Research On Synaptic Devices Based On Top Gate Transistors

Intelligent power equipment can achieve remote monitoring,intelligent control and fault detection,and its development will largely depend on efficient data collection,storage and analysis.In the face of such a large volume of data,computers based on the Von Neumann architecture are becoming increasingly difficult to meet the requirements.In contrast,the human brain has the capacity for efficient parallel computing,enabling the integration of storage and computation with low energy consumption,so efforts have been made to develop artificial synaptic devices with brain-like functions to process complex information.Top-gate transistor-based synaptic devices have been extensively investigated for their ability to modulate synaptic weights through the gate while enabling learning and memory functions.The top-gate structure,with the insulating layer on top of the semiconductor layer,has a passivating effect and improves the air stability of the device;the low contact resistance between the semiconductor layer and the source-drain electrode improves device performance.At the same time,the new generation of electronic devices also puts forward new demands on the devices such as transparency and flexibility,therefore,this paper prepares flexible transparent synaptic devices based on the top-gate transistor structure and conducts simulations on the synaptic aspects of the behaviour.The main research contents are as follows:1.Organic transistor based on top gate structure is studied.The transistor is fabricated with ordinary silicon wafer as the substrate,the mixture of IDT-BT and PCBM as the semiconductor layer of the device,PMMA as the insulating layer of the device.The electrical characteristics of the device are adjusted by changing the mixing ratio of IDT-BT and PCBM,and the performance of the top gate transistor is tested.The results show that when the mixing ratio is 2%,it shows good electrical characteristics.The switching current ratio is greater than 10~5,and the threshold voltage is about-5V.Based on the energy band structure,the working mechanism of the device is clarified.The N-type semiconductor is the acceptor and the P-type semiconductor is the donor in the semiconductor layer,and I-V double-scan test proves that it has the potential to become an artificial synapse.2.Flexible transparent synaptic transistor based on top gate is researched.The synaptic transistor takes patterned ITO-PET as the substrate,IDTBT and PCBM are blended as the semiconductor layer,PMMA is the insulating layer,and the sprayed SWCNT solution of single-walled carbon nanotubes is used as the device gate,so as to prepare the top-gate flexible transparent transistor.After testing,it is found that its performance is equivalent to that of transistors made of ordinary silicon wafers.At the same time,the transparency of the device is close to 80%,and it still has performance in bending state,and the basic synaptic behavior is successfully simulated.3.An applied study based on top-gate synaptic transistors.A pulse sequence is applied to the device as different stimuli to simulate the behaviour of a biological classical conditioned reflex,the Pavlovian dog,followed by an artificial neural network based pattern recognition algorithm for image pattern recognition,which is able to reach 81.35%recognition accuracy after multiple cycles of learning.