面向三维集成的TSV填充缺陷检测方法研究

项目名称： 面向三维集成的TSV填充缺陷检测方法研究

项目编号： No.51305179

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 机械、仪表工业

项目作者： 陆向宁

作者单位： 江苏师范大学

项目金额： 25万元

中文摘要： 随着半导体工艺和方法的改进和提高，TSV技术以其优良的性能而被广泛应用于3D封装。而TSV直径的不断减小，及深宽比（Aspect Ratio）的进一步增大，给TSV填充工艺及其缺陷检测带来了新的挑战，传统无损检测技术因分辨率较低而不能满足检测要求。本项目以小直径、高深宽比的TSV结构为研究对象，进行TSV填充缺陷的检测，提出了基于热图像超分辨率技术的主动红外探测方法，以解决微尺度下红外检测的空间分辨率问题。围绕多帧图像的超分辨率重建和基于学习的单帧超分辨率重建，以模糊辨识、增加图像动态范围、消除热噪声为核心进行深入研究，并对重构后的热图像进行特征优选和提取，应用神经网络、模糊专家系统对TSV结构缺陷进行识别，快速判断TSV填充状态及存在的缺陷，为填充工艺参数调整提供指导，并为3D TSV产品可靠性分析提供一种有效的手段和技术，也为我国IC制造提供基础理论和技术创新。

中文关键词： 贯穿硅通孔；缺陷检测；超分辨；主动红外热成像；扫描超声波显微镜

英文摘要： As semiconductor processes and methods being improved, TSV technology with the excellent performance has been widely used in 3D packaging, and becomes a trend in IC development. However,the TSV diameter is getting smaller gradually and the aspect ratio getting larger, which brings new challenges for TSV filling process and the defects inspection. The traditional nondestructive testing technology can not meet the testing requirements due to the lower resolution.This project is to study the technology of defects inspection for TSV structures with a small diameter and high aspect ratio. The active infrared testing method based on the thermal image super-resolution technology has been proposed to improve the spatial resolutions,which exist in micron scale defects inspection technology. The super-resolution reconstruction algorithms based on the multi-frame and single-frame image will be investigated. The techniques for fuzzy identification,increase of image dynamic range and the elimination of thermal noise would be used to reconstruct the thermal images with higher resolution,from which the optimal signals feature are extracted and applied into the neural networks and the fuzzy expert system to realize the identification of TSV defects,which provides the guidance for parameters adjustment in TSV filling process, a

英文关键词： TSV；Defect inspection；Super-resolution；Active thermography；SAM