Atomistic Models of Hydrodesulfurization over Co/MoS2


Robert J. Nielsen and William A. Goddard, III

Materials and Process Simulation Center, Beckman Institute (139-74), Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125


The need to meet more stringent standards limiting the sulfur content of gas oils urges a deeper understanding of the mechanism by which sulfur-containing compounds are destroyed over hydrodesulfurization (HDS) catalysts. Thiols and thioethers are readily treated by the Co- and Ni-promoted MoS2 catalysts currently employed, but satisfying the imminent restrictions will require the removal of the most refractory species, mainly alkyl-substituted polyaromatic thiophenes. Unfortunately, the nature of the catalyst's active sites is not well known. We have undertaken density functional calculations on models of pure and Co-promoted MoS2 along with relevant organic substrates to determine (1) the structural and electronic factors indispensable to the "CoMoS" active phase, (2) the mechanism by which thiophene and its derivatives react, and (3) how alkyl substituents promote and inhibit different branches of the reaction network.

Preliminary mechanistic investigations were performed on a model of a possible "corner site" of unpromoted MoS2. To survey the energetic landscape, many surface-bound intermediates in the reaction of thiophene to butadiene were considered. Thus a network of elementary steps was generated, allowing different coordinations of the reactant and different sequences of hydrogenation and hydrolysis. The most likely path through this network centers on the h5 coordination of thiophene. Though no transition states were calculated for these tentative steps, it is apparent that the replacement of the strongly bound sulfur atom initially occupying the site with a thiophene molecule poses the greatest energetic difficulty.

Models of other possible sites ("edge sites", sites bridging two metal atoms, sites incorporating promoter atoms) will spawn other mechanisms which warrant attention. This space of possible models was reduced recently by the STM observation of MoS2 nanocrystals1. To establish whether the specific pattern of reconstruction along the crystallite edges witnessed in the STM studies remains in finely dispersed MoS2 particles, we generated three- and six-metal atom clusters and studied their behavior in reconstruction and sulfidation. Exposing the reconstructed (1010) face is preferred over the (1010) face by the small clusters, as in the STM subjects. We find that for these small models it is less appropriate to discuss a "vacancy formation energy" than an "equilibrium degree of sulfidation," as the binding energies of adsorbates depends on the state of the entire crystallite's edges.