Scholarlink

This research utilises artificial neural networks (ANNs) analysis to investigate gold diffusion based on gold content weight percentage (wt%) in the intermetallic compound (IMC) layer of the ball grid array (BGA) component solder joints under various conditions. The presence of a gold element in the IMC layer on the BGA component side significantly affects the reliability of SnAgCu (SAC) solder joints. Using no rework as a control sample, the ANNs model evaluated and predicted the effects of hot air and laser rework on the gold content wt% in the BGA component side’s IMC layer. Energy-dispersive X-ray spectroscopy (EDX) measurements of the corresponding gold content wt%, soldering process parameters, and IMC layer height were included in the datasets. The conditions tested were thermal cycle test (TCT) for 0 and 900 cycles, no rework, hot air rework, and laser rework. The network design consists of three levels: an input layer, one hidden layer with different activation functions, and one output layer. With an R-Square value of 0.880 for the validation set, the ANNs model demonstrated good prediction of gold diffusion in solder joints and closely matched the actual values. By leveraging advanced computational techniques, this research provides insights into optimizing the BGA rework process and enhancing solder joint reliability in electronic manufacturing and repair applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Artificial Neural Networks Analysis for Predicting Gold Diffusion in Solder Joints Under Various Rework and Thermal Cycling Conditions

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Adlil Aizat Ismail,
  • Maria Abu Bakar,
  • Azman Jalar,
  • Mohd Ridzwan Yaakub,
  • Muhammad Iqbal Abu Latiffi,
  • Erwan Basiron,
  • Muhammad Nizam Ilias

摘要

This research utilises artificial neural networks (ANNs) analysis to investigate gold diffusion based on gold content weight percentage (wt%) in the intermetallic compound (IMC) layer of the ball grid array (BGA) component solder joints under various conditions. The presence of a gold element in the IMC layer on the BGA component side significantly affects the reliability of SnAgCu (SAC) solder joints. Using no rework as a control sample, the ANNs model evaluated and predicted the effects of hot air and laser rework on the gold content wt% in the BGA component side’s IMC layer. Energy-dispersive X-ray spectroscopy (EDX) measurements of the corresponding gold content wt%, soldering process parameters, and IMC layer height were included in the datasets. The conditions tested were thermal cycle test (TCT) for 0 and 900 cycles, no rework, hot air rework, and laser rework. The network design consists of three levels: an input layer, one hidden layer with different activation functions, and one output layer. With an R-Square value of 0.880 for the validation set, the ANNs model demonstrated good prediction of gold diffusion in solder joints and closely matched the actual values. By leveraging advanced computational techniques, this research provides insights into optimizing the BGA rework process and enhancing solder joint reliability in electronic manufacturing and repair applications.