Low-Temperature Selective Catalytic Reduction of NO x with NH3 over Mn-Ce Composites Synthesized by Polymer-Assisted Deposition.

The Mn x Ce y binary catalysts with a three-dimensional network structure were successfully prepared via a polymer-assisted deposition method using ethylenediaminetetraacetic acid and polyethyleneimine as complexing agents. The developed pore structure could facilitate the gas diffusion and accelerate the catalytic reaction for NH3 selective catalytic reduction (SCR). Moreover, the addition of Ce is beneficial for the exposure of active sites on the catalyst surface and increases the adsorption of the NH3 and NO species. Therefore, the Mn1Ce1 catalyst exhibits the best catalytic activity for NO x removal with a conversion rate of 97% at 180 °C, superior water resistance, and favorable stability. The SCR reaction over the Mn1Ce1 catalyst takes place through the E-R pathway, which is confirmed by the in situ diffuse reflectance Fourier transform analysis. This work explores a new strategy to fabricate multimetal catalysts and optimize the structure of catalysts.

Introduction

Nitrogen oxides (NO, standing for NO and NO2) are among the most harmful air pollutants which can lead to the formation of acid rain, photochemical fumes, ozone depletion, and climate change.[1−3] Governments and related institutions have paid considerable attention to reducing the emission of NO from vehicle exhaust and electric power plants.[4,5] NH3 selective catalytic reduction (NH3-SCR) of NO has been considered as an advanced technology for the elimination of nitrogen oxides from coal-fired power plants owing to the low cost and high efficiency.[6,7] At present, in the field of fixed source denitrification, the V2O5/TiO2 catalyst has attracted extensive interest due to its good selectivity, high catalytic activity at 260–450 °C, stable properties, anti-poisoning ability, and strong anti-aging performance.[8−13] Despite these considerable advantages, there are still some problems including narrow window of operation temperature, biological toxicity, and increasing energy consumption, which have impeded the practical application of the V2O5–WO3 (MoO3)/TiO2 catalyst system.[14−16]

In recent decades, researchers have investigated varieties of low-temperature denitrification catalysts, including transition metal oxide catalysts, precious metal catalysts, and ion-exchange molecular sieve catalysts.[17−19] Typically, the low-cost transition metal catalyst shows good SCR activity and high selectivity in the low-temperature range, such as Mn/TiO2, Cu/TiO2, Co/TiO2, Ni/TiO2, Ce/TiO2, FeTiO, V2O5/WO3–TiO2, Ce–Fe–Ti, Ce–W–Ti, Fe–V–Ti, and WO3/TiO2.[20−22] Among them, manganese-containing composite catalysts have emerged as a promising candidate owing to their excellent activity for SCR reaction.[23] Particularly, the MnO–CeO2 composite has been reported to be active at the low-temperature SCR of NO (100–200 °C).[24] Qi et al. synthesized MnO–CeO2 composite oxides by a co-precipitation method, which showed high NO conversion at low temperature.[25] Tang et al. found that a MnO–CeO2/AC/C composite exhibited 78% NO conversion at 100 °C.[26] Li et al. reported that after doping CeO2 with MnO, MnO–CeO can enhance the de–NO conversion, H2O resistance, and stability.[27] The morphology and structure of the MnO–CeO2 composite are of great significance in the low-temperature SCR process, which has a crucial effect on the dispersion and oxidation state of manganese.[28−30] Guo et al. prepared a CeO@MnO core–shell structure catalyst by using a citric acid method, which was of high intrinsic catalytic activity derived from the core–shell structure.[31] However, the complicated multistep reaction processes and the introduction of templates make it difficult to be practically used on a large scale.

It has been demonstrated that the synthesis methods also have a critical impact on the physicochemical properties and catalytic performance of redox catalysts.[32−34] Recently, polymer-assisted deposition, an aqueous chemical route, has been developed to prepare multicomponent metal oxides using polyethyleneimine (PEI) and ethylenediaminetetraacetic acid (EDTA) as complexing agents.[35] In the polymer-assisted deposition process, metal ions bound to polymers are used as reaction precursors. The formation of a covalent complex between the nitrogen atoms and the metal precursors of multidentate polymers makes it possible to prepare most metal precursor solutions, which is not available to obtain through commonly used chemical solution deposition techniques. Multicomponent metal oxide networks could be obtained, with oxide nanoparticles uniformly dispersed in the framework. This low-cost chemical solution approach offers a new opportunity for the development of binary metal oxides. However, there are no reports on using the polymer-assisted deposition method to synthesize MnCe binary catalysts for selective catalytic reduction of NO.

In this work, a series of MnCe binary catalysts with nanostructures are prepared through the polymer-assisted deposition method. The polymers could not only form an initial network but also inhibit the aggregation of metal oxide particles, enabling the generation of a uniform solution. MnCe binary catalysts with a three-dimensional (3D) network composed of uniformly dispersed metal oxide nanoparticles could be obtained after heat treatment. The well-developed mesopore and macropore structure could facilitate the diffusion of reaction gas and exposure of active sites of the catalyst. MnO nanoparticles with uniform size are well dispersed in the framework, which serve as the catalytic oxidation center, while the formation of cerium-based solid solution could provide more active sites for catalytic oxidation. In addition, with excellent oxygen storage capacity, CeO2 could function as an oxygen reservoir and oxygen supplier during the catalytic process. Therefore, the MnCe binary catalysts demonstrate high catalytic activity for low-temperature catalytic reduction of NO with NH3. This is the first attempt to adopt the polymer-assisted deposition route for the fabrication of MnCe binary oxides as catalysts in the catalytic reduction of NO. This simple chemical solution strategy opens up a new way for the design and construction of multicomponent metal oxides used for low-temperature SCR of NO with NH3.

Experimental Section

Catalyst Preparation

Binary oxides MnCe (x = 2, 1.5, 1, and 0.5, y = 1) were prepared by polymer-assisted deposition route by using EDTA and PEI as binding ligands.[36] In a typical synthesis of Mn1Ce1, 1.07 g of Mn(NO3)2 (50 wt %, Sinopharm), 1.30 g of Ce(NO3)2·6H2O (99%, Macklin), 3.00 g of EDTA (99%, Acros Organics), and 3.00 g of PEI (Mw-800, Aldrich) were dissolved in 60 mL of deionized water under vigorous stirring to obtain a transparent and clear precursor solution. Then, the above solution was placed in an 80 °C oven until water was evaporated to form a transparent solid. The obtained dry gel was calcined in the air with a ramping speed of 1 °C/min and then maintained at 500 °C for 4 h. Finally, the catalyst can be obtained after cooling naturally to room temperature. The MnO catalyst can be produced when no cerium source is added. The amount of Ce(NO3)2·6H2O, EDTA, and PEI was fixed at 1.30, 3.00, and 3.00 g, respectively, while the amount of Mn(NO3)2 was controlled at 2.15, 1.60, 1.07, and 0.85 g, corresponding to the Mn/Ce molar ratio of 2, 1.5, 1, and 0.5. The as-prepared binary oxide catalyst was marked as MnCe (x = 2, 1.5, 1, and 0.5, y = 1). For comparison, the Mn1Ce1-co-precipitation method (CP) composite was prepared by a traditional CP.[37]

Catalyst Characterization

The morphologies and microstructures of the as-prepared MnCe were examined by scanning electron microscopy (SEM, Gemini SEM 500), transmission electron microscopy (TEM, JEOL, JEM-2100), and field-emission TEM. The energy-dispersive X-ray spectrometry (EDS) mappings of the selected region were acquired on an energy-dispersive X-ray spectrometer attached to the SEM instrument (Nova Nano SEM 450).

The X-ray diffraction (XRD) patterns were recorded on a rotating anode X-ray powder diffractometer operated on a RigakuD/MAX 2550 diffractometer with Cu Kα radiation. Intensity data was collected over a 2 range of 10–80° with a step size of 0.05° and an accumulation time of 1 s per step.

The Raman spectra were recorded on a laser micro-Raman spectrometer (≤I0.2/cm/invia reflex) equipped with laser supplying the excitation line at 532 nm. The spectrometer was calibrated using the silicon line at 521 cm–1.

The nitrogen adsorption–desorption isotherms were conducted at 77 K on a QUADRASORB SI analyzer. The samples were first degassed under vacuum at 120 °C for 10 h before measurement. The specific surface areas and average pore diameters were determined using the Brunauer–Emmett–Teller (BET) equation and the Barrett–Joyner–Halenda (BJH) model from the desorption branch, respectively.

Hydrogen temperature-programmed reduction (H2-TPR) and NH3 temperature-programmed desorption (NH3-TPD) were performed on a Chem BET Pulsar (Automated Chemisorption analyzer) equipped with a TCD detector. For the H2-TPR, the sample was first pretreated in He (120 mL/min) at 200 °C for 40 min and then cooled to room temperature. Subsequently, the sample was heated up to 800 °C at a rate of 10 °C/min under H2 (120 mL/min). For NH3-TPD, 50 mg of sample powder was weighed and pretreated at 200 °C for 30 min in a flow of He (120 mL/min). After cooling to room temperature, the adsorption gas was saturated with NH3 for 60 min. After saturation, the sample was rinsed in pure He flow for 30 min. Finally, the sample was heated to 800 °C at a heating rate of 5 °C/min.

The surface atomic states of the catalysts were determined by X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi) using Al Kα radiation (hν = 1486.6 eV). The binding energies were calibrated using the C 1s peak at 284.6 eV as a reference and the peak deconvolution was carried out using the XPS Speak software.

In situ diffuse reflectance Fourier transform spectroscopy (DRIFTS) measurements were conducted on a Fourier transform infrared spectrometer (PerkinElmer, Model Spectrum 100) and an in situ cell (modified Harricks Model HV-DR2) with a mercury cadmium telluride detector cooled using liquid N2. The gas flow was controlled at 100 mL/min by the mass flow meter. Before the gas adsorption experiment, the sample was pretreated in nitrogen at 300 °C for 1 h to blow away the impurities. The corresponding background spectra were recorded and subjected to a subtraction process from the sample spectrum. The experimental atmosphere for in situ DRIFTS study was as follows: 500 ppm NH3/N2 or/and 500 ppm NO/N2 + 5 vol % O2/N2.

Catalytic Activity Test

The SCR catalytic performance was carried out in a fixed-bed reactor with an internal diameter of 10 mm. The temperature of the fixed-bed column was controlled by an electric heating furnace. The catalyst of 100 mg was sandwiched between the quartz cotton in the quartz tube. The simulated fumes were introduced into the reactor. The flow of simulated flue gas was controlled by a mass flow meter, which consisted of 500 ppm NO, 5 vol % O2, 500 ppm NH3, and N2 balance gas. In all reaction activity evaluation experiments, the total gas flow was 500 mL/min, with a gas hourly space velocity (GHSV) of 38,000 h–1, and the reaction temperature was controlled at 60–300 °C. The stationary state NO conversion and N2 selectivity of the sample can be calculated according to the following formulas

In the formula, NO, NO, NO2,out, N2Oout, and NH3,in indicated the inlet and outlet concentrations at steady state, respectively.

Results and Discussion

Material Characterization

The MnCe binary networks were synthesized by the polymer-assisted deposition method (Figure ). The dissolved polymers could serve not only as the initial network during the reaction process but also as a binding agent to bond with the metal ions, facilitating homogeneous dispersion of the metal oxide. After drying and calcination, EDTA and PEI decomposed to produce gases which escape from the system, remaining a bimetal oxide interconnected network with a porous structure originated from the organic polymer structure.

Figure 1

Schematic of the synthesis of the bimetal oxide network by the polymer-assisted deposition approach.

The morphology and structure of the mixed bimetal oxide networks were revealed by SEM and EDS mappings. As can be seen from Figure , the four samples are all composed of 3D porous structures formed by crosslinking of metal oxide nanoparticles. The coordination of metal ions with the polymer could enable the uniform distribution of metal ions and prevent their aggregation, resulting in a small particle size. Moreover, the particle size decreases from 25 to 3 nm with the decreasing Mn/Ce molar ratio. Compared with Mn2Ce1, Mn1Ce1 shows a tighter network structure, suggesting higher specific surface area and more exposed active sites. As the Mn/Ce molar ratio decreases to 0.5, excess cerium causes deformation and collapse of the 3D network due to the agglomeration of cerium oxides.

Figure 2

SEM images and EDS mappings of MnCe binary catalysts: (a,e) SEM images of MnO; (b,f) Mn2Ce1; (c,g) Mn1Ce1; (d,h) Mn0.5Ce1; and (i1–4) EDS mappings of Mn1Ce1.

The morphology of MnCe binary networks was further investigated by TEM as shown in Figure . The addition of Ce transforms the loose lattice interface of MnO into a dense structure, while the particle size of metal oxides is greatly reduced. Combined with SEM images, the particle size of MnO is about 25 nm, whereas the diameter of MnCe nanoparticles is as small as 3 nm. Figure e–h exhibits the high-resolution TEM (HRTEM) images of MnCe. The morphology of metal oxides in the catalyst can be inferred by measuring the lattice spacing. It is found that the lattice spacing of Mn in MnO is 0.270 nm, corresponding to the Mn2O3(222) plane, which indicates that the valence state of Mn in the MnO catalyst is Mn3+.[38] In the MnCe catalyst, the lattice spacing of 0.307 nm is related to the CeO2(111) surface.[39] In addition, the presence of 0.250 nm lattice spacing may be due to the replacement of Mn with a similar but smaller ion radius in the cubic fluorite structure, which causes decrease of the lattice spacing.[40]

Figure 3

TEM (a–d) and HRTEM (e–h) images of the mixed oxide networks: (a,e) MnO; (b,f) Mn2Ce1; (c,g) Mn1Ce1; and (d,h) Mn0.5Ce1.

The results obtained from the XRD analysis of MnCe catalysts are presented in Figure . A sharp characteristic peak of MnO is observed corresponding to α-Mn2O3 (JCPDS 31-825), indicating that manganese in the MnO catalyst is in the form of Mn2O3 which is consistent with the HRTEM result. The XRD analysis patterns for all the bimetal oxides with different proportions are basically the same. No strong or sharp peak of MnO is observed while only the broad diffraction peaks representing CeO2 can be seen, which suggests that a Mn–Ce solid solution is formed. It has been reported that when Mn/(Mn + Ce) < 0.5, the XRD analysis pattern only had the wide peak of CeO2, indicating the presence of Mn–Ce solid solution.[41] In addition, it has been reported that when a metal ion is successfully incorporated into the CeO2 lattice, the XRD peak corresponding to the oxide of that metal is not observed, which also resembles the above result.[42]

Figure 4

XRD patterns of the MnCe binary catalysts.

The Raman spectra of the samples are shown in Figure S1. The bands at 338–360 and 632–649 cm–1 can be associated with the out-of-plane bending mode of α-Mn2O3 nanocrystals.[43] As for MnCe, the peaks centered on 446, 458, and 460 cm–1, which belong to the triple degenerated F2g active mode, are typical fluorine-like cerium structures.[44] However, it is worth mentioning that the CeO2 and Mn2O3 vibration peaks shift to lower intensity with the decrease of Mn/Ce molar ratio, and the peak at 300–400 cm–1 even disappears on Mn0.5Ce1. This indicates that manganese is integrated into the crystal lattice of CeO2 and the adding of cerium species inhibits the crystallization of Mn2O3 at the same time.[45,46]

The pore structures of the MnO and bimetal oxide networks were measured by N2 isothermal adsorption/desorption, and the resultant pore parameters are listed in Table . The N2 adsorption and desorption isotherms of the four samples in Figure S2 show the IUPAC type IV mode with obvious hysteresis loops, suggesting the existence of mesoporous structures.[47] For the raw MnO, the BET surface area is only 23.0 m2 g–1 with a pore volume of 0.08 cm3 g–1. As for MnCe binary networks, an increasing trend of pore structure with the decreasing Mn/Ce molar ratio is noted for the samples, owing to the formation of a highly dispersed Mn–Ce composite.[48] The surface area and pore volume of Mn1Ce1 increase to 113.2 m2 g–1 and 0.20 cm3 g–1, respectively. The average pore size calculated by the BJH method declines to 3.5 nm, which is attributed to the smaller particle size compared with MnO. The addition of Ce can reduce the size of the metal oxide particles and increase the specific surface area of the catalyst, which is beneficial to the exposure of active sites on the catalyst surface. When further decreasing the Mn/Ce molar ratio, the BET surface area decreases slightly owing to the aggregation of metal oxide particles.

Table 1

Pore Parameters of the MnCe Binary Catalysts

sample	SBETa (m2 g–1)	Vtb (cm3 g–1)	Dpc (nm)
MnOx	23.0	0.08	6.1
Mn2Ce1	75.7	0.13	3.1
Mn1.5Ce1	87.7	0.11	3.1
Mn1Ce1	113.2	0.20	3.5
Mn0.5Ce1	107.4	0.18	3.5

Specific surface area.

Total pore volume.

Average pore diameter.

Previous studies have reported that the catalytic performance is related to the redox performance of the catalyst in the NH3-SCR process.[49] The reduction pattern of the MnCe binary metal catalysts was recorded by using H2-TPR. The corresponding TPR profiles are presented in Figure . The reduction peaks at 474 and 593 °C are observed in the MnO catalyst, which are related to the reduction of Mn2O3 to Mn3O4 and the reduction of Mn3O4 to MnO, respectively.[50,51] Compared with MnO, the reduction peaks of MnCe shift toward lower temperature range (370–550 °C). In the case of Mn2Ce1 and Mn1.5Ce1, only one reduction peak can be seen at 541 and 431 °C, respectively. However, with the decrease of Mn/Ce molar ratio, Mn1Ce1 and Mn0.5Ce1 present two overlapping peaks with temperature lower than 430 °C. It is difficult to associate the peak with the reduction of different MnO species or a specific reduction step of CeO2 because the reduction of Mn and Ce ion occurs simultaneously.[52,53] The reduction peak of the MnCe catalyst shifts to a lower temperature range compared to the MnO catalyst, indicating fairly easy reducibility.[54,55] This suggests that the Mn–Ce composite effectively enhances the storage and transfer of active oxygen and advances the catalytic cycle during the SCR process.[56] Moreover, the developed pore structure enhances the gas molecule diffusion inside the catalyst, which may be beneficial to the catalytic reaction.

Figure 5

H2-TPR patterns of MnCe binary catalysts.

The acid sites on the SCR catalyst surface determine the adsorption and activation ability of NH3 on the catalyst surface. The results of the NH3-TPD desorption study are displayed in Figure . The analytical peak at high temperatures belongs to Lewis acid sites, and the analytical peak below 200 °C belongs to Brønsted acid sites.[57] The curve of pure MnO only has an essential Lewis acid sites as shown in Figure . The peak intensity of MnCe catalysts is stronger than that of pure MnO, which may be ascribed to the higher surface area of these materials. It is interesting to note that the NH3-desorption peaks monotonically shift to high temperatures with a decrease in Mn/Ce molar ratio. However, although Mn0.5Ce1 exhibits the highest acid site temperature, its peak intensity is relatively smaller compared with that of other samples. This means the total amount of adsorbed NH3 for Mn0.5Ce1 is lower than that of others (Table ). The analytic peak temperature of Mn1Ce1 is higher than the ones of other MnCe samples, indicating that there is a larger number of acid sites. Higher NH3 adsorption capacity at Brønsted acid sites is beneficial to the NH3-SCR process at low temperature. Compared with Brønsted acid sites under similar conditions, Lewis acid centers have been revealed to maintain high NO conversion and low N2O formation.[58] Mn2Ce1 shows a new analytic peak at 487 °C. This suggests that different Mn/Ce ratios have an influence on the acid position distribution and concentration of the catalyst. Enhancing the acid site distribution of MnCe binary catalysts is conducive to improving the adsorption capacity of ammonia, which has significant benefits for improving de-NO ability.[59]

Figure 6

NH3-TPD patterns of MnCe binary catalysts.

Table 2

NH3 Adsorption Amounts of the MnCe Binary Catalysts

catalyst	NH3 adsorption amount (mmol/g)
MnOx	0.8
Mn2Ce1	1.8
Mn1.5Ce1	1.7
Mn1Ce1	2.2
Mn0.5Ce1	1.5

To determine the atomic concentration on the catalyst surface and further obtain the valence states of Mn and Ce in the MnCe binary metal oxide, XPS analysis was performed. High-resolution spectra of Mn 2p, Ce 3d, and O 1s are presented in Figure . The Ce 3d spectrum of the MnCe binary metal oxide catalyst is displayed in Figure a. By conducting peak-fitting deconvolution, the Ce 3d XPS spectrum could be separated into eight peaks including v (882.4 eV), v′ (885.3 eV), v″ (888.8 eV), v‴ (898.3 eV), u (901.1 eV), u′ (903.5 eV), u″ (906.7 eV), and u‴ (916.7 eV). It can be seen from the Ce 3d XPS spectrum that the peak “v” corresponds to the Ce 3d5/2 spin–orbit component, and the peak “u” is related to the Ce 3d3/2 spin–orbit component. Among them, six peaks of v, v″, v‴, u, u″, and u‴ (red curve) are consistent with the spectrum of Ce4+, indicating that the main morphology of Ce in MnCe binary oxide is CeO2, while v′ and u′ (blue curve) can be attributed to the surface Ce3+ atoms. Ce 3d XPS results reveal that Ce3+ and Ce4+ coexist in the manganese–cerium mixed system. In addition, the relative content of Ce3+ on the surface is evaluated and listed in Table . The superior ratio of Ce3+/Ce4+ has been proven to cause oxygen vacancies, charge imbalance, and unsaturated chemical bonds on the catalyst surface, which are conducive to the formation of chemically adsorbed oxygen on the catalyst surface.[60]

Figure 7

XPS spectra of MnCe binary catalysts (a) Ce 3d, (b) Mn 2p, and (c) O 1s.

Table 3

Concentrations of the Atoms Obtained by XPS Analysis

catalyst	Mn4+/Mn3+	Ce3+/Ce4+	Oα/(Oα + Oβ)
MnOx	1.28	/	0.29
Mn2Ce1	1.45	0.10	0.69
Mn1Ce1	1.57	0.11	0.71

The XPS spectra of Mn 2p are displayed in Figure b. The overlapping Mn 2p3/2 signal could be divided into three peaks, which correspond to Mn2+ (640.4 ± 0.2 eV), Mn3+ (641.3 ± 0.2 eV), and Mn4+ (642.1 ± 0.2 eV), respectively. The multiple valence states of Mn in the binary metal oxide catalyst make it easy to alter between valence states, promoting the transfer of electrons during the redox reaction and further boosting the catalytic oxidation performance of the catalyst. Previous work has reported that Mn4+ plays a dominant role in accelerating the oxidation of NO to NO2 in de-NO activity.[8] Thus, the relative percentage of Mn3+ and Mn4+ is calculated from the area of the corresponding characteristic peak and listed in Table . The surface Mn4+ content of Mn1Ce1 is higher than that of MnO and Mn2Ce1, owing to the electronic interactions between MnO and CeO2, resulting in an improved catalytic performance.

The peak-fitting O 1s XPS spectra for the investigated materials are presented in Figure c, which involve two categories: chemically adsorbed oxygen (or weakly chemically bonded oxygen, expressed as Oα, binding energy is located at 531.6–532.0 eV) and lattice oxygen (expressed as Oβ, binding energy is located at 529.4–529.7 eV). It has been proved from many literature studies that the chemically adsorbed oxygen on the surface possesses better catalytic activity than lattice oxygen owing to its higher mobility. Therefore, the presence of a high Oα/(Oα + Oβ) ratio is conducive to the oxidation of NO to NO2, which promotes the low-temperature activity through the “fast SCR” reaction. The relative atomic percentage of these oxygen species on the catalyst surface can be estimated from the relative areas of these sub-peaks, and the results are summarized in Table . It could be seen that the introduction of Ce significantly increases the concentration ratio of Oα, especially for Mn1Ce1 with a higher chemisorbed oxygen content of 71%.

Catalytic Performance

Figure a displays the SCR catalytic activity of catalysts with different Mn/Ce molar ratios. The pristine MnO shows low catalytic activity with the highest NO conversion of only 59%. Compared with MnO, doping of Ce remarkably increases the NO conversion within the whole temperature range owing to the synergistic effect of MnO and CeO2. It can be observed that the NO conversion on all catalysts increases first and then declines slightly with increasing temperature. The SCR activity falls in the following order: Mn1Ce1 > Mn2Ce1 > Mn1.5Ce1 > Mn0.5Ce1 > MnO. The Mn1Ce1 exhibits the highest catalytic activity of 97% at 180 °C. This phenomenon reveals that the addition of Ce effectively improves the microstructure and catalytic performance of the catalyst.[29] The significantly reduced metal oxide particle size and increased catalyst-specific surface area are conducive to the exposure of active sites on the catalyst surface, enhancing the adsorption and condensation of NH3 molecules in the gas phase. Finally, the contact between the reactant molecule and active sites are improved, which is beneficial to the catalytic performance of the catalyst. Nevertheless, the addition of excessive cerium leads to deformation of the 3D network and agglomeration of metal oxide particles, which causes active sites to be covered and the decreased catalytic activity. As a result, the Mn1Ce1 catalyst is selected as a representative for further research. The effect of calcination temperature on the catalytic activity is displayed in Figure S3, which reveals that the calcination temperature has a crucial effect on the SCR activity. The NO conversion on the MnCe mixed oxide catalyst is in the order Mn1Ce1-500 °C > Mn1Ce1-450 °C > Mn1Ce1-300 °C > Mn1Ce1-400 °C > Mn1Ce1-550 °C > Mn1Ce1-350 °C. The best calcined temperature is 500 °C. Meanwhile, Table S1 presents that the SCR performance of the Mn1Ce1 catalyst is comparable to those of other various metal-substituted catalysts reported in the literature.

Figure 8

NH3-SCR activity of (a) de-NO performance, (b) N2 selectivity of Mn1Ce1, (c) stability test of Mn1Ce1, (d)H2O resistance of Mn1Ce1, and (e) SO2 resistance of Mn1Ce1. Reaction condition: 500 ppm NH3, 500 ppm NO, 5 vol % O2, with N2 as balance gas, and GHSV 38,000 h–1.

Also, the result of N2 selectivity is tested and shown in Figure b. The Mn1Ce1 catalyst presents a relatively high N2 selectivity above 80% in the range of 60–220 °C, which slightly decreases at high temperature. The increase of intermediate product NH4NO3 in NH3-SCR reaction may be the main reason for this phenomenon. Both the higher NO conversion and N2 selectivity of Mn1Ce1 demonstrate its better de-NO performance. The catalytic stability of the Mn1Ce1 catalyst is displayed in Figure c. At 180 °C, the NO conversion of Mn1Ce1 can maintain above 95% after 70 h, indicating a good catalytic stability, which is a very important factor for practical application.

In the process of NH3-SCR reaction, the presence of water vapor affects the catalytic activity. Therefore, the effect of water vapor on Mn1Ce1 catalytic performance was explored. As shown in Figure d, the conversion of Mn1Ce1 catalyst decreases to 80% when 5% water is introduced at 180 °C, indicating that the presence of water vapor has a negative effect on the active sites of the sample, which could block active sites on the catalyst surface. After closing the steam inlet, the conversion of the catalyst quickly recovered to about 97%, which suggests that the effect of water on the NH3-SCR activity of the 3D binary metal oxide is reversible. In order to investigate the SO2 poisoning process during catalytic oxidation of NO, the Mn1Ce1 catalyst was tested, and the results are shown in Figure e. The de-NO ability remains at 97% for 60 min before SO2 gas is introduced. When SO2 is added to the simulated fumes, the catalytic activity decreases simultaneously with reaction time. After 60 min of injection, the SO2 gas is stopped, and NO conversion declines from 97 to 75%. The catalytic activity is restored to 82% after cutting off SO2 for 60 min, indicating that SO2 leads to slight deactivation of the catalyst for NH3-SCR. This phenomenon is possibly ascribed to the sulfation (e.g., formation of (NH4)2SO4 and NH4HSO4) and the accumulation of sulfates with active species (such as Ce2(SO4)3, Ce(SO4)2, and/or MnSO4), which induce the blocking of active sites or catalyst pores.[61−63]

In Situ DRIFTS

NH3 Adsorption

The chemical adsorption of NH3 on MnO and Mn1Ce1 catalysts at different temperatures was studied by in situ FTIR spectroscopy. Figure a shows the DRIFT spectra of NH3 on MnO at different temperatures. Five NH3 adsorption bands could be detected at 1572, 1525, 1434, 1346, and 1255 cm–1, respectively. These bands include the NH3 species adsorbed on the Lewis acid sites (1572, 1346, and 1255 cm–1), NH4+ species linked to Brønsted acid sites (1434 cm–1), and NH2 species (1525 cm–1).[64−66] The peak intensity gradually decreases with increasing temperature, suggesting that the adsorption of NH3 species is unstable. Similarly, different NH3 adsorption peaks are observed on the Mn1Ce1 catalyst surface in Figure b. The NH2 species appear at 1552 and 1525 cm–1.[67] The absorption peaks of NH4+ species at Brønsted acid sites appear at 1465 and 1383 cm–1, while the bands at 1327 and 1233 cm–1 are assigned to the coordinated NH3 on the Lewis acid sites. Based on the results obtained from NH3-TPD, NH3 can be adsorbed on different active sites, so Brønsted acid sites and Lewis acid sites coexist on the sample surface, which can be observed on both two samples. Moreover, it can be seen from Table that Mn1Ce1 generates more NH3 adsorption sites due to its larger surface. Therefore, stronger adsorption and activation ability of NH3 on the catalyst surface results in enhanced band intensities in Figure b. The typical peaks for Mn1Ce1 all increase with the increasing temperature, suggesting an enhanced reactivity of NH3 (Figure c).

Figure 9

NH3 adsorption DRIFT spectra over (a) MnO and (b,c) Mn1Ce1; NO + O2 adsorption DRIFT spectra over (d) MnO and (e,f) Mn1Ce1; and transient reaction of NO with pre-adsorbed NH3 over (g) MnO and (h,i) Mn1Ce1.

NO + O2 Adsorption

The steady-state adsorption spectra of NO + O2 for MnO and Mn1Ce1 catalysts show nearly the same variation in Figure d,e. For MnO, the peaks at 1346 and 1420 cm–1 belong to the species of cis-N2O22– and trans-N2O22–, respectively.[68,69] The peak at 1560–1585 cm–1 is attributed to the absorption peak of the bidentate nitrate species, while the peak detected at 1458 cm–1 is related to monodentate nitrite.[16] For Mn1Ce1, the peaks assigned to cis-N2O22– (1357 cm–1), trans-N2O22– (1428 cm–1), bidentate nitrate (1370, 1570, and 1582 cm–1) and monodentate nitrite (1542 and 1556 cm–1) become more intensive. Figure f indicates that the addition of CeO2 promotes the formation of various intermediate nitrate and nitrite species. More N2O22–, bidentate nitrate, and monodentate nitrite species on the surface of Mn1Ce1 may improve the NH3-SCR activity.

Reactions between NO + O2 and Adsorbed NH3 Species

The reactivity of NO + O2 species with pre-adsorbed NH3 was also studied on MnO and Mn1Ce1 catalysts by use of in situ DRIFTS at 180 °C, which was measured as a function of time (Figure g,h). As shown in Figure g, when the reaction gas is converted to NO + O2, the peak intensity of all NH3-adsorbed species on MnO does not change much. The bands at 1489, 1465, and 1413 cm–1 can be assigned to the NH4+ species adsorbed on Brønsted acid sites, while the band at 1359 cm–1 could be attributed to NH3 coordinated to Lewis acid sites. The band at 1266 cm–1 is detected for NH2 species. Two new bands attributed to nitrate products at 1214 and 1107 cm–1 are detected, suggesting that both NH4+ and coordinated NH3 on the MnO surface could react with gaseous NO.

For Mn1Ce1, after introducing NO + O2, the bands attributed to NH4+ species at 1490 cm–1 and ascribed to coordinated NH3 at 1572 and 1218 cm–1 decrease in intensity as displayed in Figure h. The band due to NH2 species at 1544 cm–1 increases first and then decreases as the reaction goes on. Moreover, several bands (1022, 1060, and 1503 cm–1) attributed to nitrate products appear. Based on Figure g,h, the amount of adsorbed NH3 and nitrate species on Mn1Ce1 surface are higher than that on MnO surface, suggesting that the addition of CeO2 to MnO could produce more coordinated NH3 and ionic NH4+ and enhance the low-temperature activity. As mentioned above, the adsorbed NH3 on both Lewis and Brønsted acid sites and gaseous NO species are the major characteristics in the transient SCR reaction, following the applicability of the Eley–Rideal (E–R) mechanism for it.[70]

Reactions between NH3 and Adsorbed NO + O2 Species

The reactivity of the pre-absorbed NO species on catalysts in the SCR catalytic reaction was further studied. For MnO, it can be seen from Figure a that after introduction of NH3, the peaks of monodentate nitrate at 1525 and 1454 cm–1, the cis-N2O22– species at 1362 cm–1, and the bidentate nitrate species at 1260 cm–1 are nearly unchanged, which indicates that the above-mentioned nitrate species hardly participate in the NH3-SCR reaction on MnO. Figure b displays the in situ DRIFT spectra for NH3 and adsorbed NO + O2 over Mn1Ce1 at 180 °C for different times. After the adsorption of NO + O2, the surface of Mn1Ce1 is mainly covered by trans-N2O22– (at 1430 cm–1), cis-N2O22– (at 1022 and 1358 cm–1), bidentate nitrate (at 1573 cm–1), and monodentate nitrite (at 1539 cm–1). After the further introduction of NH3, the bands of monodentate nitrite and cis-N2O22– decrease and disappear in 2 min. At the same time, the bands at 1059 and 1559 cm–1 attributed to nitrate products and coordinated NH3 at 1615 cm–1 appear after 2 min, indicating that monodentate nitrite and cis-N2O22– are reactive in the NH3-SCR reaction. On the contrary, the other pre-adsorbed NO species over the catalyst sample is still present after the introduction of NH3, suggesting the NH3-SCR reactivity over these species through the Langmuir–Hinshelwood (L–H) mechanism is low.[70]

Figure 10

Transient reaction of NH3 with pre-adsorbed NO + O2 over (a) MnO and (b) Mn1Ce1 and coadsorption/reaction of NH3, NO, and O2 over (c) MnO and (d) Mn1Ce1.

DRIFT Spectra in a Flow of NO + NH3 + O2

In order to identify the substances present in the catalyst and the reaction, the in situ DRIFT spectrum was recorded with a flow of NO + NH3 + O2 at a reaction temperature of 180 °C. As shown in Figure c, bands assigning to NH3 coordinated on Lewis acid (1352 cm–1), trans-N2O22– (1430 cm–1), and the nitrate species (1271 cm–1) from NO adsorption are all detected on the MnO surface. At the same time, bands for NH4+ ions formed due to the H2O production (1459 cm–1), bidentate nitrate (1573 cm–1), coordinated NH3 (1223 cm–1), and nitrate species (1308 and 1615 cm–1) also appear in Figure d. However, the peak intensity of Mn1Ce1 is more obvious, suggesting that the presence of Ce enhances both the adsorption and activation of NH3 and NO species on the catalyst. From the research above, the following results can be obtained. During the SCR reaction, the E–R mechanism plays a vital part in the SCR processes over catalysts; the addition of Ce brings in more acid sites, which is positive for the adsorption of NH3 species.[71] In addition, the interaction between Ce and Mn results in the enhanced surface area and the highly dispersed MnO active sites and thus improve the low-temperature activity.

As discussed above, all the ad-NO species in the transient SCR reaction only has a much lower reactivity, indicating that there is a blockage on the NH3-SCR reactions over MnCe catalysts through the L–H pathway. Meanwhile, the reaction intermediate NH2 is observed, demonstrating the occurrence of the E–R mechanism. The SCR reaction through the E–R mechanism can be described as[72]

The possible reaction pathway for selective catalytic reduction of NO with NH3 over the MnCe catalysts is proposed in Figure . First, the gas NH3 is adsorbed on Lewis acid sites and Brønsted acid sites in the form of NH4+ ions and coordinated NH3 on the catalyst surface (reaction ), whereas the reactant NO is still in the gas phase or form the weakly adsorbed state. Then, the adsorbed NH3 species could react with O2 easily to produce NH2 (reaction ). At last, the NH2 rapidly react with NO to generate N2 and H2O.

Figure 11

Proposed reaction mechanism of NH3-SCR of NO over a Mn1Ce1 catalyst surface.

Conclusions

In summary, a MnCe binary mixed oxide network with a 3D structure is prepared by using the polymer-assisted chemical solution method. Due to the effect of static electricity and hydrogen bonding, EDTA and the stable macromolecular chain of PEI form a “metal ion-EDTA-PEI” homogeneous precursor solution with an initial network structure. During the calcination, EDTA and PEI decompose to gaseous substances and escape, and the 3D network is retained at the same time. The network has a favorable mesoporous structure, which facilitates the diffusion of the reaction gas and the exposure of active sites of the catalyst. The co-doped cerium serves as an oxygen supply center for active oxygen transport. Thus, the binary mixed oxide network possesses rich active oxygen species and much more acid sites. The NO conversion of the Mn1Ce1 catalyst exhibits 97% at 180 °C and maintains above 95% after 70 h, which is comparable to some other various metal-substituted catalysts. The in situ DRIFTS studies reveal that the E–R route between gaseous NO and adsorbed NH3 is dominant in the catalytic reaction of MnCe binary catalysts. Therefore, the Mn1Ce1 catalyst prepared by the polymer-assisted deposition method for the first time presents broad application prospects in various catalytic reactions and probably becomes a potential alternative to NH3-SCR of NO under low-temperature conditions.