• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2026, Vol. 62 ›› Issue (4): 215-223.doi: 10.3901/JME.260119

• 材料科学与工程 • 上一篇    

扫码分享

计算机辅助设计多峰非球体铜焊膏及其低温烧结特性研究

朱志宏, 缪均唯, 汪宇鉴, 李万里   

  1. 江南大学智能制造学院 无锡 214122
  • 收稿日期:2025-03-24 修回日期:2025-09-13 发布日期:2026-04-02
  • 作者简介:朱志宏,男,2003年出生。主要研究方向为计算机辅助多峰金属焊膏设计与评估。E-mail:zhuzhihong9527@126.com
    李万里(通信作者),男,1989年出生,博士,副教授,硕士研究生导师。主要研究方向为电子增材制造技术、功率半导体封装技术。E-mail:li.wanli@jiangnan.edu.cn
  • 基金资助:
    再制造技术国家级重点实验室基金(61420052023WD006)和基础研究(50904010201)资助项目。

Computer-aided Design of Multimodal Non-spherical Cu Pastes and Their Low-temperature Sintering Behaviors

ZHU Zhihong, MIU Junwei, WANG Yujian, LI Wanli   

  1. School of Intelligent Manufacturing, Jiangnan University, Wuxi 214122
  • Received:2025-03-24 Revised:2025-09-13 Published:2026-04-02

摘要: 低温铜烧结互连技术凭借出色的热力学稳定性与封装可靠性,成为宽禁带半导体器件高可靠耐热互连的核心潜力技术之一。但烧结互连层的多孔结构会因孔隙率问题降低其导电导热能力,制约接头性能提升。相比于单峰铜颗粒焊膏,多峰铜颗粒焊膏虽可大幅度提高烧结互连层致密度,却面临开发效率与配比优化的挑战。如何快速确定多峰颗粒的最佳配比以实现最大堆积密度,降低烧结互连层的孔隙率,成为推动该技术实用化的关键。为此,基于蒙特卡洛模型设计了多峰颗粒随机堆积算法,用于模拟不同尺寸球体、非球体(椭圆片)颗粒的堆积过程并计算堆积密度,以获取特定尺寸下的最优颗粒配比。结果表明:单峰球体随机堆积密度稳定在0.61左右;对于双峰球体体系,在固定条件下,颗粒尺寸差异越大堆积密度越高,且在特定范围内,堆积密度峰值始终出现在小颗粒占比为30%时;对于特定尺寸的球体与椭圆片混合体系,当球体质量占比为70%~80%时,堆积密度达到最优。相关烧结实验进一步验证了该算法的实用性与可靠性,为辅助多峰非球体铜颗粒焊膏设计、加速高性能焊膏开发提供支撑,助力多峰铜焊膏技术突破。

关键词: 铜烧结互连技术, 计算机辅助设计, 蒙特卡洛, 随机堆积, 非球体颗粒

Abstract: Low-temperature Cu sinter-joining technology, with excellent thermodynamic stability and packaging reliability, is regarded as one of the core potential technologies for high-reliability and heat-resistant interconnections in wide-bandgap semiconductor devices. However, the porous structure of sintered-joints reduces its electrical and thermal conductivity due to porosity issues, restricting the improvement of joints performance. Compared with unimodal Cu particle pastes, multimodal Cu particle pastes can significantly improve the density of joints, but they face challenges in terms of development efficiency and ratio optimization. How to quickly determine the optimal ratio of multimodal particles to achieve maximum packing density and reduce the porosity of joints has become the key to promoting the practical application of this technology. For this purpose, a random packing algorithm for multimodal particles was designed based on the Monte Carlo model, which is used to simulate the packing process of spherical and non-spherical (elliptical flakes) particles of different sizes and calculate the packing density, in order to obtain the optimal particle ratio at a specific size. The results show that the random packing density of the unimodal sphere is stable at around 0.61; for the bimodal spheres system, under fixed conditions, the larger the particle size difference, the higher the packing density, and within a specific range, the peak of packing density always appears when the proportion of small particles is 30%; for the mixed system of spheres and elliptical flakes with specific sizes, the packing density reaches the optimal when the proportion of spheres is 70%~80%. The relevant sintering experiments further verified the practicality and reliability of the algorithm, providing support for assisting in the design of multimodal non-spherical Cu particle pastes, accelerating the development of high-performance pastes, and facilitating the breakthrough of multimodal Cu pastes technology.

Key words: Cu sinter-joining technology, computer-aided design, Monte Carlo model, random packing, non-spherical particle

中图分类号: