The role of vanadium species during SO₂ removal over a V₂O₅/AC catalyst

Abstract

The role of vanadium species during SO₂ removal over a V₂O₅/AC catalyst was studied in detail using in situ diffuse reflectance infrared Fourier transform spectroscopy, Raman spectroscopy, temperature programmed desorption, X-ray photoelectron spectroscopy and N₂ adsorption. Results revealed that V₂O₅ reacted with SO₂ to form a stable vanadium(V) sulfate of V₂O₃(SO₄)₂. However, V₂O₃(SO₄)₂ did not deactivate the catalyst and was catalytically active towards SO₂ oxidation, leading to the formation of H₂SO₄. Thus, the sulfur capacity of the V₂O₅/AC catalyst was determined by formation of V₂O₃(SO₄)₂ and H₂SO₄. Increasing V₂O₅ loading enhanced the formation of V₂O₃(SO₄)₂ but inhibited that of H₂SO₄ because V₂O₃(SO₄)₂ occupied the pores for H₂SO₄ storage.

---

1. Introduction

Sulfur dioxide (SO₂) is a major air pollutant.¹ SO₂ removal is a topic of ongoing research since it causes severe environmental problems such as acid rain and photochemical smog.^2,3 In this regard, activated carbon/coke (AC) has been demonstrated to remove SO₂ effectively due to its large surface area, abundant functional groups and microporous structure.^4–6

SO₂ chemical adsorption on AC contains several steps involving adsorption of SO₂ onto the surface, oxidation of SO₂ to SO₃ by O₂, reaction of SO₃ with H₂O to form H₂SO₄, and the storage of H₂SO₄ in the pores.^7–9 The rate-determining step is the oxidation of SO₂ to SO₃.⁹ Therefore, modification of AC to improve its catalytic activity towards SO₂ oxidation is a very effective way to enhance SO₂ removal. In this regard, modification of AC with metal oxides has been extensively studied. Many metal oxides, such as CuO,^10,11 Fe₂O₃,^12,13 and MnO_x,^14,15 loaded on AC have been reported to be catalytically active at low temperatures. By catalyzing with metal oxides, SO₂ is oxidized to SO₃ by O₂, which then reacts with metal oxides to form stable metal sulfates. In this way, SO₂ is fixed onto the AC. The formed metal sulfates (CuSO₄, Fe₂(SO₄)₃, MnSO₄, etc.) are inactive for SO₂ oxidation. As a result, the catalysts are deactivated. The mechanism of SO₂ removal could be expressed as follows:

MO_x + SO₂ + O₂ → M(SO₄)_x (M = Cu, Fe, Mn, etc.)

where x is determined by the valence state of the corresponding metals.

The AC modified with vanadium oxide (V₂O₅/AC) is also active for SO₂ removal. Precisely because of its high activity, abundant studies have been conducted in the past two decades. Various factors affecting the sulfur capacity^16–19 and regeneration of sulfated catalysts^20–22 have been elaborately studied. Modified with V₂O₅, the activated carbon honeycomb was more effective towards SO₂ removal compared to AC.^16,17 The sulfated catalyst regenerated in an NH₃ atmosphere exhibited even higher activity than the fresh one.²¹ All of these studies were based on the premise that the product resulting from SO₂ removal was H₂SO₄. By catalyzing with V₂O₅, SO₂ was oxidized to SO₃ by O₂. However, SO₃ reacted with H₂O to form H₂SO₄, rather than with V₂O₅ to form a stable vanadium sulfate.²³ Apparently, this mechanism is different from other AC supported metal oxides, and could be expressed as follows:

SO₂ + 1/2O₂ + H₂O → H₂SO₄

In the case of SO₂ oxidation over V₂O₅/AC, some researchers considered that SO₂ reacted with V₂O₅ to form an intermediate of VOSO₄, which readily converted into V₂O₅.^18,19,24,25 However, in our previous studies, detailed characterization of AC supported VOSO₄ revealed that VOSO₄ could not convert into V₂O₅ and H₂SO₄, but was easily oxidized into a new vanadium(V) sulfate of divanadium(v) trioxide disulfate by O₂ according to the reaction as follows:^26,27

2V^IVOSO₄ + 1/2O₂ → V^V₂O₃(SO₄)₂

Moreover, the AC modified with V₂O₃(SO₄)₂ exhibited high activity towards SO₂ removal, indicating the catalytic role of V₂O₃(SO₄)₂ in SO₂ oxidation.²⁶ This is different from other metal sulfates (CuSO₄, Fe₂(SO₄)₃, MnSO₄, etc.), which are inactive for SO₂ oxidation. Additionally, the characteristic peak of V₂O₃(SO₄)₂ during the temperature programmed process was observed over sulfated V₂O₅/AC,²⁶ implying the transformation of V₂O₅ into V₂O₃(SO₄)₂. In other words, the V₂O₅/AC catalyst might exist in another reaction for SO₂ removal, with the metal sulfate being the final product. This finding would provide a new point of view for SO₂ removal over V₂O₅/AC.

The objective of this work is to provide spectroscopic evidence for the formation of V₂O₃(SO₄)₂ and to elucidate the relationship between V₂O₃(SO₄)₂ and H₂SO₄. To achieve this purpose, the catalysts with various V₂O₅ loadings were investigated for SO₂ removal. Then, the sulfated catalysts were characterized in detail by using in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), Raman spectroscopy, temperature programmed desorption (TPD), X-ray photoelectron spectroscopy (XPS) and N₂ adsorption. Based on the information obtained, the role of vanadium species during SO₂ removal over the V₂O₅/AC catalyst was discussed.

2. Experimental

2.1 Catalyst preparation

The V₂O₅/AC catalyst was prepared by pore volume impregnation.²⁸ The volume of solution used in the impregnation was equal to the volume of the support. The coal-based AC purchased from Xinhua Chemical Plant with a surface area of 986 m² g⁻¹ was used as the support. The V₂O₅/AC catalyst was prepared by pore volume impregnation with an aqueous solution of ammonium metavanadate (99.9%, Aladdin) in oxalic acid (99.5%, Aladdin). After impregnation, the sample was dried at 50 °C for 12 h and 120 °C for 5 h, calcined in N₂ at 500 °C for 6 h, and pre-oxidized in air at 250 °C for 5 h. The V₂O₅ loadings were determined by the concentration of ammonium metavanadate solution and confirmed by inductively coupled plasma (ICP) analysis. The catalyst is expressed as Vx/AC, where x is the V₂O₅ loading. For instance, V5/AC is the catalyst with 5 wt% V₂O₅.

The V₂O₃(SO₄)₂/AC catalyst was prepared by impregnating AC with vanadyl sulfate (99.9%, Alfa) aqueous solution, followed by drying in air at 50 °C for 12 h and then at 150 °C for 5 h. During this process, vanadyl sulfate (V^IVOSO₄) was oxidized into divanadium(V) trioxide disulfate (V^V₂O₃(SO₄)₂) by O₂.^26,27 The V₂O₃(SO₄)₂ content was 3.5 wt% and denoted as VS/AC.

2.2 Catalyst characterization

2.2.1 In situ DRIFTS experiments. The formation and evolution of sulfates during SO₂ oxidation were investigated in situ by using a Bruker Tensor 27 instrument equipped with a MCT detector, operating at a resolution of 8 cm⁻¹ for 100 scans.²⁹ All experiments were carried out at a constant temperature of 120 °C. The sample was placed in a crucible in an in situ chamber. The chamber was fitted with a ZnSe window and connected to dry N₂, O₂ and SO₂via stainless steel lines. Before the measurement, the sample was pretreated in a N₂ flow at 100 mL min⁻¹ at 200 °C to remove water and impurities of physisorption. The chamber was then cooled to 120 °C and kept at that temperature for 5 min. Subsequently, the background spectrum of the sample was recorded and subtracted from the sample spectrum. The measurements were performed by exposure of the pretreated sample to a mixture of 150 ppm SO₂ and 4.5% O₂ with N₂ as the carrier. The total flow rate was maintained at 100 mL min⁻¹. The in situ DRIFTS analysis system is shown in Fig. 1.

Fig. 1 The scheme of the in situ DRIFTS analysis system. (1 and 2) mass flow controller; (3) rotameter; (4) cooling water; (5) sample chamber of the FTIR spectrometer; (6) high pressure/high temperature chamber of the diffuse reflectance accessory; (7) sweep gas analysis meter; (8) temperature controller; (9) water pump.

2.2.2 Raman spectroscopy. Raman spectra were recorded using a Horiba Labram HR800 spectrometer, operating at a spectral resolution of ca. 3 cm⁻¹. Spectral acquisition consists of 10 accumulations of 30 s. The spectra were excited by using a 514 nm Ar⁺ line.

2.2.3 TPD. TPD experiments were carried out in a fixed bed quartz reactor (i.d. 22 mm), with 0.5 g sample being loaded in the reactor. The sample was heated from room temperature to 500 °C in a N₂ flow at 400 mL min⁻¹ at a heating rate of 5 °C min⁻¹. The evolution of SO₂ was monitored using a KM 9106 Quintox flue gas analyzer.

2.2.4 XPS. XPS characterization was performed on a Thermo ESCALAB 250 spectrometer. The X-ray source was obtained using Al Kα monochromatized radiation (hν = 1486.6 eV). The high resolution V2p and S2p spectra were recorded with a pass energy of 20 eV. The binding energy (BE) scales were corrected for charging effects by assigning a value of BE = 284.6 eV to the C1s peak from the surface contamination.

2.2.5 Surface area and pore structure analysis. The textural properties of the catalysts were measured by the adsorption of N₂ at 77 K using an ASAP 2020 Micromeritics analyzer. Prior to measurement, the catalyst was dried at 150 °C for 5 h. The specific surface area was calculated by the Brunauer–Emmett–Teller (BET) equation. The Dubinin–Radushkevich (DR) method was used to obtain the micropore volume.

2.3 SO₂ removal

The temperature of flue gas was in the range of 120–200 °C after electrostatic precipitation. Moreover, low temperature was favorable for SO₂ removal. Therefore, SO₂ removal over the V₂O₅/AC catalyst was conducted at 120 °C in the same fixed bed quartz reactor (i.d. 22 mm) containing 3 g catalyst. The feed rates of N₂, O₂ and SO₂ were controlled by mass flow meters. Water vapor was introduced into the reactor by passing N₂ through a heated gas-wash bottle containing deionized water. The content of water vapor was controlled by both water temperature and the flow rate of N₂. Simulated flue gas consisted of 0.15% SO₂, 4.5% O₂, 2.5% H₂O and balanced N₂. The total flue gas rate was maintained at 400 mL min⁻¹, corresponding to a space velocity of 4000 h⁻¹. The concentrations of SO₂ and O₂ in the inlet and outlet of the reactor were analyzed using a KM 9106 Quintox flue gas analyzer.

3. Results and discussion

3.1 SO₂ removal

The results of SO₂ removal are illustrated in Fig. 2. In all runs, the measurements were terminated when SO₂ conversion decreased from 100% to 90%. The SO₂ retention was quantified by integrating curve area and expressed as mg SO₂/g catalyst. As can be seen, the AC support showed almost no SO₂ retention. After modification with V₂O₅, the breakthrough time was significantly enhanced, indicating the promoting role of vanadium oxide in SO₂ removal. Additionally, the V₂O₃(SO₄)₂/AC exhibited a sulfur capacity of 42.4 mg SO₂ per g catalyst, indicating the catalytic role of V₂O₃(SO₄)₂ in SO₂ oxidation. The result was consistent with our previous work.³⁰

Fig. 2 SO₂ breakthrough profiles over various Vx/AC catalysts at 120 °C. Reaction conditions: 0.15% SO₂, 4.5% O₂, 2.5% H₂O, N₂ balance, 4000 h⁻¹.

The sulfur content and the specific activity per mol of V atom as a function of V₂O₅ loading are illustrated in Fig. 3. The sulfur content increased with V₂O₅ loading, reaching a maximum at V5/AC. Then, it decreased with further increasing V₂O₅ loading. The specific activity per mol of V atom decreased continuously with V₂O₅ loading. The V1/AC showed an S/V molar ratio of 7.8. According to studies, SO₂ was oxidized into SO₃ over vanadium sites.^23,31 SO₃ migrated to the AC support and reacted with H₂O to form H₂SO₄ as the final product, which was stored in AC pores. The chemical form of vanadium oxides was not changed and continuously catalyzed SO₂ oxidation to H₂SO₄, leading to high specific activity per mol of V atom. However, for the catalyst with high V₂O₅ loading such as V9/AC, the S/V molar ratio was 0.8, which was almost an order of magnitude lower than V1/AC and could not be interpreted by the current mechanism. Therefore, the mechanism of SO₂ removal over V9/AC might be different from that over V1/AC.

Fig. 3 Evolution of the sulfur content and S/V molar ratio as a function of V₂O₅ loading.

3.2 In situ DRIFTS

To determine the chemical form of sulfates, the sulfated catalysts were subjected to Fourier transform infrared (FTIR) analysis. However, due to the overlap of C–O stretching and O–H bending modes of alcoholic and carboxylic groups on AC, the evolution of the characteristic peak of sulfates with V₂O₅ loading was not evident.²⁹ Thus, in situ DRIFTS was used to characterize the sulfated catalyst, because the background spectrum could be recorded to avoid the influence of oxygen groups existing on the AC surface. V1/AC, V5/AC and V9/AC were selected for in situ DRIFTS analysis.

As shown in Fig. 4, no visible peak could be observed at the beginning of the experiment. As the exposure proceeded, a broad band in the region of 900–1300 cm⁻¹ appeared over the three catalysts, and the peak intensity increased continuously with time. This band was ascribed to the stretching motion of adsorbed bisulfate (HSO₄⁻) and/or sulfate (SO₄²⁻) on the surface. As a free anion in solution, sulfate had a tetrahedral symmetry and belonged to the point group T_d. For this symmetry, there were two infrared sulfate vibrations accessible to the spectroscopic investigation using the IR technique. They were non-degenerate symmetric stretching ν₁ (980 cm⁻¹, IR inactive) and triply degenerate asymmetric stretching ν₃ bands (1100 cm⁻¹, IR active). However, if the sulfate ion was adsorbed on the surface, its symmetry could be lowered into C_3v or C_2v. As a result, the ν₁ vibration mode became IR active, and ν₃ vibration split into two peaks for C_3v and three peaks for C_2v.²³ Thus, the broad band in the region of 900–1300 cm⁻¹ was a result of the ν₁ vibration and ν₃ splitting.^23,29

Fig. 4 DRIFTS spectra of Vx/AC at 1, 10, 20, 40, 80, and 120 min during sulfation at 120 °C. Reaction conditions: 150 ppm SO₂ and 4.5% O₂ in 100 mL min⁻¹ N₂.

The band of sulfated V1/AC showed features at 991, 1078, 1153 and 1244 cm⁻¹, and the main peak located at 1153 cm⁻¹. As the V₂O₅ loading increased, the band of sulfates became more irregular and the main peak moved towards a high wave number. The characteristic peak of V5/AC is located at 1171 cm⁻¹, and that of V9/AC at 1244 cm⁻¹. The result indicated that the sulfates over the three catalysts were adsorbed on the sites with different chemical natures. For V1/AC, it was widely accepted that the sulfates were mainly linked to the AC support and existed in the form of H₂SO₄.²³ However, for the catalysts with higher V₂O₅ loading, the sulfates existed in a more complicated chemical form.

In addition, it was worth noting that a band at 1571 cm⁻¹ appeared concomitantly with the sulfates. It was resulted from the active of C=C modes in polyaromatic as a result of the adsorption of sulfates.²⁹

3.3 Raman spectroscopy

Due to the high dispersion of vanadium species and the dark material properties of AC, the characterization of vanadium species over AC was difficult. According to the study of Guerrero-Pérez,³² when the BET surface area was 1056 m² g⁻¹ and the surface density of VOx was 3 nm⁻² corresponding to a V₂O₅ loading of 48 wt%, the X-ray diffraction (XRD) pattern showed no signal for vanadium species. However, Raman spectroscopy was successful in characterizing the evolution of VOx over AC.³² Therefore, Raman spectroscopy was used to characterize the vanadium species over the sulfated catalysts (SVx/AC). The results are shown in Fig. 5.

Fig. 5 Raman spectra of SVx/AC catalysts.

For SV1/AC and SV2/AC, no characteristics of vanadium species could be detected due to the low V₂O₅ loading. When the V₂O₅ loading increased to 3 wt%, a peak at 1011 cm⁻¹ assigned to ν(V=O) appeared,^32,33 and the intensity of this band increased with V₂O₅ loading. Moreover, a new band at 535 cm⁻¹ due to ν(S–O–V), which was the characteristic peak of divanadium(V) trioxide disulfate (V^V₂O₃(SO₄)₂), was detected.³³ V₂O₃(SO₄)₂ was composed of pairs of highly distorted VO₆ octahedra connected to one another by a bridging O atom and two SO₄ tetrahedra, with characteristic bands at 1011 cm⁻¹ due to ν(V=O), 750 cm⁻¹ due to ν(V–O–V) and 535 cm⁻¹ due to ν(S–O–V).³³ The Raman spectroscopy data showed apparently the vibrations of ν(V=O) and ν(S–O–V), indicating the transformation of V₂O₅ into V₂O₃(SO₄)₂ after sulfation. Compared to ν(V=O), ν(V–O–V) in polyvanadate structures, such as (VO)₂P₂O₇ and (VO)₂O(SO₄)₄⁴⁻, was much weaker and easily smeared out due to the lower signal-to-noise ratio, resulting in the absence of ν(V–O–V).^32,34 The band at 449 cm⁻¹ was assigned to ν(SO₄²⁻), and the band at 323 cm⁻¹ was attributed to lattice vibrations.³³ Conclusively, the appearance of characteristic bands assigned to ν(V=O) and ν(S–O–V) strongly indicated the formation of V₂O₃(SO₄)₂ after SO₂ removal.

3.4 TPD study

According to studies, the SO₂-TPD pattern of the sulfated V₂O₅/AC catalyst showed a peak at 300–320 °C, which was related to the decomposition of H₂SO₄.^23,31,35 On the other hand, V₂O₃(SO₄)₂ loaded on AC formed a peak at 375 °C during the temperature-programmed process.^26,30 If V₂O₅ indeed transformed into V₂O₃(SO₄)₂ after SO₂ removal, there should be two SO₂ release peaks when the temperature-programmed process was performed on sulfated catalysts (SVx/AC). Based on this speculation, the TPD experiment was performed. The results are illustrated in Fig. 6. For comparison, the TPD result of fresh VS/AC reflecting the decomposition behavior of V₂O₃(SO₄)₂ over AC is also shown.

Fig. 6 TPD (A) and Gaussian fitting (B) results of SVx/AC catalysts. Reaction conditions: 0.5 g catalyst, 400 mL min⁻¹ N₂, heating rate of 5 °C min⁻¹.

For SV1/AC and SV2/AC, the SO₂ release peak is located at 305 °C, which was a result of the decomposition of H₂SO₄.³¹ However, a shoulder at 375 °C assigned to the decomposition of V₂O₃(SO₄)₂ was also visible. Moreover, the intensity of this shoulder increased with V₂O₅ loading, accompanied by a continuous decrease of the peak at 305 °C. For SV7/AC and SV9/AC, the peak at 375 °C became the main peak and that at 305 °C reduced to a shoulder. This result clearly demonstrated the transformation of V₂O₅ into V₂O₃(SO₄)₂ during SO₂ removal. Apparently, the formation of V₂O₃(SO₄)₂ was in accordance with V₂O₅ loading. The catalyst with higher V₂O₅ loading generated greater amounts of V₂O₃(SO₄)₂. To clearly describe the TPD results, Gaussian fitting was performed to calculate the amount of H₂SO₄ and V₂O₃(SO₄)₂. As shown in Fig. 6B, the amount of V₂O₃(SO₄)₂ increased continuously with V₂O₅ loading, while that of H₂SO₄ showed an opposite trend.

The results in Fig. 6 indicated that there were two reactions simultaneously occurring over V₂O₅/AC during SO₂ removal:

V₂O₅ + 2SO₂ + O₂ = V₂O₃(SO₄)₂ (I)

SO₂ + 1/2O₂ + H₂O = H₂SO₄ (II)

For the catalyst with low V₂O₅ loading such as V1/AC, reaction (II) was the main process for SO₂ removal. With increasing V₂O₅ loading, reaction (I) gradually took the place of reaction (II).

Numerous studies proposed that V₂O₅ was the active component for SO₂ oxidation to H₂SO₄ during SO₂ removal over the V₂O₅/AC catalyst.^{16,17,23,35,36} However, Raman and TPD results indicated the transformation of V₂O₅ into V₂O₃(SO₄)₂ after SO₂ removal. Moreover, as indicated in Fig. 2 and our previous work,^26,30 V₂O₃(SO₄)₂ was catalytically active for SO₂ oxidation. Thus, which was the active component responsible for SO₂ oxidation to H₂SO₄?

As shown in Fig. 6A, the characteristic peak of H₂SO₄ was 70 °C lower than that of V₂O₃(SO₄)₂, which avoided the overlap of two peaks and made TPD a powerful technique to study the relationship between reaction (I) and (II). Additionally, the intensity of peaks 1 and 2 over sulfated V5/AC was close to each other. Therefore, the V5/AC catalyst was selected to study the process of SO₂ removal further.

Six V5/AC catalysts were subjected to sulfation for 30, 60, 90, 120, 150 and 360 min, respectively. Then, TPD experiments were performed on sulfated catalysts to investigate formations of H₂SO₄ and V₂O₃(SO₄)₂ at different stages of SO₂ removal. The results are illustrated in Fig. 7. Note that the V5/AC catalyst was saturated with SO₂ at the sulfation time of 360 min.

Fig. 7 TPD (A) and Gaussian fitting (B) results of the sulfated V5/AC catalysts with sulfation times of 30, 60, 90, 120, 150, and 360 min. Reaction conditions: 0.5 g catalyst, 400 mL min⁻¹ N₂, heating rate of 5 °C min⁻¹.

As shown in Fig. 7A, the catalyst with the sulfation time of 30 min showed a peak at 375 °C, indicating that V₂O₃(SO₄)₂ was the main sulfate on the catalyst. Apparently, reaction (I) was the main process for SO₂ removal at the initial stage. In addition, a shoulder at 305 °C, which was the characteristic peak of H₂SO₄, was also visible. With increasing time, the intensities of both peaks 1 and 2 were enhanced. However, the intensity of peak 1 grew faster than that of peak 2. At the end of sulfation, peak 1 became the main peak while peak 2 relatively reduced to a shoulder. Similar results were also obtained over V2/AC and V7/AC, with the differences in intensities of peak 2 due to different V₂O₅ loadings.

Apparently, the formation of V₂O₃(SO₄)₂ was superior to that of H₂SO₄. H₂SO₄ mainly formed after the transformation of V₂O₅ into V₂O₃(SO₄)₂. To clearly describe this process, Gaussian fitting was performed. The fitting result was plotted versus reaction time. As shown in Fig. 7B, the amount of V₂O₃(SO₄)₂ and H₂SO₄ increased with time. At the sulfation time of 360 min, the amount of V₂O₃(SO₄)₂ corresponded to a sulfur capacity of about 34 mg SO₂ per g cat, with the S/V molar ratio being 0.97. This indicated that 97% of V₂O₅ had transformed into V₂O₃(SO₄)₂. The two lines crossed each other at about 125 min, dividing the sulfation into two stages. At stage 1, reaction (I) was the main process for SO₂ removal. At stage 2, sulfation mainly proceeded viareaction (II).

To clearly describe the formation of V₂O₃(SO₄)₂ and H₂SO₄, sulfation at 90 min was selected for further analysis. At the time of 90 min, 85% of V₂O₅ had transformed into V₂O₃(SO₄)₂. In contrast, 80% of H₂SO₄ formed in the time region of 90–360 min in which the vanadium species mainly existed as V₂O₃(SO₄)₂. In other words, H₂SO₄ mainly formed after most of V₂O₅ had transformed into V₂O₃(SO₄)₂. Considering the catalytic role of V₂O₃(SO₄)₂ and the fact that the AC was almost inactive for SO₂ oxidation, this result strongly indicated that the formation of H₂SO₄ resulted from the catalytic activity of V₂O₃(SO₄)₂ for SO₂ oxidation. During SO₂ removal over V₂O₅/AC, V₂O₅ firstly transformed into V₂O₃(SO₄)₂. By catalyzing with V₂O₃(SO₄)₂, SO₂ was oxidized to SO₃ by O₂, which then reacted with H₂O to form H₂SO₄. This in turn accounted for the two stages of SO₂ removal. At first, molecular SO₂ reacted with V₂O₅ to form V₂O₃(SO₄)₂. However, the probability of this reaction gradually decreased with the consumption of V₂O₅. In contrast, with the formation of V₂O₃(SO₄)₂, the opportunity for SO₂ oxidation to H₂SO₄ catalyzed by V₂O₃(SO₄)₂ increased. As a result, reaction (II) gradually took the place of reaction (I) and acted as the main process for SO₂ removal, dividing the whole sulfation process into two stages. It could be expected that the time for stage 1 was determined by V₂O₅ loading. The lower the V₂O₅ loading was, the shorter the time of stage 1 became.

During stage 2, the presence of H₂O was essential. H₂O could react with SO₃ to form H₂SO₄. Moreover, H₂SO₄ formed on the catalyst surface could be continuously swept by H₂O adsorbed in the pores.⁹ Therefore, H₂O significantly promoted SO₂ removal. Because SO₂ removal over the V₂O₅/AC catalyst with low V₂O₅ loading mainly proceeded via stage 2, the presence of H₂O was critical. The sulfur capacity of V1/AC was only 6 mg SO₂ per g catalyst in the absence of H₂O. When 2.5% H₂O was added in the flue gas, the sulfur capacity increased to 55 mg SO₂ per g catalyst.

Conclusively, V₂O₅ transformed into V₂O₃(SO₄)₂ during SO₂ removal over V₂O₅/AC. However, V₂O₃(SO₄)₂ did not deactivate the catalyst. The formation of H₂SO₄ resulted from the catalytic activity of V₂O₃(SO₄)₂ for SO₂ oxidation.

3.5 XPS study

To gain further insight into the sulfates, the sulfated V5/AC catalysts were subjected to XPS. Fig. 8 shows the core level S2p spectra. The peak at 168.6 eV was assigned to S⁶⁺, indicating the formation of SO₄²⁻. The intensity of the peak increased with time, which implied the accumulation of SO₄²⁻ over the catalyst surface. Additionally, an apparent shoulder at 169.7 eV was visible, indicating the existence of two types of sulfates, which was in accordance with TPD. The peak at 168.6 eV was assigned to H₂SO₄.²³ The shoulder at 169.7 eV might be assigned to V₂O₃(SO₄)₂, which was a result of the attraction of electrons by vanadium(V) spurring the peak position of S2p to move towards higher binding energy.

Fig. 8 Core level S2p spectra of the sulfated V5/AC catalysts.

Additionally, it should be noted that the evolution of two sulfates versus time revealed by XPS was not in accordance with TPD. The TPD results indicated that the formation of V₂O₃(SO₄)₂ was the main process at the initial stage of sulfation. However, XPS implied that H₂SO₄ was the main sulfate through the whole sulfation process. This was derived from the nature of the two characterization techniques. During the temperature programmed process, all of the sulfates were desorbed. By contrast, XPS is a surface characterization technique, and only the sulfates on the catalyst surface could be detected. The discrepancy between the TPD and XPS results indicated that the distribution of two sulfates was not homogeneous. Apparently, H₂SO₄ tended to store in the pores close to the surface by comparison with V₂O₃(SO₄)₂.

Fig. 9 shows the core level V2p spectra of the sulfated V5/AC catalysts. The peak of V2p_3/2 at 517.1 eV was typical for V⁵⁺,^37,38 indicating that the vanadium species were maintained in the high oxidation state. However, the full width at half maximum (FWHM) of the V2p_3/2 spectrum gradually increased from 1.22 eV to 2.25 eV with sulfation, implying that vanadium(V) became more reductive. Compared to vanadium(V), the S atom is an electron donor, and thereby led to the reduction of vanadium(V) to some extent.

Fig. 9 Core level V2p spectra of the sulfated V5/AC catalysts.

3.6 Role of vanadium species in SO₂ removal

Based on the information obtained, it was clear that V₂O₅ transformed into V₂O₃(SO₄)₂ during SO₂ removal. The sulfur capacity of V₂O₅/AC was determined by formation of V₂O₃(SO₄)₂ and H₂SO₄. For the catalyst with low V₂O₅ loading, the small amount of V₂O₃(SO₄)₂ mainly acted as an active component for SO₂ oxidation to H₂SO₄. With the V₂O₅ loading increasing, its role as the sulfate contributing to sulfur capacity was enhanced. It was accepted that H₂SO₄ formed over V₂O₅/AC mainly stored in the pores of AC, especially in micropores.¹⁹ Thus, the pore volume of AC was an important factor affecting the formation of H₂SO₄. Since the formation of V₂O₃(SO₄)₂ was superior to that of H₂SO₄, it was expected that V₂O₃(SO₄)₂ might occupy the pores of AC for H₂SO₄ storage and thus inhibit its formation. To prove this, the sulfated V5/AC catalysts were subjected to N₂ adsorption. As illustrated in Table 1, the BET surface and the pore volume decreased continuously with sulfation. In the time region of 0–90 min, the formation of V₂O₃(SO₄)₂ was the main process for SO₂ removal based on the results of TPD. The surface area and pore volume decreased obviously with time.

Table 1 Changes in textural properties of V5/AC at different sulfation times

Samples	S BET (m^2 g^−1)	V total (cm^3 g^−1)	S micro (m^2 g^−1)	V micro (cm^3 g^−1)
V5/AC	797	0.465	434	0.201
SV5/AC-30	769	0.441	413	0.189
SV5/AC-60	708	0.414	383	0.175
SV5/AC-90	667	0.387	353	0.162
SV5/AC-120	621	0.371	321	0.146
SV5/AC-150	580	0.360	295	0.138
SV5/AC-360	495	0.301	223	0.105

---

Thus, V₂O₃(SO₄)₂ played two roles in the formation of H₂SO₄. On the one hand, the catalytic activity of V₂O₃(SO₄)₂ was essential for H₂SO₄ formation. On the other hand, as a sulfate contributing to sulfur capacity, V₂O₃(SO₄)₂ occupied the pores for H₂SO₄ storage and thus inhibited its formation. The two contradictory roles might achieve a balance over V5/AC, resulting in the highest sulfur capacity under the reaction conditions applied in this work.

Accordingly, a mechanism of SO₂ removal over V₂O₅/AC is proposed in Scheme 1. The active component of V₂O₅ firstly transformed into V₂O₃(SO₄)₂. However, V₂O₃(SO₄)₂ did not deactivate the catalyst and further catalyzed SO₂ oxidation to H₂SO₄. Actually, transformation of metal oxides into metal sulfates occurred on a large number of the AC supported metal oxides. In most cases, such a transformation deactivated the catalyst. Therefore, a high sulfur capacity could be only obtained on the catalyst with high loading of metal oxide. In this case, a typical example is CuO/AC. CuSO₄ was the only sulfur-containing species over sulfated CuO/AC with a S/Cu molar ratio being not higher than 1.^10,11 By contrast, benefitting from the catalytic activity of V₂O₃(SO₄)₂, the catalyst still kept its catalytic activity towards SO₂ oxidation after transformation of V₂O₅ into V₂O₃(SO₄)₂, leading to the formation of H₂SO₄. As a result, a high sulfur capacity could be obtained over V₂O₅/AC with low V₂O₅ loading. For this reason, most previous studies paid special attention to the V₂O₅/AC catalyst with V₂O₅ loading being not higher than 2 wt%.^{16,17,31,39,40}

Scheme 1 Mechanism of SO₂ removal over the V₂O₅/AC catalyst.

4. Conclusions

In the present work, the role of vanadium species during SO₂ removal over a V₂O₅/AC catalyst was studied in detail using various techniques. Results revealed that the vanadium species had two roles in SO₂ removal. On the one hand, V₂O₅ reacted with SO₂ to form V₂O₃(SO₄)₂. In this way, a part of SO₂ was fixed on the catalyst. On the other hand, V₂O₃(SO₄)₂ did not deactivate the catalyst and acted as an active component catalyzing SO₂ oxidation to H₂SO₄. Therefore, the sulfur capacity of V₂O₅/AC was determined by formation of V₂O₃(SO₄)₂ and H₂SO₄. Because V₂O₃(SO₄)₂ occupied the pores for H₂SO₄ storage, increasing V₂O₅ loading enhanced the formation of V₂O₃(SO₄)₂ but inhibited that of H₂SO₄. Benefitting from the catalytic activity of V₂O₃(SO₄)₂, a high sulfur capacity could be obtained on V₂O₅/AC with low V₂O₅ loading, with H₂SO₄ being the main sulfate. In this case, the vanadium species mainly acted as active components for SO₂ oxidation.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors gratefully acknowledge the financial support by the National Key Research and Development Program (2017YFC0210203), the National Natural Science Foundation of China (grants 21902173 and 21978314), the Key Research and Development Program of Shanxi Province (201703D11101804), and the Shandong Province Science Foundation (ZR2017BB063).

References

1. M. Berger, P. Fioux, S. Dorge, H. Nouali, D. Habermacher, E. Fiani, M. Vierling, M. Moliere, J. F. Brilhac and J. Patarin, Catal. Sci. Technol., 2017, 7, 4115–4128.
2. J. Yun, C. Zhu, Q. Wang, Q. Hu and G. Yang, Chemosphere, 2019, 217, 18–25.
3. K. Li, L. Chen, S. J. White, K. Han, B. Lv, K. Bao, X. Wu, X. Gao, M. Azzi and K. Cen, Atmos. Res., 2017, 192, 38–47.
4. J. M. Rosas, R. Ruiz-Rosas, J. Rodríguez-Mirasol and T. Cordero, Chem. Eng. J., 2017, 307, 707–721.
5. Y. Zhu, X. Han, Z. Huang, Y. Hou, Y. Guo and M. Wu, Chem. Eng. J., 2018, 337, 741–749.
6. M. J. Lázaro, A. Boyano, M. E. Gálvez, M. T. Izquierdo, E. García-Bordejé, C. Ruiz, R. Juan and R. Moliner, Catal. Today, 2008, 137, 215–221.
7. P. Davini, Carbon, 1991, 29, 321–327.
8. F. Carrasco-Marin, E. Utrera-Hidalgo, J. Rivera-Utrilla and C. Moreno-Castilla, Fuel, 1992, 71, 575–578.
9. A. A. Lizzio and J. A. DeBarr, Fuel, 1996, 75, 1515–1522.
10. H. H. Tseng and M. Y. Wey, Carbon, 2004, 42, 2269–2278.
11. H. H. Tseng and M. Y. Wey, Chemosphere, 2006, 62, 756–766.
12. J. Ma, Z. Liu, S. Liu and Z. Zhu, Appl. Catal., B, 2003, 45, 301–309.
13. T. Grzybek and H. Papp, Appl. Catal., B, 1992, 1, 271–283.
14. S. Pan, H. Luo, L. Li, Z. Wei and B. Huang, J. Mol. Catal. A: Chem., 2013, 377, 154–161.
15. J. X. Guo, Y. F. Qu, S. Shu, X. J. Wang, H. Q. Yin and Y. H. Chu, New J. Chem., 2015, 39, 5997–6015.
16. Y. Wang, Z. Huang, Z. Liu and Q. Liu, Carbon, 2004, 42, 445–448.
17. Y. Wang, Z. Liu, L. Zhan, Z. Huang, Q. Liu and J. Ma, Chem. Eng. Sci., 2004, 59, 5283–5290.
18. J. Ma, Z. Liu, Q. Liu, S. Guo, Z. Huang and Y. Xiao, Fuel Process. Technol., 2008, 89, 242–248.
19. Y. Xiao, Q. Liu, Z. Liu, Z. Huang, Y. Guo and J. Yang, Appl. Catal., B, 2008, 82, 114–119.
20. Y. Guo, Z. Liu, Y. Li and Q. Liu, J. Fuel Chem. Technol., 2007, 35, 344–348.
21. Y. Guo, Z. Liu, Q. Liu and Z. Huang, Catal. Today, 2008, 131, 322–329.
22. Y. Guo, Z. Liu, Z. Huang, Q. Liu and S. Guo, Ind. Eng. Chem. Res., 2005, 44, 9989–9995.
23. Z. Zhu, Z. Liu, H. Niu, S. Liu, T. Hu, T. Liu and Y. Xie, J. Catal., 2001, 197, 6–16.
24. Q. Liu and Z. Liu, Fuel, 2013, 108, 149–158.
25. P. Li, Z. Liu, Q. Li, W. Wu and Q. Liu, Ind. Eng. Chem. Res., 2014, 53, 7910–7916.
26. W. Jing, Q. Guo, Y. Hou, G. Ma, X. Han and Z. Huang, Catal. Commun., 2014, 56, 23–26.
27. V. N. Krasil'nikov and V. G. Bamburov, Dokl. Chem., 2007, 416, 247–250.
28. Z. Zhu, Z. Liu, S. Liu and H. Niu, Appl. Catal., B, 1999, 23, L229–L233.
29. W. Jing, Q. Guo, Y. Hou, X. Han and Z. Huang, Korean J. Chem. Eng., 2014, 31, 794–800.
30. W. Jing, Y. Hou, Q. Guo, Z. Huang, X. Han and G. Ma, J. Fuel Chem. Technol., 2013, 41, 1223–1233.
31. X. Xing, Z. Liu and J. Yang, Fuel, 2008, 87, 1705–1710.
32. M. O. Guerrero-Pérez, J. M. Rosas, R. López-Medina, M. A. Bañares, J. Rodríguez-Mirasol and T. Cordero, J. Phys. Chem. C, 2012, 116, 20396–20403.
33. A. P. Tyutyunnik, V. N. Krasil'nikov, V. G. Zubkov, L. A. Perelyaeva and I. V. Baklanova, Russ. J. Inorg. Chem., 2010, 55, 501–507.
34. I. Giakoumelou, V. Parvulescu and S. Boghosian, J. Catal., 2004, 225, 337–349.
35. Z. Zhu, H. Niu, Z. Liu and S. Liu, J. Catal., 2000, 195, 268–278.
36. Z. Zhu, Z. Liu, H. Niu and S. Liu, J. Catal., 1999, 187, 245–248.
37. E. García-Bordejé, J. L. Pinilla, M. J. Lázaro, R. Moliner and J. L. G. Fierro, J. Catal., 2005, 233, 166–175.
38. T. Tong, J. Chen, S. Xiong, W. Yang, Q. Yang, L. Yang, Y. Peng, Z. Liu and J. Li, Catal. Sci. Technol., 2019, 9, 3779–3787.
39. X. Xing, Z. Liu and J. Wang, Fuel Process. Technol., 2007, 88, 717–722.
40. J. Wang, J. Yang and Z. Liu, Fuel Process. Technol., 2010, 91, 676–680.

---