Data-Driven Materials Science - How NOMAD is rethinking the pursuit of understanding
Concrete, industrially relevant examples were presented, including methods to describe and predict 2D topological insulators, classify metals/insulators classification, model catalytic CO2 activation, and more. Learn more...
NOMAD - modelling catalytic CO2 activation
The current Big Data revolution presents companies operating in every technology sector with new ways of developing ideas, products and services. New developments are, ultimately, what allow any company to stay competitive and flourish in the long term, regardless of their current positioning. Any business that involves materials can benefit from joining the Big Data era using the data and tools found in the NOMAD Laboratory CoE.
European, publicly funded academic research has been producing world-leading results in the Computational Materials Modelling field for many decades. While almost all of this research is publicly available in peer-reviewed journals and most was never IP-protected, the materials-related data underlying these published results have not been easy to access. Storing and re-using extremely large amounts of data used to be difficult due to high storage costs and low-access bandwidth but these storage and bandwidth problems have become progressively tractable with technological advances.
However, the bulk of materials-related data still remains practically inaccessible for two main reasons:
NOMAD Laboratory at a Glance
More information is also available under Outreach.
The current decade has seen a distinct effort by the European Commission and National Research Councils across Europe to promote Open Access to results and data from publicly-funded research. Open access to research data will immensely benefit researchers, but also industry. Within the European Materials Modelling Community, there is now much greater willingness than ever before to make data easily available for re-use, and to develop tools to make our data truly useful to others. There are various reasons for this.
Specific funding to enable an open access practice within individual EU higher education research institutions is also now widely available as a result. It is also widely recognised that data uploading should be further rewarded, by promoting appropriate citation routes. These could be designed to encourage information sharing by enabling the obvious career benefits that aggressive citation of high-quality reference data can attract. Traceability of data through automatic connections with available scientific papers could also enhance quality control (e.g., through the peer review practice already in place for the scientific papers) at the same time helping any end-users who identify their target data through standard scientific literature search tools. These benefits are far-reaching, in spite of overselling of modelling and poor validation practices having been problems in earlier industrial exposure to simulations.
Open data and software
NOMAD guarantees to keep scientific data in the NOMAD Repository for at least 10 years for free. Data is available under Creative Commons Attribution 3.0 License/CC-BY.
Software developed by NOMAD will be licensed under Apache 2.0.
Advantages for Industry
Most industries are now aware of the far-reaching potential benefits of computational materials-science. Including modelling in industrial research can help “prune” the space of resource-intensive R&D experimenting and testing on the basis of appropriate materials screening information. Identifying existing materials with a desired set of properties is most likely to be helped by browsing all the existing data on candidate materials. Crucially, the set of desired properties may have been already computed for a vast, but not exhaustive list of material compounds.
Inferring the same property for compounds composed of a much wider combinatorial space of single ingredient elements becomes, in this scenario highly desirable: and crucially it can help identify, at virtually zero cost, suitable new candidate materials for the application at hand. Any such screening of candidate materials is, however, only good if the inferred estimate of the new material’s performance is provided with a tightly controlled error bar. Appropriate inference tools should in other words ideally allow industrial R&D users to interrogate the data and be advised, in each case, on how confidently the performance of the new material can be predicted on the basis of the whole available information (estimated in millions of calculation data entries in the NOMAD Repository).
You can learn more about our interactions with industry here.
contact concerning general aspects of the CoE: Kylie O'Brien