原位光学发射光谱仪(OES)已经成为等离子体刻蚀工艺控制过程中的一种非常有潜力的在线传感器系统。采用光谱仪实时采集高密度等离子体刻蚀机中的OES光谱数据,利用BP神经网络算法对特定波长的OES数据进行分析及建模,以便对等离子体刻蚀工艺过程进行反馈控制。由刻蚀产物SiCl发出的405nm谱线被选为特征谱线来确定刻蚀工艺过程的终点。 In situ optical emission spectroscopy (OES) has become a very potential in-line sensor system for plasma etch process control. The spectrometer was used to collect the OES spectral data in the high-density plasma etching machine in real time. The BP neural network algorithm was used to analyze and model the OES data with specific wavelength in order to control the plasma etching process. The 405 nm line emitted by the etching product SiCl is selected as the characteristic line to determine the end point of the etching process.