| About this Abstract |
| Meeting |
2023 TMS Annual Meeting & Exhibition
|
| Symposium
|
Phase Stability, Phase Transformations, and Reactive Phase Formation in Electronic Materials XXII
|
| Presentation Title |
Machine Learning Models of Ultimate Tensile Strength and Elongation for Low-temperature Solder |
| Author(s) |
Yu-chen Liu, Ahmad Kholik, Shih-kang Lin |
| On-Site Speaker (Planned) |
Yu-chen Liu |
| Abstract Scope |
Low-temperature solder is a key material in solving warpage issue arising from the rather high reflow temperature in the advanced electronic packaging. Currently, Sn-Bi solder system has been viewed as a promising material system for low-temperature solder. Nevertheless, (Bi) phase coarsening after thermal aging causes the brittleness of the solder and thus decreases the reliability in real application. Element doping is typically applied in tailoring the solder properties. However, it is not economically feasible to tailor properties in the multi-component system by trial-and-error method. Therefore, this study used machine learning method to build up models in predicting ultimate tensile strength and elongation of the as-cast and aged Sn-Bi-X solders, where X represents the doping elements. Series of model evaluation methods including cross-validations and cross-plots analysis suggested that the model showed some predictive ability. The model was then used in designing promising solder system. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Other, Other, Other |