混合进化算法的Metropolis蒙特卡罗MEMS单晶硅湿法刻蚀工艺模型.PDF

49 5 Vol.49 No.5 2013 3 JOURNAL OF MECHANICAL ENGINEERING Mar. 2013 DOI 10.3901/JME.2013.05.167 Metropolis MEMS * ( 211189) Metropolis (h,k,l) Si(100) Si(110) Metropolis Metropolis O414 Hybrid Evolutionary–Metropolis Monte Carlo Algorithm for Wet Etching of Silicon MEMS Fabrication Process WANG Shuqiao XING Yan TIAN Mi QIU Xiaoli QI Jianchang (School of Mechanical Engineering, Southeast University, Nanjing 211189) Abstract A novel hybrid evolutionary–Metropolis Monte Carlo method and a new form of removal probability function are proposed for the simulation of anisotropic wet etching process of silicon. Based on the etch rates of typical orientations, the new process model has the ability to calculate the energy parameters of the atomistic Monte Carlo method automatically. The modified removal probability function uses a four indices schema to classify the configuration of surface atoms, the process of changes in the structure of the surface of the etching reaction can be described more accurately, it improved the simulation ability of Monte Carlo model. The calculated etch rates of (h,k,l) orientations at different concentration and temperature in different etching conditions show good agreement with the experiment data. Comparing with the existing model, the new removal probability function can represent the anisotropy curve of etch rate with local minima at Si(100) and Si(110) vicinal planes with high accuracy. Finally, a simulation of micro structure fabrication processes is performed to