Mutual inhibition effect of sulfur compounds in the hydrodesulfurization of thiophene, 2-ethylthiophene and benzothiophene ternary mixture

The hydrodesulphurization of a model component and equimolar ternary mixture of thiophene, 2-ethylthiophene and benzothiophene over sulphided CoMo/γ-Al2O3 and CoMo/γ-Al2O3-Nb2O5 catalysts were investigated in a fixed bed flow reactor. The catalysts were prepared by incipient wetness impregnation method and characterized by textural characteristics, total acidity and chemical species present on the catalysts surface. The characterization results showed that both CoMo/γ-Al2O3-Nb2O5 and sulphided CoMo/γ-Al2O3 catalysts exhibit a mesoporous structure with cylindrical pores open at both ends, evidenced by the IV type adsorption–desorption isotherm with a H1 hysteresis loop and have an average pore diameter between 3 and 4 nm. The chemical species present on the catalysts surface evaluated by XPS indicated that Co2+ and Mo4+ species are present in the sulfide form on both catalysts surfaces. In addition, there are also found oxidic species arising from incomplete reduction and sulphidation. The presence of niobium oxide in the catalytic support had a positive effect in leading to higher specific surface area (170 m2/g) and total acidity (0.421 meq/g) compared with CoMo/γ-Al2O3 catalyst (140 m2/g and 0.283 meq/g) respectively. The evaluation results from the hydrodesulfurization showed that CoMo/γ-Al2O3-Nb2O5 catalyst had a higher activity in hydrodesulphurization process of thiophene, 2-ethylthiophene and benzothiophene. The CoMo/γ-Al2O3-Nb2O5 catalyst eliminated sulfur from the single component feed (corresponding to 2380 ppm S) and reduced below 10 ppm, for the feed consisting in the equimolar ternary mixture of thiophene, 2-ethylthiophene and benzothiophene (2380 ppm S). The reactivity of thiophen compounds was reduced due to competitive adsorption. It was observed that benzothiophene inhibits the transformation of thiophene and 2-ethylthiophene. A mutual inhibition effect between sulfur compounds was also observed when thiophene, 2-ethylthiophene and benzothiophene were combined and tested over the CoMo/γ-Al2O3 catalyst. The inhibition effect had a lower intensity by introducing Nb2O5 in the catalyst support.

www.nature.com/scientificreports/ to 450 °C, and the reduction continued for 6.0 h. After reduction, the catalysts were activated (pre-sulfided) for 8.0 h (250 °C, 5 bar) in a solution of 1% dimethyl-disulphide solution in heptane (flowrate 1 mL/min). Both catalysts have been characterized before and after sulphidation. Textural characteristics of the catalysts (surface area, pore volume, mean pore diameter, pore-size-distribution) were determined by nitrogen adsorption at low temperature, using an Autosorb 1 Quantacrome instrument. The samples were vacuum-dried before the adsorption at 250 °C for 3 h. The total acidity of the solid catalysts was measured by temperature-programmed desorption of diethylamine (DEA-TPD) method 19 , carried out by heating the samples from 20 to 600 °C, in nitrogen atmosphere with 10 °C/ min, using a DuPont Instrument Thermal Analyst 2000/2100 coupled to a module 951 Thermogravimetric Analyzer. The total acidity was determined based on the diethylamine desorbed on the acidic sites, expressed as miliequivalents DEA per gram of catalyst. The pyridine-adsorbed FT-IR spectrum was recorded using Jasco 610 spectrometer, with a scanning range from 4000 to 400 cm −1 , a scan rate of 4 cm −1 ·s −1 and an average of 64 measurements in the final spectrum.
The X-Ray Photoelectron Spectroscopy (XPS) spectra were obtained on a SPECS spectrometer by using the Al anode (1486.6 eV) radiation.
Catalytic activity. The catalytic HDS tests of thiophene (T 0.625 wt%) 2-ethylthiophene (2-ET 0.835 wt% ) or benzothiophene (BT 1 wt%) or their equimolar mixture (T 0.21 wt%, 2-ET 0.28 wt% and BT 0.33 wt%) diluted in n-hexane over sulfided CoMo/γ-Al 2 O 3 and CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalysts, have been performed in a fixed bed flow reactor described in our previous work 19 . The experiments were carried out at a pressure of 30 bar, a liquid hourly space velocity (LHSV) of 2 h −1 , an H 2 /sulfur compounds molar ratio of 60/1 and temperatures in the range 200-275 °C. The raw materials and the reaction mixtures were analyzed by Gas Chromatography (Varian 3800) coupled with Mass Spectrometry (Varian 4000), equipped with Agilent VF-5ms capillary column (L = 30 m, D = 250 μm, d = 0.25 μm). The sulfur content for the feed (reactants) and reactor effluent was calculated based on the concentrations of the sulfur compounds from GC-MS analyses. All experiments were performed in triplicate.

Results and discussion
Catalysts characterization. Textural  The presence of niobium pentoxide into the γ-Al 2 O 3 support decreased the surface area from 230 to 192 m 2 /g and pore volume from 0.496 to 0.199 cc/g. The mean pore diameter of both supports is also between 3 and 4 nm. Before sulphidation, CoMo/γ-Al 2 O 3 -Nb 2 O 5 had a specific area of 176 m 2 /g, compared to that of CoMo/γ-Al 2 O 3 which had a specific area of 184 m 2 /g. However, after sulphidation, the specific area of the CoMo/γ-Al 2 O 3 -Nb 2 O 5   Table 2.
The total acidity of CoMo/γ-Al 2 O 3 catalyst before sulphidation (0.544 meq DEA/g catalyst) was higher than that of CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst (0.480 meq/g) ( Table 2). The γ-Alumina support has both Lewis and Brönsted acidity and is commonly used for HDS catalysts. It can be observed ( Table 2) that after sulphidation, the acidity of CoMo/γ-Al 2 O 3 catalyst decreased significantly compared to the acidity of the CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst. This behavior can be explain by the presence of niobium pentoxide acid which has both Lewis and Brönsted acidity (provided by Nb) due to the remaining -OH groups 11,20 .
he distribution of acidic sites in terms of diethylamine weight loss with increase in. temperature from 20 to 600 °C for the catalyst supports is presented in Fig. 1. emperature-programmed desorption curves of diethylamine for CoMo/γ-Al.

O. 3. catalyst before (A).
and after sulphurization (B.     www.nature.com/scientificreports/ Not only acidic strength, but also the acid types affect the HDS performances of the catalysts, since only appropriate amounts of Brönsted acid sites (B) and Lewis acid sites (L) can facilitate the adsorption of reactant molecules and the surface reactions. The amounts of Brönsted and Lewis acid sites of the catalysts, before and after sulphidation were calculated according to the Py-FTIR analyses.
The pyridine-adsorbed FT-IR spectrum showed features in the region of 1400-1560 cm −1 due to the stretching vibrations of M-N (metal-nitrogen) and N-H (pyridinium ion). The IR vibration peak at 1544 cm −1 is ascribed to Brönsted acid sites and the IR vibration peak at 1448 cm −1 is assigned to Lewis acid sites (Figs. 5 and 6) and the band around wavenumber 1488 cm −1 is due to physically adsorbed pyridine 21,22 . A significant amount of Lewis acid sites can be observed for CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst. It can be clearly seen that the use of Nb improve the Lewis acid sites of the CoMo/γ-Al 2 O 3 catalyst. This may be because Nb atoms have a special outer electronic structure, which makes it exhibit more empty orbitals, leading to the formation of more Lewis acid sites.  X-ray photoelectron spectroscopy (XPS). The chemical species present on the catalysts surface were evaluated by XPS to explain the activity in HDS of thiophene, 2-ethylthiophene and benzothiophene. In CoMo/γ-Al 2 O 3 catalyst, before and after sulphidation (Fig. 7), Co was present as Co 3+ and Co 2+ species. The binding energy of Co 3+ , with the Co 2p 3/2 profile, was 780.97 eV (before sulphidation) and 782.79 eV (after sulphidation) which could be ascribed to Co 3 O 4 and CoOOH 23 . Co 2+ was present as CoO and CoAl 2 O 4 24 with Co 2p 3/2 peak at 779.19 eV (before sulphidation) and 780.30 eV (after sulphidation). After sulphidation the Co 2p 3/2 binding energy region showed a component at 777.52 eV which may be attributed to Co-S structures, but we are unable to give the exact composition for such a species 25 . The deconvolution of Co 2p spectra also included the two characteristic www.nature.com/scientificreports/ satellite features. The Mo 3d spectra of catalysts before and after sulphidation, indicated Mo 6+ species, with the Mo 3d 5/2 peak at 230.81 eV (before sulphidation) and 231.66 eV (after sulphidation), which could be attributed to MoO 3 26, 27 . After sulphidation MoS 2 was formed 28 , shown by the presence of the Mo 3d 5/2 peak at 228.06 eV. As expected after sulphidation, in the Mo 3d spectra the S 2 s peaks appeared, corresponding to sulfide and sulfate components. From the calcination step of the catalyst preparation Co and Mo aluminates or double oxides containing Co and Mo could also be formed [29][30][31] . The S 2p 3/2 binding energy region showed a component at 160.69 eV and S 2 s at 225.08 eV that was characteristic of sulfide (S 2− ) species. Aluminium in γ-Al 2 O 3 , before and after sulphidation had a peak at 74.44 eV (Al 2p 3/2 ), and the oxygen corresponding to metal oxides had the O 1 s peak at 530 eV and 531.4 eV.
The XPS spectra of CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst (Fig. 8) indicated the presence of Mo 6+ with characteristic Mo 3d 5/2 peak at 231.7 eV before sulphidation and 230.94 eV after sulphidation. MoS 2 species, after sulphidation, were confirmed by Mo 3d 5/2 peak at 227.15 eV and by S 2 s and S 2p 3/2 peaks at 224.96 eV and 160.33 eV. Similar to the CoMo/γ-Al 2 O 3 catalyst, Co had specific peaks corresponding to Co 3+ and Co 2+ species and after   Table 3. M% represents mass percentage of different metal species (Me i ), reported to the total metal species, resulting from the curve fitting analysis.
The presence of niobium oxide in the catalytic support had a positive effect in the sulphidation degree. The sulphidation degree for each metal species was calculated from the XPS analysis, resulting in Co (37.1%), and Mo (69.6%) in the CoMo/γ-Al 2 O 3 catalyst and Co (58.7%), and Mo (98%) in the CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst.
Peak assignment was carried out according to literature sources 26,27,33 . The HDS of aromatic sulfur compounds proceeds through successive and parallel reactions of hydrogenolysis of C-S bonds. The resulting final products did not contain sulfur and H 2 S, through partial or total hydrogenation of these aromatic sulfur compounds [34][35][36] , with the formation of other heterocyclic sulfur compounds, unsaturated or saturated. Figure 10 shows the yield variations of the liquid reaction products, with temperatures, over the CoMo/γ-Al 2 O 3 catalyst. For thiophene HDS, the concentrations of the intermediate products (dihydrothiophene and tetrahydrothiophene), were less than 0.2% and therefore the variations of the yields of these products were not graphically represented, which means that for thiophene HDS over tested CoMo catalysts, the hydrogenation reactions (with the formation of saturated or unsaturated cyclic sulfur compounds) are less than of the benzothiophene HDS process.
High temperatures favored the formation of the final hydrogenated products (Fig. 10). For example, in the case of 2-ethylthiophene HDS, at 200 °C, the yield of 2-ethyltetrahydrothiophene (intermediate product) was 2.01% and the yield of hexane (final product) was 56.80%. By increasing the temperature, final product conversion rates were higher. At 275 °C, the yield of hexane reached 97.85% (Fig. 10A). The same behavior was observed in HDS of the equimolar mixture for aromatic sulfur compounds (Fig. 10B).
The HDS of benzothiophene, produces dihydrobenzothiophene as an intermediate product and ethylbenzene as the final product. The yield in dihydrobenzothiophene decreased from 1.94% at 200 °C to 0% at 275 °C, and ethylbenzene yield increased from 79.80% to 99.78%. In the HDS equimolar mixture of T, 2-ET and BT, (Fig. 10B), the product yields of BT hydrodesulphurization had lower values than the yields of these products, Table 3. Binding energies (eV) and mass percentage from fitting curve analysis, for metal species of catalysts before (A) and after sulphidation (B).  Fig. 10 are supported by the work of 37 reporting that sulfur removal from thiophene is slower than from benzothiophene. Same reactivity order is reported by Kilanowski et al. 38 .

CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst.
CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst had a higher activity in hydrodesulphurization process of T, 2-ET and BT, their conversions being higher than 90% for single component HD in the studied temperature range. At 200 °C, the conversion of T was already 91.27% and reached 100% at 275 °C. The conversion of 2-ET was 96.15% at 200 °C and 100% at 275 °C, and the conversion of BT had reached 100% at temperatures higher than 200 °C (Fig. 11A).
For the mixed species, the conversion of sulfur compounds was smaller compared to the conversion of the single component hydro-sulfurization. At 200 °C, the conversions were 31.73% (T), 56.33% (2-ET) and 97.37% (BT). However, at higher temperatures (275 °C), the conversion for each component was closer to the conversion obtained in the single component hydro-sulfurization (Fig. 11B).
The influence of temperature on the yields of reaction products for a single component and mixture of T, 2-ET and BT hydrodesulphurization, over CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst is presented in Fig. 12.
The CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst favored the formation of the sulfur-free products. For benzothiophene HDS, dihydrobenzothiophene was not identified and for 2-ethylthiothiothiophene HDS, the yield of 2-ethyltetrahydrothiophene was below 1%. www.nature.com/scientificreports/ For the HDS mixture of aromatic sulfur compounds (Fig. 12B), the yield of intermediate products increased slightly compared to the yields obtained in single component HDS. For example, the yield of 2-ethyltetrahydrothiophene was 3.70% at 200 °C and decreased with increasing temperature (Fig. 12B). However, the yield of hexane increased, from 52.64% at 200 °C to 100% at 275 °C. The yield of dihydrobenzothiophene was 1.11% at 200 °C and decreased with increasing temperature to 0%.
The sulfur concentration (ppm) after the HDS process is listed in Table 4. At 275 °C, the sulfur is completely removed from all sulfur compounds in single component HDS and is reduced to 7.23 ppm in a mixture of sulfur compounds HDS.
From the experimental results, one can note that: (1) HDS reactivity of the tested sulfur compounds decreased in the following order: BT > 2-ET > T; (2) CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst was more active in HDS for the tested sulfur compounds compared to that of the CoMo/γ-Al 2 O 3 catalyst; (3) There was a strong inhibition effect in the HDS of a ternary mixture of T, 2-ET and BT compared to that of a single component HDS species. The inhibition effect was more pronounced for T, than for 2-ET and BT. The inhibition effect decreased with increasing temperature and was less pronounced over the CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst. This behavior can be explained by two factors: electronic and steric com- www.nature.com/scientificreports/ ponents. The molecular size of the sulfur compounds (molecular critical diameter) decreases in the order benzothiophene > 2-ethylthiophene > thiophene. The size is between 4.65 Å for thiophene 39 and 6 Å for benzothiophene 40 . Consequently, the adsorption of thiophene on active sites of the catalyst should be more pronounced than that of benzothiophene, if the smaller volume of the thiophene molecule is considered.
The thiophene compounds are classified as aromatic because of their high resonance energy (29 kcal/mol for thiophene) and their properties, but the aromatic character decreases in the case of condensed rings, as in the case of benzothiophene 41 . Sulfur participates with two electrons in the formation of the aromatic electron sextet and the thiophene nucleus of the aromatic sulfur compounds shows basicity due to the six π electrons delocalized on the five-membered heterocyclic compound 41 . The molecular adsorption on the acidic sites of the catalyst is favored if the basicity of the aromatic nucleus. In the series of investigated compounds, the basicity increases in the order of thiophene < 2-ethylthiophene < benzothiophene. The higher basicity of benzothiophene is due to the weaker delocalization of π electrons on the two condensed aromatic nuclei. Subsequently, these electrons interact more strongly with the acidic centers of the catalyst, compared to that of thiophene, which has a single aromatic nucleus and a stronger aromatic character. 2-Ethylthiophene has higher basicity in the nucleus than thiophene due to the ethyl group which stabilizes the positive charge that appears in the nucleus, as a result of the interactions of the aromatic ring-acid centers in the catalyst. The higher reactivity of benzothiophene proves that the HDS process at the investigated parameters is influenced more by the electronic factors that favor the adsorption of sulfur compounds on the active centers of the catalyst than by the steric factors. The influence is also explained by the fact that the pore volume of the catalyst, 3.27 nm, is larger than the size of the molecules of the studied sulfur compounds 40 . High HDS reactivity of benzothiophene, despite the larger molecular volume, has also been reported in other the studies 40 . The higher the rate of the hydrogenation reaction of benzothiophene compared to thiophene, results in the higher reactivity of benzothiophene compared  www.nature.com/scientificreports/ to that of 2-methylthiophene and 3-methylthiophene (Brunet et al. 18 ). The relative reactivity of sulfur aromatic compounds depends on their structure. All three aromatic sulfur compounds react more slowly in HDS where the mixture is more influenced by thiophene than by 2-ethylthiophene and benzothiophene. This inhibition effect may be explained by the competitive adsorption between the three aromatic sulfur compounds on the active sites of the catalyst. The adsorption of each compound is delayed by the presence of the other two. Since the reactivity of the sulfur compounds is different, the inhibition effect is the lowest in the case of the sulfur compound with the highest in the HDS reaction with benzothiophene.
At higher temperatures (approx. 275 °C), the inhibition effect decreases considerably or is no longer present because the HDS reaction rate for all sulfur compounds increases more compared to the reaction rate at lower temperature (ex. 200 °C). Similarly, it was observed that the inhibition effect is attenuated on CoMo/γ-Al 2 O 3 -Nb 2 O 5, due to better textural characteristics and higher acidity, compared to the CoMo/γ-Al 2 O 3 catalyst.
It should be pointed out that under the conditions of industrial HDS, the petroleum fractions contain a mixture of sulfur compounds, depending on the distillation interval of these fractions; thus, it may be an important deviation from the additivity to HDS of mixtures of sulfur compounds in comparison with the individual HDS of these compounds studied in the laboratory.

Conclusions
The results presented in this paper showed that the CoMo/γ-Al 2 O 3 -Nb 2 O 5 catalyst is more active in the HDS reaction of aromatic sulfur compounds than CoMo/γ-Al 2 O 3 catalyst due to the higher specific surface area and higher acidity of the niobium catalyst after sulphidation.
The reactivity in the HDS reaction decreases in the order of benzothiophene > 2-ethylthiophene > thiophene, because the adsorption of the molecules on the acid centers of the catalyst is favored if the basicity of the aromatic nucleus is higher. Higher HDS reactivity of benzothiophene under the experimental conditions, proves that the process is influenced by the electronic factors that favor the adsorption of sulfur compounds on the active sites of the catalyst and not by the steric factors.
The experimental data has shown that for HDS of an equimolar mixture of aromatic sulfur compounds, under the experimental conditions (200-275 °C, 30 bar and LHSV 2 h −1 ), there is a strong mutual inhibition effect compared to that of the single component HDS species. The effect is more pronounced at lower temperatures (200-250 °C), thiophene being more influential, than 2-ethylthiophene and benzothiophene. The inhibition effect decreases with increasing catalyst activity.
The inhibition effect can be explained by the competitive adsorption between the three aromatic sulfur compounds on the active sites of the catalyst.