Control Design with Real-Time Implementation for Rapid Thermal Processes

快速热处理实时实现的控制设计

• 批准号: 9905729
• 负责人: A. Nazli Gundes
• 金额: $ 24万
• 依托单位: University of California-Davis
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1999
• 资助国家: 美国
• 起止时间: 1999-09-01 至 2002-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9905729&HistoricalAwards=false
• 关键词: Control; Design; Real; Time; Implementation

The long term goal of this project is to establish a multi-disciplinary research program on control theory, design, and real-time implementation for Rapid Thermal Process chambers. In the initial two years of the project, the focus will be on preliminary modeling, control design, and real-time implementation issues.Rapid Thermal Processes demand fast tracking control laws that achieve near uniform spatial temperature distributions across a semiconductor wafer during both transient and quasi steady-state phases of the process. Due to the radiative effects in a relatively small chamber volume and limitations in remote measurement techniques, achieving the stringent performance specifications for Rapid Thermal Process chambers becomes a very challenging model-based nonlinear design problem. To meet future technical challenges in microelectronics manufacturing, it is essential to develop modern control and optimization techniques with strong emphasis on real-time implementation.The ultimate goal of this effort is to pose and solve control design problems so that the control laws, when implemented on the on-site testbed (Rapid Thermal Chemical Vapor Deposition Epitaxial Reactor), result in real-time performance improvement and run-to-run repeatability. In order to pose the problem in a control setting, the most important step is to develop mathematical models of the process with varying levels of complexity: from complex validation models to simpler models suitable for control design. This step is closely linked to the real-time implementation setup since it involves chamber-specific actuator and sensor interfacing. Moreover, analytical model derivation based on first-principles is completely dependent on chamber-specific properties (such as material, geometry, actuation, measurements) and the desired reaction recipes. Therefore, this stage of the project involves direct interaction and non-trivial hardware/software interfacing issues specific to the on-site testbed. Since these research results can be immediately tested on the on-site testbed, the effort will start with control techniques based on simple analytical models, assess achievable performance limitations by implementing the designs, and then increase model and control design complexity to meet the desired process performance. Although modeling, control design and real-time implementation objectives must be carried out simultaneously to assess performance, the main effort in the initial two years will be allocated to modeling and real-time interfacing problems.The unique advantage available to the project is the on-site Rapid Thermal Process reactor, which will be used as the project testbed in collaboration with the Microelectronics Group in the Electrical and Computer Engineering Department at the University of California, Davis. The proposed research will complement the ongoing work of the Microelectronics Group by providing crucial systematic modeling approaches and controller design methods for real-time implementation. The beginning stages of this research will initiate interaction between the two research groups and will enhance the research and educational experience of graduate students in the inherently different areas of control systems and microelectronics. Through this multi-disciplinary teamwork, graduate students involved in this research will obtain valuable practical training that will prepare them for engineering challenges in future industry and research positions. It is also expected that this research will lead in the future to experimental set-tips that can be used as design exercises for graduate and undergraduate control courses.

本计画的长期目标是建立一个多学科的快速热制程制程室控制理论、设计与即时执行的研究计画。 在项目的最初两年，重点将放在初步建模，控制设计和实时实施issues.Rapid热处理要求快速跟踪控制律，实现在瞬态和准稳态阶段的半导体晶片上的空间温度分布接近均匀的过程。 由于辐射效应在一个相对较小的腔室体积和远程测量技术的限制，实现快速热处理室的严格的性能指标成为一个非常具有挑战性的基于模型的非线性设计问题。 为了满足微电子制造业未来的技术挑战，必须发展现代控制和优化技术，特别强调实时实现，这一努力的最终目标是提出并解决控制设计问题，以便控制律在现场试验台上实现时，（快速热化学气相沉积外延反应器），导致实时性能的改善和运行到运行的可重复性。 为了在控制设置中提出问题，最重要的步骤是开发具有不同复杂程度的过程的数学模型：从复杂的验证模型到适合控制设计的简单模型。 该步骤与实时实施设置密切相关，因为它涉及腔室特定的致动器和传感器接口。 此外，基于第一原理的分析模型推导完全取决于腔室特定的性质（例如材料、几何形状、致动、测量）和期望的反应配方。 因此，该项目的这一阶段涉及到现场试验台特有的直接交互和重要的硬件/软件接口问题。 由于这些研究结果可以立即在现场试验台上进行测试，因此将从基于简单分析模型的控制技术开始，通过实施设计来评估可实现的性能限制，然后增加模型和控制设计的复杂性以满足所需的过程性能。 虽然建模、控制设计和实时实施目标必须同时进行以评估性能，但最初两年的主要工作将分配给建模和实时接口问题。该项目的独特优势是现场快速热处理反应器，它将被用作与加州大学电气和计算机工程系的微电子组合作的项目测试平台，戴维斯 拟议的研究将补充微电子集团正在进行的工作，提供关键的系统建模方法和控制器设计方法的实时实施。 这项研究的开始阶段将启动两个研究小组之间的互动，并将提高研究生在控制系统和微电子学的固有不同领域的研究和教育经验。 通过这种多学科的团队合作，参与这项研究的研究生将获得宝贵的实践培训，这将使他们为未来的工业和研究职位的工程挑战做好准备。 它也预计，这项研究将导致在未来的实验设置提示，可用作设计练习的研究生和本科控制课程。