The catalytic performance of N-doped sp3@ sp2 hybrid nanocarbon
N-doped sp3@sp2 hybrid nanocarbon was prepared starting from nanodiamond (ND) which is annealed at 900°C with a nitrogen precursor. The scheme of preparation of this NC@ND-900 catalyst is presented in Fig. S1 and is characterized by a single layer discontinuous curved sp2 carbon layer on the surface (Fig. 1a). The N content is 3.7 at% (Fig. 1b) with four configurations including pyridinic N (N1, 398.5 eV), pyrrolic N (N2, 400.1 eV), graphitic N (N3, 401.1 eV) and N oxide (N4, 403.2 eV) 22. The homogeneous distribution of N species was further confirmed by energy-dispersive X-ray spectroscopy (EDS) mapping (Fig. S2) results.
The aerobic synthesis of 2-phenylbenzoxazole (P, target product) by o-aminophenol (A) and benzaldehyde (B) was chosen as model reaction (see top of Fig. 1) to explore the catalytic performance of ND-based nanocarbons catalysts. Without the catalyst (Fig. 1c), 96.1% of A could be exclusively converted into I (2-(benzylideneamino phenol, intermediate) with 99.8% of selectivity but the product P was undetected. The condensation reaction between the amino-group (A) and aldehyde-group (B) is occurring easily without catalyst, while the following dehydrogenation and annulation steps between the -OH group and imide C-H bond of I requires a catalyst. The pristine NDs (i.e. before N introduction) used as catalyst gives only traces of P (Fig. 1c). After the oxygen and nitrogen groups were introduced by HNO3/H2SO4 treatment (oND900) and followed by NH3 post-treatment (NoND900), respectively, the catalytic performance was not affected significantly. NC@ND-900 catalyst instead shows a remarkably enhanced formation of P (9.6%, structure confirmed by GC-MS in Fig. S3) compared to the NoND900, indicating the important role of N doping, but also the relevance of the methodology of its introduction. Note that substituting O2 in the reaction atmosphere with N2, no product P was detected, indicating that O2 is indispensable to the dehydrogenation step of I (Fig. 1c). To optimize the performances and understand better the catalytic behavior, the role of reaction temperature and time were further investigated (Fig. 1c and Table S1).The selectivity of P increases with reaction time and reaching the value of 97.4% at 28 h, which is comparable with the homogeneous or heterogeneous metal-based catalytic systems (Table S2). Notably, the stability of NC@ND-900 was investigated and it could be applied to 5 successive cycles with negligible loss of the catalytic efficiency (1st Yield P of 79.0% vs. 5th Yield P of 77.1%, Fig. S4). Given that similar catalytic performance of further purified NC@ND-900-HCl and extra Fe loaded Fe/NC@ND-900 samples to NC@ND-900 (Table S3), the possible contribution of residual metal impurities can be excluded (confirmed by ICP results in Table S4). This indicates that the intrinsic catalytic activity originates from the nanocarbon rather than the residual metal species.
To check the role of the different N species in the NC@ND-900 (see Fig. 1b), their amount was changed by varying the annealing temperature (Fig. S1). Four N-doped sp3@ sp2 hybrid nanocarbons were prepared and present decrease tendency of nitrogen contents (from 4.0 at. % for NC@ND-700 to 2.1 at. % for NC@ND-1300) and increase tendency of the sp2 carbon percentage through increasing the annealing temperature from 700 oC to 1300 oC (Figs 1d, S5 and Tables S5-S6). The yield of P increases from 21.5% to 27.7% as the synthesis temperatures raise from 700 °C to 900 °C and then decreased to 19.1% and 13.1 % as the temperatures increase to 1100 °C and 1300 °C, respectively (Fig. 1e), indicating the significant influence of the annealing temperature on the catalytic activity. The analysis of the relationship among different N species content, sp2 carbon percentage (derived from XPS results, Fig 1b and Table S5) and catalytic activity shows that a linear relationship is presented only between pyridinic N and sp2 carbon content normalized catalytic activity (Fig. S6) in the N-doped sp3@ sp2 hybrid nanocarbon series (Fig. 1g). This indicates a critical role of pyridinic N in the catalytic activity, but in relation to the sp2 carbon content of the catalyst.
Influence of sp2 carbon on the local structure of N species
To study the influence of sp2 carbon content on the local structure of N species, N-doped sp3@ sp2 hybrid nanocarbons with different sp2 carbon layers was synthesized by using different temperature annealed NDs (NDT, T=900 °C, 1300 °C and 1600 °C) as starting materials rather than the NDs (Fig. 2a and Fig. S1). Then, these samples were doped with N by using the same procedure for NC@ND-900. The obtained samples were referred as NC@ND-C1, NC@ND-C3 and NC@C9 when ND900, ND1300 and ND1600 were applied as starting material, respectively. Compared with the ND particles, which show typical (111) planes of diamond (Fig. 2b), one layer and 3-4 layers of curved sp2 graphitic-like shell was observed on the surface when ND900 and ND1300 was applied as precursors (Fig. 2c-2d), respectively. While the sp3 carbon have totally transformed into onion-like carbon when ND1600 was applied as precursor (Fig. 2e). The structure evolution of catalysts was also studied by the UV-Raman, as shown in Fig. 2f and Table S7, the spectrum of NDs exhibits a peak at 1622 cm-1 (overlap of G band at 1590 cm-1 and -OH bending vibrations at 1640 cm-1) and diamond sp3 mode at 1324 cm-1 30. After annealing at high temperature and introduction of N, the diamond peak (1324 cm-1) disappear while the D-mode peak centered at 1400 cm-1 appears. This peak originates from the breathing of hexagonal carbon rings with defects.30 Meanwhile, the G-mode peaks are observed and shifted from 1622 cm-1 to 1587 cm-1, indicating the transformation of sp3-hybridized C into sp2-hybridized C and the graphitization of sp2 carbon.30 The EELS results (Fig. 2g) also confirm this transformation since the intensity of π* peak (285 eV, for sp2-hybridized carbon) increased with the increasing annealing temperature and also increased slightly after the N-doping process.31 C1s spectra of XPS results (Fig 2h, Fig.S5 and Table S6) also present the same trend that the sp2/sp3 values of the samples increasing from 0.2 of NDs to 2.3 of ND1600. When the N species were introduced, the sp2/sp3 values increase from 0.5 of NC@ND-900 to 2.9 of NC@C9. These results clarify the evolution process of sp3 hybridization C transformed into sp2 hybridization C (detailed illustration in Fig. S7) and confirm N-doped sp3@ sp2 hybrid nanocarbons with different sp2 carbon layers were successfully synthesized.
The catalytic performance of above samples were investigated in the aerobic synthesis of 2-phenylbenzoxazole. As shown in Fig. 3a, around 7% yields of P was observed when the NDs was annealed under 900 °C, 1300 °C and 1600 °C (dark violet bars), suggesting the formed sp2 carbon layers in the N-free catalysts could contribute trace catalytic activity. For the N-doped samples (dark green bars), the yield of P increased from 27.7 % of NC@ND-900 to 39.2%, 42.0 % and 42.9 % of NC@ND-C1, NC@ND-C3 and NC@C9, respectively, indicating the active site role of N species (pyridinic N). By normalizing the reaction rate of P (λP) by the pyridinic N content (Fig. 3a), a clear enhancement of the value is observed on increasing the sp2 carbon layer up to about 3, while not further increasing when the sp2 carbon layers increase from 3 layers to ca. 9 layers (NC@C9). The results confirmed the contribution of sp2 carbon layers to the intrinsic catalytic activity of pyridinic N. To understand the role of inner sp3 C core, an equivalent sample to NC@ND-900 was prepared, but having as inner core SiO2 sphere (ca. 20 nm) (NC@SiO2). This sample shows 9.4% yield of P and when λP is normalized to pyridinic N, the value is very similar to that of NC@ND-900. This indicated that the sp3 C core has only the function of stabilizing and forming the surface sp2 C layer, although favouring forming more pyridinic N sites.
To address the role of the sp2 carbon layers in the enhancement of the catalytic activity, local work functions obtained from Kelvin probe force microscope (KPFM) measurements was used to evaluate the electron donating capability of N-doped samples (Fig 3b)32. Theoretically, the electron-donating capability of catalysts determines the interfacial charge transfer to the associated reactants, with a profound significance.27, 33 Fig. S8 shows the topographical and CPD (contact potential difference) images simultaneously taken on different samples. As shown in the Fig 3b, NC@ND-900 presents the highest work function (4.99 eV). When the sp2 carbon layers were increased, the work function of NC@ND-C1 and NC@ND-C3 decreased to 4.92 eV and 4.79 eV, respectively, which indicates their higher electron-donating capability. The NC@C9 presents similar work function compared to NC@ND-C3 even through NC@C9 contain more sp2 carbon layers, indicating that the further increases in the number of sp2 C layers do not influence further the work function and the interfacial charge transfer. Thus, the intrinsic catalytic activity of pyridinic N could be further improved by increasing sp2 carbon layer (up to about three) and the associated influence on the work function induces a change in the electron-donating capability.
To proof this conclusion, DFT calculations were made to investigate the enhanced catalytic activity on increasing sp2 carbon layers up to three (see Methods section for more details of calculation). Being the activation of O2 a key reaction step of the aerobic synthesis of P, attention was focused on the interaction between O2 and different model catalyst structures: a pyridine-N doped graphene coated on NDs (structure A), 2 layers graphene covered NDs (structure B) and 4 layers graphene (structure C) (Fig S9). Fig.3c and Fig S10 show the most stable optimized configurations of O2 molecule adsorbed on three model structures. We can clearly see that the most stable adsorption site for O2 is the more positively charged C atom that adjacent to the pyridinic N atom. The O2 adsorbed on structure B is more stable (-3.52 eV) than that of structure A (-2.75 eV), while similar with that of O2 adsorbed on structure C (-3.40 eV). The results indicate that, compared to the pure NDs, the activation of O2 undergoes an increase trend and then remains similar as increasing the sp2 carbon layers gradually.
To further understand the interaction between the catalysts and O2, the charge transfer of different systems were calculated by using Bader charge analysis (Fig S11). In all the analyzed systems, the electrons are transferred to the adsorbed O2, the charge transfer (Q) increasing from 0.78 of structure A to 0.89 of structure B , while slightly decreasing to 0.84 when further increasing the sp2 C layers (Fig.3c). Compared with structure A, the adjacent C atoms of pyridinic N in structure B and structure C both present more abundant electrons transferred from inner multilayers of graphene, which could promote electron transfer from the catalyst surface to O2 (Table S8). Additionally, from the calculated partial density of states (PDOS) in Fig. 3d, a stronger hybridization of the p orbital of the C atom that bind with the O2 and p orbital of the adsorbed O2 around the Fermi level in structure B and structure C with respect to structure A. This evidence further supports the mechanistic explanation of the interaction between O2 and catalyst models.
Mechanistic Study
The spin-trapping electron paramagnetic resonance (EPR) technique using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as the spin trap was performed to obtain information on the activation process of O234. As shown in Fig. 4a, the catalytic systems with mixture of substrate and catalyst or mixture of substrate and O2 remain EPR silent after adding DMPO. When substrate, catalyst and O2 were all presented, the EPR signal of the catalytic system shows the characteristic fingerprint of spin adducts DMPO−•OOH compared with the simulated DMPO−•OOH EPR signal, which resulting from trapping superoxide (O2•-) with DMPO.35 The results indicate the O2 activation by N-doped sp3@ sp2 nanocarbonsand the generation of reactive O2•- species, which could be responsible for the oxidative dehydrogenation steps during the synthesis of P. It is reasonable to speculate that the H2O2 would be formed as intermediate in the O2•- involved aerobic oxidation reactions. To proof this hypothesis, detection of H2O2 was made by using the catalyzed oxidation of N, N-diethyl-1, 4-phenylenediammonium sulphate (DPD) by Horseradish peroxidase (POD).36 The absorption peaks at around 510 and 551 nm confirm the formation of H2O2 during the reaction process (Fig. 4b). The EPR and UV-Vis results indicate that the N-doped sp3@ sp2 nanocarbon catalyzed aerobic synthesis of P was proceed via superoxide radical (O2•-) involved routine and H2O2 was formed as intermediate by hydrogen abstraction from the organic substrate directly or (in)directly. The activation process of I was also studied by ATR-IR (Fig. 4c and Fig. S12) and the red shift (from 3329 cm-1 to 3324 cm-1) of -OH group of I was observed in the presence of model catalyst. The results indicate I was activated by the catalysts and the -OH bond was weakened before the H abstraction step.
Based on the results above, a plausible reaction mechanism of intermolecular annulation reaction between 2-phenylbenzoxazole and benzylaldehyde could be proposed. As shown in Fig 4d, initially, the intermediate I forms by the condensation of o-aminophenol and benzylaldehyde. Then the formed I absorbs on the surface of catalyst and the reaction was initiated with a hydrogen abstraction step of the O-H bond from I by catalyst. Then the I* undergoes intramolecular cycloaddition of the imine to form the corresponding aminyl radical (P*). The reaction was followed by the second H abstraction from the P* due to the driving force for aromatization and ultimately generate the desired product (P). Meanwhile, after the O2 was absorbed on the surface of the catalyst, the electron transfer happened from the C atom that adjacent to the pyridinic N to the O2 to form O2•-, which subsequently react with the protons abstracted by the catalyst to form H2O2 and regenerate the active sites.
Substrate scope experiments
Having successfully achieved the aerobic oxidative synthesis of 2-phenylbenzoxazole, it is valuable to analyze how the catalytic system could be applied to the synthesis of other substituted benzoxazoles, benzothiazoles and benzimidazoles by using different substituted 2-aminophenol, 2-aminothiophenol or 2-phenylenediamine respectively with aldehydes as starting materials (Table 1). The reactions with benzaldehydes bearing electron-deficient (P2 and P3) and electron-rich groups (P4-P6) at the aromatic ring proceeds smoothly to give the desired products in good yields. When the cinnamaldehyde (P7) and 4-formyl-trans-stilbene (P8) were chosen as reactants, the desired products were produced with yields of 92.6% and 84.4%, respectively. 82.9% yield of P9 was achieved when the furfural was chosen to provide the aldehyde group. The aliphatic aldehydes were also successfully employed as substrates to give the corresponding products but with relatively lower yields (P10-P11). After the aldehydes screening, the scope of various 2-aminophenols was also investigated. Functional groups including chloro, nitro, and cyano were well tolerated under the optimal reaction conditions, and the desired products were obtained in moderate to good yields (P12-P14). The results above indicate that the reactions are insensitive to the electron-donating or election-withdrawing nature of the substitutes. A high yield of 73.3% was obtained when the benzene ring was replaced by a naphthyl group (P15). To further validate the general use of the proposed catalytic method, the 2-aminophenols was replaced with 2-aminothiophenol and 2-phenylenediamine for the synthesis of benzothiazoles and benzimidazoles. To our delight, the reactions of 2-aminothiophenol and 2-phenylenediamine with benzaldehyde, furfural and benzaldehydes bearing chloro, nitro and methoxyl groups proceed with good yield as shown in Table 1, P16-P23.
N-doped sp3@ sp2 hybrid nanocarbons are highly effective, stable and valuable metal-free carbocatalysts for the aerobic catalytic oxidative synthesis of various heterocycles including 2-substituted benzoxazoles, benzothiazoles, and benzimidazoles. They can be synthetized with high yields by one pot reaction of aldehydes with 2-aminophenole, 2-amonothiophenol and o-phenylenediamine, respectively. The correlation between catalysts characteristics and performances indicate that pyridinic N and neighboring C atoms are the active sites. The electron-donating capability of pyridinic N surface atoms, and the charge density of neighboring C atom can be modulated by increasing up to three sp2 C layers of N-doped sp3@ sp2 hybrid. Therefore, increasing the sp2/sp3 carbon ratio of the N-doped sp3@ sp2 hybrid nanocarbon could lead to a higher intrinsic catalytic activity. The mechanistic study well supports this indication and provides evidence that the carbon adjacent to pyridinic N species (with enhanced electron-donating capability) could activate O2 molecules to yield an oxygen anion radical that react with the abstracted proton on the surface of catalyst, then forming H2O2 and regenerating the catalyst simultaneously. It is the first time that it was demonstrated the possibility to enhance the intrinsic catalytic activity of pyridinic N by regulating the sp2 carbon layers of the catalysts. This study also demonstrates that the N-doped sp3@ sp2 hybrid catalytic system provides an attractive and useful methodology to achieve a green chemistry approach for the one-pot synthesis of important pharmaceutically related products.