Synthesis Of Wafer-Scale Graphdiyne/Graphene Heterostructure For Neuromorphic Applications

Two-dimensional materials,due to their unique atomic structure and excellent electronic and optoelectronic properties,provide an ideal platform for building heterostructures with various characteristics.As a new two-dimensional material,graphdiyne has a highly conjugated ? structure;Its unique network structure and excellent chemical,photoelectric and electrochemical properties,make it a promising catalytic and energy storage material.However,the application of graphdiyne in the field of optoelectronic devices and bionics has not yet been explored.In addition,due to the excellent pores and trapped energy levels in graphdiyne,it can provide an effective modulation method for the field of optoelectronic storage,laying the foundation for the realization of an efficient and stable artificial neuromorphic system.Synapses,as an important component of human neuromorphic networks,have functions such as information perception and transmission.And it usually used for brain heuristic calculations.Artificial synapses are the basis for imitating important functions of biological nerve synapses(paired impulse facilitation,short-term enhancement,long-term enhancement,spike time-dependent plasticity,etc.),and play an important part of artificial neuromorphic networks.Affected by the era of big data,the neuromorphic computing system based on the von Neumann architecture,due to its inherent shortcomings—the separation of processing and memory units,has led to a great energy consumption and time delay in data transmission.In order to build an artificial vision system with efficient visual information integration and processing capabilities like the human visual system,there is an urgent need for large-scale integrated multifunctional photoelectric synapses with a wide operating spectrum.In this thesis,a multifunctional photoelectric synapse array is designed based on a wide range of graphdiyne heterojunction films on wafer-level graphene.Among them,graphdiyne acts as a charge trapping layer to store light-excited charges;graphene acts as a channel,which is affected by and feeds back the trapped charges from graphene.The array has very small device-to-device variation(DDV),which can be simulated synaptic behaviors including excitatory/inhibitory postsynaptic current(EPSC/IPSC),paired pulse facilitation(PPF),pulse frequency dependent plasticity(SRDP)and long-term memory(LTM).At the same time,the linear conductance update behavior of the synaptic transistor is also verified,which is helpful to construct a high-precision and strong fault-tolerant artificial neural network for neuromorphic calculation.In addition,due to graphdiyne excellent charge-trapping ability and significant light absorption range(300–1000 nm),the device has associative learning,all-optical logic operations("NAND" and "NOR"),and ultraviolet-visible light range Integrated functions such as image sensing,memory and processing inside.It lays a foundation for the development of fast,efficient and low-energy artificial sensor-memory-processing systems applied to artificial intelligence and the Internet of Things in the future.