Electrical interconnects play important roles in both the performance and reliability of integrated circuits. As the operating frequency has increased and products have been exposed to harsh environments such as temperature extremes and thermal cycling, it has become essential to quantify the thermal effects on the reliability of such circuits. In particular, the effects on the signal transmission performance of the interconnects and the evolution of cracks caused by thermo-mechanical stresses need to be investigated. This study proposes that both research goals can be accomplished by S-parameter analysis alone. First, the performance variation was experimentally quantified by in situ measurements of S21 (an indicator of signal transmission performance) on bond-wire test vehicles at static temperatures between -60 °C and 120 °C. An impedance model of the interconnect was developed focusing on thermal expansion effects and it predicted the performance variation of the interconnect successfully at the given temperatures. Second, the results of ex situ measurements of S11 (an indicator of return signal) obtained using a test vehicle under thermal cycling were compared with scanning electron microscopy images of crack propagation in the interconnect. In the crack initiation phase, a resonance peak occurred on the S-parameter pattern and the peak moved toward lower frequencies as the crack propagated until failure. SPICE simulations using a suggested impedance model of the cracked interconnect confirmed the experimental results. In addition, this study demonstrates that monitoring the resonant frequency in the S-parameter pattern as a prognostic factor can predict the initiation and evolution of cracks earlier and is more sensitive than the conventional dc resistance measurement method.

中文翻译：

---

通过 S 参数分析量化恶劣温度环境下键合线互连的性能变化和裂纹演化


电气互连在集成电路的性能和可靠性方面发挥着重要作用。随着工作频率的增加以及产品暴露在极端温度和热循环等恶劣环境中，量化热效应对此类电路可靠性的影响变得至关重要。特别是，需要研究对互连信号传输性能的影响以及热机械应力引起的裂纹的演变。本研究提出，这两个研究目标都可以仅通过 S 参数分析来实现。首先，通过在 -60 °C 至 120 °C 的静态温度下对键合线测试车辆上的 S21（信号传输性能指标）进行原位测量，对性能变化进行实验量化。开发了互连的阻抗模型，重点关注热膨胀效应，并成功预测了给定温度下互连的性能变化。其次，将使用测试车辆在热循环下获得的 S11（返回信号指标）的非原位测量结果与互连中裂纹扩展的扫描电子显微镜图像进行比较。在裂纹萌生阶段，S 参数图案上出现共振峰值，并且随着裂纹扩展直至失效，该峰值向较低频率移动。使用推荐的破裂互连阻抗模型进行的 SPICE 模拟证实了实验结果。 此外，这项研究表明，监测 S 参数模式中的谐振频率作为预测因素可以更早地预测裂纹的萌生和演变，并且比传统的直流电阻测量方法更灵敏。