Defect size measurement using deep learning methods
By combining images from multiple detectors to form a pseudo-color RGB image and using a convolutional neural network, the problem of difficult defect size measurement in semiconductor manufacturing is solved, achieving fast and accurate defect size measurement and classification.
CN115516295BActive Publication Date: 2026-06-16KLA CORP
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- KLA CORP
- Filing Date
- 2021-05-17
- Publication Date
- 2026-06-16
AI Technical Summary
Technical Problem
Existing technologies struggle to accurately detect and measure defect sizes smaller than the wavelength of light in semiconductor manufacturing, and traditional methods rely on expensive and time-consuming scanning electron microscope tools.
Method used
The system, consisting of a light source, a detector, and a convolutional neural network unit, combines images from multiple detectors to form a pseudo-color RGB image, and then uses the convolutional neural network to directly measure and classify defect sizes.
Benefits of technology
It enables rapid and accurate measurement of defect dimensions, reduces reliance on expensive SEM tools, and improves detection efficiency and accuracy.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN115516295B_ABST
Abstract
A system has detectors configured to receive light beams reflected from a wafer. For example, three detectors can be used. Each of the detectors is a different channel. Images from the detectors are combined into a pseudo-color RGB image. A convolutional neural network unit (CNN) can receive the pseudo-color RGB image and determine a size of a defect in the pseudo-color RGB image. The CNN can also classify the defect into a size category.
Need to check novelty before this filing date? Find Prior Art