基于建模与优化的MOCVD温度均匀控制

Temperature Uniformity Control in MOCVD based on Modeling and Optimization

为了提高MOCVD反应室衬底温度均匀性,提出一种基于建模和优化的控制策略。采用神经网络建立反应室温度特性模型,得到输入功率与反应室衬底温度场之间的关系,在此基础上,以温度均匀性为性能指标,采用仿生优化算法对输入功率配比进行优化,实现MOCVD反应室衬底温度的均匀控制。仿真结果验证了所提出方法的有效性。

A modeling and optimization based control scheme is proposed to improve the wafer temperature uniformity in MOCVD reactor chamber.A neural network based temperature model is established by modeling the relation between the heating power and the corresponding wafer temperature.Taking the temperature uniformity as the performance index,the heating power ratio is optimized by employing the bionic optimization algorithm and the obtained temperature model.The temperature uniformity control is thus performed in real time by constantly modifying the heating power ratio.Experiments results are presented to show the effectiveness of the proposed scheme.