Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

CPP: Fachverband Chemische Physik und Polymerphysik

CPP 28: Molecular Electronics and Excited State Properties (joint session CPP/TT)

CPP 28.11: Vortrag

Mittwoch, 29. März 2023, 12:15–12:30, GÖR 226

Machine learning for molecular design of organic molecules and reaction optimization — •Julia Westermayr¹, Reinhard J. Maurer², and Detlev Belder³ — ¹Artificial Intelligence in Theoretical Chemistry Group, Leipzig University, Germany — ²Computational Surface Chemistry Group, University of Warwick, UK — ³Analytical Chemistry, Leipzig University, Germany

High-throughput screening of reaction conditions and electronic properties of molecules plays a crucial role in chemical industry and can be facilitated by automated workflows and machine learning. However, the high combinatorial complexity of the various parameters affecting molecular properties leaves unguided searches in chemical space highly inefficient and the optimization of reactions to synthesize these molecules often fails as theoretical protocols are usually decoupled from experiments. In this talk, we will show how predictive and generative deep learning models can be combined to theoretically design new molecules with potential relevance to organic electronics [1]. Further, we will present how these tools can be coupled with experiments to enable automated micro-laboratories [2] for targeted chemical synthesis. [1] JW et al. Nat. Comp. Sci, in press (2023). [2] R. J. Beulig et al., Lab Chip 17, 1996 (2017).