Introduction

Coal-fired utility boilers are major anthropogenic sources of mercury emissions (Pirrone et al. 2010). Many countries and environmental researchers have paid considerable attention to mercury (Hg) because of its high volatility, long persistence, and strong biological accumulative properties (Brown et al. 1999). In 2011, the USA announced the Mercury and Air Toxics Standard (MATS) (He et al. 2014). In January 2013, 140 nations adopted the first legally binding international treaty to set enforceable limits on emissions of mercury and exclude, phase out, or restrict some products that contain mercury (Auzmendi-Murua et al. 2014).

Many technologies have been used for mercury control in coal combustion flue gas, including activated carbon injection (ACI), catalytic oxidation, photochemical oxidation, and existing air pollution control devices (APCDs) (Srivastava et al. 2006). Catalytic oxidation of elemental mercury (Hg0) could greatly enhance mercury capture efficiency using currently available pollution control devices such as wet scrubbers (Ji et al. 2008). Compared to activated carbon injection, combining Hg0 oxidation with wet flue gas desulfurization (WFGD) is considered as a promising method for mercury removal due to its lower cost. With extensive application in coal-fired power plants, selective catalytic reduction (SCR) technology could control NOx emission as well as exhibit the co-benefit of promoting Hg0 oxidation. Many studies have evaluated the performance of the SCR catalysts on Hg0 oxidation (Cao et al. 2008; Kamata et al. 2008; Pudasainee et al. 2010; Yan et al. 2011). However, the commercial V2O5-WO3(MoO3)/TiO2 catalyst has the disadvantages of high reaction temperature (573–673 K) (Zhibo et al. 2013). Moreover, the high concentration of dust in the flue gas reduces its performance and longevity at this temperature (Zhang et al. 2012). Accordingly, in order to avoid the cost for reheating the flue gas and the deactivation by high concentration of dust, it is desirable to develop a novel SCR catalyst with high catalytic activity at low temperatures. MnOx-based catalysts (He et al. 2013; Li et al. 2012; Qu et al. 2014; Wang et al. 2014; Wiatros-Motyka et al. 2013) have been proposed as economical low temperature SCR catalysts for NOx removal and Hg0 oxidation. It has been widely studied owing to their excellent capture activity, easy manufacturing, and low cost. Ji et al. (Ji et al. 2008) synthesized a novel MnOx/TiO2 catalyst and found NO conversion and Hg0 capture could achieve 97 and 90 %, respectively. Moreover, Li et al. (2010) reported that the Hg0 oxidation efficiency over MnOx/alumina and modified MnOx/alumina catalysts was more than 90 %. Currently, CeO2 has been widely studied as an oxygen provider for heterogeneous catalytic reactions due to its oxygen storage capability by storing or releasing O via the Ce4+/Ce3+ redox couple (Shen et al. 2014a; Shen et al. 2014b; Zhang et al. 2013). Significantly, cerium oxides provide oxygen to manganese oxides at low temperatures, which promoted the oxidation activity of manganese (He et al. 2014; Larachi et al. 2002). In particular, Ce-Mn mixed oxide was found to be an excellent low-temperature SCR catalyst (Li et al. 2012; Wu et al. 2008). Li et al. (2012) reported that TiO2 supported Mn-Ce mixed oxides (Mn-Ce/Ti) was highly active for Hg0 oxidation at low temperatures (150–250 °C) under both simulated flue gas and SCR flue gas. He et al. (2014) investigated Ti-pillared-clay supported MnOx-CeO2 mixed oxide as catalysts for Hg0 capture. The results found that the catalyst exhibited high Hg0 adsorbent and catalytic oxidant in the absence of HCl. Besides, the addition of CeO2 could significantly enhance the sulfur resistance and water resistance (Wang et al. 2014). Consequently, Mn-Ce mixed oxide has been recently developed as a SCR catalyst with extraordinarily high activity for NOx removal and Hg0 oxidation.

It is worth noting that the TiO2-based SCR-DeNOx catalysts (V2O5-WO3/TiO2) usually suffer some types of deactivation during the practical application, such as poisoning, fouling, thermal degradation, and vapor compound formation (Shang et al. 2012). As a result, TiO2-based SCR catalysts are usually in operation for 3 years until the catalytic performance is finally deteriorated. Subsequently, the spent TiO2-based SCR catalyst is either disposed in landfill or recycle. The recycle of catalyst is considered as the best approach due to its cost-saving and environment-friendly benefit. Some works have been done to overcome the drawback of SCR-DeNOx catalysts, including washing by water and acid solution, thermal regeneration, and reductive regeneration (Guo et al. 2008; Khodayari and Odenbrand 2001; Shang et al. 2012). Therefore, the novel development of spent TiO2-based SCR-DeNOx catalyst as a potential catalytic support material could present a more environmentally and financially sound option for simultaneous removal of Hg0 and NO in conventional coal-fired power plants.

This work attempts to reuse the spent TiO2-based SCR catalyst. The spent TiO2-based SCR-DeNOx catalyst was used as a support to synthesize CeMn/SCR1 catalysts. The Hg0 oxidation activities were tested at 100–300 °C in simulated coal-fired flue gas without HCl. The effects of Ce/Mn mass ratios, flue gas components including NH3, SO2, H2O, and O2 were evaluated. Besides, in order to further investigate the improvement of physicochemical properties by the modified Ce-Mn mixed oxide, Brunauer-Emmett-Teller (BET), scanning electron microscopy (SEM), X-ray diffractogram (XRD), and X-ray photoelectron spectroscopy (XPS) analyses were carried out and analyzed in detail. The catalytic mechanism for simultaneous removal of Hg0 and NO was investigated in the end.

Experimental

The preparation of catalysts

The spent TiO2-based SCR-DeNOx catalysts were obtained from power plant after exposure to exhaust gases in the SCR system for 3 years. Their DeNOx activity (about 30.00 %) was much less than fresh ones. The spent TiO2-based SCR catalysts were washed in deionized water to remove poisons such as ash, alkali metals, and salt molecules from the samples. Once washing was completed, the sample was dried in an oven at 105 °C for 12 h. Finally, the samples were crushed and sieved to 100–120 mesh. The samples are denoted as SCR1 catalysts and stored in desiccator.

The Ce/SCR1, Mn/SCR1, and CeMn/SCR1 samples were prepared using cerium nitrate and/or manganese nitrate aqueous solution and SCR1 catalysts by an impregnation method enhanced by ultrasound. Specifically as follows: first, a certain amount of cerium nitrate and/or manganese nitrate was dissolved in deionized water to form the solution. Then, the SCR1 catalyst was added to the solution with stirring in a proportion. Third, the mixture was exposed to an ultrasonic bath for 1 h. Excess water was evaporated while stirring and ultrasonic bath. The obtained solid was dried at 105 °C for 12 h and calcined at 500 °C for 3 h. Finally, the obtained catalysts were ground and sieved to 100–120 mesh. The catalysts were denoted as Mn/SCR1, Ce/SCR1, and CexMn/SCR1, where x represented the Ce/Mn mass ratio (0.25, 0.67, 1.00, 1.50, and 4.00). The mass loading of (CeO2 + MnOx)/SCR1 were 1 %. The mass loading of Ce/SCR1 or Mn/SCR1 were 1 % as well.

Catalyst characterization

The analysis of surface area, average pore size, and average pore volume was carried by Micromeritics Tristar ІІ 3020 analyzer (Micromeritics Instrument Crop, USA). All samples were degassed in vacuum at 180 °C for 5 h. The specific surface area was determined using the standard Brunauer-Emmett-Teller (BET) equation. The average pore diameter and average pore volume of the samples were obtained from the desorption branches of N2 adsorption isotherm and calculated by the Barrett-Joyner-Halenda (BJH) formula.

To further analyze the morphology and surface structure of the samples, SEM analysis was performed using a Hitachi S-4800 (Hitachi Limited, Japan). The separated areas for each sample were magnified to ×100,000.

The samples were characterized by XRD using a Rigaku rotaflex D/Max-C powder diffractometer (Rigaku, Japan) with Cu-Ka radiation (l = 1.5406 A°) in the 2θ range from 10° to 80° with 0.2o/min.

XPS experiments were performed on a K-Alpha 1063 X-ray photoelectron spectrometer (Thermo Fisher Scientific, UK). XPS spectra were obtained using an Al Ka X-ray source operated at 12 kV and 6 mA. The observed spectra were calibrated with the carbon 1s electron binding energy (BE) value of 284.6 eV.

Catalytic activity measurement

The experiments were performed in a fixed-bed quartz reactor (i.d. 20 mm), which was located inside a temperature-controlled tubular furnace, as shown in Fig. 1. An Hg0 permeation tube (VICI Metronics, USA) was used to generate a constant quantity of Hg0 vapor (70.00 μg/m3), which was supplied into the gas steam. Other simulated flue gas (SFG) components including 300 ppm SO2, 500 ppm NO, 500 ppm NH3 (when used), 12 % CO2, 8 % H2O vapor (when used), 5 % O2, and balanced gas N2 were supplied by gas cylinders and were precisely controlled by mass flow controllers (MFC). A total flow rate of 1 L/min were used during the experiments, which corresponded to a gas hourly space velocity (GHSV) of 100,000/h (the catalysts mass was 0.5 g). All the gas components were mixed and preheated to the desired temperature, and then passed through the fixed-bed quartz reactor. The feed water was exactly controlled by peristaltic pump and injected into the Teflon tube that wrapped with a temperature-controlled heating tape. Afterward, the water vapor was generated. One hundred milliliters per minute pure N2 took along the water vapor to mix with the flue gas. Besides, in order to avoid adsorption of Hg0 on the inner surface and condensation of water vapor, all Teflon tubes that Hg0 and water vapor passed through were heated up to 120 °C. A RA-915M mercury analyzer (LUMEX Ltd, Russia) was employed to measure the concentration of Hg0 while NO concentration was monitored by a flue gas analyzer (MGA5, Germany). It was verified that interferences on Hg0 measurement by flue gas components such as SO2, NH3, and water vapor were negligible. Even so, before proceeding to the mercury analyzer, acid gases, NH3, and water vapor were removed by the conditioning unit and the condenser. The conditioning unit converged to a 10 % NaOH solution in which gases such as NH3 and SO2 were captured. The H2O vapor condensed into liquid water through the condenser.

Fig. 1
figure 1

Schematic diagram of the experimental setup

Full size image

Six sets of experiments were conducted, and the experimental conditions are summarized in Table 1. In set I to set II, experiments aimed at finding the optimal catalyst and operating temperature. The Hg0 oxidation activities over SCR1, Mn/SCR1, Ce/SCR1, and CexMn/SCR1 samples were evaluated under SFG (500 ppm NO, 300 ppm SO2, 5 % O2, 12 % CO2, and 70.00 μg/m3 Hg0) for 2 h. In set III to set V, the effects of individual gas components on Hg0 oxidation and the reaction pathways were explored. The experiments were conducted on the optimal sample in the presence of individual flue gases at optimal operating temperature. In set VI, the optimal sample was tested in SCR atmosphere (500 ppm NO, 500 ppm NH3, 300 ppm SO2, 5 % O2, and 70.00 μg/m3 Hg0) at each selected reaction temperature from 100 to 300 °C for investigating the simultaneous removal of NOx and Hg0.

Table 1 Experimental conditions
Full size table

At the beginning of each test for catalytic activity, the gas stream bypassed the reactor and the inlet gas was sampled to ensure a stable NO (NOin) and Hg0 (Hg0 in) feed concentration. Then, the gas flow was switched to pass through the catalysts, and the analyzer measured the gas compositions in the outlet (noted as NOout and Hg0 out). Before each test, the catalysts were first saturated with the established Hg0 in under N2 atmosphere at room temperature to avoid possible bias because of Hg0 physical adsorption (Anchao et al. 2014; Li et al. 2012; Li et al. 2011). The result indicated that Hg0 physical adsorption capacity of samples were negligible at room temperature (less than 1.0 μg/g at room temperature). Therefore, the loss of Hg0 over the catalysts should be due to the oxidation of Hg0. Hg0 oxidation efficiency (ηHg) over the catalysts was quantified by Eq. (1):

$$ {\eta}_{H\mathrm{g}}\left(\%\right)=\frac{{\mathrm{Hg}}_{\mathrm{in}}^0\hbox{-} {\mathrm{Hg}}_{\mathrm{out}}^0}{{\mathrm{Hg}}_{\mathrm{in}}^0}\times 100\% $$
(1)

Meanwhile, the NO conversion efficiency (ηNO) can be calculated using Eq. (2):

$$ {\eta}_{NO}\left(\%\right)=\frac{{\mathrm{NO}}_{\mathrm{in}}\hbox{-} {\mathrm{NO}}_{\mathrm{out}}}{{\mathrm{NO}}_{\mathrm{in}}}\times 100\% $$
(2)

Results and discussion

Characterization of the catalyst

BET

The BET specific surface area, pore volumes, and average pore sizes of the samples are listed in Table 2. Compared to the SCR1 catalyst, the addition of Mn reduced the surface areas of catalysts. The reason might be that MnOx occupied a part of free pores of supports during the impregnation process. Note that the surface areas of the sample were increased after the addition of Ce. This result suggested that Ce could improve the specific surface areas of catalysts. However, the BET surface areas of all samples had no obvious change. Furthermore, pore volumes of Ce1.00Mn/SCR1 catalyst were lower than that of Mn/SCR1 and Ce/SCR1. It has been reported that a prominent synergy effect could be exhibited between CeO2 and MnOx (Singh et al. 2013). Therefore, it could be inferred that the loss of pore volumes might be caused by deposited Ce-Mn mixed-oxide, which penetrated into the pores of carriers.

Table 2 The surface area, pore volume, and pore diameter of the samples
Full size table

SEM

The SEM images of the selective samples are shown in Fig. 2. The SEM images showed that crystal morphologies of the SCR1 catalyst were not significant influenced by the addition of metal oxide. From Fig. 2a, it could be seen that the surface of SCR1 catalyst was smooth and the size was uniform. For Ce/SCR1 or Mn/SCR1 catalyst, a few depositions of CeO2 or MnOx formed, while on the Ce1.00Mn/SCR1 catalyst surface (Fig. 2d), more depositions were observed. Since the BET surface areas had no significant change as demonstrated in Table 2, it can be concluded that the dispersion of CeO2 or MnOx on the support surface was excellent. However, a synergy effect between CeO2 and MnOx could result in particles interconnection, which promoted the formation of agglomerates.

Fig. 2
figure 2

Nanoscale of SEM analysis of a SCR1, b Ce/SCR1, c Mn/SCR1, d Ce1.00Mn/SCR1 (100,000 multiplier)

Full size image

XRD

The XRD patterns of spent SCR1, Ce/SCR1, Mn/SCR1, and Ce1.00Mn/SCR1 catalysts are shown in Fig. 3. It was found that the major reflections of XRD patterns belonged to anatase phase TiO2. The addition of CeO2 or/and MnOx did not result in significant change in the positions or the shapes of typical TiO2 diffractions. Besides, for all the samples, no obvious traces of CeO2 or/and MnOx were found in any of the catalysts during XRD analysis, which indicated that CeO2 or/and MnOx species were well dispersed throughout the support structure or their crystalline structures were too small to detect.

Fig. 3
figure 3

XRD patterns of SCR1, Ce/SCR1, Mn/SCR1, and Ce1.00Mn/SCR1

Full size image

XPS

To gain more insight into the reaction pathways, an XPS analysis was performed to determine the chemical state and the relative proportion of the elements on the surface of the fresh and spent Ce1.00Mn/SCR1 catalyst. The spent Ce1.00Mn/SCR1 catalysts were carried out on the fix bed reaction system under SCR atmosphere at 150 °C. The XPS results are shown in Fig. 4. The O1s peaks could be fitted into two peaks referred to the lattice oxygen (Oβ) at 529.3–530.0 eV and the chemisorbed oxygen (Oα) at 531.3–531.9 eV (Wu et al. 2008). As shown in Fig. 4a, the Oβ ratio for the peak at 530.0 eV (calculated by Oβ/( Oα + Oβ)) increased from 61.45 to 62.61 % compared to that of the fresh catalyst, whereas the ratio of Oα (calculated by Oα/(Oα + Oβ)) decreased from 38.54 to 37.38 %. These observations suggested that both chemisorbed oxygen and lattice oxygen participated in Hg0 oxidation and NO conversion. The increase of the Oβ ratio can be explained by two factors. First, O2 in flue gas could regenerate the lattice oxygen which consumed in the reaction. Second, the consumption of chemisorbed oxygen increased the percentage of lattice oxygen.

Fig. 4
figure 4

XPS spectra of Ce1.00Mn/ SCR1 over the spectral regions of a O 1s, b Ce 3d, and c Mn 2p (reaction condition: 70.00 μg/m3 Hg0, 500 ppm NO, NH3/NO: 1, 5 % O2, 300 ppm SO2, 500 mg of sample, 1000 mL/min total flow rate, and GHSV 100,000/h)

Full size image

The XPS spectra for Ce 3d are shown in Fig. 4b. XPS peaks denoted as u/v, u2/v2, and u3/v3 can be assigned to Ce4+ while u1/v1 correspond to Ce3+ (Wang et al. 2014). Both Ce3+ and Ce4+ existed in the fresh Ce1.00Mn/SCR1 catalyst. Based on the peak areas, the peaks corresponding to Ce4+ appeared to be dominant. After the reaction, the ratio of Ce4+/Ce3+ decreased from 2.70 to 1.54 compared to the fresh catalyst, implying a reduction of Ce4+ during the reaction. Furthermore, the presence of Ce3+ may result in a charge imbalance, which would lead to oxygen vacancies and unsaturated chemical bonds (Liu et al. 2012). This situation could generate additional chemisorbed oxygen or weakly adsorbed oxygen species on the surface of the catalyst, which is beneficial for Hg0 oxidation and the SCR reaction.

The XPS results of Mn 2p for the fresh and spent Ce1.00Mn/SCR1 catalyst are shown in Fig. 4c. For the fresh Ce1.00Mn/SCR1 catalyst, the peaks observed at 643.1 and 641.5 eV correspond to Mn4+ and Mn3+, respectively (Xu et al. 2012). The co-existence state of MnO2–Mn2O3 can promote the catalytic oxidation activity of catalyst (Xu et al. 2012). For the catalyst used at 150 °C, the ratio of Mn4+/Mn3+ decreased from 2.38 to 0.72. The result confirmed that reactions took place between the adsorbed active species and lattice oxygen (Wang et al. 2014). The lattice oxygen came from the valence state changes of MnOx. O2 in the flue gas would supply the metal oxide with oxygen to ensure that reaction can be continued. Therefore, it can be concluded that the lattice and chemisorbed oxygen on the catalyst surface are responsible for their superior performance. The reduction of Mn4+ to Mn3+ on the surface of the catalyst contributed to the Hg0 oxidation.

Performance of catalysts

Catalytic activity tests

The activity curves of the different catalysts under simulated flue gas are shown in Fig. 5. Apparently, with the addition of CeO2 or/and MnOx into SCR1, the catalytic activity was obviously enhanced, especially at low temperature. Notably, Eoxi over Ce1.00Mn/SCR1 catalyst was higher than that over Mn/SCR1 and Ce/SCR1 catalysts. This result demonstrated that there were intense interactions between manganese oxides and cerium oxides (Li et al. 2012). The interactions resulted in more surface oxygen, which is responsible for Hg0 oxidation.

Fig. 5
figure 5

Hg0 oxidation efficiency over different catalysts under simulated flue gas as a function of temperature (reaction condition: 70.00 μg/m3 Hg0, 500 ppm NO, 5 % O2, 300 ppm SO2, 12 % CO2, 500 mg of sample, 100-300 °C, 1000 mL/min total flow rate, and GHSV 100,000/h)

Full size image

The Hg0 oxidation of the Ce1.00Mn/SCR1 catalysts with different Ce/Mn mass ratios is shown in Fig. 6. According to the results of Ce1.00Mn/SCR1 catalysts, it can be seen that the ηHg first increased with the increasing of Ce/Mn mass ratios, then the ηHg decreased when the Ce/Mn mass ratios was 1.50. Remarkably, Ce1.00Mn/SCR1 catalyst showed the best activity, and its Hg0 oxidation efficiency was 92.80 % at 150 °C. These results indicated that Ce/Mn mass ratios influenced the activity of CexMn/SCR1 catalysts, which is related to the synergy between CeO2 and MnOx.

Fig. 6
figure 6

Effects of different of Ce/Mn mass ratio on Ce1.00Mn/SCR1 catalysts on Hg0 oxidation efficiency (reaction condition: 70.00 μg/m3 Hg0, 500 ppm NO, 5 % O2, 300 ppm SO2, 12 % CO2, 500 mg of sample, 100–200 °C, 1000 mL/min total flow rate, and GHSV 100,000/h)

Full size image

Effect of NH3 on Hg0 oxidation

Since NH3 is the reactant gas for selective catalytic reduction of NOx, it is necessary to study the effect of NH3 on Hg0 oxidation. Figure 7 shows the influence of NH3 on mercury oxidation over Ce1.00Mn/SCR1 catalysts. In this work, a significant inhibitive effect of NH3 on Hg0 oxidation over the catalysts was observed both in pure N2 atmosphere and in the presence of O2. NH3 (500 ppm) was added to pure N2 atmosphere and gas flow containing 5 % O2 balanced in N2, respectively. It was observed that the Hg0 oxidation efficiency decreased. In the presence of O2, Hg0 oxidation dropped to only 19.73 % over Ce1.00Mn/SCR1 catalyst, which, however, was still higher than that without O2. Some researchers believe that NH3 adsorption on catalyst most probably takes place according to the following steps (Qi et al. 2004; Zhou et al. 2013):

Fig. 7
figure 7

Effect of NH3 on Hg0 oxidation over different catalysts (reaction condition: 70.00 μg/m3 Hg0, 500 ppm NH3, NH3/NO: 0–1, 5 % O2, the optimal reaction temperature, 500 mg of sample, 1000 mL/min total flow rate, and GHSV 100,000/h)

Full size image
$$ {\mathrm{NH}}_3\left(\mathrm{g}\right)\to {\mathrm{NH}}_3\left(\mathrm{ad}\right) $$
(3)
$$ {\mathrm{NH}}_3\left(\mathrm{ad}\right)+{\mathrm{O}}^{*}\left(\mathrm{ad}\right)\to {\mathrm{NH}}_2\left(\mathrm{ad}\right)+\mathrm{O}\mathrm{H}\left(\mathrm{ad}\right) $$
(4)

where O* are active surface oxygen of the catalyst. In this process, NH3 consumed the surface oxygen (including lattice oxygen and chemisorbed oxygen) that is responsible for Hg0 oxidation and NO conversation. Gas-phase O2 regenerated the lattice oxygen, and replenished the consumed chemisorbed oxygen. Thus, the presence of gas-phase O2 offsets part of the inhibitive effect of NH3. Interestingly, the coexistence of NO with NH3 can also mitigate such inhibition. ηHg was 34.79 % in N2 + 500 ppm NO + 500 ppm NH3 + 5 %O2, which was higher than that in N2 + 500 ppm NH3. In addition, the ηHg over Ce1.00Mn/SCR catalyst was also higher than that of SCR1 catalyst. It indicated the participation of CeO2 and MnOx had certain effects on reducing the inhibitive effect of NH3.

Effect of SO2 and H2O on Hg0 oxidation

It is very important to evaluate the effect of SO2 and H2O on Hg0 oxidation because SO2 and H2O unavoidably exist in flue gas and have significant impacts on the activities of catalysts. Figure 8 presents the effects of SO2 and H2O on Hg0 oxidation over Ce1.00Mn/SCR1 catalysts at 150 °C. It can be seen that Hg0 oxidation was not promoted by SO2 itself. Otherwise, Hg0 oxidation efficiency should not rely on the existence of O2 and should rise with the increase of SO2 concentration (Li et al. 2013). As shown in Fig. 8a, introducing 200 ppm SO2 into gas flow containing 5 % O2 increased Hg0 oxidation efficiency from 60.15 to 63.10 %, which indicated that low SO2 concentration promoted Hg0 oxidation with the aid of O2. However, Hg0 oxidation efficiency gradually decreased as SO2 concentration increased from 200 to 1000 ppm. As can be seen in Fig. 8a (inset), ηHg maintained at almost the same level throughout the entire experimental period with the aid of O2. However, without the aid of O2, the Hg0 oxidation efficiency was influenced with a drop of 34.56 %, and a continuous declining trend was observed as time went by. It demonstrated that part of adsorbed SO2 reacted with catalyst to form some new species that were responsible for the enhanced Hg0 oxidation. For metal-oxide-based SCR catalysts, the new species was generally believed to be SO3 (Casapu et al. 2009; Dunn et al. 1998; Fan et al. 2010; Svachula et al. 1993). Therefore, it was inferred that SO2 would be oxidized by the lattice oxygen, reducing the active oxidant sites. The reduced active sites could not be reactivated without O2, which resulted in the decrease of Hg0 oxidation efficiency as time went by. On the other hand, with the aid of O2, the reduced oxidant sites were reoxidized by O2. Both the reactivated sites and SO3 promoted Hg0 oxidation efficiency. A promotional effect of SO2 on Hg0 oxidation can be proposed as follows (Li et al. 2008a):

Fig. 8
figure 8

Effect of SO2 and H2O on Hg0 oxidation over Ce1.00Mn/SCR1 catalysts. a Effect of SO2 concentration on Hg0 oxidation (reaction condition: 70.00 μg/m3 Hg0, 0–1000 ppm SO2, 5 % O2, 500 mg of sample, 150 °C, 1000 mL/min total flow rate, and GHSV 100,000/h). b Effect of SO2 and H2O on Hg0 oxidation (reaction condition: 70.00 μg/m3 Hg0, 300 ppm SO2, 8 % H2O, 5 % O2, 500 mg of sample, 150 °C, 1000 mL/min total flow rate, and GHSV 100,000/h)

Full size image
$$ {\mathrm{SO}}_2+\frac{1}{2}{\mathrm{O}}_2\to {\mathrm{SO}}_3 $$
(5)
$$ {\mathrm{Hg}}^0+\frac{1}{2}{\mathrm{O}}_2+{\mathrm{SO}}_3\to {\mathrm{Hg}\mathrm{SO}}_4 $$
(6)

Figure 8b illustrates the activity performance of Ce1.00Mn/SCR1 catalysts in the system with SO2 and water vapor simultaneously. As is shown, when adding 8 % H2O into the stream, the Hg0 oxidation efficiency dropped sharply. Obviously, H2O had a significant inhibitive effect on Hg0 oxidation (Li et al. 2008b). The inhibitive effect of H2O possibly came from its competitive occupancy of the available active sites and displacement of adsorbed Hg0 (Tan et al. 2012). When both H2O and SO2 were added, the inhibition was less severe than that when H2O were added individually. This indicates that H2SO4 can form when SO2, O2, and H2O co-exist, which will likely affect the behavior of Hg0 on the surface of catalyst. After the H2O and SO2 were cut off, it was found that the catalytic activity gradually recovered but not returned to the original level yet. It implies that the addition of SO2 can reduce the H2O poison. However, the catalysts were deactivated by H2SO4 somehow.

Effect of O2

To obtain higher catalytic oxidation, the presence of O2 is normally necessary, especially for metal oxide catalysts (Anchao et al. 2014). Figure 9 shows the effect of O2 on catalytic activity of Ce1.00Mn/SCR1 catalysts. The results showed that the NO conversion and Hg0 oxidation efficiency in N2 + 2 % O2 was greater than that in pure N2 atmosphere. This observation indicated that O2 is beneficial to the catalytic activity of catalyst. Meanwhile, Hg0 oxidation might comprise some chemical reactions with the presence of O2 (Fan et al. 2012; Tao et al. 2012; Yang et al. 2011a; Yang et al. 2011b). Combined with the XPS results, it was possible that gas-phase O2 achieved the supplement of the lost lattice oxygen, which facilitated Hg0 oxidation and NO conversion.

Fig. 9
figure 9

Effect of O2 concentration on simultaneous removal of Hg0 and NO over Ce1.00Mn/SCR1 catalysts at 150 °C (reaction condition: 70.00 μg/m3 Hg0, 100 ppm NO, NH3/NO: 1, 0–10 % O2, 500 mg of sample, 150 °C, 1000 mL/min total flow rate, and GHSV 100,000/h)

Full size image

Simultaneous removal of Hg0 and NO

The simultaneous removal of Hg0 and NO by Ce1.00Mn/SCR1 catalysts was also investigated and the results are illustrated in Fig. 10. As shown in the figure, 82.20 % NO conversion was obtained at 200 °C with a GHSV of 100,000/h. An obvious decrease of the Hg0 oxidation efficiency was found in the experiment compared with the NH3-free condition. Although simultaneously removal of Hg0 and NO was unsatisfactory, it is still encouraging because lower space velocity would result in higher Hg0 oxidation efficiency and NO conversion. Combined with the result of upper experiment, the problem that how to design and choose the best optimized experimental conditions for simultaneous removal of Hg0 and NO still needs further study.

Fig. 10
figure 10

Hg0 oxidation and NO conversion with respect to reaction temperatures (reaction condition: 70.00 μg/m3 Hg0, 500 ppm NO, NH3/NO: 1, 300 ppm SO2, 5 % O2, 500 mg of sample, 100–300 °C, 1000 mL/min total flow rate, and GHSV 100,000/h)

Full size image

The cooperation of Mn and Ce in the SCR catalysts

It was worthy noted that the reaction on the surface of Ce1.00Mn/SCR1 catalysts included both NO conversion and Hg0 oxidation. NO was reduced by NH3 through the catalysis of NH3-SCR while Hg0 was oxidized to Hg2+. The reaction observed in this work like previous literature (He et al. 2014; Singh et al. 2013; Xu et al. 2012) which is thought to be aided by the synergistic mechanism between the manganese and cerium oxides. They explained the synergistic mechanism as follows:

$$ 2{\mathrm{Mn}\mathrm{O}}_2\to {\mathrm{Mn}}_2{\mathrm{O}}_3+{\mathrm{O}}_{\beta } $$
(7)
$$ {\mathrm{Mn}}_2{\mathrm{O}}_3+2{\mathrm{Ce}\mathrm{O}}_2\to 2{\mathrm{Mn}\mathrm{O}}_2+{\mathrm{Ce}}_2{\mathrm{O}}_3 $$
(8)
$$ {\mathrm{Ce}}_2{\mathrm{O}}_3+\frac{1}{2}{\mathrm{O}}_2\to 2{\mathrm{Ce}\mathrm{O}}_2 $$
(9)

The co-existence of MnO2-Mn2O3 can improve the NO conversion and Hg0 oxidation. The lattice oxygen generated from manganese oxides are beneficial in promoting the oxygen cycle. Moreover, the additional oxygen generated is thought to be aided by the oxygen transfer between Mn2O3 and CeO2 (Eq. 8). Ce provided oxygen to Mn, in which why the valence and oxidation ability of Mn increased. Finally, the consumed lattice oxygen would be replenished by oxygen from the flue gas (Eq. 9).

Combining with our results, the lattice oxygen participated in the NO conversion and Hg0 oxidation. On one hand, the lattice oxygen generated via Eq. 10 is invariably responsible for the oxidation of NO to the intermediate NO2-containing species on the catalyst surface. They can further react with NH3 to form ammonium nitrite (NH2NO), which subsequently decomposes into N2 and H2O. The reaction pathway can be described by Eqs. 10–12 according to the previous study (Jin et al. 2010). On the other hand, Hg0 (ad) would react with lattice oxygen that replenished by oxygen from the flue gas to form HgO. The pathway could be explained by Eqs. 13 and 14 according to the previous research (Ji et al. 2008; Kong et al. 2011; Zhao et al. 2014).

$$ \mathrm{NO}+{\mathrm{O}}_{\beta}\to \cdots \cdots -\mathrm{O}-\mathrm{N}-\mathrm{O} $$
(10)
$$ \cdots \cdots -\mathrm{O}-\mathrm{N}-\mathrm{O}+{\mathrm{NH}}_3+{\mathrm{H}}^{+}\to {\mathrm{NH}}_4{\mathrm{NO}}_2 $$
(11)
$$ {\mathrm{N}\mathrm{H}}_4{\mathrm{N}\mathrm{O}}_2\to {\mathrm{N}\mathrm{H}}_2\mathrm{NO}\to {\mathrm{N}}_2+{\mathrm{H}}_2\mathrm{O} $$
(12)
$$ {{\mathrm{Hg}}^0}_{\left(\mathrm{g}\right)}\to\ {{\mathrm{Hg}}^0}_{\left(\mathrm{ad}\right)} $$
(13)
$$ {{\mathrm{Hg}}^0}_{\left(\mathrm{ad}\right)}+{\mathrm{O}}_{\beta}\to \mathrm{H}\mathrm{g}\mathrm{O} $$
(14)

where ⋯ ⋯ represent the catalyst surface.

Conclusions

The spent TiO2-based SCR-DeNOx catalyst was used as a support for Hg0 oxidation, which presents a more environmentally and financially sound option for simultaneous removal of Hg0 and NO in industrial application. Ce-Mn mixed oxide modified catalyst was highly active for Hg0 oxidation reaction at low temperature. The Ce1.00Mn/SCR1 catalyst exhibited high Hg0 oxidation efficiency (92.80 %) at 150 °C. The flue gas components were found to have significant effects on Hg oxidation efficiency. NH3 consumed the oxygen on the catalyst surface, hence inhibiting Hg0 oxidation. Promotional effects of SO2 on the Hg0 oxidation were observed, while the presence of H2O inhibited the Hg0 oxidation. Results also indicated that the combination of CeO2 and MnOx resulted in significant synergy for Hg0 oxidation and NO conversion.