Simulations based on density functional theory (DFT) are a valuable complementary tool to experiments because they can provide atomistic level insight into the origins of electrocatalytic activity, a domain largely inaccessible by experiments. In practice, the accuracy of such simulations is, however, hindered by both fundamental theoretical issues and the limitations of computational resources. These challenges were critically investigated in this thesis in order to create a trans-ferable and sufficiently comprehensive kinetic model for carbon nanotube (CNT) catalyzed hydrogen evolution (HER), which can quantify the effects of nanotube doping without resorting to direct experimental input. Hydrogen coverage, electrode potential, and solvation were deemed the most crucial variables to include in kinetic HER models. A baseline for the HER reactivity of pristine CNTs was established using the developed model. The Volmer-Heyrovsky mechanism was determined to be the predominant reaction mechanism on CNTs with the latter step being rate limiting. The inability of pristine CNT to efficiently catalyze HER was demonstrated by comparison to theoretical work on platinum. Introducing a substi-tutional nitrogen dopant did not result in the activation of the nearest carbon site, although a descriptor model based on adsorption energies would have suggested otherwise. By contrast, the formation of carbon 5-ring structures led to a notable activity enhancement on open-ended CNTs. The improvement was attributed to ring strain which could indicate that 5-rings found also in, e.g., fullerenes are inherently more active than regular 6-rings. The accuracy of DFT-computed activation and reaction energies were assessed by switching to an alternative theoretical framework where the reaction is described in terms of diabatic electronic states, in analogy to Marcus theory of electron transfer. This approach should mitigate spurious electron delocalization effects often plaguing DFT. Constrained DFT (CDFT) methods were implemented in this thesis to perform such simulations. The results indicated that conventional DFT not only underestimated barriers but also overestimated reaction energies. Adapting this framework more extensively was identified as an appealing future research direction. Prior to this study, the CDFT implementation was validated by reproducing reference results for gas phase charge transfer reactions. In a stringent test of the implementation's efficiency, the Marcus electron transfer parameters of a mixed-valence compound were then evaluated. To that end, a condensed phase CDFT molecular dynamics simulation was performed where the charge-localized reactant and product states were continuously monitored, which had not been attempted previously at the full electronic structure level. Marcus parameters were obtained as trajectory averages of the vertical energy gap and the electronic coupling, yielding values in satisfactory agreement with experiments and underlining the importance of accounting for explicit solvation.
|Translated title of the contribution||Elektronirakenteen vaikutus vedynkehitysreaktion katalyysiin hiilinanoputkilla|
|Publication status||Published - 2018|
|MoE publication type||G5 Doctoral dissertation (article)|
- density functional theory
- hydrogen evolution reaction
- diabatic electronic state
- carbon nanotube