Chemical synthesis of Zr-/Ce-/Sm-containing intermetallic compounds catalyzing NaBH₄-assisted hydrogenation of 4-nitrophenol

Abstract

Zr-/Ce-/Sm-containing intermetallic compounds, specifically ZrZnNi₄, CeNi₅, CeAlNi₄, CeNi₄Si, Ce(NiSi)₂, SmNi₃, SmNi₄Si, and Sm(NiSi)₂, were synthesized by reducing the metal oxides using a CaH₂ reducing agent within molten LiCl. The resultant nanopowders exhibited high specific surface areas: ZrZnNi₄ (42.0 m² g⁻¹), CeAlNi₄ (66.9 m² g⁻¹), and Sm(NiSi)₂ (25.0 m² g⁻¹). They were subsequently tested for their effectiveness in the NaBH₄-assisted hydrogenation of 4-nitrophenol. When compared to the prepared catalysts, including a conventional CeO₂-supported Ni catalyst, ZrZnNi₄ demonstrated the highest catalytic activity. Based on experimental results and density functional theory calculations, it was proposed that enhanced performance could be attributed to the formation of electron-rich Ni species in ZrZnNi₄.

---

Introduction

Nitrophenols are commonly utilized in various industrial applications, such as dyes, pesticides, herbicides, and pharmaceuticals. However, nitrophenols and their derivatives are toxic and present significant risks to both ecosystems and human health with prolonged exposure.^1,2 The conversion of nitrophenols into less harmful aminophenols via heterogeneous catalysts is a well-established method for degradation. The reduction of 4-nitrophenol (4NP) to 4-aminophenol (4AP) using NaBH₄ as a reducing agent has attracted considerable interest as a model reaction for assessing catalytic activity.³ The proposed reaction mechanism follows the Langmuir–Hinshelwood model,⁴ commencing with the hydrolysis of NaBH₄, generating B(OH)₄⁻ and H₂ gas or reactive hydrogen (H) adsorbed on the metal surface (M) as represented in eqn (1)–(6).⁵

NaBH₄(S) → Na⁺ + BH₄⁻ (1)

2M + BH₄⁻ ↔ M–BH₃⁻ + M–H (2)

M–BH₃⁻ ↔ BH₃ + M⁻ (3)

BH₃ + OH⁻ → BH₃(OH)⁻ (4)

M⁻ + H₂O → M–H + OH⁻ (5)

M–H + M–H → 2M + H₂(g) (6)

The product BH₃(OH)⁻ generated in eqn (4) is thought to exhibit reactivity similar to that of BH₄⁻ and undergoes reaction steps akin to eqn (2)–(4), leading to the formation of BH₂(OH)₂⁻. This species can engage in further reactions resulting in BH(OH)₃⁻ and, ultimately, B(OH)₄⁻. In the second step, the adsorbed reactive hydrogen reacts with the 4NP that is adsorbed onto the metal surface to produce 4AP, as depicted in eqn (7)–(9) through a direct pathway.⁶

M + 4NP → M–4NP (7)

6M–H + M–4NP → M–4AP + 2H₂O (8)

M–4AP → M + 4AP (9)

The rate-determining step in this process is the surface reduction of adsorbed 4NP to adsorbed 4AP (eqn (8)).⁴ Consequently, it is crucial to utilize catalysts that facilitate the hydrogenation of 4NP to 4AP (eqn (8)) to enhance the overall efficiency of the process.

Precious metal nanoparticles, such as Au and Ag, are recognized for their high catalytic performance in NaBH₄-assisted hydrogenation.^7,8 However, low-cost catalysts comprising Ni, Co, and Fe have also been developed to reduce catalyst expenses.^9,10 To enhance the activity of these economical metal catalysts, the incorporation of other metals into the catalyst structure has been investigated.¹¹ Low work function (WF) metals, such as W (4.55 eV), Cr (4.5 eV), and Mn (4.1 eV),¹² have the ability to readily donate electrons to higher WF metals, such as Ni (5.15 eV) and Co (5.0 eV), as long as there is tight contact between the two, leading to the creation of electron-rich Ni and Co species. Since the rate-determining step (eqn (8)) can be accelerated by increasing the surface concentration of reactive hydrogens (H) generated on these electron-rich metals (eqn (5)), designing improved catalysts that contain electron-rich active metals is a strategic approach in the NaBH₄-assisted hydrogenation of 4NP.

In this study, we developed Ni-based intermetallic compound catalysts featuring very low WF metals like Zr (4.05 eV), Ce (2.9 eV), and Sm (2.7 eV).¹² The oxidation of these metals to form oxides poses a significant challenge in the preparation of alloy catalysts with nanoscale morphologies, which are essential for achieving high surface areas. This issue was addressed by employing a CaH₂-assisted molten salt synthesis method, allowing for the reduction of metals by CaH₂ to occur in an oxygen-free molten LiCl environment,¹³ thus preventing oxidation. The versatility of this synthesis method was confirmed by the successful production of ZrZnNi₄, CeNi₅, CeAlNi₄, CeNi₄Si, Ce(NiSi)₂, SmNi₃, SmNi₄Si, and Sm(NiSi)₂ from the respective metal oxide precursors. The catalysts ZrZnNi₄, CeAlNi₄, and Sm(NiSi)₂, which possess high surface areas, were then evaluated in the context of the NaBH₄-assisted hydrogenation of 4NP.

Experimental

Preparation of intermetallic compound nanopowders

Zr-/Ce-/Sm-containing intermetallic compounds were synthesized by reducing metal oxides with CaH₂ as a reducing agent in molten LiCl at 600 °C. Initially, metal oxide precursors were prepared using the following citric acid method. For ZrZnNi₄, CeNi₅, CeAlNi₄, and SmNi₃, metal nitrate salts such as ZrO(NO₃)₂·2H₂O (97.0%, Wako Pure Chem. Corp.), Zn(NO₃)₂·6H₂O (99.9%, Wako Pure Chem. Corp.), Ni(NO₃)₂·6H₂O (98.0%, Wako Pure Chem. Corp.), Ce(NO₃)₃·6H₂O (98.0%, Wako Pure Chem. Corp.), Al(NO₃)₃·9H₂O (98.0%, Wako Pure Chem. Corp.), Sm(NO₃)₃·6H₂O (99.5%, Wako Pure Chem. Corp.) were dissolved in distilled water to obtain the stoichiometric molar ratios of ZrZnNi₄, CeNi₅, CeAlNi₄, and SmNi₃. For CeNi₄Si, Ce(SiNi)₂, SmNi₄Si, and Sm(NiSi)₂, the aforementioned metal nitrate salts were dissolved in distilled water and mixed with SiO₂ nanoparticles (99.5%, 10–20 nm, Sigma-Aldrich Co. LLC.) to achieve the stoichiometric ratios necessary for CeNi₄Si, Ce(NiSi)₂, SmNi₄Si, and Sm(NiSi)₂. Subsequently, citric acid was incorporated into the solutions at a molar ratio of total metals/citric acid = 1/1.2. After thorough mixing, the resultant solution was evaporated on a hot plate at 110 °C overnight. The resulting dried powder was initially heated in air at 250 °C for 2 hours and then gently ground in a mortar to create a homogeneous powder. Lastly, the powder was heated in air at 500 °C for 2 hours to produce the metal oxide precursors.

Next, the metal oxide precursor, CaH₂ (94.0%, JUNSEI Chem. Co.), and LiCl (99.0%, Wako Pure Chem. Corp.) were blended in a mortar in a weight ratio of 2/6/3 of metal oxide precursor/CaH₂/LiCl. This mixed powder was placed in a stainless-steel container filled with Ar gas and heated at 600 °C for 2 hours. The reduced precursors were then crushed in a mortar and rinsed multiple times with a 0.4 M NH₄Cl aqueous solution, followed by distilled water. The final dried products were designated as ZrZnNi₄, CeNi₅, CeAlNi₄, CeNi₄Si, Ce(NiSi)₂, SmNi₃, SmNi₄Si, and Sm(NiSi)₂.

The ZrNi₅ ingot for WFs measurement was prepared by arc melting a stoichiometric mixture of Zr and Ni elements on a water-cooled Cu hearth under a high-purity argon atmosphere. Annealing was subsequently conducted at 1000 °C for 20 hours to produce the single phase.

Characterization of intermetallic compound nanopowders

The crystal structure was analyzed utilizing X-ray diffraction (XRD; SmartLab, 3 kW, Rigaku Corporation) with CuK_α radiation at 40 kV and 30 mA. Porosity was assessed through N₂ adsorption and desorption at −196 °C (BELLSORP mini-II, Microtrac-BEL). The pore size distribution was derived from the measured adsorption isotherms using the Barrett, Joyner, and Halenda (BJH) method for all samples. Prior to measurement, the samples were pretreated at 150 °C for 60 minutes under vacuum. Scanning electron microscopy (SEM; JSM-7400F, JEOL Ltd) was employed to observe the morphology, with energy dispersive X-ray spectrometry (EDX) utilized for elemental analysis. X-ray fluorescence (XRF; ZSX PrimusII, 3 kW, Rigaku Corporation) was conducted in air to detect elements within the samples, except for lighter elements such as H, He, Li, N, and O. The chemical states and surface composition of the prepared sample were analyzed using X-ray photoelectron spectroscopy (XPS; PHI X-tool, ULVAC-PHI), operated with AlK_α radiation. The chemical shifts were calibrated by fixing the C1s peak of surface carbonaceous contaminants at 284.5 eV. Photoemission yield spectra were recorded in air to estimate the WFs using a RIKENKEIKI AC-2 spectrometer, which collected photoelectrons emitted from the polished surface of ZrNi₅ ingots upon UV light excitation through O₂ molecule mediation.

Catalytic test with intermetallic compound nanopowders

The catalytic reactions were performed in 20 mL glass bottles following previously documented protocols. In the catalytic tests, a 1 mL solution of p-nitrophenol (4NP) at 14 mM was combined with 10 mg of catalyst powder, 1 mL of NaBH₄ solution (0.42 M), and 7 mL of distilled water as the solvent. To ensure pseudo first-order reaction kinetics, the initial concentration of NaBH₄ (0.047 M) was set 30 times greater than that of 4NP (1.6 mM). The reactions were carried out under stirring at room temperature (25 °C) for 60 minutes. An aluminum heat sink on a hot plate was used to maintain a constant solution temperature. A small aliquot (100 μL) of the reaction solution was extracted to evaluate concentration changes throughout the reaction. The conversion of 4NP to p-aminophenol (4AP) was monitored using an ultraviolet-visible spectrometer (Shimadzu, UV-1280), through the corresponding absorbance changes at 401 and 315 nm.

For comparison, a 12 wt% NiO or 10 wt% Ni catalyst supported on commercial CeO₂ nanopowder (30 m² g⁻¹, <50 nm, Sigma-Aldrich Co. LLC) was prepared using a conventional impregnation method. Initially, Ni(NO₃)₂·6H₂O was dissolved in distilled water, and a suspension was created with CeO₂ to achieve a 12 wt% NiO loading. The resulting suspension was then evaporated on a hot plate at 110 °C overnight. The dried powder was subsequently heated in air at 500 °C for 2 hours to yield the 12 wt%NiO/CeO₂ catalyst. The metallic state nickel-loaded catalyst, denoted as 10 wt%Ni/CeO₂, was obtained after a reduction treatment of the NiO/CeO₂ catalyst at 400 °C under a H₂ flow at 100 mL min⁻¹ for 2 hours.

Electronic structure calculations

Density functional theory (DFT) periodic calculations were conducted using the generalized gradient approximation and the Perdew–Burke–Ernzerhof functional¹⁴ and, along with the projected augmented plane wave method¹⁵ implemented in the Vienna Ab initio simulation package.¹⁶ An energy cutoff of 600 eV was applied, and a gamma-centered k-mesh of 10 × 10 × 10 was utilized. A Wigner–Seitz radius of 0.146 nm was employed for the site projections of all atoms. Atomic charges were determined through Bader charge analysis,¹⁷ and the crystal structures were visualized with the VESTA software.¹⁸

Results and discussion

Preparation of intermetallic compound nanopowders

Eight intermetallic compounds—ZrZnNi₄, CeNi₅, CeAlNi₄, CeNi₄Si, Ce(NiSi)₂, SmNi₃, SmNi₄Si, and Sm(NiSi)₂—were synthesized by reducing metal oxides with a CaH₂ reducing agent in molten LiCl. Fig. 1 and Fig. S1–S5 display the XRD patterns of the metal oxide precursors, each identified in blue above the corresponding figure. Peaks associated with NiO (PDF01-071-1179) were present in all oxide precursors. Additionally, CeO₂ (PDF03-065-5923) was also detected in the oxide precursors of CeNi₅, CeNi₄Si, and Ce(NiSi)₂. Other metal oxides, including ZrO₂, ZnO, SiO₂, and Sm₂O₃, were not observed, suggesting they may exist in an amorphous form. The metal oxide precursors were subsequently reduced and washed to yield intermetallic compound nanopowders. The XRD patterns of the intermetallic compounds, shown in Fig. 1 and Fig. S1–S5 are described in orange below each figure. The peaks from the metal oxide precursors were absent, while the expected peaks for the intermetallic compounds ZrZnNi₄, CeNi₅, CeAlNi₄, CeNi₄Si, Ce(NiSi)₂, SmNi₃, SmNi₄Si, and Sm(NiSi)₂ appeared in the reduced samples. The peaks assigned to the metallic Ni were not observed in the reduced samples. Impurity peaks assigned to CaCO₃ were observed in CeNi₅ and Ce(SiNi)₂, probably due to the incomplete removal of Ca species from the samples. The peak shift was observed in CeNi₄Si, indicating the formation of deformed structure probably due to the deviation of composition in the synthesized sample. These results indicate that the reduction of the metal oxide precursors using the CaH₂ reducing agent in molten LiCl effectively formed the intermetallic compounds. The identified crystal types and crystallite sizes, calculated using the Scherrer equation based on the main peaks detected at 2θ, are summarized in Table 1. The calculated sizes ranged from 12.2 to 52.5 nm, indicating the formation of nanoscale morphologies.

Fig. 1 XRD patterns for the oxide precursors (blue) and the reduced samples (orange) of (a) ZrZnNi₄, (b) CeAlNi₄, and (c) Sm(NiSi)₂. The main peaks observed in the reduced samples are described with their index values.

Table 1 Summary of the measured properties, including BET surface area (SA), pore volume (V_p), crystallite size, and elemental composition. The crystallite sizes were calculated using the Scherrer equation applied to the main peaks detected at 2θ as described in the table

Sample	Crystal structure	Crystallinity		Porosity		Composition
		Crystallite size [nm]	2θ [degree]	BET SA [m^2 g^−1]	Vp [cm^3 g^−1]	SEM-EDX [mol%]	XRF [mol%]	XPS [mol%]
ZrZnNi4	AuBe5	12.2	44.3	42.0	0.094	Zr/Zn/Ni = 19.7/8.1/72.2	Zr/Zn/Ni = 22.4/7.9/69.8	Zr/Zn/Ni = 48.7/19.2/32.1
CeNi5	CaCu5	52.5	43.3	31.3	0.131	—	—	—
CeAlNi4	CaCu5	49.1	42.7	66.9	0.089	Ce/Al/Ni = 17.3/26.5/56.1	Ce/Al/Ni = 18.6/14.5/66.9	Ce/Al/Ni = 5.8/47.1/47.1
CeNi4Si	DyNi4Si	15.7	43.9	16.7	0.099	—	—	—
Ce(NiSi)2	Th(CrSi)2	21.9	36.7	33.4	0.187	—	—	—
SmNi3	PuNi3	41.9	36.0	6.8	0.038	—	—	—
SmNi4Si	DyNi4Si	19.9	37.0	14.7	0.138	—	—	—
Sm(NiSi)2	Th(CrSi)2	45.0	45.2	25.0	0.179	Sm/Ni/Si = 19.8/36.2/44.0	Sm/Ni/Si = 24.8/44.1/31.1	Sm/Ni/Si = 24.1/6.3/69.5

---

For representative samples, we briefly outline their crystal structure features. CeNi₅ adopts the CaCu₅-type crystal structure (hexagonal, P6/mmm, Z = 1),¹⁹ as illustrated in Fig. 2(a). This structure consists of alternating layers of Ni₃ kagome layer (A) and a closed-packed layer of CeNi₂ (B), stacked along the [001] direction in an ABAB sequence. In this arrangement, the larger Ce ion is nestled between two Ni₆ hexagons of the kagome layer. Conversely, ZrNi₅ exhibits the AuBe₅-type crystal structure (cubic, F4̄3m, Z = 4),²⁰ as presented in Fig. 2(b). In this case, Ni₃ kagome layer (A) and ZrNi₂ closed-packed layer (B) are stacked in an alternating fashion along the 〈111〉 direction, following the sequence of (AB)(AB)′(AB)″(AB)(AB)′(AB)″. Due to the smaller size of the Zr ion, the ZrNi₂ closed-packed layer displays a buckled structure, resulting in the formation of a Frank–Kasper polyhedron around the Zr atoms. Consequently, the packing density of Ni in ZrNi₅ is enhanced compared to that in CeNi₅. The stacking arrangement observed in ZrNi₅ is reminiscent of the cubic Laves-type crystal structure (XY₂, Fd3̄m, Z = 8),²¹ where the Y₃ kagome layer (A) and X₂Y closed-packed layer (B) are also alternately stacked along the 〈111〉 direction in the same pattern. In ZrNi₅, there are two distinct crystallographic sites for Ni (4c and 16e). The formation of Zr(ZnNi₄) involves substituting Ni (4c site) with Zn, leading to a 1.5% increase in the cubic lattice constant.²²

Fig. 2 (a) The crystal structure of CeNi₅, which includes the Ni₃ kagome layer and the CeNi₂ closed-packed layer, and (b) the crystal structure of ZrNi₅, composed of the Ni₃ kagome layer and a buckled ZrNi₂ closed-packed layer.

The electronic structure of ZrNi₅, which serves as a parent compound to ZrZnNi₄, was examined using DFT calculations. Fig. 3(a) and (b) illustrate the electronic structures of ZrNi₅ and ZrZnNi₄, including the energy–wave vector (E–k) diagram and the projected density of states (DOS). The energy scale is set such that the Fermi energy (E_F) is at zero. Several energy bands intersect E_F, suggesting a metallic nature of the electronic structure. A significant DOS is observed primarily from the Ni 3d orbitals in the energy range of −3.0 to +0.3 eV. Notably, there is a pseudo bandgap around +1.9 eV, as indicated in the E–k diagram. The contribution of Zr 4d character is apparent in the unoccupied band with a narrow bandwidth around +2.5 eV. According to the Badar charge analysis, the atomic charges are estimated as Zr^+1.58Ni^−0.32₅ for ZrNi₅ and Zr^+1.60Zn^+0.11Ni^−0.43₄ for ZrZnNi₄. This pattern of charges aligns well with those derived from electronegativity values (Zr 1.22, Zn 1.66, Ni 1.75).²³ Negatively charged late transition metal ions, such as Ni^δ−, typically play an important role in hydrogenation processes for hydrogen storage applications.²⁴ The substitution of Zn, which has a relatively positive character, promotes the reduction of the remaining Ni ions, resulting in an increase in E_F energy. This shift in energy is confirmed by the unoccupied Zr 4d band, which is situated around +1.7 eV within the total DOS of ZrZnNi₄ (Fig. 3(b)). A subtle indication of Zn substitution can be seen in the Zn 3d¹⁰ band located approximately at −7.2 eV, which does not contribute to the electronic structure near E_F (not shown in Fig. 3(b)). Now, let's consider the changes brought about by alloying Zr with Ni. We show calculated DOS of metallic Ni in Fig. 3(c), as a reference. The addition of Zr into Ni leads to two effects. One is the redox reaction, that is, electron transfer from Zr to Ni. The energy levels of the Zr bands are pushed lower due to the oxidation of Zr, while the Ni bands are elevated because of increased electron repulsion resulting from the reduction of Ni. Consequently, the WF of ZrNi₅ decreases compared to that of Ni metal, since the DOS at E_F in ZrNi₅ is dominated by the Ni 3d bands. This prediction is supported by the measured WF value of 4.2 eV, which is considerably lower than that of Ni metal (4.9 eV). The energy shift of WF (0.7 eV) is consistent with the shift (∼0.8 eV) of unoccupied Zr 4d band mentioned above. Another effect of addition of Zr is drastic decrease of DOS(E_F). Metallic Ni has large DOS(E_F), which often induces magnetic ordering, because of the (3d4s) electronic configuration. Electron-doping into Ni by Zr fills remaining unoccupied 3d band, resulting in the decrease of DOS(E_F), whereas the WF decreases. Thus, catalytic activity due to the electron of Ni 3d band has the tradeoff relation.

Fig. 3 (a) The calculated band structure and projected density of states for ZrNi₅. (b) The calculated total density of states for ZrZnNi₄. (c) The calculated total density of states of Ni. Magnetic ordering is not included. The calculations were performed on the primitive unit cell to reduce computational time.

Subsequently, the porosity of the prepared intermetallic compound nanopowders was assessed through nitrogen adsorption/desorption experiments. Fig. 4 and Fig. S6–S10 display the nitrogen adsorption/desorption isotherms along with the corresponding pore size distributions. The measured BET surface areas and pore volumes are collated in Table 1. A distinct hysteresis in the isotherms, with pore formation occurring at 3.8 nm, was evident in ZrZnNi₄. A negligible hysteresis was observed in CeAlNi₄, also featuring pore formation at 3.8 nm. This pore creation likely contributes to the high BET surface areas of 42.0 m² g⁻¹ for ZrZnNi₄ and 66.9 m² g⁻¹ for CeAlNi₄. Conversely, other samples such as CeNi₅, CeNi₄Si, Ce(NiSi)₂, SmNi₃, SmNi₄Si, and Sm(NiSi)₂ did not exhibit hysteresis, indicating no small pore formation below 10 nm.

Fig. 4 (a) The adsorption (blue) and desorption (orange) isotherms of nitrogen, as well as (b) the corresponding pore size distributions for ZrZnNi₄, CeAlNi₄, and Sm(NiSi)₂.

The morphologies of representative samples ZrZnNi₄, CeAlNi₄, and Sm(NiSi)₂, which contain Zr, Ce, and Sm, respectively, were examined via SEM (Fig. S11–S13). Numerous fine particles measuring less than a micrometer were distinctly visible across all samples, and the surfaces appeared to possess nanoscale structures, leading to elevated specific surface areas. Fig. 5 presents the elemental mappings for the constituent elements of ZrZnNi₄, CeAlNi₄, and Sm(NiSi)₂. The constituent elements appeared well-distributed among the particles, with overlapping distributions indicating a thorough mixture of the components. It is noted that the ideal spherical particles were observed in CeAlNi₄, which are enclosed by rectangles in Fig. 5(b). According to the elemental mappings, the core of the particles had dense concentrations of Ni, whereas Ce was distributed in the shell. Thus, the results indicated that the formation of core–shell structures.

Fig. 5 SEM images and elemental mappings for (a) ZrZnNi₄, (b) CeAlNi₄, and (c) Sm(NiSi)₂, with corresponding EDX spectra shown in Fig. S14.

The compositions of ZrZnNi₄, CeAlNi₄, and Sm(NiSi)₂ were analyzed using SEM-EDX and XRF, with the results summarized in Table 1. The spectra obtained from SEM-EDX and XRF are shown in Fig. S14 and S15–S17, respectively. In the case of ZrZnNi₄, the molar ratio of Zn was lower than the stoichiometric ratio expected for the AuBe₅-type intermetallic compound with a Ni site substituted by Zn, indicating that the Ni site was not fully occupied by a Zn atom, resulting in a partial deficiency of Zn. For CeAlNi₄, the molar ratio of Ni was slightly less than what would be expected for the CaCu₅-type intermetallic compound with an Al atom substituted for Ni, suggesting a slight deficiency in Ni. For Sm(NiSi)₂, the molar ratio of Sm/Ni/Si was nearly in accordance with the stoichiometric ratio for the Th(CrSi)₂-type intermetallic compound, indicating successful synthesis of Sm(NiSi)₂. Detailed analysis through XRF revealed very minimal amounts of impurities in ZrZnNi₄, CeAlNi₄, and Sm(NiSi)₂ (Tables S1–S3). It is noteworthy that the detected amounts of Ca were minimal in all samples, suggesting that impurities originating from CaH₂, including CaH₂, CaO, Ca(OH)₂, and CaCl₂, were effectively removed through washing with a weak acid solution during the preparation process.

Catalytic performance of intermetallic compound nanopowders

The NaBH₄-assisted hydrogenation of 4-nitrophenol (4NP) to 4-aminophenol (4AP) follows the Langmuir–Hinshelwood model, making it crucial to examine the surface species that directly facilitate this reaction. Consequently, surface-sensitive XPS measurements were conducted on ZrZnNi₄, CeAlNi₄, and Sm(NiSi)₂. Fig. 6 presents the XPS spectra calibrated by aligning the C1s peak of surface carbonaceous contaminants to 284.5 eV (see Fig. S18). The resulting surface compositions are summarized in Table 1.

Fig. 6 XPS spectra for (a) Zr 3p5/2, Zn 2p3/2, Ni 2p3/2 for ZrZnNi₄, (b) Ce 3d5/2, Al 2p, Ni 2p3/2 for CeAlNi₄, and (c) Sm 3d5/2, Si 2p, Ni 2p3/2 for Sm(NiSi)₂.

For ZrZnNi₄ shown in Fig. 6(a), a prominent peak attributed to Ni 2p3/2 was detected at 855.5 eV. Previous studies indicate that a peak associated with Ni(+2) in Ni(OH)₂ appears at 856.2 eV, while a peak for Ni(0) is found at 852.6 eV.²⁵ Hence, the oxidation state of surface nickel in ZrZnNi₄ is close to Ni(+2). Additionally, a peak corresponding to Zr 3d5/2 appeared at 181.5 eV. As reported earlier, peaks linked to Zr(+4) in ZrO₂ and Zr(0) can be found at 182.75 eV and 178.52 eV, respectively.²⁶ This suggests that the surface oxidation state of zirconium in ZrZnNi₄ is Zr(+4>, >0), or greater than zero. A peak corresponding to Zn 2p3/2 was observed at 1,021.1 eV. According to prior reports, Zn 2p lines are not highly sensitive to chemical environments, making it challenging to differentiate between Zn(0) (1,021.4 eV) and Zn(+2) (1,021.7 eV).²⁷ These findings indicate that the surface components of ZrZnNi₄ largely exist in higher oxidation states, such as Ni(+2) and Zr(+4>, >0). The measured surface composition of Zr/Zn/Ni = 48.7/19.2/32.1 deviated significantly from those obtained by SEM-EDX and XRF, with the molar ratio of Ni being considerably lower than the stoichiometric ratio; thus, Zr–Zn–O comprised the primary surface components. Because no oxide phases were observed in the XRD measurement (Fig. 1(a)), the thickness of the surface oxide layers could be very thin or/and the crystallinity was very low enough not to be detected by XRD.

In the case of CeAlNi₄, depicted in Fig. 6(b), the main peak corresponding to Ni 2p3/2 was recorded at 856.2 eV, which identifies the oxidation state of surface nickel in CeAlNi₄ as Ni(+2). A peak associated with Ce 3d5/2 was found at 881.3 eV. Previous studies identified peaks for Ce(+4) in CeO₂ and Ce(+3) in CePO₄ at 882.7 eV and 880.9 eV, respectively.²⁸ Therefore, the oxidation state of surface cerium in CeAlNi₄ approximates Ce(+3). A peak for Al 2p was detected at 73.5 eV, with prior reports indicating that peaks for Al(+3) in Al₂O₃ and Al(0) appear at 74.14 eV and 72.87 eV respectively.^29,30 Therefore, the oxidation state of surface Al in CeAlNi₄ is Al(+3>, >0), or greater than zero. These results indicate that the surface compounds in CeAlNi₄ primarily exist in higher oxidation states, such as Ni(+2), Ce(+3), and Al(+3>, >0). The measured surface composition of Ce/Al/Ni = 5.8/47.1/47.1 showed significant deviation from SEM-EDX and XRF results, with the molar ratios of Al and Ni being considerably higher than the stoichiometric ratio; consequently, Al–Ni–O emerged as the main surface components. Because no oxide phases were observed in the XRD measurement (Fig. 1(b)), the thickness of the surface oxide layers could be very thin or/and the crystallinity was very low enough not to be detected by XRD.

For Sm(NiSi)₂, as shown in Fig. 6(c), a significant peak assigned to Ni 2p3/2 was noted at 856.2 eV, indicating that the oxidation state of surface Ni in Sm(NiSi)₂ is Ni(+2). A peak corresponding to Sm 3d5/2 appeared at 1082.8 eV. Previous reports indicate that the peaks for Sm(+3) and Sm(+2) are found at 1084.0 eV and 1076.4 eV, respectively.³¹ Therefore, the oxidation state of surface Sm in Sm(NiSi)₂ appears to be close to Sm(+3). A peak corresponding to Si 2p was observed at 101.8 eV, with previous reports showing that peaks for Si(+4) in SiO₂ and Si(0) are found at 103.8 eV and 99.8 eV, respectively.³² Thus, the oxidation state of surface Si in Sm(NiSi)₂ is Si(+4>, >0). These results point to surface compounds in Sm(NiSi)₂ mostly existing in higher oxidation states, such as Ni(+2), Sm(+3), and Si(+4>, >0). Additionally, the measured surface composition of Sm/Ni/Si = 24.1/6.3/69.5 significantly differed from those obtained via SEM-EDX and XRF. The Ni molar ratio was notably smaller than the stoichiometric ratio, and thus Sm–Si–O was determined to be the main surface component. Because no oxide phases were observed in the XRD measurement (Fig. 1(c)), the thickness of the surface oxide layers could be very thin or/and the crystallinity was very low enough not to be detected by XRD.

Next, the hydrogenation of 4-nitrophenol (4NP) to 4-aminophenol (4AP) was facilitated by NaBH₄ at room temperature using ZrZnNi₄, CeAlNi₄, and Sm(NiSi)₂. The three intermetallics were chosen as representative catalysts containing very low WF metals of Zr, Ce, and Sm, respectively, with high BET surface areas available for the catalytic reactions. 12 wt%NiO/CeO₂ and 10 wt%Ni/CeO₂ were used as a reference. XRD analysis revealed that the reference catalysts of NiO/CeO₂ and Ni/CeO₂ contained minuscule peaks of NiO and Ni, respectively, alongside significant peaks corresponding to CeO₂ (Fig. S19), indicating that small particles of NiO and Ni were successfully dispersed on CeO₂, respectively. In Fig. 7(a), the change in absorbance of the reaction mixture is plotted over time with the ZrZnNi₄ catalyst. The absorbance at 401 nm, associated with the concentration of 4NP, rapidly decreased over time, while the absorbance at 315 nm, which corresponds to 4AP, increased as the reaction progressed. By converting these absorbance changes into concentration variations, the plots showing concentration changes of 4NP and 4AP were obtained, as illustrated in Fig. 7(b). It was confirmed that the conversion of 4NP to 4AP occurred in a nearly stoichiometric manner. Fig. 7(c) illustrates the normalized concentration changes (C/C₀) of 4NP with ZrZnNi₄, CeAlNi₄, Sm(NiSi)₂, and the reference catalysts of 12 wt%NiO/CeO₂ and 10 wt%Ni/CeO₂. The concentrations showed a gradual decrease with CeAlNi₄ and NiO/CeO₂, achieving nearly 80% conversion in 60 minutes. In contrast, ZrZnNi₄, Sm(NiSi)₂, and Ni/CeO₂ achieved complete conversion within 60 minutes, with ZrZnNi₄ reaching 100% conversion in just 10 minutes, indicating its superior catalytic activity. In comparison with NiO/CeO₂ and Ni/CeO₂, it was suggested that metallic state Ni was effective to achieve the quick complete conversion. From the plots of ln(C/C₀) against reaction time (Fig. 7(d)), the rate constants were determined as 0.16 min⁻¹ for ZrZnNi₄, 0.04 min⁻¹ for CeAlNi₄, 0.08 min⁻¹ for Sm(NiSi)₂, 0.04 min⁻¹ for NiO/CeO₂, and 0.06 min⁻¹ for Ni/CeO₂. Comparing these constants to our previous studies conducted under the same reaction conditions but with different catalysts (Table 2), ZrZnNi₄ demonstrated a higher catalytic performance than the alloy catalysts CrMnFeCoNi and AlCoCrFeNi, yet lower performance than 5 wt%Pd/ZnO and Al_0.2Co_1.5CrFeNi_1.5Ti_0.5.

Fig. 7 (a) Changes in absorbance and (b) the concentration shifts of 4NP and 4AP over time in the reaction solution utilizing ZrZnNi₄ catalyst. (c) Changes in normalized concentration (C/C₀) of 4NP during the reactions and (d) the linear relationship of ln(C/C₀) over time for 12 wt%NiO/CeO₂ (X), 10 wt%Ni/CeO₂ (diamond), ZrZnNi₄ (square), CeAlNi₄ (circle), and Sm(NiSi)₂ (triangle) catalysts.

Table 2 The comparison of rate constants (k) for the hydrogenation of 4NP

Catalyst	Conditions	k [min^−1]	Ref.
ZrZnNi4	25 °C; 4-NP (1.6 mM); NaBH4 (47 mM); 10 mg-cat/9 mL	0.16	This work
CeAlNi4		0.04
Sm(NiSi)2		0.08
12 wt%NiO/CeO2		0.04
10 wt%Ni/CeO2		0.06
5 wt%Pd/ZnO		0.44	33
Al0.2Co1.5CrFeNi1.5Ti0.5		0.29	34
CrMnFeCoNi		0.11	35
AlCoCrFeNi		0.03	36

---

The reaction rates were observed in the following order: ZrZnNi₄ > Sm(NiSi)₂ > CeAlNi₄, NiO/CeO₂. XPS results indicated that a minimal quantity of Ce was present on the surface of CeAlNi₄, with Al–Ni–O being the primary surface component. This suggests that the active Ni species may not have effectively interacted with Ce, limiting electron transfer from Ce to Ni and thereby resulting in reduced catalytic performance for CeAlNi₄. Conversely, XPS analysis showed that Zr and Sm were predominantly present on the surfaces of ZrZnNi₄ and Sm(NiSi)₂, respectively, implying that active Ni species could interact effectively with Zr and Sm. This efficient electron donation from Zr and Sm to Ni likely activated the Ni species, leading to higher catalytic performance in ZrZnNi₄ and Sm(NiSi)₂. Notably, DFT calculations revealed that ZrZnNi₄ acts as a reductive intermetallic compound with the capability to release electrons, which aids in the generation of reactive hydrogen and accelerates the rate-determining step (eqn (8)). Although it was suggested that the synthesized ZrZnNi₄ had a core–shell structure that very thin oxide layers covered the intermetallic-phase core, it is possible that the electron-rich Ni species (Ni^δ−), suggested by DFT calculations, may be partially exposed on surface through the pit-hole of the oxide shell, or/and the crystallized intermetallic ZrZnNi₄ existing in the core may promote the surface reaction due to the electron donation through the thin oxide shell. Therefore, the intermetallic compound ZrZnNi₄ stands out as a distinct catalyst, with its electron-rich Ni species effectively facilitating the NaBH₄-assisted hydrogenation of 4NP.

Conclusions

Eight intermetallic compounds containing Zr, Ce, or Sm were synthesized through the reduction of metal oxides. The catalysts ZrZnNi₄ (42.0 m² g⁻¹), CeAlNi₄ (66.9 m² g⁻¹), and Sm(NiSi)₂ (25.0 m² g⁻¹) were subsequently evaluated for their catalytic activity in the NaBH₄-assisted hydrogenation process of 4-nitrophenol. In comparison with the widely used Ni/CeO₂ catalyst, both ZrZnNi₄ and Sm(NiSi)₂ exhibited superior catalytic performance. Given that the WFs of Zr and Sm are considerably lower than that of Ni, which serves as the active site for hydrogenation, it was hypothesized that this high catalytic performance arises from the generation of electron-rich Ni species, particularly in the case of ZrZnNi₄.

Author contributions

Yasukazu Kobayashi: conceptualization, supervision, funding acquisition, investigation, methodology, data curation, writing – original draft, writing – review & editing. Hiroshi Mizoguchi: investigation, methodology, data curation, writing – review & editing. Koharu Yamamoto: investigation, methodology, data curation, writing – review & editing. Ryo Shoji: methodology, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

The authors declare that the data supporting the findings of this study are available within the paper and its supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6dt00513f.

Should any raw data files be needed in another format they are available from the corresponding author, Yasukazu Kobayashi, upon reasonable request.

Acknowledgements

This work was supported by JSPS KAKENHI Grant Numbers 24K08591 and 23K23440. A part of this work was supported by “Advanced Research Infrastructure for Materials and Nanotechnology in Japan (ARIM)” of the Ministry of Education, Culture, Sports, Science and Technology (MEXT). Proposal Number JPMXP125NM5175. We acknowledge Mr Koei Takagi for his help in review process.

References

1. Z. Xiong, H. Zhang, W. Zhang, B. Lai and G. Yao, Chem. Eng. J., 2019, 359, 13–31.
2. P. K. Arora, A. Srivastava and V. P. Singh, J. Hazard. Mater., 2014, 266, 42–59.
3. T. Aditya, A. Pal and T. Pal, Chem. Commun., 2015, 51, 9410–9431.
4. S. Wunder, F. Polzer, Y. Lu, Y. Mei and M. Ballauff, J. Phys. Chem. C, 2010, 114, 8814–8820.
5. R. Retnamma, A. Q. Novais and C. M. Rangel, Int. J. Hydrogen Energy, 2011, 36, 9772–9790.
6. J. Strachan, C. Barnett, A. F. Masters and T. Maschmeyer, ACS Catal., 2010, 10, 5516–5521.
7. M. I. Din, R. Khalid, Z. Hussain, T. Hussain, A. Mujahid, J. Najeeb and F. Izhar, Crit. Rev. Anal. Chem., 2020, 50, 322–338.
8. K. Zhang, J. M. Suh, J.-W. Choi, H. W. Jang, M. Shokouhimehr and R. S. Varma, ACS Omega, 2019, 4(1), 483–495.
9. P. Zhao, X. Feng, D. Huang, G. Yang and D. Astruc, Coord. Chem. Rev., 2015, 287, 114–136.
10. T. R. Mandlimath and B. Gopal, J. Mol. Catal. A:Chem., 2011, 350, 9–15.
11. M. S. Qatan, F. Arshad, M. Miskam and G. A. Naikoo, Int. J. Environ. Sci. Technol., 2024, 21, 5247–5268.
12. H. B. Michaelson, J. Appl. Phys., 1977, 48, 4729–4733.
13. Y. Kobayashi, S. Tada and R. Kikuchi, Mater. Trans., 2020, 61, 1037–1040.
14. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
15. P. E. Blöchl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 50, 17953.
16. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169.
17. E. Sanville, S. Kenny, R. Smith and G. Henkelman, J. Comput. Chem., 2007, 28, 899–908.
18. K. Momma and F. Izumi, J. Appl. Crystallogr., 2011, 44, 1272–1276.
19. F. Pourarian and W. E. Wallace, J. Less-Common Met., 1982, 87, 275–281.
20. V. Babizhetskyy, O. Myakush, A. Simon and B. Kotur, Intermetallics, 2013, 38, 44–48.
21. E. Parthe, Z. Kristallogr. – Cryst. Mater., 2006, 221, 301–304.
22. Z. Blazina, A. Drasner and Z. Ban, J. Nucl. Mater., 1981, 96, 141–146.
23. A. L. Allred and E. G. Rochow, J. Inorg. Nucl. Chem., 1958, 5, 264–268.
24. H. Mizoguchi, S. W. Park and H. Hosono, J. Am. Chem. Soc., 2021, 143, 11345–11348.
25. H. Nesbitt, D. Legrand and G. Bancroft, Phys. Chem. Miner., 2000, 27, 357–366.
26. C. Morant, J. M. Sanz, L. Galán, L. Soriano and F. Rueda, Surf. Sci., 1989, 218, 331–345.
27. M. C. Biesinger, L. W. M. Lau, A. R. Gerson and R. S. C. Smart, Appl. Surf. Sci., 2010, 257, 887–898.
28. E. Bêche, P. Charvin, D. Perarnau, S. Abanades and G. Flamant, Surf. Interface Anal., 2007, 40, 264–267.
29. P. M. A. Sherwood, Surf. Sci. Spectra, 1998, 5, 1–3.
30. C. Chen, S. J. Splinter, T. Do and N. S. McIntyre, Surf. Sci., 1997, 382, L652–L657.
31. J. Zhu, Y. Ma and S. Zhuang, Surf. Sci., 2009, 603, 1802–1811.
32. R. Alfonsetti, L. Lozzi, M. Passacantando, P. Picozzi and S. Santucci, Appl. Surf. Sci., 1993, 70–71, 222–225.
33. Y. Kobayashi, K. Yamamoto and R. Shoji, Crystals, 2024, 14(3), 278.
34. Y. Kobayashi, S. Yokoyama and R. Shoji, Metals, 2024, 14(4), 443.
35. Y. Kobayashi, H. Y. Teah, S. Yokoyama, R. Shoji and N. Hanada, ACS Sustainable Chem. Eng., 2022, 10(46), 15046–15057.
36. Y. Kobayashi, H. Y. Teah, S. Yokoyama, R. Shoji and N. Hanada, Int. J. Hydrogen Energy, 2023, 48, 30963–30973.

---