Solvent effects on dibenzothiophene hydrodesulfurization: Differences between reactions in liquid or gas phase

In order to understand the solvent effect on dibenzothiophene (DBT) hydrodesulfurization (HDS), a NiMoS/gamma-Al2O3 catalyst was tested in a trickled bed micro-reactor with model mixtures of gasoline and diesel containing 800, 500 and 300 ppm of S, using dibenzothiophene as the source of S, and heptane (C7), dodecane (C12) and hexadecane (C16) as solvents. Computer simulations using ASPENPLUS software show that the solvent chosen allows DBT HDS to be performed either in gas-solid or liquid-solid interphases. Experiments show that global DBT HDS rates are a function of solvent: -r(C16) < -r(C12) < -r(C7). Rate constant (k) and DBT-solvent adsorption constant ratios (K-DBT/K-S) were calculated in accord to the Langmuir-Hinshelwood model. For heptane, dodecane and hexadecane, a rate constant (k) of 6.2 x 10(-7) mol s (1) g (1) was found, indicating that catalytic sites are independent of the solvent. Therefore, global rate variation could be associated to a competition effect for the catalytic sites between DBT and solvent. Adsorption constant ratios (K-DBT/K-solvent) decrease in the order C7 > C12 > C16. Thus, DBT adsorption is better when the reaction is performed in gas phase using heptane as solvent than that in liquid phase using dodecane or hexadecane. Results indicate that DBT HDS selectivity depends on the solvent used. Thus, selectivity for biphenyl is larger in dodecane or hexadecane as compared to that obtained in heptane. In conclusion, the solvent effect must be taken into account to measure activity of catalysts, especially since this effect is more important under deep hydrodesulfurization conditions required to obtain gasoline and diesel fuels with a S content of 50-10 ppm.