A Hierarchical Model for Polymer Data

Information in polymer literature is often provided through a combination of text, tables, and illustrations dispersed within multiple documents, making it challenging to collate data across sources using automatic tools. To make polymer data more interoperable and reusable, we propose a model for reporting polymer data. In the model, data are grouped according to the synthetic pathway they belong to. Each synthetic pathway is encoded as a graph interconnecting multiple species. Within the graph, polymers are described by a topological-chemical-physical characterization hierarchy. The first two levels establish the molecular structure of a polymer with line notation and chemical characterizations such as tacticity. On top of that, physical characterizations such as rheology are encoded to provide information beyond single molecules. Furthermore, within each level of characterization, data is organized in another hierarchy: at the bottom level lies raw data such as GPC traces, which can be extracted into intermediate data such as the molecular weight distribution and further distilled into attributes such as the dispersity. Together, the data model provides a template that reveals inherent relationships between distinct data entries, thereby supporting data analytics downstream.