Thiophene hydrodesulfurization over CoMo/Al₂O₃-CuY catalysts: Temperature effect study

2.1 Catalysts preparation and characterization

Three bifunctional catalysts samples are prepared by loading a sulfided CoMo/γ-Al₂O₃ catalyst with a small amount of Cu-exchanged zeolites via a manual mechanical mixing. The industrial CoMo/γ-Al₂O₃ catalyst is provided by Procatalyse (MoO₃ = 14%, CoO = 3%). The obtained catalysts noted as CoMo, CoMo1 and CoMo2 contain a zeolite loading (wt%) rate of 0, 10 and 20, respectively. Before the use of the CoMo/γ-Al₂O₃ catalyst, a mass of 4 g is dried at 673 K for 1 h, then sulfided ex-situ in a fixed bed reactor with 15% of H₂S/H₂ mixture at atmospheric pressure and at the same temperature for 4 h (heating ramp of 10 K/min). Finally, the sample is flushed in a nitrogen flow. The Cu-exchanged zeolite is prepared according to the solid state ion exchange technique (SSIE) (Benaliouche et al., 2011). Prior to the ion exchange, the H form of Y zeolite is obtained by heating the commercial NH₄-Y zeolite (supplied by Zeolyst International, SiO₂/Al₂O₃ = 5.2; Na₂O = 0.2%) from room temperature to 773 K at 2 K/min in a dry nitrogen atmosphere. The sample is kept at that temperature for 6 h. After that, the obtained HY zeolite is thoroughly manually mixed with the appropriate amount of CuCl₂ salt (supplied by Merck, purity ⩾ 99.8%). The calculated amount of CuCl₂ salt is sufficient to produce the fully Cu-exchanged zeolite. After that, the obtained reactive mixture is heated at 873 K in a dry nitrogen atmosphere for 10 h. The chemical analysis of both Cu-exchanged zeolite and CoMo/γ-Al₂O₃ catalyst is performed by Inductively-Coupled Plasma spectrometer (Activa Jobin Yvon ICP-OES). The crystalline structure of the Cu-exchanged zeolite is examined by Philips X’pert PRO diffractometer. The powder X-ray diffraction (XRD) patterns are collected over a 2θ-range of 3–50° using Cu-Kα radiation (λ = 1.54030 Å). The porosity of catalysts is measured by adsorption and desorption of nitrogen at 77 K using a Quantachrome NOVA 3200 surface analyzer. The hydrogen temperature-programmed reduction (H₂-TPR) is performed after activation by sulfidation of all catalysts at atmospheric pressure. First, the sulfide sample of 60 mg is loaded in a U-shaped quartz flow reactor of a conventional setup equipped with a thermal conductivity detector (TCD). After that, the sample is heated with a temperature program from room temperature to 1323 K (heating rate of 10 K min⁻¹) under a gas flow (50 ml min⁻¹) of H₂/Ar mixture (5% H₂). In such a way, the catalyst samples are reduced and the amount of consumed H₂ is measured by a TCD detector. The amount of H₂O produced in the reduction was removed by a cold trap placed before the TCD detector. These conditions are in accordance with the literature (Dos Santos et al., 2009; Brooks, 1980; Tanaka et al., 1996).

2.2 Catalytic tests reaction

The experiments for thiophene HDS are carried out at atmospheric pressure in a continuous flow fixed-bed reactor and at different temperatures (573, 593 and 613 K). Thiophene is introduced at a rate of 9.718 μmol/s into the reactor containing sulfided catalyst (about 60 mg) by flowing hydrogen (50 ml/min) through a thiophene trap cooled at 273 K. Products are analyzed on-line by using an HP Hewlett 5890 Series II gas chromatograph analyzer equipped with a flame ionized detector (FID). The reaction is conducted after the catalyst sample reached the desired temperature and until the stabilization of the reaction rate. Before HDS model reaction, all catalysts are sulfided under the same operation conditions cited previously in H₂-TPR analysis.

3.1 Characterization of catalysts

The results of chemical analysis and textural characterization of the catalysts are reported in Table 1. The external surface area and micropore volume are calculated using the t-plot method. The total surface area is calculated according to the BET model. The results show that the commercial CoMo/Al₂O₃ has a chemical composition of 2.42 wt% Co and 9.26 wt% Mo. It has a pore volume of 0.48 cm³/g and a surface area of 184 m²/g. These results are in concordance with those obtained in the literature (Homma et al., 2005; Rana et al., 2004). For Cu-exchanged zeolite, the results reveal that H cations are almost completely exchanged by Cu²⁺ and the exchange degree is estimated at about 92% (determined by ICP elemental analysis). The Si/Al ratios for both HY, and Cu-exchanged forms obtained are nearly equal. We notice that all textural parameters values of the HY zeolite (total pore volume, the microporous volume and the total surface area) decrease after ion exchange with copper. This behavior may be explained by the fact that zeolite pore access can be partially blocked when Cu ions are in large numbers at sites inside the cages, leading to a reduction in measured volume and total surface area (Benaliouche et al., 2008). In order to analyze the effect of both the ion exchange and the sulfidation operations on the crystalline structure of the HY and CuY zeolites, we present in Fig. 1, the XRD analysis of the parent HY zeolite and its Cu-exchanged form. We also made XRD measurements for CoMo/Al₂O₃ (spectrum not shown) but no appreciable changes are observed. The diffractrogram of the parent HY zeolite shows the characteristic peaks of the faujasite framework that closely matches those from the literature (Breck, 1974; Treacy and Higgins, 2001). This result indicates the highly pure and crystalline nature of the parent zeolite. For Cu-exchanged zeolites, the XRD profiles show similar characteristic peaks of HY zeolites with a very slight decrease in intensities. These results indicate that Cu²⁺ ions seem to be well dispersed in the zeolite framework and the structure integrity of Cu-exchanged zeolite is maintained after the ion exchange and sulfidation treatment. We notice the appearance of a significant diffraction lines at 2θ values of 44° and 46° for CuY and sulfided CuY zeolites, respectively. These results suggest the formation of a small amount of CuO (ICDD #00-005-0661) and Cu₂S species (ICDD #00-053-0522). These observations may indicate that sulfidation of Cu-exchanged Y zeolites leads to the formation of sulfided Cu²⁺ and, probably, reduced Cu species.

Close

Figure 1

The XRD patterns of parent HY zeolites (a), Cu-exchanged Y zeolites (b) and sulfided Cu-exchanged Y zeolites (c).

Export to PPT

3.2 Temperature programmed reduction (H₂-TPR) analysis

The reducibility of Co, Mo and Cu sites in CoMo/Al₂O₃ catalyst and Cu-exchanged Y zeolite is studied by H₂-TPR. These analyses are carried out on sulfided samples treated after thiophene HDS test. Fig. 2I and II, show the H₂-TPR profiles of the CoMo/Al₂O₃ catalyst and Cu-exchanged Y zeolite compared with those of the fresh sulfided samples. As expected, TPR signal of CoMo/Al₂O₃ catalyst (Fig. 2I) gives two main peaks corresponding to H₂S physisorbed gas (peak below 500 K) and the hardly reducible molybdenum species, respectively (peak above 900 K). The broad band between 600 and 900 K is attributed to both Co⁰ species and Co–Mo–S phase (Ribeiro et al., 2007; Resini et al., 2003). This phase is responsible for HDS activity. We observe also a slight shift of all peaks after thiophene HDS test. The first peak is slightly shifted to higher temperature whereas the second one to lower. These results may indicate the relative stability of the active phase on the catalyst after the catalytic treatment.

Close

Figure 2

Temperature programmed reduction of CoMo/Al₂O₃ catalyst (I) and Cu-exchanged Y zeolites (II), before treatments (a) and after thiophene HDS test (b).

Export to PPT

For Cu-exchanged Y zeolite, the H₂-TPR pattern of the fresh sulfided (Fig. 2II(a)) sample is composed of three well-separated reduction peaks, at low temperatures (two peaks below 600 K) and high temperatures (one peak at 1168 K). The low-temperature peaks at 388 and 518 K can be attributed to the H₂S physisorbed gas and to the easier reducible species (Cu²⁺ to Cu⁰ and/or Cu²⁺ to Cu⁺) respectively. According to literature, these species are mostly located on the accessible sites in the supercages and/or sodalite units of the zeolite (Ribeiro et al., 2007; Resini et al., 2003). The higher temperature peak observed at 1168 K seems to be much more intense than those cited previously and can be ascribed to the formation of metallic copper (Cu⁰) from Cu⁺ ions. These species mainly situated in the hexagonal prisms are more resistant to the reduction than those located in the supercages of the zeolite. The hardly reducible copper species are known to be stabilized by the negatively charged zeolite framework (Ribeiro et al., 2007; Cid et al., 1995). The H₂-TPR profile obtained after the HDS reaction of thiophene (Fig. 2II(b)) reveals that the changes are more noticeable for high temperature reduction peaks. The intensity of this peak increases and shifts to low temperatures from 1168 to 748 K. We also note that the low temperature peaks overlapped to form one peak at 462 K with a shoulder reaching the maximum at 390 K. All these behaviors may be explained by both the reduction and the migration of more accessible Cu species from different site positions in zeolite during the HDS reaction of thiophene (presumably the migration of copper species from less accessible sites in hexagonal prism to the supercages of the zeolite). The effect of the oxidation of some copper species is not excluded. Similar results are reported by Cid, R., and al using Ni²⁺-exchanged USY zeolite (Cid et al., 1995).

3.3 Catalytic activity

The results of the thiophene HDS catalytic activity over CoMo, CoMo1 and CoMo2 catalysts are presented in Fig. 3. The values of thiophene conversion obtained after 16 h, show their dependence on both the HDS temperatures and the amount of CuY zeolites added to CoMo/Al₂O₃ catalyst. We observe that the thiophene conversion rate increases with the increase of temperature for all samples. As proposed by Topsøe et al. (1996), Brooks (1980) and Ribeiro et al. (2007), the HDS of thiophene compounds proceeds via two main pathways: (i) direct desulfurization (DDS) by the hydrogenolysis of the C–S bond and (ii) hydrogenation (HYD) of the aromatic ring followed by C–S bond hydrogenolysis. According to Topsøe et al. (1996), the increase of the thiophene HDS activity over CoMo/Al₂O₃ catalyst is attributed mainly to the formation of Co–Mo–S active site on the S-edge of MoS₂, which favors the direct desulfurization (DDS) of thiophene because of the participation of sulfur vacancy. In our case, the CoMo sample showed a high catalytic activity at 573 K of about 46 and 54% greater than CoMo1 and CoMo2 samples, respectively. This threshold value is reduced to 45% and 38% for CoMo1 and CoMo2, respectively, when the temperature is increased to 593 K. After that, widening of the gap between the estimated conversion rates is obtained at 613 K. All these behaviors may be explained by the fact that the HDS activity of the CoMo catalyst contains a contribution of both the support and the cobalt-molybdenum sulfide phases (Co–Mo–S). However, for CoMo1 and CoMo2 catalysts, we have in addition to support and the Co–Mo–S phase contributions there are a new dual functions for HDS in which both the Cu species and the acid zeolite substrate contribute to desulfurize thiophene at reaction conditions. Indeed, the presence of strong acid sites in CuY zeolite pores may cause fast deactivation (poisoning of active sites) and coke formation in the CoMo1 and CoMo2 catalysts during the HDS of thiophene particularly at high temperatures. These species can block the active sites of the catalyst leading to the reduction of HDS activity (Caeiro et al., 2006; Carsten et al., 2008; Corma et al., 1996). On the other hand, the fact that the formation of sulfided Cu⁺ and the reduced Cu species (as indicated by XRD and TPR analyses) that migrated to the supercages of the zeolite causes a harder diffusion of reactant through pores of catalyst is not excluded. The reaction products in the HDS of thiophene over CoMo/Al₂O₃ catalysts mixed with various amounts of CuY zeolites are given in Table 2. We found that C₄ hydrocarbons are formed as main reaction products in the HDS of thiophene for all catalysts. Small amounts of butane (product obtained by hydrogenation of butenes) are detected particularly for CoMo1 and CoMo2 samples. No cracking C₁-C₃ hydrocarbons products are detected. This indicates that CoMo/Al₂O₃ catalysts have low cracking activities for hydrocarbons. The selectivity results also reveal that the products formed by both hydrodesulfurization of thiophene (1-butene) and double-bond isomerization of 1-butene (trans- and cis 2 butene) are sensitive to the temperature changes particularly for samples containing zeolite loading (wt%) rate less then 10%. Indeed, the values obtained for CoMo catalyst reveal that the increase of temperature from 573 to 613 K leads to increase in the amounts of iso- and 1-butene products of about 2.5% (from 49.4% to 51.9%). Consequently, a decrease of the cis-2-butene from 29.5% to 27.9% is observed. The values obtained for trans-2-butene are nearly equal (about 21%). Nearly the same evolution is observed for CoMo1 at which the variation of hydrodesulfurization and isomerization products is estimated to be about 4.5%. However, for CoMo2 sample, only the value of butane decreases from 0.43% to 0.26% together with the temperature increase. Such behavior may be explained by the fact that the hydrogenation sites on CoMo2 catalyst are affected by the deactivation process.

Close

Figure 3

Thiophene conversion over CoMo/Al₂O₃ catalysts after 16 h: (a) CoMo, (b) CoMo1 and (c) CoMo2.

Export to PPT