Research On Memristors Based On Bi:SnO₂ And Two-dimensional Cu₃P As Functional Layers

With the rapid development of information age,the amount of information such as data,image and video is increasing.It is urgent to develop memory with low access power consumption,high storage capacity and multi-functional characteristics.At present,the mainstream storage technology is mainly flash memory.However,due to its slow erasing speed,slow writing speed and reaching the limit of miniaturization,people need to develop new storage materials and devices.Memristor(also known as resistive memory)is expected to become the next generation of nonvolatile memory due to its fast erasing write speed,low power consumption,high storage density and can be used for 3D integration technology.Moreover,since its transmission characteristics are similar to those of biological synapses,the research on neural morphology calculation based on memristor has also become a hotspot in the field of information.In this paper,the effect of Bi doping on the resistive switching characteristics of SnO2memristor was first studied.ITO/Bi:SnO₂/Ti N memristor devices were prepared by magnetron co-sputtering method and compared with undoped ITO/SnO₂/Ti N memory devices.Studies have shown that the memristor exhibits typical bipolar resistive switching characteristics,and the consistency of device performance is successfully improved and the power consumption is reduced by doping Bi element(by reducing the operating voltage and operating current).Its performance changes with the change of bismuth doping concentration,and the results show that the device with bismuth doping concentration of 4.8%has the best performance,and the power consumption is reduced by the largest extent compared with the pure tin oxide device.In this paper,the resistive switching mechanism of Bi:SnO₂memristor is further studied.The current fitting results show that ITO/Bi:SnO₂/Ti N devices before and after doping are conductive filament mechanism.Furthermore,through TEM and other characterization methods,it was observed that tin oxide crystals were formed in the middle of the resistive layer,and bismuth atoms were distributed around the tin oxide crystals.Thus,we propose a novel coaxial bismuth atom conductive filament,explain the main reason for the device performance improvement,and expand the understanding of the resistive switching mechanism of memristor.In this paper,the resistive switching mechanism of Bi:SnO₂memristor is further studied.The current fitting results show that ITO/Bi:SnO₂/Ti N devices before and after doping are conductive filament mechanism.Furthermore,through TEM and other characterization methods,it was observed that tin oxide crystals were formed in the middle of the resistive layer,and bismuth atoms were distributed around the tin oxide crystals.Thus,we proposed a novel coaxial bismuth atom conductive filament,which successfully explained the reasons for the low power consumption and performance optimization of the memristor,and expanded the understanding of the resistive switching mechanism of the memristor.The paper also focuses on the study of the photoelectric flexible memristor based on two-dimensional Cu₃P materials.The memristor has good cycle stability,large storage window and photoelectric tuning multi-stage storage characteristics.At the same time,the memristor has a bending radius of 10 mm and a bending number of 10³times,and the resistance performance does not deteriorate.The brain-like learning behavior of Cu₃P memristor was further studied.The results show that Cu₃P memristor can realize various synaptic bionic functions under photoelectric tuning,and has the human brain learning and memory function of'learning-forgetting-backtracking'.In summary,this paper obtains a low-power memristor through material optimization and structural design,and proposes a new coaxial conductive filament model and a high-performance cognitive memristor,which provides an effective method for the performance optimization,power consumption reduction and mechanical flexibility improvement of memristors.This study laid a hardware foundation for the non-volatile memory and neural morphology application of memristors.