The Dutch government’s National Growth Fund aims to ensure that the prosperity of the Netherlands is also guaranteed for the future. TU/e plays a key role in this ambition, as demonstrated when the second round of subsidies were awarded just before Easter. No fewer than eight projects involving researchers from our university were rewarded. In a series of articles we highlight the four projects with our greatest contribution. Today we focus is on Robotlab, a laboratory in which researchers use artificial intelligence to, among other things, unravel the complexity of chemical mixtures, such as paints and vaccines.
You may not expect it, but artificial intelligence (AI) is playing an increasingly important role in chemistry and materials science. Researchers are using smart algorithms to develop sustainable materials and energy-efficient solutions for data storage and data processing.
AI can also help to better understand, and even recreate, the incredibly complex molecular systems in our bodies. To boost this area of research, known as Big Chemistry, researchers from TU/e, Radboud University in Nijmegen and the University of Groningen, among others, want to build a fully automated chemical laboratory.
The origin of life
In this Robotlab, scientists will use AI to formulate new complex molecular systems, says Bert Meijer, founder of the Eindhoven Institute for Complex Molecular Systems and involved in the project on behalf of TU/e.
"We're going to do this by conducting extremely large numbers of experiments, and using the data to train algorithms. This will make them so smart that they understand chemically complex systems better than humans, which will in turn yield new designs."
An important part of the program involves unraveling the complexity of chemical mixtures, such as paints, adhesives, beverages, vaccines and cosmetics. Finding the right composition for these 'formulations' is now often based on trial and error, but this new approach makes it not only much more efficient, but also more sustainable and environmentally friendly, says Meijer. This can be done, for example, by integrating the possibilities of recycling into in the design of the new mixture. Many companies are interested in this approach.
Scientists can also use the Robolab to explore the origin of life. "Just imagine," says Willem Huck, fellow researcher at Radboud University Nijmegen. "Automation and algorithms will allow us to perform millions of years of evolution in an accelerated way in a laboratory."
In this video Willem Huck explains the purpose of the Robotlab (source: Radboud University Nijmegen)
The researchers also want to build molecular computers inspired by how our brains work. These computers would be able to think just like our brain itself.
Another possible application of the Robotlab are biomaterials that can adapt perfectly to the body. Since millions of combinations are possible, the Robotlab's AI allows it to find exactly the combinations that work best in the body.
The National Growth Fund supports the Robotlab with 96.9 million euro. Together with contributions from knowledge institutions and companies, the total investment amounts to 207.5 million euro. The program is, in a sense, a continuation of the Dutch Research Council Gravity program Functional Molecular Systems and will last seven years.
Meijer is very pleased with the support from the Growth Fund. "This grant comes at exactly the right time. The Netherlands is already a global leader in scientific research into complex molecular systems. The money from the Growth Fund will help us to convert this top position into concrete applications using AI with sustainability as the main theme."
TU/e is collaborating in this project with Radboud University Nijmegen, Fontys University of Applied Sciences Eindhoven, the AMOLF Institute in Amsterdam, the University of Groningen and companies from Brainport Campus Eindhoven and NovioTech Campus in Nijmegen, among others.
More information on complex molecular systems research can be found on the Institute for Complex Molecular Systems website.
This article was first published on 29 April by TU/e.