北京航空航天大学学报 ›› 2019, Vol. 45 ›› Issue (2): 381-387.doi: 10.13700/j.bh.1001-5965.2018.0315

• 论文 • 上一篇    下一篇

芯片互联结构断裂失效的试验研究与统计分析

陈垚君, 景博, 胡家兴, 盛增津, 张钰林   

  1. 空军工程大学 航空工程学院, 西安 710038
  • 收稿日期:2018-06-01 出版日期:2019-02-20 发布日期:2019-03-04
  • 通讯作者: 景博 E-mail:jingbo_sensor@163.com
  • 作者简介:陈垚君男,硕士研究生。主要研究方向:电子封装可靠性;景博女,博士,教授,博士生导师。主要研究方向:故障诊断与健康管理、测试性设计、传感器网络与信息融合。
  • 基金资助:
    “十三五”装备预研共用技术(41402010102)

Experimental research and statistical analysis of fracture failure for interconnected structures in electronic chips

CHEN Yaojun, JING Bo, HU Jiaxing, SHENG Zengjin, ZHANG Yulin   

  1. School of Aeronautic Engineering, Airforce Engineering University, Xi'an 710038, China
  • Received:2018-06-01 Online:2019-02-20 Published:2019-03-04
  • Supported by:
    National Defence Pre-research Foundation of China (41402010102)

摘要: 针对电子芯片互联结构失效机理复杂、寿命数据难以获取且缺乏具有理论依据支撑的寿命分布模型等问题,开展了基于退化试验数据的统计分析研究,并提出了2种统计建模方法。首先,搭建了电子芯片可靠性评估试验台,利用菊花链测试芯片为试验对象获取了互联结构的退化数据和寿命数据;其次,基于断裂力学理论,分析了互联结构断裂失效的失效机理,给出了BGA封装形式的电子芯片互联结构寿命分布服从两参数威布尔分布的理论依据,由此建立了基于寿命数据的电子芯片互联结构寿命分布模型;再次,利用退化量的分布建立了基于退化量数据的电子芯片互联结构寿命分布模型;最后,对2种统计建模方法进行比较分析。结果表明,基于退化量数据的寿命分布模型与基于寿命数据的寿命分布模型吻合度良好,2种分析方法在结果上具有一致性。

关键词: 互联结构, 威布尔分布, 退化量分布, 随机振动, 故障预测与健康管理

Abstract: Aimed at the problems of complex failure mechanism of interconnected structures in electronic chips, difficulty in obtaining lifetime data, and lack of a theoretically supported life distribution model, a statistical analysis based on degraded experimental data was conducted. First, an electronic chip reliability evaluation testbed is built, and the daisy-chain test chip is used as the test object. Degradation data and lifetime data of interconnected structures are acquired. Then, based on the theory of fracture mechanics, The failure mechanism of interconnection structure is analyzed. The theoretical basis of the two-parameter Weibull distribution of the life distribution of the interconnection structure of electronic chips in BGA package is given. The life distribution model of the interconnection structure of electronic chips based on the life data is established. Next, using the distribution of degradation, a lifetime distribution model of the interconnected structures in electronic chips is established. Finally, the two statistical modeling methods are compared. The results show that the lifetime distribution model based on degradation data is in good agreement with that based on lifetime data.

Key words: interconnected structures, Weibull distribution, distribution of degradation data, random vibration, failure prediction and health management

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发