Collaborative Research: NanoMine: Data Driven Discovery for Nanocomposites

技术摘要在过去的15年里，对纳米增强聚合物的研究取得了爆炸性的进展，提供了许多从改变热、机械、电、扩散、光学和其他性能的性能改进的例子。实验数据和模拟数据的数量，再加上单个组成相材料的数据，是惊人的。与此同时，虽然一些潜在的性质变化的机械原理已经被慢慢揭开，但我们深入理解潜在原理并从已知的加工步骤设计具有所需特性的新纳米结构聚合物的能力受到缺乏信息集成的严重限制。要确定观察到的属性变化的类型，需要手动搜索在线期刊数据库，全面阅读文章，并手动积累信息，然后综合信息。这种方法确保了许多相关的论文和文章被忽视，只允许对数据进行初步综合和了解加工-结构-性能关系。材料基因组概念的诞生为发展对材料的理解和设计新的材料概念提供了新的范式。在这个研究项目中，我们在聚合物纳米复合材料领域解决了这一挑战。虽然材料基因组方法在金属领域取得了一些成功，但聚合物领域的发展相当不发达，而且没有资源用于纳米复合材料系统。然而，由于聚合物纳米复合材料具有无限的设计空间，它是一种新的数据驱动方法的主要系统。这项工作的智力价值是将材料基因组概念应用于聚合物纳米复合材料的复杂材料系统，目的是揭示加工-结构-性能之间的关系。总体框架是将材料的响应视为加工条件、成分、相互作用和形态的函数。具体成就包括1)开发用于存储和交换聚合物纳米复合材料数据的数据资源(Nanomine)，2)开发简化描述符集以表征数据和量化结构，以及开发新的数据挖掘方法以发现潜在的材料物理，3)集成模拟工具以增强实验数据并能够探索设计概念。这项工作的更广泛影响是纳米矿数据资源本身，这是一种新的数据驱动的材料理解和发现方法，通过使用这些工具，能够在加工、产生的材料形态和性能之间建立更深层次的联系。创建一个开放源码的、可免费获取的数据资源，不仅将提供一个快速、方便的信息来源，而且还将以新的方式将研究人员联系在一起。应用于这一纳米复合材料系统的数据驱动方法，将提供可以扩展到其他材料系统的策略，极大地扩大其影响。我们还将通过跨学科研究生教育将研究和教育结合起来，包括一门基于特殊项目的课程。我们将把本科生纳入我们的研究项目。这批研究生和本科生将会有跨学科的方法来发现材料.我们还将通过NU和RPI高中外展日(如设计你的未来日和女孩职业日)接触更广泛的社区，向他们传授材料设计和数据驱动的研究。在过去15年中，纳米颗粒增强聚合物的发展创造了具有非凡性能的新材料-例如导电但透明的塑料，保持弹力更长的网球，以及更坚硬、更坚固的汽车和飞机结构塑料。然而，由于缺乏完整的信息，无论是实验数据还是模拟工具，这些先进新材料的开发都非常缓慢。目前，要了解该领域的状况，需要手动搜索在线期刊数据库，阅读全部文章，并手动积累信息，然后进行综合。目前还不存在收集数据的资源，也没有工具可以有效地挖掘收集的数据以寻找相关性，更不用说能够快速设计新材料了。在这项研究中，我们将开发一个数据资源(Nanomine)，用于存储和交换聚合物纳米复合材料数据，并开发新的数据挖掘方法，以发现潜在的材料物理。我们将集成模拟工具来增强实验数据，并允许探索设计概念。创建一个开放源码的、可免费访问的数据资源将提供一个快速和容易的信息来源，并将使人们能够从根本上对材料有新的理解，以及更有效地设计材料。应用于这一纳米复合材料系统的数据驱动方法，将提供可以扩展到其他材料系统的策略，极大地扩大其影响。我们还将在几个方面整合研究和教育，从K-12到实习工程师。

Technical AbstractIn the past 15 years, research into nanoreinforced polymers has exploded, providing numerous examples of property enhancements ranging from altered thermal, mechanical, electrical, diffusion, optical and other properties. The amount of experimental data and simulation data, compounded with the data on the individual constituent phase materials is staggering. At the same time, while some mechanistic principles underlying property changes have been slowly uncovered, our ability to both deeply understand the underlying principles and to design new nanostructured polymers with desired properties from known processing steps is severely limited by the lack of integration of the information. To determine the type of property changes that have been observed requires manual searching of online journal databases, full reading of the articles, and manual accumulation and then synthesis of the information. This approach ensures that many relevant papers and articles are overlooked and allows only rudimentary synthesis of data and understanding of the processing-structure-property relationships. The birth of the materials genome concept provides a new paradigm for developing understanding of materials and designing new material concepts. In this research project, we tackle this challenge in the domain of polymer nanocomposites. While the materials genome approach has had some success in the metals field, the polymers area is considerably less developed and no resources exist for nanocomposite systems. Yet with the infinite design space available to polymer nanocomposites, it is a prime system for a new data driven approach. The Intellectual Merit of the work is application of materials genome concepts to the complex material system of polymer nanocomposites, with the goal of uncovering the processing-structure-property relationships. The overarching framework is to consider the material response as a function of processing conditions, constituents, interactions and morphology. Specific accomplishments include 1) development of a data resource (NanoMine) for housing and exchange of polymer nanocomposite data, 2) development of reduced descriptor sets to characterize data and quantify structure, and development of new data mining methods to enable discovery of underlying material physics, 3) integration of simulation tools to augment experimental data and enable exploration of design concepts. The Broader Impacts of the work are the NanoMine data resource itself, the new data-driven approach for materials understanding and discovery, and through use of these tools the ability to make deeper connections between processing, resulting material morphology and properties. The creation of an open-source, freely accessible data resource will provide not only a fast and easy source of information, but will also link researchers together in new ways. The data driven approach applied to this one system of nanocomposites, will provide strategies that can be extended to other material systems, greatly extending its influence. We will alsointegrate research and education through interdisciplinary graduate education including a special project based course. We will include undergraduates in our research program. This cadre of graduate and undergraduate students will have an interdisciplinary approach tomaterials discovery. We will also reach out to the broader community through NU and RPI High School outreach days (such as Design Your Future Day and Career Day for Girls) and teach them about Materials Design and data driven research.Development of nanoparticle reinforced polymers in the past 15 years has created new materials with extraordinary properties - such as conducting yet transparent plastics, tennis balls that retain their bounce longer, and stiffer, stronger structural plastics for cars andairplanes. Yet the development of these advanced new materials has been very slow due to lack of integrated information, both experimental data and simulation tools. Currently, understanding the state of the field requires manual searching of online journal databases, full reading of the articles, and manual accumulation and then synthesis of the information. No resources yet exist for assembling the data, nor do tools exist to allow assembled data to be effectively mined for correlations, much less to enable rapid design of new materials. In this research, we will develop a data resource (NanoMine) for housing and exchange of polymer nanocomposite data and development of new data mining methods to enable discovery of underlying material physics. We will integrate of simulation tools to augment experimental data and enable exploration of design concepts. The creation ofan open-source, freely accessible data resource will provide a fast and easy source of information, and will enable both fundamental new understanding of materials as well as much more efficient material design. The data driven approach applied to this one system of nanocomposites, will provide strategies that can be extended to other material systems, greatly extending its influence. We will also integrate research and education in several ways from K-12 through to practicing engineers.