Skip to main content

Computation-driven rational design of MoSx-based desulphurization nanocatalysts

Final Report Summary - NANO-DESIGN (Computation-driven rational design of MoSx-based desulphurization nanocatalysts)

The combustion of Sulphur-containing compounds that are naturally present in notable quantities in biofuels, petroleum and coal leads to the formation of SO2, which has been recognised as a noxious contributor to “acid rain” and air pollution affecting populated and industrialized areas contributing to aerosol gases harmful to human health. At present, despite stringent regulated emission standards, hydrodesulphurization (HDS) catalysts are efficient but unable to remove the last few ppm of Sulphur from fuels. The purpose of this project was to rationally design a new generation of nanostructured molybdenum sulphide (MoS2 or MoSx) heterogeneous catalysts by means of computational modelling and lead the way, in close collaboration with experimental and industrial partners, towards the engineering and use of real world HDS catalysts. The Nano-DeSign project has aimed at constructing and optimizing novel computational models that are truly representative of MoS2-based model catalysts by means of state of the art computational methodology based on the Density Functional Theory (DFT) and recently developed optimization algorithms. The catalytic activity of such models towards desulfurization has been studied, establishing structure to reactivity correlations that could guide the experimental and industrial development of nanostructured materials with improved catalytic performance.

The Nano-DeSign project had three main objectives, which were fulfilled and/or adapted during the duration of the project:

a) Construct and optimize computational models that are representative of MoS2-based model catalysts:

The structure of different computational models for Au(111)-supported MoS2 nanoparticles has been optimized by means of DFT calculations. Systems featuring different shape, size, and composition have been considered. Simulated STM images have been generated for all the optimized models and core-electron binding energies have been calculated for a selected set of representative nanoparticle models. These data has been used to elucidate and rationalize experimental X-Ray Photoemission Spectroscopy (XPS) and STM data of model catalytic systems. The effect of the interaction with the Au substrate on the properties of MoS2 nanoparticles has been studied by performing calculations also for the corresponding free-standing MoS2 nanoparticles.

The properties of MoS2-based catalysts are determined by the size, shape, and composition of their nanostructured components. The structural optimization of a large number of Au(111)-supported MoS2 nanoparticle models has allowed determining the most stable species for different given conditions (S-coverage and particle size). We have found that the desulphurization of small MoS2 nanoparticles leads to the stabilization of peculiar phases of MoS2 with significantly different properties, which will potentially lead to the design of novel materials. In addition, the Au(111) support is found to play a significant role in the stability and properties of the different structures. There is a significant amount of charge transfer from the Au to the MoS2 NP and for significantly reduced nanoparticles, the corners of the nanoparticles bind strongly to the support.

The characterization of the active sites, which for MoS2-based catalysts are found at the egdes of the layered nanoparticles, is crucial to understand and improve heterogeneous catalysts. We have found that such edge sites have a distinct metallic character, which also depends on their sulphur coverage. Our determination of particular Core-Level-Shift fingerprints calculated by means of DFT-based methods has allowed monitoring the presence and evolution of the active edge sites through XPS experiments performed by our collaborators under more realistic conditions.

The electronic structure of extended monolayers of MoS2 and other transition metal dichalcogenides (TMDC) has also been studied by means of periodic DFT calculations. The calculated band structures show a remarkable agreement with those measured by menas of Angle-Resolved Photo-Emission Spectroscopy (ARPES) for monolayers of MoS2, WS2 and TaS2. We have analyzed the effect of the Au-substrate on the band structure of MoS2 in order to rationalize experimental results from ARPRES. Au is seen to significantly modify some of the out-of-plane band structure features of MoS2. These results are relevant to the emerging field of valley- and spin-tronics, for which the understanding of the properties of two-dimensional transition metal dicalchogenides is central.

b) Study the reactivity towards desulfurization of selected nanoparticle models that are representative of model catalysts:

The chemical properties of sites of the MoS2 computational models have been studied by calculation of hydrogen adsorption energies and sulphur vacancy formation energies, which have been proposed as descriptors for HDS and hydrogen evolution reactions (HER). We have found that the chemical properties of the edge (active) sites of the nanoparticles are not uniform. Their reactivity is modulated with respect to the distance to the corner sites of the nanoparticles. This finding highlights the importance of such undercoordinated corner sites in the activity of MoS2-based catalysts and the importance of using explicit nanoparticle models in computational simulations where such catalyst are studied.

c) Assess the recently developed implementation of state-of-the-art computational algorithms aiding their further development:

The genetic algorithm developed in the host group has been tested for the global optimization of Au(111)-supported MoS2 structures and is currently being further developed in order to improve it’s efficiency for relatively large and complex systems. At its current stage, the GA is inefficient in finding the global minima for the particle sizes that are relevant to the project. Improvements in the GA code that make it more efficient for ever more complex systems are expected to benefit the whole computational modeling community.

The accuracy of the method used for predicting core electron binding energies that consider final state effects has been assessed by comparing the calculated shifts with those obtained by high-resolution XPS experiments. We conclude that our implementation of this method, which takes into account final state effects and allows comparing the shifts for specific atomic species within systems of different size and composition, is sufficiently accurate to interpret experimental data qualitatively and quantitatively.

In summary, the Nano-DeSing project has contributed to the fundamental understanding of nanostructured catalysts based on MoS2 and the model systems used to investigate them. The construction of representative computational models has allowed investigating site, structure, and size- dependent properties of supported MoS2 nanoparticles, which should be useful for the rational design of improved HDS catalysts. Furthermore, during the project the fellow Albert Bruix has strongly developed and matured as a researcher, acquiring both transferable and management skills and a broad technical know-how. The experience gained has enhanced the researcher's scientific skills and will be a firm basis for developing an independent career path.