Density Functional Theory Investigation of Hydrodesulfurization Catalysis

 

 

Robert J. Nielsen and William A. Goddard, III

 

Materials Simulation Center, California Institute of Technology,

139-74, 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 current hydrodesulfurization (HDS) catalysts. Model clusters of ruthenium sulfide and molybdenum sulfide, two of the most active HDS catalysts, are being developed to describe the electronic environment of possible active sites. Our hope is that understanding the interactions of the sulfides, promoter atoms, organic species and inhibitors will clarify the step-by-step pathway by which desulfurization proceeds and identify the salient features required for high HDS activity so that the most effective catalyst may be designed.

 

 

THE MOTIVATION behind studying HDS catalysis is the ever-tightening restriction on the amount of sulfur permitted in gas oils. To meet the decreasing limit, it is becoming necessary for hydrotreating processes to remove not only those compounds which are readily destroyed with current technology, but also the small complement of sulfur-containing species which now survive. These refractory compounds, largely substituted dibenzothiophenes (Fig. 1), must be removed under conditions benign to the oil product.

 

Figure 1: 4,6-Dimethyldibenzothiophene

 

MOLYBDENUM SULFIDE, promoted by nickel or cobalt and supported on alumina, has been industry's choice in HDS catalysis for decades. General networks have been suggested, such as Fig. 2, implying indirect pathways involving hydrogenation and rearrangement of alkyl substituents are important in addition to direct desulfurization when treating certain refractory compounds. Microscopic models of where the active catalytic sites may be located have also been proposed. These often involve promoter atoms at faces and edges of MoS2 crystallites (Fig. 3.) However, we lack a step-by-step mechanism including the energetics of the relevant intermediates. The inhibition of HDS by H2 and H2S under various conditions is also not clear.

     

Figure 2: Sample of HDS reaction network1         Figure 3: Schematic of catalyst microstructure1

           Pecoraro and Chianelli have found the activity for HDS of dibenzothiophene by transition metal sulfides peaks with ruthenium (Fig. 4), and quantum chemical calculations were used by Chianelli and Harris to suggest electronic characteristics of metal-sulfur bonds that correlate to high activity.

 

 

Figure 4: Conversion of dibenzothiophene over transition metal sulfides at 400 C2

 

THE GOALS of this ab initio investigation of model clusters and their interactions with relevant organic species are clarification of a detailed, plausible mechanism for HDS over the current catalysts, and identification of the key electronic properties required in an active HDS catalyst so that the cheapest and most effective material might be suggested. First, model clusters will be chosen for studying the metal-sulfur bonds in important sulfides.

 

 

Figure 5: Octahedral RuS6 cluster

 

INITIAL CALCULATIONS involved surrounding [RuS6]8- octahedrons (Fig. 5) with an effective dielectric medium to simulate the cluster's natural environment. Unfortunately this method has failed to reproduce the known crystal dimensions. We are now investigating whether larger clusters (Fig. 6) which capture the correct geometry without the use of an external field also display the orbital configurations present in the real systems. These orbital descriptions will be used to locate trends in metal-sulfur bonding across active and inactive metal sulfides.

 

 

Figure 6: Three-ruthenium, four-sulfur cluster and highest singly occupied molecular orbital

 

Acknowledgements

 

NSF Chemistry

References

1Whitehurst, D.D., Isoda, T., and Mochida, I., Adv. Catal. 42, 345 (1998).

2Pecoraro, T. A., and Chianelli, R.R., J. Catal. 67, 430 (181).

3Chianelli, R.R., Daage, M., and Ledoux, M.J., Adv. Catal. 40, 177 (1994).