A method for detecting whether a wafer is abnormal

Abstract

The invention discloses a method of detecting normality of a wafer. The method of detecting normality of the wafer includes the following steps. Mark positions which correspond to each other are respectively arranged on all to-be-detected wafers. Data of the corresponding mark positions are theoretically maximum values of detected data of the wafers. Multiple positions which correspond one by one can be additionally arranged on the to-be-detected wafers, apart from the mark positions. Data of the set mark positions and other multiple positions of all the wafers and collected and recorded. Data collected from all the wafers are averaged and the average of all the wafers is obtained. The difference between the datum of the mark position of each wafer and the average of the wafer is obtained and judged to be within a preset range or not. If not, the wafer is normal and if yes, the wafer is not normal. The method of detecting normality of the wafer can effectively detect abnormal wafers so as to avoid large loss.

Description

technical field [0001] The invention relates to the technical field of semiconductor device manufacturing technology, and more specifically, relates to a method for detecting whether a wafer is abnormal. Background technique [0002] During the manufacturing process of semiconductor devices, process steps such as coating and etching are often involved. After the completion of each process step, there is a detection process to detect whether the wafers after each process step meet the requirements. For the wafers that meet the requirements, the next process step can be carried out. For the wafers that do not meet the process requirements It may be necessary to process accordingly and restart the process step or scrap the wafer, etc. [0003] In the prior art, after each process step is completed, the wafer production process is often monitored by a statistical process control (Statistical Process Control, SPC) method. In the detection process, it is first necessary to colle...

AI Technical Summary

Problems solved by technology

Depend on figure 2 and image 3 We can see, yes figure 1 After the original data in the process is processed by the mean value method or the range method, the abnormality of the data on the sixth wafer will not be detected, which will cause the abnormal wafer to enter the next process step, thereby causing greater losses

Images