The Steps to Creating a Materials Informatics Strategy
Blog Jul 29, 2019 | by Ruairi J Mackenzie, Science Writer for Technology Networks
When one thinks of materials science, the core concepts are very tangible and solid - crystalline structures, or nanotubules under an electron microscope. Data isn't very tangible - it's even hard to visualize (except in green Matrix-like streams of binary code), but to quote the Novel Materials Discovery (NOMAD) Laboratory, data is a crucial raw material of the 21st century.
As such, materials companies are waking up to the importance of having an informatics strategy to support their operations. Lux Research are firm proponents of this idea. We caught up with their analyst Xiao Zhong, Ph.D., who recently authored a report explaining the key steps in creating such a materials informatics strategy, to find out more.
Ruairi Mackenzie (RM): What is a materials informatics strategy?
Xiao Zhong (XZ): Materials informatics (MI) applies data science and AI methods to better understand the use, selection, development, and discovery of materials by parties across the value chain, from materials and chemicals developers to product manufacturers. Typical applications include property optimization, design of experiments, and product performance optimization.
In essence, a strategy for a chemicals and materials company here means to accelerate product development (from discovery to market) using MI methods. A strategy also includes data collection, re-organization, standardization, as well as a company’s MI talent recruitment and internal training.
RM: Why is such a strategy useful for materials companies?
XZ: Chemicals and materials companies often have a low profit margin. While there are many reasons behind this, lengthy product development cycle is a major one (if not the most important and fundamental one). A MI strategy, if deployed well, has the potential to shorten the product development cycle by a factor of two. This promises chemicals and materials companies a competitive advantage with a high barrier to entry.
RM: What has been the roadblock for materials companies deploying informatics until now?
XZ: The major roadblock is data. More specifically, companies worry about data-sharing with MI startups and/or their upstream/downstream collaborators, not to mention the fear they have of data-sharing with their peers. It is been interesting for me to observe that many companies have realized the benefit of such data-sharing (to a certain degree) but still choose not to act and set an effective MI strategy (forming a consortium with collaborators, for example). Therefore, I’d comment that besides the data, decision makers’ mindset or state of mind (taking a risky move, such as setting an MI strategy) is another roadblock. Also, the data gathered within a company often do not have the same format or level of fidelity. This further increases the barrier to set and execute an MI strategy.
RM: You’ve said that Europe is at risk of falling behind the US and Japan in their materials innovation strategies. Why have those latter two regions excelled in this area in recent years?
XZ: Europe has some great MI initiatives; the establishment of NOMAD is a good example. It is a center of excellence (CoE) that creates, collects, stores, and sorts computational materials science data. Such a center serves as an important bridge that aims to fill the gaps between academic achievements and industrial applications. On the other hand, Japan and the U.S not only have similar activities (Japan’s MI2I and the U.S.’ Center for Hierarchical Materials Design (CHiMaD) and the Materials Project, for examples), but also are further narrowing the gaps between academic and industry. Japan’s National Institute for Materials Science (NIMS) has facilitated a consortium with four Japanese chemical and materials companies to leverage data and MI. The U.S. Materials Genome Initiative (MGI) has been active in this space as well. These efforts have deeper impacts on the chemicals and materials industry than a CoE, and they are often developed on top of an existing CoE. In that sense, I think both Japan and the U.S. have made more progress on filling the academic-industry gaps using MI.
Xiao Zhong was speaking to Ruairi J Mackenzie, Science Writer for Technology Networks.