Management of reference data in materials science and engineering exemplified for creep data of a single-crystalline Ni-based superalloy

Abstract

The identification of process-structure-property relationships of materials inevitably requires the combination of research data from different measurements. Therefore, the concepts related to FAIR (findable, accessible, interoperable, reusable) data handling, increasingly reported in literature, are particularly important in the materials science and engineering domain. However, they have not yet been integrated into a single, overarching methodological framework, particularly for reference data. Here, we introduce such a framework. Our concept covers data generation, documentation, handling, storage, sharing, data search and discovery, retrieval, and usage. Furthermore, we prototypically implement it using a real dataset with creep data of a single-crystal CMSX-6 Ni-based superalloy. The presented implementation is traceable and permanently accessible through open repositories. The individual elements considered in the framework ensure the functionality and usability of the data and, thus, the adherence to the FAIR principles. In conjunction with this, we present a definition for reference data of materials. Our definition underlines particularly the importance of a comprehensive documentation, e.g., on material provenance, data processing procedures, and the software and hardware used, including software-specific input parameters, as these details enable data users or independent parties to assess the quality of the datasets and to reuse and reproduce the results. Reference data that is managed according to the proposed framework can be used to advance knowledge in the materials science and engineering domain, e.g., by identifying new process-structure-property relations.