 |
International Society of Science and Applied Technologies |
| Machine Learning Integrated with Process Capability Model for Quality Testing Model of Semiconductor Wire Bonding Process | ||||
| Author | Ching-Hsin Wang
|
|||
| Co-Author(s) | Chun-Liang Tung; Chun-Chieh Huang
|
|||
| Abstract | In recent years, Taiwan’s semiconductor industry has developed vigorously. Especially, wire-bonding plays an important role in the semiconductor industry, widely used in the welding technology of the semiconductor industry. If the quality of the wire bonding process is not good, it will directly affect the circuit conducting state or cause an open circuit. Therefore, this study employed the technology of machine learning, targeted at parameters such as gold ball size, loop height, and gold ball height for the wire bonding process, to build and train the quality control testing model using the support vector machine (SVM) of the supervised learning model as well as solved the best parameters. Finally, in the verification stage, this study developed and constructed an evaluation model of the semiconductor wire bonding process capability for the final gold wire pulling effect, which verified the reliability of the gold wire pulling force. This study proposed a convenient evaluation tool to help manufacturers improve their product quality and market competitiveness.
|
|||
| Keywords | Semiconductor Wire Bonding Process, Machine Learning, Quality Testing, process capability, support vector machine | |||
| Article #: RQD27-53 | ||||
Proceedings of 27th ISSAT International Conference on Reliability & Quality in Design |