材料科学
倒装芯片
焊接
复合材料
电子背散射衍射
微观结构
回流焊
收缩率
Crystal(编程语言)
图层(电子)
程序设计语言
计算机科学
胶粘剂
作者
Kei Murayama,Mitsuhiro Aizawa,Takashi Kurihara
标识
DOI:10.1109/ectc.2016.140
摘要
In order to accomplish high performance and miniaturization, large size chip assembly and 3-dimensional packages has been required. Stress in ultra-low k (ULK) layer and high warpage during flip chip bonding process are the most challenging issues. For this reason, a low temperature and low stress bonding process has been required. In this report, we will investigate the strain change at bump connection area during thermal compression bonding (TCB) process and compared with that of Mass reflow process using a large size chip (20x20 mm). We will also investigate the difference of strain change between two kinds of solder, Sn3.0wt%Ag0.5wt%Cu (SAC) and Sn57wt%Bi (SnBi) solder. Microstructure and crystal orientation analyses of the connection bump were observed by Electron backscattered diffraction (EBSD). We prepared two types of organic laminate, which had different shrinkage factor 1.000 and 0.999. With SAC solder, both cases of TCB and Mass reflow, the Grain Reference Orientation Deviation (GROD) value of the edge bump was higher than that of the center bump. The size of crystal grain of Sn with TCB was smaller compared with that with Mass reflow process. Using the organic laminate with 0.999 shrinkage factor, the GROD value of the edge bump showed nearly the same value of another location. By optimizing the organic laminate shrinkage factor, TCB process is expected to achieve high reliability of the micro joining compared with Mass reflow process. On the other hand, with Sn57Bi solder, in both cases of TCB and Mass reflow, the size of crystal grain of Sn was smaller compared with that of SAC. And significant difference in crystal orientation and in micro structure were not observed between TCB and Mass reflow. Sn-Bi solder joining are strong candidate materials for stress reduction for large size chip application and 3D packages.
科研通智能强力驱动
Strongly Powered by AbleSci AI