Development of non-noble Ni metal-based (Yb_1−xCo_x)₂O_3−δ catalysts for green H₂ production via ammonia decomposition

Abstract

We herein report the preparation of non-noble metal supported Yb₂O₃–Co₃O₄ inorganic catalysts for application in H₂ production via ammonia (NH₃) decomposition. The NH₃ decomposition reaction for H₂ production was studied using αwt%Ni/(Yb_1−xCo_x)₂O_3−δ catalysts according to various Co dopant ratios and Ni loadings to evaluate the dependence of oxygen vacancies as active sites. The introduction of Co₃O₄ into Yb₂O₃ with a Ni metal support improved the oxygen vacancy density of the catalyst to enhance NH₃ decomposition activity for H₂ production. The electron-deficient nature of oxygen vacancies enabled the catalysts to act as electron acceptors, facilitating the associative N desorption from NH₃ to produce N₂ in the NH₃ cracking process. The structure–property relationship established in these studies suggests that the oxygen vacancies on the αwt%Ni/(Yb_1−xCo_x)₂O_3−δ catalysts serve as active sites for NH₃ decomposition. The 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst showed the highest catalytic activity, which exhibited an NH₃ conversion of 100% at temperatures as low as 550 °C and the constant H₂ production rate of 6.23 kg per day.

---

Introduction

As environmental pollution continues to increase, the demand for eco-friendly energy is growing, leading to a stronger emphasis on regulating the use of fossil fuels.¹ Hydrogen has attracted considerable attention in this regard as an abundant, clean, and high-energy content resource, leading to potential applications in industry, transport, and power sectors. However, the storage and transportation of hydrogen pose significant challenges due to its low volumetric density.^2,3 In recent years, ammonia (NH₃) has emerged as a prominent hydrogen source due to its high volumetric energy density, low production cost, and CO_x-free hydrogen production.^4,5 Over the past decades, the catalytic conversion of NH₃ to H₂ has garnered considerable interest to obtain CO_x-free hydrogen. Various metal catalysts such as Ru, Ir, Rh, Pd, Pt, Ni, Co, and Fe supported on Al₂O₃, carbon nanotubes (CNTs), activated carbon (AC), SiO₂, MgO, ZrO₂, TiO₂, and rare earth oxides have been extensively investigated for the ammonia decomposition reaction.^6,7 Additionally, Ru-based catalysts have demonstrated excellent activity for NH₃ decomposition into hydrogen; however, the large-scale use of Ru is not feasible due to its high cost and scarcity.⁸ Therefore, derivatives of other transition metals, such as Ni, Fe, and Co have been extensively explored for NH₃ decomposition. Additionally, using CNTs or CNFs as supports and alkali promoters among Fe, Co, and Ni, results in a remarkable improvement in activity at temperatures greater than 600 °C; however, at low temperatures, i.e., below 600 °C, a considerably lower activity compared to that of Ru has been observed.⁹ Fe, Ni, and Co nanoparticles dispersed on an alumina matrix, prepared via co-precipitation, were studied at 90% loading, where a 99% conversion was achieved at 600 °C.¹⁰ Therefore, it is necessary to develop inexpensive and highly active non-noble metal-based catalysts capable of achieving 100% NH₃ conversion at temperatures as low as 550 °C for NH₃ decomposition. Among non-noble metals, Ni provides catalytically efficient active sites for NH₃ decomposition, and various Ni-based catalysts for catalytic NH₃ decomposition at low operating temperatures have been explored. For instance, 90% NH₃ conversion was achieved at 600 °C using a 10wt%Ni/CeO₂-BN (boron nitride) catalyst (GHSV = 30 000 mL h⁻¹ g_cat⁻¹).¹¹ Gan et al. reported 82% NH₃ conversion at 550 °C (GHSV = 30 000 mL h⁻¹ g_cat⁻¹) using a 40wt%Ni/La₂O₃ catalyst.¹² Okura et al. prepared several formulations of ABO₃ perovskites via a solid-state reaction, achieving NH₃ conversions of 95% and 93% at 550 °C using 40wt%Ni/BaZrO₃ and 40wt%Ni/SrZrO₃, respectively (flow rate: 80 mL min⁻¹, catalyst amount: 300 mg).¹³ However, these Ni-supported metal oxide catalysts exhibited low catalytic activity for NH₃ decomposition at temperatures as low as 600 °C and required a high Ni loading (over 40 wt%). Thus, NH₃ decomposition still necessitates high activation temperatures and a large Ni loading as the active site, highlighting the need for novel catalysts that are efficient for NH₃ decomposition at lower temperatures.

To achieve efficient NH₃ decomposition, it is important to design a novel non-noble metal (e.g., Fe, Co, and Ni)-supported catalyst that accelerates the associative N desorption from NH₃ to produce N₂, which is the rate-determining step (RDS).⁷ In this study, C-type cubic rare earth oxides were utilized as promoters for NH₃ decomposition. Rare earth oxides can form three types of crystal structures—A-type (hexagonal), B-type (monoclinic), and C-type (cubic)—depending on the ionic size of the respective rare earth element.¹⁴ Among these, the C-type cubic solid possesses large interstitial open spaces, derived from a fluorite-type structure by the removal of one-quarter of the oxide ions,¹⁵ wherein these open spaces play an important role as oxygen vacancy sites in the NH₃ decomposition process. These oxygen vacancy sites exhibit electron-deficient characteristics with a positive charge, enabling them to act as electron acceptors in the NH₃ cracking process.⁷ In a previous report, ytterbium(III) oxide (Yb₂O₃) with a C-type cubic structure was used as the catalyst for direct N₂O decomposition, wherein the highest conversion of N₂O to N₂ and O₂, occurred at a low temperature of 500 °C.^16,17 This efficiency is attributed to the oxygen vacancies in Yb₂O₃, which function as both N₂O adsorption sites and catalytic active sites. Yb₂O₃ was also used for the conversion of alcohols and dehydrogenation of tetralin. Yb₂O₃ exhibited the highest dehydrogenation and dehydration activities in comparison to those of Dy₂O₃, Gd₂O₃, and Tm₂O₃.¹⁸ It is believed that these oxygen vacancy sites and high dehydration activity of Yb₂O₃ enhance the catalytic NH₃ decomposition activity for H₂ production. Hence, Yb₂O₃ was selected as a promoter for catalytic NH₃ decomposition in this study. Yb₂O₃ has attracted considerable interest for its applicability in numerous applications; however, to the best of our knowledge, there has been no prior report on NH₃ decomposition using Yb₂O₃ as a catalyst.

In the present study, we introduced Co₃O₄ into Yb₂O₃, resulting in an Yb₂O₃–Co₃O₄ solid solution to improve NH₃ decomposition activity. This enhancement is attributed to the improved redox properties and the increased oxygen vacancies created by the substitution of low-valent Co²⁺ into Yb³⁺ sites, which serve as active sites for NH₃ decomposition. Hence, we prepared Ni-supported (Yb_1−xCo_x)₂O_3−δ catalysts and investigated their catalytic activities for NH₃ decomposition to produce H₂.

Experimental

Synthetic procedures

The (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.12) promoters were prepared using a wet impregnation method. A stoichiometric ratio of Yb(NO₃)₃ (0.1 mol L⁻¹, Sigma Aldrich, 99%) and Co(NO₃)₂ (0.1 mol L⁻¹, Sigma Aldrich, 98%) was mixed with a 3 mol L⁻¹ solution of HNO₃ (30 mL, DAEJUNG, 70%). After stirring at room temperature for 30 minutes, PVP (polyvinylpyrrolidone, DAEJUNG, 2.1 times the total molar amount of cations) was added, and the mixture was stirred at 80 °C for 6 hours. After this time, the solvent was removed on a heated agitator at 180 °C, and the resulting powder was heated at 350 °C for 4 h to remove the PVP. Finally, the powder was subjected to calcination at 500 °C for 6 h under atmospheric pressure.

Ni was loaded onto the (Yb_0.92Co_0.08)₂O_3−δ promoter by the impregnation method. The (Yb_0.92Co_0.08)₂O_3−δ promoter was impregnated with Ni by suspending the promoter powder (0.4 g) in 0.1 mol L⁻¹ Ni(NO₃)₂ aqueous solution (DAEJUNG, 98%) and adding a 3 mol L⁻¹ solution of HNO₃ (30 mL, DAEJUNG, 70%), where the Ni loading ranged from 0 to 20 wt%. After subsequent stirring at room temperature for 30 min, PVP (polyvinylpyrrolidone, DAEJUNG, 2.1 times the total molar amount of cations) was added and the mixture was stirred at 80 °C for 6 h. After this time, the solvent was removed on a heated agitator at 180 °C, and the resulting powder was heated at 350 °C for 4 h to remove the PVP. The residue was ground and then calcined at 400 °C for 4 h under atmospheric pressure and reduced at 500 °C for 2 h under a H₂ gas flow to obtain the corresponding αwt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (0 ≤ α ≤ 20) catalysts.

Materials and methods

The detailed synthesis procedure of the catalysts is described in the SI. The crystal structures of the catalysts were identified via X-ray powder diffraction (XRD, Bruker D8 Advance) using Cu-Kα radiation (40 kV, 40 mA) in the 20–80° 2θ range with a step size of 0.02° and a scan speed of 0.2° min⁻¹. Structural parameters were determined using Rietveld refinements with the RIETAN-FP program.¹⁷ The crystallite size was calculated from the measured XRD peak data using Scherrer's equation. The Brunauer–Emmett–Teller (BET) specific surface areas were measured at −196 °C using N₂ gas (TriStar 3020, Shimadzu). The O₂, H₂ and N₂ desorption sites were evaluated using the temperature-programmed desorption method using oxygen, hydrogen, and nitrogen (O₂-TPD/H₂-TPD/N₂-TPD, Micromeritics AutoChem II 2920 Chemisorption Analyzer). After preheating the sample (0.05 g) at 500 °C for 1 hour under a flow of helium gas (50 mL min⁻¹), the sample was exposed to oxygen (10 vol% O₂/90 vol% He, 50 mL min⁻¹), hydrogen (10 vol% H₂/90 vol% He, 50 mL min⁻¹) and nitrogen (10 vol% N₂/90 vol% He, 50 mL min⁻¹) at 50 °C for 30 min, followed by purging at the same temperature for 15 min under a helium flow (50 mL min⁻¹). Subsequently, the temperature was increased at a rate of 10 °C min⁻¹ under a helium flow at a rate of 30 mL min⁻¹ (BELCAT-B, Microtrac BEL). Additionally, the value of oxygen vacancy density was calculated from oxygen desorption sites and a more detailed calculation process is presented in the SI. H₂-temperature-programmed reduction (H₂-TPR) was conducted under a 5 vol% H₂/95 vol% Ar gas flow (50 mL min⁻¹) at a heating rate of 5 °C min⁻¹ (Micromeritics AutoChem II 2920 Chemisorption Analyzer). Subsequently, the oxygen storage capacity (OSC) and Ni dispersion was investigated using the reduced sample, by means of the pulse-injection method at 500 °C (Micromeritics AutoChem II 2920 Chemisorption Analyzer). X-ray photoelectron spectroscopy (XPS, AXIS SUPRA) data were acquired using Mg-Kα radiation (1253.6 eV); the spectra were fitted using the Shirley background and Gaussian–Lorentzian lineshape analysis.

The NH₃ decomposition activity was measured in a fixed-bed flow reactor consisting of a 10 mm-diameter quartz glass tube. The mixed gas comprised 10 vol% NH₃ balanced by N₂ at a total flow rate of 50 mL min⁻¹ (mass hourly space velocity = 30 000 L kg⁻¹ h⁻¹), which was passed over 0.1 g of the catalyst. The gas was supplied to the catalyst bed using a mass flow controller. The catalysts were reduced at 500 °C for 2 hours under a H₂ flow (50 mL min⁻¹) prior to the measurements. The concentration of NH₃ gas at the outlet was analyzed using a gas chromatograph (GC, GC-8AIT, Shimadzu) equipped with a thermal conductivity detector. The reaction temperature was increased from 150 °C to 500 °C in 50 °C intervals, and NH₃ conversion at each interval was obtained after stabilization for 15 minutes.

Results and discussion

Fig. 1 presents the X-ray powder diffraction (XRD) patterns of the (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.12) samples. The diffraction peaks of the samples indicate a single phase, cubic (C-type) rare-earth sesquioxide structure. The peaks of the cubic phase shift to higher angles as the Co content increases up to x = 0.08, suggesting the replacement of Yb³⁺ (0.101 nm for coordination number (CN) = 6)¹⁹ sites by Co^3+/2+ with smaller ionic sizes (Co³⁺: 0.063 nm, Co²⁺: 0.072 nm for CN = 6).¹⁹ The composition determined by inductively coupled plasma-mass spectrometry (ICP-MS) analysis of the (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.10) samples are shown in Table S3. The measured compositions of the (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.10) samples are evidently in good agreement with the feed values. The crystal structure of (Yb_1−xCo_x)₂O_3−δ was investigated using Rietveld refinement analysis (Fig. S1) for the Yb₂O₃ and (Yb_0.92Co_0.08)₂O_3−δ samples, in which the crystal structure data of Yb₂O₃ obtained from the inorganic crystal structure database (ICSD #33658) were used as a starting model. The representative XRD patterns of Yb₂O₃ and (Yb_0.92Co_0.08)₂O_3−δ samples prepared in this study are indexed to the standard C-type cubic structure data (ICSD #33658), with no impurity phases detected. The detailed structure refinement parameters from Rietveld refinement of the XRD patterns for the Yb₂O₃ and (Yb_0.92Co_0.08)₂O_3−δ samples are listed in Tables S1 and S2. The weighted profile R-factors (R_wp and R_p) and S of Yb₂O₃ are 2.528, 1.926, and 1.5307%, respectively, which verify the phase purity of the as-prepared samples. The following parameters are obtained at room temperature; a cubic system with a space group Ia3̄ (206), and lattice parameters of a = 10.4391 nm, b = 10.4391 nm, and c = 10.4391 nm, respectively (see the details in Table S2). The structural parameters listed in Table S2 exhibit lattice shrinkage upon increasing the Co content to x = 0.08 as the smaller Co^3+/2+ ions successfully substitute Yb³⁺ sites in the Yb₂O₃ crystal lattice, which decreases the lattice parameter. For the sample with x ≥ 0.10, a further peak shift is not observed compared to that of x = 0.08, where an excess Co component exists as an amorphous phase because an additional crystalline phase is not detected. Therefore, x = 0.08 is identified as the solid solubility limit for forming the single-phase C-type structure in (Yb_0.92Co_0.08)₂O_3−δ.

Fig. 1 XRD patterns of the (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.12) samples.

To investigate the effect of Co ion doping in the prepared catalysts, H₂-TPR measurements were performed, as shown in Fig. 2(a) and summarized in Table 1. The H₂-TPR analysis confirmed that the hydrogen reducibility and the amount of hydrogen consumption at low temperatures tend to increase with increasing Co dopant ratios up to x = 0.08. This increase is attributed to the higher concentration of oxygen vacancies in the Yb₂O₃ lattice due to charge compensation, which lowers the activation energy required for oxygen release, thereby facilitating H₂ reduction and consumption even at low temperatures. Additionally, the reduction of lattice oxygen at high temperatures also increased with increasing Co dopant ratios up to x = 0.08. This suggests that the concentration of oxygen vacancies has increased not only on the surface but also within the lattice. For x > 0.08, the hydrogen reduction temperature increased, and the amount of hydrogen consumption decreased. This suggests that hydrogen reduction was suppressed by the presence of the remaining amorphous phase, consistent with the solid solution limit.

Fig. 2 (a) H₂-TPR profile and (b) O₂-TPD profile of the (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.10) samples.

Table 1 H₂ consumption amount, O₂ desorption site, and oxygen vacancy density (ρ_vac) of the (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.10) catalyst

Sample	H2 consumption amount/μmol g^−1	O2 desorption site/μmol g^−1	ρ vac/×10^17 sites per m^2
Yb2O3	12	81	5.09
(Yb0.97Co0.03)2O3−δ	157	102	6.48
(Yb0.95Co0.05)2O3−δ	273	111	8.22
(Yb0.92Co0.08)2O3−δ	378	128	11.08
(Yb0.90Co0.10)2O3−δ	309	93	8.75

---

O₂-TPD analysis was performed to confirm the formation of oxygen vacancies on both the surface and within the lattice of Yb₂O₃ due to the solid solution of Co as shown in Fig. 2(b) and summarized in Table 1. It was confirmed that the amount of oxygen desorption increased with the Co dopant ratios up to x = 0.08. This increase in oxygen adsorption is due to the higher concentration of oxygen vacancies caused by the solid solution of Co. Notably, it can be seen that the amount of desorption of chemically adsorbed oxygen at high temperature is greater than that of surface-adsorbed oxygen at low temperature. This is due to the redox characteristics of Co^2+/3+. When O₂ is adsorbed on the oxygen vacancy site, Co²⁺ is oxidized to Co³⁺, and O₂ can be reduced to O²⁻ within the lattice. These findings indicate that the concentration of oxygen vacancies increased due to the solid solution of Co, particularly in the range of x ≤ 0.10, which had a greater impact on the increase in oxygen vacancies within the lattice rather than on the increase in surface oxygen vacancies. Surface oxygen vacancies primarily contribute to the physical adsorption of oxygen, but as the doping concentration of Co^2+/3+ increases, the concentration of lattice oxygen vacancies increases and electron mobility increases due to the redox capability of Co^2+/3+, thereby increasing the chemical adsorption of oxygen. Therefore, it is believed that the increased lattice oxygen vacancies can enhance the electrochemical adsorption of the catalyst and ammonia during the ammonia decomposition process. This, in turn, promotes ammonia cracking by increasing the destruction of the N–H bond due to the electron-donating properties of Co²⁺. In the range of x > 0.08, the amount of oxygen desorption decreases and the desorption temperature increases. This change occurs not only because oxygen vacancies no longer increase due to the solid solution limit, but also because oxygen adsorption is suppressed by the remaining amorphous phase. To investigate the Co^2+/3+ doping effect of the (Yb_0.92Co_0.08)₂O_3−δ catalyst in more detail, XPS analysis of Co 2p, and O 1s core levels before and after Co^2+/3+ doping was conducted, as shown in Fig. 3(a) and (b). The Co 2p_3/2 and Co 2p_1/2 peaks can be attributed to divalent Co²⁺ and trivalent Co³⁺ ions. Peaks at 781.6 eV and 796.7 eV and 780.1 eV and 795.3 eV can be attributed to Co²⁺ and Co³⁺, respectively.¹⁶ The O 1s peak can be attributed to lattice oxygen (O_L), oxygen vacancies (O_V), surface adsorbed oxygen (O_S), and surface hydroxyl oxygen (O_OH). Peaks at 529.4 eV, 531.8 eV, 533.3 eV, and 530.8 eV can be attributed to lattice oxygen (O_L), oxygen vacancies (O_V), surface-adsorbed oxygen (O_S) and surface hydroxyl oxygen (O_OH), respectively.¹⁶ As Co^2+/3+ was doped in the lattice of Yb₂O₃, the ratio of Co²⁺ increased more than that of Co³⁺, and it was confirmed that oxygen vacancies increased due to charge compensation from the introduction of Co²⁺ ions into Yb³⁺ sites within the Yb₂O₃ lattice.

Fig. 3 (a) XPS spectra and (b) surface atomic ratios for Co 2p and O 1s core-levels of Yb₂O₃, and (Yb_0.92Co_0.08)₂O₃.

Fig. 4(a) shows the temperature dependence of the catalytic NH₃ decomposition reaction over the prepared (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.12) samples. The catalytic activity increases with the Co dopant ratios up to x = 0.08, and the enhanced catalytic activity could be attributed to improved redox properties due to the induction of Co^2+/3+ valence change, which promotes electron donation for NH₃ decomposition. Additionally, the activity of (Yb_1−xCo_x)₂O_3−δ is influenced by the increased number of NH₃ adsorption sites due to the formation of oxygen vacancies caused by the partial substitution of the Yb³⁺ sites by the lower valent Co²⁺ ions. The N atom in NH₃ acts as a Lewis base site with a partial negative charge, as shown in Fig. S2, while the oxygen vacancy site, with a positive charge, is considered a Lewis acid site and an electron acceptor. This oxygen vacancy site can induce the adsorption and cracking of NH₃ on the catalyst surface. Therefore, it could be confirmed that the improved redox properties and oxygen vacancy site with increasing Co content in (Yb_1−xCo_x)₂O_3−δ (x ≤ 0.08) samples promote catalytic NH₃ decomposition activity. In contrast, the x ≥ 0.10 samples exhibit lower decomposition activity than the (Yb_0.92Co_0.08)₂O_3−δ (x = 0.08) sample, attributed to the reduced number of active sites caused by the presence of amorphous impurity phases formed by excess Co components. The highest catalytic activity is observed with the (Yb_0.92Co_0.08)₂O_3−δ (x = 0.08) specimen, achieving an NH₃ conversion of 98% at 600 °C.

Fig. 4 (a) Temperature-dependence of ammonia decomposition and (b) correlations between the specific reaction rate and TOF to the density of oxygen vacancies estimated based on the amount of O₂ derived from the O₂-TPD analysis of the (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.12) samples.

To confirm the relationship between the density of oxygen vacancies in (Yb_1−xCo_x)₂O_3−δ and the NH₃ decomposition rate, correlations between the specific reaction rate and TOF with the density of oxygen vacancies of the (Yb_1−xCo_x)₂O_3−δ (0 ≤ x ≤ 0.12) samples are shown in Fig. 4(b). The specific reaction rate (r) and TOF derived from normalization of the NH₃ decomposition rate with the density of oxygen vacancies were found to increase as the oxygen vacancy density increases, reaching 11.08 × 10¹⁷ sites per m² for the most active (Yb_0.92Co_0.08)₂O_3−δ (x = 0.08) sample, which is consistent with the catalytic NH₃ decomposition activities as given in Fig. 4(a). From these results, we identify the oxygen vacancy formation that is responsible for the high catalytic activity of (Yb_1−xCo_x)₂O_3−δ for ammonia decomposition.

Previous studies have demonstrated that the acceleration of NH₃ decomposition can be ascribed to the electron-enriched Ni transition metal acting as the active site. It has been demonstrated that the electron donation effect of Ni is beneficial for the decomposition of NH₃, which could accelerate the combinative desorption of N* from Ni surfaces.^9–13 To investigate the influence of the Ni transition metal as the active loading site on the (Yb_0.92Co_0.08)₂O_3−δ sample, αwt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (0 ≤ α ≤ 20) catalysts were synthesized, and the catalytic NH₃ decomposition activities were investigated. Specifically, the diffraction peaks in the XRD patterns of αwt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (0 ≤ α ≤ 20) catalysts before H₂ reduction, as shown in Fig. S3(a), are indexed to NiO and Yb₂O₃, with no impurity phases observed. Additionally, the XRD patterns of the αwt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (0 ≤ α ≤ 20) catalyst after H₂ reduction, as shown in Fig. S3(b), confirm the reduction of NiO to the Ni metal phase. However, no crystal structural change in Yb₂O₃ was observed. The composition determined by inductively coupled plasma-mass spectrometry (ICP-MS) analysis of αwt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (0 ≤ α ≤ 20) is shown in Table S3. The measured compositions of αwt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (0 ≤ α ≤ 20) are evidently in good agreement with the feed values. The NH₃ decomposition activities of αwt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (0 ≤ α ≤ 20) catalysts were evaluated, as shown in Fig. 5. The catalytic activity for the NH₃ conversion increases with the Ni loading of α = 16, where 100% NH₃ conversion is achieved at 550 °C. In contrast, at higher Ni loadings of α ≥ 18, the NH₃ decomposition activity at 550 °C decreases due to Ni aggregation, and a 100% NH₃ conversion is achieved at 600 °C. The optimized 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst in this study exhibited superior activity for complete NH₃ decomposition at 550 °C at a GHSV of 30 000 L kg⁻¹ h⁻¹. And as shown in Table S4 in the SI, the optimized catalyst in this study shows better ammonia decomposition performance at the same space velocity and reaction temperature compared to the performance of other reported Ni and Ru supported catalysts.

Fig. 5 Temperature-dependence of ammonia decomposition over the awt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (0 ≤ a ≤ 20) catalysts after H₂ reduction.

To investigate the active site on the prepared αwt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalysts for catalytic NH₃ decomposition, the oxygen vacancy density of the prepared catalysts was estimated based on the Ni loading using O₂-TPD analysis.

Fig. 6 presents the O₂-TPD profile recorded for the αwt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (α = 0, 2, 8, 16, 18) catalysts. As Ni was supported up to 16 wt%, the amount of oxygen desorption further increased. In particular, the highest amount of oxygen desorption was observed at α = 16 wt%. This is because, when the supported Ni metal transfers electrons to Co³⁺, Co³⁺ can be reduced to Co²⁺, oxygen (O²⁻) can be released from the surface or within the lattice according to charge compensation, and oxygen vacancies can be formed additionally compared to (Yb_0.92Co_0.08)₂O_3−δ. However, in the range of α > 16 wt%, the amount of oxygen desorption decreased. It can be inferred that the formation of oxygen vacancies can be suppressed by excessive aggregation of Ni metal in range of α > 16 wt% caused to reduce the properties of electron donors. These results are consistent with the Ni dispersion data shown in Table 2. Up to α = 16 wt%, Ni dispersion increased, but for α > 16 wt%, Ni dispersion decreased. From this O₂-TPD analysis, the density of oxygen vacancies was estimated based on the desorption amount of O₂ as shown in Table 2. Among the prepared catalysts, the 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst showed the highest oxygen vacancy density of 25.43 × 10¹⁷ site per m². This results tended to be consistent with the NH₃ decomposition activity in Fig. 5. Therefore, it can be concluded that even with the Ni-supported (Yb_0.92Co_0.08)₂O_3−δ catalyst, the oxygen vacancy sites significantly impact the high catalytic NH₃ decomposition activity.

Fig. 6 O₂-TPD profile of awt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (a = 0, 2, 8, 16, 18) catalysts after H₂ reduction.

Table 2 O₂ desorption site, oxygen vacancy density, and Ni dispersion of awt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (a = 0, 2, 8, 16, 18) catalysts after H₂ reduction

Sample	O2 desorption site/μmol g^−1	ρ vac/×10^17 sites per m^2	Ni dispersion/%
(Yb0.92Co0.08)2O3−δ	128	11.08	—
2wt%Ni/(Yb0.92Co0.08)2O3−δ	142	14.89	1.80
8wt%Ni/(Yb0.92Co0.08)2O3−δ	172	18.47	5.57
16wt%Ni/(Yb0.92Co0.08)2O3−δ	221	25.43	5.63
18wt%Ni/(Yb0.92Co0.08)2O3−δ	208	23.96	5.08

---

In order to more accurately investigate the surface active factors for ammonia decomposition in the Ni-supported (Yb_0.92Co_0.08)₂O_3−δ catalyst based on the Ni loading, XPS analysis of Ni 2p, Co 2p, and O 1s core levels according to the Ni loading was performed as shown in Fig. 7(a) and (b). The Ni 2p_3/2 peaks can be attributed to Ni metal(Ni0), NiO and Ni(OH)₂. Peaks at 852.5 eV, 853.4 eV and 855.6 eV correspond to Ni metal(Ni0), NiO and Ni(OH)₂, respectively. As the amount of supported Ni increased up to 16 wt%, the ratios of Ni metal and NiO increased, while the ratio of Ni(OH)₂ decreased. In the Co 2p core level analysis, when α = 0, the Co 2p can be attributed to only Co²⁺ and Co³⁺. However, as the Ni loading increased up to 2 wt%, Co metal was also observed with Co²⁺ and Co³⁺. This is because the supported Ni metal, acting as an electron donor, reduced Co²⁺ and Co³⁺ to Co metal. It has been reported that cobalt has a higher electronegativity of 1.88 compared to 1.11 for ytterbium, which facilitates its reduction more readily than ytterbium.²⁰ Therefore, when Ni metal supported on the surface of (Yb_0.92Co_0.08)₂O_3−δ, Co^2+/3+ ions can oxidize the Ni metal by accepting electrons from it compared to Yb³⁺ ions. As Co^2+/3+ ions are reduced to a lower oxidation state by accepting electrons from Ni metal with increasing Ni loading up to α = 16, the ratio of NiO was observed to increase due to its oxidation by lattice oxygen released from (Yb_0.92Co_0.08)₂O_3−δ as a result of charge compensation for the reduced Co^2+/3+ ions. Furthermore, as the Ni loading increased up to 16 wt%, the ratio of Co³⁺ decreased drastically and the highest ratio of Co metal was exhibited as shown in Table S5. It is inferred that the Co metal reduced from Co^2+/3+ ions by Ni metal as an electron donor through metal–support surface interactions in the Ni/(Yb_0.92Co_0.08)₂O_3−δ system may be released from (Yb_0.92Co_0.08)₂O_3−δ and exists as an Ni–Co alloy or as a single Co metal on the surface of (Yb_0.92Co_0.08)₂O_3−δ. In this regard, Efremova et al.²⁰ developed a Pt/Co₃O₄ catalyst used in the methanation reaction of CO₂, and confirmed that when Pt was supported, Co^2+/3+ of the Co₃O₄ promoter was reduced with the increase of the Co metal ratio and formation of a Co–Pt alloy. Therefore, in this study, it is inferred that the Co metal ratio increases as Ni metal is supported on the Yb₂O₃–Co₃O₄ promoter. As the Ni loading increased up to 2 wt%, based on the result of O 1s core level analysis, the ratio of oxygen vacancies increased with increasing Co metal ratio. It is because that the ratio of oxygen vacancies depend on charge compensation by reduction of Co^2+/3+. When a = 8, the ratio of oxygen vacancy was almost similar to α = 2. This is due to the similar Co metal ratio of α = 2 and α = 8. However, as the Ni loading increased up to 16 wt%, the highest ratio of oxygen vacancy was observed with the highest Co metal ratio of α = 16. When the Ni loading exceeded 16 wt%, it was confirmed that the ratios of Ni, Co metal and oxygen vacancies tended to decrease. It can be inferred that aggregation of Ni active species above 18 wt% reduced its performance as electron donors. Therefore, it is confirmed that the active sites of Ni and Co metals act as electron donors and the electron-deficient oxygen vacancies act as electron acceptors for the associative N desorption from NH₃ to produce N₂ in the NH₃ cracking process.¹¹ The correlation between the amount of Ni supported and the density of oxygen vacancies and the NH₃ decomposition rate is shown in Fig. 8(a) and (b). In Fig. 8(a), it was confirmed that the specific reaction rate and TOF continued to increase as the Ni loading increased to 16 wt%. Additionally, Fig. 8(b) shows that as the oxygen vacancy density increased, the specific reaction rate and TOF also continuously increased. Notably, the highest oxygen vacancy density and the fastest NH₃ decomposition rate were observed when the Ni loading was 16 wt%. When the Ni loading exceeded 16 wt%, the oxygen vacancy density and NH₃ decomposition rate tended to decrease. As a result, among the catalysts synthesized in this study, the 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (α = 16 wt%) catalyst achieved 100% ammonia conversion at the lowest temperature of 550 °C. It can be concluded that the highest oxygen vacancy density and Co metal content in the 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (α = 16 wt%) catalyst promoted ammonia decomposition activity in combination with the Ni metal active species. Additionally, H₂ and N₂-TPD analysis was performed for Yb₂O₃, (Yb_0.92Co_0.08)₂O_3−δ, 16wt%NiO/(Yb_0.92Co_0.08)₂O_3−δ, and 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalysts, and the results are shown in Fig. 9 and Table 3. The results showed that the optimized 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst showed the lowest desorption sites for N₂ gas compared to the Yb₂O₃, (Yb_0.92Co_0.08)₂O_3−δ, and 16wt%NiO/(Yb_0.92Co_0.08)₂O_3−δ samples. Through this, it can be confirmed that the optimized 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst had the lowest adsorption for N₂ gas compared to other samples. In addition, the desorption sites for H₂ gas showed similar values in the synthesized catalysts, and it was confirmed that there was almost no adsorption for H₂ gas. The reason why (Yb_0.92Co_0.08)₂O_3−δ had a higher adsorption for N₂ gas than Yb₂O₃ is because the adsorption of N₂ was increased at the oxygen vacancy sites due to the doping of Co^2+/3+ ions. However, as Ni metal was supported on the surface of (Yb_0.92Co_0.08)₂O_3−δ, the adsorption of N₂ on the catalyst surface decreased. This is thought to be because the bonding force between N₂ and Ni is weak. Therefore, the optimized 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst can be confirmed to have excellent desorption properties for N₂ and H₂ generated during the NH₃ decomposition process.

Fig. 7 (a) XPS spectra and (b) surface atomic ratios for Ni 2p, Co 2p and O 1s core-levels of awt%Ni/(Yb_0.92Co_0.08)₂O₃ (a = 0, 2, 8, 16, 18).

Fig. 8 (a) Correlations between the specific reaction rate and TOF to the density of Ni active site estimated based on the amount of Ni loading and (b) correlations between the specific reaction rate and TOF to the density of oxygen vacancy estimated based on the amount of O₂ derived from the O₂-TPD analysis of awt%Ni/(Yb_0.92Co_0.08)₂O_3−δ (a = 0, 2, 8, 16, 18) catalysts after H₂ reduction.

Fig. 9 (a) H₂-TPD and (b) N₂-TPD profile of Yb₂O₃, (Yb_0.92Co_0.08)₂O_3−δ, 16wt%NiO/(Yb_0.92Co_0.08)₂O_3−δ, and 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ.

Table 3 H₂ desorption site and N₂ desorption site of Yb₂O₃, (Yb_0.92Co_0.08)₂O_3−δ, 16wt%NiO/(Yb_0.92Co_0.08)₂O_3−δ, and 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalysts after H₂ reduction

Sample	H2 desorption site/μmol g^−1	N2 desorption site/μmol g^−1
Yb2O3	5.72	146.20
(Yb0.92Co0.08)2O3−δ	13.25	174.66
16wt%NiO/(Yb0.92Co0.08)2O3−δ	15.68	130.35
16wt%Ni/(Yb0.92Co0.08)2O3−δ	25.18	79.33

---

The catalyst recyclability and stability test for NH₃ decomposition was performed over an optimized 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst, as shown in Fig. 10(a) and (b). The recyclability test in Fig. 10(a) showed a slight decrease in the ammonia decomposition activity of run 2 compared to the initial run 1. However, a similar ammonia decomposition rate was maintained until run 6 without a significant decrease in activity. In the stability test for H₂ production shown in Fig. 10(b), the 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst revealed a constant NH₃ conversion of 82% at 500 °C for 80 hours, and exhibited a constant H₂ concentration of approximately 110 000 ppm converted from NH₃, which is nearly equivalent to the theoretical H₂ production amount (126 738 ppm) for an NH₃ conversion rate of 82%. This converted H₂ concentration corresponds to a high H₂ production rate of 27.4 mmol g⁻¹ min⁻¹, which is superior to most of the other catalysts reported in literature.^10–13,21 Based on these results, the H₂ production rate converted from NH₃ in one day using the optimized 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst was calculated to be 6.23 kg per day. The crystal stability of the 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalyst after recyclability and stability tests was also confirmed (Fig. S4).

Fig. 10 (a) Catalyst recycling test for ammonia decomposition over the 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalysts and (b) catalytic stability test for H₂ production via ammonia decomposition over the 16wt%Ni/(Yb_0.92Co_0.08)₂O_3−δ catalysts at 500 °C for 80 h.

Conclusions

A series of αwt%Ni/(Yb_1−xCo_x)₂O_3−δ catalysts were prepared via the wet impregnation method, and their catalytic activities toward NH₃ decomposition were evaluated. The introduction of Co^2+/3+ ions in the Yb₂O₃ lattice resulted in enhanced catalytic activity for the NH₃ decomposition with increasing oxygen vacancy sites, and the highest catalytic activity was obtained for (Yb_0.92Co_0.08)₂O_3−δ (x = 0.08). When 16wt% of Ni was supported, the proportion of Co metal increased, and an additional increase of oxygen vacancies was confirmed. Consequently, the highest catalytic activity was achieved with the 16wt%Ni-supported (Yb_0.92Co_0.08)₂O_3−δ catalyst, which obtained 100% NH₃ conversion at 550 °C. The enhancement in the activity could be attributed to the high proportion of Ni and Co metals as electron donors, and oxygen vacancies as electron acceptors, which is beneficial for the associative N desorption from NH₃ to produce N₂ in the NH₃ cracking process. Our work highlights the potential of catalytic NH₃ decomposition for H₂ production over non-noble metal catalysts with a low Ni loading of 16 wt% supported inorganic materials.

Author contributions

Yeon-Bin Choi: conceptualization, methodology, validation, investigation, data curation. Writing – original draft, writing – review & editing, visualization. Tae Wook Kang: conceptualization, methodology, writing – original draft, writing – review & editing, visualization. Seo Ra Woo: validation, investigation, data curation, visualization. Do yun Kim: conceptualization, resources, writing – review & editing. Sun Woog Kim: conceptualization, resources, writing – review & editing. Byungseo Bae: conceptualization, resources, writing – review & editing, supervision.

Conflicts of interest

There are no conflicts to declare.

Data availability

Supplementary information is available: Calculation details, Rietveld refinement, Lewis structure of NH₃, XRD patterns, measured composition, Summary of various catalysts for NH₃ decomposition, XPS result.^22–33 See DOI: https://doi.org/10.1039/D5CY00603A.

Acknowledgements

This research was financially supported by the Ministry of Trade, Industry and Energy, Korea, under the “World Class Plus Program (R&D, P177000004)” supervised by the Korea Institute for Advancement of Technology (KIAT).

Notes and references

1. F. Hayashi, Y. Toda, Y. Kanie, M. Kitano, Y. Inoue and T. Yokoyama, Ammonia decomposition by ruthenium nanoparticles loaded on inorganic electride C12A7:e⁻, Chem. Sci., 2013, 4, 3124.
2. I. Staffell, D. Scamman, A. V. Abad, P. Balcombe, P. E. Dodds, P. Ekins, N. Shahd and K. R. Warda, The role of hydrogen and fuel cells in the global energy system, Energy Environ. Sci., 2019, 12, 463.
3. L. Schlapbach and A. Züttel, Hydrogen-storage materials for mobile applications, Nature, 2001, 414, 353.
4. F. Schuth, R. Palkovits, R. Schlogl and D. S. Su, Ammonia as a possible element in an energy infrastructure: Catalysts for ammonia decomposition, Energy Environ. Sci., 2012, 5, 6278.
5. L. J. Murray, M. Dinca and J. R. Long, Hydrogen storage in metal-organic frameworks, Chem. Soc. Rev., 2009, 38, 1294.
6. S. Mukherjee, S. V. Devaguptapu, A. Sviripa, C. R. F. Lund and G. Wu, Low-temperature ammonia decomposition catalysts for hydrogen generation, Appl. Catal., B, 2018, 226, 162.
7. M. Pinzón, P. Sánchez, A. Osa, A. Romero and A. Consuegra, Recent Insights into Low-Surface-Area Catalysts for Hydrogen Production from Ammonia, Energies, 2022, 15, 8143.
8. K. Okura, T. Okanishi, H. Muroyama, T. Matsui and K. Eguchi, Promotion effect of rare-earth elements on the catalytic decomposition of ammonia over Ni/Al₂O₃ catalyst, Appl. Catal., A, 2015, 505, 77.
9. X. Duan, G. Qian, X. Zhou, Z. Sui, D. Chen and W. Yuan, Tuning the size and shape of fe nanoparticles on carbon nanofibers for catalytic ammonia decomposition, Appl. Catal., B, 2011, 101, 189.
10. Y. Q. Gu, Z. Jin, H. Zhang, R. J. Xu, M. J. Zheng, Y. M. Guo, Q. S. Song and C. J. Jia, Transition metal nanoparticles dispersed in an alumina matrix as active and stable catalysts for Co_x-free hydrogen production from ammonia, J. Mater. Chem. A, 2015, 3, 17172.
11. C. Zhou, K. Wu, H. Huang, C. F. Cao, Y. Luo, C. Q. Chen, L. Lin, C. Au and L. Jiang, Spatial Confinement of Electron-Rich Ni Nanoparticles for Efficient Ammonia Decomposition to Hydrogen Production, ACS Catal., 2021, 11, 10345.
12. Y. Yu, Y. Gan, C. Huang, Z. Lu, X. Wang, R. Zhang and G. Feng, Ni/La2O3 and Ni/MgO–La₂O₃ Catalysts for the Decomposition of NH3 into Hydrogen, Int. J. Hydrogen Energy, 2020, 45, 16528.
13. K. Okura, K. Miyazaki, H. Muroyama, T. Matsui and K. Eguchi, Ammonia Decomposition over Ni Catalysts Supported on Perovskite-Type Oxides for the on-Site Generation of Hydrogen, RSC Adv., 2018, 8, 32102.
14. A. Kahraman, H. Karacali and E. Yilmaz, Impact and origin of the oxide-interface traps in Al/Yb₂O₃/n-Si/Al on the electrical characteristics, J. Alloys Compd., 2020, 825, 154171.
15. C. M. Cho, N. Nunotani and N. Imanaka, Direct Decomposition of N₂O over C-Type Cubic Yb₂O3-Co₃O₄ Catalysts, Bull. Chem. Soc. Jpn., 2019, 92, 1148.
16. C. M. Cho, Y. Watanabe, N. Nunotani and N. Imanaka, Direct Decomposition of Nitrous Oxide Using Yb₂O₃-Pr₆O₁₁ with C-type Cubic Structure, Chem. Lett., 2018, 47, 996.
17. F. Izumi and K. Momma, Three-Dimensional Visualization in Powder Diffraction, Solid State Phenom., 2007, 130, 15.
18. R. D. Shannon, Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides, Acta Crystallogr., Sect. A, 1976, 32, 751.
19. K. S. Hui, C. W. Kwong and C. Y. H. Chao, Methane emission abatement by Pd-ion-exchanged zeolite 13X with ozone, Energy Environ. Sci., 2010, 3, 1092.
20. A. Efremova, I. Szentia, J. Kiss, A. Szamosvolgyia, A. Sapi, K. Baan, L. Olivi, G. Varga, Z. Fogarassy, B. Pecz, A. Kukovecz and Z. Konya, Nature of the Pt-Cobalt-Oxide surface interaction and its role in the CO₂ Methanation, Appl. Surf. Sci., 2022, 571, 1513.
21. C. Huang, Y. Yu, J. Yang, Y. Yan, D. Wang, F. Hu, X. Wang, R. Zhang and G. Feng, Ru/La2O3 catalyst for ammonia decomposition to hydrogen, Appl. Surf. Sci., 2029, 476, 928.
22. Y.-J. Lee and H.-J. Chae, Facile one-pot synthesis of Ni-based catalysts by cation-anion double hydrolysis method as highly active Ru-free catalysts for green H2 production via NH3 decomposition, Appl. Catal., B, 2022, 307, 121167.
23. K. Ogasawara, T. Nakao, K. Kishida, T.-N. Ye, Y. Lu, H. Abe, Y. Niwa, M. Sasase, M. Kitano and H. Hosono, Ammonia Decomposition over CaNH-Supported Ni Catalysts via an NH2–-Vacancy-Mediated Mars–van Krevelen Mechanism, ACS Catal., 2021, 11, 11005.
24. Y. Im, H. Muroyama, T. Matsui and K. Eguchi, Ammonia decomposition over nickel catalysts supported on alkaline earth metal aluminate for H2 production, Int. J. Hydrogen Energy, 2020, 45, 26979.
25. D. Sima, H. Wu, K. Tian, S. Xie, J. J. Foo, S. Li, D. Wang, Y. Ye, Z. Zheng and Y.-Q. Liu, Enhanced low temperature catalytic activity of Ni/Al–Ce0.8Zr0.2O2 for hydrogen production from ammonia decomposition, Int. J. Hydrogen Energy, 2020, 45, 9342.
26. J. Zhao, L. Deng, W. Zheng, S. Xu, Q. Yu and X. Qiu, Nickel-induced structure transformation in hydrocalumite for enhanced ammonia decomposition, Int. J. Hydrogen Energy, 2020, 45, 12244.
27. Z.-P. Hu, C.-C. Weng, C. Chen and Z.-Y. Yuan, Catalytic decomposition of ammonia to COx-free hydrogen over Ni/ZSM-5 catalysts: A comparative study of the preparation methods, Appl. Catal., A, 2018, 562, 49.
28. L. Li, F. Chen, J. Shao, Y. Dai, J. Ding and Z. Tang, Attapulgite clay supported Ni nanoparticles encapsulated by porous silica: Thermally stable catalysts for ammonia decomposition to COx free hydrogen, Int. J. Hydrogen Energy, 2016, 46, 21157.
29. W. Zheng, J. Zhang, Q. Ge, H. Xu and W. Li, Effects of CeO2 addition on Ni/Al2O3 catalysts for the reaction of ammonia decomposition to hydrogen, Appl. Catal., B, 2008, 80, 98.
30. J. Zhang, H. Xu, X. Jin, Q. Ge and W. Li, Characterizations and activities of the nanosized Ni/Al2O3 and Ni/La–Al2O3 catalysts for NH3 decomposition, Appl. Catal., A, 2005, 290, 87.
31. J. Cha, T. Lee, Y.-J. Lee, H. Jeong, Y. S. Jo, Y. Kim, S. W. Nam, J. Han, K. B. Lee, C. W. Yoon and H. Sohn, Highly monodisperse sub-nanometer and nanometer Ru particles confined in alkali-exchanged zeolite Y for ammonia decomposition, Appl. Catal., B, 2021, 283, 119627.
32. S. F. Kurtoğlu, S. Soyer-Uzun and A. Uzun, Utilizing red mud modified by simple treatments as a support to disperse ruthenium provides a high and stable performance for COx-free hydrogen production from ammonia, Catal. Today, 2020, 357, 425.
33. Z. Wang, Z. Cai and Z. Wei, Highly Active Ruthenium Catalyst Supported on Barium Hexaaluminate for Ammonia Decomposition to COx-Free Hydrogen, ACS Sustainable Chem. Eng., 2019, 7, 8226.

---