应用于微结构制造自治系统的第I类工艺缺陷诊断

在集成电路制造中,通常采用各电学参数的测试值来衡量和评价工艺效果。一种可能的工艺后果是,出现了测量结果偏离预定目标的第I类工艺缺陷。传统上的基于SPC的工艺控制技术,不含工艺诊断的内容,在寻找造成工艺失效的原因时,存在着巨大的困难。本文对集成电路制造过程中所涉及到的几类重要参数间的关系,进行了详细的分析,由此确定造成诊断问题困难的根本原因,在于测试现象与工艺故障因素之间是多对多的关系。通过正交化的方法,可以将测试参数转化成为一组广义参数。广义参数与故障因素之间存在着一对一的联系。这样就找到了一种有效的,从异常的广义参数值确定失效原因的工艺诊断方法。本文对第I类工艺诊断的技术进行了讨论。在集成电路制造自自治系统中,对于I类工艺缺陷的纠正过程,也就是工艺结果不断向工艺目标趋近的过程。

The Method Diagnosing Class I Process Failures in an Electronic Manufacturing Automation System

In the current semiconductor manufacturing industry,tested values on semiconductor parameters are usually used for evaluating process qualities.The values may differ from preset targets.Such a process failure is defined as a class I failure.It is very difficult for the traditional process control method,SPC,to find the reason for a failure,because the SPC method involves no process diagnose algorithms.After studying in detail the relationships among all the important sets of parameters that relate to the fabricating process,the reason,who makes process diagnose so difficult,is found to be that one set of parameters corresponds to another set in a multi to multi way.By orthogonalizing the tested parameters into a set of generalized parameters,one to one relationship between the generalized parameters and the process failure factors is constructed.Therefore an abnormal generalized parameter value can direct to a certain process failure factor.The method diagnosing class I process failure is discussed.In an electronic manufacturing automation system,correcting class I failures means leading the process result to target.