Investigation of the Effect of SO2 and H2O on VPO-Cr-PEG/TiO2 for the Low-Temperature SCR de-NOx

A catalyst 0.1VP(0.2)O-Cr(0.01)-PEG(1×10-4)/Ti for low-temperature SCR de-NOx was developed and the effects of SO2 and water vapor on its catalytic activity were investigated. Cr doping increases the molar ratio of V5+ to V4+ on the VPO catalyst, which promote oxidation of NO to NO2 and catalytic activity of the VPO catalyst. An appropriate amount of redox-coupled V5+/V4+ also promotes catalytic activity of the VPO catalyst. The denitration efficiency over 0.1VP(0.2)O-Cr(0.01)-PEG(1×10-4)/Ti was above 98% at 150-350℃. Cr doping also promotes the generation of Lewis acid around V5+ center and BrOnsted acid (P-OH) on the VPO catalysts. The strong surface acidity of 0.1VP(0.2)O-Cr(0.01)-PEG(1×10-4)/Ti could restrain the adsorption of and oxidation of SO2. Characterization (FT-IR, TG) results indicate that nearly no sulfate was deposited on the surface of 0.1VP(0.2)O-Cr(0.01)-PEG(1×10-4)/Ti after activity test. The competitive adsorption of water molecules and nitric oxide on the catalyst surface decreases the catalytic activity of the VPO catalyst. The addition of Cr and PEG could increase the surface area and the exposure of active site of VPO catalyst, which also increase the unoccupied active sites of VPO catalysts in the presence of water vapor. The catalyst 0.1VP(0.2)O-Cr(0.01)-PEG(1×10-4)/Ti for low-temperature SCR de-NOx shows high catalytic activity and exhibits high resistance to SO2 and water vapor.


INTRODUCTION
In addition to coal-fired flue gas, industrial waste gases such as coke oven flue gas, sintering flue gas, and garbage incineration flue gas are other major sources of NO x (Chen et al., 2015;Gamrat et al., 2016;Li et al., 2018). Among these emission sources, NO x is discharged mainly in the form of nitric oxide (NO). The selective catalytic reduction (SCR) process is currently considered to be an efficient technology for controlling NO x emissions (Niu et al., 2016;You et al., 2017). Although the V 2 O 5 -WO 3 (MoO 3 )/TiO 2 catalyst has been widely adopted as a superior catalyst for controlling NO x in coal-fired flue gas, the highest reactive activity of V 2 O 5 -WO 3 (MoO 3 )/TiO 2 is usually at the temperature range of 300-400 • C (Niu et al., 2016;You et al., 2017). However, the temperature of coke oven flue gas, sintering flue gas, and garbage incineration flue gas is below 250 • C (Chen et al., 2015;Gamrat et al., 2016). The development of catalyst for low-temperature SCR denitration is extremely urgent for the control of NO x emission from industrial flue gas.
Much effort has focused on developing catalysts for lowtemperature denitration. Different metal oxides, supporters, and polymetallic catalysts have been investigated for low-temperature denitration (Andreoli et al., 2015;Stakheev et al., 2015;Cha et al., 2016;Chen et al., 2016;Wang et al., 2016). The denitration efficiency of some of these catalysts could reach more than 95% at temperature of 150 • C. However, catalyst activities were generally inhibited by SO 2 and water vapor because of deposition of (NH 4 ) 2 SO 4 and metal sulfates on the surface of catalyst (Cha et al., 2016;Wang et al., 2016). Accordingly, the development of low-temperature denitration catalysts with excellent SO 2 resistance and water resistance is urgent.
The surface acidity of a catalyst plays an important role in the SCR denitration process (Cha et al., 2016). The Eley-Rideal mechanism indicates that NH 3 is adsorbed on the Brønsted acid site and transforms into NH + 4 ions (Yu et al., 2016). This species reacts with NO to form an activated complex and then decomposes into N 2 and H 2 O. SO 2 is an acidic gas, the adsorption of SO 2 on the surface of the catalyst would be restrained by increasing the surface acidity of the catalyst, which could promote the sulfation resistance of SCR catalyst. Vanadium is a widely used as a catalyst component because of its superior redox properties (Busca et al., 1998;Phil et al., 2008;Zhao et al., 2015). Vanadium phosphorus oxides (VPO), which is prepared by reacting vanadium oxides and phosphoric acid, is a type of heteropolyacid. Busca et al. investigated the oxidation of alkanes with VPO and the results show that there are abundant Brønsted acid sites (V-OH, P-OH) on the surface of VPO catalyst (Busca et al., 1986;Bond, 1991;Feng et al., 2015). Moreover, the polarization effect of the V-(O-P) bond could promote the formation of Lewis acid sites on the V 4+ center due to the high electronegativity of P (Bond, 1991;Benziger et al., 1997).
Water vapor is another factor which degrades the catalytic performance of low-temperature denitration catalysts. Zhang et al. proposed that the competitive adsorption of water molecules and nitric oxide on the catalyst surface is responsible for the degraded catalytic performance (Zhang et al., 2014). Adsorption and desorption equilibrium of water vapor exist on catalyst surface and the proportion of active sites occupied by water molecules is constant at a certain temperature (Melánová et al., 1999). The amount of unoccupied active sites on the surface of catalysts increases as the specific surface area of the catalyst increases. Accordingly, optimizing the structure and specific surface area of the catalyst could increase its catalytic activity in the presence of water vapor.
The surface acidity and structural properties of a catalyst are closely related to its resistance to SO 2 and water vapor (Cha et al., 2016). Vanadium phosphorus oxides (VPO) have been investigated for the oxidation of alkane by several researchers because of their excellent redox properties and surface acidities (Busca et al., 1986;Bond, 1991;Benziger et al., 1997;Melánová et al., 1999;Feng et al., 2015). However, VPO catalyst has been rarely investigated for SCR denitration and the specific surface area of VPO is usually below 20 m 2 /g. Bagnasco et al. reported that the layered crystalline VPO would be formed by isomorphous substitution with some trivalent metals (Al, Cr, Fe, etc.) and the surface morphology and structural properties of VPO could also be ameliorated by trivalent metals (Bagnasco et al., 1990). In this study, Cr modified VPO catalysts with titanium dioxide as the support material were prepared and polyethylene glycol (PEG) was used to improve the dispersion of active components. The low-temperature SCR denitration performance of VPO-Cr-PEG/TiO 2 and the effect of SO 2 and water vapor on catalytic activity of VPO-Cr-PEG/TiO 2 were investigated. Combining with activity test and characterization, the physicochemical properties of the VPO-Cr-PEG/TiO 2 were also investigated.

EXPERIMENTAL Catalyst Preparation
VPO-Cr was prepared according to a liquid-phase synthesis method. Ammonium metavanadate and chromic nitrate were added to an oxalic acid solution according to a certain mole ratio and the solution was stirred for 1 h. Subsequently, the mixture was stirred for a further 2 h after adding a certain amount of phosphoric acid (85 wt.%), HCl solution(36-38 wt.%), and PEG. Thereafter, the mixture was successively evaporated at 90 • C and it was transformed into wet gel. The wet gel was dried at 105 • C for 4 h. Afterwards, the product was calcined at 350 • C for 3 h, and the active ingredient VPO-Cr-PEG was obtained.
Using commercially available TiO 2 (specific surface area is 322 m 2 /g) as support material, the VPO-Cr-PEG/TiO 2 catalyst was prepared by impregnation method. First of all, appropriate quantities of TiO 2 and VPO-Cr-PEG were added into distilled water and the mixture was stirred for 2 h. Then the mixture was evaporated in a water bath at 70 • C and the solid obtained was calcined at 350 • C for 3 h. The catalyst prepared at different conditions such as P/V (molar ratio of P to V, x), Cr/V (molar ratio of Cr to V, y), PEG/V (molar ratio of PEG to V, z) and the weight percentage of active ingredient (w) were denoted as wVP(x)O-Cr(y)-PEG(z)/Ti.

Catalyst Characterization
The specific surface area and average pore size were measured by a surface area analyzer (V-sorbet 2008S). The morphologies of the catalysts were examined with a scanning electron microscope (SEM) operated at 30 kV (JEOLJSM-6380LV). The chemical composition of the catalysts was analyzed by X-ray photoelectron spectroscopy (ESCALAB 250). The redox performance and NO adsorption properties of the catalysts were measured with a Chembet Pulsar TPR/TPD. The crystal structure was determined by X-ray diffraction (XD-3). Infrared spectroscopy measurements were performed with an FT-IR spectrometer (Bomen MB154S).

Activity Test
The SCR denitration activities of the VPO-Cr-PEG/TiO 2 catalysts were measured in a conventional fixed-bed reactor (inner diameter = 8 mm). The diagram of the experimental system is present in Figure 1.
Flue gas was simulated by mixing 0.05 vol.% NO, 0.05 vol.% NH 3 , 6 vol.% O 2 , 0-0.1 vol.% SO 2 , and 0-16 vol.% water vapor. The mixture was balanced with N 2 . The flow rate of each gas was controlled by mass flowmeter. The total gas flow rate was set to 100 mL/min and the corresponding gas hourly space velocity was about 15,000 h −1 . Liquid water was injected into the gas pipeline located at the center of a preheater (120 • C) and it was evaporated to form water vapor. The NO x concentration at the inlet and outlet of the fixed-bed reactor were measured with a gas analyzer (MRU Varioplus). The denitration efficiency η is expressed as: (1)

SCR Denitration Activity
The NH 3 -SCR deNO x activity of the catalysts were tested at 500 ppm of NH 3 , 500 ppm of NO, 6 vol.% of O 2 , and 15,000 h −1 GHSV. The denitration efficiency of catalysts at different temperature are shown in Figure 2.
Obviously, the denitration efficiencies of the catalysts increase as temperature increases. The denitration efficiency of 0.1VP(0.2)O/Ti is 93% at 150 • C and reaches 99.3% at temperatures above 200 • C. The denitration efficiencies of Cr-PEG-modified VPO/Ti catalysts are above 98% at temperatures in the range 150-350 • C. Obviously, the addition of Cr and PEG could improve the catalytic activity of VPO/Ti, especially at low reaction temperature.

Influence of SO 2 on Catalytic Activity
Selectivity of catalyst is an important parameter for SCR denitration performance. The feed gas consisted of 0.05 vol.% NO, 0.08 vol.% SO 2 , 6 vol.% O 2 and balanced with N 2 was passed through catalyst, and the concentration of SO 2 and NO x in the outlet gas were measured. Corresponding results are presented in Figures 3, 4.

Influence of Water Vapor on Catalytic Activity
The NH 3 -SCR deNO x activity of catalysts were tested in a feed gas containing 0.05 vol.% NH 3 , 0.05 vol.% NO, 6 vol.% O 2 , and 4-8 vol.% water vapor. The effect of water vapor concentration on the NH 3 -SCR deNO x catalytic activity of catalysts is shown in Figure 6.
One can see from Figure 6 that the denitration efficiency of 0.1VP(0.2)O/Ti decreased from 99.8 to 99.3% as water vapor concentration increased from 4 to 6 vol.%. A similar trend was  Figure 7. It can be seen from Figure 7 that the catalytic activity did not decrease when the SO 2 concentration was below 400 ppm. As SO 2 concentration increased from 400 to 1,000 ppm, the denitration efficiency over 0.1VP(0.2)O/Ti decreased from 99.4 to 64.5% and the denitration efficiency over 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti descreased from 99.7 to 81%. Obviously, the Cr-PEG-modified VPO/Ti catalyst has higher catalytic activity than the VPO/Ti catalyst in the presence of SO 2 and water vapor.
The sulfation of active components and the surface deposition of ammonium sulfate of catalysts are the primary reason for the decline of the NH 3 -SCR deNO x catalytic activity in the presence of SO 2 and water vapor (Cha et al., 2016;Chen et al., 2016;Wang et al., 2016). SO 2 and water vapor were removed from the feed gas and the denitration efficiency was measured over time. These  The reason for these experimental results may be that parts of ammonia react with SO 2 and water vapor in the stainless steel tube located between preheater and reactor. Accordingly, the amount of ammonia reacted with NO x in the reactor is insufficient, which lead to the decrease of denitration efficiency. If this is the case, the denitration efficiency would increase as the ammonia concentration increases. Keeping the concentration of other gas constant as the situation presented in Figure 7, the denitration efficiency vs. NH 3 concentration was tested with SO 2 and water vapor in the feed gas, and corresponding results are present in Figure 9. Obviously, the denitration efficiency increases as NH 3 concentration increases with SO 2 and water vapor in the feed gas. When the concentration of NH 3 was above 700 ppm, the denitration efficiency over 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti is almost the same as that without SO 2 and water vapor in the feed gas. The denitration efficiency of 0.1VP(0.2)O/Ti reaches 98.3% when the NH 3 concentration is above 0.08 vol.%.
In order to further confirm the above inference, the stainless steel tube between the preheater and the reactor was heated to 120 • C by an electric heating belt, and the NH 3 -SCR deNO x activity of catalysts was tested in the presence of SO 2 and water vapor. The generation of ammonium sulfite in the stainless steel tube could be restrained by electric heating belt because the decomposition temperature of ammonium sulfite is about 60 • C. The experimental results in Figure 10 show that the denitration efficiency of 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti is almost the same as that without SO 2 and water vapor in the feed gas. Moreover, the denitration efficiency did not decrease after 20 h of testing.

BET Surface Area of Catalysts
The BET surface areas of the VPO catalysts are displayed in Table 1. It can be seen from Table 1 that Cr doping and PEG could increase the surface area of the VPO/Ti catalyst. The surface area of 0.1VP(0.2)O/Ti is about 171 m 2 /g, while that of 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti is about 217 m 2 /g. A similar tendency was observed by Wang and Balaganapathi (Wang et al., 2003;Balaganapathi et al., 2016). PEG could reduce the surface energy during the nucleation process of the   active component, which is favorable for uniform nucleation and inhibits the growth of the crystal particle. Meanwhile, PEG decomposes during the calcination of catalysts, and micropores are formed in the catalysts. Accordingly, the surface area of VPO/Ti increases by adding PEG during preparation. Competitive adsorption of water molecules and NO x on the catalyst decreases the denitration efficiency. The adsorptiondesorption balance of water vapor exists on the surface of the catalysts at any temperature. Parts of active sites on the catalyst are occupied by water vapor, which inhibits the adsorption of other reactants. Cr doping and PEG increase the surface area and the number of unoccupied active sites on the VPO catalysts in the presence of water vapor, which enhances the water vapor resistance of the VPO catalysts. The BET results are in agreement with the activity tests in Section Influence of water vapor and SO 2 on catalytic activity.

XPS Analysis
The chemical state and atomic concentration of elements on the surface of the catalysts were investigated by XPS. The characterization results are shown in Figure 11 and Table 2. Figure 11A shows that the Ti2p spectra of both 0.1VP(0.2)O/Ti and 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti could be fitted by two peaks corresponding to Ti 2p 1/2 at 464.5 eV and Ti 2p 3/2 at 458.8 eV (Reddy et al., 2006). The binding energy centered at 464.5 and 458.8 eV correspond to Ti 4+ . Obviously, Cr doping and  PEG have little effect on the binding energy of Ti 2p 1/2 and Ti 2p 3/2 in the 0.1VP(0.2)O/Ti catalyst. Figure 11B indicates that the overlapped O 1s peak could be fitted by two peaks at 530.02 and 531.2 eV, which represent lattice oxygen O β and surface chemisorbed oxygen O α , respectively (Larachi et al., 2002).  (VOPO 4 ) and the oxidation of NO to NO 2 , which is favorable for lowtemperature SCR activity (Volta, 2001;Taufiq-Yap et al., 2010). An appropriate amount of redox-coupled V 5+ and V 4+ also promotes the catalytic activity of the VPO catalyst (Centi, 1993;Stakheev et al., 2015;Ren et al., 2016;Salazar et al., 2016). Figure 11D shows the Cr 2p XPS spectra of 0.1VP(0.2)O-Cr(0.01)-PEG(1×10 −4 )/Ti. The overlapped peaks at 576-580 eV could be assigned to Cr (Ren et al., 2016). The mixed peaks in Figure 11D are difficult to deconvolute, as the Cr content is relatively low.

NH 3 -IR Analysis
FT-IR spectroscopy was used to analyze the surface acidity of 0.1VP(0.2)O/Ti and 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti due to ammonia adsorption. Figure 12 shows the spectrum of the adsorbed NH 3 species on the surface of the two catalysts. The band centered at 1,401 cm −1 is ascribed to N-H bending vibrations in chemisorbed NH 3 on Brønsted acid sites, while the bands centered at 1,619 cm −1 are assigned to NH + 4 on Lewis acid sites.
As Figure 12 shows, the addition of PEG and Cr to VPO/Ti could strengthen the intensity of the Brønsted acid sites and Lewis acid sites. This could be explained as follows: as the surface area of 0.1VP(0.2)O/Ti increases by adding PEG and Cr, the number of exposed surface acid sites also increases. Meanwhile, Cr doping leads to phosphorus deposition on the surface of 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti, which could promote the formation of a Brønsted acid (P-OH). Moreover, the addition of Cr in VPO/Ti could increase the amount of V 5+ and Lewis acid sites. A similar tendency was also observed by Pierini (Pierini and Lombardo, 2005a,b).
Brønsted acid sites on the catalyst are widely recognized as active sites for NH 3 -SCR deNO x . The NH 3 -IR results in Figure 12 are in good agreement with the activity tests in Figure 2. Furthermore, 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti with strong surface acidity could restrain the surface adsorption of SO 2 . The NH 3 -IR spectrum further confirms the experimental results in Section Influence of SO 2 on catalytic activity and Section Influence of water vapor and SO 2 on catalytic activity. Figure 13 shows the FTIR spectra of 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti with/without SO 2 and water vapor in the feed gas after the activity test. The characteristic peak centered at 502 cm −1 is attributed to symmetric stretching vibrations in TiO 2 . The band at around 1,053 cm −1 is ascribed to asymmetric stretching vibrations of V 5+ = O. The band at 1,382 cm −1 is attributed to nitrate. The bands around 1,450, 1,631, and 3,432 cm −1 could be attributed to NH 3 , an -OH group, and H 2 O, respectively. The bands centered at 2,852 and 2,923 cm −1 are ascribed to symmetric stretching and asymmetric stretching vibrations in methylene, respectively. Obviously, nearly no difference was observed in the FT-IR spectra for 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti after activity test with/without SO 2 and water vapor in feed gas. Moreover, the FT-IR spectra also indicate that no sulfate formed on the surface of catalyst after the activity test. The 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti shows excellent resistance to SO 2 and water vapor (Figure 10), which is further confirmed by the FT-IR spectra.

Thermogravimetric Analysis (TG)
The TG curves of 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti before and after the activity test with SO 2 and water vapor in the feed gas were present in Figure 14. Obviously, water is lost in the temperature range 25-90 • C before and after the activity test. Figure 14 The bound water linked to phosphate groups is lost in the temperature range 70-130 • C, and that linked to Cr 3+ and VO 3+ is lost between 150 and 250 • C. Furthermore, mass was lost in the 300-620 • C temperature range for both catalysts. This loss could be ascribed to water being coordinated to Cr in the catalysts.
Compared to the 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti catalyst before the activity test, no accelerated mass loss in the range 200-400 • C was observed in the TG curves for the 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti catalyst after the activity test. Ammonium sulfate decomposes at a temperature of 230 • C, and ammonium bisulfate decomposes at a temperature of 350 • C. The TG results further confirmed that there are no ammonium sulfate and ammonium bisulfate deposited on the 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti, which is consistent with the FTIR results shown in Figure 10. This could be explained as follows: SO 2 adsorption on the surface of 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti was restrained because of the strong surface acidity of catalysts. Meanwhile, SO 2 is hardly to be oxidize to SO 3 in the gas phase without catalyst. The ammonium sulfite generated by reaction of SO 2 with NH 3 is thermally unstable at a reaction temperature of 200 • C.

CONCLUSION
A low-temperature SCR de-NO x catalyst 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti was developed and the effects of SO 2 and water vapor on the catalytic activity of catalyst were investigated. The activity test results show that the denitration efficiency over 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti was above 98% at temperatures in the range 150-350 • C. Cr doping increases the molar ratio of V 5+ to V 4+ on the VPO catalyst, which promote oxidation of NO to NO 2 and catalytic activity of the VPO catalyst. An appropriate amount of redox-coupled V 5+ /V 4+ promotes the catalytic activity of the VPO catalyst. The addition of Cr increases Brønsted acid (P-OH) and Lewis acid sites around V 5+ on the VPO catalysts. The NH 3 -IR spectra show that the intensity of the Brønsted acid (P-OH, V-OH) and Lewis acid of the VPO catalyst increases after Cr doping. 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti catalyst with strong surface acidity could restrain the SO 2 adsorption and the oxidation of SO 2 on the catalyst. SO 2 had little effect on the catalytic activity of 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti. Characterization (FTIR, TG) results indicate that there was no ammonium sulfate deposited on the surface of 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti and the active components of 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti were not sulfated in the presence of SO 2 in feed gas. Generally, competitive adsorption of water molecules and reactant on the active site decreases the catalytic activity of low temperature NH 3 -SCR denitration catalyst. The addition of Cr and PEG increase the surface area and the unoccupied active sites in the VPO catalysts in the presence of water vapor, which enhance the water vapor resistance of the VPO catalysts. The low-temperature SCR deNO x catalyst 0.1VP(0.2)O-Cr(0.01)-PEG(1 × 10 −4 )/Ti shows high catalytic activity and exhibits good resistance to SO 2 and water vapor.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/supplementary material.

AUTHOR CONTRIBUTIONS
YJ, JJ, and JY was mainly responsible for the preparation and testing of catalysts.
LG is mainly responsible for the characterization of catalysts. MG, YZ, and YC put forward their opinions on the grammar and structure of the whole paper.