Operando XAS and Raman study on the structure of a supported vanadium oxide catalyst during the oxidation of H2S to sulphur

Document downloaded from: This paper must be cited as: The final publication is available at Copyright Additional Information http://dx.doi.org/10.1016/j.cattod.2010.02.050 http://hdl.handle.net/10251/63961 Elsevier Holgado, JP.; Soriano Rodríguez, MD.; Jimenez-Jimenez, J.; Concepción Heydorn, P.; Jimenez-Lopez, A.; Caballero, A.; Rodriguez-Castellon, E.... (2010). Operando XAS and Raman study of the structure of supported vanadium oxide during the oxidation of H2S to sulphur. Catalysis Today. 155(3-4):296-301. doi:10.1016/j.cattod.2010.02.050. 1 Operando XAS and Raman study on the structure of a supported vanadium oxide catalyst during the oxidation of H2S to sulphur J.P Holgado a, M.D. Soriano b, J. Jiménez-Jiménez c, P. Concepción b, A. Jiménez- López c, A. Caballero a, E. Rodríguez-Castellón c, J.M. López Nieto b,* a Instituto de Ciencia de Materiales de Sevilla and Dep. Química Inorgánica (CSIC- Universidad de Sevilla), Sevilla (Spain) b Instituto de Tecnología Química, UPV-CSIC, Campus de la Universidad Politécnica de Valencia, Valencia (Spain). c Dep. Química Inorgánica, Universidad de Málaga, Málaga (Spain) * To whom correspondence: Fax: +34963877809; email: [email protected] 2 Abstract The modification of crystalline phases of a vanadium oxide supported on mesoporous zirconium phosphate during the partial oxidation of H2S to sulphur has been studied by using an operando Raman-GC approach and XAS in reaction conditions. The catalyst, mainly presenting crystalline V2O5, is transformed during the oxidation of H2S at 200ºC, presenting crystals of V4O9, which is identified by the presence of a band at ca. 900 cm-1 in the Raman spectra (using a 785 nm line of an Argon ion laser) and by the presence of a pre-edge at 5469.8 eV (and a main-edge at 5482.2 eV) in XANES spectra. At the same time, it is observed a high conversion of H2S to sulphur (the main reaction product) and SO2 (as minority). Both activity and selectivity depend on the time on stream. In this way, the selectivity to SO2 decreases from ca. 5 to 1% with the time on stream. This changes could be explained on the basis of the nature of V-species: the initial presence of V5+-O-V5+ pairs and the appearance of V5+-O-V4+ pairs at high time on stream. Keyword: Operando, RAMAN, EXAFS, XANES, partial oxidation of hydrogen sulphide, vanadium oxides supported on mesoporous zirconium phosphate. 3 1. INTRODUCTION Supported vanadium oxide catalysts are interesting in a well number of reactions such as the partial oxidation of hydrocarbons or in the selective reduction of NOx [1-5]. In addition to the oxidative dehydrogenation of short chain alkanes [4, 5], an important effort to study these catalytic systems in the partial oxidation of H2S to elemental sulfur has been carried out in the last years [6-10]. At the moment, it is known that the characteristics of the metal oxide support and the nature of vanadium species strongly influence the catalytic performance and stability of supported vanadium catalysts during the partial oxidation of H2S [6-10], although the catalysts could be modified during the reaction depending on the characteristics of catalysts and/or the reaction conditions. In this way, the reducing power of reactants seems to be an important factor in the partial oxidation of H2S over V2O5 catalyst [11]. Thus, it has been reported that a reduced vanadium oxide, achieved by a reduction of V2O5 at 600 ºC for 24 hours, showed a stable reactivity and an excellent conversion of H2S and selectivity for the elemental sulfur. Recently we have observed how catalysts based in vanadium oxide supported on a mesoporous zirconium phosphate heterostructure change depending on the V-loading and the reaction conditions [12]. Accordingly, the nature of active and selective sites working during the reaction is still under discussion. Raman studies have been carried out to determine the structure of vanadium oxide catalysts [13], while X-ray absorption spectroscopy (XAS) have been used to determine the oxidation state and coordination of vanadium-based catalysts [14]. More recently, they have been also used in operando conditions in order to clarify the nature of active and selective sites [15-17]. In the present paper we show for the first time an X-ray absorption spectroscopy (XAS) and Raman study, both working in operando conditions, on the selective 4 oxidation of H2S to sulfur using a vanadium oxide supported on a mesoporous zirconium phosphate heterostructure as catalyst. In addition, the reduction or reoxidation of catalyst after the catalytic test has been also studied. 2. EXPERIMENTAL 2.1. Catalyst preparation A vanadium oxide supported on mesoporous zirconium phosphate (with 12 wt% of V-atoms, named as 12VOx/MZP) has been prepared by the wet-impregnation method of the mesoporous zirconium phosphate (MZP, SBET= 556 m2/g), using an aqueous solution of ammonium metavanadate. The mixture was rotaevaporated until complete dryness, dried overnight at 100ºC and then calcined at 550ºC for 6 h. The resulting catalyst presents a surface area of 90 m2 g-1. The mesoporous zirconium phosphate heterostructure, MZP (SBET= 556 m2/g), was prepared according to a previously reported preparation procedure [18]. For comparison, commercial VOSO4.3H2O (Aldrich), V2O5 and V4O9 have been also used. V2O5 has been prepared from ammonium metavanadate by calcination in air at 550ºC for 1h. A V4O9 sample has been prepared according to the preparation procedure previously described [19]. The sample shows an X-ray power diffraction pattern with lines at 2θ= 13.7, 21.6, 27.7, 28.1, 33.9, 35.5, 35.7, 41.2, 53.2, 56.0º in agreement to that previously reported for pure V4O9 [JCPDS-23-720]. 2.2. Catalyst Characterization The operando FT-Raman spectra were recorded with an “in via” Renishaw spectrometer, equipped with a microscope (Olympus). The samples were excited by the 785 nm line of an Argon ion laser (Spectra Physics Model 171) with a laser power of 5 2.5 mW. Spectra were collected using a backscattering geometry with a 180º angle between the illuminating and the collected radiation. A micro quartz reactor (i.d 6mm, length 200mm) located in a furnace has been used for in situ Raman studies. The furnace is provided with a small hole in order to allow focalization on the sample with the laser spot (spatial resolution of aprox 1.5µm). The reactor was positioned below the microscope objective, working with a special long working distance objective of 50X (WD=10.6mm). The flow of the reactant has been controlled by mass flow controllers using a total flow of 65 ml/min, with a H2S/air/N2 molar ratio of 1.2/5.0/93.8. The reaction has been performed on 50 mg of catalysts powder at a reaction temperature of 200 ºC. Analysis of reactants and reaction products was carried out on-line using gas chromatography and two different chromatographic columns (Molecular Sieve 5A and Porapak T). The gas line after the reactor has been heated to 150 ºC in order to avoid condensation of sulphur. Previous to venting the gases sulphur has been collected on a cold trap. Vanadium K-edge X-ray absorption spectroscopy (XAS) measurements were performed at the beamline X10DA (superXAS) located at the Swiss Light Source (SLS), Villigen, Switzerland, using a Si (110) monochromator. The measurements were performed in transmission mode using ion chambers filled with Ar/N2 as detectors. A modified commercial infrared (Specac) cell, equipped with Kapton windows was used for in operando measurements. Typically, 50 mg of catalyst powder were pressed in a stainless steel sample holder in order to obtain self-supported discs of samples with appropriated transmission properties. Due to formation of sulphur in the reaction, heating of pipes after the cell is necessary, to avoid the condensation of this element. Collection of formed sulphur in the reaction will be done in a cold trap previous to vent the gases into the exhaust. Mass flow controllers were used for dosing the gases to the cell, using a total flow of 100 ml/min, with H2S/O2/He molar ratio of 1/1/98. For 6 comparison, the spectra of catalyst before and after the catalytic tests have been also achieved at room temperature in helium. The XAS spectra were analyzed using the IFEFFIT programs package [20]. A factor of k2 was used for obtaining the Fourier Transforms (FT) of the EXAFS region of the spectra. The catalytic tests were carried out in a fixed-bed tubular reactor at 200ºC and atmospheric pressure, 50 mg of catalyst and a H2S/air/He mixture with a molar composition of 1.25/5.0/93.8 and a total flow of 130 ml min-1. The analysis of reactants and products has been carried out using the same procedure to that described previously for the operando Raman study. 3. RESULTS AND DISCUSSION 3.1 Characterization of catalysts before and after catalytic tests. Fig. 1 shows the XRD patterns (Fig. 1A), and Raman (Fig. 1B) and XANES (Fig. 1C) spectra of supported-catalyst (12VOx/MZP sample) before and after the catalytic test. For comparison, the corresponding spectra of the as-synthesized V4O9 are also included (Fig. 1, patterns c and spectra f and i). We must indicate that the used supported-catalyst was recovered after a time-on-stream 300 min. During the catalytic test the H2S conversion decreased from ca. 80 to 63%, while the selectivity to elemental sulphur was higher than 99%. The XRD patterns of fresh catalysts shows the presence of V2O5 [JCPDS: 41– 1426] and ZrP2O7 [JCPDS: 85–896]. When comparing to those achieved from used catalysts (Fig. 1A, pattern b) and pure V4O9 (Fig. 1A, pattern c), it can be concluded the disappearance of V2O5 and the formation of V4O9 [JCPDS: 23–720] after the catalytic tests [12]. 7 On the other hand, fresh catalyst is characterized by the presence of bands at 998, 706, 530, 489, 410, 305, 289, 203 and 159 cm-1 in the Raman spectrum, which indicate the presence of crystalline V2O5 [13] (Fig. 1B, spectrum d). However the presence of isolated vanadium species on the surface of the support as minority cannot be ruled out. After the catalytic tests the Raman spectrum of catalyst changes (Fig. 1B, spectrum e). Thus, a band at ca. 900 cm-1 is observed in used catalysts, while the main features of the bands related to V2O5 crystallites are not observed. This spectrum is very similar to that achieved with the as-synthesized V4O9 (Fig. 1B, spectrum f), which is also in agreement to one reported by Nilson et al. [21] during the study of the propane ammoxidation over V2O5 and one recently proposed from our group for this type of catalysts during the partial oxidation of H2S [12]. We must indicate that the Raman band at ca. 900 cm-1 characteristic of V4O9 [20] is only observed when excited with radiation of 785 nm while not observed when using an excitation wavelength of 514 nm [22]. This effect, explained by resonance effects in reduced vanadium oxides, makes their characterisation possible by properly selecting the laser excitation wavelength. Similar resonant effects have already been observed by Mestl et al. [23, 24] in partially reduced molybdenum oxide. Figure 1C shows the V K-edges XANES spectra of catalyst before (Fig. 1C, spectrum g) and after the catalytic test (Fig. 1C, spectrum h), as well as the spectrum achieved with the as-synthesized V4O9 (Fig. 1C, spectrum i). For comparison, the preedge features for the XANES spectra are shown in Table 1. According to previous XANES results of V-based compounds [14], it can be concluded that V2O5 is mainly present in fresh catalyst while V4O9 seems to be the main phase in the used catalyst. In this case, the V4O9 phase in the catalyst seems to be more amorphous than the massive oxide phase, according to the more attenuated oscillations observed by XAS. 8 3.2. Raman study on catalysts during the H2S partial oxidation. Figure 2 shows operando Raman spectra and their on-line simultaneous catalytic results for sample 12VOx/MZP during the selective oxidation of H2S with time on stream, TOS, from 0 to 180 min. The Raman spectrum of the sample at 200ºC in helium shows the same bands to that obtained for fresh catalyst (bands at 998, 706, 530, 489, 410, 305, 289, 203 and 159 cm-1) indicating the presence of bulk V2O5 (Fig. 2, spectrum b). However, after some minutes in contact with the reaction mixture, the bands related to V2O5 disappear appearing a band at ca. 900 cm-1 (Fig. 2, spectra c to j). The band at ca. 900 cm-1, observed during all the experiment (after 180 minutes of time on stream), was also observed during the characterization of the catalyst after a catalytic tests in a fixed bed reactor. Thus, and according to that previously indicated, the band at ca. 900 cm-1 can tentatively be assigned to V4O9 [21]. In addition, we can also conclude that this nonstoichiometric vanadium oxide is formed during the partial oxidation of H2S at 200ºC. On the other hand, Figure 2 also shows the catalytic results obtained during the operando Raman-GC measurements. It can be seen that the conversion of H2S decreases from 99 to 90% after 120 min. In this way, we must indicate that a similar behavior is also observed when the catalytic tests were carried out in a fixed bed reactor. In all cases, sulphur has been the main reaction product while SO2 was observed as minority. However, a small difference in the S/SO2 ratio is also observed with the time on stream. Thus, an initial selectivity of ca. 8% was observed for SO2 although it decreases with time on stream until selectivity to SO2 lower than 1% at a TOS of 120 min. In an opposite trend, the initial selectivity to elemental sulphur of ca. 92% increases until 99% after 120 min of time on stream. 15 12) M.D. Soriano, J. Jiménez-Jiménez, P. Concepción, A. Jiménez-López, E. Rodríguez-Castellón and J.M. López Nieto, Appl. Catal. B. 92 (2009) 271. 13) B.M. Weckhuysen, J.-M. Jehng and I.E. Wach, J. Phys. Chem. B 104 (2000) 7382. 14) G. Silversmit, J.A. van Bokhoven, H. Poelman, A.M.J. van der Eerden, G.B. Marin, M.F. Rayniers and R. De Gryse, Appl. Catal. A 285 (2005) 151. 15) G. Mul, M.A. Bañares, G. Garcia Cortez, B. van der Linden, S.J. Khatib and J.A. Moullijn. Phys. Chem. Chem. Phys. 5 (2003) 4378. 16) A. Bruckman, Chem. Commun. (2005) 1761. 17) S.T. Tinnemans, J.G. Mesu, K. Kervinen, T. Visser, T.A. Nijhuis, A.M. Beale, D.E. Keller, A.M.J. van der Eerden and B.M. Weckhuysen, Catal. Today, 113 (2006) 3. 18) J. Jiménez-Jiménez, M. Rubio-Alonso, D. Eliche-Quesada, E. Rodríguez-Castellón and A. Jiménez-López, J. Mater. Chem. 15 (2005) 3466-3472. 19) T. Miyake, T. Doi, Appl. Catal. A 131 (1995) 43. 20) M.Newville, J.Synchrotron Rad. 8 (2001) 322. 21) R. Nilsson, T. Limblad and A. Andersson, J. Catal. 148 (1994) 501. 22) Y. Zang, M. Meisel, A. Martin, B. Lücke, K. Witke and K.-W. Brzeinka, Chem. Mater. 9 (1997) 1086. 23) G. Mestl, J. Raman Spectrosc. 33 (2002) 333. 24) M. Dieterle, G. Weinberg and G. Mestl., Phys. Chem. Chem. Phys. 4 (2002) 812. 25) D. C. Koningeberger and R. Prins, X-Ray Absorption: Principles, Applications, Techniques of EXAFS, SEXAFS and XANES; Wiley: New York, 1988. 26) G. Grymonprez, L. Fiermans amd J. Vennik, Acta Cryst. A 33 (1977) 834. 27) H. Kwon, S. Choi and L. T. Thompson, J. Catal. 184 (1999) 236. 16 Caption to Figures Fig. 1. XRD (A), Raman (B) and XANES (C) results of fresh (a, d and g) and used (b , e and h) supported vanadium oxide catalyst (12VOx/MZP). For comparison, the corresponding spectra of the as-synthesized V4O9 are also included (c, f and i). Symbols: V2O5 (■); ZrP2O7 (○); V4O9 (▲); elemental sulfur (□). Fig. 2. Operando Raman spectra (left) of 12VOx/MZP catalyst during partial oxidation of H2S and the corresponding catalytic results, i.e. H2S conversion (■) and selectivity to elemental sulphur (□) (right). Spectra: a) at r.t. in He; b) at 200ºC in He; c) during reaction at a TOS of 15 min; d) during reaction at a TOS of 30 min; e) during reaction at a TOS of 45 min; f) during reaction at a TOS of 60 min; g) during reaction at a TOS of 75 min; h) during reaction at a TOS of 90 min; i) during reaction at a TOS of 120 min; j) during reaction at a TOS of 180 min. Reaction conditions: 50 mg of catalyst, total flow 65 ml min-1; H2S/air/He molar ratio of 1.2/5.0/93.8. Fig. 3. Normalized absorption spectra (XANES) for the 12VOx/MZP catalyst at different time on stream (A). For comparison, the enlarged pre-edge region of V-O bond is also included (B). Spectra: at room temperature in He (a); at 200ºC in He (b); spectra in reactions conditions at a TOS (in min) of 15 (c), 30 (d), 45 (e), 60 (f), 75 (g), 90 (h), 120 (i). Fig. 4. . Normalized absorption spectra (XANES) for the V2O5 catalyst at different time on stream (A). For comparison, the enlarged pre-edge region of V-O bond is also included (B). Spectra: at 200ºC in He (a); spectra in reactions conditions at a TOS (in min) of 15 (b), 30 (c), 45 (d), 60 (e), 75 (f), 90 (g). 17 Fig. 5. Fourier transforms of the EXAFS region obtained for the 12VOx/MZP catalyst before (dashed line) and after a treatment in reaction conditions at a TOS of 120 min (solid line). Reaction conditions as in Fig. 3. Fig. 6. Fourier transforms of the EXAFS region obtained for the V2O5 (A) and V4O9 (B) catalyst before (dashed line) and after a treatment in reaction conditions at a TOS of 120 min (solid line). Reaction conditions as in Fig. 3. 18 Table 1. Pre-edge and main peak observed in XANES spectra of 12VOx/MZP catalysts and references. Sample Treatment 1 pre-edge peak position (eV) Main-edge position (eV) Δ (eV) 2 V2O5 He/rt 5470.6 5483.7 13.1 V 4 O 9 He/rt 5469.8 5482.2 12.4 VOSO 4 He/rt 5469.8 5480.1 10.3 12VOx/MZP He/rt 5470.4 5483.2 12.8 12VO x /MZP He/200ºC 5470.4 5482.9 12.5 12VO x /MZP RC, after a TOS of 120 min 5469.9 5482.1 12.2 1) Treatment of samples: in Helium at room temperature (rt) or at 200ºC; in reaction conditions (RC) as in Figure 2. 2) Difference between main-edge position and pre-edge peak position.