Enabling Ultra Deep Hydrodesulfurization by Nanoscale Engineering of New Catalysts

The HYDECAT project was initiated to make a targeted effort in the field of hydrodesulfurization (HDS), which is the process where sulfur is removed from crude oil by addition of hydrogen to form hydrocarbons and hydrogen sulfide. This PhD thesis represents my share in the project.
Due to the adverse environmental and societal effects of sulfur emissions from on-road transportation, legislation has been continuously tightened, pushing oil refiners to produce ultra-low sulfur diesel (ULSD), with a maximum sulfur content of 15 ppm. Since these specifications are expected to be further tightened, the existing HDS catalysts fall short. Experiments were performed on a setup dedicated to testing minute amounts of well-defined catalytic systems in the ambient pressure gas phase HDS of the model compounds dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-DMDBT). An existing µ-reactor platform connected to a high resolution time-of-flight mass spectrometer (TOF-MS) was modified and optimized for this specific reaction. The µ-reactor has a reaction volume of only 240 nL and can be operated between 0.1-5 bar and temperatures up to 400 ◦C. Only 0.01 % of the mixed gas flows from the two inlets, O1 and O2, is bypassed through the reaction chamber and exposed to the catalyst. A channel terminated by a narrow capillary ensures that the entire reaction gas volume can be directed into the TOF-MS by probing only 5·1014 molecules/s.
The low vapor pressure of both DBT and 4,6-DMDBT complicated the process of introducing them in their gaseous form into the µ-reactor at ambient pressure, and a specially designed flange with an incorporated ion source and internal heat tracing was implemented. 
HDS of DBT and 4,6-DMDBT at 800 mbar on six mass-selected Pt samples were conducted. Two Pt samples of ∼3 nm (185 kamu) and two samples of ∼6 nm (1500 kamu) all showed that only the direct desulfurization (DDS) pathway was followed, hence resulting in biphenyl (BiPhe) and 3,3'-dimethylbiphenyl (3,3'-
DMBiPhe), respectively. The same was observed for two samples of Pt single atoms. One 1500 ,kamu sample reached full conversion and was used to derive a sensitivity factor, x , relating the DBT and BiPhe signals, since most ionization cross sections were unknown. This was applied in all the following data interpretation. Large deviations between the, in theory, identical samples made it difficult to see any clear trends, and it was estimated that a reaction temperature difference of 30 ◦C could have induced the different activities observed.
Four NiMo-based samples were tested in the HDS of DBT. Two metallic NiMo samples of ∼3.5 nm (134 kamu), and two in-flight sulfided NiMoS_x samples - 
one of ∼5 nm (200 kamu) and one of∼6.5 nm (440 kamu). X-ray photoelectron spectroscopy (XPS) and activity measurements emphasized the need for a sulfidation step prior to the reaction, since exposure to air revealed the formation of an oxide layer. Scanning transmission electron microscopy (STEM) images of the in-flight sulfided NiMoS_x samples showed flat lying platelet-like particles in the 200 kamu sample and upright standing particles in the 440 kamu sample. When normalized to the amount of metal in each sample, the activity of the flat lying particles were exceeded by the activity of the upright standing particles by an order of magnitude, indicating that more active edge sites are exposed in the latter and thereby enabling better HDS activity.