ZIF-67-Derived Co−N−C Supported Ni Nanoparticles as Efficient Recyclable Catalyst for Hydrogenation of 4-Nitrophenol

Abstract

In this study, a novel, highly efficient, environment friendly, and low-cost nanocatalyst, denoted as Ni(x)@Co−N−C, was successfully developed by encapsulating Ni nanoparticles into N-doped porous carbon derived from ZIF-67. A variety of techniques including powder X-ray diffraction (XRD), nitrogen adsorption/desorption, scanning electron microscopy (SEM), transmission electron microscopy (TEM), and X-ray photoelectron spectrometer (XPS) were used to characterize the prepared materials. The TEM images reveal that the nanoparticles were distributed homogeneously in the carbon support. The N atoms in the carbon support serve as the sites for the nucleation and uniform growth of Ni nanoparticles. The catalyst was used for the degradation of environmentally harmful 4-nitrophenol (4-NP) to 4-aminophenol (4-AP). Among all the catalysts investigated, Ni(10)@Co-N-C exhibited the highest catalytic activity for the hydrogenation of 4-NP, with a specific reaction rate of 6.1 × 10⁻³ s⁻¹, activity parameter of 31 s⁻¹g⁻¹, and turn over frequency (TOF) of 1.78 × 10¹⁹ molecules g_metal⁻¹s⁻¹. On the other hand, the specific reaction rate and TOF value were 1.7 × 10⁻³ s⁻¹ and 6.96 × 10¹⁸ molecules g_metal⁻¹s⁻¹, respectively, for Co−N−C. This suggests that Ni(10)@Co−N−C is about three times more catalytically active than the Co−N−C catalyst. The superb activity of Ni(10)@Co−N−C in comparison to Co−N−C can be ascribed to the homogeneous dispersion of small-sized Ni nanoparticles, the interconnected three-dimensional porous arrangement of the support Co−N−C, the presence of N atoms in the carbon framework that stabilize metal nanoparticles, and the synergistic electronic effect between Ni and Co. The Ni(10)@Co−N−C catalyst maintained consistent catalytic activity over multiple cycles, which suggests that porous N-containing carbon support can effectively prevent aggregation and leaching of metal nanoparticles. The ICP-AES analysis of the recycled Ni(10)@Co−N−C revealed a slight reduction in metal content compared to the fresh sample, suggesting almost negligible leaching of metal nanoparticles.

1. Introduction

One of the important issues that has generated significant concerns recently is water pollution because of its harmful consequences on human health caused by the direct disposal of various industrial, agricultural, and residential wastewater into water bodies [1,2]. 4-nitrophenol (4-NP) is water soluble, belongs to a highly toxic, carcinogenic, and poorly biodegradable compound that is extensively found in the industrial wastewater from sectors such as dye manufacturing, textile, drug synthesis, and herbicides production [3,4]. The presence of 4-NP in water can affect the nervous system, internal organs, and blood of human and animals [5,6]. On the other hand, 4-aminophenol (4-AP) has lower toxicity and serves as an important intermediate in the development of various products across different industries [7]. Thus, it is crucial to convert 4-NP to 4-AP based on the harmful effects of 4-NP. Several remedial methods have been employed to convert 4-NP to 4-AP, including adsorption [8], photocatalytic degradation [9], microbial degradation [10], microwave assisted catalytic oxidation [11], electrochemical degradation [12], and catalytic hydrogenation [13]. Among these methods, catalytic hydrogenation has proven effective from the perspective of sustainable development and green chemistry, as it produces high-value-added products without generating acid waste significantly, thereby minimizing the impact on the environment.

The catalytic hydrogenation approach that utilizes reducing agents like NaBH₄ and hydrazine hydrate with metal-based catalysts is the most cost-effective and straightforward method to convert 4-NP to 4-AP [14]. Although the use of noble metal-based catalysts such as Ag, Au, Pd, and Pt for the reduction of 4-NP exhibit excellent catalytic activity [15,16,17], their practical applications are often hindered due to their high costs and scarce supplies. As a result, the catalysts based on non-noble metal-based nanoparticles have garnered significant attention due to their ample supplies, low costs, plentitude of active surface sites, and distinct electronic structures [18,19,20]. Among various metal nanoparticles, Ni nanoparticles are particularly intriguing because they are cost effective and exhibit high catalytic activity and selectivity in hydrogenation reactions [21,22,23]. However, metal nanoparticles tend to agglomerate and change their shape easily during hydrogenation due to their high surface energy, and thus affect their catalytic activity and stability. To alleviate the agglomeration problems caused by high surface energies, metal nanoparticles are typically dispersed or immobilized on the surface of supports such as metal oxides [24], polymers [25], and carbon-based materials [26,27], or by other means such as coordinating ligands [28] and ionic liquids [29].

Porous carbon (PC)-based supports have received particular interest for immobilizing metal nanoparticles due to their excellent textural properties, outstanding hydrothermal stability, low corrosion potential, and resistance across a wide pH range. Furthermore, incorporating heteroatoms including sulfur (S), boron (B), oxygen (O), nitrogen (N), and phosphorus (P) can significantly enhance the chemical and electronic properties of carbon materials. Metal–organic frameworks (MOFs), consisting of uniformly dispersed metal ions or metallic clusters coordinated with electron-donating organic ligands [30], have recently emerged as promising templates for developing porous carbons through pyrolysis. Their tunable pore structures, diverse morphologies, and homogeneous distribution of heteroatoms make them particularly attractive for the development of PC [31,32]. MOF-derived carbons display stable morphologies and exhibit large internal surface areas and uniform pore sizes. They also exhibit high crystallinity, facilitate the easier incorporation of heteroatoms, and offer enhanced chemical and thermal stability [30,33]. Because of these advantages, MOF-derived porous carbons are widely used in various applications, including catalysis and energy storage devices [34,35]. Additionally, a variety of carbon materials with diverse functional groups can be generated by utilizing different MOFs as precursors. Zeolitic imidazolate frameworks (ZIFs), a subgroup of MOFs, are considered as attractive precursors for fabricating porous carbons due to their low costs, simple synthesis procedures, large specific surface areas, and tunable morphology and porosity. In addition, porous carbons derived from ZIFs are inherently N-doped, which enhances the hydrophilicity of the catalyst [36,37]. Yusran et al. [38] developed Co@N-C nanocatalyst using a Co-based MOF precursor and utilized them for the catalytic reduction of 4-NP. The catalyst displayed excellent catalytic activity due to the combined effect of active sites provided by Co nanoparticles and the establishment of a C-N matrix that acted as the strong Lewis base. MOF-derived N-doped porous carbons are viewed as promising supports for immobilizing metal nanoparticles due to their ultra-high porosity. The uniformly dispersed N atoms serve as the active sites for the emergence of the homogeneously distributed metal nanoparticles within the carbon support since N-doping enhances the bonding between the support and the metallic nanoparticles. Tian et al. has developed novel N-doped carbon-supported Ru nanoparticles (Ru-NC) using 2D Zn-MOF-Ru(cptpy)₂ as the metal precursor and dicyandiamide as the N dopant [39]. The material thus obtained was used as the catalyst for the reduction of nitroaromatic compounds and rhodamine B, and exhibited excellent conversion efficiency and stability. Jiang et al. developed a Ru-loaded ZIF template-assisted mesoporous carbon catalyst and utilized it for ammonia borane hydrolysis and p-nitrophenol reduction [40]. The catalyst showed superb catalytic activity for both reactions, which was attributed to the good structure of the catalyst and its hydrophilic nature, as well as the synergistic effect of bimetallic RuCo. In addition to dispersion, the catalytic activity of carbon-supported metal nanoparticles is significantly influenced by the choice of the metal precursors and synthesis process [41]. The N-doped porous carbon-supported metal nanoparticles are mainly synthesized by three different methods: (i) a post-loading method where a metal precursor is deposited onto pre-synthesized N-doped porous carbon, (ii) incorporating metal and N precursors simultaneously onto a pre-synthesized carbon support, and (iii) the in situ formation of metal and/or metal oxide on a N-doped carbon material [42]. Using the post-loading technique, the hierarchically porous structures of N-doped carbons and the amount of metal loading can be adjusted [43]. Even though the post-loading of metal nanoparticles on pre-synthesized N-doped porous carbon has various advantages, such as control over nanoparticles size and distribution, flexibility in the choices of materials, improved stability, and the efficient use of nanoparticles in comparison to the other two approaches, it has not been extensively explored to synthesize nanocatalysts.

With the aim to create a highly efficient, environmentally friendly, and low-cost catalyst for the reduction of 4-NP, herein, we synthesized a series of nanocatalysts by uniformly dispersing Ni nanoparticles in N-doped porous carbon derived from ZIF-67, named as Ni(x)@Co−N−C, where x is the weight percentage of Ni in the composite, using the post-loading approach. The catalytic performance of the Ni(x)@Co−N−C was assessed by its ability to reduce 4-NP to 4-AP in the presence of NaBH₄ as the reducing agent. Carbonized ZIF-67 was chosen as the support because it contains N atoms, which acted as nucleation sites for the growth of the Ni nanoparticles. Also, some Co nanoparticles remained within the porous carbon structure after the carbonization of ZIF-67, aiding in the formation of bimetallic NiCo nanoparticles during thermal reduction. The synergistic effect between Ni and Co has the potential to enhance reaction kinetics in the 4-NP reduction process. Moreover, both Ni and Co are cost-effective and possess excellent redox properties, enabling efficient electron transfer during the reduction of 4-NP to 4-AP. The structural characteristics of the developed nanocomposites were analyzed by various characterization techniques. To the best of our knowledge, the immobilization of Ni nanoparticles onto N-doped porous carbon derived from ZIF-67 by the post-loading technique and their utilization as a catalyst for the hydrogenation of 4-NP has not been reported yet.

2. Results and Discussion

2.1. Structural Characterizations of Ni(x)@Co-N-C

The phases and crystallite sizes of the prepared ZIF-67, ZIF-67-600, Co−N−C, and Ni(x)@Co−N−C nanocomposites were determined from the analysis of X-ray diffraction (XRD) patterns and are presented in Figure 1. The characteristic peaks of ZIF-67 displayed in Figure 1A matched well with the simulated pattern published in the literature, indicating the successful synthesis of ZIF-67 [44]. The XRD pattern of ZIF-67 pyrolyzed at 600 °C (ZIF-67-600) showed peaks at 2θ = 43.2, 51.5, and 75.9°, which are the characteristics of metallic Co (JCPDS No. 15-0806) [45]. ZIF-67 was also calcined at 500 and 700 °C to observe the effect of temperature on the crystallinity and hence on the catalytic activity. The peak intensity became stronger at relatively higher temperatures, as shown in Figure S1 (Supplementary Materials), due to the high degree of crystallization of Co. At higher pyrolysis temperatures, Co atoms might acquire enough energy to migrate and aggregate, resulting in the formation of larger particles. However, the peak positions shifted slightly towards higher angles, and the intensity of the Co characteristic peak at 2θ = 44.2° significantly decreased after treating ZIF-67-600 with HNO₃, as can be seen from the XRD pattern of Co−N−C. This change in the XRD pattern suggested the presence of some small-sized Co nanoparticles. They might have located in regions of the carbon matrix that were not easily accessible to HNO₃, making it difficult for acid to reach, and etched them entirely. However, the intensity of the peak at 2θ = 26° increased after acid treatment of ZIF-67-600, indicating a higher degree of graphitization in Co−N−C. The distance between the adjacent planes of the carbon atoms was approximately 0.341 nm, hinting the formation of well-graphitized Co−N−C. The XRD patterns of Ni(x)@Co−N−C catalysts shown in Figure 1B were almost identical to the Co−N−C material, implying that the framework of Co−N−C was not distorted by the incorporation of Ni nanoparticles. The diffraction peaks at 2θ = 44.2, 51.2, and 76.0° corresponded to the (111), (200) and (220) planes of the well-crystallized face-centered cubic (fcc) phase of Ni (JCPDS card No. 04-0850) [46]. These results suggested that Ni²⁺ cations were converted into Ni metals during the thermal reduction process. The interplanar spacing (d₁₁₁ = λ/2sinθ) and lattice constant (a₀ = d₁₁₁√3) of (111) facet of Ni were estimated to be about 0.204 nm and 3.54 Å, respectively. A schematic illustration of the structural evolution of the Ni(x)@Co−N−C catalyst is presented as Scheme 1.

Average crystallite size was determined from the most intense peak using Scherrer formula,

$$\mathrm{D}={\displaystyle \frac{\mathsf{\lambda }}{\mathrm{B}\cos \mathsf{\theta }}}$$

(1)

where λ is the X-ray wavelength for Cu Kα, B is the full width at the half-maximum (FWHM) of the peak in radians, and θ is the diffraction angle associated with the most intense peak. The average crystalline sizes were estimated to be 3.6, 5.9, 7.8, 12.4, and 12.9 nm for Co−N−C, Ni(5)@Co−N−C, Ni(10)@Co−N−C, Ni(20)@Co−N−C, and Ni(30)@ Co−N−C, respectively. At high Ni precursor concentrations, nanoparticles formation is more abundant than at lower Ni concentrations due to the increased availability of metal ions. However, the number of N atoms on the porous carbon support may become insufficient or saturated for Ni nanoparticles immobilization, reducing their ability to prevent aggregation, resulting in larger crystallite sizes.

The surface morphologies of ZIF-67, Co−N−C, and Ni(x)@Co−N−C were examined in a SEM and are presented in Figure 2. The SEM image of ZIF-67 illustrates the rhombic dodecahedron shape featuring a uniform size of approximately 500 nm. After the pyrolysis and HNO₃ treatments, the product particles inherited the overall crystal shape of ZIF-67, as can be seen from the SEM image of Co−N−C. However, the surface became rougher and the crystal size decreased due to the decomposition of the organic linkers in ZIF-67 during the pyrolysis. The morphology of Co−N−C remained largely unchanged after the inclusion of Ni nanoparticles, as demonstrated in the SEM images of Ni(5)@Co−N−C, Ni(10)@Co−N−C, Ni(20)@Co−N−C, and Ni(30)@ Co−N−C.

Figure 3 shows the TEM images of ZIF-67, Co−N−C and Ni(x)@Co−N−C. It revealed that ZIF-67 has a particle size of approximately 500 nm (Figure 3a) and exhibited a characteristic rhombic dodecahedron shape. After pyrolysis, Co−N−C maintained the ZIF-67 backbone structure with some deformation and displayed a uniform particle size of about 300 nm with a rough surface, as observed in the TEM image of Co−N−C, shown in Figure 3b. The reduction in particle size compared to ZIF-67 was due to the skeleton shrinkage during carbonization. The organic linkers in ZIF-67 decomposed at the high pyrolysis temperature, which can be realized from the sharp weight loss of about ~60 wt.% within the temperature range of 270 to 320 °C in the thermogravimetric analysis (TGA) plot of ZIF-67 presented in Figure S2. The decomposition of the organic linker led to the shrinkage of the ZIF-67 skeleton after carbonization. The observation of darker dots in the TEM image of Co−N−C at some places within the carbon matrix indicated the presence of Co nanoparticles. It suggested that the HNO₃ treatment might not be very effective to remove all of the Co nanoparticles from the carbon framework since some Co nanoparticles might have been completely covered by the graphitic carbon layer. During the high temperature pyrolysis in an inert atmosphere, the 2-MI organic ligands in ZIF-67 promptly converted and developed to nitrogen-rich porous carbon. Concurrently, the Co nodes within the framework were reduced by the nitrogen-rich carbon and produced metallic Co nanoparticles. Most of these Co nanoparticles were accessible to reactants and worked as active centers for the catalytic reaction. Some self-grown carbon nanotubes on the surface of the carbon matrix were also detected. The Co node in ZIF-67 could act as the catalyst during carbonization and facilitate the nucleation and growth of the nanotubes by providing the catalytic sites [47]. The formation of carbon nanotubes during the carbonization process could improve the electronic conductivity of the composites [48], which is beneficial for the catalytic performance of the nanocomposite. The TEM images of Ni(5)@Co−N−C, Ni(10)@Co−N−C, Ni(20)@Co−N−C, and Ni(30)@ Co−N−C presented in Figure 3c–f display the homogeneous distribution of a large amount of nanoparticles in the carbon matrix. The average sizes of the nanoparticles were determined to be around 4, 9, 13, and 20 nm for Ni(5)@Co−N−C, Ni(10)@Co−N−C, Ni(20)@Co−N−C, and Ni(30)@Co−N−C, respectively. Furthermore, at high Ni precursor concentrations, a few relatively large nanoparticles were observed on the surface. During the thermal reduction, Ni atoms can diffuse and interact with Co nanoparticles in Co−N−C support. This interaction can alter the growth kinetics of both Ni and Co nanoparticles and alter their particle size. In addition, some existing Co nanoparticles coalescence due to thermal reduction at high temperatures and, as a result, their size increased. The particle size distribution curve shown in the insets of Figure 3c–f evidenced the formation of nanoparticles of different sizes. The larger particle size of the nanoparticles measured with the TEM images as compared to the crystallite sizes estimated from the XRD data indicated the presence of multiple crystallites separated by grain boundaries within a single particle. The average crystallite size determine by using the Scherrer equation is a theoretical value that may not estimate the particle size accurately as it also depends on particles’ shape.

The high-resolution TEM (HRTEM) image of Ni(10)@Co−N−C in Figure 4a confirmed that the particles were well crystallized. The interplanar distance of the well-resolved lattice fringes were measured to be 0.204, 0.176, and 0.126 nm, matching with the (111), (200), and (220) planes of the face-centered cubic (fcc) structure of Ni [49]. The selected area electron diffraction (SAED) displayed in Figure 4b showed that diffraction spots superimposed on the rings, which revealed the formation of polycrystalline materials. The distinct circular patterns in particular corresponded to the (111), (200), (220), and (311) lattice planes of Ni, suggesting the growth of well-crystallized metal nanoparticles. Co and Ni have very similar atomic weight and lattice constants, differing by less than 1%, making it challenging to distinguish between them within the measurement errors from the lattice constant estimation. Therefore, elemental mapping was performed on Ni(10)@Co−N−C, which indicated that all the Ni, Co, C, N, and O elements were homogeneously distributed over the entire carbon framework (Figure 4c). During thermal reduction, Ni atoms could diffuse into the residual Co nanoparticles within the Co−N−C framework, indicating the potential for forming bimetallic NiCo nanoparticles. The actual amounts of Ni and Co present in the composite determined by ICP-OES are presented in

Table 1. Structural and textural properties of Ni(x)@Co-N-C.

Sample	Surface Area (m2g−1)	Pore Volume (cm3g−1)	Pore Size (nm)	d111 a (nm)	a0 b (nm)	Metal Loading (wt.%) c
						Co	Ni
Co−N−C	100	0.14	7.8	0.206	0.357	31.0	-
Ni(5)@Co−N−C	259	0.44	4.1	0.204	0.355	29.7	2.9
Ni(10)@Co−N−C	316	0.38	4.1	0.205	0.355	27.1	5.6
Ni(20)@Co−N−C	312	0.29	4.1	0.205	0.355	26.1	11.1
Ni(30)@Co−N−C	109	0.19	4.1	0.205	0.355	24.2	16.8

^a Interplanar spacing. ^b Lattice constant. ^c Amount of metal loading determined by ICP-OES.

. It was observed that after the introduction of Ni, Co loading decreased. This could be attributed to the competition between Ni and Co ions for coordination sites in the support during thermal reduction. As the Ni concentration increased, more Co was replaced by Ni, leading to a reduction in Co content in the composite. However, the actual Ni content in the catalyst was approximately half of the theoretical value. The lower Ni loading than theoretical value suggested that some Ni precursors did not successfully anchor onto the porous carbon support, likely due to the presence of residual oxygen-containing functional groups. Although HNO₃ treatment was performed to remove these groups, complete elimination may not have occurred, possibly due to the mild reaction temperature. These remaining functional groups might have decomposed and weakened the metal–support interaction, causing some Ni precursors to absorb weakly onto the carbon surface. As a result, they were washed away easily during washing and filtration, leading to a reduction in the actual metal content. The elemental composition and chemical states of the prepared nanocomposites were tested through X-ray photoelectron spectroscopy (XPS) analysis and the results are shown in Figure 5 and Figure S3. Figure S4 displays the survey spectrum of Ni(10)@Co−N−C, revealing the characteristic peaks of C 1s, N 1s, O 1s, Co 2p, and Ni 2p. This corroborated well with the existence of C, N, O, Co, and Ni in the nanocomposites. The high-resolution C 1s spectrum of Ni(10)@Co−N−C, shown in Figure 5 exhibited a strong peak at a binding energy value of 284.6 eV, along with smaller peaks at 285.4, 286.8, and 289.0 eV, which corresponded to C–C/C–C, C–N, C–O, and C=O, respectively [50]. The occurrence of the C–N bond indicated that N species were embedded in the carbon framework of Ni(10)@ Co−N−C. The distinct peaks located at 398.0, 400.8, 403.5, and 406.2 eV in the high resolution N 1s spectrum of Ni(10)@Co−N−C could be assigned to N_Pyridinic, N_Pyrrolic, N_Quaternary, and N_Oxide, respectively [51]. Pyridinic N, which can significantly modulate the electronic structure of the carbon matrix, was identified as the main N state in the nanocomposite. The pyridinic N and pyrrolic N species could also serve as nucleation sites for the growth of the Ni nanoparticles, promoting their homogeneous dispersion within the nanocomposite. This process consequently increased the binding energy between the support and the metal nanoparticles [52]. The pyridinic-N in the carbon framework can improve the surface wettability of the composite [53], which is highly important for the enhancement of catalytic activity of the catalyst. The high resolution Co 2p spectrum displayed two distinct peaks at binding energies of 781 and 797 eV, corresponding to Co 2p_3/2 and Co 2p_1/2 of metallic Co [54]. The deconvulated Co 2p spectrum showed peaks at binding energies of 778.4, 780.6, and 783 eV, corresponding to Co⁰, Co-N_x, and Co²⁺, respectively. The shake-up satellite at 786.8 eV was from the spin orbit component (Co²⁺ combined with O) [55]. The presence of Co²⁺ was due to the unavoidable surface oxidation of Co during the characterization [56]. The observation of a peak corresponding to Co-Nx suggested a co-ordination of Co with pyridinic-N and quaternary-N. As for the deconvoluted Ni 2p spectrum, the principal peaks at 855.8 and 873.8 eV with a spin-energy separation of 18.0 eV corresponded to Ni 2p_3/2 and Ni 2p_1/2 of metallic Ni, respectively. The spectra showed distinct peaks that matched to the metallic Ni (853.3 eV), Ni²⁺ (855.4 eV), and Ni³⁺ (858.8 eV), along with shakeup satellite peaks at 862.6 and 885.7 eV [57]. The Ni satellite peak originated from the multielectron excitations (shake-up) [58,59]. The emitted photoelectron during XPS excitation excited the valence electrons, which led to the energy loss and appearance of satellite peak at a higher binding energy site. The large percentage of Co and Ni satellite peaks can be attributed to the presence of unpaired d-electrons in valence shells of both Ni and Co. When an electron is excited to these higher unoccupied states during XPS excitation, interactions between the remaining core hole and the unpaired d-electrons led to multiple splitting, resulting in the appearance of high-intensity satellite peaks located at a higher energy than the main core-level peaks [60]. The atomic composition of each metallic component was estimated from the deconvoluted peaks and determined to be 30.6, 37.6, and 31.8% for Ni⁰, Ni²⁺, and Ni³⁺, respectively. The contributions from Co⁰, Co²⁺, and Co³⁺ were 33.3, 33.4, and 33.3%, respectively. Although XRD analysis could not detect the oxidized Ni species, the presence of Ni²⁺ and Ni³⁺ indicated that some Ni nanoparticles were not completely covered by the carbon layer, resulting in partial oxidization upon exposure to air. As observed, the binding energies for Co 2p in Ni(10)@Co−N−C were on the higher side relative to those in Co−N−C, as presented in Figure S5. The shifting demonstrated that electrons were exchanged between Co and Ni, hinting the formation of bimetallic NiCo nanoparticles. When Co atoms closely interact with Ni, electrons are transferred from Co to Ni to equilibrate the Fermi level due to the difference in their work functions (5.01 eV for Ni and 5.0 eV for Co), resulting in the formation of NiCo nanoparticles [61]. No substantial shift in the peak positions was observed due to the close values of work function of both the materials.

The physicochemical properties of Ni(x)@Co−N−C were evaluated through the N₂ adsorption/desorption isotherm analysis. Figure 6 presents the N₂ adsorption–desorption isotherms of Ni(5)@Co−N−C, Ni(10)@Co−N−C, Ni(20)@Co−N−C, and Ni(30)@Co−N−C. The isotherms of all samples displayed a mixture of type I and type IV isotherms with hysteresis loops, indicating the co-existence of micropores and mesopores in the nanocomposites [62]. As ZIF-67 was used as the precursor, its intrinsic porosity and framework arrangement worked as the template for the generation of the pores in Co−N−C. During carbonization, the organic linkers in ZIF-67 decomposed and the released volatile gases created the voids, which contributed to the formation of the pores in Co−N−C [63]. The estimated specific surface area, pore size, and pore volume are summarized in Table 1. As seen in Table 1, the specific surface area and pore volume of Co−N−C were much lower than the parent ZIF-67 (surface area = 1266 m²g⁻¹, pore volume = 0.65 cm³g⁻¹), which could be credited to the partial collapse of the microporous structure resulting from the high-temperature treatment. The BET surface area increased after incorporation of Ni nanoparticles due to the formation of more voids and rough surfaces after thermal reduction. However, Ni(30)@Co−N−C with a higher Ni content showed lower surface area and smaller pore volume, which indicated that excessive incorporation of Ni could block the pore channels. Although the surface area changed after the immobilization of Ni nanoparticles, the pore size was uniform, which suggested that metal nanoparticles might have resided on the external surface or within larger mesopores of the Co−N−C support. In addition, the N atom in the support facilitated the uniform dispersion of the nanoparticles. The high BET specific surface area and larger pore volume of the catalyst were beneficial as they could enhance catalytic activity by providing more active sites for the reactants. The Ni(10)@Co−N−C was particularly favorable for catalytic reactions, providing ample of active sites for the reactants due to its larger surface area and pore volume. The porous nature of the nanocomposite could facilitate rapid diffusion of the reactants during catalysis by providing uniform channels for substrate adsorption.

2.2. Evaluation of the Catalytic Activity of the Catalyst

The reduction of 4-NP to 4-AP in the presence of an excess quantity of NaBH₄ was employed as a model reaction to evaluate the catalytic activity of Ni(x)@ Co−N−C. 4-NP is a phenolic compound that has a nitro (–NO₂) group at the opposite position of the hydroxyl (–OH) group on the benzene ring. The size of a benzene molecule is generally about 0.6 nm. The –NO₂ and –OH functional groups extended slightly beyond the benzene ring, adding to the overall size. Considering these factors, the molecular dimension of 4-NP could be estimated as about 0.6 to 0.8 nm. The 4-NP could diffuse into the reaction system as the surface area, pore size, and pore volume of the catalyst Ni(x)@Co−N−C were significantly larger than the molecular dimension of 4-NP. The addition of NaBH₄ deprotonated the OH group of 4-NP and shifted the absorption peak of 4-NP from 317 nm to 400 nm due to the formation of 4-nitrophenolate ions under the basic conditions [64], as seen in Figure S6A. The intensity of the peak at 400 nm remained stable for hours (Figure S6B) without the catalyst, indicating that the presence of NaBH₄ alone could not speed up the reduction process of 4-NP. In contrast, the intensity of the peak dropped gradually after the addition of the catalyst Ni(x)@Co−N−C and a small shoulder peak simultaneously emerged at 300 nm, indicating the formation of 4-AP. Figure 7a–e illustrates the absorbance curves of 4-NP reduction with Co−N−C and Ni(x)@Co−N−C catalysts as a function of reaction time. As demonstrated in Figure 7, Co−N−C needed the significantly longer time of 18 min to complete the reduction reaction compared to Ni(5)@ Co−N−C and Ni(10)@Co−N−C. The reaction time reduced drastically to 7 min and the peak at 400 nm disappeared with the addition of a small amount of Ni(10)@Co−N−C nanocatalyst into the reaction system, as shown in Figure 7c. The reaction time was much shorter than those of Ni(20)@Co−N−C (16 min) and Ni(30)@Co−N−C (18 min). Given that the amount of NaBH₄ was excess as compared to 4-NP and the catalyst, the reaction kinetics can be assumed to follow a pseudo-first-order reaction that can be expressed as,

$$\ln \left({\displaystyle \frac{{\mathrm{C}}_{\mathrm{t}}}{{\mathrm{C}}_0}}\right)=-{\mathrm{k}}_{\mathrm{app}}\mathrm{t}$$

(2)

where k_app represents specific rate constant, t denotes reaction time, and C₀ and C_t are concentrations of 4-NP at time t = 0 and t = t, respectively. C₀ and C_t can be directly determined from the absorbances at the beginning (A₀) and at the reaction time (A_t) of 4-NP, respectively, following Beer-Lamberts law given by,

$$\mathrm{A}=\mathsf{\epsilon }\mathrm{C}\mathrm{l}$$

(3)

where A is absorbance at particular wavelength, ɛ molar absorption co-efficient, C concentration of the absorbing solution, and l is path length of the cuvette. Since the molar absorptivity and cuvette length were the same throughout the reduction experiments, absorbance would be directly proportional to concentration. Thus, the ratio of C_t/C₀ can be directly determined from the corresponding A_t/A₀ ratio where A₀ and A_t are the absorbance at 400 nm for 4-NP solution at t = 0 and reaction time t = t. The plots of ln(C_t/C₀) versus reaction time (t) for different catalysts are shown in Figure 7f. The specific rate constants were estimated from the slopes of these plots and the results are listed in

Table 2. Catalytic activities of Co−N−C and Ni(x)@ Co−N−C for 4-NP to 4-AP conversion.

Catalyst	t a (min)	kapp b (s−1)	κ c (s−1gcat.−1)	κ d (s−1gmetal−1)	TOF (Molecules gmetal−1s−1)
Co−N−C	18	1.7× 10−3	9	27.4	6.96 × 1018
Ni(5)@Co−N−C	8	5.2 × 10−3	26	79.7	1.56 × 1019
Ni(10)@Co−N−C	7	6.1 × 10−3	31	93.2	1.78 × 1019
Ni(20)@Co−N−C	16	3.0 × 10−3	15	40.3	6.75 × 1018
Ni(30)@Co−N−C	18	2.5 × 10−3	13	30.5	5.69 × 1018

^a Reaction time. ^b Specific rate constant. ^c Activity parameter obtained by considering the total mass of catalyst. ^d Activity parameter obtained by considering total amount of metal (Ni + Co) in the catalyst.

. Among the nanocatalysts investigated, the Ni(10)@Co−N−C nanocatalyst exhibited the best catalytic activity with a specific reaction rate of 6.1 × 10⁻³ s⁻¹. The 4-NP degradation efficiency of Ni(10)@Co−N−C was also examined at different stirring speeds, 250, 400, and 550 rpm, to evaluate the mass transport limitation. The specific rate constants were determined from the corresponding ln(C_t/C₀) versus time (t) plot (Figure S7) and estimated to be 5.6 × 10⁻³, 6.15 × 10⁻³, and 6.18 × 10⁻³ s⁻¹ for stirring speeds of 250, 400, and 550 rpm, respectively. Due to faster stirring speed, 4-NP molecules were evenly distributed throughout the solution, allowing the reactant molecules to reach the active sites more efficiently. The near-constant specific reaction rates at stirring speeds above 400 rpm indicated that mass transfer limitations could be effectively overcome under the current experimental conditions. However, the specific rate constant was not entirely reasonable to compare different supported catalysts with different loading amounts. Therefore, considering the amount of catalyst, the activity parameter (κ) was also estimated for a quantitative comparison of the catalytic activity using the relationship κ=k_app/M, where M is the mass of the catalyst. It is evident from Table 2 that Ni(x)@Co−N−C catalysts exhibited significantly higher catalytic activity compared to Co−N−C, with a maximum activity parameter of 31 s⁻¹g⁻¹ for Ni(10)@Co−N−C under the identical experimental conditions. The specific reaction rate of Ni(10)@Co−N−C was about 3.5 times higher than that of Co−N−C. The sequence of specific rate constants from high to low was Ni(10)@ Co−N−C > Ni(5)@Co−N−C > Ni(20)@Co−N−C > Ni(30)@Co−N−C > Co−N−C. It showed that catalytic activity decreased with the increase in Ni metal loading. The decrease in catalytic activity with the increase in Ni metal loading could be primarily attributed to the increase in particle size and reduction in pore volume of the catalyst. The TEM image analysis revealed that particle size increased with the increase in Ni loading. Since larger particles have smaller surface areas compared to smaller ones, the number of available active sites for catalytic reaction was less. Thus, a longer reaction time was required for the complete reduction of 4-NP to 4-AP. Also, large nanoparticles occupied the pores in the Co−N−C support, significantly reducing the pore volume. This reduced pore volume hindered the diffusion of reactant molecules, prolonging the reaction time for complete reduction and ultimately reducing catalytic activity. The lower specific reaction rate constant and activity parameter of the catalyst with higher Ni loading were thus directly linked to the extended time required for the complete reduction of 4-NP to 4-AP due to larger particle size and low pore volume. The activity parameters of Ni(10)@Co−N−C−500 and Ni(10)@Co−N−C−700 were determined as 5 and 24 s⁻¹g⁻¹, respectively. It is important to note that apparent specific rate constants and activity parameters of Ni-containing Co−N−C were much higher than those of the Co−N−C support, highlighting the positive impact of Ni on the catalyst’s performance. The activity parameter κ of Ni(x)@Co−N−C was also determined by considering the actual content of Co and Ni in the catalyst and these are listed in Table 2. It was noticed that by loading 10 wt.% Ni in the composite, the activity parameter can be enhanced by 70%. Turnover frequency (TOF), a kinetic parameter manifesting the number of 4-NP molecules reduced by one gram of catalyst (Ni(x)@ Co−N−C) per second time, was also estimated using the Equation (4) [65] considering total amount of metal and the results are presented in Table 2.

$$\mathrm{TOF}={\displaystyle \frac{{\mathrm{Conc}}_{4-\mathrm{NP}}\times \mathrm{Abs}.\mathrm{decline}\mathrm{\%}\times {\mathrm{N}}_{\mathrm{A}}}{\mathrm{M}\mathrm{e}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{w}\mathrm{e}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t}\left(\mathrm{\%}\right)\times \mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}}}$$

(4)

where ${\mathrm{C}\mathrm{o}\mathrm{n}\mathrm{c}}_{4-\mathrm{N}\mathrm{P}}$ is the initial concentration of 4-NP and N_A is Avogadro’s Number. TOF of the best catalyst, i.e., Ni(10)@Co−N−C, was 1.78 × 10¹⁹ molecules g_metal⁻¹s⁻¹, which was much higher than that of Co−N−C (6.96 × 10¹⁸ molecules g_metal⁻¹s⁻¹), suggesting the excellent reductive capabilities of the catalyst for 4-NP. As there was no significant difference in the total metal content for Ni(5)@Co−N−C and Ni(10)@Co−N−C catalysts, a minor variation in the TOF value was observed. The TOF values of the catalyst shown in Table 2 suggested that the catalytic activity of the catalysts for 4-NP reduction to 4-AP were highly dependent on their size, distribution, and composition. The superb catalytic activity of Ni(10)@Co−N−C could be attributed to the modification of the electronic structure and synergistic effects between Ni and Co. The catalytic activity of Ni(x)@Co−N−C was much higher than Ni-based catalysts supported on other carbon supports, such as RGO, carbon black, mesoporous carbon, etc., presented in Table S1. The improved activities could be attributed to the cumulative effects resulting from the homogeneous distribution of the nanoparticles, smaller particle size, metal loading, formation of bimetallic NiCo, and favorable textural properties of the Co−N−C support containing N within its framework. The catalytic activity of a catalyst for converting 4-NP to 4-AP is heavily dependent on its ability to absorb 4-NP molecules and the electron transfer rate from the donor BH₄^¯ to the acceptor 4-nitrophenolate. According to the literature, it has indicated that the presence of N species in the support have a significant impact on the catalytic activity of metal nanocatalyst embedded in the N-containing support [66]. The dominance of pyridinic N could also alter the electronic structure of the carbon by contributing a lone-pair of electrons, and thereby enhanced the interaction between the support and the substrate during the 4-NP reduction. This is particularly significant as 4-NP exhibits high charge and spin density. The improved interaction reduced the dissociation energy and enhanced the conversion rate. In addition, the large surface area of the catalyst offered numerous active sites for the reaction to occur. The three-dimensional porous structure of the support was also beneficial for the transport of the reactants to the surface of metal nanoparticles, and thus boosted the specific reaction rate [67]. Additionally, it could confine metal nanoparticles within the pores, preventing further growth of metal nanoparticles [68]. The overall catalytic conversion process preceded as follows: At first, reactant 4-NP and the BH₄^¯ penetrated into the porous support Co−N−C through the pores and adsorbed onto the surface of metal nanoparticles. Some electrons from the NiCo nanoparticles migrated to the carbon support due to strong metal-support interactions facilitated by the presence of N in the carbon support. This electron transfer rendered the NiCo nanoparticles electron-deficient and electrophilic. Upon the addition of NaBH₄, the solution pH increased, leading to the deprotonation of 4-NP into 4-nitrophenolate ions. Both 4-nitrophenolate and BH₄⁻ ions, being negatively charged, adsorbed onto electrophilic NiCo nanoparticles. The adsorbed BH₄⁻ ions then reacted with water molecules on the catalyst and generated active hydrogen species. The 4-nitrophenolate ions adsorbed onto NiCo nanoparticles, which received these active hydrogen species and produced 4-AP via the formation of the 4-hydroxylaminophenol intermediate and the elimination of two water molecules [69,70]. Finally, 4-aminophenol detached from the NiCo nanoparticles surface, and allowed the active sites for the next catalytic cycle. A schematic illustration of the conversion mechanism is shown in Scheme 2. The metal nanoparticles played a significant role by storing the electrons after electron transfer from the hydride, and thereby influenced the catalytic activity of the nanocomposite [71]. The improved activity of the catalyst could also be attributed to the formation of bimetallic NiCo nanoparticles. In bimetallic nanoparticles, electron density on the surface is higher than that of the monometallic nanoparticles because of the synergistic electronic effect between Ni and Co. According to the d-band center theory, the energy level of the d-band center (E_d) determines the adsorption intensity between the active site on the catalyst surface and the adsorbing molecules [72]. The d-band center of Ni locates at −1.4 eV and that of Co at −1.3 eV. It has been reported that a higher energy level corresponds to a stronger adsorption capacity [73]. As a result of the formation of bimetallic NiCo, the d-band centers of both metal components would be rearranged, leading to a stronger binding energy for 4-NP molecules and ultimately resulting in higher catalytic activity. According to the magnetic saturation computation, the d-band energy of Ni atoms have 0.6 holes (d-band hole), which can interact with 4-NP, leading to the formation of chemical adsorption bonds [74]. The nitro group could be effectively reduced by the metal nanoparticles through a direct mechanism to the nitroso group during electron transfer, and then converted into hydroxylamine intermediate, which was further hydrogenated to form an amino group at the final stage [75]. The disparity in electronegativity between the two metal atoms could also prompt electron transfer. The electronegativity of Ni atom is 1.91 and that of Co is 1.88 [76]. Due to electronegativity differences, electrons tend to move from active Co nanoparticles encapsulated by graphitic carbon to Ni. The redistribution of electron density could modify the surface and electronic structure of the bimetallic NiCo. The modification of the electronic structure could affect the binding energies of the reactant molecules, intermediates, and products which could improve the adsorption efficiency of NiCo nanoparticles, generate sufficient amounts of hydrogen, and subsequently enhance the reaction kinetics [77]. The XPS analysis revealed the presence of higher valence states of Ni and Co alongside their zerovalent forms, which contributed in the catalytic activity enhancement. During the 4-NP reduction process, NiO acted as the reducing agent, while CoO worked as an oxidizing agent. NiO donated electrons to CoO, reducing it to metallic Co, which served as additional catalysts. It thus improved the overall reaction kinetics and accelerated the 4-NP degradation rate.

An important benefit of the N-doped porous carbon-supported metallic nanoparticles as heterogeneous catalysts is their easy recyclability and reusability. The reusability of Ni(10)@Co−N−C was explored for the reduction of 4-NP to 4-AP, and the performance of the catalyst after being used for five cycles is depicted in Figure 8A. A gradual decline in the activity was observed with the catalyst after consecutive reaction cycles. The activity remained at approximately 93% after the fifth cycles, indicative of the excellent durability of the catalyst. The TEM image of the recycled catalyst, presented in Figure 8B, revealed an increase in the particle size (11 nm) compared to the fresh sample (9 nm), suggesting some degree of agglomeration among the metal nanoparticles during the catalytic reduction of 4-NP to 4-AP. As the size of NiCo nanoparticles grew, the number of exposed active sites decreased, which led to a decline in catalytic activity over multiple cycles. To assess the potential structural changes in the catalyst after five cycles, WXRD analysis was performed on the reused Ni(10)@Co−N−C catalyst. The analysis revealed no notable changes in structural properties, as indicated by the similarity between the WXRD patterns of the recycled Ni(10)@Co−N−C and fresh Ni(10)@Co−N−C, as shown in Figure 8C. Additionally, ICP-AES analysis of the recycled Ni(10)@Co−N−C revealed a slight reduction in metal content (Co ~26.8 wt.%, Ni ~5.1 wt.%) compared to the fresh sample (Co ~27.1 wt.%, Ni ~5.6 wt.%), suggesting minimal metal leaching. The excellent stability of the catalyst can be attributed to the presence of N-containing basic sites in Co−N−C, which promoted the uniform dispersion and stabilization of the metal nanoparticles within the support. The reduction in catalytic activity over successive cycles can be partly attributed to the loss of some catalysts during the separation process.

3. Materials and Methods

3.1. Materials

4-nitrophenol (4-NP), cobalt nitrate hexahydrate [Co(NO₃)₂⋅6H₂O], and nickel nitrate hexahydrate [Ni(NO₃)₂⋅6H₂O] were purchased from Alfa Aesar (Heysham, UK). 2-methylimidazol (2-MI) and NaBH₄ were obtained from Aldrich (St. Louis, MO, USA) and methanol (CH₃OH) was provided by Avantor (Norway).

3.2. Synthesis of Carbonized ZIF-67 Supported Ni Nanoparticles (Ni@ Co−N−C)

ZIF-67 was prepared at first to synthesize the carbonized ZIF-67-supported Ni nanoparticles. In the typical synthesis of ZIF-67, a metal solution was prepared by dissolving Co(NO₃)₂⋅6H₂O (3 mmol) in methanol (30 mL). Separately, the 2-methylimidazole (2-MI, 12 mmol) solution was prepared by dissolving 2-MI (12 mmol) in methanol (30 mL). The 2-MI solution was then gradually added to the metal precursor solution and stirred continuously for 10 min at room temperature. The solution was then thermally aged in a Teflon autoclave at 100 °C for 20 h. Once the solution had cooled down to room temperature, the purple solid precipitates were collected by centrifugation, washed with methanol, and dried at 70 °C to yield ZIF-67. The as-synthesized ZIF-67 was heated at 600 °C at a rate of 1 °C min⁻¹ for 6 h in a nitrogen environment. The carbonized product at this stage was denoted as ZIF-67-600. The ZIF-67-600 was further added into HNO₃ (30%) solution and stirred continuously at 60 °C for 24 h to remove the free or loosely bounded Co nanoparticles. The solids were separated from HNO₃ via centrifugation, washed with de-ionized water, and dried at 70 °C. The final product was named Co−N−C. The HNO₃ treatment can also have the potential to enhance the hydrophilicity and surface chemical activity of Co−N−C.

The prepared Co−N−C was used as the support to immobilize Ni nanoparticles through wet impregnation followed by calcination in an inert environment. In the typical metal nanoparticles immobilization process, Co−N−C (100 mg) was dispersed in an aqueous solution of Ni(NO₃)₂⋅6H₂O (0.01M) and stirred continuously at room temperature for 3 h. The concentration of Ni(NO₃)₂⋅6H₂O was varied from 0.05 to 0.3 mmol to prepare Ni nanoparticles with various loadings. The solid phase was subsequently separated by centrifugation, washed with de-ionized water, and then dried at room temperature for 24 h. The Co−N−C impregnated by Ni²⁺ ions was then thermally reduced at 600 °C under a flow of Ar/H₂ (95%:5%) gas for 6 h to produce Co−N−C supported Ni nanoparticles. The resulting product was designated as Ni(x)@Co−N−C, where x represents the weight percentage of Ni in the composite. Although the actual metal loading measured by ICP-AES was about half of the theoretical value (e.g., optimal catalytic composition of 10% corresponds to an actual measured value of 5.6%), the theoretical value is used for sample name consistency.

3.3. Characterizations

The degree of crystallinity and phase of Ni(x)@Co−N−C were determined using powder X-ray diffractometer (Lab-x XRD-6000 SHIMADZU, Kyoto, Tapan). The morphological and microstructural properties of the prepared materials were studied via field emission scanning electron microscope (FESEM, Bruker Nova NanoSEM 230, Karlsruhe, Germany) and transmission electron microscope (TEM, JEOL JEM2100F, Tokyo, Japan). The chemical states of Ni(x)@Co−N−C were analyzed using Thermo VG scientific sigma probe X-ray photoelectron spectrometer (XPS) (Thermo Fisher Scientific, Waltham, MA, USA). The actual amount of Ni and Co metals present in the Ni(x)@Co−N−C nanocomposite were estimated using a Jarrell-Ash, Thermo Fisher Scientific Massachusetts, USA) ICAP 9000, an inductively coupled plasma optical emission spectrometer (ICP-OES). The textural properties of the composites were explored from the N₂ adsorption–desorption isotherms acquired through Autosorb iQ2, Florida, USA at 77 K. The Brunauer–Emmett–Teller (BET) method was employed to estimate the specific surface area and the Barrett–Joyner–Halenda (BJH) method to estimate the pore size and volume.

3.4. Catalytic Reduction of 4-NP to 4-AP

The catalytic performance of Ni(x)@Co−N−C was evaluated based on its capability to reduce 4-NP to 4-AP in the presence of NaBH₄. The reaction was selected as the model for studying catalytic performance because it could proceed under mild conditions. To perform the experimental procedure, fresh aqueous solutions of 4-NP (0.09 mM, 22.5 mL) and NaBH₄ (0.0306 M, 5 mL) were prepared and thoroughly mixed. The solution was then added to an aqueous mixture containing 0.2 mg of Ni(x)@Co−N−C and 12 mL of H₂O and stirred at a speed of 400 rpm. At regular intervals, a small fraction of the reaction solution was removed from the system to track the reaction progress using UV-Vis spectrophotometer (T90+, PG Instruments, Leicester, UK). After the reaction was completed, the catalyst was collected by centrifugation, rinsed with de-ionized water, and air-dried. It was then reused for 4-NP reduction reaction by adapting similar procedures. The stability, phase purity, and physical changes of the recycled Ni(x)@Co−N−C were also investigated.

4. Conclusions

In summary, a highly efficient catalyst consisting of Ni nanoparticles embedded in N-doped porous carbon derived from the carbonization of ZIF-67 was developed for the reduction of 4-NP to 4-AP. The N doping in the carbon framework of the present catalyst provided the active sites for the growth of homogeneously dispersed small-sized Ni nanoparticles. In addition, its presence could also enhance the support–substrate interaction during the 4-NP reduction. These characteristics, along with large surface area, three-dimensional porous support, and the synergetic effect between Ni and Co, endowed Ni(10)@Co−N−C with excellent catalytic activity, achieving a specific reaction rate of 6.1 × 10⁻³ s⁻¹, 3.5 times higher than that of Co−N−C. The catalyst demonstrated excellent durability in effectively converting 4-NP to 4-AP, which was attributed to the N-containing basic sites in Co−N−C that facilitated the stabilization of Ni nanoparticles within the support. The easy synthesis, superb catalytic performance, and remarkable reusability of Ni(x)@Co−N−C makes it a promising nanocatalyst for various reactions.