Berlin 2018 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

MM: Fachverband Metall- und Materialphysik

MM 51: Topical Session (Symposium MM): Big Data in Materials Science - Managing and exploiting the raw material of the 21st century

MM 51.1: Talk

Thursday, March 15, 2018, 10:15–10:30, H 0107

Towards a Roadmap of Chemical Space — •Johannes Margraf and Karsten Reuter — Technische Universität München

Data-driven research in materials science and chemistry is typically focused on the properties of individual compounds. If the value of some property is known for a representative part of chemical space, the remaining values can be estimated by interpolation. In this context, it is essential to understand what compounds make up the chemical space of interest. This can be trivial to define (e.g. all binary octet semiconductors) or require significant effort (e.g. all organic molecules up to a certain size). With this contribution, we want to shift the focus from individual compounds to chemical reactions. This means that chemical space is transformed from a list of compounds to a network of reactions connecting them (i.e. a map).

To this end, we will show how the space of chemical reactions can be constructed and organized in a systematic and exhaustive manner. We will further discuss how these reaction networks can be applied in the prediction of reaction mechanisms. Finally, we will compare how machine learning of reaction energies performs in reaction and compound space.