Automated Reaction Modeling and Measurement
The Jakob Dahl Group combines experimental chemistry, physical modeling, and machine learning to understand and predict complex chemical reactions, such as nanocrystal formation and catalytic processes. The aim is to transform these multi-step reactions from systems governed largely by trial and error into predictable and designable chemical processes. [
Research
We develop computer models that combine reaction kinetics with quantitative descriptions of how a molecule's structure influences reactivity. Using approaches adapted from machine learning, these models are trained and validated. Because the models stay physically meaningful, a better fit not only improves predictive accuracy but also provides insights into how a reaction proceeds, including hard-to-observe steps such as catalyst breakdown.
These models guide a self-driving laboratory that performs experiments and selects the most informative to make next. We perform these experiments on in-house automated high-throughput and in-situ synthesis platforms, coupled to UV-VIS and IR spectroscopic readout, with other analysis performed in shared facilities at DTU.
A common thread in our research is the interface between organic and inorganic chemistry, where organic ligands bound to inorganic cores control reactivity. In homogeneous catalysts, we study changes in these ligands make metal catalysts faster and more selective. Transferring this approach to nanomaterials, we investigate how ligands tune the size, shape, and stability of nanocrystals for applications in solar energy, displays, and photocatalysis.
Contact
Jakob Dahl Assistant Professor Department of Chemistry jakda@kemi.dtu.dk