Method for detecting wafer edge edge-wash boundary

Abstract

The invention relates to a method for detecting wafer edge edge-wash boundary, relating to a semiconductor manufacturing process. The method comprises the steps of: providing a wafer scanning machinetable comprising a detector; locating a lens of the detector to allow the lens of the detector to scan an image of a special thickness [Delta]d in the thickness direction of the wafer, rotating the wafer to allow the detector lens to scan a circle of the wafer by taking the special thickness [Delta]d as a scanning range, and performing approximate rectangular processing for the scanned images to obtain a rectangular scanning image with a width of [Delta]d' and a length of the wafer perimeter C; dividing the rectangular scanning image into M units, dividing each unit into N pixels, performing convolution processing of the N pixels to obtain an edge-wash boundary feature database, wherein M>1, and N>1; setting a wafer edge-wash boundary threshold value, performing difference calculation of two adjacent rows of data of the edge-wash boundary feature database, wherein when the difference value of the two adjacent rows is larger than the threshold value, the boundary is considered as an edge-wash boundary and the edge-wash boundary is reported to accurately detect the wafer edge edge-wash boundary.

Description

Technical field [0001] The present invention relates to a semiconductor manufacturing process, and in particular, to a method for detecting the edge-wash boundary of a crystal. Background technique [0002] In the semiconductor manufacturing process, as the size of semiconductor manufacturing technology continues to shrink, in the wafer manufacturing process, the special structure of the crystal edge often leads to defects such as peeling of the crystal edge film. Currently, the monitoring of crystal edge defects is mainly based on optical For on-chip offline monitoring, for patterned wafers, especially in the middle and later stages of the process, there is a lot of noise when detecting crystal edge defects due to the difference in photolithography edge cleaning distance and the increase in film growth levels. [0003] For the detection of the edge wash boundary of the crystal, the current method is to collect the original image of the crystal edge, confirm the edge wash bo...

AI Technical Summary

Problems solved by technology

However, this method can effectively detect known wash boundaries with obvious features, such as figure 1 As shown, it is a schematic diagram of the result of detecting the edge washing edge of the prior art, such as figure 1 As shown, only part of the edge wash boundary 100 is detected, but the unknown defects cannot be detected

Images