Preparation, characterization and catalytic activity of WO₃ supported on sulfated tin oxide catalysts

2.1 Materials and catalyst preparation

Tin oxide gel was prepared by a dropwise addition of ammonia solution (15 wt.%) to 0.5 M solution of SnCl₄·5H₂O (Riedel-de Haen), up to a final pH value of 8, with continuous vigorous stirring for 4 h. The gel was left overnight then washed by decantation with 1% ammonium acetate (Matsuhashi et al., 1990; Khder et al., 2008) until all chloride ions were eliminated (silver nitrate test). The resulting tin hydroxide gel was mixed vigorously for 4 h with an appropriate amount of 1 M H₂SO₄ solution giving the final ${\text{SO}}_4^{2-}$ of 15 wt.%. The obtained sulfated tin hydroxide gel was dried at 120 °C for 24 h. The dried sample was designated as SS.

To load the sulfated tin hydroxide material with WO₃, appropriate amounts of ammonium paratungstate (APT) solution (30 g/L) (Prolabo) was added to an accurately weighed dried SS material, the resulted gels were vigorously stirred for 4 h, then dried at 120 °C overnight.

The resulting prepared catalysts were calcined at 400, 500, 650 and 800 °C for 4 h. The samples were designated as xWSS-y, where x is the weight percent of WO₃ loaded and y is the temperature of calcination; for example, 25 wt.%WO₃ loaded on 15 wt.% ${\text{SO}}_4^{2-}$ /SnO₂ calcined at 500 °C is expressed as 25WSS-500.

2.2 Techniques

Thermal analysis (TG–DTA) of uncalcined samples was performed in air atmosphere using Linseis Thermal Analyzer, Type STA PT-1600 (Germany). A weight of 40–60 mg was placed in a crucible of 100 μL capacity, and the heating was programmed from 25 °C up to 1000 °C at a constant heating rate of 10°/min.

The X-ray powder diffractograms, of all the prepared samples calcined at 800 °C as well as for all calcination products of 25WSS catalysts, were recorded on Philips Diffractometer Type PW (1830). The pattern was obtained with Ni-filtered CuKα radiation (λ = 1.5418 Å) at 40 kV and 30 mA, at a scanning speed of 1° in 2θ, the scanning range 2θ of 20–70°. The spacing d corresponding to 2θ of the peaks were calculated and correlated with those of ASTM to determine the existing phases. The average crystallite size d of the strongest peaks at 2θ = 26.15° (110), 33.49° (1 0 1) and 51.35° (2 1 1) were determined from Scherrer equation (Ristic et al., 2002):

The specific surface areas of all the samples were determined by nitrogen physisorption at 77 K, applying BET method, using Gemini III 2375 Surface Area Analyzer apparatus. Prior to any adsorption measurement, the sample was degassed at 200 °C for 6 h under a reduced pressure of 10⁻⁴ Torr.

The FT-IR spectra of the samples were recorded using Jasco FT-IR-460 spectrophotometer in the range of 1200–1700 cm⁻¹, at a resolution of 4 cm⁻¹; by mixing 0.005 g of the sample with 0.1 g KBr in 13 mm diameter self supporting disks were used.

The total acidity of the solid samples was measured by means of potentiometric titration method (Rao et al., 2006; Bennaradi et al., 2007). The dry solid (0.1 g) was suspended in 10 mL acetonitrile (Lab-Scan), and agitated for 3 h. The suspension was then titrated with 0.1 N n-butylamine (Merck) at 0.05 mL/min. The electrode potential variation was measured with Inolab Digital pH-mV model using a double junction electrode. The nature of acid sites presented on the surface of the catalyst was determined with FT-IR transmission spectra of adsorbed pyridine (Scharlau). Prior to the pyridine adsorption (Khder et al., 2008; Khder and Ahmed, 2009), the samples were degassed at 200 °C for 3 h under high vacuum followed by suspending in a dried pyridine. Then, the excess pyridine was removed by evaporation at70 °C. A mixture of 0.005 g sample with 0.1 g dried KBr was pressed into 13 mm disk, and the FT-IR spectra of the samples were recorded using Jasco FTIR-460 spectrophotometer within the range of 1200–1700 cm⁻¹, at a resolution of 4 cm⁻¹.

2.3 Determination of the catalytic activity

The catalytic activity of the prepared catalysts was tested for the esterification of propionic acid (Merck) with n-butanol (SRL). The esterification reaction was carried out in a 100 mL flat-bottomed flask, equipped with a reflux condenser, containing a stirring mixture of propionic acid (0.05 M), n-butanol (0.10 M) and the catalyst (0.2 g). The stirring rate in all the experiments was 600 rpm, at 90 °C for 4 h. The reaction mixture was then immediately filtered and quenched to stop the reaction. Liquid samples of 0.5 mL volume were withdrawn and the amount of unreacted acid was titrated with 0.1 N NaOH. The effects of reaction temperature, catalyst amount, the initial molar ratio between the acid and alcohol (PA:B), WO₃ content and calcination temperature were all studied.

3.1 Thermal analyses

Figs. 1 and 2 illustrate the TG and DTA curves of the original prepared samples. All TG curves for all samples show three characteristic steps at 50–200, 430–550 and 560–760 °C with weight loss of 3.4–5.6 wt.%, 7.7–9.4 wt.% and 0.9–4.2 wt.%, respectively. These steps accompanied in DTA curves with endothermic peaks that observed at 110–135, 460–500 and around 630 °C, respectively. These thermal effects are due to desorption of physically adsorbed water, partial dehydroxylation and decomposition of sulfate species bounded to the tin oxide surface, respectively (Gutierrez-Baez et al., 2004; Hino et al., 2007; Khder et al., 2008).

Close

Figure 1

TG curves of SS and WSS catalysts.

Export to PPT

Close

Figure 2

DTA curves of SS and WSS catalysts.

Export to PPT

In addition to the above thermal effects, the TG patterns of SS, 5WSS and 15WSS samples show a gradual step in the range of 200–400 °C, which appear in DTA curve as an endothermic peak located at 229 °C for SS sample and shifted to a higher temperature (∼276 °C) with an increase in wt.%WO₃ loading. This thermal effect is due to partial dehydroxylation and removal of some ammonia as a result of partial decomposition of tungstate species. The TG curves of the samples that loaded with ⩾25 wt.%WO₃ show two steps at 200–300 and 300–400 °C with weight loss of 2.3–2.4 wt.% and 2.2–4.0 wt.%, respectively. These steps are accompanied in DTA with endothermic effects, which their peaks are located at 280–288, ∼340 and 368–380 °C which correspond to the decomposition of APT and evolution of ammonia, and formation of amorphous ammonium tungsten bronze (ATB) (French and Sale, 1981; Khder and Ahmed, 2009). The DTA curves of 35WSS and 45WSS samples show two exothermic effects at 410 and 425 °C attributed to the decomposition of ATB and the formation of crystalline WO₃, respectively (Khder and Ahmed, 2009; Fait et al., 2008). The thermal effects at 270–370 °C may be due to two effects, the first is endothermic which is resulted from the loss of water molecules and dehydroxylation, and the second effect is exothermic as consequence of tin oxide crystallization (Gutierrez-Baez et al., 2004), the latest effect does not appear because the endothermic effect is more pronounced. It was proved that the interaction between sulfate ions and tin oxide stabilizes the crystal size against sintering and inhibiting the crystallite aggregate (Gutierrez-Baez et al., 2004). Moreover, the addition of ATP to sulfated SnO₂ was also found to hinder the crystallization of SnO₂ (Khder and Ahmed, 2009). These were proved in this work, with no sharp exothermic signals observed indicating a distinct crystallization process of WO₃ phase, which may have resulted from the high dispersion of WO₃ on the SnO₂ surface.

Table 1 shows the values of TG% at different calcination temperatures. It is clear that the TG% depend on the percentage of WO₃ loading which shows similar trend on decreasing at all calcination temperatures up to a certain wt.% of WO₃ then increases, whereas, at 500 °C and 600 °C it is decreased then increased. These trends may be due to the variation in the amount of chemically bonded water, the decomposition of ATP and sulfate ions.

3.2 X-ray diffraction

The X-ray diffractograms of SS and xWSS samples calcined at 800 °C are shown in Fig. 3. All samples show peaks at 2θ = 26.15°, 33.49° and 51.35° which may correspond to tetragonal cassiterite structure (JCPDS No. 41-1445). The intensities of these peaks are gradually decreased, while their widths increased progressively with the increase in the percentage of WO₃ loaded. The characteristic peaks of WO₃ phase at 2θ = 23.12°, 23.59° and 24.38° (JCPDS No. 46–1096) were only observed for samples loaded with ⩾25 wt.%WO₃, at lower loading percentage, however, these peaks do not appear which mean that WO₃ crystals transformed into very small WO₃ crystals or dispersed completely over SS support.

Close

Figure 3

XRD patterns of xWSS catalysts calcined at 800 °C.

Export to PPT

The effect of calcination temperature on the crystallinity of 25WSS sample is shown in Fig. 4. The products calcined at 400 and 500 °C showing amorphous structure with no diffraction lines that characterize WO₃ were observed, whereas, products calcined at temperature ⩾650 °C, exhibit characteristic peaks of cassiterite structure that increase in intensities with increasing the calcination temperature.

Close

Figure 4

XRD patterns of 25WSS catalyst calcined at different temperatures.

Export to PPT

The calculated average crystallite size of the peaks appeared at 26.15°, 33.49° and 51.35° that correspond to 110, 101 and 211 planes are listed in Table 2. The results show that SnO₂-800 phase existed as nano-crystalline phase, measuring the crystallite size indicates a value of 23.06 nm for 110 planes and 17.30 nm for 101 and 211 planes. Increasing the loading of sulfated tin oxide with 5–25 wt.%WO₃ leads to a decrease in tin oxide crystallinity, further increase of WO₃ content above 25 wt.% leads to an increase in the crystallite size which may be due to the formation of detectable WO₃-like species.

3.3 FT-IR spectra

The FT-IR spectra of the original samples and the 400 °C products of xWSS are shown in Figs. 5 and 6, respectively. The spectra show vibration bands at 588 and 659 cm⁻¹ which are assigned to the stretching vibration of O–Sn–O bonds in the framework (Zhang and Gao, 2004; Salavati-Niasari et al., 2010; Sarkar et al., 2010). The bands at 800–806 cm⁻¹ are assigned for stretching vibrations of O–W–O groups, with shoulders appeared at 753 cm⁻¹ assigned for the W–O vibration of the surface tungstate species (Armendairiz et al., 2003), and the 960 cm⁻¹ bands are assigned for the stretching vibration of octahedrally coordinate species on the support surface (Khder and Ahmed, 2009). The spectra show that bands at 1260–900 cm⁻¹ are attributed to chelating bidentate sulfate on the SnO₂ support (Yamaguchi et al., 1986). Peaks at 982 and 1029 cm⁻¹ are assigned to asymmetric stretching frequencies of partially ionized S⚌O and S–O bonds, the bands at 1259 cm⁻¹ are assigned to stretching vibration of S–O bonds, and bands at 1394 cm⁻¹ are assigned for asymmetric stretching vibration of S⚌O at sulfate ion bonded to the tin oxide surface (Sohn et al., 2006). The bands appeared at 1618–1628 cm⁻¹ are attributed to the bending vibration of water (Cui et al., 2007). The broad bands and shoulders that appeared at ∼3030, 3136–3149 and 3406–3423 cm⁻¹ are attributed to the stretching vibrations of water and OH groups, that coexist with the bending vibration bands of adsorbed water at ∼1618 cm⁻¹ (Salavati-Niasari et al., 2010; Triwahyono et al., 2003), which are all decreased when the samples calcined at 400 °C. This results is in agreement with TG-DTA results, since the first step in TG curves and first endothermic peak in DTA curves are attributed to the desorption of water molecules.

Close

Figure 5

FT-IR spectra of the uncalcined xWSS catalysts.

Export to PPT

Close

Figure 6

FT-IR spectra of xWSS catalysts calcined at 400 °C.

Export to PPT

Fig. 7 shows the FT-IR spectra in the range of 400–2000 cm⁻¹ for 25WSS catalyst calcined at different temperatures. The spectra reveal that the raising of the calcination temperature to values ⩾650 °C results in the disappearance of the characteristic bands of S–O and S⚌O which appear at 1000–1500 cm⁻¹, presumably due to the decomposition of sulfate ions bonded with SnO₂, which is also in agreement with thermal analysis results.

Close

Figure 7

FT-IR spectra of 25WSS catalyst calcined at different temperatures.

Export to PPT

3.4 Specific surface area

The specific surface areas of the prepared catalysts were determined from the analysis of nitrogen adsorption data performed at 77 K according to the Bruner–Emmett–Teller method (BET method). The calculated S_BET values are presented in Table 3, in which by examination illustrate the following points: (i) Thermal treatment at 400 °C causes an increase in S_BET for SS and xWSS catalysts, the highest increase was observed for 25 wt.%WO₃ loading. (ii) The increase of the calcination temperature from 400 to 500 °C was accompanied with a slight decrease in the specific surface area. Further increase of the calcination temperature causes a drastic decrease in S_BET for SS sample which is due to the decomposition of sulfate species, crystallization and sintering of SnO₂, but for WO₃ loaded samples the decrease in S_BET becomes less as the WO₃ content increased, except for the sample loaded with 45 wt.%WO₃ which shows an increase in S_BET up to 500 °C then decreased slightly with further increases in the calcination temperature. (iii) The increase of WO₃ loading causes a gradual decrease in the surface area for 400 °C and for 500 °C products, whereas for products prepared at ⩾ 650 °C the S_BET increased gradually with increasing WO₃ content to show maximum at 25 wt.%WO₃ loading. Increasing the loading to 35 wt.%WO₃ exhibits a decrease in S_BET value which then increased again for 45 wt.%WO₃ loading.

The loaded WO₃ forms a surface over layer that reduces the surface diffusion of SnO₂ and inhibits sintering (Iglesia et al., 1993) which leads to an increase in the surface area. For WO₃ loading at ⩾ 35 wt.%, the formation of crystalline WO₃ phase leads to a decrease in the surface area (Gregoio and Keller, 2004).

3.5 Surface acidity measurements

The surface acidity measurements of the prepared catalysts by means of potentiometric titration with n-butylamine (Bennardi et al., 2007; El-Sharkawy et al., 2007) were performed to estimate the amount of acid sites and their relative acid strengths. n-Butylamine is a strong base and can be adsorbed on to the acid sites of different strengths and types, thus it titrates both Lewis and Brønsted sites (Tanabe, 1981). On the other hand, the acid strengths of these sites may be classified according to the following scale (Cid and Pecchi, 1985; Sharma et al., 2004): E_i > 100 mV indicates very strong acid sites, 0 < E_i < 100 mV indicates strong acid sites, −100 < E_i < 0 indicates weak acid sites and E_i < −100 mV indicates very weak acid sites. Representative titration curves obtained for some catalysts are shown in Fig. 8. The computed amounts of the acid sites (mmol/g) as well as the values of E_i are listed in Table 3.

Close

Figure 8

Potentiometric titration curves of some selected catalysts.

Export to PPT

The results of the acidity measurements indicate that the strengths of the acid sites and the total surface acidity are both depend on calcination temperature and wt.%WO₃ loading. The results reveal the following points: (i) SS and xWSS samples possess very strong acid sites, with E_i values in the range of 107–525 mV, except for SS-800 sample which possesses strong acid sites, E_i = 75 mV. (ii) The acid site strengths of SS and 5WSS samples are maxima for 500 °C products, whereas for the other catalysts the acid site strengths are decreased with an increase in the calcination temperature. (iii) The amount of the surface acidity show maxima at 400 °C for SS, 5WSS and 35WSS samples, whereas for 15WSS and 25WSS samples exhibit minimum acidity values at 500 °C, and in the same extend the surface acidity of 45WSS samples decreased gradually with raising the calcination temperature. (iv) Loading SS with 5 wt.%WO₃ for 400 and 500 °C products causes an increase in the surface acidity; further increase of WO₃ loading up to 25 wt.% causes a decrease of the surface acidity, which then shows maximum acidity value with 35 wt.%WO₃ loading. (v) The amounts of surface acidity values for 650 and 800 °C products show maximum for 25 wt.%WO₃ and 15 wt.%WO₃ loading, respectively.

The FT-IR spectra of pyridine adsorbed on the prepared catalysts with different WO₃ loadings that calcined at 500 °C as well as for 25WSS catalysts calcined at different temperatures are shown in Figs. 9 and 10, respectively. The spectra show absorption bands at 1400–1440 cm⁻¹ for adsorbed pyridine by hydrogen bonds (Maksimov et al., 2000; Pizzio et al., 2003). The absorption bands appeared at 1458, 1509, 1562, 1580 and 1620 cm⁻¹ are attributed to the pyridine adsorbed on Lewis acid sites (Tanabe, 1981; Villabrille et al., 2002; Rao et al., 2006). The bands at 1544, 1640 and 1650 cm⁻¹ are due to the pyridine adsorbed on Brønsted acid sites with the formation of pyridinum ion (Tanabe, 1981; Villabrille et al., 2002; Pizzio et al., 2003; Khder and Ahmed, 2009). The entire spectra show a characteristic peak at 1490 cm⁻¹ attributed to the adsorption of pyridine on Brønsted or/and Lewis acid sites (Khder and Ahmed, 2009; Tyagi et al., 2009). These results revealed that all the samples contain both of Lewis and Brønsted acid sites at different ratios depending on the WO₃ content as well as on the calcination temperature. The same results show clearly that the intensities of the bands at 1544 and 1490 cm⁻¹ are the highest for 25WSS catalyst. This sample was chosen hereafter to study the effect of calcination temperatures (Fig. 10). It is clear that the intensities of the bands at 1544 and 1640 cm⁻¹ which attributed to pyridine adsorbed on Brønsted sites as well as the bands at 1490 cm⁻¹ are maxima for the products calcined at 400 and 500 °C, also the bands at 1460 and 1580 cm⁻¹ which ascribed the pyridine adsorbed on strong Lewis acid sites are maxima for the products calcined at 500 °C.

Close

Figure 9

FT-IR spectra of pyridine adsorbed on xWSS catalysts calcined at 500 °C.

Export to PPT

Close

Figure 10

FT-IR spectra of pyridine adsorbed on xWSS catalyst calcined at different temperatures.

Export to PPT

The generation of strong Lewis acidity by sulfate species may be due to the presence of surface sulfur complex which has a covalent S⚌O bond that acts as electron-withdrawing species followed by the inductive effect. Thus, the inductive effect of S⚌O makes the Lewis acid strength of Sn⁴⁺ stronger (Furuta et al., 2004), (see structure A in Scheme 1). In addition, the presence of water or OH-surface groups on the catalyst give Brønsted acid sites, and the adsorption of water molecules would convert Lewis acid sites into Brønsted acid sites. Furthermore, the addition of WO₃ to the support (SS catalyst) followed by calcination, causes changes in the textural and acidic properties that lead to tungsten-support interaction. This interaction could occur by different ways. For example, WO₃ could replace sulfate groups and interact directly with SnO₂ surface and/or interact with sulfate groups, as illustrated in structures B and C in Scheme 1 (Khder and Ahmed, 2009). In view of the TG curves, Fig. 1, the final weight loss at temperatures between 500 and 800 °C is due to the decomposition of sulfate species that leads to the ultimate formation of SO₃ moieties (Xia et al., 2002; Ahmed et al., 2008). The WO₃ content clearly affects the decomposition of sulfate ions, increasing the WO₃ loading decreases the percent weight loss. The same result was proven by the surface acidity and S_BET values of the catalysts, the samples with higher WO₃ content (⩾35 wt.%WO₃) or higher calcination temperatures (⩾ 650 °C) show loss of acidity and S_BET that is probably due to the agglomeration of crystalline WO₃ on the surface.

Close

Scheme 1

Possible structures of the catalyst surface.

Export to PPT

3.6 Catalytic activity

The esterification of propionic acid with n-butanol is an electrophilic substitution reaction, and it is extremely slow since the rate is dependent on the autoprotolysis of the carboxylic acid (Lilja et al., 2002), and for the acceleration acid catalyst is generally utilized. The esterification reaction has been recorded to be a straight forward reaction that is subjected to general Brønsted acid catalysts (Sharma et al., 2004; Lilja et al., 2005; Ali et al., 2007), and in some cases recorded to be catalyzed by Lewis acid sites (Barbosa et al., 2006; Rao et al., 1979), or by both of Brønsted and Lewis acid sites (Wang and Li, 2004; Samantaray and Parida, 2003; Khder et al., 2008).

In general, the esterification reaction is dependent on: (i) temperature of reaction, (ii) catalyst amount, (iii) reactant proportion, (iv) stirrer speed and (v) operating pressure (Rao et al., 1979). During the course of this research, optimization of conditions i.e. reaction temperature, catalyst amount, reactant molar ratio, the content of WO₃ in the catalyst and calcination temperature were all presented hereafter for the esterification of propionic acid with n- butanol over the prepared catalysts. All reactions were performed previously at optimized stirrer speed of 600 rpm (Ali et al., 2007).

3.6.4

3.6.4 Effect of WO₃ content and calcination temperature

The effect of WO₃ loading on 15 wt.%SO₄/SnO₂ catalyst was studied for the catalysts calcined at 400 and 500 °C. The reactions are performed for acid/alcohol molar ratio of 1:2 using 0.2 g catalyst at 90 °C reaction temperature for 4 h, and the resulting data are listed in Table 5 and represented in Fig. 11. The results can be summarized as follows; (i) SS catalyst shows high catalytic activity, this activity decreases slightly from 76.72% to 65.08% upon raising the calcination temperature from 400 to 500 °C, respectively. The high catalytic activity is due to the presence of very strong Brønsted acid sites. (ii) The catalytic activities of xWSS catalysts for 400 °C products are higher than those of 500 °C products. (iii) The xWSS catalysts calcined at 400 °C show a slight decrease in the percentage conversion of propionic acid from 77.69% to 71.38% when the wt.%WO₃ loading increases from 5 wt.% to 15 wt.%. Further increase in WO₃ loading to 25 wt.% causes an increase in the conversion to 83.02%, with a slight change for more WO₃ loading. Furthermore, catalyst prepared at 500 °C and loaded with 5 wt.%WO₃ causes a slight increase in the conversion, but further increase in WO₃ loading causes a gradual decrease in the conversion of propionic acid to show minimum at 35 wt.% loading. (iv) The conversion of propionic acid over 25WSS catalysts that calcined at different temperatures (Fig. 12) always decreases with the raise of calcination temperature up to 650 °C to attain 1.55%, with no more change observed after. The low catalytic activity for ⩾ 650 °C products is due to the loss of Brønsted acid sites, the decomposition of sulfate and the sintering effect. The above results indicate that the esterification reaction of propionic acid with n-butanol is catalyzed mainly by Brønsted acid sites and to a lesser extent by Lewis acid sites.

Table 5 Effect of WO₃ loading on the conversion of propionic acid after 4 h (PA:B = 1:2; catalyst weight = 0.2 g; reaction temperature = 90 °C).

Products calcined at 400 °C		Products calcined at 500 °C
Catalyst	Con.%	Catalyst	Con.%
SS	76.72	SS	65.08
5WSS	77.69	5WSS	70.41
15WSS	71.38	15WSS	50.04
25WSS	83.02	25WSS	38.40
35WSS	77.20	35WSS	29.18
45WSS	81.57	45WSS	35.01

Close

Figure 11

Effect of WO₃ content on catalytic activity over xWSS catalysts calcined at 400 and 500 °C.

Export to PPT

Close

Figure 12

Effect of calcination temperature on the catalytic activity over 25WSS catalyst.

Export to PPT

The prepared xWSS catalysts especially that calcined at 400 °C are more efficient for esterification of propionic acid with n-butanol than other reported solid acid catalysts. For example, the formation of n-butyl propionate over cation exchange resins (Amberlyst) and zeolite HZSM-5 catalysts (Liu and Tan, 2001) attains 30–56% conversion, whereas applying the same reaction conditions (PA:B = 1:1, reaction temperature 110 °C for 4 h) using 25WSS-400 catalyst prepared during the course of this research attains 66.61% conversion. Also, the formation of n-butyl propionate over heteropolyacid supported onto natural alumina (Sharma et al., 2004) measures 26–33% conversion, whereas over 25WSS-400 catalyst attains 41.76% conversion at the same reaction conditions (PA:B = 2:1, reaction temperature of 80 °C for4 h).