Building a knowledge graph framework for metamaterials
The field of materials engineering has entered a new era of design and fabrication, in which physics-based modeling and machine learning techniques enable researchers to investigate new design spaces and develop materials with unprecedented mechanical properties. However, in the realm of polymer nanocomposites and structured materials, the data required to enable this design approach is not readily accessible in an annotated or reusable format.
In order to provide this missing functionality, a team of investigators from four universities (Duke, Northwestern, RPI and Caltech) has been developing open-source prototype data resources for nano-composite materials (NanoMine) and structured materials (MetaMine). These web-based databases contain a nascent set of centralized data, a data schema for additional curation, and visualization and analysis tools to enable informed design and scientific discovery. NanoMine has already undergone initial beta testing by a portion of the user community, using curated data from over 1200 samples. These tests have identified critical user and robustness improvements for NanoMine, readying it to be generalized as a materials data framework. A key component of the work entails demonstration of the framework's extensibility to a different material domain: structured materials (aka, metamaterials).
This will be demonstrated through the development of MetaMine, a public data resource that will extend the schema and ontology of NanoMine for metamaterials, with the addition of metamaterial-focused topological descriptor tools, tiling tools, and design tools. The Schmidt Academy is collaborating with the Daraio group on the MetaMine project to develop the site's platform and infrastructure, and on integrating additional design tools and data.