Selective Catalytic Reduction of NO by NH 3 Using a Combination of Non-Thermal Plasma and Mn-Cu / ZSM5 Catalyst

: Dielectric barrier discharge (DBD) could generate non-thermal plasma (NTP) with the advantage of fast reactivity and high energy under atmosphere pressure and low-temperature. The presented work investigated the selective catalytic reduction (SCR) of nitric oxide (NO) using a combination of NTP and an Mn-Cu / ZSM5 catalyst with ammonia (NH 3 ) as a reductant. The experimental results illustrate that the plasma-assisted SCR process enhances the low-temperature catalytic performance of the Mn-Cu / ZSM5 catalyst signiﬁcantly, and it exhibits an obvious improvement in the NO removal e ﬃ ciency. The reaction temperature is maintained at 200 ◦ C in order to simulate the exhaust temperature of diesel engine, and the 10% Mn-8% Cu / ZSM5 catalyst shows the highest NO removal performance with about 93.89% at an energy density of 500 J L − 1 and the selectivity to N 2 is almost 99%. The voltage, frequency and energy density have a positive correlation to NO removal e ﬃ ciency, which is positively correlated with the power of NTP system. In contrast, the O 2 concentration has a negative correlation to the NO removal, and the NO removal e ﬃ ciency cannot be improved when the NO removal process reaches reaction equilibrium in the NTP system.


Introduction
The environmental hazards of nitric oxides (NO x ) have been proved to be related to the photochemical smog and greenhouse effects closely in the last several decades. The wanton emission of NO x poses great threats to the ecosystems, which not only causes a series of environmental problems, but also endangers human health [1][2][3][4]. The mobile sources emission is the main source of NO x , especially the high-power diesel engines, and nitric oxide (NO) constitutes a great majority of NO x . The selective catalytic reduction (SCR) of NO x with ammonia (NH 3 ) is regarded as one of the most effective strategies for NO x emission control owing to its reliable stability, which has been widely used in diesel engine [5][6][7][8]. However, the NH 3 -SCR has a high operating temperature window (>350 • C), and the exhaust temperature of the diesel engine is usually 200 • C. The increasingly stringent emission regulations demand the development of high-performance NH 3 -SCR available for low-temperature conditions [9,10].

Characterizations of Mn-Cu/ZSM5 Catalysts
The characterizations of Mn-Cu/ZSM5 catalysts were carried out in this section, which could reveal the composition of catalysts concretely. In order to observe the micromorphology of the catalyst in detail, the 10% Mn-8% Cu/ZSM5 was selected as the representative to get SEM, HRTEM and EDS images, and the results are shown in Figure 1. It shows that the 10% Mn-8% Cu/ZSM5 catalyst is mainly constituted by pseudospherical particle. The HRTEM micrograph reveals that the lattice fringe of 10% Mn-8% Cu/ZSM5 catalyst was 2.5 Å, which could be assigned to the (111) crystal planes of crystalline CuO in accordance with the standard CuO card (JCPDS 01-1117). It illustrates that the target Cu was loaded on the surface of the 10% Mn-8% Cu/ZSM5 catalyst. Further, the EDS images of O, Al, Si, Mn and Cu elements were performed on the 10% Mn-8% Cu/ZSM5 catalyst, which indicates that the active metal elements Mn and Cu were uniformly dispersed on the surface of the ZSM5 carrier. The Al and Si elements were derived from the ZSM5 carrier, and the O element was derived from the active metal oxides of the 10% Mn-8% Cu/ZSM5 catalyst in addition to the ZSM5 carrier. The Brunauer-Emmett-Teller (BET) surface area, pore volume, and pore diameter of ZSM5 and Mn-Cu/ZSM5 catalysts are summarized in Table 1. Obviously, the BET surface area of the prepared catalysts decreased because the ZSM5 was impregnated with Mn and Cu metal elements, and the pore volume was reduced while the pore diameter increased. Li et al. investigated the relationship between the catalytic activity and specific surface areas, and he discovered that the larger specific surface area promoted the NO x adsorption significantly because it could provide more acid sites for the catalytic reaction [37]. The XRD pattern of Mn-Cu/ZSM5 catalysts with different loadings of 4%, 6%, 8%, and 10% Cu are shown in Figure 2. The Mn was not found in the XRD pattern. On the one hand, a small amount of Mn-based metal oxide catalysts were present in the amorphous state. On the other hand, the Mn n+ ions were incorporated into the Cu oxide lattice, and then it tends to form a solid solution during the calcinations based on related research [38,39]. Tang et al. found that the amorphous MnO x played an important role in the removal of NO x [40]. Boningari et al. revealed that amorphous MnO 2 was of higher activity if compared with crystalline MnO 2 for NO reduction [41]. Fang et al. reported that some Mn oxides are strongly associated with Cu oxides to form CuMn 2 O 4 , which was an active oxidation catalyst because of the easy electron transfer between Mn and Cu ions without a change in lattice, improving the lattice oxygen mobility and redox quality [42]. The (111) crystal planes of crystalline CuO were detected at 37 • . Zhao et al. reported that the chemical forms of Cu species were present nearly exclusively in the form of oligomeric species and mononuclear ions [43].  Figure 3 shows the XPS pattern of Mn-Cu/ZSM5 catalysts with different loadings of 4%, 6%, 8%, and 10% Cu, respectively. As depicted in the picture, the Mn 2p 3/2 spectra could be separated into three peaks, including Mn 2+ , Mn 3+ , and Mn 4+ .  [43]. According to the relevant research, the redox cycle between Mn and Cu with different chemical states and NO oxidation mechanism could be described by the following reactions [45][46][47][48][49].
Mn 4+ + NO → Mn 3+ + NO ads NO ads + + O ads − → NO 2 (4) Figure 3. The XPS pattern of Mn-Cu/ZSM5 catalysts with different loadings of 4%, 6%, 8%, and 10% Cu, respectively.  Figure 4 displays that the voltage was positively correlated with the NO removal efficiency in range of 10-20 kV at a frequency of 1000 Hz. The tested concentration of NO and NH 3 were 300 ppm and 60 ppm, respectively. The NO removal efficiency was just 39.16% at a voltage of 20 kV only in the presence of NTP, and NO removal efficiency was 92.08% in combination of NTP and 10% Mn-8% Cu/ZSM5 catalyst, increased by 52.92%. The N 2 selectivity remained stable at nearly 99%. Du et al. found that more energy would be injected into the NTP system with the increase of voltage, and a lot of active species, free radicals, and high-energy electrons could generate under this condition [50]. On one hand, high-energy electrons could collide with H 2 O and O 2 in the air. The ultimate result is that lots of active species and free radicals would generate in the NTP system, such as ·OH, ·O, and so on. On the other hand, high-energy electrons could collide with pollutant molecules directly or indirectly, and then break the chemical bonds. Clearly, the electric field strength was gradually increased with the voltage increasing. Therefore, the electric field intensity and quantity of plasma generated per unit time increased to the NO removal efficiency.  Figure 5 illustrates the effect of frequency on NO removal efficiency at a voltage of 20 kV. The tested concentration of NO and NH 3 were 300 ppm and 60 ppm, respectively. Obviously, the selectivity to N 2 was almost 99%, and the NO removal efficiency was proportional to the discharge frequency, which indicates that the NO removal efficiency was greatly affected by frequency. Zhu et al. found that the energy injected into the NTP system was positively correlated with frequency [51]. Compared with only plasma, the NO removal efficiency increased in the plasma-catalyst system. Among all the catalysts, the 10% Mn-8% Cu/ZSM5 catalyst shows the highest activity for NO removal. Meanwhile, it suggests that there was a significant synergistic effect between the NTP and Mn-Cu/ZSM5 catalyst. However, the NO removal efficiency tended to be gentle at high frequency, which indicates that the energy consumption of NO decreased compared to a low frequency. Yan et al. revealed that the gaseous benzene removal efficiency would increase with the increase of frequency because the energy imported to the reactor in unit time increased [52].

Effect of Energy Density on NO Removal
The discharge power is measured according to the V-Q Lissajous figure method [53], which is shown in Figure 6. The discharge power was given by Equation (5), and the energy density was calculated using Equation (6).
In the equation, T is the period of discharge cycle, in s; u(t) represents the high voltage applied between electrodes, in kV; i(t) is the current changes with time, in A; f represents power frequency, in Hz; u c (t) is the voltage at both ends of C m , in kV; C m represents the measuring capacitance, which is 0.12 µF; S is the area of Lissajous graphics. ε is the energy density, in J L −1 ; and Q is the gas flow, in L min −1 . Figure 7 shows the variation of energy density and the NO removal efficiency. The tested concentration of NO and NH 3 was 300 ppm and 60 ppm, respectively. The NO removal efficiency was climbing up with the increase of energy density, and Mn-Cu/ZSM5 catalysts enhanced the NO removal efficiencies in the presence of NTP, which presents a straight line that approximated the proportional increase. The NO removal efficiency was only 41.66% without the catalyst at an energy density of 500 J L −1 , and NO removal efficiency was 93.89% in combination of NTP and 10% Mn-8% Cu/ZSM5 catalyst, increased by 52.23%, and the selectivity to N 2 is almost 99%. The results indicate that the 10% Mn-8% Cu catalyst with plasma exhibits the highest catalytic activity.

Effect of O 2 Concentration on NO Removal
The relationship between energy density and NO removal efficiency under different O 2 concentrations (0, 2%, 4%, 6%, and 8%) is shown as Figure 8, which is in combination with NTP and the 10% Mn-8% Cu/ZSM5 catalyst. The tested concentration of NO and NH 3 was 300 ppm and 60 ppm, respectively. The energy density varied from 100 to 500 J L −1 . The result indicates that the NO removal efficiency increased gradually with the energy density increasing, and then decreased with the increase of O 2 concentration under the same energy density. The N 2 selectivity shows a similar trend. This result is consistent with Zhao's results, and Zhao et al. revealed that the NO x removal efficiency decreased with the O 2 concentration increasing, and he found that there was a critical O 2 concentration to make the NO x removal efficiency zero [54]. According to relevant research, large quantities of active species would be generated when the O 2 concentration increased in the NTP system, such as ozone (O 3 ) and so on. Then, lots of NH 3 would be oxidized to NO molecules, which play a negative role in the removal of NO. As a final result, the NO and NH 3 were partially oxidized to NO 2 [55,56]. Therefore, the NO removal process is suppressed and the NO removal efficiency decreased. In the presence of O 2 and O 3 , the following reactions would occur in the NTP system based on the relevant research [57][58][59]. The relationship between energy density and NO 2 concentration is presented in Figure 9. In the presence of O 2 molecules, the NO 2 concentration increased slightly when the energy density was over 100 J L −1 , and then the NO 2 concentration remained stable with the energy density continually increasing. Ruggeri et al. revealed that NO 2 was a key reaction product in the catalytic reaction and the SCR reaction dominated the NO reduction with a catalyst below 200 • C [60]. The electric field of the reaction system was strengthened with the increase of energy density, and more NH 3 molecules participated in the reactions with O 2 and ·O, which led to new NO 2 molecule production. On the one hand, the active species O 3 , ·O, and ·N formed in the plasma-catalyst system reacted with NO to form different NO x . On the other hand, O 3 and ·O could react with NO to oxidize NO to NO 2 . Koebel et al.
found that the NO would be oxidized to higher valence oxides because O 3 was more oxidative than O 2 . So, the NO 2 was the main product in the presence of O 2 [61]. The oxidation of NO to NO 2 could be summarized as the following reactions [62][63][64][65].

Effect of NH 3 Concentration on NO Removal
The effect of NH 3 concentration and energy density on NO removal efficiency is presented in Figure 10. The tested NO concentration was 300 ppm and the NH 3 concentration varied from 20 ppm to 100 ppm. The result indicates that the NO removal efficiency was affected by the addition of NH 3 significantly. In the presence of NH 3 , there was higher probability of collision between NO molecules and high-energy electrons in the NTP system. As a result, the numbers of active species and free radicals were improved with the increase of energy density. When the NH 3 concentration increased in the NTP system, it is generally believed that the ammonia radical (·NH 2 ) plays an important role in the NO x removal. Wang et al. reported that the collision reactions of NO molecules were enhanced with the increase of NH 3 concentration, and this is because an enormous amount of free radicals would generate in the NTP system, such as NH 2 ·, NH·, and H· [66]. As a result, more NO molecules are removed under the same energy density. Therefore, the energy consumption decreased in the presence of NH 3 in the NTP system. However, the NO removal efficiency tends to stabilize at high NH 3 concentrations, which indicates that there is an optimal NH 3 concentration in combination of NTP and the Mn-Cu/ZSM5 catalyst. On the whole, the NO removal efficiency cannot be improved when the NO removal process reaches reaction equilibrium in the NTP system, even if the NH 3 concentration increased continuously. The main chemical reactions could be concluded as follows [67,68].

The Catalytic Long-Term Stability of the 10% Mn-8% Cu/ZSM5 Catalyst
The long-time stability is an important index to evaluate the NO removal efficiency of the 10% Mn-8% Cu/ZSM5 catalyst in the NTP system, which could provide basic research data for the practical applications. The tested concentration of NO and NH 3 was 300 ppm and 60 ppm, respectively. The reaction temperature was maintained at 200 • C, and the O 2 concentration was 0 and the energy density of plasma was 500 J L −1 . As can be seen from Figure 11, a 25 h continuous test was performed in this section, and the long-term stability of the 10% Mn-8% Cu/ZSM5 catalyst at an energy density of 500 J L −1 was explored. The NO removal efficiency of the 10% Mn-8% Cu/ZSM5 catalyst was rather stable with a slight fluctuation around ± 1% after the 25 h of continuous testing, and the NO removal efficiency remained at nearly 93%. At the same time, the N 2 selectivity remained stable at nearly 99%, which illustrates that there was almost no deactivation that occurred during the 25 h test at an energy density of 500 J L −1 , and further demonstrates that the 10% Mn-8% Cu/ZSM5 catalyst had excellent long-term stability in the NTP system. Figure 11. The catalytic long-term stability of the 10% Mn-8% Cu/ZSM5 catalyst.

Catalyst Preparation
ZSM5 (Si/Al = 25) was purchased from Nankai Catal. Corp. The Mn-Cu/ZSM5 catalysts were synthesized using the incipient wetness impregnation method. The Mn(NO 3 ) 2 and Cu(NO 3 ) 2 were dissolved in deionized water, and the corresponding proportion ZSM5 was added in solution. The active component Mn loading was 10 wt % and the target Cu loadings were 4 wt %, 6 wt %, 8 wt %, and 10 wt %, respectively. The ZSM5 and mixture of metal nitrates was heated for 2 h in the ohmic heating equipment, and then stirred for 5 h continuously. The obtained solid samples were dried at 100 • C, and then calcined for 5 h at 500 • C in air. Immediately, these catalysts were cooled down to room temperature under N 2 atmosphere. Finally, the Mn-Cu/ZSM5 catalysts were pressed, crushed, and sieved to 40-60 mesh to obtain suitable particles for the following experiments, which were noted as 10% Mn-4% Cu/ZSM5, 10% Mn-6% Cu/ZSM5, 10% Mn-8% Cu/ZSM5, and 10% Mn-10% Cu/ZSM5, respectively.

Catalyst Characterization
The structural and morphological properties of synthesized solid catalyst samples were characterized using scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HRTEM), energy dispersive spectroscopy (EDS), X-ray diffraction (XRD), and X-ray photoelectron spectroscopy (XPS), respectively. The S80 JEOL (SEM, S80 JEOL, Tokyo Japan) was used to examine the microscopy of Mn-Cu/ZSM5 catalysts surface, and the HRTEM images were recorded on a Tecnai G2 F20 S-TWIN microscope (FEI, Hillsboro, OR, USA) at an accelerating voltage of 200 kV. The specific surface area and pore size of Mn-Cu/ZSM5 catalysts were calculated by the Brunauer-Emmett-Teller (BET) method and the Barrett-Joyner-Halenda (BJH) method. Prior to the measurements, these catalysts were outgassed overnight under vacuum. The EDS for the O, Al, Si, Mn, and Cu elements of Mn-Cu/ZSM5 catalysts were measured using an electric refrigeration X-ray energy spectrometer connected to SEM. The XRD measurement was carried out by an XRD-7000 S system (Shimadzu Corporation, Kyoto, Japan). The XPS was implemented on the surface analysis system (VG Multilab2000) with Al Kα radiation (hv = 1486.6 eV) as the excitation source at 150 W, and the C1 s line at 284.6 eV was taken as the standard.

Experimental Setup
The schematic of the DBD plasma experiment is shown as Figure 12. The Mn-Cu/ZSM5 catalyst (2 mL, 40-60 mesh) was placed at the center of the DBD plasma reactor. The concentrations of Ar, O 2 , NO, and NH 3 in the experimental system were regulated by the mass flow controller (MFC), and then the gases flowed into the buffer tank. After they were mixed thoroughly, the gases were tested with the DBD plasma reactor, which could generate a NTP with the advantage of fast reactivity and high energy at atmosphere pressure and low-temperature. The total gas flow was kept constant at 2 L min −1 , and the Ar was used as a carrier gas. The gas hourly space velocity (GHSV) was 60,000 h −1 . The temperature of the mixed gas was maintained at 200 • C in the NTP system. The plasma reactor was connected to an AC power, and the output waveform of DBD plasma was measured by Tektronix-TBS1102 digital oscilloscope, which was equipped with the current loop and P6015A high voltage probe. The concentrations of NO x (NO + NO 2 ) and N 2 O in the outlet gas were analyzed by an FT-IR gas analyzer (GASM ET DX 4000, Finland).  Figure 13 presents the DBD plasma reactor, which adopts the coaxial line structure. The dielectric tube was made of quartz and the sealing gaskets at both ends were made of polytetrafluoroethylene. The copper wire mesh was covered in the quartz, and it was used as a ground electrode. The diameter of dielectric tube was 20 mm, and the thickness of the dielectric tube was 2 mm. The length of the DBD plasma reactor was 500 mm. The NO removal efficiency and N 2 selectivity were calculated by the following Equations (29) and (30). The analysis of experimental results was made in triplicate, and the error bar was used as an index to evaluate the difference between the average and measurement. The following Equations (31) and (32) were the average and error bar, respectively. (32) where NO in is the concentration of NO at the inlet of the DBD plasma reactor, N 2 O out is the concentration of N 2 O at the outlet of the DBD plasma reactor, and NO xout is the concentration of NO x (NO + NO 2 ) at the outlet of the DBD plasma reactor. N a represents the average of three experimental results, N 1 , N 2 , and N 3 are the result of three experiments, respectively. N i and N a are the measured and averaged result, respectively.

Conclusions
The presented work proved that the plasma-assisted catalytic for NO removal was an effective method, which could provide a new strategy for the removal of NO x from the diesel engine and it has huge potential for industrial applications in the future. The NO removal efficiency in the NTP system over Mn-Cu/ZSM5 catalysts was investigated, and the results indicate that the combination of NTP and Mn-Cu/ZSM5 catalyst promoted NO removal efficiency significantly. In addition, the influence parameters of voltage, frequency, energy density, and O 2 and NH 3 concentration on the NO removal efficiency were explored. The detailed experimental results are shown as follows: (1) The plasma-assisted SCR process exhibited an obvious improvement in NO removal efficiency, and 10% Mn-8% Cu/ZSM5 shows the highest catalytic activity with about 93.89% at an energy density of 500 J L −1 , and the selectivity to N 2 was almost 99%. (2) The voltage and frequency were positively correlated with the power of the NTP system. The NO removal efficiency increased with the voltage, frequency, and energy density increasing. (3) The O 2 concentration was negatively correlated with the NO removal in the NTP system. With the O 2 concentration increasing, the NO removal efficiency decreased significantly under the same energy density. (4) The NH 3 concentration was related to the NO removal closely, and the NO removal efficiency could not be improved when the NO removal process reached reaction equilibrium in the NTP system.
Author Contributions: All authors have contributed in various degrees to the analytical methods used, to the research concept, to the experiment design, to the acquisition of data, or analysis and interpretation of data, to draft the manuscript or to revise it critically for important intellectual content. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors confirm that this article content has no conflict of interest.