The Co-N-C Catalyst Synthesized With a Hard-Template and Etching Method to Achieve Well-Dispersed Active Sites for Ethylbenzene Oxidation

Biomass obtained from organic residues gradually becomes one of the optimal renewable feedstock of value added chemicals. Herein, the Co-N-C catalyst was prepared via a hard-template and etching method using the casein as C and N sources, magnesium oxide as the template, and cobalt porphyrin as the metal precursor. The obtained Co-N-C catalyst exhibited excellent catalytic performance for selective oxidation of ethylbenzene with a conversion rate of 96.5% under mild conditions. Moreover, the catalysts were investigated by techniques such as BET, XRD, Raman, transmission electron microscopic (TEM), and X-ray photoelectron spectroscopy (XPS). The results showed that the etching progress could improve the dispersion of Co and the exposure of active sites. Herein, the efficient oxidation of ethylbenzene was attributed to the well-dispersed Co-N species and the increased specific surface area.


INTRODUCTION
The catalytic conversion of carbon feedstock like petroleum into high value chemicals has been of great importance in the chemical industry (Yang et al., 2015;Den et al., 2018;Pliekhov et al., 2018). And one of the most important chemical reactions in the chemical industry is the selective oxidation of aromatic hydrocarbon. Over the past decades, the precious-metal catalyst has accounted for the major share of the market of selective oxidation of aromatic hydrocarbon (Rosser et al., 2016;Wang et al., 2017). However, the ever-rising cost limited the industrial production and large-scale application of the precious metal catalysts (Chung et al., 2017;Wei et al., 2017). To date, an enormous amount of effort has been devoted to exploiting the noble-metal-free catalysts. Through long-term development, transition metals (Co, Fe, Ni), and nitrogen co-doped carbon materials (M-N-C) have broad application in the hydrogenation and oxidation reactions (Cheng et al., 2015;Liu et al., 2017;Jiang et al., 2018). Among them, a series of Co-N-C catalysts has been investigated in selective oxidation of aromatic hydrocarbon and exhibit excellent catalytic performance (Gutmann et al., 2013;Jie et al., 2017). At present, the high temperature pyrolysis method has been the main way to synthesize the Co-N-C catalysts. Generally, the dispersion and the size of active sites and the specific surface area are the key factor for catalytic performance. Nevertheless, Co species, the active sites of the catalyst, were more likely to aggregate during the SCHEME 1 | Schematic diagram of process to prepare M-Co-N-C-X catalysts.
pyrolysis, which had a great effect on the catalyst performance. In order to tackle the problem, it was of great necessity to synthesize a kind of Co-N-C catalyst with high dispersion clusters as active sites.
Herein, cobalt porphyrin (CoTPP), 1-butyl-3methyllimidazolium chloride and casein were chosen as metal, carbon and nitrogen precursors, and the magnesium hydroxide (Mg(OH) 2 ) were introduced as the pore former . The detailed preparation process was shown in the Scheme 1. Generally, the 1-butyl-3-methyllimidazolium chloride was suitable to be used in the conversion of casein, which was rich in nitrogen, into porous carbon materials (Jia et al., 2017;Ding et al., 2019). The ionic liquids acted as a reaction medium and a porosity-directing regulator (Lee et al., 2010;Guo et al., 2013;Zhang et al., 2015c;Martinaiou et al., 2018). At the same time, Mg(OH) 2 could be used to prevent the aggregation of metal particles and enlarge the specific surface areas when used as hard templates Liu W. G. et al., 2016;Liu X. B. et al., 2016. Furthermore, the unique Co-N 4 coordination structure of the CoTPP could enhance the dispersion of metal nanoparticles and provide plenty of Co-N species Chen et al., 2018). Thought the hard-template and etching method, the catalyst containing the highly dispersed active sites with nano-size was prepared and showed the excellent catalytic performance in the selective oxidation of ethylbenzene. In addition, the active sites of the catalysts and the effect of acid etching on catalytic activity were explored by BET, XRD, Raman, transmission electron microscopic (TEM), and X-ray photoelectron spectroscopy (XPS).

Preparation of Cobalt (II) Porphyrin
The cobalt porphyrin was prepared as illustrated in the reported method. A mixture of benzaldehyde (0.05 mol, 5.6 g) and propanoic acid (180 mL) was added to 500 mL three-necked flask and heated to 130 • C with stirring continuously, and then freshly distilled pyrrole (4.7 g) was added dropwise and kept the mixture refluxed for an hour. After the mixture cooled down to room temperature, 100 mL deionized water was inserted to the flask and then refrigerated overnight. The obtained precipitate was washed, filtered and purified by column chromatography to obtain the purple powder and denoted 5, 10, 15, 20tetraphenylporphyrin (TPP). Subsequently, the TPP (1.0 g) was dissolved in 100 mL N, N-dimethyl-formamide (DMF) and heated to reflux, and then CoCl 2 ·6H 2 O (2.5 g) was added until the porphyrin was exhausted. Following that, the production was diluted with deionized water (100 mL) and then refrigerated overnight. The product was washed with deionized water until the filtrate became neutral and then dried at 80 • C for 12 h. The brown powder was obtained and denoted as CoTPP.

Preparation of the Catalysts
The 1-butyl-3-methyllimidazolium chloride (BmimCl), casein and Mg(OH) 2 were used to synthesize Co, N co-doped carbon materials, which served as carbon source, nitrogen source and hard template, respectively. In this typical synthetic procedure, BmimCl (2.0 g) was added to a 50 mL round flask and heated to 120 • C with constantly stirring. And then, casein (0.4 g) was dissolved into the ionic liquid and obtained a homogeneous solution. Subsequently, CoTPP (0.1 g) and Mg(OH) 2 (0.4 g) were mixed and then dispersed in the solution. The black viscous liquid was transferred to a quartz boat and then heated to 500 • C in nitrogen atmosphere at a ramp rate of 5 • C/min. The tube furnace was insulated for 2 h at this temperature, and the obtained powder was labeled as M-Co-N-C. The sample treated with 2 mol/L HCl solution for different length of time was denoted as M-Co-N-C-X (X means the acid treatment time). Similar to the above progress, the catalyst prepared without Mg(OH) 2 or CoTPP was denoted as Co-N-C-X or M-N-C-X respectively. Beyond that, the M-Co-N-C-9 was recovered via readily centrifugation and then reused for several times to test the reusability and durability. The catalyst recycled for 5 times was donated as M-Co-N-C-9-R.

Catalyst Characterization
Nitrogen adsorption-desorption isotherms were measured on Nova 1000e apparatus from Quanta Chrome Instruments at 77 K. The samples were outgassed at 200 • C for 3 h prior to the measurements. In the relative pressure ranging from 0.05 to 0.98, the specific surface areas (S BET ), the mesoporous volume (V total ) and the pore size distribution (D P ) were calculated using the Brunauer-Emmett-Teller (BET) and the Barrett-Joyner-Halenda (BJH) formula. Raman was conducted on Mono Vista 2560 Spectrometer with a 532 nm (2.33 eV) laser. High-resolution transmission electron microscope (HRTEM, JEOL-2100F) operating at 200 kV was carried to measure the morphology of samples. The mapping was conducted to determine the local elemental composition. X-ray photoelectron spectroscopy (XPS) was measured on a PHI 5000 CESCA system (Perkin Elmer) using Al Kα radiation (1486.6 eV). The X-ray diffraction (XRD) analysis was performed on a Japan XRD-6100 analyzer using Ni-filtered Cu Kα radiation with a scanning angle (2θ) ranging from 10 • -80 • , operated at 50 kV and 10 mA. The mass spectrometry analysis was performed on GCMS-QP2010ultra.

Catalyst Test
To investigate the catalytic performance of the catalysts for selective oxidation of alkanes, the selective oxidation of ethylbenzene with TBHP as oxidant at 80 • C was conducted. Before the reaction test, the catalyst (15 mg) and substrate (1.0 mmol) were put into the reaction tube. Then deionized water (3 mL) and TBHP (3.5 mmol, 70 wt% in water) were added sequentially. The mixture was heated to 80 • C with magnetic stirring and kept for 12 h. After the reaction, ethyl acetate (9 mL) was added into the mixture to extract the filtrate and n-dodecane (100 µL) was added to the system as an internal standard. The obtained sample was quantitatively analyzed by using GC analysis. The catalyst was recovered, washed with ethanol and then dried in vacuum at 80 • C.

RESULTS AND DISCUSSION
Nitrogen Adsorption/Desorption Isotherms N 2 adsorption-desorption measurements operating at 77 K were performed to detect the samples. According to the Figure 1 and Figure S1, the N 2 adsorption-desorption isotherm and pore size distribution of Co-N-C-9, M-Co-N-C, and M-Co-N-C-X revealed the pore textural properties. The isotherm of the samples showed a relatively slow increase in the range of 0.05 < P/P 0 < 0.80, and remarkable growth after P/P 0 > 0.80, which indicated the N 2 adsorption-desorption isotherms were attributed as type II curves (Thommes et al., 2015). As shown in the Figure 1, it was evidently clear that the hard template and etching progress were successful to form the pore structure. From the pore size distribution obtained from the Barrett-Joyner-Halenda (BJH) method, it was found that the M-Co-N-C-9 had the concentrated distribution of mesopores structure. The detailed information of the specific surface area and total pore volume of the samples were displayed in Table 1. It was noted that the specific surface area and total pore volume of M-Co-N-C-X grew gradually with the etching time increasing, and those of M-Co-N-C-9 were larger than other samples. As a general rule, enlarging the specific surface area of catalyst was conducive to exposing more active sites and strengthening the diffusion of substrates into the reaction regions, which might improve the catalytic performance of the samples (Zhang et al., 2015b;Wan et al., 2016;Wu et al., 2019). However, when the etching time was longer than 9 h, the specific surface areas and total pore volumes of the M-Co-N-C-12 decreased and it possessed more complicated pore size distribution, which might be on account of the damage of mesoporous structure.

X-Ray Diffraction (XRD) Patterns Analysis
The crystalline structure of the M-Co-N-C-X (X = 3, 6, 9 and 12) was investigated by XRD. As shown in Figure 2

Raman Spectroscopy Analysis
Raman spectroscopy was selected to detect the graphitization process and structural defects of the catalysts (Figure 3). An obvious D band at ∼1,338 cm −1 associated with structure defects and a G band at ∼1,572 cm −1 corresponding to graphitic carbon were obtained in the Raman spectroscopy (Guo et al., 2015;Yang et al., 2016b). Compared with the M-Co-N-C, the value of I D /I G of the M-Co-N-C-X was higher. Moreover, the etching progress between carbon atoms and metal particles produced more defects with properly extending the time, which would promote the catalyst performance (He et al., 2017;Lin et al., 2018). In addition, it could be found that the D bands in M-Co-N-C-3, M-Co-N-C-6, and M-Co-N-C-9 were broader than M-Co-N-C and M-Co-N-C-12, suggested that the content of the doped N atoms in M-Co-N-C-3, M-Co-N-C-6, and M-Co-N-C-9 was higher than it in M-Co-N-C and M-Co-N-C-12. It might because the N atoms could change the electronic structure of the graphitic networks (Song et al., 2016).

Transmission Electron Microscopic (TEM) Analysis
The morphology and structure of the M-Co-N-C-X composites were examined by TEM, high resolution transmission electron microscopy (HRTEM) and the elemental mapping (Figure 4). As seen from the HRTEM and TEM images, the number and diameter of nanoparticles on the carbon layer gradually decreased as the etching time increased. When the etching time was 9 h, no metal nanoparticles could be detected, which was because the etching reaction could effectively remove large metal particles such as MgO, Co 3 O 4 , and so on. Compared with MgO, the special structure of cobalt porphyrins, in which the Co atoms were strongly coordinated with N atoms, can effectively immobilize the cobalt atoms on the carbon layer, thus the Co species can be retained. As shown in Figures 4D,E 2016a; Wang et al., 2018). The result was in good agreement with the XRD analysis. Furthermore, it could also find the (102) facet of Co 0 crystal with the d-spacing values of 0.43 nm in Figure 4E could also be found, but no nanoparticle in Figure 4F. Obviously, the large cobalt nanoparticles were etched and the remnants of cobalt particles were hard to be detected. In addition, elemental mapping analysis indicated that the existence of cobalt species and the uniform distribution of Co and N in M-Co-N-C-6, and revealed that the cobalt species have smaller scale, thus the effect of etching was confirmed. The high dispersion of active sites would be beneficial to the catalytic performance.

X-Ray Photoelectron Spectroscopy (XPS) Analysis
X-ray photoelectron spectroscopy (XPS) was used to further explore the surface chemical composition of M-Co-N-C-X. As shown in Figure 5, the elements of Co, C, N, O, and Mg could be found in the full survey spectra of M-Co-N-C-X, which also revealed the relative contents of these elements. As shown in the Figure 5A, the full survey spectra of M-Co-N-C-X revealed the existence of C, N, O, Co, and Mg, indicated the trend of different elements with the etching time increasing. XPS survey spectra revealed that the content of Mg and O dropped rapidly, while the Co and N grew gradually when the sample was treated with more processing time. From the high-resolution C 1s spectrum (Figure 5B), three peaks at the binding energies of 284.8, 286.6, and 288.6 eV were assigned to C=C, C-N and Frontiers in Chemistry | www.frontiersin.org  O-C=O, respectively (Su et al., 2014;Lin et al., 2016). The N 1s XPS spectrum ( Figure 5C) could be deconvoluted into two peaks: pyridinic N (N1, 398.6 eV), pyrrolic N (N2, 400.5 eV), respectively (Wu et al., 2014;Fu et al., 2015;Sun et al., 2017). And the Figure 5D presented the Co 2p XPS spectrum, where there were two broad binding energy peaks located at 779.5 and 795.5 eV corresponding to the Co 2p3/2 and Co 2p1/2 binding energy, respectively (Chao et al., 2014;Aijaz et al., 2015;Wu et al., 2017). The high-resolution of Co 2p3/2 could be fitted into Co-O (780.3 eV) and Co-N (782.6 eV). It was worth noting that it might be that Co-O or Co-N played a decisive role in in the reaction process. Among the element composition of the M-Co-N-C and M-Co-N-C-6 as displayed in Table 2, it could be found that the content of O and Mg got a tremendous decline (54 and 99%), and the content of Co also had the obvious decline (16.5%), which indicated that the magnesium species and cobalt species, especially for magnesium oxide and cobalt oxides,  were easily removed. When the etching time was more than 6 h, the content of Mg in the samples remained substantially unchanged, indicating the Mg atom has been basically cleared. But, comparing M-Co-N-C-6 and M-Co-N-C-9, the content of Co and N increased significantly (35 and 33%), which may be because that the carbon layer was destroyed when the etching time over 6 h and large numbers of Co, N were exposed. As literatures illustrated, the element of N, acting as ligands, and metal ions may form the complex Co-N x , generated excellent catalytic performance for the reaction (Puello-Polo and Brito, 2007;Datsyuk et al., 2008;Wei et al., 2015;Chen et al., 2017). The content of Co-N progressively increased with the etching time increasing from 0 to 9 h, but the content of Co-O reduced. It might be ascribed to the structure of Co-N was more stable than Co-O under the etching progress. When the etching time arrived 12 h, the content of the Co-N decreased which was attributed to the destroyed of the structure of Co-N.

Catalytic Performance of M-Co-N-C-X
The catalytic performance of the as-prepared catalysts was tested in selective oxidation of ethylbenzene with TBHP as oxidant and water as solvent, because the efficient oxidation of ethylbenzene account for a highly important position in the chemical industry. The analytic conditions of GC and gas chromatogram of selective oxidation of ethylbenzene using M-Co-N-C-9 as catalyst were shown in Figure S4, and the detailed data of all samples were listed in Table 3. Obviously, the catalytic activity of the sample prepared without CoTPP (M-C-9) was well controlled with only 62.8% of ethylbenzene conversion ( Table 3, entry 1). Comparing M-C-9 and M-N-C-9 (Table 3, entry 1 and 2), the catalyst prepared via pyrolyzing porphyrin (TPP), casein and Mg(OH) 2 presented relative high catalytic activity, suggesting that the doped of N atoms took an important part in promoting the reaction. Remarkably, the M-Co-N-C-9 showed more excellent catalytic performance than M-N-C-9 (Table 3, entry 7 and 2), which illustrated that the Co atom was also critical to catalytic performance. However, the activity of Co-N-C-9 without utilizing Mg(OH) 2 as pore former sharply decreased ( Table 3, entry 3). And combined with the result of BET analysis (shown in Table 1), it could deduce that the higher specific surface area and total pore volume could improve the catalytic performance. To thoroughly investigate the influence of the etching, the catalysts (M-Co-N-C-X) were used to perform the same experiments ( Table 3, entry 4-8). Interestingly enough, the catalytic performance of the samples first increased and then decreased with the increase of etching time, but the M-Co-N-C-0 was the worst which indicated that the Co 3 O 4 (detected in Figure 2 of XRD and Figure 4D of TEM) was not the active species. In contrast, the M-Co-N-C-9 had the optimum performance. According to the XPS analysis (shown in Table 2), the variation regularity of the catalytic performance of M-Co-N-C-X was in line with the content of Co-N, indicating that the Co-N species were the active sites. Combined with the abovementioned characterizations and the experiments, it proved that the etching progress could expose more Co-N active sites which exerted decisive catalytic effect for the selective oxidation of ethylbenzene .

Applicability to Other Arylalkanes
The heterogeneous catalyst of M-Co-N-C-9 has been used to investigate the catalytic oxidation with a wide series of organic compounds containing C-H bonds with TBHP under similar conditions. The catalytic performance of the catalyst on these substrates displayed excellent results and some of the rules have been summed up. As shown in Table S1 (entries 2-4), the α-C connected with electrondonating or electronwithdrawing functional groups showed great activity during the catalytic oxidation process. In addition, for the p-substituted derivatives of ethylbenzene, the substituent played a decisive role in the reaction and displayed excellent conversion and selectivity to the corresponding ketones (Table S1, entries 5-6). It was amazing that, the selectivity of corresponding ketones of substrates containing two benzene rings reached up to 99%. In contrast, the catalytic activity for catalytic oxidation of naphthenic hydrocarbons was more moderate, because the C-H of cyclanes was inactive.

CONCLUSIONS
In summary, the noble-metal-free catalyst of M-Co-N-C-9 with extremely tiny metal particles has been successfully prepared through an effective hard-template and etching method. A high catalytic activity over M-Co-N-C-9 with 96.5% conversion of ethylbenzene and 97.3% selectivity to acetophenone was observed for oxidation of ethylbenzene with TBHP as oxidant under the mild conditions, which may explain that the doping of ionic liquid and the etching of MgO made a critical contribution to form the unique defective structure and enlarge the specific surface area, so that massive active sites were exposed to the reaction substrate. According to a series of characterizations, the Co-N species were considered to be the vital active sites for enhancing the catalytic performance.

DATA AVAILABILITY
All datasets generated for this study are included in the manuscript and/or the Supplementary Files.

AUTHOR CONTRIBUTIONS
CS and SJ designed and finished the synthesis and characterization analysis of materials and wrote the research paper. ZL and HC supervised the project, helped design the experiments, evaluated the data, and wrote the manuscript. The results of the manuscript were discussed by all authors.