Selective Catalytic Reduction of NO over Cu–Mn/OMC Catalysts: Effect of Preparation Method

Ordered mesoporous carbon (OMC) was used as support for CuOx and MnOx, and the effects of preparation method on selective catalytic reduction (SCR) of NO with NH3 were investigated. The Cu–Mn/OMC prepared by solvent evaporation-induced self-assembly method, named as self-assembly synthesis (S), exhibited higher NO conversion and N2 selectivity than the catalyst prepared by ultrasound-assisted impregnation (I) or mechanical mixing (M). The structural and surface properties of catalysts were characterized by various techniques. XRD and TEM results showed good dispersion of active phases on Cu–Mn/OMC(S). XPS analysis suggested that the surface of Cu–Mn/OMC(S) had the maximum amount of O–C=O groups and chemisorbed O. The strongest acidity and largest amount of oxidative species were further illustrated by NH3-TPD and H2-TPR profiles, which were consistent with the XPS results. Accordingly, these favorable properties may be the main reasons for the outstanding performance of Cu–Mn/OMC(S) in NH3-SCR reaction. Thus, selfassembly synthesis can be considered an effective method for the preparation of OMC–supported catalysts.


INTRODUCTION
Anthropogenic emission of NO x has raised wide concern in recent years for its contribution to environmental pollution, including photochemical smog, acid rain, ozone depletion, and haze formation (Kang et al., 2007).Selective catalytic reduction (SCR) with NH 3 has been considered the most efficient technology for controlling NO x emissions from stationary resources and diesel vehicles (Li et al., 2011).However, the optimal operation temperature of commercial SCR catalyst (V 2 O 5 -WO 3 (MoO 3 )/TiO 2 ) is in a narrow window of 300 to 400°C, and the formation of N 2 O over V 2 O 5 -based catalyst is significant at high temperature (Chen et al., 2010).Therefore, many studies have been investigated for the low-temperature SCR catalyst, which could contribute to low energy consumption and improve economics for flue gas cleaning (Li et al., 2012).
The specific surface area of OMC can be as large as 1000-4000 m 2 g -1 , and the narrow pore size distribution associate with controllable morphology can promote mass transfer, which are also highly desired for catalysis applications (Liu et al., 2012;Florent et al., 2013;Wang et al., 2015).Great efforts have been made to modify the structure and properties of OMC, and transitional metal oxides (MnO x (Liu et al., 2013;Su et al., 2013), CuO x (Ouzzine et al., 2008;Amanpour et al., 2013;Chen et al., 2014), CeO x (Shen et al., 2013;Zhang et al., 2013), FeO x (Yang et al., 2012;Zhang and Qu, 2015), etc.) are introduced to prepare metal-containing OMC catalysts based on their enhancement in catalytic performance as active phases.Liu et al. (2012) introduced a one-pot aqueous self-assembly strategy to disperse Cu on OMC, which exhibited a higher activity and selectivity than the traditional impregnated Cu/AC catalyst.Kiani et al. (2014) investigated the ratios of MnO2/OMC composite synthesized by impregnation method and found that the chemical properties were remarkably improved with the addition of MnO 2 .Moreover, Mn-based and Cu-based catalysts, such as MnO x /TiO 2 (Sultana et al., 2012), MnO x /CNTs (Su et al., 2013), CuMn 2 O 4 (Fang et al., 2014), CuO x /AC (Li et al., 2012), exhibit relatively high SCR activities especially at low temperature.Therefore, a series of CuO x and MnO x containing OMC with different metal loadings were prepared for NH 3 -SCR of NO in our previous study (Cao et al., 2015), which illustrated that the catalyst with 5 wt.%Cu and 5 wt.%Mn exhibited the best catalytic performance.In addition, a possible synergetic mechanism of the redox cycle between Cu and Mn ions was proposed.
As is well known, the preparation method is crucial in controlling the structural and surface properties of catalysts, which have a significant impact on catalytic activity.Lian et al. (2015) found that VO x /CeO 2 prepared by a simple homogeneous precipitation method showed higher NH 3 -SCR activity and higher SO 2 resistance than catalysts prepared by incipient wetness impregnation or sol-gel method.In situ prepared MnO x -CeO x /CNTs was also reported to display a better SCR activity than catalysts prepared by impregnation or mechanical mixing (Zhang et al., 2013).Therefore, a systematic study for the effect of preparation method on SCR performance of Cu-Mn/OMC catalysts would be worthwhile.In the present work, Cu-Mn/OMC catalysts with 5 wt.%Cu and 5 wt.%Mn were prepared by three different methods: self-assembly (S), ultrasound-assisted impregnation (I), and mechanical mixing (M).The performance of the prepared catalysts for NH 3 -SCR of NO was evaluated in a quartz reactor from 100°C to 300°C.The structural and surface properties of the obtained catalysts were characterized by N 2 adsorption-desorption, XRD, TEM, XPS, H 2 -TPR, and NH 3 -TPD.The effects of preparation method on the catalytic activity of Cu-Mn/OMC were discussed in detail.The correlation between the catalytic performance and the surface properties was also investigated.

Self-Assembly Synthesis
Cu-Mn/OMC catalyst was obtained through solvent evaporation-induced self-assembly method, as reported by our group (Cao et al., 2015).In a typical process, 4.80 g of F127 was dissolved in ethanol, mixed with 6.24 g of TEOS, 0.567 g of Cu(NO 3 ) 2 , 0.668 g of Mn(AC) 2 , and 12 g of 25 wt.% resol ethanolic solution under magnetic stirring for 2 h.Then the mixture was transferred into a dish to evaporate ethanol followed by drying at 110°C for 24 h and calcination at 850°C under inert atmosphere for 1 h.After washing and drying, the obtained catalyst was denoted as Cu-Mn/OMC(S).

Ultrasound-Assisted Impregnation
Briefly, 4.80 g of F127 was dissolved in ethanol, mixed with 6.24 g of TEOS and 12 g of 25 wt.% resol ethanolic solution under magnetic stirring for 2 h.Then, the mixture was transferred into a dish to evaporate ethanol, followed by drying at 110°C for 24 h and calcination at 850°C under inert atmosphere for 1 h.An OMC support was obtained after washing and drying.For the ultrasound-assisted impregnation method, 1.0 g of OMC, 0.188 g of Cu(NO 3 ) 2 , and 0.223 g of Mn(AC) 2 were mixed in deionized water.After stirring for 6 h, the obtained mixture was then dried at 110°C overnight, followed by calcination at 500°C for 5 h under inert atmosphere.The obtained catalyst was denoted as Cu-Mn/OMC(I).

Mechanical Mixing
OMC support was prepared as mentioned above.For the mechanical mixing method, 1.0 g of OMC, 0.188 g of Cu(NO 3 ) 2 , and 0.223 g of Mn(AC) 2 were mixed and ground in an agate mortar.The mixture was then calcined at 500°C for 5 h under inert atmosphere, and the obtained catalyst was denoted as Cu-Mn/OMC(M).

Characterization of Catalysts
N 2 adsorption-desorption experiment was conducted using a Micromeritics Instrument Corporation (Norcross, Georgia) ASAP 2020 system.The samples were degassed at 250°C for 3 h before measurement.The specific surface areas were calculated by Brunauer-Emmett-Teller (BET) equation, while the pore volume and average pore diameters were determined by the Barrett-Joyner-Halenda (BJH) method using the desorption branches of the isotherms.
Powder X-ray diffraction (XRD) patterns were obtained on a PAN-alyticalX'Pert PRO XRD system, using Cu Kα radiation (λ = 0.15418 nm).To analyze the crystal phases of catalysts, wide-angle XRD were employed at scattering angles (2θ) ranging from 10° to 80° with a step length of 0.0167°.Small-angle XRD patterns were also recorded for 2θ between 0.5° and 10° to verify the ordered mesoporous structures.
Transmission electron microscopy (TEM) images were obtained using a Hitachi H-9500 microscope operated with an acceleration voltage of 300 kV.The samples were prepared by a suspension in ethanol, followed by evaporating onto a carbon-coated copper grid.Energy dispersive X-ray spectroscopy (EDX) data were recorded to determine the elemental contents.
X-ray photoelectron spectroscopy (XPS) were recorded by a Thermo ESCALAB 250Xi using Al Kα as the radiation source at constant pass energy of 1486.6 eV to collect the surface information of the samples.The binding energies for C 1s, O 1s, Cu 2p, and Mn 2p were calculated with reference to the energy of C 1s peak at 284.6 eV.
Temperature programmed desorption of ammonia (NH 3 -TPD) was performed using a Micromeritics AutoChem II 2920 chemisorption analyzer.Prior to desorption, 20 mg of samples were heated in a He stream (30 mL min -1 ) from room temperature to 200°C and held for 1 h.After cooling to 100°C, the samples were saturated with NH 3 (40 mL min -1 ) for 1.5 h.The TPD profiles were recorded from 120°C to 600°C at a heating rate of 10 °C min -1 .
Temperature programmed reduction of hydrogen (H 2 -TPR) was carried out with the same catalyst characterization system for NH 3 -TPD.The experiments were performed in 10% H 2 /Ar with a flow rate of 30 mL min -1 and at a heating rate of 10 °C min -1 .Briefly, 20 mg of samples were pretreated at 200°C under a flow of helium (30 mL min -1 ) for 2 h followed by an increase in temperature from 100°C to 800°C.Isopropanol and liquid nitrogen was used to collect water formed during the process.The amount of H 2 consumption was measured by the corresponding thermal conductivity detector (TCD) signals.

NH 3 -SCR Activity Tests
The NH 3 -SCR catalytic activity was evaluated using a quartz flow tube reactor heated by a temperature-controlled furnace.Briefly, 200 mg of the catalyst (30-60 meshes) was placed in the reactor and heated to 100°C under N 2 flow for 1 h to purify the sample surface.The total flow rate was 60 mL min -1 , corresponding to a gas hourly space velocity (GHSV) of 36000 h -1 .The gas mixture contained 500 ppm NO, 500 ppm NH 3 , 5% O 2 , and the balance was N 2 .All data were recorded at a chosen temperature range between 100°C to 300°C, after the reaction reached a steady state.The inlet flow rate was controlled by a mass flow controller and the concentration of NO, NH 3 , N 2 O and NO 2 in the outlet gas was continuously analyzed with a gas analyzer (Gasmet DX-4000).NO conversion and N 2 selectivity of SCR were calculated according to the following expressions: where the [NO]

N 2 Adsorption-Desorption Isotherms
The N 2 adsorption-desorption isotherms of Cu-Mn/OMC catalysts in Fig. 1(A) exhibited typical type-IV isotherms and H 1 hysteresis loops, indicating the existence of mesostructure with a narrow pore size distribution.The hysteresis loops of Cu-Mn/OMC(I) and Cu-Mn/OMC(M) occurred approximately at the relative pressure P/P 0 = 0.50-0.85,owing to the capillary condensation of mesoporous structure.While the hysteresis loop of Cu-Mn/OMC(S) slightly shifted to a smaller relative pressure (P/P 0 = 0.45-0.75),which implied a smaller pore diameter, this finding was also confirmed by the BJH adsorption pore size distribution as shown in Fig. 1(B).
The BET surface area, total pore volume, and average pore diameters are summarized in Table 1.An OMC support was also investigated for comparison, which exhibited a specific surface area of 1057 m 2 g -1 and a pore volume of 1.29 cm 3 g -1 , and both decreased after incorporation of Cu and Mn.This finding might be due to pore blocking caused by the loading of active phases (Alegre et al., 2014;Cao et al., 2015).No obvious differences in structural properties were observed between Cu-Mn/OMC(I) and Cu-Mn/OMC(M); however, the catalyst prepared by self-assembly synthesis possessed relatively low specific surface area and pore volume.This may be attributed to the distribution of carbon matrix with the incorporation of metal precursors during the self-assembly process, which could have a more remarkable impact on the pore structure (Gao et al., 2008).

XRD
The small-angle XRD patterns of ordered pore structures of Cu-Mn/OMC catalysts are presented in Fig. 2(A).The featured diffractions of Cu-Mn/OMC(M) were very weak, and a broad peak at 2θ = 2°-4° was detected.A strong peak at 2θ = 0.5°-1° was detected for Cu-Mn/OMC(S) and Cu-Mn/OMC(I), revealing the existence of ordered 2D p6mm hexagonal mesostructure (Tripathi et al., 2014).The peak intensity of Cu-Mn/OMC(S) was weaker than that of Cu-Mn/OMC(I), illustrating a decline in the structural order of catalysts, which was consistent with the N 2 adsorptiondesorption results.
The Cu-Mn/OMC catalysts were further characterized by wide-angle XRD to evaluate the crystallinity and dispersion of active phase, as shown in Fig. 2(B).The broad diffraction peaks at 2θ = 20°-24° of all samples could be indexed as amorphous carbon.Diffraction peaks of Cu 0 (at 43.5°, 50.8°, 74.3°, JCPDS 65-9743) and Cu 2 O (at 36.4°,42.3°, 61.3°, 73.9°, JCPDS 05-0667) were observed in the patterns, whereas no CuO phase was found.This suggested that Cu 2+ was only present on the surface of Cu-Mn/OMC, which formed a thin amorphous outer shell (Yin et al., 2005), and Cu 2+ in precursors might also be reduced to Cu 0 and Cu + during the carbonization process (Liu et al., 2012).In addition, the diffraction peaks of Cu-Mn/OMC(S) corresponding to Cu and Cu 2 O were weaker than that in other samples, which could be attributed to the better dispersion of Cu phases (Sultana et al., 2012).No diffraction peaks of Mn oxides were found for Cu-Mn/OMC(S) and Cu-Mn/OMC(I), indicating that Mn oxides were highly dispersed on the surface of the OMC support.For Cu-Mn/OMC(M), the diffraction peaks at around 32.4°, 38.1°, 34.9°, 40.5°, 58.7° and 70.2° were observed owing to the lattice plane of Mn 3 O 4 (103) (211) (JCPDS 18-0803) and MnO (111) (200) (220) (311) (JCPDS 07-0230), which indicated the agglomerated Mn oxides on the surface of OMC support (Wang et al., 2015).The agglomeration of metal oxides may have destroyed the ordering of pore structure on the surface, which was consistent with the SAXRD results shown in Fig. 2(a).The XRD results showed that Cu and Mn oxides could be dispersed better through self-assembly synthesis than the other two methods, and such better dispersion was supposedly associated with the high activity in NH 3 -SCR reaction.

TEM Images and EDX
The micro-morphology and constituents of the catalysts were determined by TEM and EDX analysis.TEM images depicted in Fig. 3 show strip-like and hexagonally arranged patterns in large domains, which confirmed the preservation of 2-D hexagonal mesoporous structure (p6mm) after metal loading (Wang et al., 2015).For Cu-Mn/OMC(S) in Fig. 3(A), the dark spots with uniform size were observed

Samples
BET surface area (m 2 g -1 ) Total pore volume (cm 3 g -1 ) Average pore diameter (nm)  111) and ( 200) directions, respectively.In addition, lattice spacing of 2.57 nm was also found, which correspond to MnO in (111) directions, indicating the presence of Cu and Mn oxide phases on the OMC support.
The relative atomic contents determined through EDS are listed in Table 1.The contents of active phase in three samples were approximate to the theoretical value (5%), indicating little metal loss during catalyst preparation process.In addition, the population of oxygen in Cu-Mn/OMC(S) was much higher than that in catalysts prepared by the other two methods, which might be due to the oxidation of carbon support with in-situ addition of precursor acid radical during preparation and could lead to relatively high catalytic performance (Teng et al., 2001).

XPS
The catalysts prepared by three methods were further analyzed by XPS to determine the chemical state of species on the surface, and the obtained C 1s, O 1s, Mn 2p, and Cu 2p spectra are shown in Figs.4(A)-4(B).The spectrum of C 1s could be divided into four single peaks corresponding to C-C (284.4 eV), C-O (285.4 eV), O-C=O (287.0 eV), and π-π* (289.9 eV) (Cao et al., 2015), and the contents of groups are summarized in Table 2.The Cu-Mn/OMC(S) had the largest proportion of O-C=O groups, which are usually ascribed to carboxyl and ester groups.The content of C-O groups in Cu-Mn/OMC(S) was less than that in the other two samples, which might be caused by the transformation of C-O groups into O-C=O (Levi et al., 2015).As the carboxyl had the strongest acidity among fundamental groups (Teng et al., 2001), the increase of O-C=O might lead to a higher surface acidity and superior catalytic performance (Zhao et al., 2014).The O 1s spectra were fitted into three peaks assigned to lattice oxygen in the metal oxides (530.6 eV), C-OH or C-O-C (532.2 eV), and chemisorbed O (533.3 eV), respectively (Biniak et al., 1997;Zafra et al., 2014).It was distinct that most of O atoms on the surface of catalysts existed in the form of C-O, and the proportion of chemisorbed O on Cu-Mn/OMC(S) was higher than that on Cu-Mn/OMC(I) or Cu-Mn/OMC(M).The high mobility of chemisorbed O was beneficial for the adsorption of NO and could increase SCR activity.
The Mn 2p and Cu 2p spectra of the catalysts are shown in Figs.4(C)-4(D), and the relative percentages of Mn and Cu are listed in Table 3.The Mn 2p 3/2 could be deconvolved into Fig. 3. TEM images of (A) Cu-Mn/OMC(S), (B) Cu-Mn/OMC(I), and (C) Cu-Mn/OMC(M) and HRTEM images of (D) Cu-Mn/OMC(S), (E) Cu-Mn/OMC(I), and (F) Cu-Mn/OMC(M).three characteristic peaks representing the Mn 2+ (640.1 eV), Mn 3+ (641.1 eV), and Mn 4+ (642.5 eV), respectively (Yang et al., 2011).Table 3 show that the molar ratio of Mn 4+ /Mn 3+ on Cu-Mn/OMC(S) was much higher than that on Cu-Mn/OMC(I) or Cu-Mn/OMC(M).Previous studies illustrated that Mn species with higher oxidation state is more active in redox reactions over Mn-based catalysts (Wang et al., 2012), and Mn 4+ exhibits the highest activity among different manganese oxides (Kapteijn et al., 1994).The increase in Mn 4+ /Mn 3+ was beneficial to the oxidation of NO to NO 2 , which could enhance the low-temperature SCR activity (Liu et al., 2009).For Cu 2p spectra, the two characteristic peaks at 932.4 eV and 934.6 eV were attributed to Cu + and Cu 2+ , respectively, by performing peak-fitting deconvolutions.Cu 2+ was considered to adsorb and activate NO molecules (Fang et al., 2014), and would favor NH 3 -SCR reaction by the redox cycle between Cu 2+ and Cu + .Given that the Cu-Mn/OMC(S) had a significantly higher ratio of Mn 4+ /Mn 3+ and Cu 2+ /Cu + than Cu-Mn/OMC(I) or Cu-Mn/OMC(M), together with a stronger surface acidity and a larger amount of active oxygen owing to high content of O-C=O and chemisorbed O, the Cu-Mn/OMC(S) might exhibit the best catalytic performance in NH 3 -SCR reaction.

NH 3 -TPD
The acid properties of Cu-Mn/OMC samples were investigated because selective reduction of NO by NH 3 is significantly affected by the strength of acid sites on the surface of catalysts (Fang et al., 2014).The NH 3 -TPD profiles of three catalysts are shown in Fig. 5, and the amounts of NH 3 desorbed from per unit area of catalysts are calculated in Table 4.The desorption peak of NH 4+ bound to weak Brφnsted acid sites is detected at low temperature, whereas the peak at high temperature is thought to be caused by NH 3 desorption from medium acid sites (Mhamdi et al., 2009;Shen et al., 2013).We could notice that the profile of   Cu-Mn/OMC(I) or Cu-Mn/OMC(M) showed two desorption peaks at around 167 and 425°C, corresponding to weak Brφnsted acid sites in the range of 120-300°C and medium acid sites in the range of 300-600°C, respectively.For Cu-Mn/OMC(S), the peak of weak acid sites shifted to a higher temperature at 182°C, suggesting an intensification of weak Brφnsted acid sites on the surface of the catalyst.Two desorption peaks in the high temperature range were also observed at 333 and 382°C, indicating the presence of at least two kinds of medium acid adsorption sites.In addition, the NH 3 uptake of Cu-Mn/OMC(S) was larger than that of the other two samples at both weak and medium acid sites.Given that the quantity of NH 3 uptake was directly proportional to surface acid amount and the peak temperature was correlated with surface acid strength, it could be concluded that the Cu-Mn/OMC(S) catalyst had a stronger acidity than Cu-Mn/ OMC(I) or Cu-Mn/OMC(M) especially at low-temperature range of weak Brφnsted acid sites, which might be attributed to the highly dispersed active phases on OMC and the strong interaction between Cu and Mn oxides (Zhang et al., 2013).This finding was consistent with the XPS results and might contribute to the best catalytic activity of NH 3 -SCR reaction.

H 2 -TPR
The surface redox properties of the catalysts synthesized by different methods were characterized by the H 2 -TPR, and the amounts of H 2 consumption by per unit area of catalysts are summarized in Table 4.The H 2 -TPR profiles as a function of reduction temperature are presented in Fig. 6. Results of Cu-Mn/OMC(S) showed a strong peak at 196°C, which was higher than the reduction temperature of Cu/OMC at 173°C (Cao et al., 2015) and could be attributed to the reduction of Cu 2+ → Cu + in the presence of Mn oxides (Fierro et al., 1994;Kim et al., 2011).Two weak peaks at 356 and 538°C were also found for Cu-Mn/OMC and could be interpreted as the reduction of Mn 4+ → Mn 3+ → Mn 2+ (Wu et al., 2008).For Cu-Mn/OMC(I) and Cu-Mn/OMC(M), the position of the strong peaks shifted to lower temperatures at 176°C and 162°C, respectively.The Cu-Mn/OMC(M) catalyst had the lowest H 2 reduction temperature, indicating the highest reactivity of surface oxidation species, which might decrease N 2 selectivity in NH 3 -SCR reaction.Although the activity of oxidation species on the surface of Cu-Mn/ OMC(S) was lower than that of Cu-Mn/OMC(I) or Cu-Mn/OMC(M), the Cu-Mn/OMC(S) had the largest amount of H 2 consumption (Table 4), suggesting a high quantity of surface oxidation species which was consistent with the XPS results.Redox capability of catalyst is crucial to the catalytic cycle in the NH 3 -SCR reduction and in increasing the activity of catalysts (Lian et al., 2015).Considering that the metal loading and dispersity of active phase was strongly related to the redox property (Lin and Bai, 2014), the highest redox capability of Cu-Mn/OMC(S) might be due to the well-dispersed of Cu and Mn oxides, as indicated by the XRD and TEM results.

Catalytic Performance
The SCR activity of NO with NH 3 was evaluated as a function of temperature from 100 to 300°C, as presented in Fig. 7(A).The conversion of NO over all of the catalysts first increased with the increase in temperature and reached the maximum at 250°C, however, the catalytic activity decreased with further increase of reaction temperature.Cu-Mn/OMC(S) displayed the best SCR activity among the three catalysts and the highest NO conversion of 85% was reached at 250°C with T 50 (referring to the reaction temperature at which NO conversion achieved 50%) at 200°C.However, the activities of Cu-Mn/OMC(I) and Cu-Mn/OMC(M) were much lower, with highest NO conversion of 59% and 45%, respectively.In addition, the activity of Cu-Mn/OMC(M) was a little higher than that of Cu-Mn/OMC(I) at low reaction temperature, whereas Cu-Mn/OMC(I) exhibited a better performance at high temperatures (> 250°C).This phenomenon might be due to the strong activity of oxidation species on Cu-Mn/OMC(M) corresponding to the lowest consumption peak temperature in H 2 -TPR, which was beneficial for SCR reaction at low temperature range.
N 2 selectivity was measured to evaluate the SCR  performance of catalysts.For all of the catalysts, N 2 O was not formed until 200°C and the NO conversion decreased with the increase in temperature.In addition, Cu-Mn/OMC(S) had the optimum selectivity performance in SCR reaction with no less than 86% of N 2 selectivity, followed by Cu-Mn/OMC(I) and Cu-Mn/OMC(M).N 2 O is an important by-product resulting from nonselective oxidation of NH 3 .The outlet concentration of N 2 O was measured, and it was worth noticing that the N 2 O concentration of Cu-Mn/OMC(S) was the lowest.NH 3 oxidation was also evaluated given that N 2 selectivity is closely related to the oxidizing capacity of catalysts (Roy et al., 2008).As shown in Fig. 7(C), the oxidation of NH 3 by the three catalysts followed the order: Cu-Mn/OMC(M) > Cu-Mn/OMC(I) > Cu-Mn/OMC(S).This finding is consistent with the N 2 selectivity results.Cu-Mn/OMC(M) exhibited the strongest NH 3 oxidizing capacity and the lowest N 2 selectivity, which might be attributed to the highest reactivity of surface oxidation species as indicated by the lowest reduction temperature in H 2 -TPR.The stability of catalyst was further tested, and it could be seen from which indicated a good stability in activity and selectivity of Cu-Mn/OMC(S) in NH 3 -SCR of NO.Morphology of active phase, amount of active oxygen, redox ability, and surface acidity are affected by preparation methods, and these properties could have an important influence on SCR performance (Kobayashi and Miyoshi, 2007;Satsuma. et al., 2015).As demonstrated by the N 2 adsorption-desorption characterization, the decrease of specific surface area for Cu-Mn/OMC(S) was more obvious than that for the other catalysts, possibly because of better combination of active phases with the OMC support, which was beneficial for catalytic activity.The XRD and TEM results suggest that the difference in dispersion of metal oxides could cause the differences in catalytic activity, and the highly dispersed active phases on the surface of Cu-Mn/OMC(S) samples was beneficial to NO conversion (Lian et al., 2015).In the XPS analysis, more Mn 4+ /Mn 3+ and Cu 2+ /Cu + were observed in Cu-Mn/OMC(S), which could be beneficial to the oxidation of NO and favorable for the NH 3 -SCR reaction.For monometallic catalyst of Cu/ OMC, the redox cycle of Cu 2+ ↔ Cu + was proposed during SCR reaction.With the addition of Mn, the interaction between Cu and Mn could enhance the redox process of Cu 2+ + Mn 3+ ↔ Cu + + Mn 4+ (Cao et al., 2015), which would decrease the electron transfer energy and improve catalytic performance.The adsorption and activation of NH 3 on the surface of catalyst, together with the adsorption of NO in the neighborhood, were considered the two crucial steps of NH 3 -SCR reaction (Sun et al., 2009).NH 3 -TPD and H 2 -TPR results demonstrated that Cu-Mn/OMC(S) not only performed the strongest acid strength, but also showed the largest quantity of oxidation species.Strong acid strength of catalysts could increase the adsorption of NH 3 , and the oxidizing capacity was important for NH 3 activation (Lee et al., 2015).Based on these favorable properties, it was obvious that Cu-Mn/OMC(S) exhibited the best activity in NH 3 -SCR reaction.

CONCLUSIONS
OMC was used as support for Cu and Mn oxides in NH 3 -SCR of NO, and the influence of preparation methods on catalytic activity was investigated in this study.Results showed that Cu-Mn/OMC prepared by self-assembly synthesis displayed a higher NO conversion and N 2 selectivity at 250°C than the catalysts prepared by impregnation or mechanical mixing.The N 2 adsorption-desorption isotherms and small-angle XRD results showed that the uniform structure of OMC was maintained successfully when Cu and Mn species were introduced.Wide-angle XRD and TEM images demonstrated good dispersion and small crystallite size of active phases on Cu-Mn/OMC(S), whereas slight agglomerations were found on Cu-Mn/OMC(I) and Cu-Mn/OMC(M).NH 3 -TPD and H 2 -TPR profiles illustrated that Cu-Mn/OMC(S) exhibited the strongest acidity and largest amount of active oxygen species, which were consistent with the XPS results.Therefore, the outstanding SCR performance of Cu-Mn/OMC(S) might be attributed to the highly dispersed active phases, strong surface acidity, and abundant oxidative species.In view of the advantages discussed above, self-assembly synthesis is an effective method for preparation of Cu-Mn/OMC catalyst for lowtemperature SCR of NO with NH 3 .
in and [NO] out indicate the inlet and outlet concentrations of NO, respectively, and the [NO 2 ] out and [N 2 O] out are the outlet concentrations of NO 2 and N 2 O, respectively.

Table 1 .
Textural properties and elemental content of catalysts.

Table 2 .
Deconvolution of XPS core-level C 1s and O 1s of catalysts.

Table 3 .
Deconvolution of XPS core-level Mn 2p and Cu 2p of catalysts.

Table 4 .
Amount of acidic sites determined from NH 3 -TPD and H 2 consumption during H 2 -TPR of catalysts.