Highly Dispersed Pt Nanoparticles on N-Doped Ordered Mesoporous Carbon as Effective Catalysts for Selective Hydrogenation of Nitroarenes

Highly-dispersed Pt nanoparticles supported on nitrogen-modified CMK-3 mesoporous carbon (Pt/N-CMK-3) were first fabricated by a two-step impregnation route. The influences of N content on the catalyst porous structure, Pt nanoparticle size, surface properties, and interaction between Pt species and the support were investigated in detail using N2 sorption, X-ray diffraction (XRD), transmission electron microscopy (TEM), and X-ray photoelectron spectra (XPS). The N species acted as anchoring sites for the stabilization of Pt particles. Benefiting from the formation of ultrafine metal nanoparticles, the Pt/N-CMK-3 exhibited excellent catalytic activity and selectivity for the selective hydrogenation of nitro aromatics to the corresponding anilines with hydrogen. The Pt/N-CMK-3 catalyst could be reused eight times and keep its catalytic performance.


Introduction
Substituted aromatic amines are crucial industrial intermediates for the production of various fine chemicals, such as dyestuffs, agrochemicals, pharmaceuticals, and polymers, and most of them are synthesized by catalytic reduction of corresponding nitro aromatics [1][2][3]. Particularly, sheterogeneous catalytic reduction over supported metal catalysts as an environment-friendly and efficient protocol attracts much interest [4,5]. The selective reduction of nitro aromatics over supported metal catalysts was widely adopted with different hydrogen sources, such as hydrazine hydrate, sodium borohydride, gas hydrogen, formic acid, ammonia borane, and so on [6-9]. Among them, H 2 , a low cost, non-toxic, and the cleanest hydrogen donor, is recognized as the most ideal reducing agent for the hydrogenation of nitro aromatic compounds in industrial production.
Heterogeneous noble metal-based catalysts such as Ru, Rh, Pt, and Pd have been reported to be efficient for the hydrogenation of nitroarenes [10,11]. Downsizing the noble metal particles to a few nanometers can dramatically improve their catalytic activity, due to the increasing surface-to-atom ratio [12,13]. However, the supported small metal nanoparticles often suffer from serious aggregation because of the high surface energy [14][15][16]. In addition, it remains challenging to keep other reducible groups, especially the halogen groups (F, Cl, Br, and I), intact at high conversion rates when using noble metal catalysts with H 2 as the hydrogen source [17][18][19]. Generally, the catalytic performance is closely linked to the metal particle sizes, the structure or surface properties of the support, and the interaction effect between the metal and support. The activity and selectivity of the catalyst for

Catalyst Characterization
N 2 adsorption-desorption isotherms and corresponding pore size distribution profiles of the CMK-3, Pt/CMK-3, and Pt/N-CMK-3-x samples were recorded. As depicted in Figure 1a, all of the Pt/N-CMK-3-x samples exhibited typical IV isotherms with distinct hysteresis at relative higher pressure (P/P 0 > 0.4), ascribing to the characteristics of mesoporous structure, similar to that of the CMK-3 support. This result demonstrated that the mesoporous structure of CMK-3 was still kept after being incorporated with Pt, C, and N components. The similar pore size distributions of the samples in Figure 1b also proved this point, and the pore sizes were primarily in the range of 3-4 nm. The detailed textural parameters and physical properties of the samples are shown in Table 1. With the increased amount of 2-methylimidazole, the specific surface area, pore volume, and pore size of the Pt/N-CMK-3-x showed a remarkable decline, implying the successful incorporation of C and N from the pyrolysis of 2-methylimidazole. The final contents of Pt and N in the samples were determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) and CHN elemental analyzer ( Table 1). As can be seen, the Pt contents were all located at 1.90 ± 0.05 wt%, while the N contents showed obvious increase from 2.1 to 3.5 wt% with increasing the amount of 2-methylimidazole. However, when the amount of 2-methylimidazole was continuously increased, the Pt/N-CMK-3-3 showed only a small increase in N contents about 0.2 wt%. Maybe the excess 2-methylimidazole did not incorporate in the CMK-3 frameworks, due to N loss during the synthesis at 800 • C. determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES) and CHN elemental analyzer (Table 1). As can be seen, the Pt contents were all located at 1.90 ± 0.05 wt%, while the N contents showed obvious increase from 2.1 to 3.5 wt% with increasing the amount of 2-methylimidazole. However, when the amount of 2-methylimidazole was continuously increased, the Pt/N-CMK-3-3 showed only a small increase in N contents about 0.2 wt%. Maybe the excess 2-methylimidazole did not incorporate in the CMK-3 frameworks, due to N loss during the synthesis at 800 °C. The XRD patterns of the CMK-3 and Pt/N-CMK-3 materials are displayed in Figure 2. All samples showed broadened diffraction peaks at 2θ = 43 • , indicating the presence of graphitic carbon. For the Pt/CMK-3 catalyst, an obvious diffraction peak at 2θ = 39.8 • was assigned to (111) lattice planes of metal Pt (PDF 70-2431). When the N species were incorporated in the matrix, the diffraction peak for metal Pt became weak and wide. As the N content was further increased, the diffraction peak of Pt (111) in Pt/N-CMK-3-2 and Pt/N-CMK-3-3 was disappeared completely, indicating that the Pt species were highly dispersed in the N-CMK-3 matrix. These results revealed that the incorporation of N in the carbon matrix benefited the dispersion of Pt nanoparticles on the surface, due to the complexing and stabilizing effect.  The XRD patterns of the CMK-3 and Pt/N-CMK-3 materials are displayed in Figure 2. All samples showed broadened diffraction peaks at 2θ = 43 o , indicating the presence of graphitic carbon. For the Pt/CMK-3 catalyst, an obvious diffraction peak at 2θ = 39.8 o was assigned to (111) lattice planes of metal Pt (PDF 70-2431). When the N species were incorporated in the matrix, the diffraction peak for metal Pt became weak and wide. As the N content was further increased, the diffraction peak of Pt (111) in Pt/N-CMK-3-2 and Pt/N-CMK-3-3 was disappeared completely, indicating that the Pt species were highly dispersed in the N-CMK-3 matrix. These results revealed that the incorporation of N in the carbon matrix benefited the dispersion of Pt nanoparticles on the surface, due to the complexing and stabilizing effect.  All the samples displayed obvious order pore channels, similar to pure CMK-3 (not shown). For Pt/CMK-3, plenty of darker metal nanoparticles with average particle sizes of 5.5 nm were distributed on the support. As the N species were introduced, the average size of Pt nanoparticles was remarkably decreased to 2.9 nm in Pt/N-CMK-3-1, which was attributed to the significant role of N in anchoring metal particles. It can be seen from Table 1 that the Pt particle sizes of Pt/N-CMK-3 and Pt/N-CMK-3-1 counted by TEM were approximately equal to the mean sizes of Pt crystallites by XRD. When the N contents were continuously increased, the mean sizes of the Pt particles reached a minimum value of about 1.2 nm in Pt/N-CMK-3-2 catalyst. These results, in agreement with the XRD result in Figure 2, confirmed that the N content had significant effects on  All the samples displayed obvious order pore channels, similar to pure CMK-3 (not shown). For Pt/CMK-3, plenty of darker metal nanoparticles with average particle sizes of 5.5 nm were distributed on the support. As the N species were introduced, the average size of Pt nanoparticles was remarkably decreased to 2.9 nm in Pt/N-CMK-3-1, which was attributed to the significant role of N in anchoring metal particles. It can be seen from Table 1 that the Pt particle sizes of Pt/N-CMK-3 and Pt/N-CMK-3-1 counted by TEM were approximately equal to the mean sizes of Pt crystallites by XRD. When the N contents were continuously increased, the mean sizes of the Pt particles reached a minimum value of about 1.2 nm in Pt/N-CMK-3-2 catalyst. These results, in agreement with the XRD result in Figure 2, confirmed that the N content had significant effects on the sizes of metal Pt crystallites or particles formed, which will be analyzed in detail in subsequent XPS analysis. In addition, the high-angle annular dark-field scanning TEM (HAADF-STEM) and the elemental mapping images demonstrated that Pt and N species were uniformly dispersed in Pt/N-CMK-3-2 ( Figure 3e).
Catalysts 2020, 10, x FOR PEER REVIEW 4 of 13 the sizes of metal Pt crystallites or particles formed, which will be analyzed in detail in subsequent XPS analysis. In addition, the high-angle annular dark-field scanning TEM (HAADF-STEM) and the elemental mapping images demonstrated that Pt and N species were uniformly dispersed in Pt/N-CMK-3-2 ( Figure 3e). The valence states of Pt and N types of the prepared materials were characterized by XPS spectra. As depicted in Figure 4a, the Pt 4f XPS spectra of Pt/CMK-3 showed symmetric doublet peaks at the binding energy of 71.7 and 75.2 eV, respectively, which were attributed to Pt 0 . However, this binding energy value was higher than the binding energy of bulk Pt (4f7/2 = 71.2 eV) [28][29][30], indicating the interaction between Pt particles and the CMK-3. When the CMK-3 support was treated with 2-methylimidazole and the N species were introduced into the matrix, the binding energies of Pt 0 4f were shifted to higher values with increasing N content, implying the existence of an interaction effect between Pt and N species due to a charge-transfer. The N species in the Pt/N-CMK-3-2 was further identified by the N 1s XPS spectrum in Figure 4b. Three peaks at 397.5, 399.4, and 400.7 eV were assigned to pyridinic-type, pyrrolic-type, and graphitic-type N [31][32][33][34], respectively. It has been reported that the pyrrolic and pyridinic N sites act as anchoring sites for the stabilization of Pt particles and suppressing their agglomeration [35,36]. The relative peak area percentage of each type of N in Pt/N-CMK-3-x catalysts is listed in Table 2. The surface N compositions were close to those determined by the CHN elemental analyzer in Table 1. When the N content was increased, the relative peak intensity for graphitic N was strengthened, and in the meanwhile, the relative peak intensity for pyrrolic N was found to be decreased. However, the peak intensities for the pyridinic N showed no obvious change. Therefore, the total content of pyrrolic and pyridinic N for the Pt/N-CMK-3-2 showed a maximum value, due to the combination of the two factors of the increases in N contents and the decrease in pyrrolic N on the catalyst surfaces. The valence states of Pt and N types of the prepared materials were characterized by XPS spectra. As depicted in Figure 4a, the Pt 4f XPS spectra of Pt/CMK-3 showed symmetric doublet peaks at the binding energy of 71.7 and 75.2 eV, respectively, which were attributed to Pt 0 . However, this binding energy value was higher than the binding energy of bulk Pt (4f 7/2 = 71.2 eV) [28][29][30], indicating the interaction between Pt particles and the CMK-3. When the CMK-3 support was treated with 2-methylimidazole and the N species were introduced into the matrix, the binding energies of Pt 0 4f were shifted to higher values with increasing N content, implying the existence of an interaction effect between Pt and N species due to a charge-transfer. The N species in the Pt/N-CMK-3-2 was further identified by the N 1s XPS spectrum in Figure 4b. Three peaks at 397.5, 399.4, and 400.7 eV were assigned to pyridinic-type, pyrrolic-type, and graphitic-type N [31][32][33][34], respectively. It has been reported that the pyrrolic and pyridinic N sites act as anchoring sites for the stabilization of Pt particles and suppressing their agglomeration [35,36]. The relative peak area percentage of each type of N in Pt/N-CMK-3-x catalysts is listed in Table 2. The surface N compositions were close to those determined by the CHN elemental analyzer in Table 1. When the N content was increased, the relative peak intensity for graphitic N was strengthened, and in the meanwhile, the relative peak intensity for pyrrolic N was found to be decreased. However, the peak intensities for the pyridinic N showed no obvious change. Therefore, the total content of pyrrolic and pyridinic N for the Pt/N-CMK-3-2 showed a maximum value, due to the combination of the two factors of the increases in N contents and the decrease in pyrrolic N on the catalyst surfaces.

Catalytic Reaction
Nitrobenzene was first conducted as a model compound over the Pt/N-CMK-3-2 to optimize the reaction conditions, and the results are listed in Table 3. Firstly, the hydrogenation of nitrobenzene with H2 was investigated in different solvents. All the solvents like ethylacetate, toluene, ethyl ether, methanol, and ethanol showed excellent catalytic activity, but ethanol

Catalytic Reaction
Nitrobenzene was first conducted as a model compound over the Pt/N-CMK-3-2 to optimize the reaction conditions, and the results are listed in Table 3. Firstly, the hydrogenation of nitrobenzene with H 2 was investigated in different solvents. All the solvents like ethylacetate, toluene, ethyl ether, methanol, and ethanol showed excellent catalytic activity, but ethanol (nitrobenzene/ethanol = 4:1 (mmol mL −1 )) gave the highest nitrobenzene conversion (entries 1-7). Secondly, the results showed that the nitrobenzene conversion steadily increased with increasing H 2 pressure from 0.5 to 2 MPa, indicated that the H 2 pressure influenced the dissolution of H 2 in the solvent. The reaction rate did not change obviously when increasing H 2 pressure from 2 to 6 MPa (entry 5 and entries 10,11), implying the absence of a hydrogen transport limitation at high pressure. Finally, the nitrobenzene conversion continually increased with the temperature without loss of selectivity (100%) (entries [12][13][14]. Table 3. Catalytic performance of Pt/N-CMK-3-2 catalysts for hydrogenation of nitrobenzene a . Catalysts 2020, 10, x FOR PEER REVIEW 6 of 13 (nitrobenzene/ethanol = 4:1 (mmol mL −1 )) gave the highest nitrobenzene conversion (entries 1-7). Secondly, the results showed that the nitrobenzene conversion steadily increased with increasing H2 pressure from 0.5 to 2 MPa, indicated that the H2 pressure influenced the dissolution of H2 in the solvent. The reaction rate did not change obviously when increasing H2 pressure from 2 to 6 MPa (entry 5 and entries 10,11), implying the absence of a hydrogen transport limitation at high pressure. Finally, the nitrobenzene conversion continually increased with the temperature without loss of selectivity (100%) (entries 12-14). We compared the initial conversions of nitrobenzene as a model compound over Pt/N-CMK-3-x catalysts with H2 in ethanol for 10 min at 40 °C, and the results are summarized in Table 4. As can be seen, the CMK-3 and N-CMK-3-2 materials provided no activity (entries 1 and 2). When the CMK-3 was treated with 2-methylimidazole, the Pt/N-CMK-3-1 showed a nitrobenzene conversion of 8.0%, which was much higher than that of Pt/CMK-3 (3.9%). The Pt/N-CMK-3-2 with the minimum particle size showed the highest nitrobenzene conversion. The turnover frequencies (TOFs) for the Pt/N-CMK-3-2 exhibited high values of 18,819 h -1 . However, when the N content was further increased, the initial conversions of nitrobenzene of Pt/N-CMK-3-3 showed a decline. These variations matched well with the results of TEM. These results demonstrated that the activity of the Pt/N-CMK-3-x for the selective hydrogenation of nitrobenzene was significantly influenced by the N content inducing changes in the size of the Pt nanoparticles. We compared the initial conversions of nitrobenzene as a model compound over Pt/N-CMK-3-x catalysts with H 2 in ethanol for 10 min at 40 • C, and the results are summarized in Table 4. As can be seen, the CMK-3 and N-CMK-3-2 materials provided no activity (entries 1 and 2). When the CMK-3 was treated with 2-methylimidazole, the Pt/N-CMK-3-1 showed a nitrobenzene conversion of 8.0%, which was much higher than that of Pt/CMK-3 (3.9%). The Pt/N-CMK-3-2 with the minimum particle size showed the highest nitrobenzene conversion. The turnover frequencies (TOFs) for the Pt/N-CMK-3-2 exhibited high values of 18,819 h −1 . However, when the N content was further increased, the initial conversions of nitrobenzene of Pt/N-CMK-3-3 showed a decline. These variations matched well with the results of TEM. These results demonstrated that the activity of the Pt/N-CMK-3-x for the selective hydrogenation of nitrobenzene was significantly influenced by the N content inducing changes in the size of the Pt nanoparticles.
The scope of Pt/N-CMK-3-2 in hydrogenation of nitroarenes, a series of nitro compounds with diverse substituent groups, were tested under the optimized reaction conditions, and the results are summarized in Table 5. To our great delight, the Pt/N-CMK-3-2 exhibited high activity and selectivity for the hydrogenation of substituted nitroarenes. Apart from nitrobenzene ( Table 5, entry 1), the substituted nitrobenzenes having nonreducible groups like -CH 3 , -NH 2 , and CH 3 O-were also furnished with excellent yield (>99%) (entries 2-7). It has been reported that supported noble catalysts, such as Pt, Pd, Rh, etc., display poor chemoselectivity to the hydrogenation of the nitro group when halogen groups exist in the same molecule [37]. Herein, no obvious dehalogenation product was observed in the selective hydrogenation of the halogen-substituted nitroarenes (entries [8][9][10][11][12][13]. Moreover, other reducible groups such as -COOCH 2 CH 3 , -COOH, -CN, and -CHO on the nitrobenzene were also well tolerated to give the corresponding amines in high selectivity (entries [14][15][16][17]. Also, for heterocyclic nitroarenes containing N element, full conversion and high selectivity of >99.0% was achieved (entries [19][20][21][22]. In contrast, the hydrogenation of p-chloronitrobenzene over Pt/CMK-3 showed not only low catalytic activity but also poor selectivity (entry 23). The high chemoselectivity of the Pt/N-CMK-3-2 catalysts for the hydrogenation of nitroarenes to anilines was likely due to the higher reactivity of nitro group than other functional groups. The stability and reusability of the Pt/N-CMK-3-2 catalyst were further investigated by the hydrogenation of nitrobenzene at 40 • C. As presented in Figure 5, the Pt/N-CMK-3-2 exhibited a nitrobenzene conversion of 80.1% in the first cycle. After completion of the reaction, the catalyst was separated by filtration, washed with ethanol three times, and dried overnight at 60 • C. Then the recovered catalyst was directly used for the next run without any reactivation or purification. The nitrobenzene conversion remained at 78.7% for the eight runs, and the aniline selectivity was kept at 100%. After each cycling reaction, the Pt contents in the product solution were determined by ICP-AES. It was found that the solution hardly contained the Pt element (<1 ppm). As can be seen in Table 1, the total Pt content in the spent Pt/N-CMK-3-2 was 1.86 wt%, which was very close to that of before the reaction. It was demonstrated that the Pt/N-CMK-3-2 possessed good recyclability and has great potential for practical applications in the selective hydrogenation of nitroarenes in the future.

Catalyst Preparation
Ordered mesoporous silica SBA-15 was obtained using P123 as a structure directing agent and TEOS as the silica source under acidic conditions according to the document [38]. CMK-3 was  1 likely due to the higher reactivity of nitro group than other functional groups. The stability and reusability of the Pt/N-CMK-3-2 catalyst were further investigated by the hydrogenation of nitrobenzene at 40 °C. As presented in Figure 5, the Pt/N-CMK-3-2 exhibited a nitrobenzene conversion of 80.1% in the first cycle. After completion of the reaction, the catalyst was separated by filtration, washed with ethanol three times, and dried overnight at 60 °C. Then the recovered catalyst was directly used for the next run without any reactivation or purification. The nitrobenzene conversion remained at 78.7% for the eight runs, and the aniline selectivity was kept at 100%. After each cycling reaction, the Pt contents in the product solution were determined by ICP-AES. It was found that the solution hardly contained the Pt element (<1 ppm). As can be seen in Table 1, the total Pt content in the spent Pt/N-CMK-3-2 was 1.86 wt%, which was very close to that of before the reaction. It was demonstrated that the Pt/N-CMK-3-2 possessed good recyclability and has great potential for practical applications in the selective hydrogenation of nitroarenes in the future. 1.5 >99.9 99.9 2 likely due to the higher reactivity of nitro group than other functional groups. The stability and reusability of the Pt/N-CMK-3-2 catalyst were further investigated by the hydrogenation of nitrobenzene at 40 °C. As presented in Figure 5, the Pt/N-CMK-3-2 exhibited a nitrobenzene conversion of 80.1% in the first cycle. After completion of the reaction, the catalyst was separated by filtration, washed with ethanol three times, and dried overnight at 60 °C. Then the recovered catalyst was directly used for the next run without any reactivation or purification. The nitrobenzene conversion remained at 78.7% for the eight runs, and the aniline selectivity was kept at 100%. After each cycling reaction, the Pt contents in the product solution were determined by ICP-AES. It was found that the solution hardly contained the Pt element (<1 ppm). As can be seen in Table 1, the total Pt content in the spent Pt/N-CMK-3-2 was 1.86 wt%, which was very close to that of before the reaction. It was demonstrated that the Pt/N-CMK-3-2 possessed good recyclability and has great potential for practical applications in the selective hydrogenation of nitroarenes in the future. 2.0 >99.9 99.7 3 likely due to the higher reactivity of nitro group than other functional groups. The stability and reusability of the Pt/N-CMK-3-2 catalyst were further investigated by the hydrogenation of nitrobenzene at 40 °C. As presented in Figure 5, the Pt/N-CMK-3-2 exhibited a nitrobenzene conversion of 80.1% in the first cycle. After completion of the reaction, the catalyst was separated by filtration, washed with ethanol three times, and dried overnight at 60 °C. Then the recovered catalyst was directly used for the next run without any reactivation or purification. The nitrobenzene conversion remained at 78.7% for the eight runs, and the aniline selectivity was kept at 100%. After each cycling reaction, the Pt contents in the product solution were determined by ICP-AES. It was found that the solution hardly contained the Pt element (<1 ppm). As can be seen in Table 1, the total Pt content in the spent Pt/N-CMK-3-2 was 1.86 wt%, which was very close to that of before the reaction. It was demonstrated that the Pt/N-CMK-3-2 possessed good recyclability and has great potential for practical applications in the selective hydrogenation of nitroarenes in the future. The stability and reusability of the Pt/N-CMK-3-2 catalyst were further investigated by the hydrogenation of nitrobenzene at 40 °C. As presented in Figure 5, the Pt/N-CMK-3-2 exhibited a nitrobenzene conversion of 80.1% in the first cycle. After completion of the reaction, the catalyst was separated by filtration, washed with ethanol three times, and dried overnight at 60 °C. Then the recovered catalyst was directly used for the next run without any reactivation or purification. The nitrobenzene conversion remained at 78.7% for the eight runs, and the aniline selectivity was kept at 100%. After each cycling reaction, the Pt contents in the product solution were determined by ICP-AES. It was found that the solution hardly contained the Pt element (<1 ppm). As can be seen in Table 1, the total Pt content in the spent Pt/N-CMK-3-2 was 1.86 wt%, which was very close to that of before the reaction. It was demonstrated that the Pt/N-CMK-3-2 possessed good recyclability and has great potential for practical applications in the selective hydrogenation of nitroarenes in the future. The stability and reusability of the Pt/N-CMK-3-2 catalyst were further investigated by the hydrogenation of nitrobenzene at 40 °C. As presented in Figure 5, the Pt/N-CMK-3-2 exhibited a nitrobenzene conversion of 80.1% in the first cycle. After completion of the reaction, the catalyst was separated by filtration, washed with ethanol three times, and dried overnight at 60 °C. Then the recovered catalyst was directly used for the next run without any reactivation or purification. The nitrobenzene conversion remained at 78.7% for the eight runs, and the aniline selectivity was kept at 100%. After each cycling reaction, the Pt contents in the product solution were determined by ICP-AES. It was found that the solution hardly contained the Pt element (<1 ppm). As can be seen in Table 1, the total Pt content in the spent Pt/N-CMK-3-2 was 1.86 wt%, which was very close to that of before the reaction. It was demonstrated that the Pt/N-CMK-3-2 possessed good recyclability and has great potential for practical applications in the selective hydrogenation of nitroarenes in the future. The stability and reusability of the Pt/N-CMK-3-2 catalyst were further investigated by the hydrogenation of nitrobenzene at 40 °C. As presented in Figure 5, the Pt/N-CMK-3-2 exhibited a nitrobenzene conversion of 80.1% in the first cycle. After completion of the reaction, the catalyst was separated by filtration, washed with ethanol three times, and dried overnight at 60 °C. Then the recovered catalyst was directly used for the next run without any reactivation or purification. The nitrobenzene conversion remained at 78.7% for the eight runs, and the aniline selectivity was kept at 100%. After each cycling reaction, the Pt contents in the product solution were determined by ICP-AES. It was found that the solution hardly contained the Pt element (<1 ppm). As can be seen in Table 1, the total Pt content in the spent Pt/N-CMK-3-2 was 1.86 wt%, which was very close to that of before the reaction. It was demonstrated that the Pt/N-CMK-3-2 possessed good recyclability and has great potential for practical applications in the selective hydrogenation of nitroarenes in the future.

Catalyst Preparation
Ordered mesoporous silica SBA-15 was obtained using P123 as a structure directing agent and TEOS as the silica source under acidic conditions according to the document [38]. CMK-3 was prepared using SBA-15 as the template and sucrose as the carbon source, and then was carbonized at 900 • C for 6 h under nitrogen, as described by Ryoo et al. [39]. The silica template was removed using 5 wt% hydrofluoric acid aqueous solutions at room temperature. The CMK-3 product was obtained after filtering, washing, and drying. The N-CMK-3 was prepared by an impregnation method with the following steps: 6 g of CMK-3 and a certain amount of 2-methylimidazole were dissolved in 60 mL deionized water at room temperature; the amount of 2-methylimidazole was 1, 2, and 3 g in the synthesis of N-CMK-3-1, N-CMK-3-2, and N-CMK-3-3, respectively; then, the mixture was stirred at 60 • C and the water was vaporized slowly; and finally, the obtained solid was calcined in N 2 atmosphere at 800 • C for 6 h with a heating rate of 2 • C min −1 .
Pt/N-CMK-3-x (x represents the amount of 2-methylimidazole with 1, 2, and 3 g) catalyst with 2 wt% of Pt was synthesized by ultrasound-assisted traditional wetness impregnation method. In brief, 6 g of N-CMK-3-x powder and 6.2 mL H 2 PtCl 6 ·6H 2 O aqueous solution (0.1 mol L −1 ) were mixed with deionized water (60 mL) to form a homogeneous suspension under ultrasound conditions. Then, the mixture was stirred at 40 • C until water was evaporated. Finally, the obtained solid was calcined in a flow of 30 vol.% H 2 in N 2 at 200 • C for 3 h with a ramp rate of 2 • C min −1 . For comparison, Pt/CMK-3 was prepared by the identical route.

Catalyst Characterization
Nitrogen adsorption analysis was carried out at liquid nitrogen temperature (−196 • C) by using an ASAP2020 analyzer (Norcross, GA, USA). Prior to measurement, samples were degassed at 200 • C for 10 h. The specific surface areas of the samples were calculated by the Brunauer-Emmett-Teller (BET) method. The pore volume was calculated at relative pressure P/P 0 of 0.991. The pore size distribution plot was calculated using the Barrett-Joyner-Halenda (BJH) formula. The metal Pt loading amount of the catalysts was determined on ICP-AES (Waltham, MA, USA). The N content in the catalysts was measured by using the PerkinElmer 2400 CHN elemental analyzer (Waltham, MA, USA). XRDpatterns were performed on a Rigaku D/MAX-2200 (Billerica, MA, USA) apparatus with a Cu Kα source (40 kV, 40 mA) at room temperature in the 2θ range of 10-90 • . Transmission electron microscopy (TEM) and HAADF-STEM micrographs were obtained on a JEM-2010F (JEOL, Beijing, China) with an element energy-dispersive X-ray spectrometer operating at 200 kV. XPS of the catalysts were recorded with an ESCALAB 250xi spectrometer (Thermo Fisher Scientific, Waltham, USA) equipped with Al Kα radiation source (hν = 1486.6 eV). Binding energies of all elements were calibrated by C1s at 284.6 eV.

Catalytic Reaction and Product Analyses
The chemoselective hydrogenation of nitro aromatics was carried out in a 100 mL stainless-steel autoclave with a stirring controller. In a typical experiment, the autoclave was charged with 80 mmol of nitro aromatics, 40 mg of catalysts, and 20 mL of green solvent ethanol. Before starting the reaction, the reactor was flushed three times with 0.5 MPa of hydrogen to remove the air, and then sealed tight and pressurized to 2 MPa H 2 . The stirring speed was kept at 800 rpm. Then the hydrogenation reaction was proceeded at 40 • C for a certain time. After the reactor was cooled to room temperature, the remaining H 2 was carefully released. One hundred microliters of the mixture were isolated by filtration for further analysis.
Each reaction was repeated more than three times to reach the carbon balance of more than 98% and obtained the nitroarenes conversion with an error ascertained to be within 5%. The qualitative and quantitative analyses of the products were done by gas chromatography (GC)-mass spectrometry (GC-MS, Shimadzu GCMS-QP 2010 Plus, Shanghai, China) and GC (Varian CP-3800, Palo Alto, California, USA) with n-decane as the internal standard.
For the recycling study, the hydrogenation reaction was performed in the same reaction conditions as mentioned above. The catalyst after each run was filtered, and washed several times with ethanol, and dried at 60 • C. Then the recovered catalyst was directly used for the next run without any reactivation or purification. Considering the catalyst loss during the filtration, washing, and drying, the amount of catalyst changed throughout each cycle. However, the Pt/N-CMK-3-2/nitrobenzene/ethanol ratio was always kept the same as it was in the first cycle.

Conclusions
In summary, highly-dispersed Pt nanoparticles supported on nitrogen-modified CMK-3 mesoporous carbon were successfully synthesized by a facile two-step impregnation route. TEM results revealed that ultrafine Pt nanoparticles were uniformly dispersed on the N-doped mesoporous carbon. The prepared Pt/N-CMK-3-2 was found to exhibit much higher catalytic activity for the hydrogenation of various nitro aromatics as compared to the Pt/CMK-3 prepared without the incorporation of N species. The existence of N species in carbon matrix facilitated high metal dispersion and prevented the agglomeration of Pt nanoparticles, due to the interaction effect between Pt and N atoms, which resulted in high catalytic activity. In addition, the Pt/N-CMK-3-2 could completely transform various substituted nitro aromatics to the corresponding aromatic amines with excellent selectivity, even for the case of the halogenated nitrobenzene. The Pt/N-CMK-3-2 catalyst was highly stable and could be reused for the selective hydrogenation of nitrobenzene without obvious loss of catalytic performance.