**1. Introduction**

Currently, environmental protection is more stringent than ever before. The large quantities of nitrogen oxides (NOx) produced by the burning of fossil fuels are a major cause of atmospheric pollutants. Carbon monoxide (CO) is another atmospheric pollutant in flue gases. Thus, the reduction of NO by the CO produced by incomplete combustion in the flue gas can remove toxic CO and NO simultaneously and economically (CO-SCR) [1–3]. However, the high price and low catalytic activity at low temperature (more than 50% NO conversion below 250 ◦C) of efficient noble metal catalysts seriously limit their further application. Therefore, it is necessary to develop catalysts with low temperature, high performance, low cost and that are green [4].

For the CO-SCR reaction, the ideal catalyst should not only be economical, easy to prepare, long-term stable and so on. In addition, a low reaction temperature [5,6], high selectivity [7,8] and NO conversion rate [9] are required. Noble metals are frequently used in CO-SCR reactions to prepare noble-metal catalysts. However, the scarce resources, high price and high temperature instability limit its large-scale application. As a result, many studies have focused on the development of nonprecious metals. NO reduction occurs through a redox reaction mechanism. Therefore, the reducibility and oxygen migration ability of the catalyst are two key factors that determine the catalytic performance of the catalyst for NO removal. At present, metal oxides have become a hotspot of heterogeneous catalysis research because of their low price and large reserves, such as CoO<sup>x</sup> [10,11], CuO<sup>x</sup> [12–14], MnO<sup>x</sup> [15,16] and CeO<sup>2</sup> [17]. Among them, copper oxides and manganese oxides have attracted much attention due to their good redox properties. Manganese oxides show a variety of valences (Mn2+, Mn3+, Mn4+) and abundant reactive oxygen species

(vacancy oxygen and adsorbed oxygen), which imply their potential in low-temperature CO-SCR catalysis [18–20].

In related reports, the reducibility and oxygen migration ability of MnO<sup>x</sup> could be improved by proper cation doping. These include MnCe [21], MnCu [22–24], MnCo [25], MnNi [26,27] and MnFe [28]. Because of its excellent oxidation–reduction performance and strong synergistic effect between binary metal oxides, doping copper into the catalyst can effectively improve the removal rate of the catalyst. Wan et al. [29] found that the Mn2O3 modified CuO/γ−Al2O<sup>3</sup> catalyst showed significant catalytic efficiency, and they attributed the increase in activity to the establishment of a Cu2+ + Mn3+ Cu<sup>+</sup> + Mn4+ oxidation– reduction cycle. In addition, the addition of Cu to Mn-based catalysts is beneficial to the dispersion of MnOx. The performance of copper oxides is affected by many factors in the NO + CO reaction. Ivanka Spassova [24] reported that CuCo2O<sup>4</sup> and Cu1.5Mn1.5O<sup>4</sup> mixed oxides supported on DFS were responsible for enhancing activity. The results showed that Liu [30] suggested that copper-modified manganites had higher catalytic activity for CO oxidation and selective catalytic reduction of NO than pure MnOx. Therefore, it is further expected that CuO and MnO<sup>x</sup> form a strong coupling at the nanointerface, which will lead to a change in the Mn4+ octahedral environment, thereby further improving the CO−SCR performance of MnOx.

This article reports that foam-like CuxMn3−xO<sup>4</sup> spinels were prepared by using a citrate-based modified pechini method and applied to the CO-SCR reaction in the temperature range of 100–400 ◦C. It was characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), X-ray diffraction (XRD), BET surface area (BET), H<sup>2</sup> temperature programmed reduction (H2−TPR) and X–ray photoelectron spectroscopy (XPS). The structure-activity relationship between the physical chemistry properties and the catalytic performance of the CuxMn3−xO<sup>4</sup> catalyst with different concentrations of Mn4+ was studied. The purpose of this work is to investigate the relationship between the active phase of spinel and the bulk properties of CuxMn3−xO<sup>4</sup> (x = 0, 1, 1.5, 2, 3) catalysts prepared with different CuO/MnO<sup>x</sup> contents.

#### **2. Results and Discussion**

#### *2.1. XRD Analysis of Catalysts*

XRD patterns were tested to identify the crystal structure of Mn2O3, CuO and synthesized CuxMn3−xO<sup>4</sup> spinels. As shown in Figure 1a, for the Mn2O<sup>3</sup> sample, (200), (211), (222), (123), and (440) planes of Mn2O<sup>3</sup> (JCPDS#01-076-0150) could be observed at 18.5◦ , 23.1◦ , 33.0◦ , 35.7◦ and 55.0◦ , respectively. The diffraction peaks of 32.5◦ , 35.5◦ , 38.6◦ , 48.9◦ , 53.4◦ , 58.2◦ , 61.5◦ , 66.3◦ , 67.7◦ , 68.0◦ , 72.3◦ and 82.6◦ were assigned to the (110), (−111), (111), (−202), (020), (202), (−113), (−311), (113), (220), (311) and (−313) planes of cubic phase CuO (JCPDS#01-080-0076). A CuxMn3−xO<sup>4</sup> mixed oxide with a spinel structure was found in the Cu1Mn2O<sup>4</sup> (JCPDS#01-074-2422), Cu1.5Mn1.5O<sup>4</sup> (JCPDS#01-070-0260) and Cu2Mn1O<sup>4</sup> catalysts. XRD patterns show that the diffraction peak (I peak) can match spinel Cu1.5Mn1.5O<sup>4</sup> (Figure 1b). Compared with other samples, the intensity of the "I" diffraction peak of the Cu1.5Mn1.5O<sup>4</sup> sample is the strongest, indicating that the Cu1.5Mn1.5O<sup>4</sup> sample contains a spinel active structure (Cu-O-Mn) [30,31]. As for Cu1Mn2O<sup>4</sup> and Cu1.5Mn1.5O4, they showed identical diffraction patterns to Mn2O<sup>3</sup> but only with a slight shift in the peak position of Mn2O<sup>3</sup> toward high values, implying the insertion of Cu atoms with smaller radius than Mn atoms into the lattice of Mn2O3. It is also noticed that the crystallinity of Cu1.5Mn1.5O<sup>4</sup> becomes higher in comparison with that of Cu1Mn2O<sup>4</sup> and Cu2Mn1O4, implying that excessive Cu doping is not conducive to the formation of Cu-O-Mn structure (Table 1). The lattice parameters of the synthesized CuxMn3−xO<sup>4</sup> catalyst were calculated by XRD, as shown in Table 1. Compared to CuxMn3−xO<sup>4</sup> spinels, the lattice parameters of CuxMn3−xO<sup>4</sup> spinels became smaller after doping with increased copper contents. The results also prove the above conclusions.

spinels became smaller after doping with increased copper contents. The results also

**Table 1.** Crystal sizes, lattice parameters, actual molar ratios of Cu to Mn and BET surface areas of

**Lattice Parameter <sup>a</sup> nm**

Mn2O<sup>3</sup> 63.07 a = b=c = 0.9423 - 18.2 Cu1Mn2O<sup>4</sup> 42.69 a = b = c = 0.8290 0.93:2.05 18.9 Cu1.5Mn1.5O<sup>4</sup> 31.79 a = b = c = 0.8284 1.46:1.54 19.7 Cu2Mn1O<sup>4</sup> 31.76 a = b = c = 0.8282 1.97:1.02 18.7

<sup>a</sup> Calculated 2*θ* = 33.0° by the XRD patterns using the Debye–Scherrer equation. <sup>b</sup> Obtained by the

**Actual Molar Ratios of Cu:Mn <sup>b</sup>**

<sup>c</sup> <sup>=</sup> 0.5135 - 28.9

**BET Surface Area m<sup>2</sup> g–1 <sup>c</sup>**

prove the above conclusions.

**Sample Crystal Size**

**nm**

CuO 42.43 <sup>a</sup> <sup>=</sup> 0.4687, b <sup>=</sup> 0.3427,

ICP results. <sup>c</sup> Surface area derived from the BET equation.

CuxMn3−xO4.

**Figure 1.** XRD patterns (**a**), local enlargement of (I) in XRD (**b**) of the as-synthesized CuxMn3−xO<sup>4</sup> samples. **Figure 1.** XRD patterns (**a**), local enlargement of (I) in XRD (**b**) of the as-synthesized CuxMn3−xO<sup>4</sup> samples.

*2.2. N<sup>2</sup> Sorption Analysis of Catalysts* Figure 2 illustrates the obtained N<sup>2</sup> adsorption-desorption isotherm and pore size **Table 1.** Crystal sizes, lattice parameters, actual molar ratios of Cu to Mn and BET surface areas of CuxMn3−xO<sup>4</sup> .


ples have mesoporous and macroporous structures with a large distribution range of pore [32]. It should be pointed out that of the Cu1.5Mn1.5O<sup>4</sup> catalyst own the largest BET surface <sup>a</sup> Calculated 2*θ* = 33.0◦ by the XRD patterns using the Debye–Scherrer equation. <sup>b</sup> Obtained by the ICP results. <sup>c</sup> Surface area derived from the BET equation.

area and most mesoporous among the CuxMn3−xO<sup>4</sup> catalysts, Cu doping leads to the for-

#### mation of more Cu1.5Mn1.5O<sup>4</sup> spinel structures, resulting in irregular changes in grain size. *2.2. N<sup>2</sup> Sorption Analysis of Catalysts*

The specific surface areas of Cu−Mn spinel oxides with different Cu/Mn ratios are recorded in Table 1. The corresponding results conform to the XRD analysis of the catalysts. Figure 2 illustrates the obtained N<sup>2</sup> adsorption-desorption isotherm and pore size distribution of all the catalysts. The CuxMn3−xO<sup>4</sup> samples have type IV isotherms, which also proves that the samples possess a mesopores and significant macropores structure, and that the results of mesopores or macroporous foamy network structure are consistent with that of SEM. The low-pressure part of the near-linear middle part of the isotherm curve can be attributed to the unsaturated adsorption of single or multilayers, which also proves the existence of a macroporous structure. However, the hysteresis loops in the high p/p<sup>0</sup> range are related to capillary condensation in the mesopores, indicating that there are mesopores on the wall of the macropores. In addition, the corresponding Barrett–Joyner–Halenda (BJH) pore-size distribution curves in Figure 2b show that the CuxMn3−xO<sup>4</sup> samples have mesoporous and macroporous structures with a large distribution range of pore [32]. It should be pointed out that of the Cu1.5Mn1.5O<sup>4</sup> catalyst own the largest BET surface area and most mesoporous among the CuxMn3−xO<sup>4</sup> catalysts, Cu doping leads to the formation of more Cu1.5Mn1.5O<sup>4</sup> spinel structures, resulting in irregular changes in grain size. The specific surface areas of Cu−Mn spinel oxides with different Cu/Mn ratios are recorded in Table 1. The corresponding results conform to the XRD analysis of the catalysts.

**Figure 2.** N<sup>2</sup> adsorption-desorption isotherms (**a**) and pore size distributions (**b**) of the assynthesized CuxMn3−xO<sup>4</sup> samples. **Figure 2.** N<sup>2</sup> adsorption-desorption isotherms (**a**) and pore size distributions (**b**) of the as-synthesized CuxMn3−xO<sup>4</sup> samples.

#### *2.3. SEM and TEM Observation 2.3. SEM and TEM Observation*

The morphology and structural characteristics of the as-prepared catalysts at different molar ratios of Cu/Mn were characterized, as shown in Figure 3. Figure 3a,b show SEM images of pure Mn2O<sup>3</sup> at different magnifications. The Mn2O<sup>3</sup> sample is mainly composed of a foam structure with a diameter of 5–20 μm. The magnified SEM image further revealed that the surface of these particles had a hierarchical porous structure. In addition, with increasing Cu doping content, the surface of CuxMn3−xO<sup>4</sup> catalyst particles becomes irregular, and the foam-like particles are broken into a uniform particle structure with a smaller particle size in Figure 3c–j. The mapping of CuxMn3−xO<sup>4</sup> sample images is displayed in Figure 3k1–k4. It can be clearly observed that copper and manganese elements The morphology and structural characteristics of the as-prepared catalysts at different molar ratios of Cu/Mn were characterized, as shown in Figure 3. Figure 3a,b show SEM images of pure Mn2O<sup>3</sup> at different magnifications. The Mn2O<sup>3</sup> sample is mainly composed of a foam structure with a diameter of 5–20 µm. The magnified SEM image further revealed that the surface of these particles had a hierarchical porous structure. In addition, with increasing Cu doping content, the surface of CuxMn3−xO<sup>4</sup> catalyst particles becomes irregular, and the foam-like particles are broken into a uniform particle structure with a smaller particle size in Figure 3c–j. The mapping of CuxMn3−xO<sup>4</sup> sample images is displayed in Figure 3k1–k4. It can be clearly observed that copper and manganese elements are uniformly dispersed on the entire catalyst surface.

are uniformly dispersed on the entire catalyst surface. Figure 4 shows the morphologies and microstructures of the Cu1.5Mn1.5O<sup>4</sup> catalyst at different magnifications. Combined with the SEM results, spherical nanoparticles with particle sizes ranging from 20 to 40 nm were formed in the Cu1.5Mn1.5O<sup>4</sup> sample. According to the equipped Cu1.5Mn1.5O<sup>4</sup> standard card (JCPDS#01-070-0260), the 0.48 and 0.25 nm lattice fringes can be matched to the (111) and (311) crystal planes of the Cu1.5Mn1.5O<sup>4</sup> spinel structure, respectively. It is worth noting that there was a strong synergistic interaction between Cu and Mn oxides in the active components of the spinel structure. Compared with Cu2Mn1O<sup>4</sup> spinel, Cu1.5Mn1.5O<sup>4</sup> has low crystallinity and can provide more oxygen vacancies, which may improve the catalytic performance of Cu-Mn catalysts in CO-SCR Figure 4 shows the morphologies and microstructures of the Cu1.5Mn1.5O<sup>4</sup> catalyst at different magnifications. Combined with the SEM results, spherical nanoparticles with particle sizes ranging from 20 to 40 nm were formed in the Cu1.5Mn1.5O<sup>4</sup> sample. According to the equipped Cu1.5Mn1.5O<sup>4</sup> standard card (JCPDS#01-070-0260), the 0.48 and 0.25 nm lattice fringes can be matched to the (111) and (311) crystal planes of the Cu1.5Mn1.5O<sup>4</sup> spinel structure, respectively. It is worth noting that there was a strong synergistic interaction between Cu and Mn oxides in the active components of the spinel structure. Compared with Cu2Mn1O<sup>4</sup> spinel, Cu1.5Mn1.5O<sup>4</sup> has low crystallinity and can provide more oxygen vacancies, which may improve the catalytic performance of Cu-Mn catalysts in CO-SCR [30].
