Macromolecules Database Creation for Polymer Properties Roselyne 1

Macromolecules Database Creation for Polymer Properties Roselyne 1, 3 Tchoua , Mladen 2 Rasic , Kyle 3 Chard , Jian 2 Qin , Debra 4, Audus Juan de 2 Pablo , Ian 1, 3 Foster {1 Department of Computer Science, 2 The Institute for Molecular Engineering, 3 The Computation Institute, University of Chicago and Argonne National Laboratory}, 4 The National Institute of Standards and Technology DESIGN GOALS Create a database of Flory-Huggins (χ) values extracted from thousands of publications available in peer-reviewed journals. Provide online access to miscibility information of polymers: χ values, relevant tables, figures and equation from publications. Combine literature mining results with modeling capabilities to facilitate the materials design process for molecular engineers. Create an environment of sophisticated predictive models for materials design. Database Creation Predictive Modeling PROBLEM There is a wealth of important materials properties locked within scientific publications. This information is unstructured and not machine accessible. One example is the Flory-Huggins interaction parameter χ which gives a measure of the interaction of polymer chains with solvent molecules and polymer-polymer interactions. This parameter is used by molecular engineers during their material design process. METHOD: • Crawl journals and classify those that contain Flory-Huggins and other related keywords: χ , ODT, miscibility, scattering function, etc. • Automate knowledge extraction from text, equations, tables and figures. • Leverage knowledge from crowdsourced experts to aid classification and knowledge extraction. • Create a curated database for a range of polymer properties. Sub-5 nm Domains in Ordered Poly(cyclohexylethylene)-block-poly(methyl methacrylate) Block Polymers for Lithography [2] Equations A variety of computational tools to facilitate sophisticated materials design, including but not limited to: Input Parameters Flory-Huggins Phase Diagram Voorn-Overbeek Phase Diagram Lattice Cluster Theory Phase Diagram RPA Structure Factor Simplified SAFT Eo. S Critical Point calculation ______________________________________________________ Flory-Huggins Parameters Lattice Cluster Theory Parameters RPA Structure Factor Parameters Output _______________________________________________________ Paper Metadata Figures Tables Generate Temperature. Dependent Phase Diagram Query the Database for χ Data and system parameters… RESEARCH PLANS for 2015 • • • Explore input from reviewers in order to automate extraction of χ-relevant text, tables, equations and figures. Integrate the extraction of chi parameter data into phase diagram software Implement additional tools to further characterize material systems. Explore the generalization of this system to more journals and more parameters Optimize software for faster and more robust numerical calculations. CHi. Mad is supported by U. S. Department of Commerce, National Institute of Standards and Technology under financial assistance award number 70 NANB 14 H 012.