Enhancement of CeO2 modified commercial SCR catalyst for synergistic mercury removal from coal combustion flue gas

CeO2 modified commercial SCR (selective catalytic reduction) catalysts with different CeO2 content were prepared and researched for synergistic mercury removal from coal combustion flue gas in this study. The characterization analyses on the catalysts indicated that the introduction of CeO2 increased the surface area, the dispersity of the metal oxides on the TiO2 support and the redox behavior of the catalyst, which was beneficial to the catalytic activity. The experimental results confirmed that the CeO2 loading improved the catalytic efficiencies over the commercial SCR catalyst. The catalyst with a CeO2 content of 4% displayed the optimal performance for NO and synergistic Hg0 removal, of which the NO conversion and Hg0 removal efficiency reached 90.5% and 78.2%, respectively, at 300 °C in simulated coal-fired flue gas. The Hg0 removal activity, the independence of Hg0 removal from HCl concentration and the effects of SO2, NO and NH3 on Hg0 removal efficiency all became positive over the modified catalyst compared to over the raw one, which was mainly due to the sufficient chemisorbed oxygen derived from the synergy of V2O5 and CeO2 and the redox transformation between Ce3+ and Ce4+ on the catalyst surface. The CeO2 modification generated a significant enhancement on the catalytic performance and made the commercial SCR catalyst more suitable to be employed for NO and synergistic mercury removal in a coal combustion power plant.


Introduction
Mercury is a kind of extremely harmful pollutant in the ecological environment. It poses a serious threat to human health due to its hypertoxicity, persistence and bioaccumulation. 1 According to the Global Mercury Assessment 2018 issued by the UN Environment Programme, the global anthropogenic mercury emission reached 2150 tons in 2015, which increased by 12% compared to that in 2010. 2 Signicant coal burning is one of the main reasons for the growth of mercury emissions. And coal combustion power plants are considered as the major anthropogenic source of mercury release. 3 As the Minamata Convention on Mercury came into force in August 2017, the limit on mercury emission from coal-red power plants will be more rigorous on the basis of the existing regulations. 4 Therefore, it is urgent to pay extensive attention to mercury emission control of coal combustion power plants under the dual pressure of environmental protection and convention fulllment.
Mercury in coal-red ue gas exists mainly in the types of elemental Hg (Hg 0 ), oxidized Hg (Hg 2+ ) and particle bound Hg (Hg P ). Hg 2+ and Hg p can be respectively captured by wet ue gas desulfurization (WFGD) and particulate matter control device (PMCD) of power plant because of their physical properties, while Hg 0 is difficult to be controlled by the single pollutant control equipment due to its volatility and water insolubility. 5,6 So the key to the control of mercury emission from coal combustion power plant is the removal of Hg 0 . Similarly with mercury, NO x is also a sort of hazardous contaminant with great harm to environment that coal burning releases, and NO occupies about 95% among NO x . [7][8][9] Currently, the method of selective catalytic reduction (SCR) is generally used by coal-red power plants for NO removal. Besides, the SCR catalyst has the capacity of oxidizing Hg 0 to Hg 2+ due to the existence of active oxygen on its surface, followed by Hg 2+ being removed in the downstream WFGD. 10,11 Compared with other Hg removal plans such as sorbent injection, utilizing SCR catalyst to synergistically remove Hg is remarkably cost-effective and meanwhile benecial to the avoiding of secondary mercury pollution. 12 Hence, it is promising for coal-red power plant to adopt this approach to deal with the Hg removal from ue gas. And the research on the synergistic Hg 0 oxidation with SCR catalyst has attracted more attention in recent years.
The commercial SCR catalyst that is currently used by coal combustion power plants is the TiO 2 -supported V 2 O 5 -WO 3 /TiO 2 catalyst. A series of studies have been made on the Hg 0 oxidation over the V 2 O 5 -WO 3 /TiO 2 catalyst. The results indicated that the V]O bond on the catalyst surface could participate in Hg 0 oxidation as the active sites. The Hg 0 removal efficiency over the catalyst could reach 60-80% in general, and sometimes the efficiency was even higher than 90%. 13,14 The increases of V 2 O 5 loading, surface area and reaction temperature are in favor of the Hg 0 oxidation activity. 15 Especially, the existence of HCl in the ue gas had an obvious promotion on the Hg 0 oxidation over the V 2 O 5 -based catalysts. Hg 0 removal efficiency of V 2 O 5 -WO 3 /TiO 2 was close to 100% at 380 C with 4.5 mmol m À3 HCl contained in the reaction gas. 16 The SiO 2 -TiO 2 -V 2 O 5 catalyst likewise showed a Hg 0 removal efficiency of nearly 100% in the co-presence of O 2 and HCl. 17 And the facilitation of HCl on the efficiency of commercial SCR catalyst was also testied by kinetic analysis. 18 However, though the commercial V 2 O 5 -WO 3 / TiO 2 catalyst displays certain Hg 0 removal capacity under the appropriate conditions, it has apparent drawbacks such as the narrow working temperature range and the limited Hg 0 removal efficiency at the SCR operating temperature. 16,19 Meanwhile, the effectiveness of Hg 0 removal depends heavily on the HCl concentration. The efficiency could be as high as 90% in the ue gas derived from burning high-rank coal, while in ue gas of burning low-rank coal only less than 30% was observed. 17,20,21 This condition is distinctly disadvantageous to those power plants that combust sub-bituminous coal or lignite. So it is necessary to make modication on commercial SCR catalyst to improve its catalytic properties. In recent years, CeO 2 -based catalysts have gradually come into view of researchers due to its prominent catalytic activity. Related studies demonstrated that element Ce would help enhance the oxygen storage capacity of the catalyst, which led to the superior performance on NO and Hg 0 removal. Illustratively, Gao et al. 22 prepared CeO 2 /TiO 2 catalyst by sol-gel method and found the NO conversion of the catalyst reached 93.4-98.6% in the wide temperature range of 250-450 C; Li et al. 23 investigated Hg 0 removal activity of CeO 2 / TiO 2 in simulated coal-red ue gas and conrmed the optimal efficiency could attain 94%, and efficient Hg 0 oxidation could be achieved even in the absence of HCl; Fan et al. 24 acquired that the zeolite supported CeO 2 /HZSM-5 catalyst exhibited Hg 0 removal efficiency of more than 95% among the range of 120-320 C; Wang et al. 25 loaded CeO 2 on Ti-based pillared interlayered clays to examine the simultaneous NO and Hg 0 removal efficiency over the catalyst, and the results showed that the NO conversion was almost 100% at 350 C while Hg 0 removal efficiency also reached higher than 50% in the same condition. In view of the advantage of the activity of catalyst containing CeO 2 , it is reasonable to speculate that using CeO 2 to modify the V 2 O 5 -WO 3 /TiO 2 catalyst will make a signicant improvement on the catalytic properties of the catalyst. Zhao et al. 19 has previously modied the TiO 2 support with CeO 2 and synthesized V 2 O 5 -WO 3 /TiO 2 -CeO 2 catalyst, and the experimental study conrmed the enhancement of Hg 0 removal performance of the catalyst, such as the efficiency and sulfur-resistance, resulted from the addition of CeO 2 . Some literatures also prepared the CeO 2 modied V 2 O 5 -WO 3 (MoO 3 )/TiO 2 to investigate the NO removal activity specically, and the satisfactory NO conversions, sulfurresistance and alkali metal resistance were obtained over the catalysts. [26][27][28] Nevertheless, few literatures have made investigations on the effectiveness of employing CeO 2 to directly modify the commercial SCR catalyst of power plant for synergistic Hg 0 removal so far, which is of great value and close correlation to practical application. Moreover, the present commercial SCR catalyst is not replaceable in the short term, though some researched novel catalysts such as Mn-based, Cubased, noble metal and perovskite structure catalysts displayed considerable Hg 0 removal efficiency in the lab-scale tests. [29][30][31][32] Thus, it can be seen that it is of great signicance to examine the synergistic Hg 0 removal performance of the CeO 2 modied commercial V 2 O 5 -WO 3 /TiO 2 catalyst.
Based on the above presentations, this work takes CeO 2 modied commercial SCR catalyst as the researching object, and conducts the experiments in simulated coal combustion ue gas (SFG). NO removal performance of the catalysts with different CeO 2 loadings were rst tested considering the primary purpose of SCR. Then the Hg 0 removal activity of the catalysts was investigated in detail. Hg 0 removal efficiencies over different CeO 2 -loading catalysts at different temperatures were evaluated, and the effects of individual ue gas components in SFG on the efficiency were detected as well. The characterization analyses of X-ray uorescence (XRF), Brunauer-Emmett-Teller (BET), X-ray diffraction (XRD), H 2 -Temperature Programmed Reduction (H 2 -TPR) and X-ray photoelectron spectroscopy (XPS) were carried out to understand physicalchemical properties of the catalysts and explore the modication mechanism of CeO 2 . The study results of this work will present application prospect of the CeO 2 modication on commercial SCR catalyst for improving the catalytic performance.

Catalyst preparation
The honeycomb commercial SCR catalyst employed in this study was got from a catalyst corporation of China which professionally produces SCR catalyst of coal-red power plant. The CeO 2 modied catalysts were prepared by the solution impregnation method. The honeycomb catalyst was grinded to powder rst and sieved with a 200 mesh sier. Then a certain amount of the sieved ne catalyst powder was placed in a beaker, followed by the Ce(NO 3 ) 3 aqueous solution which contained the desired quantity of Ce(NO 3 ) 3 being lled into the beaker. The obtained slurry was stirred for 1 h and then exposed to an ultrasonic bath for 2 h. Aer the mixture was dried at 110 C for 12 h and calcinated in air at 500 C for 4 h sequentially, the nal CeO 2 modied commercial SCR catalyst was acquired. The mass fractions of CeO 2 of 1%, 2%, 4% and 7% in the modied catalysts were designed. In the process of preparing the catalysts with different CeO 2 loadings, the weight of the original catalyst powder was remained unchanged, and the CeO 2 loading was controlled by the solvend amount of the added Ce(NO 3 ) 3 aqueous solution. The CeO 2 modied catalysts were abbreviated as (x)CeO 2 -SCR (x represents the mass fraction of CeO 2 ) in the later sections, and the catalyst without modication was designated as raw SCR. Additionally, the pure CeO 2 catalyst was also prepared for comparison, which used Ce(NO 3 ) 3 as the precursor as well to maintain the consistency.

Catalyst characterizations
The characterization methods of XRF, BET, XRD, H 2 -TPR and XPS were carried out over the fresh and spent catalyst samples in order to understand the physical and chemical properties of the catalysts and analyze the CeO 2 modication mechanism. The XRF analysis was conducted with an EAGLE III focusing uorescence spectrograph which was operated at 38 kV. The measurement of the BET surface was accomplished on an ASAP 2020 porosimeter by means of N 2 adsorption. The XRD analysis was performed using an X'Pert PRO diffractometer (Cu Ka radiation) of which the working voltage and emission current were 40 kV and 40 mA, respectively, with the scanning angle ranging from 10 to 80 (2q). The test of H 2 -TPR was carried out on an Autochem 2920 analyzer with the operating temperature raised from 30 C to 850 C at a rate of 10 C min À1 , and the reaction gas was 50 mL min À1 10% H 2 /Ar. The XPS technique was implemented on a PerkinElmer PHI 5100 ESCA system with Al Ka X-ray source (hn ¼ 1486.6 eV) to study the valence states of the elements and using the C 1s binding energy value of 284.6 eV for the spectra calibration.

Catalytic activity measurement
The experimental system used in this work was similar to that employed in our previous studies, 33-35 as described in Fig. 1. Briey, the ue gas components (N 2 , O 2 , HCl, SO 2 , NO, and NH 3 ) came from standard cylinder gases and their gas ow was accurately controlled by the corresponding calibrated mass owmeter, respectively. Water vapor (H 2 O) was produced by a steam generator. The continuous feed of Hg 0 vapor of approximately 60 mg m À3 was generated from a Hg 0 penetration tube (VICI, Metronics Inc., Santa Clara, CA) which was placed in a U-tube and heated by a water bath, with N 2 carrying the Hg 0 vapor into the ue gas. The catalytic reaction was made to occur in a xed bed reactor with a temperature controller to set the reaction temperature. The NO and Hg 0 concentrations in the ue gas were measured by a gas analyzer (AFRISO, Multilyzer STe, M60) and a Hg 0 online monitor (Ohio Lumex, RA-915M), respectively. And the N 2 O and NO 2 concentrations were monitored with a FTIR analyzer (Gasmet Dx4000). Several specic gas-washing bottles were added for eliminating the acid gas to prevent corrosion and interferences on the monitoring devices. The gas line of the system was heated by electric heating belt to avoid any possible adsorption of the gas components on the line before the measurement. The exhaust gas was puried by active carbon before discharged to atmosphere.
The experiments of this work were carried out under the condition of simulated coal-red ue gas of which the composition was 4% O 2 , 10 ppm HCl, 800 ppm SO 2 , 400 ppm NO, 400 ppm NH 3 , 8% H 2 O and 60 mg m À3 Hg 0 with N 2 to balance unless otherwise noted. The total ow of the ue gas was controlled at 1 L min À1 . The catalyst dosage was 0.5 g for each test, and the space velocity (GHSV) was correspondingly about 50 000 h À1 . In each test, the ue gas was rst introduced to the bypass, and the concentrations of NO and Hg 0 at the inlet of the reactor were acquired when the readings of the monitoring devices reached stability. Then the gas stream was switched to pass through the catalyst until the stable NO and Hg 0 concentrations at the outlet of the reactor were obtained as well. The stability was dened as the uctuation of the concentrations being no more than 5% for a period of at least 30 min. Aer each step of the experiment, the spent catalyst was replaced by fresh sample before starting the next test. The NO conversion, N 2 selectivity and Hg 0 removal efficiency adopted to evaluate the catalytic activity of the catalyst were respectively calculated by eqn (1)-(3) as follows.
The subscript "in" and "out" in the equations represented the gas concentrations at the inlet and outlet of the reactor, respectively. As the outlet Hg 0 concentration was read when it achieved a stable value, the catalyst was in the state of Hg saturated adsorption at this time and all the removed Hg was gaseous Hg 2+ . Additionally, the researched catalysts were veri-ed to have almost no capacity for Hg 0 removal at room temperature. So the physical adsorption of Hg 0 was negligible, and the dened Hg 0 removal efficiency here was equal to Hg 0 oxidation efficiency.

Results and discussion
3.1. Characterization of the CeO 2 -SCR catalysts 3.1.1 XRF analysis. XRF analysis was adopted to investigate the element compositions and contents of the catalysts. The results were summarized in Table 1. Before the loading of CeO 2 , the content of V 2 O 5 which was the active component and the content of WO 3 using for improving the thermal stability and surface acidity in raw SCR catalyst were 0.98% and 6.63%, respectively. Both the values were among the ranges of the contents of V 2 O 5 and WO 3 in usual honeycomb commercial SCR catalyst, which were respectively 0.5-3% and 2-10%. The activity of SCR catalyst was generally in proportion to the content of V 2 O 5 . But exorbitant vanadium content would lead to the growing SO 2 /SO 3 conversion. 36 The V 2 O 5 content of the raw SCR catalyst employed in this work was a moderate percent of about 1%, indicating this catalyst was well typical and representative. Small amount of SiO 2 was also detected to contain in the catalyst, which was helpful for boosting the mechanical strength. For the CeO 2 modied catalysts, the practical contents of CeO 2 in the catalysts with different CeO 2 loadings were very close to the corresponding designed values, which testied the accuracy of the preparation of the catalysts. Meanwhile, the addition of CeO 2 did not cause apparent variations on the contents of V 2 O 5 , WO 3 and SiO 2 in the catalysts.
3.1.2 BET analysis. The surface structural properties of the CeO 2 modied commercial SCR catalysts tested by BET analysis were listed in Table 2. According to the results, the surface area of the raw catalyst was at a relatively low level of 18.64 m 2 g À1 , which might result from the specic preparation process of the catalyst corporation. The introduction of CeO 2 made a signicant enhancement on the surface area and pore volume of the catalyst. The surface area increased dramatically from 18.64 m 2 g À1 to 69.23 m 2 g À1 with the loading of only 1% CeO 2 . The increase of surface area could raise the amount of the active sites available for Hg 0 and other reactants on the catalyst surface, thereby it usually being benecial to the catalytic activity. 35,37 And the enlargement of pore volume was also in favor of the Hg 0 removal capacity of the catalyst. The surface area showed a slight declined trend as the CeO 2 loading augmented, which was probably due to the blockage of some surface micropores caused by the increasing CeO 2 loading. 38,39 It's worth noting that the surface area of the CeO 2 modied catalysts was much closer to that of pure CeO 2 than to the raw SCR catalyst, indicating that the surface area was obviously altered and controlled by CeO 2 though it occupied only a minor proportion in the catalysts. By contrast, the pore size of the catalyst was not distinctly affected by the addition of CeO 2 , and the change was small.
3.1.3 XRD analysis. The crystalline characteristic in the catalysts was investigated by XRD analysis, and the result was shown in Fig. 2. On the patterns of raw SCR catalyst and pure CeO 2 , only the peaks corresponding to anatase TiO 2 and CeO 2 were discovered respectively. 23,30,40 With CeO 2 doped into the commercial SCR catalyst, the peak intensity of TiO 2 became weak gradually, and meanwhile the peak standing for CeO 2 was not detected when the CeO 2 content was lower than 4%. This   phenomenon suggested that there existed some interaction between TiO 2 and CeO 2 in the catalysts. 33,41,42 CeO 2 was well dispersed and in the form of amorphous phase on the catalyst surface. As the CeO 2 content reached 4%, a peak corresponding to CeO 2 emerged on the pattern at 28.6 , indicating that the present load amount has made the dispersion of CeO 2 on the catalyst reach the critical point of saturation. Further increasing the CeO 2 loading would lead to the conversion of the doped CeO 2 from amorphous phase to crystalline state. The emergence of distinct characteristic peaks corresponding to CeO 2 on the prole of 7% CeO 2 -SCR conrmed this inference. In addition, the peaks of V 2 O 5 and WO 3 were not discovered on any catalyst pattern, displaying an amorphous distribution as well.
More active substance existed in the amorphous phase was considered to be advantageous for the catalytic activity of the catalyst, while the appearance of the crystal of the active species was adverse to the catalytic performance. 43,44

NO removal performance of the CeO 2 -SCR catalysts
Considering the primary function of SCR catalyst was to remove NO for coal combustion power plant, NO removal activity of the CeO 2 modied commercial SCR catalysts in simulated coal-red ue gas was rst examined prior to the investigation on Hg 0 removal performance. The experimental results were shown in Fig. 3. The NO conversions of the catalysts showed a growing trend as the reaction temperature increased from 150 C to 400 C. The optimal temperature range was 300-400 C which was consistent with that of literature report. 41,45,46 NO conversion over the raw SCR catalyst in this range was 74.6-84.3%, which was a little lower than the efficiencies monitored in power plants. This might be attributed to the higher GHSV in the lab reactor than that under the practical conditions (2000-3000 h À1 ), 47 which led to the shorter contact time between ue gas and catalyst. As CeO 2 was added into the catalyst, NO conversion was apparently promoted. And the catalyst with the CeO 2 loading of 4% exhibited the best activity for NO removal. The NO conversions were 90.5%, 92.5% and 89.3%, respectively, at the temperature points of 300-400 C over 4% CeO 2 -SCR. Besides, the efficiency of 4% CeO 2 -SCR could also reach nearly 80% at 250 C. Thus, the CeO 2 modication not only improved NO conversion of commercial SCR catalyst, but also broadened the working temperature and enhanced the medium-low temperature activity of the catalyst. The superior NO removal performance of 4% CeO 2 -SCR was associated with the higher content of CeO 2 dispersed in the amorphous phase, while the slightly decreased NO conversion over 7% CeO 2 -SCR compared to that over 4% CeO 2 -SCR might be due to the generation of CeO 2 crystal in the catalyst. Additionally, the surface area was also a possible inuence factor for the NO removal activity because the variation trend of the surface area was very close to that of the NO conversion among 4% CeO 2 -SCR, 7% CeO 2 -SCR and the raw catalyst. Therefore, the experimental acquirement was in good agreement with the characterization results. The efficiency of pure CeO 2 was in a poor level among the testing temperature range, indicating that the element V was still responsible for the nice NO removal activity of CeO 2 -SCR though CeO 2 generated modication effects on the catalysts. To sum up, the CeO 2 modication led to an advancement upon the property of the commercial SCR catalyst and made it own prominent NO removal activity, which established a solid foundation on the utilization of the catalyst for synergistic Hg 0 removal. As another important evaluation index for NO removal performance, N 2 selectivity was measured over the 4% CeO 2 -SCR catalyst which exhibited the highest NO conversion, and the results were shown in Fig. 4. Under SFG, the N 2 selectivity over the catalyst reduced slightly with the increase of the reaction temperature, which was caused by the generation of a small amount of N 2 O and NO 2 during the reaction. The detected concentrations of N 2 O were much higher than those of NO 2 . So the decrease of the N 2 selectivity was mainly due to the N 2 O generation at the higher temperatures. Nevertheless, the N 2 O generation was lower than 15 ppm in the whole temperature range of 150-400 C, and even the poorest N 2 selectivity measured at 400 C reached as high as 90.5%. Hence, the catalyst displayed great N 2 selectivity in the NO removal process, further conrming the excellent NO removal performance of the CeO 2 -SCR catalyst in the simulated coal-red ue gas.  This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 25325-25338 | 25329 3.3. Hg 0 removal performance of the CeO 2 -SCR catalysts 3.3.1 Hg 0 removal efficiency under different temperatures in SFG. Hg 0 removal performance of the CeO 2 modied commercial SCR catalysts was then investigated as the emphasis. First, the Hg 0 removal efficiencies of the catalysts in simulated coal-red ue gas were measured under different reaction temperatures, and the results were shown in Fig. 5. As the temperature increased, the variation trend of the Hg 0 removal efficiencies of the CeO 2 -SCR catalysts was opposite to that of the NO conversions, and it was a descending tendency. The possible reason for this phenomenon was that the lower temperature was benecial to the Hg 0 adsorption on the catalyst which was an essential procedure for Hg 0 removal, and the Hg 0 oxidation was realized mainly through the form of adsorbed Hg 0 (Hg 0 ad ). 35,42,48 The introduction of CeO 2 into the catalyst accelerated the Hg 0 removal efficiency apparently. Analogously to the testing results of NO removal activity, the optimal sample for Hg 0 removal was 4% CeO 2 -SCR as well, which corresponded to the characterization results again. Hg 0 removal efficiency of 4% CeO 2 -SCR achieved more than 90% in the temperature range of 150-250 C. Even at 300 C which was among the conventional operating temperature of SCR catalyst (300-400 C), 4% CeO 2 -SCR also exhibited the efficiency of as high as 78.2% on the basis of NO conversion guaranteed at 89.3%. So the catalyst showed remarkable activity for simultaneous NO and Hg 0 removal. The prominent performance for synergistic Hg 0 removal was mainly owed to the sufficient chemisorbed oxygen (O ad ) of 4% CeO 2 -SCR led by the existence of Ce 3+ /Ce 4+ ion pair and the oxygen transfer between them in the catalyst, 38,49 which would be conrmed by the subsequent XPS analysis. The abundant O ad would facilitate Hg 0 oxidation to generate HgO as the active species. The related reaction process was described by eqn (4)- (6). As the efficiencies of the raw catalyst and pure CeO 2 were no more than 38.3%, the superior performance of the CeO 2 modied commercial SCR catalyst was also primarily resulted from the synergy of V 2 O 5 and CeO 2 in the catalyst. 50 In addition, considering the GHSV was much higher in the experimental condition than in actual ue gas of power plant, the catalytic efficiencies might be preferable in practical application. Hence, the SCR catalyst manifested to be more competent and promising for commercial use aer the CeO 2 modication.
Hg 0 (g) / Hg 0 ad (4) Hg 0 ad + O ad / HgO 3.3.2 Effects of the ue gas components on Hg 0 removal efficiency. Effect of each ue gas component on the Hg 0 removal efficiency of the CeO 2 -SCR catalyst was then investigated to reveal its role in Hg 0 removal process. And the results were made comparison with those of the raw SCR catalyst to explore the reasons for the modication effect of CeO 2 on the catalyst for Hg 0 removal in simulated coal-red ue gas. Because the optimum catalytic efficiencies were implemented at 300 C over 4% CeO 2 -SCR with the NO conversion and synergistic Hg 0 removal efficiency being 89.3% and 78.2%, respectively, the experiments of this part were carried out at 300 C using 4% CeO 2 -SCR as the catalyst sample. The reaction atmosphere was SFG with the concentration of the investigated component changed and the others constant.
3.3.2.1. Effect of HCl. As the important oxidant for Hg 0 oxidation in coal combustion ue gas, effect of HCl on the Hg 0 removal efficiency of the catalysts was examined, and the results were shown in Fig. 6. For the raw catalyst, Hg 0 removal efficiency was disadvantaged in the absence of HCl, and the highest value was only 27%. Even though 10 ppm HCl was added into the ue gas, the efficiency was still maintained at a low level since it was below 40% in the whole temperature range. Only when the HCl concentration increased from 10 ppm to 30 ppm did the Hg 0 removal efficiency of the raw catalyst show a signicant improvement. It increased by 35.5% and 45.4%, respectively, at 250 C and 300 C as the instances. The above results veried the viewpoint in the literatures that the commercial SCR catalyst was qualied to be utilized in the ue gas derived from burning bitumite with high HCl content while not appropriate to work under low HCl concentration caused by  using low-rank coals for Hg 0 removal. 20,23 By contrast, aer CeO 2 modication, 4% CeO 2 -SCR exhibited much more prominent Hg 0 removal efficiency than raw SCR catalyst under the same HCl concentration. The performance over 4% CeO 2 -SCR was even better without HCl than that over the raw catalyst in the presence of 30 ppm HCl. A limited increase of the efficiency of 4% CeO 2 -SCR was observed as the HCl concentration raised. Nevertheless, the catalyst displayed satisfactory Hg 0 removal activity when exposed to 10 ppm HCl. The Hg 0 removal efficiencies were excellent at 150-300 C. Therefore, the CeO 2 modication weakened the dependence of Hg 0 removal activity of the catalyst on HCl content of the ue gas. This was really good news for power plants combusting sub-bituminous coal and lignite which occupied the majority of all items. The reason for the superior Hg 0 removal efficiency of 4% CeO 2 -SCR under low HCl concentration was also due to the improved content of O ad on the catalyst surface. More HCl could be converted by the abundant O ad to form active Cl (Cl*) which had strong oxidation, followed by Hg 0 being oxidized to HgCl 2 by Cl*. 51,52 Through this way, the introduced CeO 2 enhanced the HCl utilization of the catalyst. The involved reactions could be described by eqn (7) and (8). Meanwhile, this was also one of the main reasons for the higher Hg 0 removal efficiency over 4% CeO 2 -SCR compared to that over raw SCR in the simulated coal-red ue gas besides the direct oxidation by O ad .

Effect of SO 2 .
Effect of SO 2 on the Hg 0 removal efficiency was shown in Fig. 7. The variation trends of Hg 0 removal efficiency of raw SCR catalyst and 4% CeO 2 -SCR were almost the same with the rising SO 2 concentration. The efficiency was promoted rst as the SO 2 content in the ue gas increased from 0 to 800 ppm. The promotion could be explained by SO 3 generated from SO 2 oxidation, and then Hg 0 reacted with SO 3 to form HgSO 4 , 10 as described by eqn (9) and (10). The increase range of the efficiency was a little larger over 4% CeO 2 -SCR than over raw catalyst, which was probably because the adequate O ad in 4% CeO 2 -SCR converted more SO 2 to SO 3 that had the ability Fig. 7 Effect of SO 2 on Hg 0 removal efficiency of raw SCR catalyst and 4% CeO 2 -SCR in simulated coal-fired flue gas (reaction gas: SFG with 0, 400, 800, 1200 ppm SO 2 ).  This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 25325-25338 | 25331 to oxidize Hg 0 and facilitated the proceeding of eqn (10). As SO 2 content further increased to 1200 ppm, the efficiency suffered slight inhibition, which might be due to the generation of vanadium sulfate and/or cerium sulfate under the high SO 2 concentration that caused the deactivation of the catalyst to some extent. 53,54 Compared to the dramatic decrease of the Hg 0 removal efficiency over Mn-based catalysts in the presence of SO 2 , 34,55 the commercial V-based catalyst exhibited the advantage of owning excellent sulfur-resistance distinctly.
Hg 0 + SO 3 + O ad / HgSO 4 (10) 3.3.2.3. Effects of NO and NH 3 . NO and NH 3 were the principal reactants of the SCR deNO x reaction. Effects of NO and NH 3 in the ue gas on Hg 0 removal efficiency were important factors for determining the performance of a catalyst for synergistic Hg 0 removal. The testing results on raw SCR and 4% CeO 2 -SCR were shown in Fig. 8. The increase of NO concentration without the injection of NH 3 generated the inuence of promoting rst and then restraining on the efficiencies of both the catalysts, as shown in Fig. 8(a). NO could be oxidized by chemisorbed oxygen on the catalyst to NO 2 which had the capacity to oxidize Hg 0 to Hg(NO 3 ) 2 . 47 The related reactions were presented by eqn (11) and (12). And it was the reason for the improvement of the Hg 0 removal efficiency with the raise of NO concentration. As the NO content further increased aer it has reached 400 ppm, the excessive NO would lead to the generation of materials such as nitrite which had no Hg 0 oxidation capacity and easily caused pore plugging on the catalyst surface besides NO 2 , 56 resulting in the diminishment of the Hg 0 removal efficiency. Under the condition of NH 3 added, the proceeding of SCR deNO x reaction removed NO in the ue gas, and the actual concentration of NO was shrunken. Thus, it showed a gradual increase trend of the efficiency as NO content lied from 0 to 600 ppm, and the inhibition was not formed. Similarly to the effect of SO 2 , the promotion of NO on the efficiency of 4% CeO 2 -SCR was more evident than on the efficiency of raw catalyst, which was owed to the more sufficient O ad in 4% CeO 2 -SCR accelerating the proceeding of eqn (11) and (12) as well. The existence of NH 3 suppressed Hg 0 removal efficiency apparently. This judgment could be viewed more intuitively from the results in Fig. 8(b). The increase of the ratio of NH 3 /NO in the ue gas led to obvious inhibitive effect on the efficiencies over both the raw and modied catalysts. NH 3 was considered to form intense competitive adsorption with Hg 0 on the surface, hindering the  necessary Hg 0 adsorption process and also the following Hg 0 oxidation. 42,57,58 It was worth noting that the inhibition of NH 3 on the Hg 0 removal efficiency was weaker over 4% CeO 2 -SCR than over the raw catalyst. The reasonable explanation was that the modied catalyst owned stronger NO removal activity. More NH 3 was expended in NO removal reaction so that the inhibition on Hg 0 removal was weakened. In this view, the CeO 2 modication made the catalyst display better NH 3 -resistance in Hg 0 removal process, and the property of the catalyst for synergistic Hg 0 removal was thereby reinforced.

Effect of H 2 O.
A certain amount of water vapor (H 2 O) was contained in coal-red ue gas since water was one of the components of coal. Effect of H 2 O on the Hg 0 removal efficiency was investigated, and the results were shown in Fig. 9. H 2 O generated an unfavorable inuence on the efficiency. It declined by a close extent over the raw catalyst and 4% CeO 2 -SCR as 8% H 2 O was added into the ue gas. The inhibitive action could be attributed to the competitive adsorption between H 2 O and the reactants of Hg 0 oxidation such as Hg 0 and HCl on the catalyst. 59 As H 2 O content was augmented from 8% to 12%, the downward trend of the efficiency was visibly diminished, which was perhaps because the common adsorption sites for Hg 0 , HCl and H 2 O were limited and the further increase of H 2 O concentration would not aggravate the inhibition. 34 Based on the results, the inhibition of H 2 O on the Hg 0 removal efficiency was not intense in general.

Modication mechanism of CeO 2 explored by XPS analysis
According to the above experimental results, the CeO 2 modication generated excellent results on the NO and Hg 0 removal performance of commercial SCR catalyst. The characterization results of BET and XRD could present the related reasons for the modication effects in a certain degree. In order to further explore the modication mechanism of CeO 2 on the catalyst, H 2 -TPR and XPS analyses were carried out to detect the redox behavior and valence states (or types) of the elements in the raw and modied catalysts. 3.4.1 H 2 -TPR analysis. H 2 -TPR analysis was implemented over the raw SCR and 4% CeO 2 -SCR catalysts, and the results were shown in Fig. 10. On the prole of the raw catalyst, the peaks emerged at 485 C and 568 C could be attributed to the reduction of V 5+ and surface oxygen, respectively, and the broad shoulder peak at around 720 C was corresponded to the overlap of the reduction of W 6+ and lattice oxygen. 44,60,61 By contrast, a reduction peak was observed at 461 C on the prole of 4% CeO 2 -SCR. As Ce 4+ was reported to reduce at about 470 C, this peak was considered to be the overlapped reduction peak of V 5+ and Ce 4+ . 62 It was evident that the temperature of this peak was lowered and the intensity was strengthened dramatically compared to the peak of the raw catalyst at 485 C, which indicated that the synergy of element V and Ce reinforced the reactivity of the catalyst. In addition, the reduction peak of surface oxygen of 4% CeO 2 -SCR at 563 C was much stronger than that of the raw catalyst, so it demonstrated the existence of Ce enhanced the oxygen storage capacity of the catalyst. Combining the above factors, the integral area of the reduction prole was obviously larger over 4% CeO 2 -SCR than over the raw catalyst, suggesting the improved redox behavior of the catalyst led by the CeO 2 modication. The superior redox behavior was favorable to the NO and Hg 0 removal performance, which was one of main reasons for the prominent catalytic efficiencies of the CeO 2 modied commercial SCR catalyst.
3.4.2 XPS analysis. The XPS spectra of the elements for the fresh catalysts, together with the tting results of the curves, were shown in Fig. 11. For the spectra of O 1s, the tting peaks were assigned to lattice oxygen (O latt ), chemisorbed oxygen (O ad ) and oxygen of hydroxyl and free water (O hyd ) in sequence at the binding energies from small to large, 25,63 as shown in Fig. 11(a). And the tting peaks of V 2p at the binding energies of approximately 516.4 eV and 517.6 eV could be distributed to V 4+ and V 5+ , respectively, 64,65 which was shown in Fig. 11(b). In addition, the analysis on the spent catalyst sample of 4% CeO 2 -SCR aer reacted in simulated coal-red ue gas was conducted as well. The obtained curves of Ce 3d, O 1s and V 2p were made comparisons with those of the fresh catalyst, and the results were shown in Fig. 12. On the curves of the element Ce as shown in Fig. 12(a), the tting peaks of u, u2, u3, v, v2 and v3 were attributed to Ce 4+ , while the peaks of u1 and v1 were corresponded to Ce 3+ . 38,66 And the spectra of O and V for the spent catalyst were shown respectively in Fig. 12(b) and (c). The ratios of each elemental type or valence state in the corresponding elements of the catalysts were acquired through integrating the tting peaks and calculating the peak area. The calculation results for the elements of the fresh and spent catalysts were summarized in Tables 3 and 4, respectively.
According to the testing results, the addition of CeO 2 into the catalyst improved both the surface atomic content of O and the proportion of O ad , which led to the increase of the content of O ad on the catalyst, as the data listed in Table 3. It could be judged from the results of Ce 3d of 4% CeO 2 -SCR shown in Fig. 12(a) and Table 4 that Ce 3+ and Ce 4+ coexisted in the modied catalysts. The presence of Ce 3+ with a proportion of 16.6% could create charge imbalance and unsaturated chemical bonds on the surface, which was favorable for the generation of chemisorbed oxygen, thereby raising the content of O ad and boosting the oxygen storage capacity of the catalyst. 38,61,67 O ad was the active oxygen species that could participate in the catalytic reactions. 4% CeO 2 -SCR owned the highest content of O ad among the catalysts, which was another important reason for its optimal NO and Hg 0 removal performance. As the CeO 2 loading increased from 4% to 7%, the O ad content on the catalyst declined and it was even lower than that of the raw catalyst. This result could be associated with the conversion of CeO 2 to the crystalline phase in 7% CeO 2 -SCR according to the XRD results, which made it disadvantaged for the forming of O ad from the loaded CeO 2 , and meanwhile the forming of crystalline CeO 2 might consume a number of the intrinsic O ad on the surface. Besides O ad , the intensity of the V 5+ peak and the ratio of V 5+ were also enlarged with the introduction of CeO 2 . The increase of the V 5+ proportion might be attributed to part of V 4+ being oxidized by the abundant O ad to V 5+ on the modied catalysts. V 5+ was the active species in V-based catalyst as well, which had good oxidation and was benecial to Hg 0 removal activity. So the adequate O ad was also presented in the form of V 2 O 5 . As the content of O ad on the surface of pure CeO 2 did not show an advantage, it further demonstrated the superior oxygen storage capacity was the result of the synergy of CeO 2 and V 2 O 5 in the CeO 2 -SCR catalysts.
Aer the 4% CeO 2 -SCR catalyst was reacted in SFG, the XPS spectra of O 1s and V 2p for the spent catalyst were compared with those for the fresh one. The results indicated that the intensity of both the O ad and V 5+ peaks reduced apparently aer the reaction, as shown in Fig. 12(b) and (c). The variation could be observed more intuitively by the results in Table 4 that the ratios of O ad and V 5+ in the corresponding elements decreased from 32.5% to 27% and from 60.8% to 54.7%, respectively, in the reaction process, while the ratio of Ce 3+ increased from 16.6% to 20.5%. The variation trends of the ratios of O ad and Ce 3+ on the catalyst were in accordance with those in the related literatures aer the catalysts were spent. 19,68 The decline of the ratio of O ad demonstrated it indeed participated in the catalytic reactions as the active substance. And the decrease of the proportion of V 5+ suggested the redox behavior between element V and Ce on the catalyst during the reactions, which could be expressed by eqn (13). Combining the eqn (13) with the previous eqn (5), it could be seen that it occurred the process of the redox transformation between Ce 3+ and Ce 4+ on the surface, in which the chemisorbed oxygen was generated. The formed O ad then involved in the catalytic reactions such as eqn (6), (7) and (9)- (12) so that the performance of the catalyst for synergistic Hg 0 removal in SFG was improved. Besides O ad , the ratio of O hyd increased by 7.4% in the spent catalyst. On one hand, H 2 O contained in the ue gas adsorbed on the catalyst and formed hydroxyl during the reaction, which caused the competitive adsorption with Hg 0 and led to the inhibition of H 2 O on Hg 0 removal efficiency; on the other hand, the increased proportion of O hyd might also be due to the generated H 2 O of eqn (7), thereby further demonstrating the occurrence of this reaction.
Combining the experimental results and the XPS analysis conclusions, the modication effects of CeO 2 on commercial SCR catalyst was mainly owed to the more sufficient chemisorbed oxygen which derived from the interaction between element V and Ce and the redox transformation between Ce 3+ and Ce 4+ on the catalyst surface. The abundant O ad improved the catalytic activity of the catalyst and the promotion of related ue gas components such as HCl on the Hg 0 removal efficiency. Integrating these factors, the catalytic property for synergistic Hg 0 removal was enhanced by the CeO 2 modication. The modication process was described more vividly and specically by the illustration shown in Fig. 13.

Conclusions
CeO 2 modied commercial SCR catalyst was prepared and investigated for NO and synergistic Hg 0 removal. The research results indicated that the CeO 2 loading improved a series of properties of the catalyst. Concretely, the BET surface area, the dispersity of the metal oxides on TiO 2 support and the redox behavior were increased with the introduction of CeO 2 into the catalyst, which was favorable to the catalytic activity. The catalyst with the CeO 2 content of 4% exhibited the optimal performance for simultaneous NO and Hg 0 removal. The NO conversion of 4% CeO 2 -SCR was as high as 90.5% at 300 C in SFG with excellent N 2 selectivity, while the synergistic Hg 0 removal efficiency could reach 78.2% under the same condition. Owing to the abundant chemisorbed oxygen generated from the synergy of V 2 O 5 and CeO 2 and the redox transformation between Ce 3+ and Ce 4+ , the Hg 0 removal activity, the HCl utilization and NH 3 -resistance in Hg 0 removal process and the promotion of SO 2 and NO on Hg 0 removal efficiency were improved over 4% CeO 2 -SCR compared to over the raw catalyst. On account of these factors, the CeO 2 modication made an enhancement on the synergistic Hg 0 removal performance of the commercial SCR catalyst in simulated coal-red ue gas, especially under low HCl concentration. Therefore, the CeO 2 modied commercial SCR catalyst was a potential candidate to be practically applied in coal combustion power plant for NO and synergistic mercury removal.

Conflicts of interest
There are no conicts to declare.