Scholarlink

<p>Silicon carbide (SiC), a third-generation semiconductor material, represents mainstream technology for future markets. However, SiC’s inherent hardness and brittleness create significant processing challenges. Conventional dicing saw cutting suffers from chipping defects and limited throughput. In contrast, laser cutting introduces heat-affected zones (HAZ) that deteriorate chip mechanical strength through recast layer formation and thermal stress propagation. This study optimizes 4&#xa0;H-SiC wafer cutting parameters using a 355&#xa0;nm nanosecond laser and develops a finite element method (FEM) model achieving &lt; 10% prediction error. Space-Filling Design, Gaussian Process Regression Surface (GPRS), and Analysis of Variance (ANOVA) investigated parameter and thermal effects on cutting quality. Quality metrics were converted to quantitative scores for Best Known Parameter (BKP) determination. The optimized BKP parameters (5.28&#xa0;W laser power, 64.16&#xa0;mm/s scanning velocity, single-pass) achieved 102&#xa0;μm cutting depth, 12.4&#xa0;μm width, 2.3&#xa0;μm HAZ width, and 75.57 quality score. Compared to baseline parameters, results demonstrated 23.43% chip strength enhancement, 17.67% surface roughness improvement, 13.76% fracture surface roughness reduction, 34.21% decreased edge residue adhesion, and 46.89% reduction in subsurface defects.</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

Finite element modeling and multi-objective optimization of 4 H-SiC wafer cutting using nanosecond UV laser: quality prediction and experimental validation

  • 期刊论文

  • 发表日期：

    2025年12月13日

查看全文
  • Cheng-Tang Pan,
  • Ju-Yueh Lin,
  • Yu-Chi Liu,
  • Shao-Yu Wang,
  • Chen-Han Lin,
  • Ming-Chan Lee

摘要

Silicon carbide (SiC), a third-generation semiconductor material, represents mainstream technology for future markets. However, SiC’s inherent hardness and brittleness create significant processing challenges. Conventional dicing saw cutting suffers from chipping defects and limited throughput. In contrast, laser cutting introduces heat-affected zones (HAZ) that deteriorate chip mechanical strength through recast layer formation and thermal stress propagation. This study optimizes 4 H-SiC wafer cutting parameters using a 355 nm nanosecond laser and develops a finite element method (FEM) model achieving < 10% prediction error. Space-Filling Design, Gaussian Process Regression Surface (GPRS), and Analysis of Variance (ANOVA) investigated parameter and thermal effects on cutting quality. Quality metrics were converted to quantitative scores for Best Known Parameter (BKP) determination. The optimized BKP parameters (5.28 W laser power, 64.16 mm/s scanning velocity, single-pass) achieved 102 μm cutting depth, 12.4 μm width, 2.3 μm HAZ width, and 75.57 quality score. Compared to baseline parameters, results demonstrated 23.43% chip strength enhancement, 17.67% surface roughness improvement, 13.76% fracture surface roughness reduction, 34.21% decreased edge residue adhesion, and 46.89% reduction in subsurface defects.