Speaker
Description
The recent discovery of rapid and high-quality synthesis of Graphene (GR) at liquid Cu catalysts was an important finding towards large-scale commercialization. At present though, there is little mechanistic understanding of what precisely leads to the improved Gr quality as compared to the more traditional synthesis on solid Cu catalysts, which typically employs highly comparable growth conditions just at ~100 K lower temperature. The absence of extended defects, generally higher mobility of atoms, clusters, and flakes, or an increased carbon dissolution at the liquid Cu surface are but a few possible aspects that immediately come to mind.
Compared to its solid state, the characterization of Cu’s catalytic properties in its liquid state is a major challenge both experimentally and computationally. For the latter, the main obstacles are the large length and time scales which are necessary to reliably emulate the temporal evolution of the liquid. Corresponding molecular dynamics simulations require large simulation cells and need to span into the nanosecond timescale – an endeavor presently intractable via first-principles methods. Owed to the complexity of the catalytic process, the predictive quality of first-principles methods is, however, a prerequisite. In this work we therefore use computationally efficient machine-learning interatomic potentials (MLIPs) trained to density-functional theory data in order to extrapolate first-principles predictability to the required scales.
Detailed benchmarking confirms that the MLIP captures the involved physics well, accurately simulating X-ray reflectivity determined density profiles and graphene adsorption heights as well as other observables of Gr and liquid Cu. On basis of this data we compare the Gr adsorption on liquid Cu to solid Cu in strain-free, large-scale simulation cells to find a surprising similarity in adsorption strength and geometry. This suggests that no electronic effects but only the absence of irregularities on the liquid surface lead to high quality Gr on the liquid Cu surface. We apply the MLIP to further study the catalytic mechanism of graphene synthesis. Exploiting the computational efficiency of our MLIPs, we perform metadynamics simulations to obtain free energy barriers of possible rate determining steps which can be compared to the distinct reaction kinetics found experimentally. By comparing to the apparent activation energy of Gr formation we identify carbon attachment as the rate limit step of the catalytic Gr growth process on liquid Cu. This is in stark contrast to Gr growth on solid Cu, where precursor activation was previously found to be the rate determining step. Methodologically, our work draws a path for the use of accurately trained MLIPs as a multiscale modeling technique to explore previously unchartered computational problems. In that we provide new insight into the domain of liquid metal catalysts which are generally still poorly understood at the atomic scale.
| Abstract Number (department-wise) | TH 10 |
|---|---|
| Department | TH (Reuter) |
