe using UVO treatment with different temperature and time are also discussed.These experimental investigations provide methods and mechanisms for development high-performance memristor using optical-induced modification.With the modified 2D InSe and Te,vertical nano-devices with two terminals are fabricated to investigate the variations of electrical performance.The laser-irradiated2 D InSe and Te exhibit performance of digital-type volatile memristor,while pristine materials have no switching properties.The nano-devices based on UVO modified InSe demonstrate nonvolatile memristive with high performance and stability.These results have proved the application of optical-induced modification in development of memristor.Besides,the mechanism of nonvolatile InSe memristor are discussed.Finally,it is approved that optical-induced modification of 2D materials also has promising application on strain engineering and flexible devices in combination with the nano-manipulation technology.The nonvolatile memristor based on UVO modified InSe demonstrates tunable resistance state and outstanding stability.Hence,the applications of memristors in convolutional computing system and artificial neural network(ANN)are discussed.Convolutional computing is performed based on memristor crossbar array with tunable resistance state.With low calculation error,the memristor crossbar array can be used to process the digital images,including smoothing and edge detection.Furthermore,nonvolatile InSe memristors have stable and fast switching under the electrical pulses,so that memristors can also simulate the behavior of synapse.Hence,the accuracy of handwritten image recognition using ANN based on nonvolatile modified InSe memristor is calculated to reveal the application in the field of inmemory computing. |