Modeling hydrodesulfurization catalysis (HDS) by solid-state synthesis, homogeneous desulfurization and heterogeneous probe reactions.

Abstract

Three approaches towards modeling the HDS catalyst are discussed: structural investigations and catalytic measurements of intercalation compounds of $\rm MoS\sb2,$ reactivity studies of soluble bimetallic clusters, and HDS of probe molecules over heterogeneous catalysts. In order to test the viability of the pseudo-intercalation model for the structure of the HDS active site, intercalation compounds of $\rm MoS\sb2$ were synthesized. Different Co species and Co loadings were introduced into the $\rm MoS\sb2$ host by flocculation of aqueous suspensions of single-layer $\rm MoS\sb2.$ The structures of the intercalation compounds were determined by EXAFS, electron diffraction, and powder X-ray diffraction. The $\rm MoS\sb2$ layers in these materials are distorted 1T-type structures consisting of linked trimolybdenum clusters. After using these materials as catalysts for the HDS of thiophene, the initial structures of the intercalates were found to be unstable: the sulfide phases 2H-$\rm MoS\sb2$ and $\rm Co\sb9S\sb8$ were the only phases detected by EXAFS. These results strongly suggest that the pseudo-intercalation model would not be thermodynamically nor kinetically stable under the conditions of HDS catalysis. The reactivity of the butterfly cluster $\rm Cp\sb2MO\sb2CO\sb2S\sb3(CO)\sb4$ (7b, $\rm Cp = \eta\sp5$-$\rm C\sb5H\sb5)$ for the desulfurization of thiols was investigated as a homogeneous model for the HDS mechanism. Both cyclopropylmethyl thiol (CPMT) and 5-hexenethiol (HET) were stoichiometrically desulfurized at $\rm 110\sp\circ C$ by cluster 7b to form the cubane cluster, $\rm CP\sb2MO\sb2Co\sb2S\sb4(CO)\sb2,$ and rearranged organic products expected of radical clocks (1-butene and methylcyclopentane, respectively). In neat HET, the reaction with cluster 7a resulted in a mixture of rearranged and unrearranged (1-hexene) desulfurization products which allowed the calculation of the rate of hydrogen abstraction $\rm (k\sb{abs} = 3.7\times 10\sp6\ M\sp{-1}s\sp{-1}).$ Thus the reagent thiol was identified as the hydrogen atom donor. The free radical mechanistic model was tested by reacting these same thiols over a conventionally prepared heterogeneous HDS catalyst. The initial product from the HDS of CPMT was 1-butene, while the HDS of HET formed a mixture of hexenes. In this case, the rearranged product (methylcyclopentane) was not formed. A free radical mechanism is proposed with the rate of hydrogen abstraction faster in the heterogeneous system than in the homogeneous one.