Aliphatic amines are widely applied in the production of pharmaceuticals, agrochemicals, surfactants, dyes and polymers^[1−2]. Selective catalytic hydrogenation of amides to the corresponding amines using the hydrogen molecular is an environmentally friendly route as the only byproduct is water. However, the hydrogenation of amides is very challenging because of the resonance stabilization caused by a delocali-zation of the nitrogen electronic lone pair. In addition, the hydrogenation of amides always produces more by-products like alcohols and N-substituted products^[3−4]. Therefore, it is a great challenge to develop efficient catalytic systems for hydrogenation of amides.

Homogeneous systems using metal complex and acidic or basic additives have been reported to be active for amide hydrogenation^[5−9]. However, it is difficulty in the product separation and catalysts recycling. While heterogeneous catalysts can avoid such difficulties and are more suitable for industrial applications. Recently, heterogeneous noble metal catalytic systems have been extensively studied in amide hydrogenation. Bimetallic catalytic systems which consist of a noble metal and an oxophilic metal such as Rh-Mo^[10], Rh-Re^[11], Ru-Re^[11], Pt-Re^[12−13], Pt-V^[14], Rh-V^[15], Ru-W^[16], Ru-Mo^[17−19] and Ir-Mo^[20] have been proved to be effective for amide hydrogenation. The use of precious metals will undoubtedly increase the cost for the hydrogenation process. And heterogeneous non-noble catalysts for amide hydro-genation were rarely reported. Shen et al. prepared 60%Ni/ LaAlSiO catalysts for the hydrodeoxygenation of N,N-dimethyl-formamide (DMF) to trimethylamine (TMA) in a fixed-bed reactor in 2016^[21]. Hu et al. recently reported that NiMo nitrides catalysts can facilitate the hydrogenation of aliphatic acyclic amides under the mixed atmosphere of H₂ and NH₃^[22]. In conclusion, the development of efficient non-noble hetero-geneous catalysts which allow for selective hydrogenation of amides is still highly demanded.

It has been reported that heterogeneous cobalt-based catalysts such as reduced Co₃O₄ or N-doped carbon supported Co catalysts have a strong ability to activate the C=O bonds in the hydrodeoxygenation of biomass-derived compounds^[23−26]. And the activity depends on the reducibility and dispersion of supported cobalt. We anticipate that the supported Co catalysts may be also applicable for the hydrogenation of aliphatic amides to amines. In this work, the nitrogen-doped ordered mesoporous carbon supported Co nanoparticles (Co/MNC) were prepared via an ion exchange-pyrolysis strategy^[27]. The supported Co catalysts pyrolyzed at 500 ℃ possessed the uniform dispersion of Co nanoparticles and the co-existence of Co⁰ and Lewis acidic CoO_x species, showed good catalytic activity and selectivity in hydrogenation of aliphatic amides, affording a maximum yield of cyclohexylmethylamine to 70.6% at 180 ℃ under 3 MPa H₂. Moreover, this catalytic system was available to primary, secondary and tertiary aliphatic amides.

1 Experimental Section 1.1 Catalyst preparation

Mesoporous nitrogen-doped carbon (MNC) was prepared according to the literature [27]. Generally, 3.08 g of 2,4-dihydroxybenzoic acid, 0.6 g of ethylenediamine, 0.934 g of hexamethylenetetramine and 3.5 g of P123 were dissolved in 80 mL of deionized water. And the mixed solution was stirred at room temperature for 2 h. Then, the above solution was transferred into a Teflon-lined stainless-steel autoclave. After keeping at 130 ℃ for 4 h, the autoclave was cooled down to room temperature, and the obtained orange-red polymer gel was washed three times with deionized water.

Nitrogen-doped ordered mesoporous carbon supported Co nanoparticles were prepared through an ion exchange-pyrolysis strategy^[27−28]. Typically, the orange-red polymer was first prepared according to the above-suggested method and was redispersed in a mixture of 96 mL of H₂O and 24 mL of ammonium hydroxide solution (28.0%~30.0%), followed by the addition of 1.345 g of Co(NO₃)₂·6H₂O. After stirring at 50 ℃ for 6 h, the suspension was centrifuged, and a brown solid was obtained after washing the precipitate with deionized water three times and drying at 50 ℃ under vacuum for 8 h. Then, the brown solid was treated under a H₂/N₂ (5%/95%) atmosphere at different temperatures for 2 h with a heating rate of 2 ℃∙min⁻¹. After that, 1% O₂/Ar was introduced to passivate the surface of catalysts. The obtained samples were labeled as Co/MNC-T, in which T indicates the reduction temperature. Cu/MNC and Ni/MNC were obtained through the similar preparation process when using Cu(NO₃)₂·6H₂O and Ni(NO₃)₂·6H₂O as the metal precursor.

1.2 Catalyst characterization

The X-ray powder diffraction (XRD) patterns were obtained by using a D8 ADVANCE X-ray with Cu Kα radiation (λ=0.154 05 nm). Transmission electron microscopy (TEM) was characterized on a JEOL 2100 electron microscope operated at 200 kV. Nitrogen adsorption-desorption isotherms were measured at −196 ℃ using a Micromeritics ASAP 2460 analyzer. The specific surface areas were calculated according to the Brunauer-Emmet-Teller (BET) method. The pore size distribution was determined by Barret-Joyner-Halenda (BJH) model. X-ray photoelectron spectroscopy (XPS) measurements were recorded on a Thermo Scientific K-Alpha working in the constant analyzer energy mode with Al Kα radiation as the excitation source. The carbonaceous C 1s line (284.8 eV) was used as the reference to calibrate the binding energies (B.E.). Fourier-transform infrared (FT-IR) experiments were conducted on Thermo Scientific Nicolet-iS50 FT-IR spectrometer, and the Raman spectra were carried out on Thermo Scientific DXR 3Xi with a 532 nm excitation wavelength. Inductively coupled plasma-Mass Spectrometry (ICP-MS) was conducted on NexION 350X.

1.3 Catalytic test

In a typical reaction, 50 mg of catalyst, 0.5 mmol of substrate and 3 mL solvent was added in a 10 mL high-pressure stainless-steel autoclave and purged with H₂ for three times at room temperature. After the autoclave was purged with H₂ to desired pressure, it was heated to 180 ℃ for 12 h. It was then let to cool down to room temperature and depressurized, and the reaction medium was centrifuged. Conversions and yields were determined by GC with dodecane as an internal standard. The GC sensibility was calibrated using the commercial products. The amide conversion (Conv.) and product selectivity (Sel.) were calculated with the follow equations:

$ \mathrm{C}\mathrm{o}\mathrm{n}\mathrm{v}.\left(\mathrm{\%},\mathrm{\ a}\mathrm{m}\mathrm{i}\mathrm{d}\mathrm{e}\right)=\frac{n\left(\mathrm{a}\mathrm{m}\mathrm{i}\mathrm{d}\mathrm{e}\mathrm{\ }\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{e}\mathrm{d}\right)}{n\left(\mathrm{a}\mathrm{m}\mathrm{i}\mathrm{d}\mathrm{e}\mathrm{\ }\mathrm{f}\mathrm{e}\mathrm{d}\right)}\times100 $

$ \mathrm{S}\mathrm{e}\mathrm{l}.\mathrm{ }(\mathrm{\%},\mathrm{\ C}\mathrm{y}\mathrm{C}\mathrm{H}_2\mathrm{N}\mathrm{H}_2) = \frac{n\left(\mathrm{C}\mathrm{y}\mathrm{C}\mathrm{H}_2\mathrm{N}\mathrm{H}_2\mathrm{\ p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{e}\mathrm{d}\right)}{n\left(\mathrm{a}\mathrm{m}\mathrm{i}\mathrm{d}\mathrm{e\ }\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{e}\mathrm{d}\right)} \times 100 $

$ \mathrm{S}\mathrm{e}\mathrm{l}.\left(\mathrm{\%},\mathrm{\ C}\mathrm{y}\mathrm{C}\mathrm{H}_2\mathrm{O}\mathrm{H}\right)=\frac{n\left(\mathrm{C}\mathrm{y}\mathrm{C}\mathrm{H}_2\mathrm{O}\mathrm{H}\mathrm{\ }\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{e}\mathrm{d}\right)}{n\left(\mathrm{a}\mathrm{m}\mathrm{i}\mathrm{d}\mathrm{e}\mathrm{ }\mathrm{\ c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{e}\mathrm{d}\right)}\times100 $

$ \begin{aligned} \mathrm{S}\mathrm{e}\mathrm{l}.\left( \mathrm{\%},\ \left( \mathrm{C}\mathrm{y}\mathrm{C}\mathrm{H}_2\right)_2\mathrm{N}\mathrm{H}\right) & = 2 \times \frac{n\left(\left(\mathrm{C}\mathrm{y}\mathrm{C}\mathrm{H}_2\right)_2\mathrm{N}\mathrm{H\ }\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{e}\mathrm{d}\right)}{n\left(\mathrm{a}\mathrm{m}\mathrm{i}\mathrm{d}\mathrm{e}\mathrm{ }\mathrm{\ c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{u}\mathrm{m}\mathrm{e}\mathrm{d}\right)} \times\\ & \;\;\;\;\;\;\;\;100\end{aligned} $

$ \begin{aligned}\quad\;\;\;\;\;\mathrm{S}\mathrm{e}\mathrm{l}.\left(\mathrm{\%},\mathrm{O}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{r}\mathrm{s}\right)& = 100-\mathrm{S}\mathrm{e}\mathrm{l}.\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }\mathrm{ }(\mathrm{\%},\mathrm{C}\mathrm{y}\mathrm{C}{\mathrm{H}}_{2}\mathrm{N}{\mathrm{H}}_{2})-\\ &\;\;\;\;\;\;\;\;\;\;\;\;\;\;\mathrm{S}\mathrm{e}\mathrm{l}.\left(\mathrm{\%},\mathrm{C}\mathrm{y}\mathrm{C}{\mathrm{H}}_{2}\mathrm{O}\mathrm{H}\right)-\\ &\;\;\;\;\;\;\;\;\;\;\;\;\;\;\mathrm{S}\mathrm{e}\mathrm{l}.\left(\mathrm{\%},{\left(\mathrm{C}\mathrm{y}\mathrm{C}{\mathrm{H}}_{2}\right)}_{2}\mathrm{N}\mathrm{H}\right) \end{aligned}$

2 Results and Discussion

Co/MNC-T was prepared by an ion exchange process, followed by pyrolysis for removing P123 and reducing cobalt partially (Fig. 1(a)). The structure of Co/MNC-T samples were characterized by X-ray diffraction (XRD). As for all samples, a board peak at 22° could be ascribed to C(002) diffraction (Fig.1(b)). On increasing the temperature from 500 to 700 ℃, the diffraction of Co(111) strengthened gradually, probably ascribing to the increased particle size. Subsequently, the nitrogen adsorption–desorption measurements for the Co/MNC samples were conducted to determine the pore structure and surface area. As shown in Fig.1(c), the N₂ adsorption/ desorption isotherms of all samples could be considered as type Ⅳ plots with a type H3 hysteresis loop starting from about P/P₀ = 0.45, indicating the samples have mesopores. The pore size distribution curve (Fig.1(d)) further demonstrated the dominant presence of mesoporous in these samples. The surface areas and pore sizes and volumes are summarized in Table 1. The BET surface area of the Co/MNC-500 is 464 m²∙g⁻¹ and the pore volume is 0.33 cm³∙g⁻¹. As the pyrolysis temperature increased, both the surface areas and pore volumes slightly decreased. The reason for the decrease of the surface area might be due to the partial collapse of the mesoporous structure under an extra high heating temperature.

Transmission electron microscopy (TEM) studies were carried out to investigate the morphology and particle size. The TEM image of the Co/MNC-500 sample showed uniform nanoparticles with an average size of about 3.30 nm on the support (Fig.1(j)). To identify whether the nanoparticles belong to certain Co species, a high-resolution TEM analysis (HR-TEM) of Co/MNC-500 was conducted. As shown in Fig.1(k), two distinct lattice planes with the lattice spacing of 0.203 and 0.213 nm, which could be ascribed to the Co(111) and CoO(200) planes, indicating that the co-existence of CoO and Co⁰. The scanning TEM (STEM) image and the corresponding energy-dispersive X-ray (EDX) elemental mapping images of the Co/MNC-500 sample are shown in Fig.1(f)−(i). The images clearly illustrated the uniform distribution of C, N, O, and Co elements on the sample. The Fig.1(j)−(o) showed that the average nanoparticle size increased with the enhanced pyrolysis temperature due to the aggregation of Co nanoparticles at higher pyrolysis temperature.

FT-IR measurements were carried out to check the functional groups present on the sample surface. The spectra of Co/MNC-T are displayed in Fig. 2(a). Both samples contained broad peaks with varying intensities at around 3 420 cm⁻¹, which could be assigned to the O―H stretching vibration. The stretching of the C=N group for both samples appeared at 1 590cm⁻¹. Raman spectroscopy was used to further investigate the graphitic character of the samples. As shown in Fig. 2(b), two intense peaks appeared at 1 369 and 1 600 cm⁻¹ for all samples which are assigned to the D band (defects in carbon matrix) and G band (in-plane stretching vibration of sp² hybridization of C atoms). For Co/MNC-T samples, the I_D/I_G ratio calculated based on the peak area are 0.66~0.76, demonstrating that higher pyrolysis temperature might lead to more defects. Furthermore, XPS was carried out to explore the elemental state of the series of catalysts. As shown in Fig. 2(c), Co 2p spectra are fitted into four peaks at 778.8, 780.5, 782.5, and 786.7 eV, which belong to Co⁰, Co³⁺, Co²⁺, and satellite peaks, respectively^[28−30]. The proportions of Co species in each catalyst are shown in Table 2. With the enhanced pyrolysis temperature, the proportion of Co³⁺ and Co²⁺ decreased, while the proportion of Co⁰ increased. Four peaks of N 1s are observed at 398.5, 400.0, 401.0, and 403.0 eV (Fig. 2(d)), which belong to pyridinic N, pyrrolic N, graphitic N, and NO, respectively. The relative abundance ratio of pyridinic N was decreased obviously as the pyrolysis temperature increased. Based on the above results, cobalt species were reduced during the pyrolysis process, thus resulting in the co-existence of Co⁰ and Lewis acidic CoO_x species on the surface of catalysts.

As shown in Table 3, the above Co/MNC catalysts were investigated for the hydrogenation of amides. Cyclo-hexanecarboxamide (CyCONH₂, 1a was chosen as a representative substrate for primary amides, and all the catalysts were evaluated at 180 °C and 3 MPa H₂. Cu/MNC and Ni/MNC were also prepared for the hydrogenation of amides. However, these two catalysts showed no activity towards the reaction. It is shown that Co/MNC-500 exhibits the best catalytic performance among all the investigated catalysts, with 92.4% conversion of CyCONH₂ and 76.5% selectivity of CyCH₂NH₂. However, further increasing the pyrolysis temperature resulted in 58.6% CyCONH₂ conversion for Co/MNC-600. These results indicate that the pyrolysis temperature plays a key role in CyCONH₂ hydrogenation to CyCH₂NH₂. Combined with results of TEM and XPS (Fig. 1(e)−(o), Fig. 2(d)), we conclude that uniform small size of Co nanoparticles and an appropriate ratio of Co⁰ and Lewis acidic CoO_x species are important for the hydrogenation of amides.

Since the Co/MNC-500 catalyst gave rise to the highest activity and selectivity, we then investigated the reaction stability and substrate applicability. As shown in Fig. 3(a), the conversion was around 92.4% and the selectivity to CyCH₂NH₂ was 76.5% for the first runs. However, the conversion decreased to 30.9% for the second runs. The loss of catalytic activity occurred to some extent during the repeated runs. When the spent catalyst was thermal treated in 5% H₂/N₂ at 500 ℃ for 1 h, the CyCONH₂ conversion recovered to 91.6%, indicating that the spent Co/MNC-500 could be regenerated through such a simple and reduction process. The Co leaching experiments were also investigated by ICP-MS. After each reaction, the content of Co in the filtrate was less than 1 mg·L⁻¹. Form the TEM pattern of reused Co/MNC-500 catalyst (Fig. 3(b)), there is no obvious change about the particle size of Co nano-particles, indicating the catalyst was stable during the reaction. In addition, it exhibited good activity and selectivity for a variety of primary, secondary and tertiary amides to the corresponding amines (Table 4).

3 Conclusion

Efficient nitrogen-doped ordered mesoporous carbon supported Co nanoparticles were prepared via an ion-exchange and pyrolysis strategy for the hydrogenation of amides. The optimal pyrolysis temperature of 500 ℃ under hydrogen atmosphere was found to result in the highest activity and selectivity, while both Co/MNC-600 and Co/MNC-700 were not efficient for amide conversion. Characterization showed that the suitable pyrolysis temperature could lead to uniform dispersion of small Co nanoparticles and the co-existence of both the Co⁰ and Lewis acidic CoO_x species over Co/MNC. The synergy between two cobalt species was responsible for the high reactivity of Co/MNC-500 in the hydrogenation of aliphatic amides.