Experimental Study On Damage Mechanism And Prediction Model Of Silicon Micro/Nano Structure Fabricated By Focused Helium Ion Beam

Focused helium ion beam is an emerging nano-processing technology in the field of ion beams.With its extremely high processing accuracy and imaging capabilities,it is widely used in materials modification,semiconductor processing and other industries.During the ions incidence process,light ions such as helium(He)ions incident in the substrate cause a smaller sputtering yield,a greater implantation depth,and special processing effects.For threedimensional substrate materials,the use of focused helium ion beam processing will cause a wide range of materials inside the substrate to become amorphous and produce a large number of helium bubbles.This paper takes the damage effects of single crystal silicon(Si)substrate with a focused helium ion beam as the research object,analyzes its damage formation mechanism through the movement of ions on the substrate,carries out experiments to quantitatively analyze the damage effects,and establishes a damage model for damage prediction of the range of effects,and analysis and discussion of experimental and simulation results.The main research contents are as follows:(1)Select experiment materials and equipment,formulate experiment procedures and experiment plans,and conduct focused He ion beam processing experiments on single crystal Si.It was found that a unique shape of amorphous and completely amorphous processing damage appeared in the processed Si substrate,and helium bubbles were formed in the completely amorphous damage area.The various damage phenomena appearing in the experiment were calibrated,and then the evolution process of amorphization damage with process parameters was quantitatively analyzed.In addition,the number of helium bubbles in the damage area of the helium bubble is calibrated,and the evolution of the number of helium bubbles changing with its diameter range is given.(2)Based on the physical process of the interaction between incident ions and substrate atoms,a theoretical model of interaction between particles is used to describe the movement of He ions and explain the law of He ions movement in the substrate.Based on the principle of energy deposition of He ions in the substrate,the energy loss and distribution of ions are simulated by software to explain the cause of damage.(3)Based on the experimental results and rules,separately consider the effects of nuclear stopping and electronic stopping on ions movement,establish the function form of the processing damage range,and calculate the function coefficients through genetic algorithms.The function model was used to simulate the most important profile among the damage profiles,namely the outer profile of the completely amorphous damage,and then the damage morphology prediction accuracy and the simulation accuracy of the critical dimensions were analyzed and discussed.The results proved the function has good applicability for the prediction of He ion beam processing damage and its evolution.(4)In order to simulate multiple damage effects on the substrate more quickly and accurately,a more convenient,faster and more predictive model was established.The model can simultaneously simulate the amorphous,completely amorphous and helium bubble damage area,and complete the prediction of the helium bubble distribution.Firstly,the damage image is preprocessed to obtain the gray image of the processing damage,and the distribution function of the helium bubble distribution is selected and fitted.Integrate pixel information and equation coefficients to build a deep neural network model.The neural network model can predict the range profile of multiple damage effects at the same time,and can simulate the change of the number of helium bubbles with its diameter range.After comparative analysis of experimental and prediction results,the model shows extremely high prediction ability and simulation accuracy.This article explains the damage formation mechanism of the Si substrate processed by the focused He ion beam,and quantitatively analyzes the law of various processing damages with process parameters.The established damage function model and the prediction model based on the deep neural network show extremely high simulation accuracy.The methods and research results used in this study have greatly promoted the development of experimental and simulation research in the field of focused helium ion beam processing.