All the data reported previously have been reevaluated to determine the quantitative reaction network for dibenzothiophene hydrodesulfurization at about 100 atm and 300/sup 0/C. This network shows the high selectivity of the catalyst for simple hydrodesulfurization (biphenyl + H/sub 2/S formation). When methyl groups are present in the 4 and 6 positions of dibenzothiophene, the primary hydrogenation reaction is approximately as fast as the primary hydrodesulfurization reaction. Ni and Mo or Ni and W used instead of Co and Mo in the catalyst, or H/sub 2/S in the reactant mixture, also favor hydrogenation relative to hydrodesulfurization. Previous results have been evaluated in light of the recent literature to provide a new interpretation of the reaction mechanism of catalytic hydrodesulfurization on surfaces of sulfided Co-Mo/Al/sub 2/0/sub 3/. It has often been assumed that catalytic hydrodesulfurization of thiophene and related compounds proceeds via a one-point end-on adsorption involving bonding of the sulfur atom with Mo ions at an anion vacancy on the catalyst surface. This interpretation is inadequate, failing to account for deuterium-exchange results, the reactivities of benzothiophene and dibenzothiophene, and the small steric effects of methyl substituents affecting the adsorption and reaction of compounds in the thiophene, benzothiophene, and dibenzothiophene families …

Flow reactor studies of HDS of dibenzothiophene at about 300/sup 0/C and 100 atm have shown a new dependence of catalyst activity on H/sub 2/S concentration at low values. We infer that the catalyst undergoes structural changes during reaction, sometimes becoming sulfur deficient and relatively inactive. It is evidently necessary to maintain the catalyst in a properly sulfided state during reactor startup and operation by avoiding its contacting too little H/sub 2/S. The relative activities of Ni--Mo/Al/sub 2/O/sub 3/, CO--Mo/Al/sub 2/O/sub 3/, and Ni--W/Al/sub 2/O/sub 3/ catalysts for HDN of acridine have been determined, showing that Ni--Mo/Al/sub 2/O/sub 3/ is the most active. The results are interpreted inthe context of the acridine reaction network, in which both the hydrogenation and hydrogenolysis (cracking) reactions are kinetically important.

Kinetics data have been determined for the catalytic hydroprocessing of the acidic fractions of a heavy distillate of a liquid derived from Powhatan No. 5 coal. A commercial, sulfided Ni-Mo/..gamma..-Al/sub 2/O/sub 3/ catalyst was used in the experiments, carried out at 350/sup 0/C and 120 atm with the coal liquid fractions dissolved in cyclohexane. The feed and hydrotreated products were analyzed by gas chromatography/mass spectrometry. The data were analyzed with group-type methods for compound classes, and results were also obtained for some individual organooxygen compounds. Catalytic hydroprocessing leads to a large increase in the number of compounds and a shift to lower boiling ranges. The data are broadly consistent with reaction networks determined with pure compounds; the most important reactions include aromatic ring hydrogenation, hydrodeoxygenation, and hydrodemethylation. Pseudo first-order rate constants for conversion of the predominant organooxygen compounds are on the order of 10/sup -4/ L/(g of catalyst.s); the reactivity decreases in the order cyclohexylphenol > dimethylhydroxyindane > tetrahydronaphthol > phenylphenol > 1-naphthol. 12 references, 15 figures, 5 tables.

A new method of structural analysis is applied to a group of hydroliquefied coal samples. The method uses elemental analysis and NMR data to estimate the concentrations of functional groups in the samples. The samples include oil and asphaltene fractions obtained in a series of hydroliquefaction experiments, and a set of 9 fractions separated from a coal-derived oil. The structural characterization of these samples demonstrates that estimates of functional group concentrations can be used to provide detailed structural profiles of complex mixtures and to obtain limited information about reaction pathways. 11 references, 1 figure, 7 tables.