Nitrogen oxide (NOx), one of the main precursor substances of photochemical smog, the greenhouse effect, acid rain, ozone depletion, and PM2.5, is an atmospheric pollutant in coal-fired flue gas that causes great harm to the ecosystem and human life [1]. According to a recent survey, NOx emissions in China increase annually. Thus, in 2010, the emission load was about twice that recorded in 2000 and is expected to reach 19.7 Mt by 2020 [2]. Additionally, coal-fired power plants are considered to be the largest source of NOx emissions, discharging ~46% of the total NOx emissions [1]. Therefore, reduction of the NOx generated by coal-fired power plants has become imperative in the environmental field.
The current NOx control technologies can be classified into three categories: combustion, pre-combustion, and post-combustion NOx control [3]. Pre-combustion control refers to the selection of low-nitrogen fuel to inhibit the formation of NOx from the source. Owing to fuel limitations, this method is rarely used in industry. Combustion control, also known as low-nitrogen oxide combustion technology, is mainly employed to reduce NOx emissions by changing the combustion conditions and improving the production technique. This method requires relatively low cost. However, its NOx removal rate can only reach 15%-30%, making it difficult to meet the current requirements for environmental protection [4]. Post combustion control is the use of flue gas denitrification technology. It refers to the installation of denitrification devices in the flue tail by which the NOx in the flue gas is converted into harmless nitrogen and other substances through physical and chemical reactions. Owing to its good NOx removal performance and simple operation, this technique has been extensively utilized in coal-fired power plants.
Flue gas deNOx technology can be divided into selective catalytic reduction (SCR), non-SCR (NSCR), selective non-catalytic reduction (SNCR), SNCR/SCR mixed technology, oxidation, solid adsorption, absorption, catalytic decomposition, electron beam irradiation, etc. [5, 6]. Among them, SCR has become the most widely used NOx emission control method for flue gas treatment in coal-fired power plants, globally. This can be attributed to its high NOx conversion efficiency and relatively mature process [7]. Its reaction principle is to reduce NOx (including NO and minor NO2) to N2 selectively by ammonia, urea, H2, HC, or other reductants present on a particular catalyst [1, 7-9]. This method has the advantages of no by-product formation, simple equipment structure, high deNOx efficiency, reliable operation, convenient maintenance, and low one-time investment. Ammonia is the most commonly used reductant in the SCR unit of a coal-fired power plant. The main NH3-SCR reactions are as follows [10] :
As the core of the SCR technology, catalysts are used to reduce the activation energy and reaction temperature of NOx decomposition, increase the N2 selectivity in NOx reduction products, avoid the occurrence of side reactions, and thereby improve the reaction efficiency. The selection of catalysts is important. Specifically, the qualified SCR catalysts should possess the following characteristics:
(1) High deNOx activity;
(2) Strong anti-poisoning ability;
(3) High mechanical strength;
(4) Suitable operating temperature range.
Many catalysts have been proven to be active for SCR reactions. The main active components are noble and transition metal oxides, while TiO2, Al2O3, SiO2, zeolite, and carbon are often used as carriers. These catalysts possess different deNOx properties and advantages and disadvantages, which are summarized in Table 1.
Among the numerous denitrification catalysts, V2O5-WO3 (MoO3)/TiO2 catalysts have been the most widely used SCR catalysts in coal-fired power plants for many years [11-13]. Their NOx conversion rates are > 90%, while N2 selectivity among the production processes is high when the temperature is > 300 ℃. TiO2 is employed as the carrier for these catalysts. It provides a large specific surface area and porosity, which are beneficial to the catalytic reaction and adsorption of the reactants. As the main active substance in the catalysts, V2O5 exhibits strong catalytic activity and stability. It has a remarkable catalytic effect on NO reduction and performs even better when supported on TiO2. However, the vanadium content in the catalyst is usually low (0.8%-1.2%) because SO2 can be oxidized to undesirable SO3 by V2O5 [44]. Thus, WO3 or MoO3 is used as adjuvant. By interacting with anatase TiO2, WO3 shows high catalytic activity [45], increases the Lewis sites on the catalysts [46], inhibits TiO2 sintering [47], improves SO2 resistance [48], and extends the reaction temperature window [49].
The specific NOx removal mechanism over vanadium-titanium-based SCR catalysts has been studied in recent years [50-55]. It is generally accepted that adsorption of the reducing agent (NH3) on the catalyst surface is strong, while NO adsorption is negligible. Thus, the deNOx process is considered as the reaction of gaseous NO and adsorbed NH3, which follows the Eley-Rideal mechanism. According to previous studies, both Brönsted and Lewis acid sites have been found on the surface of vanadium-titanium-based catalysts [56]. Most researchers consider that Brönsted acid sites play an important role in the deNOx reaction. They believe that the NH3 adsorbed on the Brönsted acid sites would form NH4+ ions that are then oxidized into -NH3+ by V5+=O, while H-O-V4+ is generated. Subsequently, -NH3+ reacts with gaseous NO to produce -NH3+NO and decomposes to N2 and H2O. Moreover, H-O-V4+ is reoxidized by O2 in the flue gas and thus, returns to its initial state (Fig. 1) [57]. However, a few people insist that NH3 is first adsorbed on the Lewis acid sites. Thus, NH2 is formed after activation and dehydrogenation of NH3 and reacts with gaseous NO to generate nitrosamines (NH2NO) as the intermediate [51]. Considering these two points of view, the SCR reaction would occur following the oxidative dehydrogenation of NH3, regardless of the adsorption sites of ammonia [58].
The vanadium-titanium-based SCR catalyst has been proven to oxidize elemental mercury (Hg0) into oxidized mercury (Hg2+) in flue gas [59-62]. However, its Hg0 oxidation efficiency is unsatisfactory under most conditions. Only in the presence of sufficient HCl and O2 can the oxidation rate reach 80%-90% [63]. The reaction is as follows:
Two possible mechanisms have been proposed for this reaction, namely, the Eley-Rideal and Langmuir-Hinshelwood mechanisms. The former considers that gaseous Hg0 reacts with HCl, which adsorbs on the vanadium-titanium-based catalyst surface to form HgCl2 [64-68]. On the other hand, the latter considers that the reaction takes place between the adsorbed Hg0 and HCl [69, 70]. The vanadium oxides in the catalyst also play a major role in the reaction process [71], while tungsten and molybdenum can enhance the mercury oxidation performance [71, 72]. Commercial SCR catalysts can oxidize Hg0 in the flue gas as HCl is added [66, 73-75] and the generated Hg2+ would subsequently be removed by the desulfurization system. To date, this is the most economical and feasible method available to control the mercury discharged from coal-fired power plants. However, inadequate HCl in the flue gas would significantly inhibit Hg0 oxidation [69]. Therefore, to achieve the synergistic removal of mercury when using low chlorine coal as a fuel, the stability and mercury oxidation effect of the SCR catalyst should to be further improved.
The operating conditions of the SCR system in coal-fired power plants are very complex. Many factors may affect the deNOx efficiency of the SCR catalysts, mainly in the following areas.
An increase in flue gas velocity would accelerate the abrasion of the catalyst, as well as decrease the contact time of the catalyst and flue gas. This limits diffusion, adsorption, and the full reaction of the reactants, thereby reducing the NOx removal rate. On the other hand, excessive contact time can lead to the oxidation of ammonia, which would also reduce the NOx removal rate. The actual flue gas velocity of the SCR system in a coal-fired power plant is usually set to 3000-5000 h-1.
Reactions (1)-(4) suggest that the NOx removal efficiency improves as the NH3/NOx ratio increases. However, the addition of an excessive amount of NH3 does not change the NOx removal rate significantly. Instead, more NH3 is directly discharged into the atmosphere, causing secondary pollution. Thus, the NH3/NOx ratio in the SCR system is usually maintained in the range of 0.8-1.2. Furthermore, the flue gas must be thoroughly mixed with the reductant before entering the SCR system to ensure that NOx is adequately reduced by NH3. Limited reduction would also be detrimental to NOx removal from the coal-fired power plant.
Based on previous studies, the operating temperature window of industrial vanadium-titanium-based catalysts is relatively narrow (300-400 ℃) [29, 39]. In fact, increasing the temperature to its optimum value requires high energy consumption that cannot be reached in some power plants; thus, the actual deNOx efficiency is rather low. When the operating temperature is < 320 ℃, the activity and selectivity of the catalysts decrease and SO2 poisoning becomes more significant. In addition, high temperatures (> 400 ℃) can lead to TiO2 sintering [76] and cause some side reactions, such as the production of NO and N2O from NH3 oxidation.
Denitration efficiency decreases gradually with the aging of the catalysts. Thus, high-temperature sintering, catalyst pore plugging by particles, and the existence of SO2, water vapor, alkali metals, alkaline earth metals, lead, and arsenic [29, 77-79] cause SCR catalyst deactivation during the SCR process. SO2 inhibition of the catalytic activity is the most prevalent. In the NH3-SCR process, the following SO2 poisoning mechanism is supposed: SO2 is oxidized on the catalyst surface. The generated SO3 reacts with the reductant to produce NH4HSO4 and (NH4)2SO4. These two substances deposit in the catalysts pores, covering some active sites and leading to a sharp decline in the specific surface area of the catalysts [80]. Under normal SCR operating conditions, the SO2 conversion rate is in the range 1%-2% [81]. Studies have shown that the V=O bond in the vanadium-titanium-based catalyst plays a key role in SO2 oxidation [82]. Therefore, the V2O5 content in the catalyst is usually set to be quite small to suppress SO3 generation. Water appears in the form of water vapor in the flue gas. The water vapor condensed on the catalyst surface not only exacerbates poisoning by the soluble salts of alkali metals, such as K and Na, but also vaporizes and swells as the temperature increases. This impairs the fine structure and leads to cracking of the catalyst. Water vapor has been reported to compete with NO and NH3 for physical adsorption on the catalyst surface, thereby decreasing the deNOx activity [48]. The alkali metals in fly ash (K, Na) also degrade the catalytic performance since they become deposited in the active centers of the catalyst. Thus, long-term accumulation can block the pores on the catalyst and lead to catalyst poisoning [79].
Ammonia slip, the residual ammonia present in the downstream flue gas after the SCR system, is an important factor that influences the normal operation of the flue gas treatment system. A high gas velocity, high NH3/NOx ratio, uneven NH3 and flue gas mixing, as well as a decrease in SCR catalytic activity all lead to the addition of slip ammonia [83]. This increase would severely affect the stability of the exhaust system. This occurs because slip ammonia further promotes the formation of the sulfate, leading to catalyst poisoning, a reduction in the heat transfer efficiency of the air preheater, corrosion, and blocking of the catalyst pores [84]. Moreover, studies have shown that NH3 competes for the active sites with Hg0. Hence, the Hg0 oxidation performance decreases with an increase in ammonia concentration [68]. Therefore, considering both the Hg0 conversion efficiency and the potential impact of ammonia emission, ammonia slippage from the SCR system should be minimized and is generally controlled to ~3 × 10-6. During actual operation, a reduction in the NH3/NOx ratio has been considered as the main reducing method of ammonia slippage from coal-fired power plants. Nevertheless, as previously discussed, this method is not conducive to the SCR reaction.
In view of the abovementioned influence factors, the vanadium-titanium-based catalysts in the NH3-SCR system have been modified in many studies to adapt to the various coal-fired power plant operating conditions to remove NO stably and efficiently. The main research directions are as follows: (1) broadening the reaction temperature of the catalysts, especially expansion to the low-temperature region, (2) improving the thermal stability, (3) enhancing the co-oxidation efficiency of elemental mercury, (4) controlling the slip ammonia, (5) reducing the rate of SO2 oxidation, (6) increasing the alkali resistance, and (7) others. In this paper, the research status of modified SCR catalysts in these aspects is analyzed and summarized.
When the boiler starts and stops or is operated under low load, the flue gas temperature may be lower than the optimum operating temperature of a conventional SCR catalyst. Improving the low-temperature deNOx efficiency of the SCR catalyst is beneficial to adapt to the actual operating conditions and cut down the energy consumption and deNOx cost. Furthermore, a high temperature would promote SO2 oxidation, which is not expected to happen in an SCR system. Accordingly, many researchers have recently attempted to develop low-temperature (< 250 ℃) SCR catalysts [84-87]. As they are placed before the desulfurization unit, these catalysts must withstand the high concentrations of SO2 present in the flue gas. This SO2 would lead to the destruction of the active components in the catalyst, while the surface-active sites would be covered by the generated metal and ammonium sulfates, which will completely inactivate the catalyst. Thus, poor sulfur resistance is considered to be the biggest bottleneck of most novel low-temperature SCR catalysts. Conversely, vanadium and other substances present in traditional SCR catalysts usually show excellent activity in high concentrations of SO2. Modification based on these components may be more feasible to develop ideal low-temperature SCR catalysts with good SO2 resistance.
Deng et al. [45] explored the influence of WO3 content on the low-temperature deNOx performance of SCR catalysts. According to their study, the V-W-Ti catalyst with increased WO3 content exhibited higher NH3-SCR activity at low temperatures compared to the vanadium-titanium catalysts. Particularly, the catalytic efficiency was > 80% at 180 ℃ when 7% tungsten was loaded. The highly dispersed V2O5 and WO3 on the catalyst surface and the exposed acidic sites may be responsible for this increase.
Tables 2 and 3 show some modified vanadium-titanium-based SCR catalysts with deNOx activity at low temperatures, including both metal and non-metal doping catalysts. The preparation method, modification effect, and anti-poisoning performance are summarized.
Owing to the strong redox ability and oxygen storage capacity of CeO2, ceria-based catalysts have been attracting increasing attention worldwide [88-90]. Ce can significantly increase both the low-temperature deNOx efficiency and reaction rate of vanadium-titanium-based SCR catalysts when doped into the carrier [91] or loaded as an additive [87, 92]. Current studies generally believe that Ce and V may work synergistically. Ce3+ species would cause charge imbalance, vacant sites, and unsaturated chemical bonds, which increase the amount of chemisorbed oxygen and NO oxidation. Rich surface acidity, strong redox properties, and active intermediate species are the main reasons for the high deNOx activity of Ce-doped SCR catalysts [87, 91-97]. Furthermore, Zhang et al. [94, 95] and Li et al. [96] respectively chose TiO2-ZrO2 and TiO2-carbon nanotubes (CNT) to replace raw TiO2 as the carrier for Ce-doping catalysts. In these cases, higher low-temperature deNOx activity was exhibited. However, the V2O5-CeO2/TiO2-ZrO2 catalysts possessed poor stability and H2O resistance despite the good resistance to SO2 poisoning. Additionally, the SO2 resistance of V2O5-CeOx/TiO2-CNTs varied greatly at different temperatures; thus, it was excellent at 250 ℃ but poor at 200 ℃. In contrast, the Zr-CeVO4/TiO2 catalyst prepared from TiO2 nanosheets (TiO2-NS) as the support showed the desired stability and H2O/SO2 durability in low-temperature SCR reactions [98]. This may occur because the formation of NH4NO3 and (NH4)2SO4/NH4HSO4 was significantly inhibited and abundant active oxygen species and Brönsted acid sites were not covered in the presence of H2O and SO2. Similarly, Pr, another rare earth metal, was also proven to be a good modifier [99]. When the doping content of Pr6O11 was 4 wt%, 98% of the NOx was converted into N2 at 220 ℃ by V2O5-MoO3/Pr6O11-TiO2. Strong SO2 and H2O resistances were also exhibited.
Doping the common transition metals Fe [100] and Cu [101, 102], which would increase the amount of adsorbed oxygen and acid activity centers on the catalyst surface, often shows good low-temperature SCR efficiency and anti-poisoning performance. The transition metals and vanadium would form a redox cycle, thus improving the redox behavior of the catalyst. For example, the denitrification activity of the V2O5-WO3/Fe2O3/TiO2 nanoparticle catalyst was enhanced in the range 200-400 ℃ because of the promotional effect of V4+ + Fe3+ ↔ V5++ Fe2+ (Fig. 2) [100]. Moreover, the presence of Cu2+ + V4+ ↔ V5+ + Cu+ (Fig. 3) promoted the Cu-doped V2O5/WO3-TiO2 catalyst to reach a deNOx efficiency of 84% and a N2 selectivity of ~95% at 250 ℃ [102]. However, the V-Fe/TiO2 nanoparticle catalysts did not show good SCR activity at low temperatures in the presence of 20 vol% H2O. Their denitrification efficiencies were < 60% at 200 ℃ [103]. In addition, although loading on TiO2 can improve the SCR activity at low temperatures (150-250 ℃) [104, 105], when added into V2O5/TiO2 as promoter, MnOx had a negative effect on the SCR performance and slightly reduced N2O generation [92].
Sb is another promising modifier for commercial SCR catalysts. Compared to other modified SCR catalysts, the Sb-promoted SCR catalyst performed better during NOx conversion at 150-400 ℃, and was still effective at 230 ℃ after 90 h of SO2 deactivation (Fig. 4) [106]. Since more ammonium bisulfates were formed on 10 wt% W-V2O5/TiO2 than on 2 wt% Sb-V2O5/TiO2 [106], antimony may be a good substitute for tungsten to act as the auxiliary of the catalysts in low-temperature or low-sulfur content environments. In addition, Sb-Ce co-doping [93, 97] of the SCR catalysts also achieved a desirable NOx reduction efficiency at low temperatures. Processing strong surface acidity and redox performance, 2 wt% Sb-2 wt% V2O5-10 wt% Ce/TiO2 revealed the highest NOx conversion rate of 84% and excellent N2 selectivity (> 95%) at 200 ℃ [93]. Moreover, after SO2 processing at high temperatures, excellent deNOx activity was still revealed [97].This was attributed to the substantial Ce3+ ions and acid sites existing on its surface after the sulfidation treatment.
The impregnation sequence of the promoters and V also influences the SCR activity and N2O selectivity [92]. According to research on Ce-, W-, Zr-, and Mn-SCR, the catalyst first added vanadium showed higher low-temperature deNOx activity. This superiority can be attributed to the formation of more V2O5 aggregates on the sample, which provides valuable information for the development of new SCR catalysts with excellent SCR performance.
Common non-metals such as F [107-111], N [112], and S [113] can also improve the low temperature denitrification performance of vanadium-titanium-based catalysts (Table 3). They are usually doped into the TiO2 supporter by a sol-gel method to change the catalyst structure. Anatase TiO2 is more favorable to the SCR reaction [114]. Fortunately, these doped non-metals can inhibit the phase transition of TiO2 from anatase to rutile according to a series of characterization methods such as XPS, TPD, and EPR [107-113]. Moreover, the surface acidity, oxygen vacancies, and active sites can also be increased. The electron donors NO and NH3 reduce vanadium, and superoxide ions (O2-) are then formed at the V sites. These superoxide ions promote the SCR reactions so that the SCR activity of the catalysts is directly proportional to the concentration of these superoxide ions. However, the coexistence of H2O and SO2 would synergistically inhibit the formation of V4+ and superoxide ions, thus restricting the SCR reaction [109].
These experimental results were also confirmed by theoretical calculations. Zhang et al. [115] investigated the cluster molecules of F-doped V2O5-WO3/TiO2 catalysts based on first-principle molecular dynamics and density functional theory. They found that atomic F replaced the bridge oxygen in TiO2 and formed an oxygen vacancy with high activity. Tungsten could interact with the oxygen vacancy, which can be promoted by vanadium, thus making the formation of superoxide ions advantageous.
Tran et al. [116] prepared TiO2-SiO2 supports using Ti-bearing blast furnace slags (BFS). On loading with V2O5 and WO3, their low-temperature deNOx performance was significantly better than that of V2O5-WO3/TiO2. The presence of SiO2 and some impurities such as CaO, MgO, Al2O3, Fe2O3, and SO42- increased the surface area and acidity and promoted adsorption of the reactant.
Heteropoly acids (HPA), including H3PW12O40, H4SiW12O40, and H3PMo12O40 [117], were also used as a modifier to enhance the low-temperature deNOx performance of commercial SCR catalysts. All the 15-wt% HPA-V2O5/TiO2 catalysts achieved ~100% deNOx efficiency at 300 ℃ when the V2O5 content was 3-6 wt%, which was better than that achieved by 10 wt% WO3-V2O5/TiO2. The addition of HPA could add acid sites and V4+ ions on the catalysts and promote the resistance to alkali poisoning. Further, neither N2O generation nor non-selective NH3 oxidation was increased.
Following long-time running at high temperatures, conventional SCR catalysts would be sintered and TiO2 would convert from the anatase to the rutile phase. Therefore, the surface area and the catalytic activity of the catalysts can be greatly reduced [118]. To mitigate this problem, researchers have committed to improving the thermal stability of SCR catalysts.
Adding silicon to the carrier may be an available measure. The unfavorable phase transformation of TiO2 and the decrease in catalytic surface area can both be inhibited by Si doping. After hydrothermal treatment at 750 ℃ in 10 vol% H2O/air for 24 h, the stable structure of a TiO2-WO3-SiO2 supporter caused the vanadium oxide to form the most active polymer, other than V2O5 crystals, (Fig. 5) and maintain a good dispersion [119]. However, Casanova et al. [120] hold that Si doping alone could not efficiently promote the stability of vanadium-containing catalysts. They added rare earth and transition metals to the V2O5/TiO2-WO3-SiO2 catalyst and studied the resulting SCR performance [120-124]. Among these doping metals, Fe, Ce, La, Nd, and Pr were relatively sensitive to heat treatment, while Tb, Er, Dy, Sm, and Gd showed outstanding anti-deactivation properties and high deNOx activity after thermal aging at 750 ℃ in air [120, 122, 123]. The formed rare earth vanadates inhibited the rutilization of anatase TiO2. These components only partially decomposed at high temperatures and thus, improved the thermal resistance of the catalysts [123]. Similarly, the FexEr1-xVO4/TiO2-WO3-SiO2 catalyst possessed better thermal stability than the FeVO4/TiO2-WO3-SiO2 catalyst due to the presence of erbium-iron vanadate after thermal aging at 750 ℃ (Fig. 6) [124].
Similarly, zirconium was used as a modifier to increase the thermal stability of the vanadium-containing catalysts by changing the vanadium speciation [125]. After aging treatment at 750 ℃ for 12 h, the 10 wt% Zr-1 wt% V2O5/TiO2-WO3 catalyst showed high SCR efficiency at 150-400 ℃. However, the deNOx activity on the fresh catalysts was reduced by zirconium doping at temperatures < 400 ℃, which can be attributed to the decreased V5+=O and Brönsted active sites.
The presence of tungsten and barium would also influence the thermal stability of the SCR catalysts [126]. After thermal aging at 600 ℃ for 6 h, the deNOx efficiency of 6 wt% W-2 wt% V2O5/sulfated TiO2 did not change significantly since TiO2 conversion was low. However, the 2 wt% Ba-2 wt% V2O5/sulfated TiO2 was dramatically deactivated due to the generation of V-O-Ba, an inactive vanadium species for the SCR reaction.
Yu et al. [127] found that the deposited potassium could form V-O-K species, inhibited the aggregation of vanadium oxides, and thus, improved the thermal stability of the catalysts (Fig. 7). Compared to unmodified vanadium-titanium-based catalysts, this species could increase the deNOx efficiency threefold after treatment at 800 ℃ for 4-5 h.
These studies may help us find more detail on the sintering and rutilization processes of TiO2, and are conducive to improving the stability and denitrification activity of SCR catalysts to increase their service life.
Mercury is a volatile and poisonous trace element that exhibits migration, persistence, high biological enrichment, as well as nerve toxicity. It can easily cause environmental pollution and serious health hazards to human beings and thus, has been recognized as one of the persistent environmental pollutants to be preferentially controlled globally [128, 129]. In 2013, the EPA updated the Mercury and Air Toxics Standard (MATS) requirements, in which the total mercury emissions from new units using low-rank raw coal as a fuel should be controlled to levels < 0.003 lb/GWh [130].
Currently, coal combustion accounts for ~45% of the anthropogenic mercury sources [131]. According to the literature, the mercury in coal-fired flue gas usually exists in three forms: elemental mercury (Hg0), oxidized mercury (Hg2+), and particulate mercury (Hgp) [132], of which Hg0 is the most difficult to handle because of its high volatility and low water solubility. Among the current Hg0 removal technologies from coal-fired flue gas, the use of catalysts to convert Hg0 to Hg2+, which can be removed by existing desulphurization equipment, is economical and practical. This technique has good future application prospects. In this method, the ratio of Hg2+ in the flue gas determines the mercury removal efficiency. Therefore, it is of great significance to improve the catalytic oxidation performance of Hg0.
Kamata et al. [67] investigated the Hg0 oxidation performance of metallic oxide-modified TiO2 catalysts. According to their study, the mercury oxidation efficiency is of the order of MoO3 > V2O5 > CrO3 > Mn2O3 > Fe2O3 > CuO > NiO in the absence of NH3. This indicates that V2O5, as the main active component in commercial SCR catalysts, is one of the most effective metal oxides for Hg0 catalytic oxidation. In fact, many mercury oxidation catalysts have been developed [133], such as Au/TiO2[134], CuO/TiO2 [135], CeO2/TiO2 [136], MNOx-CeO2/TiO2 [137, 138], CuO-CeO2/TiO2 [139, 140], V2O5/ZrO2-CeO2 [141], MNOx-CeO2-ZrO2 [142], Co-Ce-ZrO2 [143], Mn/γ-Al2O3 [144], CuCl2/γ-Al2O3 [145], Nb-Co-Ce/Al2O3 [146], Fe-Ti-Mn spinel [147], and Co-MF[148]. Although all these catalysts demonstrated a high mercury oxidation efficiency, their components are quite different from the traditional vanadium-titanium-based catalysts and thus, their application to deNOx flue gas is uncertain. Moreover, the addition of these mercury oxidation catalysts may require transformation of the existing flue gas treatment equipment, which will result in a significant increase in cost. Therefore, adding other substances without altering the existing components of the SCR catalysts, to enhance the Hg0 oxidation performance while maintaining high deNOx efficiency, has aroused great concern.
According to previous studies, modification by transition metals (Fe [149], Cu [150], Ag [151], Ru [152, 153], Ce [154-156], etc.), metal chlorides (CaCl2 [157], CuCl2 [158], etc.), and non-metallic species (Si [159], etc.) can improve the elemental mercury oxidation properties of V2O5-WO3/TiO2 catalysts, which are summarized in Table 4. The reaction conditions are also listed here since they may dramatically affect the mercury oxidation efficiency.
Huang et al. [149] investigated the influence of transition metal oxide doping on the mercury oxidation of traditional SCR catalysts. Fig. 8 reveals that in the presence of 8 × 10-6 HCl, ZnO, Y2O3, NiO, and ZrO2 tend to inhibit this process, while the reverse holds for Co3O4, Fe2O3, and MnO2. Their promotional effects can be ordered as Fe2O3 > MnO2 > Co3O4. Overall, 1% wt Fe2O3/SCR exhibited the best Hg0 oxidation efficiency (> 90%) at 350 ℃. Even in the presence of 400 × 10-6 SO2, its mercury oxidation effect was only slightly reduced to 80%. Hg0 may be oxidized to Hg0 by Fe2O3 on the Fe2O3/SCR catalyst surface [149]. When HCl was added to the flue gas, the intermediate (Hg-OH-Fe-Cl) was generated. Chemisorbed oxygen and Fe3+ would then oxidize HCl to active chlorine by the Mars-Maessen mechanism. Subsequently, the active chlorine then reacted with the adsorbed HgO and generated HgCl2 by the Langmuir mechanism. This process is presented in Fig. 9. Studies have shown that the CuO-V2O5-WO3/TiO2 catalyst exhibited excellent Hg0 oxidation properties, even in the absence of HCl. The redox cycle of V4++ Cu2+ ↔ V5++Cu+ would generate abundant chemisorbed oxygen (Oα), which could oxidize Hg0 into HgO (Fig. 10) [150].
Ag doping significantly improved the Hg0 catalytic oxidation performance of the vanadium-titanium-based catalyst [151]. More than 90% Hg0 can be converted in the simulated flue gas, which contains 5 × 10-6 HCl at 350 ℃. This may be attributed to the addition of Ag that weakened the strength of the V-O bond and reduced the desorption activation energy of the surface oxygen species. Vanadium existed in a higher oxidation state so that the oxidation ability of the catalyst was improved and the reaction temperature was reduced. The possible Hg0 oxidation mechanism of this catalyst was also investigated. At low temperatures, amalgamate Hg0 and adsorbed Hg0 reacted with the adsorbed HCl and formed HgCl, the intermediate before HgCl2 formation. The active chlorine or chlorine atom would also react with gaseous Hg at high temperatures.
Ru/SCR can also increase Hg0 oxidation to > 90% with the aid of 5 × 10-6 HCl at 350 ℃. Yan et al. [152, 153] considered the Deacon reaction as its main mechanism. In fact, the Deacon reaction of RuO2 is quite different from that observed for the transition metal oxides (e.g., MnO2). The gaseous oxygen is easily adsorbed on the coordinatively unsaturated Ru site (cus-Ru) and activated to reactive oxygen, which then captures hydrogen from HCl. Therefore, HCl was activated and formed Cl2 [161, 162]. Hg0 can be adsorbed at two original cus-Ru sites and then chlorinated at the cus-Ru sites near the chlorine atoms. Moreover, gaseous Hg0 can even react directly with atomic chlorine. The reaction is as follows [152]:
Li et al. [159] modified the commercial SCR catalyst with Si, a common non-metallic element. They suggested that Hg0 oxidation follows the Eley-Rideal mechanism. In this mechanism, HCl, NO, and NO2 are first adsorbed on the V2O5 active sites and then react with gaseous Hg0. According to this study, the SiO2-TiO2-V2O5 catalyst revealed a higher Hg0 oxidation efficiency than that observed with SiO2-V2O5, possibly because the V-O-Ti bond is more active than the V-O-Si bond. Particularly, in the presence of 10 × 10-6 HCl, the maximum Hg0 oxidation efficiency of the prepared catalyst reached 99% when the weight percentages of TiO2 and V2O5 were 6% and 5%, respectively. This demonstrated that non-metallic doping can efficiently improve the mercury oxidation performance and even performs better than widely studied metals. Thus, this process should be studied systematically and deeply.
Since the presence of HCl plays a vital role in the oxidation of Hg0, researchers have started to select metallic chlorides as modifiers to improve the Hg0 oxidation efficiency of low chlorine coal-fired flue gas. For the CaCl2-SCR catalyst, 1.0 wt% CaCl2 addition increased the mercury oxidation efficiency from 7.7% to 78.8% in the absence of HCl at 350 ℃. Hg0 oxidation could be attributed to both the active chlorine and the HCl produced by CaCl2 hydrolysis [157]. In general, the active chlorine is considered as an important substance in mercury oxidation by chloride-modified catalysts [163]. Fig. 11 illustrates that gaseous Hg0 would react with active chlorine on the surface and generate gaseous HgCl with low energy (a); some active chlorine could remove the surface binding and form adsorbed HgCl if the mercury atom was adsorbed on the active site near V2O5 (b). These two types of HgCl are both oxidized to HgCl2 by active surface chlorine or chlorine atoms and then removed from the surface. Therefore, an increase in active chlorine species could significantly enhance Hg0 oxidation. For the CuCl2-SCR catalysts, both active Cl species and Cu2+ ions present on its surface played roles in Hg0 oxidation. Fig. 12 illustrates that the Cl species released by CuCl2 react with the adsorbed Hg0 to form HgCl2, while CuCl2 is reduced to CuCl. The formed CuCl would react with O2 and generate the intermediate copper oxychloride (Cu2OCl2). This could be re-chlorinated to CuCl2 by the adsorbed HCl on the catalyst surface when HCl is added to the flue gas [158]. Therefore, doping of CuCl2 significantly improved the Hg0 oxidation efficiency.
Almost all the studies on CeO2-modified SCR catalysts showed that > 80% Hg0 efficiency could be achieved in the absence of HCl [154-156]. CeO2 and V2O5 would work synergistically in these catalysts. The formed redox cycle of Ce4+ + V4+ ↔ Ce3+ + V5+ not only greatly improved the NO conversion rate but also promoted Hg0 oxidation. This Hg0 oxidation process follows the Mars-Maessen mechanism, where the adsorbed Hg0 combines with the chemisorbed or lattice oxygen, produced by O-O bond breakage on the V2O5 surface, and forms the weakly bonded Hg-M-Ox-1 (M = Ce or V). HgO is generated thereafter (in Fig. 13). The presence of gaseous O2 would promote the regeneration of the lattice and chemically adsorbed oxygen; thus, a reaction cycle is formed. Additionally, V2O5-WO3/TiO2-CeO2 showed good resistance to SO2 and H2O [154, 155]. On the other hand, the presence of SO2 and H2O would reversibly inhibit NO transformation and Hg0 oxidation over the V2O5-CeO2/TiO2 catalyst and thus, limits its application [156].
The co-modification of Ce and Cu was also found to be effective in improving the Hg0 oxidation in the absence of HCl [160]. These two dopants form the redox cycle Cu++Ce4+ ↔ Cu2++Ce3+ in Ce-Cu/SCR catalysts, which promotes the formation of more chemisorbed oxygen and thus, exhibits good redox properties and Hg0 oxidation performance.
Notably, over all these catalysts, ammonia exhibits the most significant inhibitory effect on mercury oxidation, which is similar to the phenomenon observed for traditional SCR catalysts. To some extent, the NO reacted with most of the NH3, thereby reducing this inhibitory effect. However, the slip ammonia that is not consumed in the denitration reaction should be removed to ensure good Hg0 oxidation performance.
It is generally recognized that automotive ammonia can be effectively converted to harmless N2 by dual layer (noble metal-SCR) catalysts [164, 165]. However, to date, no suitable solution to mitigate slip ammonia from the SCR system of coal-fired power plants has been found. Unfortunately, reducing the NH3/NO ratio, as the main existing control method, would significantly reduce the deNOx efficiency. Therefore, some researchers have focused on the preparation of new types of catalysts to inhibit slip ammonia while maintaining satisfactory deNOx performance.
Owing to the inherent characteristics of coal-fired flue gas, the catalysts suitable for slip ammonia treatment in power plants require an excellent ammonia conversion efficiency and nitrogen selectivity at high SO2 concentrations and ~350 ℃. In general, the temperature of the catalytic decomposition of ammonia is > 500 ℃ [166], while the selective catalytic oxidation (SCO) of ammonia requires relatively low temperatures, thus may be an alternative process.
Chen et al. [152] found that a Ru-modified V2O5-WO3/TiO2 catalyst would perform well on slip ammonia oxidation. When installed behind the commercial SCR catalyst, it could effectively oxidize the slip ammonia in the flue gas as keeping a relatively high NO conversion efficiency. Additionally, its N2 selectivity can reach ~90%. The addition of Mo could further promote the ammonia oxidation performance of Ru-SCR catalysts [153]. At 350 ℃, the Mo-Ru/SCR catalyst can remove > 95% slip ammonia, of which 97% is converted to harmless N2. Its ammonia oxidation efficiency can still reach > 70% in the presence of SO2. Simultaneously, > 99% Hg0 could be oxidized when the HCl concentration in the flue gas was only 5 × 10-6. Thus, Hg0, NH3, and NOx can be efficiently removed simultaneously when a Mo-Ru/SCR catalyst is used as a SCR-plus catalyst (Fig. 14).
The unsaturated Ru (cus-Ru) and reactive oxygen species on the Ru/SCR catalyst surface play important roles in ammonia oxidation. The NH3 adsorbed on the surface would form the intermediate product with high activity, which is much more easily oxidized (Eqs. (7)-(12) [153]). The addition of Mo further increases the amount of cus-Ru, and thus, promotes the oxidation of slip ammonia.
However, addition of the noble metal ruthenium greatly increases the cost of the catalyst, which is not conducive to commercial application. To date, no other studies of vanadium-titanium-based catalysts for controlling slip ammonia have been reported. Undoubtedly, there are wide areas and great application values to seek more economical and efficient SCO catalysts for slip ammonia from coal-fired power plants.
In addition to burning low-sulfur coal, mixing coal, and spraying calcium- or magnesium-based absorbents, scholars have tried to control SO3 generation from coal-fired power plants through the modification of commercial SCR catalysts. Changing the physical properties of the catalyst and doping other compounds into the traditional vanadium-titanium-based catalyst can both reduce the SO2 oxidation efficiency.
DeNOx reaction and SO2 oxidation occur at different parts of the vanadium-based catalysts. According to chemical kinetics studies, the SCR reaction, controlled by external diffusion, mainly occurs in the 0.1-mm surface of the catalyst. On the other hand, the SO2 oxidation process occurs in the whole catalyst and thus, the SO2 conversion rate is proportional to the wall thickness of the SCR catalyst [167]. Therefore, decreasing the wall thickness or using flat catalysts can reduce the formation of SO3 [168]. Others hold that corrugated SCR catalysts may have a lower oxidative effect on SO2 than other shapes of catalysts [169].
With the growth of the catalyst service life, the SO2 oxidation rate of SCR catalysts would increase [170]. In addition to K2O/TiO2, all the TiO2-loaded metal oxides can oxidize SO2 into SO3. Their SO2 oxidation performance is as follows: V2O5/TiO2 > Fe2O3/TiO2 > Re2O7/TiO2 > CrO3/TiO2 > Nb2O5/TiO2 > MoO3/TiO2 > WO3/TiO2 [171]. Since V2O5/TiO2 displays a strong SO2 oxidation ability [114], reducing the vanadium content in the SCR catalyst can inhibit SO3 generation; however, the deNOx efficiency will also significantly decrease [171].
As early as 1981, Morikawa et al. [172] have studied the SO2 oxidation of vanadium-titanium-based catalysts doped with metal oxides. Their oxidation activities are of the order of: WO3-V2O5-TiO2 > MoO3-V2O5-TiO2 > V2O5-TiO2 > Ta2O5-V2O5-TiO2 > Y2O3-V2O5-TiO2 > GeO2-V2O5-TiO2 > ZnO-V2O5-TiO2. Thus, WO3 or MoO3 doping can enhance the SO2 oxidation of vanadium-titanium-based catalysts, while GeO2 and ZnO would inhibit it. Some studies have suggested that tungsten can promote SO2 oxidation [173, 174]. However, others argue that doping with WO3 or MoO3 would increase the surface acidity, which can effectively reduce the SO2 conversion rate [175]. When the loading mass fraction of MoO3 is > 9%, the SO2 conversion rate could be < 1% [176]. Moreover, the decomposition of NH4HSO4 can be markedly enhanced by WO3 doping, which promotes the reaction of NH4HSO4 and NO and inhibits N2O production during the reaction [177].
Shi et al. [178, 179] doped the transition metal Cu into a traditional SCR catalyst. A 0.6% addition reduced the SO2 conversion rate from 0.73% to 0.68%, while maintaining a denitrification efficiency of 94.47%. They found that the amount of V4+ ions in the doped catalyst increased, while those of the V5+ ions and lattice oxygen decreased. This may abate the transfer of lattice oxygen, thereby inhibiting SO2 oxidation. Theoretical calculations have shown that the oxygen charge in the V-O-M oxygen bridge changed from negative to positive in Me-doped catalysts. This indicated that the basicity of oxygen decreased and thus, SO2 oxidation was inhibited.
Li et al. [180] found that doping BaO, MgO, SrO, and their mixtures would reduce the SO2/SO3 conversion rate of the SCR catalysts in high-sulfur flue gas, wherein the SO2 oxidation rate of BaO addition is the lowest. Adding SiO2 into the three single-doped catalysts can significantly improve the denitrification efficiency, which is originally slightly reduced. Their SO2 oxidation rate was only ~0.5%, while the denitrification efficiency was > 80% at 300-440 ℃. Some studies found that vanadium tended to be present in a lower oxidation state when BaO [181] or SiO2 [173, 174, 182] was added. The redox capacity and SO2 adsorption sites were also reduced and thus, SO2 oxidation was inhibited. For example, the V2O5-BaO/TiO2 catalyst, which contains 2% BaO, showed less than half the SO2 oxidation efficiency of an unmodified or tungsten-added catalyst. Further, when the Ti/Si molar ratio in the V2O5/TiO2-SiO2 catalyst was 8/2, a high deNOx efficiency was maintained, while the SO2 conversion was only about a third of that of the V2O5 /TiO2 catalyst [174].
Rare earth, niobium, and silicon-aluminum composite oxides have also been used simultaneously as additives of vanadium, titanium, and tungsten to prepare a new catalyst for low SO3 generation [183]. This prepared catalyst showed the same deNOx activity as the conventional SCR catalyst, while the SO2 oxidation rate was reduced by 20%-40%. However, its inhibitory mechanism has not been revealed.
In general, doping of the oxidizing substances can improve the Hg0 and slip NH3 oxidation efficiency of the SCR catalyst. However, SO2 oxidation may also be promoted. This will lead to catalyst poisoning, thereby causing a reduction in the Hg0 and NOx removal efficiencies. However, the additives that inhibit SO2 oxidation tend to reduce the oxidation properties of the catalysts. Moreover, NH3 oxidation is an adverse reaction of the SCR process, which may reduce the denitrification efficiency. Therefore, both the oxidation capacity and the selectivity of the catalysts should be taken into account when choosing the appropriate modified compounds. As a result, it is rather complicated to inhibit SO2 oxidation while efficiently reducing NOx, oxidizing mercury, and slip ammonia at the same time. A comprehensive selection of catalyst formulations and preparation methods should therefore be made.
Studies have shown that alkali metals in the flue gas would deposit on the surface of the catalysts used in coal-fired power plants [184]. This can lead to catalyst deactivation when these substances are in excess [185, 186]. These alkali metals would significantly decrease both the surface acidity and reducibility of the vanadium-titanium-based catalysts, change the V=O groups, and thus, inhibit the SCR reaction [185, 187, 188].
Fortunately, the modification of ceria would improve the alkali poisoning resistance of the V2O5-WO3/TiO2 catalyst. Fig. 15 reveals that the deNOx activity of V2O5-CeO2-WO3/TiO2 was much higher, even after doping with K and Na [185]. This phenomenon may be due to the good oxygen storage capacity of ceria and the high reducibility produced by the redox properties of Ce4+ to Ce3+. Additionally, the adsorption amount of alkali metals on the active V species was reduced since the K or Na atoms would easily bond to the CeO2 (110) surface [187]. The co-doping of Sb and Nb can also enhance the K2O poisoning resistance of the V2O5/TiO2 catalyst. The K2O-doped 1 wt% V2O5-Sb-Nb/TiO2 (atomic ratio of Sb/V = 0.5, Nb/V = 1) catalyst even demonstrated higher deNOx performance than fresh vanadium-titanium catalysts [189]. However, the relevant mechanism was not explained in this study. Moreover, the addition of Nb possibly facilitated the dissociation of water, which generated surface hydroxyl species and increased the surface acidity. Therefore, the V2O5-Sb-Nb/TiO2 catalyst showed higher water resistance than the vanadium-titanium catalysts.
In fact, compared to the study of the SO2 and H2O resistances of the catalysts, the influence of alkali metal poisoning has rarely been investigated. This could be a promising research direction for the industrial application of SCR catalysts.
In addition to the above five types of catalysts, some other modified catalysts have also been studied. For example, modification by Mn [190], Si [191], nanotubes [192], and cordierite supports [14] would further increase the denitrification activity; low-cost Fe [193, 194] and Cu [195] enhance N2 selectivity; addition of BaSO4 [196] improves the SO2 resistance; and cordierite-loaded SCR catalysts [14] possess excellent mechanical strength and wash coating stability. These catalysts are summarized in Table 5.
Although for several years, new types of SCR catalysts have aroused great interest, vanadium-titanium-based catalysts remain unchallenged owing to their excellent performance and experience in application. However, their low-temperature deNOx efficiency, thermal stability, and synergistic mercury oxidation performance, still leaves much to be desired and the problems of SO2 oxidation and ammonia slip need to be urgently addressed.
By doping different materials such as noble metals (e.g. Ag, Ru), transition metals (e.g. Mn, Fe, Cu, Zr, etc.), rare earth metals (e.g. Ce, Pr), and other metal chlorides (e.g. CaCl2, CuCl2) and non-metals (F, S, Si, N, etc.), these modified vanadium-titanium catalysts have been proven to exhibit a corresponding promoted performance. Their reported advantages and disadvantages are summarized in Table 6.
In fact, owing to equipment and experimental condition limitations, the N2O selectivity of many of these catalysts was not considered in previous studies, and the impact of flue gas composition such as SO2 and H2O on the catalyst properties is still unclear. Further studies on their shortcomings and practicality are also needed.
In summary, more studies are needed to further improve the comprehensive performance of modified vanadium-titanium-based SCR catalysts to adapt to various operating conditions in coal-fired power plants. Future research may involve the following aspects:
(1) Great effort should be made to explore vanadium-titanium-based catalysts with high deNOx performance, high N2 selectivity, good SO2 and H2O resistance, and a wide operating temperature window. Considering both catalytic activity and cost, transition metal oxides, rare earth metal oxides, and non-metallic elements are still the main modifiers. Moreover, modification of the carrier is also a possible research direction.
(2) As the key to the synergistic control of mercury pollution by SCR catalysts, Hg0 oxidation performance and its enhancing method should be thoroughly investigated. Moreover, slip ammonia in the tail of the SCR system significantly inhibits Hg0 oxidation. The reaction of slip ammonia and the generated SO3 would affect the stable operation of the catalyst and the downstream flue gas equipment and should therefore be inhibited. Much effort should be made to realize the cooperative control of multiple pollutants in the SCR system.
(3) Mechanisms affecting the catalytic properties of the modified constituents should be revealed. Many experimental results have shown that doping suitable metallic oxides and non-metallic elements can significantly enhance the NH3-SCR performance of vanadium-titanium-based SCR catalysts. This is attributed to many aspects, including a larger surface area, increased acidic sites, and added surface oxygen. However, their detailed mechanisms are still unclear. The mechanism of mercury oxidation and poisoning of vanadium-titanium-based SCR catalysts should also be studied intensively. Moreover, blind experimental screens would waste resources and research efforts, which could be avoided by theoretical calculation of possible modified components.
(4) Larger pilot tests are also needed to develop modified SCR catalysts suitable for commercial application. To date, most of the catalysts and the flue gas produced at the laboratory level are somewhat different from the actual situation, and thus, there is insufficient data to exhibit the actual effect of the catalysts in practical applications. Studies that more closely simulate the actual industrial conditions are needed to further investigate the value of using these catalysts.
2018,
Vol. 39 Issue (8): 1347-1365 
PDF
2018,
Vol. 39 Issue (8): 1347-1365 PDF