Enhanced Oxygen Vacancies in a Two-Dimensional MnAl-Layered Double Oxide Prepared via Flash Nanoprecipitation Offers High Selective Catalytic Reduction of NOx with NH3

A two-dimensional MnAl-layered double oxide (LDO) was obtained by flash nanoprecipitation method (FNP) and used for the selective catalytic reduction of NOx with NH3. The MnAl-LDO (FNP) catalyst formed a particle size of 114.9 nm. Further characterization exhibited rich oxygen vacancies and strong redox property to promote the catalytic activity at low temperature. The MnAl-LDO (FNP) catalyst performed excellent NO conversion above 80% at the temperature range of 100–400 °C, and N2 selectivity above 90% below 200 °C, with a gas hourly space velocity (GHSV) of 60,000 h−1, and a NO concentration of 500 ppm. The maximum NO conversion is 100% at 200 °C; when the temperature in 150–250 °C, the NO conversion can also reach 95%. The remarkable low-temperature catalytic performance of the MnAl-LDO (FNP) catalyst presented potential applications for controlling NO emissions on the account of the presentation of oxygen vacancies.


Introduction
Nitrogen oxides (NO x ) emitted from automobile exhaust gas and the industrial combustion of fossil fuels are air pollutants [1][2][3]. The selective catalytic reduction (SCR) of NO x with NH 3 is currently the most effective method to control NO x emissions [4]. Supported Mn-based catalysts have attracted widespread attention because of the large amount of free oxygen on their surface-this plays a key role in the field of low-temperature SCR reactions. There are many choices of supports for Mn-based catalysts at low temperature, including Mn/zeolite [5]. Mn/activated carbon [6,7], Mn/TiO 2 [8,9], Mn/Al 2 O 3 [10], Mn/ZrO 2 [11], and Mn/Vermiculite (VMT) [12]. In general, a suitable catalyst carrier should have unique properties, including strong surface acids, a high specific surface area, high thermal stability, and strong mechanical strength [13].
Mn-based mixed metal oxides (MMOs) have also attracted much attention. Studies of metal oxides as SCR catalysts have suggested that carrier-free metal oxides still offered good denitrification

Catalyst Preparation
The traditional coprecipitation (CP) preparation of Mn-Al hydrotalcite-like compounds used Mn(CH 3 COO) 2 ·4H 2 O as the Mn source and Al (NO 3 ) 3 ·9H 2 O as the Al source dissolved into distilled water to obtain the metal precursor solutions. Here, the Mn:Al molar ratio was 1:1, and the content of Mn was about 29 wt %. Next, a 1 mol/L NaOH solution was slowly added as the precipitant until the samples were pH = 9. The resulting slurry was aged at room temperature for 8 h and washed with deionized water until the pH reached neutrality; it was then dried at 80 • C. This sample was named MnAl-LDH (CP). The hydrotalcite precursor was then calcined at 550 • C in air for 6 h and it was termed MnAl-LDO (CP).
The FNP approach also used Mn(CH 3 COO) 2 ·4H 2 O and Al (NO 3 ) 3 ·9H 2 O, dissolved into distilled water to obtain the metal precursor solutions (solution A); a 1 mol/L NaOH precipitant served as solution B. Solutions A and B were injected into the impinging stream reactor via the injection needle of the FNP equipment [31,32]. The liquid flow rate was adjusted over several iterations to control the pH of the product at 7. The resulting slurry was aged at room temperature for 8 h, and washed with deionized water until the pH was neutral. The samples were then dried at 80 • C. The product was named MnAl-LDH (FNP). This hydrotalcite precursor was calcined at 550 • C in air for 6 h and termed MnAl-LDO (FNP). The preparation schematic diagram is shown in Figure 1a.

Material Characterization
The specific surface area and pore size distribution of the catalyst were measured by a US ASAP2020C (Micromeritics Instrument Co., Norcross, GA, USA) N 2 adsorption-desorption instrument. The catalyst was first vacuum degassed at 200 • C, cooled to room temperature, and then subjected to N 2 desorption experiments at −196 • C. The particle size distribution was measured by a laser particle analyzer (NanoPlus-3, NanoPlus, Gerbrunn, Germany). X-ray diffraction (XRD) was used to characterize the crystal form of the catalyst metal oxide. The instrument model was a Bruker D8 ADVANCE (Bruker Biosciences Co., Billerica, MA, USA), and the ray source was Cu-Kα. The test conditions were at 40 mV and 40 mA, with a scanning range of 10-90 • . H 2 temperature programmed reduction (H 2 -TPR) (temperature-programmed reduction) was used to determine the reduction performance of the catalyst. The instrument used was Micrometic ASAP 2720 Multisorbent (Micromeritics Instrument Ltd., Norcross, GA, USA). The test conditions were as follows: the catalyst was purged with nitrogen for 20 min at 100 • C, and then it was fed with a 10% H 2 /Ar gas mixture at a flow rate of 38 mL/min. The thermal conductivity detector automatically collected data from room temperature to 650 • C. The transmission electron microscopy (TEM) using a JEOL-2010F electron microscope (JEOL, Tokyo, Japan), operated at 200 kV. We used X-ray photoelectron spectroscopy (XPS) to analyze the electronic states and the atom amounts on the sample surfaces using a Kratos AXIS Ultra DLD (Kratos Analytical Inc., Manchester, UK). The binding energy of C 1 s of 284.8 eV was taken as a reference to the binding energies of the spectra.

Material Characterization
The specific surface area and pore size distribution of the catalyst were measured by a US ASAP2020C (Micromeritics Instrument Co., Norcross, GA, USA) N2 adsorption-desorption instrument. The catalyst was first vacuum degassed at 200 °C, cooled to room temperature, and then subjected to N2 desorption experiments at −196 °C. The particle size distribution was measured by a laser particle analyzer (NanoPlus-3, NanoPlus, Gerbrunn, Germany). X-ray diffraction (XRD) was used to characterize the crystal form of the catalyst metal oxide. The instrument model was a Bruker D8 ADVANCE (Bruker Biosciences Co., Billerica, MA, USA), and the ray source was Cu-Kα. The test conditions were at 40 mV and 40 mA, with a scanning range of 10-90°. H2 temperature programmed reduction (H2-TPR) (temperature-programmed reduction) was used to determine the reduction performance of the catalyst. The instrument used was Micrometic ASAP 2720 Multisorbent (Micromeritics Instrument Ltd., Norcross, GA, USA). The test conditions were as follows: the catalyst was purged with nitrogen for 20 min at 100 °C, and then it was fed with a 10% H2/Ar gas mixture at a flow rate of 38 mL/min. The thermal conductivity detector automatically collected data from room temperature to 650 °C. The transmission electron microscopy (TEM) using a JEOL-2010F electron microscope (JEOL, Tokyo, Japan), operated at 200 kV. We used X-ray photoelectron spectroscopy (XPS) to analyze the electronic states and the atom amounts on the sample surfaces using a Kratos AXIS Ultra DLD (Kratos Analytical Inc., Manchester, UK). The binding energy of C 1 s of 284.8 eV was taken as a reference to the binding energies of the spectra.

Activity Measurement
A micro-structured fixed bed reactor evaluates the catalytic performance of the powdered catalysts in the NH 3 -SCR reaction at the lab-scale. The NH 3 -SCR catalytic activity tests of the synthesized catalysts were carried out at atmospheric pressure in a fixed-bed stainless steel reactor with an internal diameter of 10 mm. The stainless steel reactor was installed in a vertical split-tube furnace. For each test, 0.1 g catalyst was charged. A typical composition of the simulated flue gas was as follows: 500 ppm NH 3 , 500 ppm NO, 5 vol % O 2 , N 2 as the balance gas, and the total volume flow was 100 mL/min. Its SCR reaction performance was evaluated at different temperatures under flowing simulated flue gas after a 30 min initialization period. The initial NO concentration in the simulated flue gas was 500 ppm, and this was referred to as [NO] in . The output gases were evaluated every 50 • C from 50 to 400 • C by Fourier transform infrared spectroscopy (FTIR) (Thermofisher IS10, Thermo Fisher Scientific, Waltham, MA, USA), and the NO content was [NO] out . The NO conversion was calculated from the following equations: The N 2 selectivities were calculated according to the input concentrations of NO x and NH 3 , and the output concentrations of N 2 O, NO 2 , NO x , and NH 3 as follows:

Results and Discussion
Transmission electron microscopy (TEM) images of the samples are shown in Figure 1b,c. These data show that the MnAl-LDO (FNP) catalyst particles were uniform and of micro-diameter distribution. Figure 1d shows that the average particle diameter of the as-obtained catalyst was smaller (114.9 nm) than MnAl-LDO (CP) catalyst (294.1 nm)-this was because of fast settling during preparation. Figure 1e shows the N 2 adsorption-desorption isotherms and the Barrett-Joyner-Halenda (BJH) pore size distribution curves of MnAl-LDO (CP) and the MnAl-LDO (FNP) catalysts. The adsorption isotherms have a well-defined H 2 -type hysteresis loop at a p/p 0 close to saturation-this is characteristic of the plateau/inflection of the mesoporous materials. The pore size distribution of the samples showed two sharp and obvious peaks that were located at 13 nm and 7 nm for MnAl-LDO (CP) and MnAl-LDO (FNP) catalysts, respectively [33]. The Brunauer-Emmett-Teller (BET) surface area, pore size, and pore volume are slightly reduced, as shown in Table 1. The high resolution TEM (HR-TEM) images of these catalysts (Figure 2a,b) showed abundant irregular spots on the surface of the MnAl-LDO (FNP) catalyst. This was the pore space. The MnAl-LDO (CP) catalyst had an obvious black shadow because of slight agglomeration. To study particle composition, HR-TEM images with a 5 nm gauge are shown in Figure 2c [12]. In addition, there is no evidence of an XRD-detectable Al-related oxide phase in any of the catalysts, which suggests an amorphous character for alumina, and/or an Al-bearing mixed oxide phase.   [12]. In addition, there is no evidence of an XRD-detectable Al-related oxide phase in any of the catalysts, which suggests an amorphous character for alumina, and/or an Al-bearing mixed oxide phase.  [12]. In addition, there is no evidence of an XRD-detectable Al-related oxide phase in any of the catalysts, which suggests an amorphous character for alumina, and/or an Al-bearing mixed oxide phase.  The elementary oxidation states and surface compositions of the samples were investigated by XPS. Figure 3b shows that the Mn 2p spectra of the samples were fitted with the Mn 2p 3/2 and Mn 2p 1/2 of MnO x . Previous research has shown that peaks at 642.7 eV and 653.4 eV were Mn 3+ from Mn 2 O 3 , the peaks at 644.3 eV and 655.1 eV were Mn 4+ of MnO 2 , and the peaks at 641.5 eV and 652.3 eV were attributed to Mn 2+ . As shown in Table 2, The Mn 3+ content of both samples was similar, but the Mn 4+ /Mn ratio in MnAl-LDO (FNP) was higher than in MnAl-LDO (CP). Previous reports proved that a large amount of Mn 4+ improved NH 3 adsorption, and the strong interaction between Mn 3+ and Mn 4+ amplified catalytic activity in the SCR reaction at low temperature. Exposure to Mn ions could increase oxidation capacity, and provide more reactive oxygen species.  The O 1s spectra of both catalysts were fitted into three peaks: The peak at 530.9 eV corresponded to lattice oxygen, the peak at 531.4 eV was related to the adsorbed oxygen, and the peak at 532.4 eV was attributed to the surface oxygen. Previous reports have shown that in low-temperature SCR reactions, free NH 3 is absorbed on the lattice oxygen and then activated into an amino group. This generates intermediate products [34,35]. Therefore, the increased lattice oxygen content could enhance the oxidation capacity on the surface of the catalyst and promote a "fast SCR" reaction. Here, the relative lattice oxygen concentration of MnAl-LDO (CP) and MnAl-LDO (FNP) catalysts were 30.42% and 35.44%, respectively (Table 2). Similarly, adsorbed oxygen could be converted to lattice oxygen, and the enhanced adsorbed oxygen content could accelerate the SCR reaction cycle. This would improve the catalytic activity, and the O ads could promote oxidation of NO into NO 2 , and consequently facilitating the "fast SCR" reaction [10].
During the formation of composite metal oxides, Al 2 O 3 supported a large number of active sites and it improved the specific surface area. Figure 3d shows that the Al 2p diffraction peak of the MnAl-LDO (FNP) catalyst shifted to lower binding energies compared with the MnAl-LDO (CP) catalyst. Previous reports have suggested that the binding energies decreased, due to the interactions with transition metals. These interactions enhanced the number of oxygen vacancies, and when forming a MnAl layered double oxide, the lattice distortion between Mn and Al could promote formation of oxygen vacancies [36,37].
Catalyst surfaces with oxygen vacancies are more active and they promote adsorption near the oxygen vacancies on the surface [38]. Metal vacancies can also tune the electronic structure of the surface and improve the catalytic activity, because of their electron and orbital distributions. The presence of metal vacancies in the catalyst reaction increased the valence state of nearby metal centers [39,40]. When an oxygen vacancy was formed in a metal oxide, the two electrons were singled out, and they could be transferred to neighboring metal ions, reducing the valence of the metal ions [41]. So, the MnAl-LDO (FNP) catalyst had better NO conversion.
The surface redox property was a key characteristic of SCR catalysts. Figure 4 shows the H 2 -TPR profiles of MnAl-LDO (CP) and MnAl-LDO (FNP) catalysts. The reduction curves of the MnAl-LDO (CP) catalyst showed a distinct reduction peak at 180-550 • C. This could be fitted with three peaks: the peak at 280 • C represented MnO 2 to Mn 2 O 3 , the peak at 349 • C represented Mn 2 O 3 to Mn 3 O 4 , and the peak at 450 • C was the further reduction of Mn 3 O 4 to MnO, agreeing well with the results of XPS analysis [42,43]. Generally, the curves corresponded to the amount hydrogen uptake over the catalysts, and the peak locations mean how easy the catalyst species are reduced. Moreover, the peak area in the profile of MnAl-LDO (FNP) catalyst was larger than that in the profile of MnAl-LDO (CP) catalyst. The enhancement was mainly due to the synergistic effect between the Mn and Al, which can create oxygen defects and oxygen vacancies, which suggested that the mobility of surface oxygen was enhanced and beneficial to the SCR reaction [44]. Meanwhile, the strong reduction performance of MnAl-LDO (FNP) catalyst offered richer active sites on the surface. All these features indicated that the modification with the new method and rich oxygen vacancies could greatly increase the redox ability of the MnAl-LDO (FNP) catalyst, which was favorable to enhance its SCR activity. can create oxygen defects and oxygen vacancies, which suggested that the mobility of surface oxygen was enhanced and beneficial to the SCR reaction [44]. Meanwhile, the strong reduction performance of MnAl-LDO (FNP) catalyst offered richer active sites on the surface. All these features indicated that the modification with the new method and rich oxygen vacancies could greatly increase the redox ability of the MnAl-LDO (FNP) catalyst, which was favorable to enhance its SCR activity. The catalytic performance of both catalysts at a GHSV of 60,000 h −1 is shown in Figure 5. The NO conversion of MnAl-LDO (FNP) was 48% at 50 °C (MnAl-LDO (CP) was only 24%). At 250 °C, the NO conversions of both catalysts increased gradually and reached a maximum at 250 °C, and then they decreased with increasing temperature. The MnAl-LDO (FNP) catalyst had a broad temperature window of 100-400 °C, in which NO conversion was over 80%. This window was only 200-300 °C for MnAl-LDO (CP). Table 3 details a two-dimensional MnAl-layered double oxide prepared via FNP. It had remarkable NO conversion. Figure 5b shows the N2 selectivity; when the temperature was below 250 °C , the N2 selectivity of MnAl-LDO (FNP) catalysts maintained 90%, and it then it sharply decreased, because of the production of NO2 and N2O. The MnAl-LDO (CP) catalysts had N2 selectivity that rapidly decreased above 100 °C, due to the high amounts of byproducts, because of the lower production of N2O and NO2. Figure 5c shows the N2O yield. The curve was on the upward trend, when the temperature was between 50-300 °C, the N2O generation increases with fastest rate, when the temperature between 300-400 °C, the N2O generation decreases. Figure 5d shows the NO2 yield; when the temperature between 50-250 °C, the curve leveled off, NO2 formation changed a little, and were the lowest at 200 °C. When the temperature was between 250-400 °C, the curve was on the rise, and going up very fast, reaching the highest at 400 °C. 250 °C was equivalent to a cut-off point; the temperature was less than 250 °C, and the NO2 generation was very low; when the temperature was higher than 250 °C, the NO2 generation increased significantly. The catalytic performance of both catalysts at a GHSV of 60,000 h −1 is shown in Figure 5. The NO conversion of MnAl-LDO (FNP) was 48% at 50 • C (MnAl-LDO (CP) was only 24%). At 250 • C, the NO conversions of both catalysts increased gradually and reached a maximum at 250 • C, and then they decreased with increasing temperature. The MnAl-LDO (FNP) catalyst had a broad temperature window of 100-400 • C, in which NO conversion was over 80%. This window was only 200-300 • C for MnAl-LDO (CP). Table 3 details a two-dimensional MnAl-layered double oxide prepared via FNP. It had remarkable NO conversion. Figure 5b shows the N 2 selectivity; when the temperature was below 250 • C, the N 2 selectivity of MnAl-LDO (FNP) catalysts maintained 90%, and it then it sharply decreased, because of the production of NO 2 and N 2 O. The MnAl-LDO (CP) catalysts had N 2 selectivity that rapidly decreased above 100 • C, due to the high amounts of byproducts, because of the lower production of N 2 O and NO 2 . Figure 5c shows the N 2 O yield. The curve was on the upward trend, when the temperature was between 50-300 • C, the N 2 O generation increases with fastest rate, when the temperature between 300-400 • C, the N 2 O generation decreases. Figure 5d shows the NO 2 yield; when the temperature between 50-250 • C, the curve leveled off, NO 2 formation changed a little, and were the lowest at 200 • C. When the temperature was between 250-400 • C, the curve was on the rise, and going up very fast, reaching the highest at 400 • C. 250 • C was equivalent to a cut-off point; the temperature was less than 250 • C, and the NO 2 generation was very low; when the temperature was higher than 250 • C, the NO 2 generation increased significantly.  Before the SCR reaction, the analog flue gas had no NO2 and N2O. However, the output gas contained numerous byproducts with increasing concentrations. The N2O concentration rose sharply above 150 °C due to the following reaction: 4NO + 4NH3 + 3O2 → 4N2O + 6H2O [45]. Of note, the N2O concentration gradually decreased, due to the strong reduction property of NH3 at high temperature. At temperatures over 250 °C, the concentration of NO2 increased dramatically, and this could lead to the following: 2NO + O2 → 2NO2 and 2NH3 + 7O2 → 2NO2 + 3H2O [46]. In the presence of oxygen, NO is oxidized to NO2, and then NO2 is reduced in the presence of a reductant to N2, and the reaction rate would be increased if NO2 was generated in the reaction system. This corresponds to the fast-SCR reaction shown in the equation. 4NH3 + 2NO + 2NO2 → 4N2 + 6H2O [47]. Furthermore, the content of NOx (NO + NO2 + N2O) was higher than the original NO concentration (500 ppm) at 350 °C, which suggested that the NH3 nitrogen source was oxidized to form the NO or NO2 at high temperature, as shown in Figure 5c,d.  Before the SCR reaction, the analog flue gas had no NO 2 and N 2 O. However, the output gas contained numerous byproducts with increasing concentrations. The N 2 O concentration rose sharply above 150 • C due to the following reaction: 4NO + 4NH 3 + 3O 2 → 4N 2 O + 6H 2 O [45]. Of note, the N 2 O concentration gradually decreased, due to the strong reduction property of NH 3 at high temperature. At temperatures over 250 • C, the concentration of NO 2 increased dramatically, and this could lead to the following: 2NO + O 2 → 2NO 2 and 2NH 3 + 7O 2 → 2NO 2 + 3H 2 O [46]. In the presence of oxygen, NO is oxidized to NO 2 , and then NO 2 is reduced in the presence of a reductant to N 2 , and the reaction rate would be increased if NO 2 was generated in the reaction system. This corresponds to the fast-SCR reaction shown in the equation. 4NH 3 + 2NO + 2NO 2 → 4N 2 + 6H 2 O [47]. Furthermore, the content of NO x (NO + NO 2 + N 2 O) was higher than the original NO concentration (500 ppm) at 350 • C, which suggested that the NH 3 nitrogen source was oxidized to form the NO or NO 2 at high temperature, as shown in Figure 5c,d.

Conclusions
The MnAl-LDO successfully prepared by a co-precipitation method and a flash nanoprecipitation method, as an excellent NH 3 -SCR catalyst. XRD, XPS, and TPR analyses demonstrated that a series of Mn-Al hydrotalcite were good precursor catalysts. The best catalysts, such as MnAl-LDO (FNP), showed much higher catalytic activity than the MnAl-LDO (CP) catalyst. The maximum NO conversions is 100% at 200 • C, and it had a narrow activity temperature window of 150-300 • C (NO conversion > 90%). MnAl-LDO (CP) exhibits more inferior NO conversion, and the highest conversion of NO is only 74.58% at 200 • C. MnO 2 is the main active components. In addition, the high concentrations of Mn 4+ should also be partly responsible for the good performance of the MnAl-LDO (FNP) catalyst, with rich oxygen vacancies also being beneficial to NO conversion. H 2 -TPR analyses indicated that MnAl-LDO (FNP) possesses a higher reducibility than the MnAl-LDO (CP) catalysts.