HDS Reactions

The comprehensive review on deep HDS published by Whitehurst et provides detailed accounts of the HDS mechanism of a wide range of the model S-containing compounds. This review deserves attention in spite of the fact that it deals predominantly with the y-Al203-supported hydroprocessing catalysts. 

all aspects that are relevant for deep HDS are covered and discussed in great detail. 

Note that the direct removal of a sulfur atom from dibenzothiophene requires 2 molecules of hydrogen, while the indirect removal of sulfur from a hindered dibenzothiophene requires 5 molecules of hydrogen per sulfur atom. This is important when estimating hydrogen requirements for deep- 

---

Although the precise mechanism of the HDS reaction is still under debate, we deliberately chose this scheme because it illustrates the kinetics of processes involving two kinds of sites. Consequently, two site balances exist ... 

This paper describes the successful incorporation of molybdenum and molybdenum-nickel clusters into zeolites with 12-membered ring by aqueous ion exchange and application of the resulting materials to HDS reaction of benzothiophene. Stoichiometry of the ion exchange was examined by elemental analysis. UV-visible spectroscopy and EXAFS measurements were carried out to investigate the structure of molybdenum species loaded on zeolites. 

Study on applicability of MoNi/zeolite catalysts to the HDS reactions of other sulfur compounds is ongoing in our laboratory. 

The major contributions of the Co sites in CoSx-MoSx/NaY to the HYD and HDS reactions were corroborated by a FTIR study of NO adsorption. Figure 7 shows the IR spectra of NO adsorbed on CoSx/NaY (2.1Co/SC) and CoSx-MoSx/NaY (2.1Co 2.1Mo/SC). Nitric oxide... 

CoSx-MoSx/NaY exhibited doublet bands at 1867 and 1807 cm, accompanying a weak shoulder peak at ca. 1880 cm. These signals are apparently assigned to those of NO molecules adsorbed on Co sulfides. No peaks ascribable to e NO adsorption on Mo sulfide sites were detected at all. What is important in Fig.7 is that in CoSx-MoSx/NaY, coordinative unsaturation sites are present only on the Co sites in spite of the coexistence of the same amount of Mo sulfide species in the zeolite cavities. These results clearly support that the Co sites in CoSx-MoSx/NaY play major roles in the HYD and HDS reactions. 

A New Improved Synthesis of Tricycle Thienobenzazepines Apphcation of chemistry recently developed by Knochel" combined with the well-described halogen dance (HD) reaction, allowed preparation of our key intermediate A in only three synthetic transformations (Scheme 6.4). In this respect, treatment of 2-bromo-5-methylthiophene with hthium diisopropylamide followed by dimethylformamide afforded aldehyde 11 in good yield, lodo-magnesium exchange with conunercial 4-iodo-3-nitro anisole followed by reaction with 11 afforded the thiophene catbinol 12. Dehydroxylation of 12 provided our key intermediate A which presented the requisite functionality to examine our approach to the construction of the seven-member ring system. 

Fig. 18. The experimental total differential cross-section for the F + HD reaction over the full range of collision energies.

One property of the F + HD reaction which is particularly unique is the nearly complete absence of direct reaction pathway at energies below about Ec = 1 kcal/mol.26,27,31 At these low energies the reaction, and all of its observable characteristics, is mediated through a reactive resonance. The total DCS presented so far is a highly averaged quantity, the actual data obtained from the Doppler-selected TOF measurement is however the state-to-state DCS. To illustrate the effect of reactive resonance at the state-to-state level of details, let us focus on the low energy reaction. Figure 20... 

Fig. 3. The normalized excitation functions in A2 versus collision energy for the two isotopic channels for the F+HD reaction. The solid line is the result of quantum scattering theory using the SW-PES. The QCT simulations from Ref. 71 are plotted for comparison. The experiment, shown with points, is normalized to theory by a single scaling factor for both channels. Also shown in (a) is the theoretical decomposition of the excitation function into direct and resonant contributions using the J-shifting procedure.

In summary, the H + HD reaction shows little sign of resonance scattering in the ICS. Furthermore, the product distributions without angle resolution show no unusual behavior as functions of energy that might indicate resonance behavior. On the other hand, the forward peaking in the angular product distribution does appear to reveal resonance structure. Since time-delay analysis is at present not possible in a molecular beam experiment, it is the combination of a sharp forward peak with the unusual... 

The 0(1D) + HD reaction was also studied using the same Rydberg tagging TOF technique described above. The experiment was carried out at a collision energy of 1.7kcal/mol, which is below the 1.8 kcal/mol barrier for the abstraction mechanism inferred from previous experimental studies.63 The purpose of this experimental study is to investigate the isotope effect on this reaction. 

The mechanism proposed involves hydrogenation of the C2 C3 double bond, formation of 2-vinylthiophenol by an E2 elimination, and hydrocarbon elimination by homolysis of the S—Caryi bond. This pathway rationalizes the primary formation of (104) observed in some HDS reactions of (102) over Co/Mo/S catalysts, as well as the kinetic evidence that the rate-determining step on real catalysts is the removal of surface sulfur.158-160... 

For example, a simple and possibly too naive picture of the HDS reaction would be that a sulfur vacancy at the edge of an MoS2 slab is the active site. A sulfur-containing hydrocarbon such as thiophene adsorbs with its sulfur atom towards the exposed molybdenum. Next, the molecule becomes hydrogenated, the two C-S bonds in thiophene break, the sulfur-free hydrocarbon desorbs and the catalytic site is regenerated by the removal of sulfur by hydrogen. We refer to Prins et al. [50] for a discussion of reaction mechanisms. 

Figure 9.19 In situ Mossbauer emission spectra of 57Co in (left) a series of sulfided Co-Mo/A1203 catalysts and (right) MoS2 particles doped with different amounts of cobalt, corresponding to Co/Mo ratios of a) about 3 parts per million, b) 0.05 and c) 0.25. The Co-Mo-S phase, active in the HDS reaction, has a spectrum unlike that of any bulk cobalt sulfide and is most clearly observed in the spectra of Co-Mo/Al203 catalysts of low Co content, and in the MoS2 particles doped with ppms of cobalt (from Wivel et al. [70] and Topspe et al. [71]).

Figure 9.20 Correlation between the activity of a series of Co-Mo/AI203 catalysts for the HDS reaction, expressed in the reaction rate constant /cT, and the cobalt phases observed in Mossbauer spectra (left) as well as the NO adsorption sites probed with infrared spectra of adsorbed NO (right) (left figure from Wivel et al. [70], right figure adapted from [49] and [74]).

Irrespective of the exact configuration around the promoter atom, we have a detailed picture of the Co-Mo-S phase on the atomic scale. Figure 9.23 summarizes schematically what a working Co-Mo/A1203 hydrodesulfurization catalyst looks like. It contains MoS2 particles with dimensions of a few nanometers, decorated with cobalt to form the catalytically highly active Co-Mo-S phase. It also contains cobalt ions firmly bound to the lattice of the alumina support, and it may contain crystallites of the stable bulk sulfide Co9S8, which has a low activity for the HDS reaction [49]. 

Therefore, the data indicate that Co-Mo-S can be considered as a M0S2 structure with Co atoms located in edge positions. As discussed below, these Co atoms play a direct role in the catalysis. Furthermore, it is generally accepted that the HDS reaction involves adsorption on sulfur vacancies. The low sulfur coordination number (large coordinative unsaturation) estimated from the Co EXAFS may, in fact, reflect that active sites (vacancies) are associated with the Co atoms. 

---