论文部分内容阅读
由于ＩＣ制造过程的高度复杂非线性，利用传统的方法进行工艺分析难以获得满意的结果，现描述一种新的途径——神经网络反向传播算法（ＢＰ算法），并介绍了神经网络ＢＰ算法在ＰＥＣＶＤＳｉ３Ｎ４钝化工艺中的应用。通过建立ＰＥＣＶＤ淀积工艺的神经网络模型，对ＰＥＣＶＤＳｉ３Ｎ４钝化工艺进行了计算机模拟，讨论了相关工艺条件对Ｓｉ３Ｎ４介质膜特性的影响。所编写的神经网络应用程序已用Ｃ语言在计算机上得到了实现
Due to the highly complex non-linearity of the IC manufacturing process, it is difficult to obtain satisfactory results by the traditional method for process analysis. Now, a new way is described, namely BP neural network back propagation algorithm (BP algorithm), and the neural network BP algorithm Application in PECVDSi3N4 passivation process. Through the establishment of neural network model of PECVD deposition process, the PECVDSi3N4 passivation process was simulated by computer, and the influence of relevant process conditions on the properties of Si3N4 dielectric film was discussed. The prepared neural network application has been implemented on a computer using C language