Advancing Research Through a Robust Data Infrastructure

Researchers in the Inorganic Chemistry Department at the Fritz Haber Institute, together with partners in the National Research Data Infrastructure project FAIRmat, have developed a repository for standardized experimental catalysis data. The platform enables reliable data sharing and provides a strong foundation for artificial-intelligence-driven analysis.

Key Aspects:

• New Catalysis App: A repository for standardized experimental catalysis data has been implemented within the NOMAD platform.
• The motivation behind the app: Catalysts are important for the chemical industry and essential for decarbonizing our economy. The application of modern data science methods could accelerate the development of improved and new catalysts. However, published experimental data are often unsuitable or require extensive curation due to inconsistencies arising from a lack of uniform standards.
• Outlook: The Catalysis App enables sharing of standardized data and establishes a strong foundation for further AI-driven analysis and utilization. The authors encourage the community to explore it and give feedback, since its value increases significantly with the volume, diversity, and quality of the data it hosts.

Handling large amounts of research data

Digital research data forms an essential foundation for scientific work. Whether in the laboratory or on the computer, vast amounts of data are generated in science. Modern research institutions therefore need effective concepts for managing their data to ensure the quality, traceability, and long-term reusability of data in accordance with the FAIR principles. FAIR stands for Findability, Accessibility, Interoperability and Reusability of digital assets.

From a science policy perspective, the primary goal of research data management is to promote transparency and traceability in research and to avoid duplication of work through the reuse of research data. Thus, the management of research data is a central component of good scientific practice.

New Catalysis App within the NOMAD platform

The exploration and development of new catalysts is an important area of research in chemistry and materials science. However, the search for new catalysts is often costly and labor-intensive. Modern data science could speed the discovery and optimization of new and improved catalysts. Archiving and provision of experimental research data is therefore of particular interest. The lack of machine-readable experimental data nowadays though impedes data-driven discoveries in catalysis research. Although repositories for catalysis data already exist, they are not yet sufficiently structured and equipped with metadata to make them easily usable by AI.

The open-source, web-based NOMAD platform (https://nomad-lab.eu/) is ideally suited to setting up a repository for standardized experimental catalysis data. Within this platform, the team has created an AI-ready catalysis plugin that will ultimately enable FAIR data in catalysis. The Catalysis App is now ready to use — see https://nomad-lab.eu/prod/v1/gui/search/heterogeneouscatalyst.

How to use the Catalysis App

The user interface of the catalysis plugin in NOMAD has been designed so that information can be retrieved from different perspectives. For example, it is possible to ask which chemical elements or compositions are good catalysts for a desired chemical reaction. Alternatively, it is possible to ask which products are formed from specific starting compounds using which catalyst. In addition, many other parameters can be filtered, such as the synthesis method, the form of the catalyst, such as supported catalyst, shaped body, or thin film, or the reaction conditions. This allows users to search only for high-pressure reactions or only for low-temperature applications.

A key advantage of this app is its data visualization feature. Some plots have been predefined in the app, but users can also design their own graphical representations.

Data can be uploaded both manually and automatically via the application programming interface (API). The authors developed schemas, i.e., data structures, for typical datasets in catalysis. Furthermore, they designed Excel spreadsheet templates that can be directly parsed and allow users to enter their data in a format that is familiar to most researchers.

Example data for frequently studied catalytic reactions are available in the app. The authors would like to thank the numerous catalysis researchers for their support in creating these.

Outlook and call for participation

The new repository supports the exchange of catalysis data and lays a solid groundwork for AI-driven analytics. Yet the value of a platform such as the Catalysis App increases significantly with the volume, diversity and quality of the data it hosts. Hence, the authors encourage the community to explore the tool, test its capabilities and build upon it. They also welcome constructive feedback to help develop the app further and ensure that it continues to meet the evolving needs of researchers.

More about NOMAD

“NOMAD” stands for “Novel Materials Discovery” and is a project of research data management. It offers a free, open-source web interface to manage and share data from materials science. The project began in 2014 with the NOMAD Repository and was founded by Claudia Draxl from HU Berlin and Matthias Scheffler, former director of the Theory Department of FHI. Since then, the project has grown steadily and is now being hosted and further developed as part of the FAIRmat project, which is the German consortium for research data management (Nationale Forschungsdaten-Infrastruktur, NFDI) in the fields of condensed matter and the chemical physics of solids. https://nomad-lab.eu/nomad-lab/nomad-lab.html