All chemical agents were commercially available. They were used without further purification. The reactant 4-CBA were obtained from Aladdin Chemical Reagent (Shanghai, China). Palladium chloride was purchased from Sigma-Aldrich. Activated carbon was provided by Sinopec Shanghai Petrochemical Company Limited. Zinc nitrate hexahydrate, 2-methylimidazole, phosphoric acid (H₃PO₄, 85 wt% in H₂O), ammonia solution (NH₃ ▪ H₂O, 25 wt% in H₂O) were purchased from Sinopharm Chemical Reagent Co., Ltd.

Firstly, ZIF-8 was supported on AC according to the previous report with few modifications (Li J. et al., 2016). Typically, the activated carbon (C) was dispersed in the solution of 0.6 g Zn(NO₃)₂·6H₂O with 8 ml ethanol and 2 ml water. 1.31 g 2-methylimidazole dissolved in the mixture of 8 ml ethanol and 2 ml water. Then the solution of 2-methylimidazole was added into the above salt solution with stirring for 20 min. The mixed solution was transferred into a 100 ml Teflon-lined stainless autoclave. After sealed, the autoclave was heated to 120°C and maintained for 2 h. The hydrothermal product was filtered and washed thoroughly by water, and finally dried at 110°C. The above obtained samples (C@ZIF-8) were carbonized at different temperatures (500°C, 600°C, 700°C, and 800°C) for 2 h in nitrogen atmosphere. The carbonized samples were treated by HCl aqueous solution and washed by water. After drying at 120°C, the final samples were denoted as NC-T, where T represented the calcination temperature of 500°C, 600°C, 700°C, and 800°C.

Powder X-ray diffraction (XRD) measurement was carried out using Bruker D8 Advance diffractometer, with Cu Kα radiation (g = 0.15406 nm) at step scan 0.02° from 5° to 80°. The nitrogen adsorption/desorption isotherms were obtained at 77 K on a Micromeritics ASAP 2020 analyzer. The total porous volume (V_total) was estimated from the adsorbed capacity of nitrogen at a relative pressure P/P₀ of 0.98. The microporous volume (V_micro) was determined using the t-plot method. X-ray photoelectron spectroscopy (XPS) analysis was performed with an AXIS Ultra DLD photoelectron spectrometer (Kratos Analytical, United Kingdom) using monochromatic Al Kα radiation. The C1s peak at 284.6 eV was applied as reference to calibrate the position of the other peaks. Palladium dispersion on NC samples were tested by hydrogen titration of chemisorbed oxygen (H₂-O₂ titration) in a Micromeritics Autochem II 2920 instrument with a thermal conductivity detector (TCD). Before the tests, the Pd catalysts were firstly reduced under a hydrogen atmosphere at 200°C for 60 min. Typically, 400 mg catalyst sample was pre-reduced at 200°C by hydrogen for 120 min. After pre-reduction, hydrogen absorbed on the surface of catalyst was titrated by 15 pulses of oxygen until full saturation. Furthermore, the resulting monolayer of adsorbed oxygen was titrated using 15 pulses of hydrogen until constant TCD signal detected. The content of Pd loading was tested by an inductively coupled plasma emission spectrometer (ICP-OES, Varian 725ES). The hydrogen temperature programmed desorption (H₂-TPD) experiments was carried out using a Micromeritics Autochem II 2920 instrument equipped with a thermal conductivity detector (TCD). HRTEM images and elemental mapping of Pd catalysts were conducted on a Tecnai 20 S-TWIN electron microscope operated at 200 kV.

The Hydrogenation of 4-CBA was performed in a 100 ml stainless autoclave with a Teflon liner. Generally, 50 mg catalyst and 0.15 g 4-CBA were dispersed in 50 ml de-ionized water. After the autoclave was sealed, it was purged by hydrogen to remove the air. Then it was pressured with 0.8 MPa hydrogen. Finally, the autoclave was heated to 80°C for hydrogenation reaction. After the completion of this reaction, the obtained liquid products were treated with 14% NH₃·H₂O and 17% H₃PO₄ aqueous solutions respectively. Finally, the liquid products were analyzed through high-performance liquid chromatography (HPLC, Agilent 1,200). The HPLC was equipped with a UV detector and a C-18 column (4.6*250 mm, 5 μm). The hydrogenation of 4-CBA over Pd catalysts leaded to primary product 4-hydroxymethyl benzoic acid. The TOF (turnover frequency) of Pd catalysts was calculated as the number of moles of 4-CBA converted per mole of Pd per minute.

All of NC samples in present work were synthesized in similar procedure but at different calcination temperatures ranging from 500 to 800°C to tailor different N species. Firstly, XRD patterns of C@ZIF-8 and NC samples were shown in Figure 1. Before calciantion, the precursor C@ZIF-8 showed the diffraction of ZIF-8 (cubic space group), corresponding to the ZIF-8 (011) (002) and 112) diffractions, respectively (Zhang et al., 2014). After calcination at different temperature, four NC samples exhibited similar diffraction peaks at around 25° and 43°, corresponding to the graphitic carbon (002) and (100) diffraction peaks (Fu et al., 2017). Further, nitrogen adsorption-desorption isotherms of NC samples were tested and the structure parameters of these materials were summarized in Table 1. NC-500 exhibited the lowest BET surface area (678.1 m²/g). NC-600 showed higher BET surface area (701.6 m²/g) than that of NC-500. NC-700 and NC-800 have similar BET surface area with NC-600. These results demonstrated that nitrogen species of NC can be regulated by different pyrolysis temperature at 600–800°C, while similar surface area and porous volume of NC were preserved.

XPS characterization was performed to analyze N types and N contents of NC samples (Peng et al., 2014; Fu et al., 2017; Yang et al., 2018). As shown in Figure 2, the high-resolution N1s XPS spectra of four NC samples revealed there were mainly three types of N species: pyridinic N (N1) at 398.5 eV, pyrrolic N (N2) at 399.8 eV and graphitic N at 401.9 eV (N3) (Xiong et al., 2014; Chen et al., 2015). The typical nitrogen types in nitrogen doped carbon materials were shown in Scheme 1 (Li M. et al., 2016; Cao et al., 2017; Quílez-Bermejo et al., 2019). In Figure 3, the total N content decreased from 2.58 to 2.19 at.% with the increasing of calcination temperature. The relative contents of different N species in Figure 3 were calculate based on the integration areas of each N1s peak component from the N1s XPS spectra in Figure 2. The relative percentage of pyridinic N increased from 30.5% to 50.5%. Differently, the percentage of pyrrolic N decreased continuously with increasing calcination temperature. Similar results have been reported earlier for N doping carbon materials, which revealed that pyrrol functional groups were not stable at high temperature and partly transformed to pyridinic N (Arrigo et al., 2008; Sun et al., 2017).

The change of the ratio of Pd⁰/Pd²⁺ was related with the different N content and N types of supports NC (Figure 5). As mentioned in Figure 3, the N content of NC decreased along with the increase of calcination temperature from 500°C to 800 C. However, the catalytic activities of Pd/NC increased with the increasing of calcination temperature (Figure 6). The variation trend of TOF value of Pd/NC did not agree with the effect of N content for 4-CBA hydrogenation. Thus, the N content was not the primary factor to influence hydrogenation activity, but the moderate N content may be a positive effect on this hydrogenation reaction. It was noticed that the ratio of Pd⁰/Pd²⁺ in Pd/NC catalysts increased almost linearly along with the relative percentage of pyridinic N (Figure 5). The highest ratio of Pd⁰/Pd²⁺ showed the best hydrogenation performance. Owing to its electron donating properties, pyridinic N can enhance the interaction between Pd and NC through its lone pairs as metal coordination sites, preventing the reoxidation of Pd⁰. These results were in line with some experimental and DFT studies of pyridinic nitrogen strong interaction with metal nanoparticles (Li et al., 2016c; Nagpure et al., 2016; Jia et al., 2021). Besides, H₂-O₂ titration results indicated that Pd nanoparticles loading on NC supports maintained better dispersion compared with that of un-doped one.

To further clarify the important roles of nitrogen doping on catalytic activity of Pd/NC catalysts, the relationships of TOF of four Pd/NC catalysts and the ratio of Pd⁰/Pd²⁺ as well as Pd dispersion were depicted in Figure 7. A strong dependence of the TOF of Pd/NC catalysts with the ratio of Pd⁰/Pd²⁺ as well as Pd dispersion was observed. The higher the ratio of Pd⁰/Pd²⁺, the higher the catalytic activity of Pd/NC catalysts. Besides, the higher Pd dispersion also resulted in higher activity for Pd/NC catalysts. The above linear relationship may be attributed to the more exposure active sites Pd⁰ over Pd/NC.

It was well known that the hydrogen dissociation/activation ability has a profound effect on catalytic hydrogenation performance (Liu et al., 2016; Bi et al., 2019). The H₂ activation/dissociation ability of Pd catalysts were studied via H₂-TPD. As shown in Figure 8, all catalysts showed a similar hydrogen desorption peak at 322°C, indicating that similar type of active sites existed on Pd catalysts surface. In addition, there was an obvious desorption peak at 433°C for Pd/NC-700. Pd/NC-800 showed two desorption peaks at low temperature of 322°C and high temperature of 521°C. The differences of desorption peaks indicated there were different types of active Pd sites on Pd/NC-700 and Pd/NC-800. The desorption peak at low 322°C was assigned to chemisorbed hydrogen on surface Pd active sites, and the high temperature peaks at 433 and 521°C were assigned to strongly chemisorbed hydrogen on different types of Pd active sites. Furthermore, the total hydrogen desorption of Pd/NC-800 were obviously larger than that of other Pd catalysts. Therefore, H₂-TPD results further concluded that the enhanced activation capability of H₂ molecular in NPC-800 supported Pd catalyst also contributed to its higher hydrogenation activity. Furthermore, the morphologies of Pd NPs and the EDS elemental mapping of C, N, O and Pd of Pd/NC-800 was shown in Figure 9. It was evident that N and Pd were dispersed uniformly on the surface of Pd/NC-800. In addition, the used Pd/NC-800 catalyst was centrifuged and washed with ethanol for further TEM analysis. The particle size distributions of Pd in fresh and used Pd/NC-800 were shown in Figure 10. These results indicated that Pd nanoparticles in used Pd/NC-800 still maintained well dispersion and remained similar particle size distribution compared with that of fresh Pd/NC-800. In addition, Pd content in used Pd/NC-800 was also analyzed by ICP-OES technique. The results confirmed that the leaching of Pd was negligible. These results indicated Pd/NC-800 catalyst was quite stable. Therefore, the present NC materials can be promising catalyst supports to immobilize Pd nanoparticles (Li et al., 2013; Yu et al., 2021).