Abstract
In the move towards a carbon-zero energy economy, molecular H2 offers great promise as a clean energy storage and transport mechanism. The production of H2 in a carbon-neutral and environmentally sustainable manner is therefore a critically important step towards mitigating anthropogenic climate change. Electrocatalytic splitting of water is an ideal route to clean H2 production on an industrial scale, but this requires an effective and affordable catalyst. In this thesis, density functional theory calculations are used to explore two different nanostructured catalysts in this role. The optimisation of molybdenum disulfide (MoS2), which is an Earth-abundant and low-cost material, is explored in depth. Several structural modifications which significantly improve its activity are reported. Following this, a preliminary investigation into nanostructured tantalum nitrides (TaN and Ta3N5) as related binary catalysts is also performed.
First, the mechanism for hydrogen evolution on MoS2 is studied using a detailed model of the electrochemical cell that accounts for both solvent and potential. The effects of two common support materials (graphene and Au(111)) below MoS2 are also considered. It is found that hydrogen evolution proceeds by a Volmer-Heyrovský type process, where H first adsorbs to an S atom on the edge of MoS2, diffuses to an Mo atom, then reacts with another proton from solution to form H2. This dominant mechanism stays the same with MoS2 on either a graphene or Au(111) support. The different supports primarily influence the reaction rate through changing the adsorption energy of H, which determines the reaction thermodynamics. The overall activation barrier is calculated to be about 1.30 eV (at 0 V vs the standard hydrogen electrode) for both supported cases. However, the support is also found to strongly influence how the barrier responds to changes in potential, with the barrier on MoS2/Au(111) dropping much more rapidly as the potential becomes more negative.
Unfortunately, neither the Au(111) nor graphene support notably improve the hydro- gen evolution activity of MoS2. Thus, in a following study, we screen across a wide range of different support materials for the catalyst, all based on doped graphene derivatives. Here, the free energy of H adsorption (∆GHads) is examined, as this was found to capture the majority of change in the reaction rate in response to different supports in the previous chapter. A high degree of tunability of ∆GHads is observed on both the MoS2 basal plane and the edge sites. Critically, it is found that an N-doped graphene support yields ideal thermoneutral (0 eV) H adsorption on the Mo-edge, suggesting this would make a very active catalyst for hydrogen production. For the basal plane, no support is able to bring ∆GHads close to 0 eV with the lowest value being about 1.4 eV. However, the changes observed in ∆GHads are rationalised using density of states analyses, which show that ∆GHads decreases towards 0 eV with the addition of supports that shift the energy of the filled S atom states towards the Fermi level. A different pattern is observed for the MoS2 edge, where ∆GHads is found to increase towards 0 eV when supports that donate more electron density to MoS2 are added. The discovery of these trends assists in understanding the key factors contributing to catalyst activity, and aids in the design of novel catalysts.
To explore further improvements to the activity of the MoS2 basal plane, which makes up a large portion of the material but was not strongly activated by the addition of supports, we also examine both armchair and zigzag nanotube forms of MoS2. Here, the strain induced by coiling significantly lowers ∆GHads , such that smaller (more highly strained) nanotubes adsorb H more strongly. With the induction of S-vacancy defects in the nanotubes, which are known to commonly occur during synthesis, a thermoneutral ∆GHads value is found for nanotubes around 20 Å in diameter. The trends in ∆GHads with nanotube diameter are once again rationalised using density of states arguments, showing that ∆GHads decreases alongside the energy gap between filled S atom p-states and unfilled Mo d-states.
Lastly, an exploratory study into the nanoparticulate structures of TaN and their hydrogen evolution activity is performed. Possible nanoparticle structures are determined using a recently developed machine-learning-assisted global optimisation algorithm. Two compositions, TaN and Ta3N5, and two sizes, 16 and 24 atoms, are studied. For the 1:1 composition, regular Ta core structures are observed with adsorbed N atoms around the nanoparticle surface. In contrast, at the 3:5 composition notably more open structures are found, where N integrates into the centre of the cluster. The nanoparticle structures adsorb H much more strongly than the (100) and (111) surfaces of TaN, and H favours different adsorption sites, suggesting that their hydrogen evolution properties are very different from the bulk. However, future work will need to perform a more detailed study at higher H coverages and considering competing reactions.
The work in this thesis contributes to our understanding of the factors governing nanoscale catalyst activity for the hydrogen evolution reaction. Several synthetically accessible modifications to the MoS2 catalyst that could greatly improve its activity are suggested, representing a step towards affordable and sustainable hydrogen production on an industrial scale.