Robert Strothmann receives Kekulé scholarship
The PhD candidate in the Theory Department will receive a fellowship from the Chemistry Industry Fund starting April 1, 2022. With the help of machine learning, he will try to predict new molecules and optimize molecular design.
The Chemistry Industry Fund (Fonds der Chemischen Industrie) - the sponsorship arm of the German Chemical Industry Association (Verband der Chemischen Industrie) - has been supporting basic research, young scientists and also chemistry teaching in schools since 1950. To get one of its prestigious Kekulé doctoral scholarships, you have to study quickly and with outstanding grades - like Robert Strothmann. The young chemist from Krefeld in North Rhine-Westphalia studied chemistry at Freie Universität Berlin, in close proximity to the Fritz Haber Institute. Despite the Corona pandemic and a stay abroad at the University of Vienna, he quickly finished his studies and took up his doctorate in the Theory Department of the FHI just one month after completing his master's degree at RWTH Aachen University.
Since he had already done research in the field of theoretical chemistry during both his bachelor's and master's theses, it was clear to him early on that he wanted to continue working in this field. With Prof. Karsten Reuter as his supervisor, Robert Strothmann now conducts research in the area of machine learning in the field of chemistry since 2021, with a focus on molecular design and using photoswitch molecules as an example.
"The situation in chemistry is somewhat paradoxical," Strothmann tells us. "On the one hand, we have enormous amounts of data at our disposal. At the same time, the 'chemical space' (i.e., the set of conceivable molecules) is so large that you can usually find very few examples in the literature for a specific application. So the challenge is to extract general patterns and rules from the data in order to apply them to the concrete case."
The basic idea of his PhD is to combine the large amount of known molecules in the form of structural databases with a suitable machine learning model to speed up the prediction of new more efficient molecules. This could be used to produce materials with better properties, for example for technical devices or other industrial products, ideally contributing to a resource-efficient economy. Dr. Johannes Margraf's group, whose team conducts research on data-efficient machine learning in chemistry, is the ideal place for this project. "We are very pleased that Mr. Strothmann has joined us. His project is important for harnessing the applications of artificial intelligence to discover new molecules. And not - as before - only on the computer, but also in the laboratory," says Margraf.
That's because machine learning should not be a purely theoretical endeavor, Strothmann agrees. "Theoreticians and experimentalists must work together. Modern computational methods and machine learning in combination with experimental expertise and high-quality measurement data - that's the silver bullet. That's how we can create a generalizable approach to molecular design."
In this important endeavor, Robert Strothmann gets full support from the Fritz Haber Institute. "It is a great privilege to work here. I am the first person in my family that went to university. To be able to do a PhD at such a renowned institute is a unique feeling."