玻璃覆晶的封装互连性能检测方法及倒装设备研究

发布时间：2018-07-17 15:31

【摘要】：近年来，液晶显示器（LCD）行业发展迅速，制造更大尺寸、轻薄化、低功耗、高分辨率的的LCD成为产业技术发展的方向。LCD模组普遍采用各向异性胶互连（ACF）的玻璃覆晶（COG）封装方式，其芯片凸点间距已达25μm，封装密度高，工艺参数复杂。在ACF-COG互连工艺设计与优化过程中，互连电阻因其不稳定性和易于退化的特点已成为评估产品互连性能的关键指标，容易引发严重的可靠性问题。互连电阻的数值与产品邦定工艺、设备特性、封装材料特性、COG产品设计等诸多参数相关，影响因素众多。如何确保在相邻凸点不短路的情况下，互连电阻小、均匀并且可控是COG封装领域研究的热点。目前评估互连电阻的主要方法是通过制造测试芯片，采用电阻监测及人工经验相结合的方式，费时费力，且缺乏有效的定量分析手段。因此，对于后续更高密度的玻璃覆晶模组互连性能评估而言，亟需一种互连电阻的自动估计及定量分析方法。本文在分析国内外COG封装工艺及互连电阻模型研究现状的基础上，针对ACF-COG互连电阻自动估计方法开展研究，重点解决基于颗粒导电面积的ACF互连电阻模型、以及凸点导电面积的自动获取方法，为更高密度的LCD模组封装自动化电阻监测及工艺优化决策提供支持。考虑到国内高密度COG倒装设备研发的迫切性，本文结合COG封装工艺特点，开展倒装设备的设计与研发工作，并在该设备上进行了互连电阻自动估计方法的测试验证。主要研究工作及取得的成果如下：一、提出了基于导电面积的互连电阻估计思路，并以此为出发点，提出了基于导电面积的ACF-COG封装导电颗粒电阻模型，该模型中包括颗粒与电极/凸点接触电阻、颗粒薄壁体电阻以及接触面上的隧道电阻。通过模型测试与数据比对发现，与其他已发表的电阻模型相比，该导电颗粒互连电阻模型预测结果与实测试据吻合度更高。基于该导电颗粒的电阻模型，提出了多导电颗粒情况下凸点与电极的互连电阻模型，该模型综合了因颗粒间的电场交互及边界电势效应引入的附加电阻，为后续获得导电颗粒的接触面积和位置后求解凸点互连电阻提供了途径。二、提出了基于机器视觉的COG凸点互连电阻估计方法，解决了凸点导电面积自动获取的难题。根据凸点边界直线概率较大的特点，本文提出了基于边界特征提取、能量梯度、凸点几何特征信息融合的凸点边界提取及区域截取方法，可以自动实现凸点角度纠偏及凸点区域提取，该方法对于图像中颗粒团聚或气泡等噪声的干扰具有强鲁棒性。通过图像形态学操作，可实现凸点导电颗粒的识别与导电面积的计算，并利用电阻模型，准确估计凸点互连电阻。经凸点图片测试分析发现，利用本文提出的凸点边界提取及区域截取方法，凸点检出率超过95%。在凸点导电颗粒计数方面，，通过人工比对，准确率达到94%，特别在因压力过大或者颗粒识别噪声较大的颗粒难识别的情况下，本方法有明显优势。在互连电阻估计方面，该方法可以通过精确的导电面积及颗粒位置计算获得较准确的凸点互连电阻。三、根据ACF-COG封装封装密度高、工艺流程复杂、对位精度高的特点，提出了一种高密度高精度倒装机新结构形式。该倒装机由IC上料转台、玻璃上料机械手、精密对位预压台以及双头本压台等关键部分组成。设计并优化了邦定压头结构，以控制压头的平面温差及平面度，通过结构热传递与热应变仿真分析并调整结构参数，实现了热压面温差与平面度分别为3.39C、2μm，达到设备热压要求。根据高密度高精度COG倒装机的方案设计，对预压对位、芯片拾取的视觉系统进行了设计与选型，设计了IC芯片上料机构、邦定压头、运动平台，并最终完成了设备机械结构设计。四、结合高密度高精度COG倒装设备的功能与结构设计，设计了设备的控制系统方案，分析并优化了系统控制时序，实现了双片生产时间12s（11.75s）生产工序规划与控制。针对设备的移动定位精度、视觉对位精度、邦定压头热压稳定性进行了测试，结果显示，设备的定位和热压精度达到了预期的设计目标。采用LCD玻璃基板和芯片进行邦定，采集测试模组的228个凸点图片进行了互连电阻估计及统计测试，实验测试结果表明，经方法估计的凸点互连电阻及邦定压力与实测值吻合，在准确性和凸点性能分析方面有优势。该方法可获取凸点导电面积、凸点下导电颗粒数目、颗粒导电面积等信息数据并进行统计分析，同时可快速定位缺陷凸点，为COG封装工艺分析与优化提供了便捷的手段。
[Abstract]:In recent years , liquid crystal display ( LCD ) industry has been developing rapidly , making LCD of larger size , light thinning , low power consumption and high resolution becoming the direction of industrial technology development .

On the basis of analyzing the current research situation of COG packaging technology and interconnection resistance model at home and abroad , this paper focuses on the research of ACF - COG interconnection resistance automatic estimation method , and focuses on the ACF interconnection resistance model based on the particle conductive area and the automatic acquisition method of the conductive area of the bump . In view of the urgency of research and development of the high - density COG flip - chip equipment , the design and research and development of the flip - chip equipment are carried out in consideration of the characteristics of the COG packaging technology . The test verification of the automatic estimation method of the interconnection resistance is carried out on the equipment . The main research work and the achievement obtained are as follows :

Based on the resistance model of the conductive particles , the interconnection resistance model of the bump and the electrode in the case of multi - conductive particles is proposed , which combines the electric field interaction with other published resistance models and the additional resistance introduced by the boundary potential effect , which provides a way for the subsequent obtaining of contact area and position of the conductive particles and solving the bump interconnection resistance .

This paper presents a method for estimation of bump boundary extraction based on machine vision , which solves the problem of auto - acquisition of convex - point conductive area . Based on the feature of large - line probability of convex point boundary , this paper presents the method of convex point boundary extraction and region interception based on boundary feature extraction , energy gradient and convex point geometry information fusion .

3 . According to the characteristics of high packing density , complicated process flow and high contraposition precision of ACF - COG package , a new structure of high - density high - precision pouring machine is proposed .

Based on the function and structural design of high - density high - precision COG flip - chip equipment , the control system scheme of equipment is designed , the system control timing is optimized , and the process planning and control of 12s ( 11.75s ) production process are optimized . The results show that the positioning and hot - pressing accuracy of the equipment reach the expected design goal . The results show that the method can obtain the information data such as the conductive area of the bump , the number of conductive particles under the convex point , the conductive area of the particles , and make statistical analysis . At the same time , the defect bump can be quickly located , which provides a convenient means for the analysis and optimization of COG packaging process .
【学位授予单位】：上海交通大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN873.93

【参考文献】