2.1 Catalysts preparation

2.2 Characterization of the catalysts

Chemical compositions were determined by analysis of the solutions obtained by attack of the solids by a mixture of concentrated fluorhydric and perchloric acid, then by nitric acid. Na content was obtained by atomic absorption, Ti content by colorimetry and S by gravimetry.

The specific surface areas, pore and micropore volumes of the samples were determined by nitrogen physisorption at −196 °C using a Micromeritics ASAP 2000 instrument. The samples were out gassed in vacuum during 5 h at 200 °C prior to nitrogen physisorption.

The X-ray diffraction patterns were recorded on a CGR theta 60 instrument using monochromatized CuKα radiation at λ = 1.54 Å. The position of the d₀₀₁ line in X-ray diffraction gave the interlayer distance and consequently the pillar height.

The X-ray photoelectron spectra (XPS) were performed at room temperature with a Surface Science Instruments SSX-100 model 206 spectrometer with a monochromatised AlKα source, operating at 10 kV and 12 mA. Samples calcined at 400 °C for 3 h were compressed in a small cup under a 5 kg/cm² pressure for 30 s and supported on a holding carousel. The binding energies (BE) of W4f, V2p, S2p and Ti2p lines were referenced to the C1s band at 284.8 eV.

Temperature-programmed desorption of ammonia (NH₃-TPD) was performed in a micro-reactor system with an on-line mass spectrometer (GC–MS) to examine the surface acidity of the catalysts. In each experiment and before adsorption of ammonia a sample weight about 0.150 g calcined and at 400 °C was placed in the cell, evacuated at 400 °C for 2 h under a flow of helium and then cooled to ambient temperature. Pure ammonia gas (50 mL/min) was adsorbed at 100 °C for 20 min. Then after sweeping NH₃ by He (50 mL/min) and stabilization of the base line of the detector, the temperature was progressively increased up to 400 °C at a constant rate of 10 °C/min. The desorbed NH₃ amount was collected and trapped in a boric acid solution (20 g/L), then titrated by sulphuric acid (5 × 10⁻³ M) for quantitative determination.

Temperature programmed H₂-TPR was performed to examine the redox ability of the catalyst. TPR was measured by monitoring hydrogen consumption from a hydrogen dilute in inert gas mixture (5% H₂/He) while increasing the temperature of sample at a constant rate 10 °C/min and maintaining a constant flow rate (60 mL/min). The inlet and outlet gas composition were measured using a quadrupole mass spectrometer QMC 311 Balzers coupled to the system.

2.3 Catalytic activity measurement

The SCR experiments were carried out in a tubular continuous-flow fixed-bed reactor between 100 and 400 °C operating under atmospheric pressure, with a GHSV 46.000 h⁻¹. The total flow rate was controlled by flow meter (100 mL/min). The gas mixture was: NO (0.1 vol.%), NH₃ (0.1 vol.%) and O₂ (2.5 vol.%) in He. The catalyst amount of 0.150 g calcined at 400 °C and the space velocity were kept constant for all experiments. Each catalyst was heated for 1 h at 100 °C under helium before adding the charge. All reported conversions were measured at steady state, under our experimental conditions, after 1 h at a given temperature. They were analyzed using a quadrupole mass spectrometer (QMC 311 Balzers).

3.1 Characterization of the catalysts

The STi-PILC, used as a support for vanadia and tungsten, exhibited a high basal spacing of 22 Å with a large surface area of 226 m²/g and pore volume of 0.170 cm³/g significantly higher than those of the initial clay (Na⁺-Mont.) indicating a successful pillaring process (Table 1). The chemical compositions of the initial clay and the sulphated Ti-pillared clay support are also reported in Table 1. The increase of TiO₂ and sulphur contents at the detriment of Na₂O content results of pillaring and indicates the incorporation of sulphated Ti-polycations between the silicate layers upon pillaring. Table 2 clearly shows that a significant loss of surface area and micropore volume occurs upon vanadia addition. This effect can be explained by the blocking of small pores due to the building up of the vanadia layer. The basal spacing is not significantly affected by impregnation of the support with the vanadium solution [13]. When tungsten and vanadium were deposited at the same time, the surface area also decreased as sharply as the catalyst with only vanadium. The comparison of the thermal stability of the samples was performed under similar operating conditions (Table 2). For STi-PILC, the BET surface area, total pore volume and micropore volume decreased with the increase of the thermal treatment temperature but remained appreciable after calcinations at 400 °C. The decrease of the BET surface area is mainly due to the dehydratation and to the dehydroxylation of the Ti pillars and the sulfate species. Since microporosity is introduced by pillaring, the decrease of the micropore volume indicates a sintering of small pores upon calcinations. In contrast, for the samples containing vanadium and tungsten, the BET surface areas and micropore volumes increased after calcinations at 400 °C. This result could be explained by the structural changes occurring during the heat treatment. The dehydroxylation of vanadia and tungsten species with the increase of the calcinations temperature to 400 °C induces a discharge of small pores filled with vanadium and tungsten species leading to an increase of the surface areas. For all samples, the surface areas are above 150 m²/g at 400 °C; characterizing the thermal stability of the investigated catalysts.

Table 3 summarizes the XPS binding energies of W, V, S and Ti present on the surface of the samples. The W4f_7/2 binding energy measured for the sample containing tungsten occurred at 36.5 eV and corresponds to tungsten in the 6+ oxidation state WO₃ [17]. An additional BE at 247.5 eV was observed for W4d_5/2 suggesting the existence of tungsten as W(VI) only [17]. No XPS peak at 32 eV characteristic of metallic W was observed. The binding energy at 517.1 eV typical of V2p_3/2 was detected for all samples containing vanadium. The literature data show that the binding energy for V⁵⁺ in V₂O₅ is between 517.4 and 516.4 eV, while it is between 515.7 and 515.4 eV for V⁴⁺ in V₂O₄ [17]. Therefore, the observed values indicate that the vanadium is present on the surface of all catalysts investigated in the 5+ valence form, as V₂O₅. Binding energy at near 169 eV was measured for the S2 p [17]. According to the literature data, this binding energy is typical of sulphur in the S⁶⁺ oxidation state as in Na₂SO₄ or Fe₂(SO₄)₃ [18]. This BE was observed for all samples investigated indicating that sulphur remain in 6+ oxidation state, in the form of SO₄²⁻, on the surface of the catalysts after vanadium and tungsten addition. The peaks at 161–162 eV belonging to sulphide and at 164 eV to elemental sulphur were not observed. The Ti-pillared clay material showed a Ti2p binding energy values around 459 eV. This suggests the presence of Ti⁴⁺, which agrees well with the data reported in literature [17]. This similar BE identical for all the samples indicates that the Ti⁴⁺ oxidation state remains unchanged before and after addition of tungsten and vanadium. Note that the intensities of the bands decreased in the case of V2p, S2p and Ti2p after W addition indicating an increasing coverage of V₂O₅, SO₄²⁻ and TiO₂ by WO₃ (data not shown). The amounts of sulphur in the Ti-pillared clay catalysts measured by the atomic surface composition evaluated as S/(Ti + V + W) atomic ratio calculated from each atomic value obtained by XPS measurements are also shown in Table 3. It is clear that the surface atomic ratio of sulphur to the sum of the cation shows a large dependency with the metal impregnation. When only vanadia was deposited, the value of S/(Ti + V + W) atomic ration decreased. It could be explained by the decrease of exposed sulphate species on the surface of the vanadia catalyst. When tungsten and vanadia were deposited at the same time, the S/(Ti + V + W) atomic ratio decreased sharply. This result could be explained by the decrease of exposed sulfate species due to the larger coverage by V₂O₅ and WO₃. In conclusion, the XPS results show that the catalysts contain the mixed oxide elements in the highest oxidation state, W(VI), V(V) and Ti (IV), most probably as WO₃, V₂O₅ and TiO₂ with SO₄²⁻ species on the surface of the pillared material catalyst.

The NH₃-TPD patterns of the clay samples are shown in Fig. 1. The shape of desorption pattern obtained for the STi-PILC support is very different from those recorded for the samples containing WO₃ and/or V₂O₅. However, the NH₃-TPD spectra of these solids contain at least two maxima. The first peak is centered at about 230 °C, while the second one, much broader, at temperature above 300 °C. This ammonia desorption suggest that there are at least two distinct NH₃ species adsorbed on the sulphated Ti-pillared clay surface. The low temperature peak can be related to the desorption of ammonia either physisorbed or linked to weak Bronsted acid sites, while the second peak at high temperature can be attributed to NH₃ desorbed from Lewis acid sites. It is known that the acidity of STi-PILC is essentially related to the presence of sulfate. In our previous work we have shown that the addition of sulfate to Ti-pillared clay increased significantly the total acidity and essentially the Bronsted acid sites [16,19,20]. The addition of vanadium into the STi-PILC increased the intensity of the first peak which correspond to the amount of weakly adsorbed ammonia and decreased the quantity of stronger chemisorbed NH₃ molecules. Thus the addition of vanadia to STi-PILC significantly changed the surface acidic properties of the support. This is in good agreement with measurements reported in earlier study that the strength of Bronsted acidity increased with V₂O₅ loadings, whereas Lewis acidity, mainly observed on the Ti-PILC (sulphur free support), remained almost constant [21]. For the catalyst containing WO₃ and V₂O₅ the major desorption of NH₃ occurred between 150 and 300 °C, the peak due to weak acid sites was intense and centered at about 220 °C with a decrease of the intensity of the second peak. The change in the TPD profiles can be explained by the progressive transformation of the surface acidity after V₂O₅ and WO₃ addition. When WO₃ and V₂O₅ exist simultaneously on the surface of STi-PILC, the amount of Bronsted acid sites further increases, along with a decrease in Lewis acid sites. This suggests that the addition of WO₃ and V₂O₅ produces more Bronsted acid sites while some of Lewis acid sites are covered by tungsten and vanadia. The Lewis acidity observed in all Ti-pillared clay samples was related to the acidic property of TiO₂ [19,20]. The addition of V₂O₅ to STi-PILC induces an increase of the total amount of NH₃ adsorbed on acid sites indicating that V₂O₅ contributes to the improvement of the total acidity (Table 4). In the case of the WO₃–V₂O₅/STi-PILC sample, the slight decrease of the total amount of NH₃ adsorbed may be explained by the coverage of the strong acid sites, arising from sulfate species, by relatively weak acid sites of WO₃. This interpretation agrees the detection of few amount of sulphur on the surface catalyst as observed by the decrease of S/(Ti + V + W) XPS atomic ratio for the WO₃–V₂O₅/STi-PILC sample.

The TPR profile of the STi-pillared clay support displays only large peak, starting near 400 °C and centered at 580 °C (Fig. 2). It was shown and well discussed in our earlier work that the hydrogen consumption over sulphated Ti-pillared clay is due to the reduction of sulfate to essentially SO₂ with little H₂S formation [22]. For the V₂O₅/STi-PILC catalyst, the threshold of hydrogen consumption began at 400 °C and two peaks were found centered at 506 and 530 °C. The first reduction peak, shifted towards lower temperature after vanadia addition, is mainly due to the sulphate reduction into essentially SO₂ with few H₂S and the second hydrogen consumption is attributed to the reduction of V⁵⁺ on V₂O₅ [22]. The reduction of V₂O₅/STi-PILC occurs at a temperature much lower than that of STi-PILC, showing an enhancement of sulphate reduction after vanadia addition. The TPR profiles clearly indicate that V₂O₅/STi-PILC is more easily reduced than STi-PILC. Since the WO₃/TiO₂ catalyst exhibits maxima above 700 °C, the reduction maximum at higher temperature was attributed to the reduction of TiO₂ and WO₃ [23]. In the case of WO₃–V₂O₅/STi-PILC, the TPR profiles displays only a large peak, starting near 400 °C and centered at 515 °C whereas the small peak observed at 530 °C for V₂O₅/STi-PILC disappeared. However, the reduction of sulfate completely hides the reduction of vanadium. Therefore the modification in reduction properties of sulfate and of vanadia evidences the existence of a strong interaction between tungsten, vanadia and sulfate on the surface of the WO₃–V₂O₅/STi-PILC catalyst leading to better red-ox properties of this material.

3.2 SCR-NO reaction

The NO conversions for the samples investigated are compared in Fig. 3. For the sulphated Ti-pillared clay, the NO conversion began at 200 °C and increased with reaction temperature. The SCR-NO activity of the STi-PILC support is essentially related to the presence of sulfate which enhances the Bronsted acidity of the catalyst and contributes to the improved activity at high temperature. As mentioned in our previous work, the sulphated Ti-pillared clay exhibits higher NO removal activity at high temperature than the non-sulfated ones [16]. We have shown a direct correlation between the SCR-NO activity and Bronsted acidity enhanced by the sulfation. After vanadia addition onto the STi-PILC, the SCR activity began at 100 °C, increased with temperature up to 300 °C then declined slightly at high temperature due to NH₃ oxidation by O₂ giving NO. Several researches have explained that the decrease of NO conversion is not caused by the decline of the catalytic activity but by the generation of NO by NH₃ oxidation [1,2]. Note that no NO₂ formation was observed for all the SCR experiments using our catalysts. The N₂O formation was observed only on catalysts with high vanadia loading (5 wt.%) and above 300 °C as indicated in our earlier work [22].

In the case of WO₃–V₂O₅/STi-PILC and V₂O₅/STi-PILC with the same vanadium content (3 wt.%), the SCR-NO activity is similar for the two catalysts, therefore the promoting effect of WO₃ is not found at high temperature (Fig. 3). The identical results were observed for the same catalysts containing 1.5 wt.% of vanadium (data not shown). The highest NO conversion (97%) at 350 °C, was obtained only for the samples with 3 wt.% of vanadium. When tungsten was deposited with vanadium at the same time on the surface of sulfated Ti-pillared clay, the sulfate species seem to play a more decisive role for NO removal activity than WO₃ species. These results are in contradiction with previous report results on effects of WO₃ supported by sulphur free V₂O₅/TiO₂ catalysts [2,12]. The literature also reported that the SCR-NO activities on 4.4%V₂O₅ + 8.2% WO₃/TiO₂-PILC (sulphur-free TiO₂-PILC) was higher than that on 4.4%V₂O₅/TiO₂-PILC and these two catalysts were highly active in SCR-NO than 4.4%V₂O₅ + 8.2% WO₃/TiO₂ and 4.4%V₂O₅/TiO₂, especially at higher temperature [12]. In this later study the authors concluded that the addition of WO₃ and V₂O₅ further increased the activities of the TiO₂-PILC catalysts. From our study, it can be concluded that the contribution of WO₃ on V₂O₅/STi-PILC to the SCR activity is negligible in presence of sulfate. Hence, when both tungsten and sulfate exist simultaneously on the surface of V₂O₅/STi-PILC, the sulfate species seem to play a more important role for NO removal activity than WO₃.

It has been already reported that the addition of V₂O₅ and WO₃ provides a wider temperature window than the V₂O₅ addition due to the improved reducibility at low temperature and the increased acidity at high temperature [5]. In this work, the surface area of all samples is similar at 400 °C. Accordingly, the change of SCR reactivity should be explained by just considering the TPD and TPR results. NH₃-TPD showed that the addition of vanadium and tungsten increased the total acidy. Also H₂-TPR indicated that the addition of V and W increased the red-ox properties of the material. However, the reasonable explanation of the similarity in activity of the WO₃–V₂O₅/STi-PILC and V₂O₅/STi-PILC catalysts having the same vanadia loading (3 wt.%) is the dispersion between the acid and red-ox sites. TopsØe et al. studied the V₂O₅/TiO₂ catalyst and based on TPD, TPR and FTIR they proposed two separate catalytic functions, namely acid (V⁵⁺-OH) and red-ox (V⁵⁺O) [3]. The authors suggested a dual site mechanism taken place over V⁵⁺–OH and V⁵⁺O. From this mechanism, it was suggested that the V⁵⁺O be reduced to V⁴⁺–OH by the ammonia adsorbed on the neighboring V⁵⁺–OH and V⁴⁺–OH be reoxidized by gaseous oxygen.

Now we try to compare our catalysts with those reported in the literature and especially the commercial SCR catalysts which consist of TiO₂ as support and WO₃–V₂O₅ as active phase [1,2]. Nowadays, deactivation of the commercial SCR catalysts applied in bio fuel plants is a major problem due to the higher amounts of potassium in bio fuel compared to other fuels. The commercial catalyst lifetime can be prolonged by sulfation with SO₂ or H₂SO₄ [7,8]. Washing with H₂SO₄ (0.5 M) was the most effective regeneration method and the bio-modified catalyst regained 92% of its initial activity [7]. Commercial SCR catalysts containing about 0.5–1% of sulphur that mainly exists in the form of sulfate on the surface were suggested [24]. It was also shown and well discussed in our earlier work that the undesired reaction of NH₃ oxidation to NO was inhibited by the sulfate species leading to an increase of the SCR-NO activity at high temperature [22].

The present study has employed pillared inter-layered clays intercalated by titanium as a support to replace the conventional TiO₂, well known with a low surface area and with only Lewis acidity compared to Titanium-pillared clay (Ti-PILC). From these reports it can be suggested that Ti-PILC is a promising support to replace TiO₂ for SCR catalyst and the NO removal activity can be enhanced by the increase of Bronsted acidity induced from the sulfate (1 wt.%). The experimental findings of this study suggest that the WO₃–V₂O₅/STi-PILC or V₂O₅/STi-PILC catalysts may be preferably applied to SCR processes for NO containing exhaust gases from power plants which strongly require the catalysts with a high SCR activity at around 350 °C.