中国电子学会电子制造与封装技术分会会刊

中国半导体行业协会封测分会会刊

无锡市集成电路学会会刊

导航

电子与封装

• 综述 •    下一篇

多学科协同机械跨界驱动异构集成应用

张墅野,胡星宇,何鹏   

  1. 哈尔滨工业大学材料结构精密焊接与连接全国重点实验室,哈尔滨  150001
  • 收稿日期:2026-01-26 修回日期:2026-02-09 出版日期:2026-03-03 发布日期:2026-03-03
  • 通讯作者: 张墅野
  • 基金资助:
    松山湖材料实验室开放课题基金(2023SLABFN33)

Multidisciplinary Collaboration, Mechanical Transboundary Driven Heterogeneous Integration Applications

ZHANG Shuye, HU Xingyu, HE Peng   

  1. State Key Laboratory of Precision Welding & Joining of Materials and Structures, Harbin Institute of Technology, Harbin 150001, China
  • Received:2026-01-26 Revised:2026-02-09 Online:2026-03-03 Published:2026-03-03
  • Contact: Shuye Zhang

摘要: 异构集成是后摩尔时代提升集成电路性能的关键路径,但其高密度与多材料特征引发了严峻的电-热-力多物理场耦合挑战。围绕多学科协同与机械跨界驱动,系统综述了异构集成的协同设计方法、跨尺度建模仿真技术及工程应用进展。研究揭示,机械因素通过热膨胀系数失配诱导的残余应力与界面翘曲,深度耦合电迁移与焦耳热动力学,已成为决定互连稳定性与系统能效边界的核心变量。为有效应对耦合瓶颈,总结了一种跨层级、双向闭环的协同设计范式。以系统功耗、带宽等为目标正向驱动封装-互连-材料协同选型,并以多物理场仿真结果反向约束架构布局与工艺窗口,同时借助跨尺度建模实现关键参数的双向传递与快速收敛。展望未来,指出需进一步融合人工智能驱动的下一代电子设计自动化(EDA)工具链,推动从设计协同向标准化的生态协同迈进,为芯粒及三维集成电路(3D IC)的高可靠性演进提供物理支撑。

关键词: 异构集成, 多学科协同设计, 多物理场模拟, 机械驱动

Abstract: Heterogeneous integration technology represents a key pathway for enhancing integrated circuit performance in the post-Moore's Law era. However, its high density and multi-material characteristics pose severe challenges in coupled electrical-thermal-mechanical multi-physics.  Centering on multidisciplinary collaboration and mechanical cross-boundary drivers, this paper systematically reviews co-design methodologies, multi-scale modeling and simulation techniques, and engineering application advancements in heterogeneous integration. The research reveals that mechanical factors such as residual stresses and interface warping induced by Coefficient of Thermal Expansion mismatch are deeply coupled with electromigration and Joule heating dynamics. These factors have become core variables that determine interconnect stability and system energy efficiency boundaries. To effectively address these coupling bottlenecks, this paper summarizes a cross-hierarchical and bidirectional closed-loop co-design paradigm. This framework uses system power consumption and bandwidth as primary targets to drive the forward co-selection of packaging, interconnects, and materials. Meanwhile, multi-physics simulation results provide reverse constraints for architectural layouts and process windows, while multiscale modeling enables the bidirectional transfer of critical parameters and rapid convergence. Looking forward, the article suggests the further integration of Artificial Intelligence (AI)-driven next-generation Electronic Design Automation (EDA) toolchains. This will promote the transition from design synergy to standardized ecosystem collaboration, providing the necessary physical support for the high-reliability evolution of Chiplets and Three-Dimensional Integrated Circuits (3D ICs).

Key words: heterogeneous integration, multidisciplinary co-design, multiphysics simulation, mechanical-driven