Você está na página 1de 9

Applied Catalysis A: General 353 (2009) 296304

Contents lists available at ScienceDirect

Applied Catalysis A: General


journal homepage: www.elsevier.com/locate/apcata

Combined effect of noble metals (Pd, Au) and support properties on HDS activity of Co/SiO2 catalysts
` A.M. Venezia a,*, R. Murania b, G. Pantaleo a, V. La Parola a, S. Scire c, G. Deganello b
a

Istituto per Lo Studio dei Materiali Nanostrutturati (ISMN)-CNR, via Ugo La Malfa 153, 90146 Palermo, Italy ` Dipartimento di Chimica Inorganica e Analitica Stanislao Cannizzaro, Universita di Palermo, Viale delle Scienze, Parco dOrleans, 90128 Palermo, Italy c ` Dipartimento di Scienze Chimiche, Universita di Catania, Viale A. Doria 6, 95125 Catania, Italy
b

A R T I C L E I N F O

A B S T R A C T

Article history: Received 24 September 2008 Received in revised form 30 October 2008 Accepted 5 November 2008 Available online 17 November 2008 Keywords: Hydrodesulfurization HDS Co catalysts Pd Au effect Mesoporous supports

Cobalt-based catalysts supported on different types of SiO2 are studied in the hydrodesulfurization of thiophene. Amorphous silica and siliceous MCM-41 and HMS, characterized by different texture and surface acidity are used as carriers. The effects due to the modication of the support by impregnation with palladium precursor and to the co-impregnation of cobalt and gold are considered. The catalysts are characterized by N2 physisorption (BET), XRD, TPR and XPS techniques. The use as supports of the ordered mesoporous silica with higher surface area with respect to amorphous silica, produces a better dispersion of the cobalt oxides particles. Moreover, the addition of palladium to the supports inhibits the interaction between cobalt and silica and favors the reducibility of the cobalt oxide. Co-impregnation of gold and cobalt decreases the Co3O4 crystallite size. Due to the combination of these effects an increase in the cobalt activity is observed. The extent of the improvement is highly dependent on the support characteristics. As compared to the monometallic Co catalysts, the monometallic Pd catalysts supported on amorphous SiO2 and on MCM-41 are less active. However quite unexpectedly the Pd/HMS exhibits the best performance among all the studied samples. Catalytic tests of HDS of tetrahydrothiophene (THT) support the direct route of the CS bond cleavage as the rate-determining step of the reaction. 2008 Elsevier B.V. All rights reserved.

1. Introduction In the search for more efcient hydrotreating catalysts, especially suitable for deep hydrodesulfurization, novel systems based on new type of supports and on species other than Mo promoted with elements from Group VIII have been investigated [1]. Particularly, as alternative to the conventional alumina supported CoMo catalysts, suffering from sintering and coke deactivation, systems based on mesoporous supports and noble metals have been explored in different laboratories [26]. Recently, Al and Ti modied HMS materials used as supports for NiMo and CoMo HDS catalysts have been shown to enhance the activity in the HDS of thiophene and dibenzothiophene [4]. Mesoporous MCM-41, and MCM-48 doped with Al or Zr have been also used as support for catalysts in hydrotreating processes [57]. The activity enhancement determined by the use of the mesoporous supports is largely due to the high surface area which favors the dispersion of the active species. For HDS application the majority of studies have focused on Mo as the

* Corresponding author. Tel.: +39 0916809372; fax: +390916809399. E-mail address: anna@pa.ismn.cnr.it (A.M. Venezia). 0926-860X/$ see front matter 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.apcata.2008.11.005

main active species and Co or Ni as activity promoters. However the use of monometallic Co catalyst for HDS has also been considered and in some cases, such as Co over carbon, higher activity of Co/C as compared to Mo/C was obtained [8]. Recently in our group, monometallic Co/MCM-41 with HDS activity comparable to CoMo on amorphous silica was obtained [3,9]. It was shown that, depending on the type of the support, very interacting like alumina or inert like silica, it was possible through the choice of an appropriate Co supporting method to enhance the activity of the catalyst by favoring the formation of the active species Co3O4 [9]. As a possible way to avoid or limit the catalyst deactivation by coke formation during the hydrotreatment reaction, addition of noble metals (Rh, Pt, and Pd) to Cosaponite has been considered [10 12]. Moreover an increased HDS activity of noble metal modied cobalt catalysts was attributed to the enhanced reducibility and metal dispersion. Very recently, addition of gold to cobalt catalysts supported on amorphous and mesoporous (MCM-41) silica was reported to determine an increase of the HDS activity [13]. Again the explanation was the improved Co dispersion and the increased Co reducibility, more evident in the case of the mesoporous support. In that study two different procedures such as, consecutive impregnation and co-impregnation of Au and Co were also investigated. It was concluded that the large effect on the

A.M. Venezia et al. / Applied Catalysis A: General 353 (2009) 296304

297

reducibility of the Co was achieved by the consecutive impregnation in which cobalt was deposited after modication of the support by gold addition [13]. However it was also shown that the contemporary addition of the two metals, Au and Co produced a decrease of the active Co3O4 crystallite size. Based on these results and also as continuation of our research interest in the use of gold and palladium for hydrotreatment reactions [2,14], in the present work we explored the possibility to enhance the catalytic performance of cobalt catalysts by using mesoporous silica as supports and by adding two different noble metals, gold and palladium for the purpose of hydrodesulfurization. The strategy of the study was to improve the reducibility of the Co by adding a metal like Pd which is regarded as a good hydrogen activator and at the same time to increase the catalyst dispersion by addition of Au. The obtained catalysts were tested in the HDS of thiophene and tetrahydrothiophene. Their catalytic behaviour was related to structural and surface properties of the materials as determined by XRD, TPR and XPS analyses. 2. Experimental 2.1. Support and catalyst preparation The commercial silica was obtained from Merck (amorphous silica gel 60). The MCM-41 was prepared by hydrothermal method according to the method reported by Choma et al. [15]. The surfactant, hexadecyl trimethylammoniumbromide (CTAB), was dissolved in a solution of aqueous ammonia (28 wt.%), distilled water and ethanol. After stirring this mixture for 10 min, the tetraethyl orthosilicate (TEOS) was added. The molar composition of the gel was: 0.09 TEOS:1.12 NH4OH:4.86 EtOH:0.04 CTAB:15.62 H2O. This mixture was homogenized by stirring for 2 h at ambient temperature before heating at 100 8C for a total of 5 days under static conditions. Thereafter the product was calcined at 650 8C for 6 h in air. Attainment of the ordered mesoporous structure was conrmed by the XRD pattern showing the typical reection at 2u % 2.68 and by the typical type IV N2 adsorptiondesorption isotherm [5,16]. The mesostructured HMS material was synthesized according to a published procedure [17]. Basically, HMS was assembled from 4:1 molar mixtures of (TEOS) (Aldrich) as the inorganic precursor and dodecylamine (DDA) (Aldrich) as the structure-directing surfactant in 90:10 (v/v) water/ethanol. About 49 mmoles of DDA was dissolved in 50 ml of ethanol and 450 ml of H2O. The surfactant solution was heated to 60 8C and 196 mmoles of TEOS was added to give the reaction mixture. The gel mixture was kept in a closed Teon vessel and stirred at 60 8C for 20 h. The reaction product was ltered, washed with distilled water and dried at room temperature for 24 h. The surfactant was removed by calcination in air at 600 8C for 4 h. Attainment of the ordered mesoporous structure was conrmed by the XRD pattern showing a single diffraction line at 2u % 28 and by the typical type IV N2 adsorptiondesorption isotherm [17,18]. The cobalt catalysts with 5 wt.% Co, were prepared by incipient wetness impregnation of the supports, with an aqueous solution of Co(NO3)26H2O (from Aldrich) followed by 2 h drying at 120 8C and calcination at 400 8C for 2 h. The monometallic palladium catalysts were prepared by wetness impregnation of the supports with an aqueous solution of PdCl2 (Aldrich) followed by drying at 120 8C and calcination at 400 8C for 2 h. The second impregnation of the Pd catalysts with Co(NO3)26H2O solution, followed by the same drying and calcinations procedure, yielded the Pd promoted Co catalysts. In the case of the tri-element catalysts, the Pd promoted supports were co-impregnated with aqueous solution of cobalt and gold precursors (AuCl3 from Aldrich) and thereafter were dried and calcined as previously. The 5 wt.% loading for Co and the 1 wt.% each for Pd and Au was ascertained by X-ray uorescence analyses.

2.2. Catalyst characterization 2.2.1. X-ray diffraction X-ray diffraction measurements for the structure determination were carried out with a Philips vertical goniometer using Niltered Cu Ka radiation. A proportional counter and 0.058 step sizes in 2u were used. The assignment of the various crystalline phases was based on the JPDS powder diffraction le cards [19]. The particle sizes of the metals and oxides were estimated as volumeaverage crystallite dimension, through the line-broadening (LB) of the available reection peaks, using the Scherrer equation according to the method reported in the literature [20]. The instrumental broadening was determined by collecting the diffraction pattern of the standard, lanthanum hexaboride LaB6. 2.2.2. BET analyses The microstructural characterization was performed with a Carlo Erba Sorptomat 1900 instrument. The fully computerised analysis of the adsorption isotherm of nitrogen at liquid nitrogen temperature, allowed obtaining, through the BET approach, the specic surface area of the samples. By analysis of the desorption curve, using the BJH calculation method, the pore size volume distribution was also obtained [18]. 2.2.3. X-ray photoelectron spectroscopy (XPS) The X-ray photoelectron spectroscopy analyses were performed with a VG Microtech ESCA 3000 Multilab, equipped with a dual Mg/Al anode. The spectra were excited by the unmonochromatised Al Ka source (1486.6 eV) run at 14 kV and 15 mA. The analyser operated in the constant analyser energy (CAE) mode. For the individual peak energy regions, a pass energy of 20 eV set across the hemispheres was used. Survey spectra were measured at 50 eV pass energy. The sample powders were analysed as pellets, mounted on a double-sided adhesive tape. The pressure in the analysis chamber was in the range of 108 Torr during data collection. The constant charging of the samples was removed by referencing all the energies to the C 1s set at 285.1 eV, arising from the adventitious carbon. The invariance of the peak shapes and widths at the beginning and at the end of the analyses ensured absence of differential charging. Analyses of the peaks were performed with the software provided by VG, based on non-linear least squares tting program using a weighted sum of Lorentzian and Gaussian component curves after background subtraction according to Shirley and Sherwood [21,22]. Atomic concentrations were calculated from peak intensity using the sensitivity factors provided with the software. The binding energy values are quoted with a precision of 0.15 eV and the atomic percentage with a precision of 10%. Contact of the samples with air was minimised during sample loading; particular care was adopted for samples analysed after the H2/H2S treatment. In this case the samples were kept in n-eptane until being transferred into the XPS instrument. 2.2.4. Fourier transform infrared spectroscopy (FTIR) The FTIR spectra of adsorbed pyridine provided information on the acidity of the supports, HMS, MCM-41 and commercial SiO2. The spectra were recorded with a Perkin Elmer System 2000 FT-IR spectrophotometer with a resolution of 2 cm1, according to the previously described procedure [23]. The reported spectra have been obtained by subtracting the spectrum of the sample before the admission of pyridine and are normalised to the same amount of catalyst per cm2. 2.2.5. Zero point charge determination (ZPC) The ZPC of the various supports was determined by mass titration [23]. According to this method, the variation of pH of a

298

A.M. Venezia et al. / Applied Catalysis A: General 353 (2009) 296304

water solution containing increasing amount of solid was monitored until the steady state value of pH (ZPC) was reached. 2.2.6. Hydrogen temperature programmed reduction (H2-TPR) TPR measurements were conducted with a Micromeritics AutoChem 2910 Automated Catalyst Characterization System, equipped with a thermal conductivity detector (TCD). About 0.1 g of sample was used for each measurement. The samples were pretreated with a mixture of 5 vol.% O2/He at 50 ml/min, heating up (10 8C/min) to 400 8C and holding at this temperature for 30 min. After lowering the temperature down to 50 8C, the gas mixture of 5 vol.% H2/Ar was introduced at 30 ml/min into the sample tube and was also used as a reference gas. During the analysis, the temperature was increased up to 900 8C at a rate of 10 8C/min. The efuent gas was analysed with a TCD. 2.3. HDS reaction The hydrodesulfurization of thiophene was carried out in the vapour phase using a continuous ow microreactor [3]. An amount of 200 mg of catalyst (sieved fraction 210430 mm), diluted with inert particles of SiC (in a weight ratio of 5:1 with respect to the catalyst) was used for each test. The samples were sulded in situ with a mixture of 10 vol.% H2S/H2, at 50 ml/min, while raising the temperature up to 400 8C at a rate of 7 8C/min and were maintained at this temperature for 2 h. After purging with nitrogen, the HDS of thiophene was carried out at 340 8C with 5.3 vol.% thiophene in H2 and WHSV = 7500 h1. The reaction products were analysed by online gas chromatography (Carlo Erba GC 8340 gaschromatograph). Fractional conversions were calculated from the ratio of the peak area of the C4 products over the sum of the peak areas of the products and thiophene. The reaction rate for HDS (kHDS) was calculated from the fractional conversion at the steady state conditions, reached after 6 h on stream, assuming a rst order reaction in thiophene. Measurements of the rate constants at the temperatures of 355 8C, 370 8C and 395 8C were performed to determine the apparent activation energy for each catalyst. The activity tests with tetrahydrothiophene (THT) were carried out using 1.2 vol.% substrate in H2. The error on the catalytic activity data, such as rate constants and activation energies was estimated as 10%, mostly arising from the oven temperature uctuation. 3. Results and discussion 3.1. Support and catalyst characterization In Table 1 the supports with their textural properties and zero point charges are given. The mesoporous silica are characterized by larger surface area, smaller pore sizes and smaller ZPC values as compared to amorphous SiO2, with the HMS exhibiting the highest total surface acidity (lowest ZPC value). The nitrogen adsorption desorption isotherms of the MCM-41 and the HMS are shown in Fig. 1a and b. The isotherms are of type IV, characteristic of mesoporous materials. According to the IUPAC classication, the hysteresis loops of the MCM-41 and the HMS can be classied as H1 and H3 types respectively [18]. The H1 is associated with a more uniform pore distribution as compared to the H3 type. In
Table 1 Surface area (S), average pore diameters (dp) and zero point charge (ZPC) of the supports. Supports SiO2 MCM-41 HMS S (m2/g) 316 810 790 dp 5 2.6 3.4 ZPC 6.5 5.6 3.6

Fig. 1. Nitrogen adsorption isotherm of (a) MCM-41 and (b) HMS.

Fig. 2 the distribution of the pore sizes for the two materials is plotted. According to the distribution and as given in Table 1, the MCM-41 and the HMS are characterized by an average diameter of pores of 2.6 nm and 3.4 nm respectively. IR spectroscopy of the adsorbed pyridine was used to determine Brnsted and Lewis acidity. The FTIR spectra of the investigated silicas after adsorption of pyridine and outgassing at room temperature are shown in Fig. 3. The spectra present two main bands at 1447 cm1 and 1599 cm1. According to the literature [2325] these bands can be attributed respectively to the 8a and 19b modes of pyridine molecules interacting via H-bonding with the acidic surface hydroxyl-groups (silanols) of SiO2. The shift of both bands with respect to those of not-bonded pyridine is not, in fact, large enough to indicate Lewis acidity, both bands being also completely pumped off at 150 8C. Moreover, the absence of a band at 1540 cm1, indicates that silanol groups are not able to protonate pyridine. From the comparison of the spectra of the three silicas a clear difference in the intensity of the pyridine bands can be noticed, with HMS exhibiting stronger bands as compared to the other two. No substantial difference is instead observed in the frequency of the bands. This behaviour suggests an increasing support acidity in terms of increasing number of silanol groups but not in terms of acid strength.

A.M. Venezia et al. / Applied Catalysis A: General 353 (2009) 296304

299

Fig. 2. Distribution of pores in HMS and MCM-41.

Fig. 5. XRD patterns of calcined Pd and Au promoted Co catalysts on different silica supports.

Fig. 3. FTIR spectra of the silica supports after adsorption of pyridine and subsequent outgassing at room temperature.

In Fig. 4 the XRD patterns of the Co and PdCo systems supported on amorphous and on HMS silica are shown. All the patterns contain the typical reections of the Co3O4 crystalline phase with the additional PdO lines for the Pd promoted catalysts. In Fig. 5 the XRD patterns of the trimetallic systems supported on the three different silicas are displayed. The diffractograms exhibit the additional lines typical of metallic gold. From the line-broadening analyses of the most intense peaks, using the Sherrer equation [20], the crystallite sizes of the main phases were calculated. The obtained values are listed in Table 2. It follows that the Co3O4 particle sizes in the Co catalysts are somehow affected by the supports, being smaller in Co/MCM-41 and Co/HMS as compared to CoSiO2. For all the supports, pre-impregnation with Pd determines a decrease of the cobalt oxide particle size. The co-impregnation of cobalt with gold, in the tri-element systems, contributes to a further decrease of the Co3O4 crystallite sizes. As it is suggested by the absence of any PdO-related peak in the XRD of the MCM-41 supported catalysts and as it is indicated in Table 2, the particle sizes of the PdO crystallites in such samples must be less than the XRD detection limit of $3 nm. The particular texture of the MCM41 with small pores (less than 3 nm) may account for the formation of such small PdO particles.
Table 2 Crystal particle size (d) of catalysts as obtained from XRD pattern. Samples SiO2 series Co PdCo Pd(AuCo) Pd MCM-41 series Co PdCo Pd(AuCo) Pd HMS series Co PdCo Pd(AuCo) Pd dCo3 O4 (nm) 13 12 11 dAu (nm) 16 dPdO (nm) 7 7 10

9 7 5

20

<3.0 <3.0 <3.0

Fig. 4. XRD patterns of calcined Pd promoted and unpromoted Co catalysts on different silica supports.

9 8 7

14

8 7 6

300

A.M. Venezia et al. / Applied Catalysis A: General 353 (2009) 296304

Fig. 6. TPR patterns of HMS supported catalysts.

Fig. 8. TPR patterns of SiO2 supported catalysts.

Temperature programmed reduction is a powerful mean for the determination of the oxide chemical species. As reported in the literature, the reducibility of cobalt oxide is strongly dependent on the support, loading and promoter effect [9,2628]. In Figs. 68 the TPR proles of the catalysts supported on HMS, MCM-41 and on amorphous silica are shown respectively. According to the literature [29], unsupported Co3O4 contains one or two peaks close to each other in the temperature range of 200400 8C. The reduction process is assumed to occur in two steps, a lower temperature reduction of Co3+ to Co2+ and a slightly higher temperature reduction of Co2+ to Co0. Due to a variety of effects such as particle size, morphology, or insufcient hydrogen partial pressure, the shape of the peaks may not be well dened. As shown in Fig. 6 the reduction prole of Co supported on HMS is characterized by a main feature, split into two peaks, one at 276 8C and the other at 343 8C attributed to the two reduction steps of the Co3O4. The extra peak at high temperature (833 8C) is due to high temperature driven formation of CoOxSiO2 species [28]. The presence of such hard to reduce cobalt species on SiO2 has also been attributed to Co2+ migration into the silica framework [30]. The PdCo catalyst in which the cobalt was added after support impregnation with Pd has a quite different pattern. Indeed, the

Fig. 7. TPR patterns of MCM-41 supported catalysts.

peak at high temperature due to the hard to reduce mixed oxides CoOxSiO2 has disappeared, suggesting that the palladium is inhibiting the strong interaction between the cobalt and the carrier. Moreover, a new peak at low temperature (127 8C) appears. Furthermore in the TPR pattern of the monometallic Pd catalyst (not shown here) in the considered range of temperature (50 900 8C) no hydrogen consumption peak was detected, probably due to the reduction of PdO at lower temperature than the starting of our ramp. Only a negative peak at around 60 8C attributed to the decomposition the Pd b-hydride was observed. The lack of such peak in the TPR patterns of the palladium promoted cobalt catalysts may be attributed, in agreement with Ref. [31], to restricted hydrogen transport through the Co oxide. The peak at low temperature (127 8C) could in principle contain some contribution of the reduction of Pd2+ which the interaction with Co oxide may shift towards high temperature [31]. However since the intensity of the peak is much higher than what expected on the basis of the Pd content, the main contribution to the peak arises from the reduction of the cobalt species shifted to lower temperature. Indeed, in accordance with several studies on noble metal promoted Co catalysts [11,31,32], the effect of palladium can be ascribed to the hydrogen spillover [33]. According to this mechanism the hydrogen would at rst adsorb and dissociate on the noble metal sites producing highly reactive hydrogen species, i.e. H+, H, and H3+, which then spill over to Co oxides reducing them more easily. As shown in Fig. 6 similar TPR pattern is obtained for the tri-element system in which gold and cobalt are co-impregnated on the palladium modied support. Such similarity conrms the negligible effect played by gold [4] as compared to the effect played by palladium on the cobalt oxide reduction. In Fig. 7 the TPR patterns of the Co and the Pd(AuCo) on the MCM-41 are shown. The pattern of Co/MCM-41 catalyst is characterized again by two major features, one at $300 8C and the other above 700 8C [9]. The low temperature peak can be attributed to the unresolved two steps reduction of the Co3O4. The high temperature peak quite broad and intense is attributed to cobalt species strongly interacting with the support. Again, in the case of the palladium containing samples, the high temperature peak disappears and a new peak at low temperature shows up. The TPR patterns of the Co and Pd(AuCo) on amorphous silica, given in Fig. 8, are similar to the patterns of the corresponding samples supported on HMS and MCM-41. However in this case the peak at high temperature, due to the composite species CoOxSiO2, is much smaller as compared to the other supported catalysts. Such

A.M. Venezia et al. / Applied Catalysis A: General 353 (2009) 296304 Table 3 Tmax and hydrogen consumption Va (ml/gcat) of TPR peaks obtained for different samples. Catalyst Tmax I peak SiO2 series Co Pd(AuCo) II peak III peak V I peak II peak III peak Table 4 XPS binding energy values and atomic ratios. Catalyst SiO2 series Co PdCo Pd(AuCo) Pd MCM-41 series Co PdCo Pd(AuCo) Pd HMS series Co PdCo Pd(AuCo) Pd Co 2p3/2 780.6 780.3 780.5 Pd 3d5/2 337.5 337.5 337.9 Au 4f7/2 85.2 Co/Si

301

Pd/Si

261 128

402 251

711

8 8

12 20

0.01 0.01 0.02

0.006 0.006 0.006

MCM-41 series Co PdCo 197 Pd(AuCo) 165 HMS series Co PdCo Pd(AuCo)
a

307 381 369

740 519

8 11

15 12 17

12 6

780.6 780.8 780.4

337.9 337.5 337.9

85.2

0.01 0.02 0.03

0.002 0.002 0.003

276 127 143

343 281 258

833 454 495

12 7 6

6 16 15

8 1 2

The estimated error on the volumes is 10%.

780.8 780.7 780.8

337.5 337.9 337.8

85.2

0.01 0.02 0.02

0.002 0.003 0.004

difference in the TPR patterns has already been reported and according to the literature, higher surface area supports, like MCM41 or HMS are generally more interacting with the supported metal as compared to the lower surface area supports [27,32]. In Table 3 the temperatures of the reduction peaks with the corresponding hydrogen consumption volumes are reported. Within the experimental errors of 10%, mainly related to the integration procedure, the total amount of consumed hydrogen may be considered in accordance with the Co3O4 stoichiometric amount and is fairly similar and independent from the presence of the noble metal. In fact, the possible contribution of palladium and gold reduction in the temperature range considered is negligible. The surface composition and the surface chemical states of the precursor cobalt oxide species were investigated by X-ray photoelectron spectroscopy. In Fig. 9 the Co 2p spectra of the calcined HMS supported catalysts are shown. The spectra, typical of all the analysed Co containing samples, are characterized by the two Co 2p3/2 and Co 2p1/2 spinorbit components separated by 15.5 eV and by a shallow feature in between. By the curve tting routine, the Co 2p3/2 binding energy at 780.5 eV is obtained. The value is intermediate between the binding energies for Co3+ and Co2+ species [34]. In agreement with the XRD data, the spectra are attributed to the Co3O4 phase. The shallow feature is a shake up satellite associated with some Co2+. In Table 4 the XPS results in terms of binding energies and atomic ratios are summarized. The presence of noble metal does not modify the Co 2p spectra. The Pd

3d spectra of all samples are characterized by Pd 3d5/2 component at $337.5 0.2 eV typical of highly oxidized Pd. The Au 4f7/2 binding energy at $85.2 eV is indicative of oxidized gold. As the quantitative XPS analyses are concerned, in spite of the different support texture, no relevant changes of the Co/Si atomic ratios with the type of silica are observed. Generally an increase of this ratio is observed in the multimetalllic catalysts, in accordance with the slight decrease of the cobalt crystallites upon addition of the noble metals. Concerning the Pd/Si atomic ratios, their values are affected by the support texture. Indeed concerning the monometallic Pd catalysts, much smaller values are obtained with the high surface area MCM-41 and HMS supports as compared to the amorphous silica. Such results may be attributed to a preferential localization of the small PdO particles inside the support pores, with the consequent shadowing of the XPS signal. Moreover, in the case of the mesoporous supports, the Pd/Si ratio further decreases upon the successive deposition of the other two elements, probably due to coverage effect. In Fig. 10 the Co 2p spectra of two selected samples, Co/MCM-41 and PdCo/MCM-41 in the calcined state and after H2S/H2 treatment are shown. The spectra of the calcined samples are characteristic of the Co3O4 as described above. In the H2S treated samples the spectra show evidence of two components, one at 778.8 eV and the other one at 781.1 eV. The peak at 778.8 eV can be attributed to cobalt sulde or to cobalt metal, since they are undistinguishable. The binding energy of the Co 2p3/2 level in Co metal is indeed close to that in Co9S8 and both species do not show any characteristic satellite structure [9,35]. However, as already reported for similar systems [9], detection of a broad S 2p signal at $162 eV with the atomic ratio S/Co (considering only the Co 2p component at 778.8 eV) approximately close to 1 suggests formation of the thermodynamically favored Co9S8 [35]. The high energy peak at $781 eV with a rather broad satellite feature in the range of 785 787 eV is attributed to Co(II). The incomplete suldation of the Co species is ascribed to H2S diffusion limitation in the large size Co3O4 particle [36]. 3.2. Catalytic tests In Table 5 the catalytic data in terms of reaction rate and activation energy for the HDS of thiophene and reaction rate for the HDS of tetrahydrothiophene (THT) are summarized. The histograms of the thiophene conversion rate as function of the catalyst supports and catalyst formulation are shown in Fig. 11. A slight increase of the activity is observed in the case of cobalt supported on mesoporous silicas as compared to cobalt on amorphous SiO2. As already published in a previous paper [9] such increase of activity is attributed to the smaller Co3O4 particle size. However the positive effect of the better cobalt dispersion is rather limited

Fig. 9. Co 2p spectra of HMS supported catalysts.

302

A.M. Venezia et al. / Applied Catalysis A: General 353 (2009) 296304

Fig. 10. Co 2p spectra of Co/MCM-41 and PdCo-MCM-41: (a) calcined and (b) after sulfuration.

because it is counter-balanced by a stronger metalsupport interaction taking place in the high surface area silica. Indeed as it was shown previously [9], the cobalt species interacting with the support are also very difcult to sulfurize as compared to the Co3O4 with a consequent loss of activity. With the exception of the amorphous silica supported catalyst, the modication of the support by addition of palladium has a benecial effect on the catalyst activity. The behavior can be explained on the basis of the TPR results. The large reduction peak at temperature >700 8C observed for the Co/MCM-41 and Co/HMS and attributed to cobalt species highly interacting with the supports is absent in the TPR of

the corresponding PdCo catalysts. If their formation is inhibited by the Pd pre-impregnation, an increase of the activity arises. Beside this indirect role played by Pd, the lowering of the reduction temperature of the Co3O4 by the hydrogen spillover would also contribute to the benecial effect played by Pd. In the amorphous silica catalysts, where the interaction between the Co and the support is not so strong to produce a high temperature reduction peak [37] no signicant catalytic effect due to palladium is observed. The introduction of gold has a small positive inuence related to a small decrease of the Co3O4 particle sizes. An exception is given by the HMS supported system where the Pd(AuCo) catalyst is slightly less active as compared to the PdCo catalyst. Within the three series of samples, the monometallic Pd catalysts are

Table 5 Thiophene HDS reaction rate (kTP) and apparent activation energy (Eapp) and tetrahydrothiophene HDS reaction rate (kTHT) of the supported catalysts. Samples kTP (mmole s1 g1) Eapp (kJ/mole1) kTHT (mmole s1 g1)

HMS series Co PdCo Pd(AuCo) Pd MCM-41 series Co PdCo Pd(AuCo) Pd SiO2 series Co PdCo Pd(AuCo) Pd

0.50 0.82 0.70 1.42

34 34 40 19

1.59 2.37 5.25

0.42 0.70 1.00 0.19

27 32 33 33

1.09 2.44 0.93

0.35 0.31 0.44 0.21

35 33 33 30

Fig. 11. Thiophene HDS rate constants of promoted Co catalysts as a function of different catalyst supports.

A.M. Venezia et al. / Applied Catalysis A: General 353 (2009) 296304

303

generally the least active with the exception of the HMS supported one. Indeed Pd/HMS exhibits three times the activity of the corresponding cobalt catalyst. The reason for such high activity is not clear, especially in consideration of the small differences existing between the two mesoporous supports. As shown in Table 1 the HMS has slightly larger pores and also a more acidic ZPC as compared to MCM-41. Moreover as obtained from the FTIR of adsorbed pyridine, the HMS contains a larger number of silanol groups which however are not so acidic to protonate the pyridine. The exposure to acid molecules such as SO2 or NO2 has been reported to increase the acidity of the silanol groups, however H2S was shown to be not sufciently acidic to induce proton transfer of a pre-adsorbed base [38]. Therefore a sort of bifunctional mechanism involving both Pd particles and support silanols, similar to that proposed for Pt supported on aluminum modied MCM-41 [39] is not likely. The surprisingly high activity of the Pd/HMS deserves a deeper investigation which at the moment is beyond the purpose of the present study. Pd catalysts are generally used as hydrogenation catalysts because of their capability to activate the molecular hydrogen. On the other hand cobalt catalysts are suitable for the hydrodesulfurization processes due to the tiophene activation on the sulphur vacancies of the Co9S8 leading to the cleavage of the CS bond [40]. Moreover as shown from the TPR results, the noble metal enhances the cobalt reducibility. As observed in the present study, the combination of the noble metals and cobalt produces a synergetic effect on the enhancement of the HDS activity. It is worth mentioning that the activation energy values in Table 5 are comparable with those obtained recently with other Co-based catalysts [9] and are lower than the values obtained with CoMo or NiMo systems [14,41]. To discriminate between the two HDS routes, either hydrogenation followed by the CS cleavage or direct CS rupture, occurring on the different systems, catalytic tests with tetrahydrothiophene (THT) were carried out on the selected samples of Co, Pd and Pd(AuCo) over the two mesoporous silicas. If the rst route were preferentially followed on Pd and the second on Co, catalytic activity ratio for the HDS of the tetrahydrothiophene of the Pd and the Co catalysts should drastically change because the hydrogenation activity of the Pd would not play any important role. On the other hand, if the direct CS bond rupture was the rate-determining step, no changes would be expected. In Fig. 12 the corresponding reaction rates as a function of the formulation are given. The comparison with the data in Fig. 11 suggests indeed that, under the present experimental conditions, the thiophene hydrodesulfurization

proceeds through the direct CS bond cleavage regardless the catalyst formulation. 4. Conclusion Ordered mesoporous silicas (MCM-41 and HMS) with respect to the amorphous SiO2 improve the dispersion of the supported cobalt oxide particles, slightly enhancing their catalytic activity in the hydrodesulfurization of thiophene. The modication of the support by the addition of palladium increases the reducibility of the supported Co3O4 with consequent increase of the catalytic activity. The co-impregnation of cobalt and gold over the Pd modied silica, further promotes the Co activity, by the combined effect of an increased cobalt reducibility and increased Co3O4 particle dispersion. Comparative tests in the HDS of two substrates, thiophene and tetrahydrothiophene (THT) suggest the direct CS bond cleavage as the preferential route for the thiophene reaction. Acknowledgements Support by European Community, Network of Excellence (NoE) IDECAT (Integrated Design of Catalytic Nanomaterials for Sustainable Production) and COST D36 action is acknowledged. References
[1] H. Topse, B.S. Clausen, F.E. Massoth, in: J.R. Anderson, M. Boudart (Eds.), Hydrotreating Catalysis, Springer-Verlag, Berlin, 1996. [2] A.M. Venezia, V. La Parola, G. Deganello, B. Pawelec, J.L.G. Fierro, J. Catal. 215 (2003) 317. [3] V. La Parola, G. Deganello, A.M. Venezia, Appl. Catal. A 260 (2004) 237. [4] T.A. Zepeda, B. Pawelec, J.L.G. Fierro, A. Olivas, S. Fuentes, T. Halachev, Micropor. Mesopor. Mater. 111 (2008) 157. [5] A. Corma, A. Martinez, V. Martinez-Soria, J. Catal. 169 (1997) 480. [6] M. Hussain, S.-K. Song, J.-H. Lee, S.-K. Ihm, Ind. Eng. Chem. Res. 45 (2006) 536. [7] E. Rodriguez-Castellon, A. Jimenez-Lopez, D. Eliche-Quesada, Fuel 87 (2008) 1195. [8] J.P.R. Vissers, V.H.J. de Beer, R. Prins, J. Chem. Soc., Faraday Trans. 1 (83) (1987) 2145. [9] A.M. Venezia, R. Murania, G. Pantaleo, G. Deganello, J. Mol. Catal. A 271 (2007) 238. [10] M.A. Al-Saleh, M.M. Hossain, M.A. Shalabi, T. Kimra, T. Inui, Appl. Catal. A 253 (2003) 453. [11] M.M. Hossain, M.A. Al-Saleh, M.A. Shalabi, T. Kimura, T. Inui, Appl. Catal. A 278 (2004) 65. [12] M. Hossain, Chem. Eng. J. 123 (2006) 15. [13] A.M. Venezia, R. Murania, G. Pantaleo, G. Deganello, Gold Bull. 40 (2007) 130. ` [14] A.M. Venezia, V. La Parola, V. Nicol, G. Deganello, J. Catal. 212 (2002) 56. [15] J. Choma, S. Pikus, M. Jaroniec, Appl. Surf. Sci. 252 (2005) 562. [16] A. Wang, Y. Wang, T. Kabe, Y. Chen, A. Ishihara, W. Quian, J. Catal. 199 (2001) 19. [17] N. Marin-Astorga, G. Pecchi, T.J. Pinnavaia, G. Alvez-Manoli, P. Reyes, J. Mol. Catal. A 247 (2006) 145. [18] S.J. Gregg, K.S. Sing, Adsorption, Surface Area and Porosity, 2nd ed., Academic Press, San Diego, 1982. [19] JCPDS Powder Diffraction File Int. Centre for Diffraction Data, Swarthmore, 1989, File No. 42-1467. [20] H.P. Klug, X-ray Diffraction Procedure for Polycrystalline and Amorphous Materials, Wiley, New York, 1954. [21] D.A. Shirley, Phys. Rev. B 5 (1972) 4709. [22] P.M.A. Sherwood, in: D. Briggs, M.P. Seah (Eds.), Practical Surface Analysis, Wiley, New York, 1990, p. 181. ` [23] V. La Parola, G. Deganello, S. Scire, A.M. Venezia, J. Solid State Chem. 174 (2003) 482488. [24] E.P. Parry, J. Catal. 2 (1963) 371. [25] G. Busca, Catal. Today 41 (1998) 191. [26] M. Vo, D. Borgmann, G. Wendler, J. Catal. 212 (2002) 10. [27] L.B. Backman, A. Rautiaien, M. Lindblad, O. Jylha, A.O.I. Krause, Appl. Catal. A 208 (2001) 223. [28] G. Jacobs, T.K. Das, Y. Zhang, J. Li, G. Racoillet, B.H. Davis, Appl. Catal. A 233 (2002) 263. [29] R.L. Chin, D.M. Hercules, J. Phys. Chem. 86 (1982) 3079. [30] K.E. Coultier, A.G. Sault, J. Catal. 154 (1995) 56. [31] A. Sarkany, Z. Zsoldos, Gy. Steer, J.W. Hightower, L. Guczi, J. Catal. 157 (1995) 179. [32] J.G. Panpranot, A. Goodwin Jr., Sayari, Catal. Today 77 (2002) 269.

Fig. 12. Tetrahydrothiophene HDS rate constants of promoted Co catalysts on mesoporous supports.

304

A.M. Venezia et al. / Applied Catalysis A: General 353 (2009) 296304 [38] A.A. Tsyganenko, E.N. Storozhevam, O.V. Manoilova, T. Lesage, M. Daturi, J.-C. Lavalley, Catal. Lett. 70 (2000) 159. [39] Y. Kanda, T. Kobayashi, Y. Uemichi, S. Namba, M. Sugioka, Appl. Catal. A 308 (2006) 111. [40] A. Chica, K.G. Strohmaier, E. Iglesia, Appl. Catal. B 60 (2005) 223. [41] S.L. Gonzales-Cortes, S.M.A. Rodulfo-Baechler, T. Xiao, M.L.H. Green, Catal. Lett. 111 (2006) 57.

[33] J.H. Sinfelt, Bimetallic Catalysts: Discoveries, Concepts and Applications, Wiley, NY, USA, 1983. [34] M.A. Stranick, M. Houlla, D.M. Hercules, J. Catal. 103 (1987) 151. [35] Y. Okamoto, T. Imanaka, S. Teranishi, J. Catal. 85 (1980) 448. [36] K. Inamura, T. Takyu, Y. Okamoto, K. Nagata, T. Imanaka, J. Catal. 133 (1992) 498. [37] M.M. Hossain, M.A. Al-Saleh, M.A. Shalabi, T. Kimura, T. Inui, Appl. Catal. A 274 (2004) 43.

Você também pode gostar