1. Introduction

Pollutants, such as sulfur dioxide (SO₂), nitric oxides (NO_x), trace heavy metals and particulate matters (PMs) are typical components in industrial exhaust flue gas, which are harmful to both the environment and human health. SO₂ and NO_x are the main precursors of acid rain and toxic chemical smog [1, 2]. Elemental mercury (Hg⁰) is among the most toxic heavy metals owing to high toxicity, volatility and bio-accumulation [3]. The flue gas contains three mercury species: Hg⁰, oxidized mercury (Hg²⁺), and particle bound mercury (Hg_p) [4]. Most of the Hg_p and Hg²⁺ can be effectively collected by conventional air-pollution control devices (APCDs) such as electrostatic precipitators, fabric filters, and wet flue-gas desulfurization (WFGD) equipment, which are broadly installed in plants around the world [5]. However, Hg⁰ is relatively stable in the atmosphere, and (due to its chemical inertness, water insolubility, and volatility) is hardly removed by existing APCDs [6].

The current techniques for removing SO₂, NO_x, and Hg⁰ are selective catalytic reduction (SCR) [7, 8], WFGD [9], and activated-carbon injection [10, 11]. However, individual pollutant-treatment systems are expensive to construct and operate, complex, prone to instability, and require a high occupational area [12]. The current technologies for simultaneous removal of SO₂, NO_x and Hg⁰ include carbon-based material adsorption [13-15], photocatalytic oxidation [16, 17], wet scrubbing [18-20], NH₃-SCR [7, 8], and non-thermal plasma (NTP) technologies [21, 22]. Carbon-based material adsorption and wet scrubbing have the disadvantages of secondary pollution and high investment costs. Photocatalytic oxidation and NTP have the disadvantage of limited industrial applications. Compared with these technologies, NH₃-SCR catalytic technology has the advantages of high efficiency, low energy consumption and low operating cost. This makes the catalytic method to remove SO₂, NO_x and Hg⁰ has become the main research direction for air pollution control. Initially, this research focused on SO₂, NO_x, and Hg⁰ removal by catalysts without carriers, but the removal effect was non-ideal [23]. In later studies, the SO₂, NO_x, and Hg⁰ removal was improved by loading active components (transition metals, rare metals, precious metals) on various carriers (carbon materials, metal oxides, metal–oxide frameworks (MOFs), zeolites) [24-28]. This review concludes the advantages and disadvantages of simultaneous removal methods, introduces the catalysts prepared by conventional and new carriers, and discusses the low-temperature performances of the catalysts. The mechanism of pollutant removal by catalysts is also discussed, as well as the possible research directions and future prospects.

2. Removal technologies

Simultaneous removal technologies include carbon-based material adsorption [13-15], photocatalytic oxidation [16, 17], wet scrubbing [18-20], NH₃-SCR [7, 8], and non-thermal plasma (NTP) technologies [21, 22]. The removal techniques are summarized in Fig. 1. In carbon-based material adsorption, pollutants are often removed by modified activated carbon, active coke, or activated-carbon fibers, which have porous structures and strong adsorption capacities. However, carbon-based material adsorption is limited by high operating cost and quality degradation of fly ash [29]. Under UV/visible radiation, pollutants are oxidized by positively charged holes on the photocatalyst surface. A photocatalytic oxidation system can be worked at relatively low temperature and pressure [30], but the UV light source raises the energy consumption. In wet scrubbing technology, the oxidant solution in the scrubbing system (such as NaClO₂, KMnO₄/NaOH or H₂O₂) can oxidize NO and Hg⁰ to soluble NO₂ and Hg²⁺, which can then be removed by chemical solutions. However, this method produces large volumes of wastewater and high secondary pollution. NTP is a recently developed gas-phase oxidation method that removes multi-pollutants from flue gas [31, 32]. NTP technologies remove pollutants quickly and efficiently, operating at atmosphere pressure and room temperature with no secondary pollution [33, 34]. Their defects are high energy consumption and limited amount of flue gas; moreover, the high-voltage electricity poses a safety hazard that limits its industrial application [35]. Synergistic removal of NO and Hg⁰ by existing NH₃-SCR systems promises to overcome the disadvantages of previous simultaneous multi-pollutant-removal technologies. In the NH₃-SCR system, NO is reduced to N₂ and H₂O by NH₃, and Hg⁰ is oxidized to Hg²⁺ by O₂. The generated Hg²⁺ attaches to PM surfaces and is finally removed by electrostatic precipitators (ESP) or wet flue-gas desulfurization (WFGD) [36]. The NH₃-SCR system improves the purification efficiency and lowers the operating costs of industrial exhaust flue-gas purification, without generating secondary pollution [12, 37, 38]. The present article systematically reviews the catalyst types, the key factors controlling the reaction processes, and the proposed reaction mechanisms of NH₃-SCR technologies.

3. Simultaneous catalytic removal

The existing catalysts for simultaneous removal of NO and Hg⁰ have been studied in the absence [23, 39, 40] and presence of carriers. Carriers can be divided into two main categories: conventional and new. Carbon-based materials [15, 41-43, 44], metal oxides [26, 27, 45-47] and silica [48, 49] are among the conventional carriers, whereas molecular sieves [50-53], MOFs [28, 54] and pillared interlayered clays (PILCs) [24, 55] are new carriers. This paper reviews the differences between the conventional and new supported catalysts in the simultaneous removal of SO₂, NO and Hg. Most supported metal–oxide catalysts are prepared by the impregnation method. In the presence of different carriers, the same active component can exhibit very different activities. However, as most catalysts are studied only in the laboratory, more research on their practical applications is needed.

3.1. Catalysts 3.1.1. Unsupported catalyst

Many catalysts used to remove NO_x, SO₂ and Hg⁰ have been reported in the literature. Among the catalysts without carriers [23, 39, 40] are the CeO₂-TiO₂ (CeTi) catalysts of Li et al. [39], which oxidize elemental mercury. They found that when the weight ratio of CeO₂/TiO₂ is 1:2, the CeTi catalyst exhibits higher Hg⁰ oxidation activity in 150−250 ℃, indicating that their superior performance was conferred by the high focuses of surface cerium and oxygen. Zhang et al. [23] researched TiO₂-Al₂O₃-CeO₂ materials synthesized by the sol-gel method, which simultaneously remove NO and Hg⁰. They found that TiAl₁₀Ce₂₀ nanoparticles supplemented by 10 wt% Al₂O₃ removed NO and Hg⁰ with superior efficiency (93.41% and 80.54%, respectively) in an SCR atmosphere at 300 ℃. Song et al. [56] studied the effect of Mn, Ti and Cu doping on simultaneous removal of NO and SO₂ of Fe_2.5M_0.5O₄ catalyst. The results show that the doping of the three metal oxides can significantly increase the removal rate of NO, but has no effect on the removal rate of SO₂, because SO₂ can be completely removed under most conditions. Sun et al. [57] prepared a series of MFe₂O₄ (M=Mn, Zn, Cu, Ni, Co) catalysts to remove NO and SO₂ by hydrothermal method. They found that the removal rate of SO₂ at room temperature can reach 100%, and the CuFe₂O₄ catalyst has the highest activity compared to other catalysts. Shi et al. [58] studied the natural ferrous manganese ore to simultaneously remove NO, Hg⁰ and toluene from flue gas. Activity evaluation experiments show that the maximum efficiencies of NO, Hg⁰ oxidation are 95.29%, 83.0%. The characterization results show that the oxidation activity of manganese ore is mainly provided by Mn⁴⁺ and Fe³⁺ in it.

3.1.2. Catalysts based on conventional supports

Carbon-based catalysts: Carbon-based materials are the most commonly used adsorbents, which are often used to remove multi-air-pollutant from flue gas, including activated carbon, activated coke and activated carbon fiber [59-61]. Using an activated-carbon carrier, Song et al. [62] prepared 1%M-6%Y/AC (where M is Ni, Al or Mo) by equal-volume impregnation. The catalytic activity of simultaneous SO₂ and NO_x removal was maximized on the 1%Ni-6%Y/AC catalyst. Sumathi et al. [63] prepared a palm shell activated carbon (PSAC) adsorbent and loaded several metal oxides (Ni, V, Fe and Ce) on the ride to simultaneously remove SO₂ and NO_x. Among their sorbents, CeO₂/PSAC yielded the maximum sorption capacity for the simultaneous removal of SO₂ and NO_x. Sun [64] studied the catalytic activity of copper and iron-loaded microwave coconut shell activated carbon (MCSAC) with different loadings for simultaneous denitrification and mercury removal. The 15%Cu-5%Fe/MCSAC catalyst exhibited the best removal efficiency. The NO removal rate was 86% and the mercury removal rate was 61% at 225 ℃. In addition, effect of NH₃ concentration on removal of NO and Hg⁰ was investigated, as shown in Fig. 2a.

Zhang et al. [65] studied the removal effect of FeCeO_x/CNTs catalyst for NO and Hg and the SO₂ tolerance. They found that the NO and Hg removal rates of the catalyst were only 15.44% and 16.13%. Coating FeCeO_x/CNTs catalyst with a layer of SiO₂ can significantly improve the activity and SO₂ tolerance of the catalyst. Li et al. [66] used activated coke as a medium to simultaneously remove SO₂ and NO_x from flue gas. It is found that AC still has high desulfurization and denitration efficiency after regeneration cycle. Zhao et al. [67] employed a CuO-MnO_x/AC-H catalyst for the simultaneous removal of Hg⁰ and NO from flue gas. They pointed out the significant role of the reaction temperature in physisorption and chemical oxidation. The optimal temperature for simultaneous removal of Hg⁰ and NO_x was 200 ℃. At this temperature, the 8%CuO-5%MnO_x/AC-H catalyst achieved simultaneous removal efficiencies of 90% for mercury and 78% for NO_x. Zhu et al. [68] synthesized a series of V₂O₅-xCeO₂/AC catalysts to investigate the capabilities of the catalysts to simultaneously remove Hg⁰ and NO in the temperature range from 120 ℃ to 240 ℃. They discovered through experiments that compared to 1 V/AC and 1 V-8Ce/AC, 8Ce/AC catalyst has the highest Hg⁰ removal performance (E_Hg > 98%) at 120−240 ℃. The 1 V-8Ce/AC catalyst showed the best NO conversion performance compared to 1 V/AC and 8Ce/AC, 8Ce/AC, and the E_NO increased gradually from 55% to 86.9% with the increase of temperature (120−240 ℃), as shown in Fig. 2b. Zhu et al. [44] used ammonium halide modified rice husk char to remove gaseous mercury. They found that the RHC adsorbent impregnated with ammonium bromide has excellent mercury removal performance, and can still reach an efficiency of about 80% for removing gaseous mercury. In addition, they found that this enhanced HgO removal adsorbent can synergistically remove SO₂ and NO. Gao et al. [69] studied a series of CoO_x-CeO₂ loaded biomass activated carbon (CoCe/BAC) derived from maize straw. When removing both NO and Hg⁰ at 230 ℃, the 15%Co_0.4Ce_0.6/BAC sample achieved 96.8% Hg⁰ removal efficiency and 84.7% NO removal efficiency. Based on characterization analyses, the author attributed the best performance of 15%Co_0.4Ce_0.6/BAC to its largest specific surface area, lowest crystallinity and strongest redox ability than other samples in the series. The NO removal was unaffected by low SO₂ concentration in the gas stream, and the NO conversion efficiency reached 80%, as shown in Fig. 2c.

Metal-oxide-based catalysts: Titania has stable physical and chemical properties, is non-toxic, and its surface area is below 100 m²/g [70]. TiO₂ has a certain acidity which enables it to be the main site of NH₃ adsorption in the NH₃-SCR reaction and thus beneficial to the removal of NO_x. Therefore, TiO₂ is a common support widely used in NH₃-SCR reaction. The most prevalent commercial catalyst is V₂O₅-WO₃/TiO₂, which delivers high activity but is active only within a narrow temperature window (300–400 ℃) [71]. Zhao et al. [46] found that adding CeO₂ to the V₂O₅-WO₃/TiO₂ catalyst can improve its catalytic activity for Hg⁰ oxidation without using HCl. The co-catalyst CeO₂ shows satisfactory oxidation activity and improved water oxidation resistance. The promoted catalysts showed satisfactory oxidation activity and improved H₂O resistance. Wang et al. [72] studied the oxidation catalytic performance of VWTi catalysts with different loadings of CuO. The results show that the 7%Cu/VWTi exhibited high Hg⁰ oxidation and a desired NO removal efficiency at 280−360 ℃. Meng et al. [73] prepared a series of Mn-V-W/TiO₂ (Mn-VWT) catalysts by impregnation to simultaneously remove Hg⁰ and NO. The results show that Mn-VWT-400 has the best simultaneous removal performance. Its removal efficiency of Hg⁰ and NO reached 100% and 82%, respectively. Yang et al. [74] used CeO₂ to modify V₂O₅/TiO₂ for simultaneous removal of NO and Hg⁰. Experiments show that CeO₂ modification not only significantly improves the SCR performance, but also significantly enhances the oxidation of Hg⁰. Wang et al. [75] investigated the sulfated CuCl₂/TiO₂ catalyst to simultaneously remove NO and Hg. By characterizing the catalyst before and after vulcanization, it was found that the CuSO₄ generated on the catalyst increased the number of acid sites on the catalyst surface, thereby improving the SCR activity. CuCl₂(S₁₈)/TiO₂ catalyst showed simultaneous catalytic oxidation of Hg⁰ (73.9%) and NO (87%) at the temperature of 325 ℃. Wang et al. [76] synthesized CeO₂ supported on anatase TiO₂ with high-energy facets and used it in the NH₃-SCR reaction. In the activity evaluation experiment, it was found that compared with Ce/P25 catalyst, it has high SCR activity. Through the characterization, it is concluded that the unique feature of the active energy (001) facet improves the thermal stability of CeO₂, and the presence of Ti³⁺ on the surface of TiO₂ effectively promotes the development of the SCR process, both of which lead to the remarkable catalytic performance for the catalyst. Chen et al. [77] studied the SCR activity of MnCe oxide-supported titanate nanotubes (TNT). And compared with the conventional MnCe/TiO₂ and V₂O₅/TiO₂ nanoparticle catalysts, it is found that MnCe/TNTs catalysts have higher SCR activity, sulfur and water resistance. The characterization results show that the BET surface area of MnCe/TNTs is larger, and the adsorption capacity of NH₃ is greatly enhanced, which is beneficial to the SCR activity. Liu et al. [78] designed the MnSmCo/Ti catalyst to simultaneously remove NO and Hg⁰ at low temperature. The experimental results show that the catalyst exhibits NO conversion rate of 80% and Hg⁰ removal rate of 100% at a gas hourly space velocity of 100, 000 h⁻¹ in the temperature range of 150−250 ℃. Li et al. [79] prepared a series of molybdenum (Mo)-modified vanadium (V)-based selective catalytic reduction (SCR) catalyst samples using the impregnation method to remove elemental mercury (Hg⁰) and nitrogen oxides (NO) at the same time. Through experiments, it is found that the metal loading and the order of impregnation have no effect on the denitration activity of the catalyst, and have a particularly large effect on the mercury removal activity. The dipping sequence experiment found that MoO₃(7) + WO₃(3) → V₂O₅(0.5)/TiO₂ exhibited the highest mercury removal activity. The metal loading experiment found that V₂O₅(0.5) + MoO₃(7)/TiO₂ catalyst exhibited the highest mercury removal activity. Liu et al. [80] studied the simultaneous removal of NO and Hg with Sb-modified Mn/TiO₂ catalyst under low-temperature SCR conditions. The results show that the MnSb/TiO₂-0.25 catalyst can achieve NO removal (> 90%) and Hg⁰ oxidation (> 80%) at 200−300 ℃. Li et al. [26] studied The effect of flue gas composition on the removal of Hg⁰ and NO by a series of CeO₂ (ZrO₂)/TiO₂ catalysts. At 250–400 ℃, the CeO₂(ZrO₂)/TiO₂ catalyst exhibited high Hg⁰ catalytic oxidation efficiency (> 95%) and high denitration efficiency (> 95%). Liu et al. [45] prepared Co/TiO₂ catalysts by the sol–gel method, and determined the optimal Co loading for Hg⁰ catalytic oxidation (7.5%). Yang et al. [81] prepared TiCe_0.25Sn_0.25O_x catalyst for synergetic removal of NO and Hg⁰ from flue gas. Through experiments they found TiCe_0.25Sn_0.25Ox has excellent NO and Hg⁰ removal performance, and also has satisfactory SO₂ tolerance, and revealed the synergistic removal mechanism and the effects of flue gas components on synergetic removal of NO and Hg⁰, as shown in Fig. 3a. Zhang et al. [82] loaded Co-Ce oxide on rod-like MnO₂ to simultaneously remove NO and Hg. They found that the Co-Ce/MnO₂ catalyst has excellent SCR activity and Hg capture efficiency. In addition, the Co-Ce/MnO₂ catalyst has a high tolerance to SO₂. Finally, the excellent activity and high sulfur resistance of the catalyst are explained through characterization techniques, as shown in Fig. 3b. Zhang et al. [83] researched Ce-Mo/TiO₂ catalyst to simultaneously remove NO and Hg. Ce-Mo/TiO₂-0.3 catalyst exhibited an excellent Hg⁰ oxidation and NO reduction efficiencies of over 90% at above 250 ℃. In addition, Ce-Mo/TiO₂-0.3 catalyst also presented an excellent resistance to SO₂ and H₂O poisoning and a good stability and recyclability. The evaluation data of the catalytic activity of Ce-Mo/TiO₂ is shown in Fig. 3c. Liu et al. [84] found that loading Nb on MnTiO_x can significantly improve the conversion of NO and the oxidation of Hg by the catalyst. The removal efficiency of different catalysts for NO and Hg are shown in Fig. 3d. In addition to TiO₂, MnO₂ is also used as a carrier for simultaneous removal of NO and Hg⁰ from flue gas.

As a stable crystalline powder, Al₂O₃ has good thermal stability, acid and alkali corrosion resistance, easy to form a fixed shape, and the surface area is less than 200 m²/g. These ideal properties have attracted extensive attention in the field of simultaneous removal of SO₂, NO_x, and Hg⁰. Wang et al. [47] prepared Cu and Fe-loaded Al₂O₃ catalysts for simultaneous removal of NO and Hg⁰. The characterization results show that Al₂O₃ as a support significantly increases the surface area and pore volume of the catalyst, improves the ability of the catalyst to adsorb gaseous reactants and therefore improves their catalytic efficiency. Moreover, Al₂O₃ increased the absorbed oxygen amount, indicating an increased number of active oxidative sites and a stronger oxidation ability for Hg⁰. Yue et al. [85] prepared a series of Cu_aCe_bZr_cO₃/ γ-Al₂O₃ catalysts by impregnation method to simultaneously remove NO_x and Hg⁰. Their research results show that 15% Cu_1.4Ce_0.55Zr_0.25O₃/γ-Al₂O₃ has the highest removal efficiency of NO (93%) and Hg⁰ (98%).

Silica based catalysts: Mesoporous silica has a large specific surface area, unique open-framework, a large pore volume and a high intrinsic surface activity, providing many adsorption sites for NO, SO₂ and Hg⁰. Chiu et al. [86] used the impregnation method to load V, Mn and Cu oxides onto SiO₂ to remove NO, SO₂ and Hg. The results show that the CuO_x/SiO₂ catalyst exhibits the greatest Hg⁰ and NO removal performance. In addition, he also studied the removal of Hg⁰ and NO by MnO_x impregnated V₂O₅-WO₃/TiO₂-SiO₂ catalyst [87]. The catalysts loaded with 10% MnO_x showed the best activity for the removal of Hg⁰ and NO. Liu et al. [49] loaded MnO_x-CeO_x on SiO₂ to remove NO and Hg. Through the activity evaluation of different loadings of MnO_x(a%)–CeO_x(b%)/SiO₂, it is found that 20%MnO_x-10%CeO_x/SiO₂ has the best catalytic activity, and its mercury removal efficiency is 96% at 100−300 ℃. Secondly, they also found that the introduction of CeO_x can broaden the SCR activity temperature range of MnO_x/SiO₂ catalysts, as shown in Figs. 4a and b. Gu et al. [88] studied the catalytic NO removal efficiency of MnFeO_x/SiO₂ catalysts under different Fe/Si conditions. They found that adding an appropriate amount of FeO_x can significantly widen the temperature window of the catalyst. Li et al. [89] prepared a TiO₂-SiO₂ photocatalyst for the catalytic oxidation of Hg. Studies have shown that the catalyst has excellent catalytic oxidation activity for Hg, and the Hg removal mechanism is summarized, as shown in Fig. 4c. Chen et al. [90] used the recycled Cu modified rice-husk-derived SiO₂ to remove NO and Hg. Studies have shown that 50% Cu/SiO₂ has the greatest Hg removal efficiency (88.2%), and 50%Cu-10%Ce/SiO₂ has 85% NO conversion at 100−300 ℃. Lin et al. [91] synthesized Cu-Mn-Ce oxide-incorporated Mesoporous Silica by hydrothermal exfoliation of silicate to remove NO and Hg. The research results show that Cu₂Mn₈ exhibits the highest Hg removal efficiency, and Ce doping can improve the NO removal efficiency.

3.1.3. Catalysts based on new supports

Molecular sieves-based catalysts: Zeolite-based materials have large specific surface area, good thermal stability, unique framework structure, excellent mechanical strength and adsorption, so they are considered as ideal carriers. Zeolite molecular sieves can chemically adsorb SO₂, Hg and NO on the surface. The metal elements loaded on the surface of the zeolite molecular sieve and embedded in its pores can promote the oxidation or reduction of SO₂, Hg and NO. Cao et al. [53] loaded Fe₃O₄, silver and V₂O₅ on HZSM-5 by dipping. The experimental results show that at 150 ℃, the removal rate of Hg by the catalyst is 97%, the removal rate of Hg⁰ and NO is 80%. Finally, the mechanism of simultaneous removal of NO and Hg by the catalyst is explained, as shown in Fig. 5a. Yuan et al. [92] prepared nano-iron loaded ZSM-5 and combined vaporized H₂O₂ and Na₂S₂O₈, and Ca(OH)₂ solution to simultaneously remove SO₂, NO and Hg⁰. Their research results show that when the mass of the catalyst is 0.3 g, the removal rate of SO₂, NO and Hg⁰ reaches the highest, respectively 99.9%, 96.5% and 90.1%. Fan et al. [93] prepared Cu/HZSM-5 to simultaneously remove Hg⁰ and NO. The experimental results show that the loading of copper significantly improves the activity of the catalyst, and the catalyst with a loading of 6% shows the best activity. Zhang et al. [67] used ZSM-5 modified with Mn-Fe to remove Hg at low temperatures and found that the introduction of Mn/Fe can effectively improve the mercury removal efficiency of ZSM-5. The catalyst with a Mn/Fe molar ratio of 1:2 has a 100% removal rate for Hg at 200−250 ℃. Applying an immersion method, Chui et al. [51] used the CuCl₂/MCM prepared by the impregnation method is used for the simultaneous removal of SO₂ and NO. The experimental results show that the loading of CuCl₂ reduces the removal efficiency of SO₂ by 4%, but increases the removal efficiency of NO by 11%, as shown in Fig. 5b.

Silico–aluminum–phosphate (SAPO) is a microporous crystal composed of three kinds of tetrahedral units, namely, PO₄⁺, AlO₄⁻ and SiO₂. Its pore structure can be controlled by changing different synthesis conditions and silicon content [94, 95]. SAPO are widely used as adsorbents, catalysts, and catalyst carriers. It is a new type of catalytic material with excellent shape selectivity, thermal stability and damp-heat stability. Because the framework is negatively charged, it has exchangeable cations and also exhibits proton acidity. SAPO can adjust the amount of acid by introducing various impurity atoms, so it can exhibit a catalytic performance ranging from medium strong acid to strong acid. It is well known that copper additives can improve the activity of SAPO-34, and copper-infused SAPO is very promising in industrial applications. Fickel et al. [94] showed that the NO conversion efficiency of Cu-infused SAPO approaches 100% at 200–400 ℃. Wang et al. [50] found that the content of Si and Al affects the content of Cu²⁺ species on the Cu/SAPO-34 catalyst. In addition, the content of Si can affect the content of active sites (Si or Al) and the number of acid sites of the catalyst. To affect the SCR activity of the catalyst. Pang et al. [52] prepared Mn-exchanged SAPO-34 catalysts for low-temperature SCR of NO with NH₃. Among their samples, 4Mn/SAPO-34 achieved the highest SCR activity in the 120–210 ℃ temperature range. The characteristic results suggested that species related to Mn³⁺ and Mn⁴⁺ were the active sites of the catalyst.

MOFs-based catalysts: MOFs are intriguing porous materials with the advantage of adjustable pore size that are self-assembled by organic ligands and metal ions [96]. The adsorption performance of MOFs is boosted by sufficient porosity, large surface area, and versatile structure. In addition, MOFs promote the dispersion of active components, providing a transportation channel for catalytic reactants. The metal centers in MOFs are evenly distributed and highly dispersed, concentrating their activity in certain catalytic reactions. The advantages of MOF carriers have been exploited in the removal of SO_2, NO and Hg⁰ from flue gas.

Wang et al. [97] prepared a MOF called MIL-100(Fe) and used it in the SCR reaction. Experimental results show that the NOx conversion rate of MIL-100(Fe) exceeds that of conventional V₂O₅-WO₃/TiO₂ catalysts below 300 ℃. Zhang et al. [54] investigated the MnO_x-loaded MIL-100(Fe) was prepared by the impregnation method. The experimental results show that the catalyst supported by MIL-100(Fe) has higher Hg⁰ adsorption and oxidation capacity than the catalyst supported by Fe₂O₃ at all test temperatures. The Hg⁰ removal efficiency was 77.4% at 250 ℃. In addition, the catalyst can be easily recovered from fly ash using magnets. The catalyst activity evaluation diagram and the removal mechanism diagram are shown in Fig. 6a. Zhang et al. [28] prepared Mn-Ce supported MOF (MnCe@MOF) catalyst by impregnation method. When removing Hg⁰ and NO from the flue gas, the loading of Mn and Ce significantly enhanced the activity of the catalyst. Characteristic analysis shows that the MOF scaffold has a developed porous structure and a high surface area. This makes the active phase highly dispersed and thus enhances the activity of the catalyst. The sulfur and water resistance experiments show that MnCe@MOF has high resistance to SO₂ and H₂O. Fig. 6b shows the removal efficiency of Hg⁰ and NO on the MnCe@MOF catalyst and the morphology of the catalyst. The stronger adsorption of Hg⁰ and NH₃ on MOF catalysts than on conventional materials might be explained by the special structure and Lewis acidity of MOF [98, 99].

PILCs-based catalysts: PILCs is a new type of catalytic material with a two-dimensional pore-like molecular sieve structure. Schematic representation of the clay structure is shown in Fig. 7a. It has the characteristics of adjustable pore structure, high porosity, high thermal stability, hydrothermal stability and exchangeable cations, so it is widely used to make catalysts and adsorbents [55, 100-103]. Chae et al. [55] synthesized a novel SCR catalyst with a Ti–PILC support, which demonstrated higher performance than conventional catalysts. Shen et al. [24] innovatively loaded MnO_x in PILCs for low-temperature NH₃–SCR denitration research. At the optimal loading of MnO_x/Ti–PILC (10%) and 80 ℃, the NO conversion efficiency approached 100%, with good H₂O resistance. Next, Shen et al. [104] systematically studied the low-temperature denitration activity of doped Mn–MO_x/Ti–PILC. They found that Ce and La additives promoted the denitrification activity [104]. He et al. [105] prepared a series of innovative Ce-Mn/Ti columnar clay (Ce-Mn/Ti-PILC) catalysts by the impregnation method for trapping elemental mercury (Hg⁰) at 100−350 ℃. The experimental results show that: in the temperature range of the study, the 6%Ce-6%MnO_x/Ti-PILC catalyst performance is greater than 90% Hg⁰ removal rate. The activity evaluation data of the catalyst and the Hg removal mechanism are shown in Fig. 7b. It is found through experiments that in the first stage of the reaction, the main Hg⁰ capture mechanism of the catalyst is adsorption. As the reaction progresses, the oxidizing ability of Hg⁰ is greatly improved, which is contributed by the hydroxyl oxygen and lattice oxygen on the catalyst surface. Wang et al. [25] prepared a series of Mn-Ce/Ti-PILCs catalysts for simultaneous removal of NO and elemental mercury in simulated flue gas. The experimental results show that 6%Mn-6%Ce/Ti-PILCs exhibit excellent NO conversion rate (> 95%) and Hg⁰ removal efficiency (> 90%) at 250 ℃. The characterization results show that the removal activity of NO and Hg⁰ is strongly affected by the conversion between oxygen adsorbed on the catalyst surface and lattice oxygen. The ratio of Mn⁴⁺/Mn³⁺ and Ce⁴⁺/Ce³⁺ on the surface of the catalyst affects the conversion rate of NO and the removal of Hg⁰. The modified PILC behaves like a catalyst support. The NO and Hg removal efficiencies from 100 ℃ to 300 ℃ are shown in Fig. 7c.

Catalyst with the same active component(s) shows very different activities in the simultaneous removal of SO₂, NO, and Hg⁰ catalysts. Table 1 [24, 25, 28, 47, 49, 54, 64, 65, 69, 75, 80, 82, 106-111] lists the NO and Hg⁰ removal efficiencies and SO₂ resistances of carriers with the same active components. Comparing the active metallic components among the carriers, we find that catalysts with Mn–Ce active sites are more resistant to SO₂ than catalysts with other metallic active sites. Therefore, catalysts with an Mn–Ce active component are potentially valuable in industrial applications. Furthermore, Mn–Ce@MOF is more SO₂-resistant than Mn–Ce/SiO₂, possibly because the MOF specific surface area is higher in the former than the latter. Finally, we find that new supported catalysts are more SO₂-resistant than catalysts based on conventional supports and unsupported catalyst.

3.2. Influence of gas composition on removal effect 3.2.1. Single component effect

Real industrial exhaust gas contains species such as O₂, NO, H₂O, NH₃, HCl, and fly ash, whose concentrations depend on the industrial type and operational conditions. Because such variable gas concentrations remarkably affect the NO reduction and Hg⁰ oxidation of catalysts, the impact of single flue-gas components on SO₂, NO, and Hg⁰ removal has attracted much interest.

Effect of O₂ : O₂ is an important component in the flue gas and plays a decisive role in most catalytic reactions. [112, 113].Fig. 8a summarizes the effects of O₂ on SO₂, NO, and Hg⁰. In the absence of O₂, Hg can react with the chemisorbed and lattice oxygens on the catalyst surface. As the reaction proceeds, the limited oxygen on the catalysts is gradually depleted, and the mercury removal efficiency is low. When O₂ is plentiful, the Hg⁰ removal rate increases because the depleted lattice and chemisorbed oxygens are replaced with gas-phase O₂ [114, 115]. Increasing the O₂ concentration does not significantly affect the Hg⁰ removal efficiency. However, However, a large increase in O₂ concentration will increase the rate of SO₂ oxidation to SO₃. The excessive generation of SO₃ will not be conducive to the progress of the SCR reaction, because it reacts with NH₃ and H₂O to produce ammonium sulfate and viscous ammonium bisulfate, gradually occupying the active sites of the catalyst and deactivating the catalyst.

Effect of NH₃:NH₃ is one of the common reducing agents in industrial applications. The NH₃ content in the flue gas has a great influence on the reduction of NO and the oxidation of Hg⁰. Increasing the NH₃/NO ratio can promote the reduction of NO in the NH₃ SCR reaction, but the increase in NH₃ content will cause it to compete with Hg⁰ for adsorption on the catalyst, thereby inhibiting the oxidation of Hg⁰ [26, 113, 116]. Wang et al. testified that NH₃ inhibits Hg⁰ oxidation over a CuO-MnO₂-Fe₂O₃/ γ-Al₂O₃ catalyst [117]. However, other studies have reported a significantly positive effect of NH₃ on Hg⁰ oxidation [118]. Li et al. [119] found that NH₃ consumes surface oxygen and limits the Hg⁰ adsorption, thereby inhibiting Hg⁰ oxidation over the catalyst. However, when the NH₃ supply is removed and O₂ is present, the inhibited Hg⁰ oxidation activity is completely recovered. Therefore, adding O₂ at an appropriate concentration to a mixture of NO, SO₂ and Hg will preserve (to some extent) the Hg⁰ oxidation by NH₃. The mechanism by which NH₃ affects the SO₂, NO, and Hg⁰ conversion is schematized in Fig. 8a.

Effect of H₂O: Water vapor is a main component in flue gases, and often leads to catalyst deactivation.Fig. 8a shows the effects of H₂O on SO₂, NO, and Hg⁰. Water vapor decreases the Hg⁰ oxidation and NO reduction over the supported catalyst [120], because it strongly competes with the reactive species for adsorption on the active sites of the catalyst [121]. When H₂O is removed from the simulated system, the efficiencies of NO reduction and Hg⁰ oxidation are restored to some level slightly below the initial value.

Effect of HCl: HCl is regarded as the flue-gas component that most strongly affects the Hg⁰ oxidation, because the main oxidized mercury species in coal-combustion flue gas is HgCl₂ [122].Fig. 8a summarizes the effect of HCl on SO₂, NO, and Hg⁰. HCl is a critical oxidant for Hg⁰ oxidation, especially by metal-based catalysts [123]. Although the introduction of HCl can significantly promote the oxidation of Hg⁰, it cannot directly oxidize elemental mercury because it exists in a naturally reduced state [124]. Hg⁰ oxidation by HCl requires the assistance of oxygen. According to several studies [125, 126], In the presence of HCl, HCl and O₂ will produce intermediate Cl₂ and active chlorine Cl^* on the catalyst, making the catalyst oxidize Hg⁰ through the Chlor-Deacon process. [127-129]. Wang et al. found that when present at low concentrations (10 ppm), HCl barely affects the NO conversion. However, at high concentrations, HCl prohibits the NO conversion because the chloride ions react with the metal on the catalyst surface, forming volatile metal-Cl compounds. These volatiles destroy the active sites and consequently deactivate the SCR reaction [130].

3.2.2. Interactive effect of SO₂, NO, and Hg⁰

The above-mentioned flue-gas components are mixed with various pollutants (SO₂, NO, and Hg⁰), which might influence each other during the catalytic reaction. Therefore, to simultaneously remove multiple pollutants, many researchers have investigated the interactive effects of SO₂, NO, and Hg⁰ during the pollutant-removal process. Their results are summarized in Fig. 8b.

Interactive effect of SO₂ and NO: The effect of SO₂ on NO removal in the SCR process is contentious. Some researchers [131, 132] believe that SO₂ inhibits NO removal throughout the reaction system. Proponents of this view consider that SO₂ reacts with metal oxides and NH₃ in the presence of O₂ to form metal sulfates and ammonium sulfates (such as NH₄HSO₄ and (NH₄)₂SO₄). Especially, ammonium sulfates formed at low temperature on the catalyst surfaces occupy the active sites, lowering the SCR catalytic activity [133]. Other researchers [134, 135] considered that the sulfate formed by SO₂ and metal oxides can enhance the acidity of the catalyst and help improve the low-temperature SCR activity of the catalyst, and some studies have found that introducing SO₂ into the SCR system can increase the removal rate of NO. Thus far, the influence of NO on SO₂ removal during the simultaneous removal of NO and SO₂ has not been investigated [14, 136], possibly because the SO₂ reaction dominates over the NH₃ and O₂ reactions, meaning that NO does not affect the SO₂ removal.

Interactive effect of SO₂ and Hg⁰: The effect of SO₂ on Hg⁰ removal from the flue gas is complicated and inconclusive. Both promotional and inhibitory effects are reported in the literature [39, 137-139]. Li et al. [39] studied the effect of SO₂ on the oxidation of mercury on CeO₂-TiO₂ catalyst. Through experiments, it is found that in the absence of oxygen in the simulated flue gas, SO₂ greatly inhibits the oxidation of Hg⁰, but once O₂ is introduced, SO₂ can promote the oxidation of Hg⁰. Tao et al. [140] explained the anoxic experimental phenomenon as competition by Hg⁰ and SO₂ for similar active sites on the catalyst surface; therefore, when no O₂ exists in the flue gas, SO₂ inhibits the Hg⁰ conversion. When O₂ is present, the oxidation of SO₂ to SO₃ is mediated by the chemisorbed oxygen generated by the Ce³⁺-related charge imbalance on the catalyst. However, the effect of Hg on SO₂ removal appears to be unreported in the literature. We speculate that Hg affects SO₂ removal by the same process as SO₂ affects Hg removal.

Interactive effect of NO and Hg⁰: Wang et al. [15] studied the effect of NO on Hg⁰ removal over an MnCe/TP catalyst. When the NO concentration increases from 200 ppm to 500 ppm, the removal efficiency of Hg⁰ will slightly decrease. Wang believed that the Hg⁰ oxidation was inhibited by other nonactive species such as nitrite, which might cover the active sites. Tao et al. [140] similarly found that in the absence of oxygen, increasing the NO concentration did not substantially change the Hg⁰ removal efficiency. When O₂ was introduced to a flue gas containing NO, the Hg⁰ removal efficiency slightly increased and a large amount of oxidized mercury was observed in the gas flow downstream of the sample. Li et al. [39] reported the same effect of NO on Hg oxidation over a CeTi catalyst. NO alone exerted a slightly propulsive effect on Hg⁰ removal, especially at higher concentrations. The limited oxygen stored on the catalyst is consumed by both NO and Hg⁰ oxidation [141]. The propulsive phenomenon [69] might be explained by weak NO absorption to the sample surface. Some of these absorbed NO might react with the O₂ on the metal–oxide catalyst surface, generating species such as NO⁺ and NO₂ that can oxidize Hg⁰ [142].

Whether the Hg⁰ removal also affects NO removal is a pertinent question. Several researchers [25, 69, 112, 143] who studied the effect of Hg⁰ on NO removal reached a consensual conclusion: Hg⁰ exerts only a slight effect on the reduction of NO. The inhibitory effect has been attributed to two possible causes. First, Hg⁰ might compete with NO and NH₃ for adsorption to catalytic sites [69]. Second, the oxidation reaction of Hg⁰ might accumulate HgO on the sample surface. If the accumulated Hg⁰ covers some active sites, the NO removal efficiency will be slightly reduced [23, 106, 112, 144].

3.3. Removal mechanism

Because the supported catalyst carrier generally has a rich specific surface area and a good pore structure, the supported active component is evenly dispersed on the surface and through the pores of the carrier. A supported catalyst ensures high catalytic activity for flue-gas removal. Thus far, no complete SO₂, NO, and Hg⁰ removal mechanism has been proposed. The mechanism of removal of SO₂, NO and Hg⁰ under the catalyst action is summarized in Fig. 9. The pathways of SO₂, NO and Hg⁰ removal are demonstrated in Fig. 9 with equations.

4. Conclusions and prospects

This paper presented the modern supported catalysts used in the removal of SO₂, NO and Hg⁰ from flue gas. The advantages and disadvantages of the formation technologies, the effects of crucial factors, and the removal mechanisms of the supported catalysts were carefully discussed. The technologies of carbon-based material adsorption technologies, photocatalytic oxidation technologies, wet scrubbing technologies, and NTP technologies face several problems: Large energy consumption, large equipment footprint, and secondary pollution. Solid-catalytic technology, which allows higher utilization of apparatus, incurs lower operating costs, and requires less investment than multi-pollutant simultaneous removal technologies, has been widely investigated, and many supported catalysts have been researched in the past decade. MOF carriers for the low-temperature removal of NO and Hg⁰ provide much higher SO₂ resistance than traditional supported catalysts (carbon-based materials, metal oxides, and silica). In addition, the composition of real flue gas is very complicated and remarkably affects the NO reduction and Hg⁰ oxidation. The removal efficiencies of NO, SO₂, and Hg⁰ are enhanced in the presence of O₂. NH₃ inhibits Hg⁰ oxidation and NO reduction in the absence of O₂, but promotes these reactions in the presence of O₂. H₂O inhibits Hg⁰ oxidation and NO reduction by competitively adsorbing to the active sites. After exploring the surface chemical processes, this review clarified the mechanisms of SO₂, NO and Hg⁰ removal, and proposed a theory of catalytic multi-pollutant purification by new supported catalysts. Reviewing the past solid catalytic technology and the current technological development direction, we finally proposed the future research direction. At present, catalysts based on new supports have advantages in sulfur resistance and denitrification and mercury removal compared with catalysts based on conventional supports. Among the catalysts based on the new carrier, the new MOFs material as the carrier has higher sulfur resistance at low temperature than other traditional materials such as carbon materials and metal oxides. This is because its regular pore structure and large surface area make the supported metal oxides highly dispersed. Compared with other materials, its sulfur resistance is poor at high temperature. This is because the thermal stability of MOFs at high temperatures is poor. The molecular sieve material has a synergistic removal effect of thimerosal compared to other materials at high temperature. In summary, the high-temperature activity of MOFs-based supported catalysts and the low-temperature activity of molecular sieve-based supported catalysts are worthy of in-depth study in the gas purification technology for removing SO₂, NO_x and Hg⁰ from industrial exhaust gas.

Declaration of competing interest

The authors declare no competing financial interest.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Nos. 52000093, 51968034, 41807373 and 21667015), National Key R & D Program of China (No. 2018YFC0213400), China Postdoctoral Science Foundation (Nos. 2020T130271, 2019M663911XB) and Open Fund of National Engineering Laboratory for Mobile Source Emission Control Technology (No. NELMS2019B03).