Advanced Metallurgical Reliability Issues in Flip Chip Technology

本研究探讨覆晶焊点所受之电性力、热机械力及化学力之交互作用。先前的NSF项目确定了倒装芯片焊点中的电迁移成为硅器件互连系统中的薄弱环节的基本原因。 当电迁移与热应力和固态老化（Kirkendall空洞形成）相结合时，电子公司遇到了许多新的和意想不到的倒装芯片焊点可靠性问题。 这些独特的冶金可靠性问题，倒装芯片焊点研究在此授权。 从研究中获得的科学理解不仅增强了焊点可靠性的基础知识，而且还可以优化先进的制造工艺和材料。 这并不是因为低熔点焊料合金中的原子扩散比Al或Cu中的原子扩散快。 实际上，在器件工作温度下，Al中的晶界扩散和Cu上的表面扩散接近于焊料合金中的体扩散。 相反，这是因为焊料合金的电迁移的临界积比Al和Cu中的电迁移的临界积小两到三个数量级。 因此，电迁移可以通过比使Al或Cu线失效所需的电流密度小两到三个数量级的电流密度使焊点失效。 此外，还有其他几个因素会增强焊点中的电迁移。 倒装芯片焊点的独特的线-凸点结构在线-凸点接触处产生了严重的电流拥挤现象。 硅芯片角部的连接处处于非常大的热应力下，并且应力集中与电流拥挤重叠。 焊料与凸块下金属化反应非常快，并且由于极性效应，反应速率和产物可以受到电迁移的影响;阴极处的反应不同于阳极处的反应。 在用于先进电子消费产品的高密度系统级封装中对倒装芯片技术的需求正在迅速增长。 具有焊料凸块的区域阵列的倒装芯片技术是能够满足用于直接芯片附接的非常大量的输入/输出互连的要求的唯一现有技术。 今天，在工业中进行的基础材料研究很少，尽管电子设备中的材料加工和可靠性问题非常微妙和复杂。 过去，电子公司能够击败国际半导体技术路线图（由半导体工业协会发布）中预测的时间表，并获得市场领先时间。如今，将低k材料集成到Si器件中的时间表正在缩短！ 同样，由于缺乏对无铅焊料可靠性的了解，大型计算机制造商获得了豁免，允许在2010年之前使用高铅焊料。 这两个案例揭示了材料研究在电子制造相关大学中的重要作用。在不久的将来，没有其他技术可以提供大量的输入/输出芯片上的互连作为倒装芯片技术。 这项对焊点在电、机械和化学力组合下的基本行为的研究将有利于美国的大型计算机和消费电子产品。

In this grant, interactions among the electrical force, thermo-mechanical force, and chemical force on flip chip solder joint are examined. A previous NSF project established the basis reasons why electromigration in flip chip solder joints becomes a weak link in the interconnect system on silicon devices. When electromigration is combined with thermal stress and solid-state aging (Kirkendall void formation), many new and unexpected reliability problems of flip chip solder joints are encountered by electronic companies. These unique metallurgical reliability issues in flip chip solder joints are studied in this grant. Scientific understanding from the study not only enhances basic knowledge of reliability of solder joints, but also enables optimization of advanced manufacturing processes and materials. It is not because atomic diffusion in the low melting point solder alloy is faster than that in Al or Cu. Actually, grain boundary diffusion in Al and surface diffusion on Cu is close to that of bulk diffusion in solder alloys at the device operation temperature. Rather it is because the.critical product of electromigration of solder alloys is two to three orders of magnitude smaller than that in Al and Cu. So electromigration can fail a solder joint by a current density that is two to three orders of magnitude less than that needed to fail Al or Cu lines. In addition, there are several other factors that enhance electromigration in solder joints. The unique configuration of line-to-bump of the flip chip solder joint produces a serious current crowding phenomenon at the contact between the line and the bump. The joints at the corners of a Si chip are under a very large thermal stress, and the stress concentration overlaps the current crowding. The solder reacts very fast with under-bump metallization and the reaction rate and products can be affected by electromigration due to the polarity effect; the reaction at the cathode is different from that the anode. The demand for flip chip technology in high-density system-in-packaging for advanced electronic consumer products is growing rapidly. Flip chip technology with an area array of solder bumps is the only existing technology that can meet the requirement of a very large number of input/output interconnects for direct chip attachment. Today, little basic materials research is conducted in industry, although materials processing and reliability problems in electronic devices are quite subtle and complicated. In the past, electronic companies were able to beat the timetable projected in the International Semiconductor Technology Roadmap (published by the Semiconductor Industry Association) and gained market lead-time. Today, the schedule for integrating low-k materials into Si devices is pulling back! Similarly, due to the lack of knowledge on reliability of Pb-free solders, mainframe computer makers have obtained exemptions for allowing the use of high-Pb solder until 2010. These two cases reveal the important role of materials research in universities related to electronics manufacturing. In the near future, there is no other technology that can deliver the large number of input/outputs on-chip interconnections as flip chip technology. This study of the basic behavior of solder joints under a combination of electrical, mechanical and chemical forces will benefit mainframe computers as well as consumer electronic products in the US.