My contribution
I developed and implemented EnsembleMTD, an ensemble RMSD-driven metadynamics workflow for electrolyte-reduction pathway discovery. I performed GFN2-xTB simulations and targeted DFT calculations, analyzed species populations and reaction networks, and prepared the first-authored manuscript with supervision from Andreas Heuer and Diddo Diddens.
The Python workflow connects enhanced-sampling trajectories to automated reaction-network analysis. Independent runs are aggregated into ranked product candidates rather than treating one observed event as the whole reaction mechanism.
Workflow decisions
The workflow separates trajectory generation, species and reaction analysis, ensemble aggregation, and targeted electronic-structure checks. A systematic parameter screen informed reusable default settings.
That separation makes the result inspectable: a candidate pathway can be traced back to its contributing trajectories and analysis choices. The public repository provides the implementation and run instructions.
Status and limits
The implementation is public under the MIT license. The associated manuscript, “Ensemble RMSD-Driven Metadynamics for Electrolyte-Reduction Pathway Discovery,” is unpublished. Local documents differ on its submission stage, so this portfolio does not assign a more specific status.
Discovered pathways depend on the model chemistry and sampling choices. They are mechanistic candidates that require interpretation and targeted validation, not an exhaustive catalog of all electrolyte reactions.