CDI-Type I: Complex Catalyst Enabled via Computational Thinking

0940768 Vlachos CDI主题：理解复杂性和从数据到知识PI提出了一个计算框架，用于合理设计（表面合金）催化剂，可用于化学，石化，环境和能源应用。催化系统在空间上是扩展的，紧急的，多尺度的，需要不同的方法来设计。该框架的发展可以指导实验工作，选择和合成具有理想性能的材料，并对各个学科和行业产生影响。 尽管催化和反应工程在工业和环境部门以及美国经济中的重要性，但催化剂合成基本上是一门艺术而不是科学。开发一个合理的催化剂设计框架面临着许多挑战，包括从周期表中选择多种元素和合适结构的组合爆炸，催化反应系统的多尺度性质，催化剂对环境的响应性（它们的结构以及因此它们的性能可以响应于环境条件而变化），异质外延结构的固有复杂性和突发行为，以及同时满足多个目标的需要（催化剂活性，选择性，稳定性）学术价值PI提出的计算框架克服了上述挑战，并有可能实现变革性的变化，催化剂设计实践，通过计算思维制定新的催化材料。这一框架将被应用于一个新兴的，复杂的材料家族（表面合金催化剂）的Co-PI（陈）之一开创。这个新的框架首先开发一个详细的，多尺度动力学模型的反应，使用最近推出的（由Vlachos的小组）分层多尺度框架，解开复杂的，多尺度的催化反应的行为。一个系统的方法将被用来达到最佳的材料性能，沿着与相应的不确定性，使用多尺度动力学模型。将通过第一原理方法为各种材料结构和元素周期表的元素建立一个大型描述符数据集。从这个数据集的数据挖掘将与系统的方法相结合，以查看一小部分合适的材料。将引入新的多尺度建模工具。最后，将对材料进行实验测试，以评估模型预测并提供反馈。整体方法利用了理解和建模系统的复杂性和紧急行为（主题1）与数据挖掘（第一原理模型）知识（主题II）的共生关系，以获得候选催化材料。更广泛的影响这项工作有可能开发一种系统的方法来设计新产品，并可能影响化学，石化，环境和能源部门。计算框架跨越多个学科，从催化，表面科学，反应工程，系统工程（优化），应用数学和计算，以达到高经济和社会利益的化学品和能源生产和减排的工业从业者。它是复杂、紧急、多维系统的产品工程原型。PI计划利用由Vlachos领导的催化科学与技术中心（CCST）的工业网络和由特拉华州能源研究所（UDEI）领导的Chen通过通讯，年度报告，年度研讨会，出版物和网络研讨会进行接触。将开发以以太网为基础的课程和研讨会传播方式。还建议通过各种方式对各级学生和博士后研究员进行多学科教育，并让代表性不足的群体的学生参与。他们还将利用工程学院与当地初中和高中之间的现有网络，展示数学，科学和工程在解决社会问题中的重要性。

0940768Vlachos CDI Themes: Understanding Complexity and From Data to KnowledgeThe PIs propose a computational framework for rational design of (bimetallic surface alloy) catalysts that can be used in chemical, petrochemical, environmental, and energy applications. Catalytic systems are spatially extended, emergent, and multiscale and demand a different approach for their design. The development of this framework can guide experimental efforts toward selection and synthesis of materials with desirable properties and can have an impact on various disciplines and industries. Despite the importance of catalysis and reaction engineering in the industrial and environmental sectors and the US economy, catalyst synthesis has by-and-large been an art rather than a science. The development of a rational framework for catalyst design has many challenges, including the combinatorial explosion in choosing multiple elements from the periodic table and suitable architectures, the multiscale nature of catalytic reaction systems, the responsiveness of catalysts to their environment (their structure and thus their performance could vary in response to the environmental conditions), the inherent complexity and emergent behavior of heteroepitaxial structures, and the need to meet multiple objectives simultaneously (catalyst activity, selectivity, and stability).Intellectual Merit The PIs proposed computational framework overcomes the above challenges and has the potential to enable transformative changes in catalyst design practices by formulating novel catalytic materials via computational thinking. This framework will be applied to an emerging, complex family of materials (surface alloy bimetallic catalysts) pioneered by one of the Co-PIs (Chen). This new framework starts by developing a detailed, multiscale kinetic model for a reaction at hand, using a recently introduced (by Vlachos' group) hierarchical multiscale framework, to unravel the complex, multiscale behavior of catalytic reactions. A systems' approach will be used to arrive at optimal material properties, along with the corresponding uncertainty, using a multiscale kinetics model. A large dataset of descriptors will be built via first principle methods for various materials architectures and elements of the periodic table. Data mining from this dataset will be coupled with the systems' approach to look at a small set of suitable materials. New multiscale modeling tools will be introduced. Finally, the materials will be tested experimentally to assess and provide feedback to the model predictions. The overall approach capitalizes on the symbiosis of understanding and modeling the complexity and emergent behavior of the systems (Theme 1) with data mining (of first principle models) to knowledge (Theme II) to arrive at candidate catalytic materials.Broader Impact This work has the potential to develop a systematic approach for designing new products and could impact the chemical, petrochemical, environmental, and energy sectors. The computational framework crosses multiple disciplines, ranging from catalysis, surface science, reaction engineering, systems' engineering (optimization), and applied mathematics and computation, to reach out to industrial practitioners for chemicals and energy production and emissions reduction of high economic and societal interest. It serves as a prototype for product engineering of complex, emergent, multidimensional systems. The PIs plan to leverage the industrial network of their center for catalytic science and technology (CCST) led by Vlachos and the University of Delaware Energy Institute (UDEI) led be Chen to reach out via newsletters, annual reports, annual symposia, publications, and webinars. Ethernet-based dissemination of courses and seminars will be developed. Multidisciplinary education of students at various levels and of postdoctoral fellows via various means and involvement of students from underrepresented groups are also proposed. They will also utilize existing networks between the college of engineering and local middle and high schools to demonstrate the importance of mathematics, science, and engineering in solving societal problems.