SBIR Phase II: Simulation of Rapid Thermal Processing in a Distributed Computing Environment

该小型企业创新研究(SBIR)第二阶段项目将利用第一阶段的研究成果，继续开发和演示用于在分布式计算环境中详细模拟快速热处理(RTP)的计算工具。RTP已成为制造先进半导体器件的关键技术。随着晶片尺寸的增大和芯片尺寸的减小，通过高性能计算的数值模拟来理解辐射换热、瞬变流体流动和换热以及化学反应等高度耦合的物理过程，是设计、优化和控制RTP反应器的关键。在第二阶段，将建立一个基于改进的离散传递法(MDTM)的三维表面辐射模型，将光源和过程腔中的辐射传递作为一个整体来处理。通过有限体积法严格求解时间域麦克斯韦方程，可以考虑详细的图样效应。通过在Navier-Stokes方程中加入Burnett项来模拟低压RTP中的稀薄气体动力学。包含各种多学科物理模型的控制方程将用三维非结构有限体积法来求解。为了满足计算密集型三维模拟的需要，将在求解过程中实施有效的并行策略。并行处理机之间的数据通信将由消息传递接口(MPI)库进行。为了加快整体解的收敛速度和提高并行性能，采用代数多重网格(AMG)方法对每个处理器中的离散方程进行求解。所提出的模拟工具可用于系统地研究RTP系统中发生的潜在物理现象，并有助于RTP电抗器的设计、优化和控制。该模拟工具将对需要详细了解多模和高耦合输运现象的半导体制造设备行业有很大的帮助。潜在的应用包括RTP反应堆和许多其他制造和材料处理系统的设计、优化和控制。

This Small Business Innovation Research (SBIR) Phase II project will continue to develop and demonstrate a computational tool for detailed simulation of Rapid thermal processing (RTP) in a distributed computingenvironment by taking advantages of the findings in Phase I. RTP has become a key technology in thefabrication of advanced semiconductor devices. As wafers get larger and chip dimensions smaller, theunderstanding of the highly coupled physics such as radiative heat transfer, transient fluid flow and heattransfer as well as chemical reactions through numerical modeling using high-performance computing isthe key to the design, optimization, and control of RTP reactors. In Phase II, A 3D surface radiationmodel based on the modified discrete transfer method (MDTM) will be developed to treat radiativetransfer in the lamphouse and process chamber as a whole process. The detailed pattern effects will betaken into account by rigorously solving time-domain Maxwell's equations through a finite volumeapproach. The rarefied gas dynamics in low pressure RTP will be modeled by adding Burnett terms intothe Navier-Stokes equations. The governing equations that contain various multi-disciplinary physicalmodels will be solved by a 3D unstructured finite volume method. To address computationally intensive3D simulation needs, an efficient parallel strategy will be implemented in the solution procedure. Datacommunication among parallel processors will be conducted by the Message Passing Interface (MPI)library. To accelerate the overall solution convergence and improve the parallel performance, thealgebraic multi-grid (AMG) method will be used to solve the discretized equations in each processor. It isexpected that the proposed simulation tool can be used to systematically investigate the underlyingphysics occurring in RTP systems, and to help in the design, optimization, and control of RTP reactors.The proposed simulation tool will significantly benefit the semiconductor manufacturing equipment industries that require a detailed understanding of multimode and highly coupled transport phenomena. The potential applications include the design, optimization, and control of RTP reactors and many other manufacturing and materials processing systems.