X-ray diffraction (XRD, Bruker D8) analysis was performed to determine the crystalline phase of the samples. Hydrogen temperature programmed reduction (H₂-TPR, AutoChem1 II 2920) analysis was performed with a thermal conductivity detector (TCD) to study the reduction behavior of the catalyst precursor. In H₂-TPR, the prepared samples were pretreated by a programmed ramp-up at 10°C/min to 300°C dry under 30 ml/min Ar inert gas protection, kept purged for 1 h and then cooled to room temperature, followed by heating from 80°C to 800°C at 10°C/min in a 10% H₂/Ar mixture at 30 ml/min, and the amount of H₂ consumed was characterized by a thermal conductivity detector (TCD). X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha) was performed to analyze the metallic elements. The surface morphology and metal distribution of the catalysts were obtained using scanning electron microscopy (SEM, Zeiss MERLIN Compact). The metal particle size of the catalyst was obtained using transmission electron microscopy (TEM, JEM 2100F).

A simple impregnation method was used to prepare NiMo/AC catalysts. First, 1 g of activated carbon (AC) was dispersed in 10 ml of deionized water, and nickel chloride hexahydrate and ammonium molybdate tetrahydrate were added at a total mass fraction of 15% and stirred for 24 h. Then, the pH value was adjusted to eight and stirred for 3 h. The mixture was washed 3–5 times using a centrifuge to ensure that the supernatant was free of chloride ions. The cleaned samples were dried in a vacuum drying oven at 60°C. The samples were homogeneously ground with an agate mortar and then heated in a tube furnace at 10% H₂/N₂ for 3 h at a certain temperature to obtain the NiMo/AC catalyst.

In a 100 ml high-pressure reactor, 2 g of NEC, 18 ml of decalin and 0.4 g of catalyst were added. Argon was used to displace the gas in the reactor. After heating to 150°C, hydrogen gas was introduced to pressure the reactor to 8 MPa. The temperature and pressure data of the reactor were recorded by sensors on a computer. The reaction was stopped after 4 h regardless of the extent of the reaction. After the reactor was cooled to room temperature, the product was collected to measure the composition using gas chromatography, and then the amount of hydrogen uptake was identified.

The chemical composition and elemental profiles of the catalysts were obtained using SEM in combination with an EDS analyzer. The SEM images in Figures 1A, C show some morphological features of the Ni₄Mo/AC catalyst. The surface of the material was rough, which was a common feature observed in the SEM images of the carbon support. The EDS spectrum in Figure 1B shows the composition of the prepared materials. In the sample, C, Ni, and Mo were found to be present in large amounts. Both Ni and Mo had double peaks, and C had only a single peak, showing the relative amounts of these elements. C has the most intense peak, as it was the main component of the carrier (Naushad et al., 2019). The different colors in the mapping image (Figures 1D–F) of Ni₄Mo/AC showed the distribution of Ni and Mo in the prepared catalyst. The distribution revealed that Ni and Mo were homogeneously dispersed in different regions of the carbon-supported surface without obvious agglomeration or segregation. Such a uniform distribution also implied that an alloy of NiMo was formed (Wang et al., 2013). These better morphologies are rare in Ni-based catalysts made by gas reduction (Shanmuganandam and Ramanan, 2016; Wu et al., 2019b). In our preparation, proper metal loading based on the support and stable conditions contributed to this result together.

To examine the crystal morphology and structure, the prepared Ni₄Mo/AC samples were studied by TEM. Figures 2A, B shows TEM micrographs of the Ni₄Mo/AC catalyst. The images illustrate the particle size and metal dispersion of Ni₄Mo/AC. It is obvious that the metal particles of Ni₄Mo/AC exhibited high homogeneity, as shown in the SEM image of Figure 1B. The statistical analysis of particle size is given as an inset in Figure 2A, showing that for the Ni₄Mo/AC catalyst, the average particle size was approximately 9.75 nm within a narrow range. In addition, the high-resolution image (Figure 2C) showed a lattice spacing of 0.210 nm, which was consistent with 0.210 nm for the Ni₄Mo (111) facet (Wang et al., 2021). This indicated the formation of an ordered Ni₄Mo nickel-molybdenum alloy in the catalyst. The mapping image (Figures 2D-F) further demonstrated the uniform dispersion of Ni and Mo on the carrier surface and the formation of an alloy structure.

The powder XRD pattern of Ni₄Mo/AC is shown in Figure 3A. The diffraction peaks at 2θ values of 44.398°, 51.704°, and 76.264° corresponded to the (111), (200), and (220) facets of the Ni₄Mo alloy, respectively. The diffraction peak of Ni₄Mo was shifted to a small angle compared to the standard card of Ni (PDF#04-0850), which was due to the doping of molybdenum with a larger atomic weight than that of Ni. In addition, comparing the XRD patterns of Ni₄Mo/AC and Ni/AC (Figure 3A), it could be seen that the half-peak width becomes larger, showing that the doping of a certain amount of Mo reduced the crystallinity of the pure nickel phase. These results indicated that an ordered Ni₄Mo alloy was formed on the activated carbon after doping with a small amount of molybdenum.

Temperature-programmed reduction (TPR) was used to study the interactions between the active components of the loaded catalysts. To evaluate the interactions between nickel and molybdenum, Ni₄Mo/AC, Ni/AC and Mo/AC samples before reduction were subjected to H₂-TPR (Figure 3B). Under the same conditions, all metals could be reduced. For Mo/AC (Supplementary Figure S1), two major peaks were observed, one at 360°C and the other at 470°C, as an inverse peak, which was due to the reaction of the sample to form MoC during the temperature-programmed reduction (Qu et al., 2003). For Ni/AC, the maximum reduction temperature showed a broad band centered at 250°C and 350°C. In contrast, the H₂-TPR pattern for Ni₄Mo/AC could be observed with a peak at 340°C and a lap-shoulder peak at 250°C. The H₂-TPR of Ni/AC and Ni₄Mo/AC samples before reduction were analyzed after peak fitting. In the former sample, the reduction peaks at 260°C, 296°C, and 344°C were attributed to the progressive reduction of metallic nickel, and the reduction peak at 485°C was attributed to the reduction of surface and bulk oxygen in activated carbon (Le et al., 2017). In the latter samples, the reduction peaks at 224°C, 332°C, and 397°C were attributed to the reduction of metallic nickel, and the reduction peak at 437°C was attributed to the reduction of metallic molybdenum. The above split-peak results indicated that the reduction process of nickel is changed after doping with molybdenum metal. These results suggested that the mixing of Ni-Mo metal differs from the reduction process of elemental Ni overlying elemental Mo and that there was an interaction between the two metals, meaning that a new alloy was formed.

XPS was used to study the detailed information on the distribution of elements Ni and Mo. The XPS spectra (Supplementary Figure S2) confirmed that only the expected elements Ni, Mo, C, and O from metal oxidation were present in the Ni₄Mo/AC catalyst. The corresponding Ni2p spectra and Mo3d spectra of Ni₄Mo/AC are shown in Figures 3C, D, where 853.66, 856.37, 860.13, and 862.38 eV corresponded to Ni, Ni²⁺, Ni³⁺ and satellite peaks, 228.53 and 231.58 eV corresponded to Mo, and 232.97 and 236.02 eV corresponded to Mo⁶⁺ (Xiao et al., 2021). The possible reason for the presence of Ni and Mo oxides was the oxidation caused by the catalyst transfer process, in which it is easy to form an oxide layer on the surface owing to the active metal properties of nanosized Ni and Mo.

The catalytic performance of the prepared catalysts is shown in Figure 4 and Table 2. In Figure 4A, it can be seen that Mo/AC has no hydrogenation performance. Ni/AC had a lower hydrogenation activity with only 58.46% conversion and 23.6% selectivity for the final product at 4 h. In contrast, Ni₄Mo/AC had a high catalytic activity, and finally, the 4-h conversion reached 100%. The final product selectivity was 98.73%, and the hydrogen uptake amount reached 5.77 wt%. This indicated that the hydrogenation performance of the catalyst was greatly improved after doping with a certain amount of Mo, which has no activity in hydrogenation. As we mentioned above, the catalyst formed a Ni₄Mo alloy, which could significantly reduce the dissociation energy of hydrogen (Wang et al., 2021). This resulted in an enhancement of the contact within active hydrogen atoms and LOHC molecules, thus promoting the formation of C-H bonds in the N-heterocycles.

The reduction temperature of the loaded catalyst affects the morphology and particle size, which in turn affects the catalytic performance. The catalyst precursors were reduced for comparison at 400°C, 450°C, and 500°C, and a series of characterizations were performed. The three catalysts were also tested for the hydrogenation of NEC (Figure 4B). It can be realized that the effect of temperature is non-linear. The catalytic performance was best at a reduction temperature of 450°C, followed by 400°C and 500°C. Although the reduction temperature only differs by 50°C, the reactivity more than doubles. The activity of the catalyst prepared at 500°C is not as good as that of Ni/AC, while the catalyst prepared at 400°C is only similar to that of Ni/AC. Characterizations were carried out to analyze the cause of this phenomenon. XRD patterns (Figure 5) showed that the diffraction peaks of Ni were most significantly shifted to a small angle at a reduction temperature of 450°C, which indicated the best doping effect of molybdenum. The peaks in Ni₄Mo/AC-400°C did not exhibit a significant shift. The XPS patterns (Figure 6) of the three catalysts showed a slightly higher Ni content in the metallic state at a reduction temperature of 450°C. However, the valence distribution of Mo is more significantly affected by temperature, which directly reflects the alloying degree of NiMo. Compared to Ni/AC, it can be claimed from the XPS fitting data (Table 3) that Ni₄Mo/AC had the highest alloying at 450°C. The excessively high 0-valent Mo content at 500°C indicates that this temperature may be beneficial to the in situ reduction of metallic Mo, making the segregation of Mo more obvious. This can also be corroborated by its poorer activity, which is closer to that of Mo/AC. The low 0-valent Mo content at 400°C indicates that the reduction process of the catalysts was not carried out thoroughly enough. The reduction rate of Mo at lower temperatures could not match the formation of Ni, so better alloying could not be formed, which can also be confirmed from the H₂-TPR of Mo/AC. In addition, TEM (Figure 7) of the three catalysts showed that the particle size of the catalysts gradually increased with increasing temperature, but the larger particle size did not favor the catalytic hydrogenation process of NEC (Yu et al., 2020). At 500°C, the higher reduction temperature led to a certain extent of aggregation of metals. The active sites are thus reduced, which also affects reactivity. Therefore, an appropriate reduction temperature of 450°C is an important condition to ensure catalyst activity by affecting alloying.

It was found that NiMo/AC catalysts with different Ni-Mo ratios also had different catalytic effects. Three catalysts with different Ni-Mo ratios, Ni₃Mo/AC, Ni₄Mo/AC, and Ni₅Mo/AC, were prepared, and they were subjected to characterization and NEC hydrogenation tests. The hydrogenation curves of the three catalysts are shown in Figure 4C. The best catalytic effect of Ni₄Mo/AC and the worst catalytic effect of Ni₅Mo/AC were shown, while Ni₃Mo/AC was slightly more active than Ni₅Mo/AC. Similar to what we mentioned above, this could also be attributed to the alloying degree of nickel and molybdenum at different ratios. In the XRD patterns of all three catalysts (Figure 8), distinct nickel diffraction peaks could be observed, but the nickel diffraction peaks in Ni₄Mo/AC were most clearly shifted to a small angle, indicating that the best alloying effect was achieved at a Ni-Mo ratio of 4, followed by Ni₃Mo/AC and then Ni₅Mo/AC. The XPS patterns (Supplementary Figure S4) of the three catalysts showed approximately the same peak splitting results, and the difference also came from the valence distribution of Mo under different NiMo ratios. The reactivity is linearly related to the content of 0-valence Mo, that is, the degree of possible segregation. In addition, H₂-TPR tests were performed on the prepared samples before reduction. The H₂-TPR patterns (Supplementary Figure S5) of the three catalysts have similar shapes, but the highest reduction peaks were found for Ni₄Mo/AC, which may be responsible for the high alloying of the catalysts. It can be realized that the difference in alloying degree caused by the ratio change is not as significant as the effect of temperature, so the catalyst activity changes little by the ratio.