Scholarlink

<p>A methodology for wafer grinding to produce pre-designated wafer profiles is proposed. This methodology includes methods for grinding trajectory simulation, chuck table identification, wafer profile quantification, and an inverse approach based on machine learning algorithms—specifically, determining the grinding setup parameters required to achieve desired wafer profiles. Issues regarding the influence of chuck table shape on wafer profiles and how to reduce its impact when creating desired wafer profiles are addressed. In particular, the chuck shape is identified during grinding, which accounts for uncertainties caused by grinding forces on the chuck table, such as table deformation. Experimental results show that the maximum deviation between the pre-designated wafer profile and that produced by the proposed methodology is approximately 0.4&#xa0;μm, and the wafer total thickness variation (TTV) was reduced from 3&#xa0;μm to 1.2&#xa0;μm after implementing the proposed grinding parameter design methodology, which validates the effectiveness of the proposed methodology in practical manufacturing scenarios. This study provides the wafer manufacturing industry with a methodology for rapidly designing process parameters to achieve pre-designated wafer profiles, thereby significantly enhancing wafer shape control capability and process stability in wafer grinding.</p>

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

Wafer grinding parameter designs with pre-designated wafer profiles

  • 期刊论文

  • 发表日期：

    2026年2月27日

查看全文
  • Tzu-Fan Chiang,
  • Sheng-Min Guo,
  • Wen-Nan Cheng,
  • Chih-Chun Cheng,
  • Chung-Yu Tsai

摘要

A methodology for wafer grinding to produce pre-designated wafer profiles is proposed. This methodology includes methods for grinding trajectory simulation, chuck table identification, wafer profile quantification, and an inverse approach based on machine learning algorithms—specifically, determining the grinding setup parameters required to achieve desired wafer profiles. Issues regarding the influence of chuck table shape on wafer profiles and how to reduce its impact when creating desired wafer profiles are addressed. In particular, the chuck shape is identified during grinding, which accounts for uncertainties caused by grinding forces on the chuck table, such as table deformation. Experimental results show that the maximum deviation between the pre-designated wafer profile and that produced by the proposed methodology is approximately 0.4 μm, and the wafer total thickness variation (TTV) was reduced from 3 μm to 1.2 μm after implementing the proposed grinding parameter design methodology, which validates the effectiveness of the proposed methodology in practical manufacturing scenarios. This study provides the wafer manufacturing industry with a methodology for rapidly designing process parameters to achieve pre-designated wafer profiles, thereby significantly enhancing wafer shape control capability and process stability in wafer grinding.