Adsorption and dehydrogenation of ammonia on Ru₅₅, Cu₅₅ and Ru@Cu₅₄ nanoclusters: role of single atom alloy catalyst

Abstract

Hydrogen production by the catalytic decomposition of ammonia (NH₃) is an important process for several important applications, which include energy production and environment-related issues. The role of single Ru-atom substitution in a Cu₅₅ nanocluster (NC) has been illustrated using the NH₃ decomposition reaction as a model system. The structural stability of Ru@Cu₅₄ NC has been evaluated using Ru₅₅ and Cu₅₅ NCs for comparison. Ru@Cu₅₄ prefers an icosahedron structure (I_h), like Ru₅₅ and Cu₅₅ NCs, with almost comparable average binding energies of −5.55 eV per atom. The adsorption of NH_x (x = 0–3) on different adsorption sites of the icosahedron Ru@Cu₅₄ NC has also been studied and the corresponding adsorption energies have been estimated. The site-preference investigation suggested that NH₃ prefers to adsorb vertically to the Ru@Cu₅₄. The stable geometries of the N and H atoms on the high symmetry adsorption sites of Ru@Cu₅₄ NC have been studied. Although the N atom favours top and hollow sites, the H atom prefers to stay in the Ru–Cu bridge site along with the hollow sites. The adsorption energy of N on the Ru@Cu₅₄ NC fcc site is found to be −5.42 eV, which is very close to the optimal value (−5.81 eV) of the ammonia decomposition volcano curve. The reaction energies for stepwise H atom elimination from an adsorbed NH₃ molecule have been estimated. Finally, NH₃ adsorption and decomposition on Ru@Cu₅₄ have been illustrated in terms of electronic structure analysis. The energetics calculations for the dehydrogenation of NH₃ suggest that Ru@Cu₅₄ NC can be a suitable catalyst.

---

Introduction

In recent times, the research and development of molecular hydrogen generation reactions have become an active area of scientific exploration.^1,2 An economy based on hydrogen fuel is expected to reduce dependency on fossil fuels and reduce environmental pollution. Hence, it is now obvious that hydrogen is going to be the fuel for on-board vehicular applications and portable devices in coming decades.^3,4 Conventional hydrogen production processes, such as partial oxidation, steam reforming of hydrocarbons, etc., are associated with the emission of CO_x (x = 1, 2) as by-products. As NH₃ is produced in large scale worldwide with an existing huge distribution infrastructure, hydrogen production by NH₃ decomposition could counter the above-mentioned problems. Ammonia, with its 17.8% hydrogen carrying capacity, has been found to be a potential source for CO_x (x = 1, 2) free hydrogen production through its decomposition.⁵ The decomposition of NH₃ can produce high-purity hydrogen for fuel cells and has good practical applications.

The fundamental reaction mechanism of NH₃ decomposition on catalytic surfaces has become very important to design novel catalysts with higher chemical reactivity and selectivity. Ru and Ir, among the Pt group metals, are regarded as the most active catalysts for the decomposition of NH₃.^6,7 Mainer et al. reported a combined experimental and theoretical study in which they demonstrated NH₃ adsorption and dehydrogenation on a Ru (0001) surface.⁸ However, the high cost and limited availability of Ru and Ir catalysts created the necessity of finding a suitable catalyst for this purpose. Less-expensive and easily available alternatives such as Ni-, Co-, Fe-, Cu-, and Pd-based catalysts posses’ good catalytic properties for NH₃ adsorption and decomposition. A variety of such metal surfaces, like Ni,^9–12 Co,¹³ Fe,^13,14 Cu,^15,16 Pd^17–19etc., have been studied extensively as catalysts for this purpose.

Cu is regarded as a good catalyst for the purpose of gas-molecule adsorption and decomposition on its surface. The interaction of NH₃ with Cu surfaces has been investigated by several authors, both experimentally and theoretically.^20–24 Bartels et al. investigated NH₃ molecule adsorption on a Cu(111) surface and inferred that multiple electronic excitation of the ammonia-substrate bond can lead to the desorption of the molecule from the substrate.²⁰ Using scanned-energy-mode photoelectron diffraction, Baumgartel et al. studied the local structure of a Cu(111) surface with ammonia adsorbed and found that the NH₃ molecule occupies the top site with a Cu–N bond length of 2.09 Å.²¹ In addition to experimental studies, several computational studies based on quantum mechanical principles on NH₃ dehydrogenation process on a pure Cu surface are also available.^22–24 Jiang et al. reported the adsorption configuration and dissociation mechanism of NH₃ on Pd(111) and Cu(111) surfaces.²² It was revealed that the adsorption energy trend of NH_x (x = 0–3) is NH₃ < NH₂ < NH < N and that the ammonia dissociation over Pd-based and Cu-based catalysts is a structure-sensitive reaction. In addition, Robinson and Woodruff²³ and Xing²⁴ have studied the complete dissociation of NH₃ on a Cu(111) surface.

As per the study of Jiang et al.,²⁵ NH dissociation is the most energy-intensive step in NH₃ dehydrogenation on clean Cu(111) surfaces, making it the rate-determining step. Subsequently, effort has been made to improve the catalytic activity of Cu by alloying it with other transition metals such as Pd, Ni, etc.^26,27 Jiang et al. systematically investigated NH₃ dehydrogenation on Pd–Cu(111) and Cu–Pd(111) surfaces using density functional theory (DFT) with the aim of exploring the effect of the surface composition of the catalysts and the role of the dopant metal on the catalytic activity of NH₃ dehydrogenation.²⁶ They showed that as compared to Pd(111) and Cu(111) surfaces, the synergistic effect exists in different layers of the catalyst surfaces, which may help to design an optimal catalyst for ammonia dissociation by doping suitable atoms. The main limitation of the large energy barrier for NH dissociation has not been reduced significantly or eliminated despite all these efforts.

So far, most of the NH₃ dehydrogenation studies are limited to metal or bimetallic surfaces, and very few are available on nanodomains. In the nano regime, the step-by-step NH₃ dehydrogenation process has been computationally studied on 13-atom Cu-, Ni- and Cu-doped Ni clusters by Chen et al.²⁷ They reported that adsorption energy of N on the Ni₁₂Cu cluster is very close to the optimum value of the volcano curve, making it a suitable catalyst for NH₃ dehydrogenation. The heat of reaction for the NH₃ dehydrogenation process for Ni₁₂Cu is found to be in between those of Ni₁₃ and Cu₁₃ clusters. X. Chen et al. investigated the NH₃ decomposition reaction on three types of M@Ni core–shell nanoparticles (M = Fe, Ru, Ir) with 13 core M atoms and 42 shell Ni atoms. Regarding NH₃ decomposition, they found that the Ru@Ni core–shell nanocluster possesses catalytic performance comparable to that of single metal Ru, but the catalytic activity of Fe@Ni and Ir@Ni core–shell NCs was found to be unsatisfactory as compared to the active metal Ru.²⁸ In addition to these studies, Lanzani et al. studied the bonding and dissociation of NH₃ and its fragments on a nanosized icosahedral Fe₅₅ cluster using spin-polarized DFT.²⁹ They suggested that the catalytic activity of iron surfaces towards ammonia-like molecules is enhanced when the metal is in the nanostructured phase.

Both Ru and Cu catalysts are widely known for their use as catalysts in various catalytic applications. The catalytic activity of these catalysts increases as we move from the bulk to the surface to the nano domain due to the increase in surface area, which are obvious in surface catalytic reactions. Nanoclusters have drawn considerable attention from researchers in the fields of physics, chemistry, materials science and especially catalysis, as they play an important role in bridging the gap between isolated atoms and bulk material.^30,31

There are few catalytic studies on Ru₅₅ nanoclusters, which may be because their limited availability and very high cost makes them a poor choice in catalysis. One of the important applications of Ru₅₅ nanoclusters as a catalyst is described in ‘Ru₅₅ nanoparticle catalyze the dissociation of H₂O monomer and dimer to produce hydrogen: A comparative DFT study’.³² Here, the authors reported the H₂O monomer and dimer dissociation followed by H₂ release on a Ru₅₅ nanocluster using DFT calculations. Su et al. used a Ru₅₅ nanocluster to enhance the hydrogen evolution reaction (HER) performance as compared to the conventionally used catalysis.³³ They used a Ru₅₅ nanocluster catalyst deposited on O-doped graphene decorated with single metal atoms (e.g., Fe, Co and Ni) for HER. Consistent with theoretical predictions, these hybrid catalysts show outstanding HER performance, much superior to other reported electro-catalysts, such as the state-of-the-art Pt/C catalyst. Ajaml et al. reported the relationship between particle size, intermediate structure and energies of water reduction to produce hydrogen (H₂) on Ru_x (x = 6, 13 and 55) clusters on a g-CN support using DFT calculations as well as by experiment.³⁴ Ungerer et al. investigated the electron distribution and magnetic properties of Ru fcc nano-dots, both in icosahedral (13 and 55 atoms) and cubic shape (13 and 63 atoms).³⁵

Similarly, the Cu₅₅ nanocluster is also known for its several applications as a catalyst. Among other applications, it is used in acetylene hydrogenation, methanol and ethanol formation from syngas, conversion of CO₂ to hydrocarbon fuels, CO oxidation, and as a degradant agent for SF₆ molecules.^36–42 Zhao et al. reported the effect of the size of the Cu nanocluster on the selectivity and activity of acetylene-selective hydrogenation to ethylene using DFT calculations.³⁶ They have shown that on the Cu₅₅ cluster, C₂H₂ is easily hydrogenated to form C₂H₄via a CHCH₂ intermediate, and C₂H₄ prefers desorption over its hydrogenation, suggesting that the Cu₅₅ cluster exhibits a good selectivity towards C₂H₄ formation. The catalytic activity of catalysts greatly depends on their local structure. The local structure changes of the Cu₅₅ nanocluster during heating were investigated by Lin et al. using MD simulation.³⁷ They reported how the local structure changes accompanying the atom packing affect the internal energy with increasing temperature by calculating the radial distribution function. Zhang et al. reported the formation of C₂ oxygenates from syngas on Cu clusters of different sizes (Cu₁₃, Cu₃₈ and Cu₅₅) and the size effect of Cu clusters on the catalytic performance of C₂ oxygenates in order to search for a novel Cu-based catalyst with improved activity and selectivity towards C₂ oxygenates.³⁸ Lim et al. studied the conversion of CO₂ into hydrocarbon fuels (CH₄, CO and HCOOH) on defective-graphene-supported Cu nanoparticles (Cu₅₅ clusters) using a first-principles method.³⁹

However, in the above paragraphs, we mentioned that the use of Ru₅₅ or Cu₅₅ alone is not effective for the dehydrogenation of NH₃ to H₂ and N₂, although a combination of these two elements can be useful for this purpose. Hence, the single-atom catalyst Ru@Cu₅₄ has been chosen for the study of NH₃ dehydrogenation.

To the best of our knowledge, earlier studies on the adsorption and dehydrogenation of ammonia have mainly focused on pure Ru and Cu metal surfaces. Limited systematic studies have been carried out to understand the effect of dopant metals on the dehydrogenation of ammonia. Also, few reports are available in the literature on the finite size effect in this dehydrogenation process. Therefore, it is of interest to probe the effect of the dopant, structure and size of bimetallic single-atom alloy catalysts on the adsorption and dehydrogenation of ammonia. In the present study, an effort has been made to incorporate the effect of the Ru atom by using it to replace a Cu atom in Cu₅₅ NC and study the NH₃ dehydrogenation mechanism process using the DFT method, with the aim of clarifying the favourable reaction mechanisms and the effect of the single Ru atom on the Cu surface. To determine the exact mechanism of NH₃ decomposition on this surface, the elementary steps have been calculated, along with the adsorption energy (E_ads) and reaction energy (ΔE). The regeneration of the catalyst in terms of nitrogen removal from the NCs has also been investigated for its reuse. This study will help to provide a fundamental understanding of the structural, energetic and catalytic properties of the Ru@Cu₅₄ catalyst for NH₃ dehydrogenation and explain why this catalyst can be useful for the production of H₂ molecules via NH₃ decomposition.

Computational details

All calculations were performed within the spin-polarized density functional theory using the plane wave-pseudopotential approach as implemented in the Vienna ab initio simulation package (VASP).^43–45 The electron–ion interaction and the exchange correlation energy were described under the projector-augmented wave (PAW) method and the generalized gradient approximation (GGA) of Perdew–Burke–Ernzerhof (PBE), respectively.^46–48 The energy cut-off for the plane wave basis set for assuming a good energy convergence was fixed at 500 eV. We considered a large number of initial geometries for the clusters with intermediates and found the lowest-energy structures. A large number of spin states were also considered for all these configurations. Ionic optimization was carried out using the conjugate gradient scheme, and the forces on each ion were minimized to 5 meV Å⁻¹.^49,50 A suitable simulation box of size 20 × 20 × 20 Å was considered for all the calculations. The k-point sampling in the Brillouin zone (BZ) was treated with the Monkhorst–Pack scheme using the gamma point.⁵¹ The total energies of each relaxed structure using the linear tetrahedron method with Blochl corrections was subsequently calculated in order to eliminate any broadening-related uncertainty in the energies.⁵² The average binding energy (E_b) of the Ru@Cu₅₄ NC was estimated from the total energy calculations using eqn (1)where, E_Ru@Cu54 = total energy of the Ru@Cu₅₄ NC, E_Ru-atom = total energy of the Ru atom, and E_Cu-atom = total energy of the Cu atom.

The adsorption energies of the adsorbate on Ru@Cu₅₄ NC, E_ads, was calculated using eqn (2)

E_ads = E_{adsorbate+cluster} − (E_adsorbate + E_cluster) (2)

E_{adsorbate+cluster} = total energy of adsorbate and cluster, E_adsorbate = total energy of adsorbate and E_cluster = total energy of cluster.

Results and discussion

Structural stability of metal clusters

Ru has been regarded as a very good catalyst for ammonia (NH₃) decomposition, as Ru lies in the optimum position of the volcano curve that depicts the choice of materials for NH₃ decomposition. However, Ru has limitations due to its higher cost and low availability. On the other hand, Cu is a versatile catalyst used in different catalytic processes. Like Ru, Cu also has the limitation of a higher energy requirement during the last NH dissociation, which makes this the rate determining step. To rationalize the cost effectiveness of materials and reduce energy expenses, a combination of Cu and Ru atoms has been considered. Nanoclusters may possess better catalytic activity as compared to their bulk counterparts. The stability of 55-atom Cu₅₅ and Ru₅₅ nanoclusters (NCs), along with a Cu₅₅ species with a single Ru atom substitution (Ru@Cu₅₄), has been investigated using DFT. It is found that, like Cu₅₅ and Ru₅₅ NCs, the substituted Ru@Cu₅₄ species possesses an icosahedron (I_h) structure in its ground state, as shown in Fig. 1(a–c). The cohesive energies and interatomic bond distances of the Cu₅₅, Ru@Cu₅₄ and Ru₅₅ NCs are summarized in Table 1. From the table, it can be seen that Ru@Cu₅₄ has more or less the same cohesive energy as the Cu₅₅ and Ru₅₅ NCs. Hence, stability-wise, Ru@Cu₅₄ does not differ much from the 55-atom NCs of its constituent elements. The spin moments of these NCs are also listed in Table 1. We calculated the stability of the Ru@Cu₅₄ nanocluster with Ru atom at the top and edge positions. The configuration with Ru at the top position was found to be 0.12 eV more stable compared to that at the edge position in terms of average binding energy. Hence, the Ru@Cu₅₄ nanocluster with the Ru atom at the top position was used throughout our entire study. We assumed that the Ru atom should be on the surface of the Ru@Cu₅₄ nanocluster, as it would facilitate NH₃ adsorption followed by its dehydrogenation through direct interaction with the NH₃ molecule. Investigation of the structural stability with the Ru atom at the inner position of Ru@Cu₅₄ nanocluster is of no use here, as there would be no direct interaction between the Ru atom and NH₃ molecule. Hence, the structural stability of the Ru@Cu₅₄ nanocluster with Ru as an inner atom has not been studied here.

Fig. 1 Optimized structures of the (a) Cu₅₅, (b) Ru₅₅ and (c) Ru@Cu₅₄ nanoclusters. (Atom colours: Sky blue = Ru, Brown = Cu).

Table 1 Average binding energy (eV per atom), average bond lengths (Å) and spin moment of the Cu₅₅, Ru@Cu₅₄, and Ru₅₅ nanoclusters

Species	Average binding energy	Average bond length (Å)			Spin moment (μB)
		Cu–Cu	Cu–Ru	Ru–Ru
Cu55	−5.59	2.67	—	—	14
Ru@Cu54	−5.55	2.68	2.79	—	14
Ru55	−5.54	—	—	2.70	13

---

It has been reported that the N atom adsorption energy is one of the important factors for the NH₃ dehydrogenation catalytic reaction. In order to determine the catalytic activity of the Ru@Cu₅₄ NC for NH₃ dehydrogenation, the adsorption energy of the N intermediate was compared with the optimum value of the famous volcano curve (−5.81 eV), and it was found that the adsorption energy of N on the Ru@Cu₅₄ NC fcc site (−5.42 eV) (see Table 2) is very close to that value. In a similar way, the N-adsorption energy of the Cu@Ru₅₄ NC was also calculated and found to be −1.85 eV (see Table 2), which is far from the optimum value of the volcano curve (−5.81 eV). Hence, Cu@Ru₅₄ was not considered further for NH₃ dehydrogenation.

Table 2 Adsorption energy E_ads (eV), M–N bond length d_M–N (Å), (M = Ru, Cu), N–H bond length d_N–H (Å) and H–N–H bond angle <H–N–H (in degrees) for Ru₅₅, Ru₅₄Cu, Cu₅₅, Ru@Cu₅₄ nanoclusters with adsorbed NH₃ and N (at top site t1)

Species	E ads	d M–N	d N–H	<H–N–H
Ru55–NH3	−1.05	2.78	1.03	107.58
Cu@Ru54–NH3	−0.87	2.04	1.03	107.31
Cu55–NH3	−0.84	2.06	1.03	107.35
Ru@Cu54–NH3	−0.85	2.23	1.02	107.7
Ru55–N	−5.03	1.67
Cu@Ru54–N	−1.85	1.75
Cu55–N	−4.36	1.84
Ru@Cu54–N	−5.42	1.65

---

Adsorption of reaction intermediates

After establishing the equilibrium structures of Ru@Cu₅₄, the adsorption of an NH₃ monomer, NH_x (x = 0–2) and H atoms on it were studied. The Ru@Cu₅₄ cluster has twenty faces made up of Cu and Ru atoms. For the NH₃, NH_x (x = 0–2), N and H species, there are six types of adsorption sites, one top site t1, one edge site t2, two bridge sites b1 and b2, and two hollow sites h1(fcc) and h2(hcp). The optimized structures of the Ru@Cu₅₄ NC with adsorbed NH₃, NH_x (x = 0–2), H, 2H, H₂, 2N and N₂ are shown in Fig. 2(a–o). The adsorption energy and geometric parameters are summarized in Tables 2 and 3.

Fig. 2 (a)–(o) Optimized structures of different fragments of NH₃ on Ru@Cu₅₄ nanoclusters: (a) NH₃-t1 (one NH₃ located at the t1 site of Ru@Cu₅₄), (b) NH₃-t2, (c) N-t1, (d) N-FCC, (e) NH₂-t1, (f) NH₂-bridge, (g) NH-t1, (h) NH-bridge, (i) H-bridge, (j) H-FCC, (k) H-HCP, (l) 2H, (m) H₂, (n) 2N and (o) N₂. (Atom colours: Sky blue = Ru, Brown = Cu, Red = N and Blue = H).

Table 3 Adsorption energy (E_ads in eV), geometrical parameters Ru–N (d_Ru–N), Cu–N (d_Cu–N), N–H (d_N–H), Ru–H (d_Ru–H) and Cu–H (d_Ru–H) bond lengths (Å) and H–N–H bond angle (in degrees) of intermediates on Ru@Cu₅₄ nanoclusters

Species	Adsorption site	E ads	D M–N/M–H (M = Ru, Cu)	d N–H	<H–N–H
NH3	Top(t1)	−0.85	2.23 (Ru–N)	1.02	107.7
	Top(t2)	−0.48	2.12 (Cu–N)	1.02	107.7
NH2	Top(t1)	−4.55	1.90 (Ru–N)	1.02	111.23
	Bridge(B)	−4.48	2.08 (Ru–N)	1.02	111.5
			2.0 (Cu–N)
NH	Top(t1)	−4.05	1.77	1.03	—
	Bridge(B)	−4.34	1.93 (Ru–N)	1.03	—
			1.93 (Cu–N)
N	Top(t1)	−5.38	1.65 (Ru–N)	—	—
	FCC	−5.42	1.72 (Ru–N)	—	—
			1.92 (Cu–N)
H	Bridge(B)	−2.70	1.70 (Ru–H)	—	—
			1.79 (Cu–H)
	FCC	−2.54	1.80 (Cu–H)	—	—
	HCP	−2.41	1.75 (Cu–H)	—	—

---

For the adsorption of NH₃ molecule on Ru@Cu₅₄, calculation shows that the NH₃ molecule is preferentially adsorbed at the top site t1 and edge site t2 with binding energies of −0.85 and −0.48 eV, respectively, through the formation of Ru–N (for t1) and Cu–N (for t2) bonds (shown in Fig. 2(a) and (b)). Here, the Ru–N and N–H bond distances were found to be 2.23 and 1.02 Å for NH₃ adsorbed at the t1 site, whereas the Cu–N and N–H bond distances were found to be 2.12 and 1.02 Å for the t2 site, respectively. Upon NH₃ adsorption at the t1 site, the H–N–H bond angle was calculated to be 107.7°, which is comparable to the value obtained in the case of NH₃ adsorption on the Ru(110) surface.¹² The obtained geometric parameters, such as the Ru–N and N–H bond lengths and the H–N–H bond angle of NH₃ adsorbed on Ru@Cu₅₄, which are listed in Table 2, are comparable with those in the pure Ru₅₅ NC.

For the generation of H₂ molecules, NH₃ should be dissociated into N and H atoms on the Ru@Cu₅₄ surface. Thus, it is important to determine the stable positions of the intermediate N, H, NH₂ and NH species on Ru@Cu₅₄ NC. First, we investigated the stability of the N atom at different sites of Ru@Cu₅₄. For those adsorptions, six initial configurations were considered. In these six configurations, the N atom positions were at top site t1, edge site t2, the two bridge sites b1 and b2 and the two three-atom hollow sites h1(fcc) and h2 (hcp). Among these, two sites, i.e., the top (t1) and three-atom hollow fcc site, were found to be stable sites for N atom adsorption, as shown in Fig. 2(c and d). The adsorption energy of the reaction intermediate N on the Ru@Cu₅₄ fcc site is found to be −5.42 eV (Table 3), which is very close to the optimal value (−5.81 eV) of famous ammonia decomposition volcano curve.⁵³

With respect to NH₂ adsorption on the 55-atom Ru@Cu₅₄ NC, NH₂ can be adsorbed on both the Ru-top and Ru–Cu bridge site, as shown in Fig. 2(e) and (f). From Table 3, it is seen that NH₂ at the top site (−4.55 eV) on Ru@Cu₅₄ was found to be more stable than that at the bridge site (−4.48 eV). For pure Ru₅₅, the NH₂ adsorption energy is −5.04 eV for the bridge configuration, whereas it is −4.67 eV for the top (t1) site. For NH₂ adsorbed at the t1 site on Ru@Cu₅₄ NC, the Ru–N and N–H bond lengths were found to be 1.90 Å and 1.02 Å, respectively, while the H–N–H cone bond angle is 111.23°. Similarly, for the bridge site, the Ru–N and Cu–N bond lengths were optimized to be 2.08 and 2.00 Å, respectively, with N–H bond lengths of 1.02 Å and an H–N–H bond angle of 111.5°.

For NH adsorption on Ru@Cu₅₄, two types of stable adsorption positions (top and bridge sites) were examined. The adsorption of NH on the top and bridge sites of Ru@Cu₅₄ NC is shown in Fig. 2(g) and (h), respectively. NH at the Ru–Cu bridge site is found to be more stable as compared to that on the top site by 0.29 eV. The NH adsorption energy on the Ru@Cu₅₄ NC bridge was calculated to be −4.34 eV (Table 3). In the bridge configuration, the Ru–N, Cu–N and N–H bond lengths were found to be 1.93, 1.93 and 1.03 Å, respectively.

To determine the stability of the H atom on Ru@Cu₅₄, the H atom was placed in all the probable sites, i.e., the top, edge, and Ru–Cu and Cu–Cu bridge sites, as well as the fcc and hcp three-atom hollow sites. Even after placing the H atom at the t1 and t2 sites, it was found that it went to the Ru–Cu bridge position after optimization. The stable structures with the H atom at the bridge, FCC and HCP sites of Ru@Cu₅₄ NC are shown in Fig. 2(i–k), and the corresponding adsorption energies are summarized in Table 3. From Table 3, it can be seen that the H atom of the Ru–Cu bridge site has higher stability compared to those at the hollow sites. The Ru–H and Cu–H bond lengths are found to be 1.70 and 1.79 Å, respectively, for the bridge configuration shown in Fig. 2(i).

The release of hydrogen molecules from the nanoclusters depends on the binding of H atoms on them. To determine the hydrogen release behaviour, the H atom stability in presence of an N atom on Ru@Cu₅₄, as well as on Cu₅₅ and Ru₅₅, was studied. The binding energy of the H atom on Cu₅₅–N, Ru@Cu₅₄–N, Ru₅₅–N was calculated to be −2.63, −2.65 and −2.81 eV, respectively. From the H atom binding energy, it can be stated that hydrogen release from Ru@Cu₅₄ NC will be almost the same as that from Cu₅₅ NC, but easier than that from Ru₅₅.

After the dehydrogenation of NH₃ into N and H atoms, the associative recombination of two H atoms followed by the desorption of the H₂ molecule is an important step for the production of hydrogen. Desorption of the H₂ molecule will depend on the diffusion of H atoms on the surface, which in turn is proportional to the limiting strength of the H atom on the surface. For the formation of H₂, the most stable co-adsorption configuration was taken as the initial state, in which two H atoms adsorb on adjacent fcc sites, as shown in Fig. 2(l), for Ru@Cu₅₄. For Cu₅₅ and Ru₅₅, the co-adsorption of two H atoms also occurs on adjacent fcc sites. The potential energy diagram for H₂ desorption from the Ru₅₅, Ru@Cu₅₄ and Cu₅₅ NCs are shown in Fig. 3. In the first step of the reaction, two H atoms adsorbed on the stable fcc sites combine to form an H₂ molecule (Fig. 2(m)) with a reaction enthalpy of 0.25, 0.30, 0.54 eV for Cu₅₅, Ru@Cu₅₄, and Ru₅₅, respectively. In the second step of the reaction, the adsorbed H₂ molecule desorbs from the catalyst surface with a reaction energy of 0.77, 0.82 and 1.03 eV for Cu₅₅, Ru@Cu₅₄, and Ru₅₅, respectively. From Fig. 3, it can be seen that for Ru@Cu₅₄, the enthalpy of the reaction falls between those of the Cu₅₅ and Ru₅₅ NCs. It can be stated that substitution of one Ru atom in the Cu₅₅ nanocluster will be useful for the production of H₂ molecules.

Fig. 3 Potential-energy diagram of H₂ desorption processes on the Ru₅₅, Ru@Cu₅₄ and Cu₅₅ nanoclusters.

Thermochemistry of NH₃ decomposition

In the earlier section, it was mentioned that the adsorption energy of the intermediate N on Ru@Cu₅₄ NC is very close to the optimal value of the volcano curve. To further understand the NH₃ decomposition behaviour on Ru@Cu₅₄, the potential energy diagram of NH₃ decomposition on this NC has been calculated. The potential energy diagram of NH₃ decomposition on the Ru₅₅, Cu₅₅ and Ru@Cu₅₄ NCs is shown in Fig. 4. The most stable adsorption configuration of Ru@Cu₅₄ NC, which is icosahedral, has been chosen as the initial state. We have considered the local interactions between Ru and Cu atoms like other bimetallic systems.

Fig. 4 Potential-energy diagram of NH₃ decomposition on the Cu₅₅, Ru@Cu₅₄ and Ru₅₅ nanoclusters. (Atomic colours: Sky blue = Ru, Brown = Cu, Red = N and Blue = H).

In the first step of dehydrogenation, NH₃ adsorbs on the top site on Ru@Cu₅₄ with an adsorption energy of −0.85 eV and becomes dissociated into NH₂ and H intermediates with the breakage of one of the three N–H bonds. The NH₂ intermediate remains in the Ru–Cu bridge site, and the H atom moves to the hcp hollow site, as shown in Fig. 4. The dehydrogenation reaction is an endothermic process with a reaction energy of 0.39 eV on Ru@Cu₅₄. The reaction energy for the same dehydrogenation process on Cu₅₅ is 0.98 eV, which is higher than that on Ru@Cu₅₄. For the second step of dehydrogenation on Ru@Cu₅₄, the NH₂ adsorbed at the bridge site is considered; it becomes dissociated into NH and H fragments. Although NH remains on the bridge site, the H atom moves to the hcp hollow site, as shown in Fig. 4. The corresponding reaction energy is 0.97 eV, while the value for Cu₅₅ is 0.61 eV. Similarly, for the third step of dehydrogenation, NH sits at the bridge site initially and is then dissociated into N (fcc) and H fragments (hcp), as shown in Fig. 4. This process is endothermic with a reaction energy of 0.30 eV on the Ru@Cu₅₄ NC. The same process of dehydrogenation is also endothermic on the Cu₅₅ NC with a comparatively higher reaction energy of 1.16 eV. This suggests that the dehydrogenation of the NH intermediate is the rate-determining step for Cu₅₅ due to the high positive reaction heat (Fig. 4). The lower reaction energy of the last step of dehydrogenation for Ru@Cu₅₄ makes it a better catalyst for NH₃ decomposition compared to Cu₅₅.

By comparing the data obtained in this work, the results show that the overall reaction energy for NH₃ dehydrogenation on Ru@Cu₅₄ NC (1.66 eV) is comparable to that on Ru₅₅ NC (1.2 eV), but lower as compared to that on pure Cu₅₅ NC (2.75 eV), which is consistent with the order of N adsorption energies. As expected, the Ru@Cu₅₄ NC is thermodynamically feasible for the reaction of NH₃ decomposition and can act as a better catalyst than pure Cu₅₅ NC for this purpose.

In catalysis, the activation energy is a very important parameter, and is the one of the characteristics of the activity of catalysts. Recently, it has been established that for the dissociative chemisorption of a number of molecules, the activation energy depends linearly on the adsorption energy.^54,55 The Brønsted–Evans–Polanyi (BEP) relation has been found to be applicable for this type of surface reactions. In the NH₃ dehydrogenation process, Duan et al. showed that activation energy for the associative desorption of an N₂ molecule and the N-adsorption energy follow a linear relationship for the Fe(110), Co(111), Ni(111) and Cu(111) surfaces.¹³ Here, the N-adsorption energy for the Cu₅₅, Ru₅₅ and Ru@Cu₅₄ NCs was calculated to be −4.36, −5.03 and −5.42 eV, respectively. From this, we can conclude that the activation energy for the associative desorption of an N₂ molecule on Ru@Cu₅₄ NC will be within the range of Ru₅₅.

Regeneration of catalyst

Catalyst regeneration is a very important step in catalysis for reuse and economic viability. Hence, N removal is necessary after the release of a hydrogen molecule from Ru@Cu₅₄. The stability of the two adsorbed N atoms located at different fcc and hcp sites was investigated. Two N atoms were found to be more stable at adjacent fcc sites, as shown in Fig. 2(n), and the corresponding adsorption energy was found to be −5.48 eV per N atom. In the next step, the two N atoms are combined into an N₂ molecule vertically on the catalytic surface (Fig. 2(o)) with a reaction enthalpy of 0.12, 0.35, and 0.75 eV for Cu₅₅, Ru@Cu₅₄, and Ru₅₅, respectively. N₂ then desorbs from the catalyst surface with a reaction energy of 0.43, 0.76, and 0.63 eV for Cu₅₅, Ru@Cu₅₄, and Ru₅₅, respectively. The distance of N₂ molecule from the nearest Ru/Cu atom was found to be 1.91, 1.90, and 1.94 Å for Ru@Cu₅₄, Cu₅₅ and Ru₅₅, respectively. The potential energy diagram and the minimum energy path for the conversion of N₂(g) from the two adsorbed N atoms is shown in Fig. 5. From Fig. 5, it can be seen that N₂ removal from Ru@Cu54 will be easier, as the enthalpy of the reaction lies between those of Cu₅₅ and Ru₅₅.

Fig. 5 Potential-energy diagram of N₂ desorption processes on the Ru₅₅, Ru@Cu₅₄ and Cu₅₅ nanoclusters.

Electronic structure analysis

Since the N atom adsorption energy is one of the important factors in the catalytic NH₃ dehydrogenation reaction, in order to interpret the N atom adsorption energy difference, the projected density of states must be calculated for the nanoclusters under study. The interaction between the adsorbate and substrate has been studied by performing the density of states spectra and d-band calculations for the Cu₅₅, Ru@Cu₅₄ and Ru₅₅ nanoclusters, which are shown in Fig. 6. From the figure, it can be seen that the d-bands are more populated near the Fermi level for Cu₅₅ and Ru@Cu₅₄ as compared to Ru₅₅. This facilitates the N atom adsorption on Ru@Cu₅₄, as for metals, a d-band centre close to the Fermi energy can facilitate a stronger interaction between the adsorbate and substrate.⁵⁶ The calculated d-band centres of Ru₅₅, Ru@Cu₅₄ and Cu₅₅ nanoclusters were found to be −2.85, −2.82 and −3.96 eV, respectively. As the d-band centre of Ru@Cu₅₄ is nearest to the Fermi level, the N atom adsorption energy will be higher for Ru@Cu₅₄ as compared to Ru₅₅ and Cu₅₅. In the previous section, it was reported that the N atom adsorption energy is the highest for Ru@Cu₅₄ among the three nanoclusters.

Fig. 6 Density of states projected onto the d-band of the Ru@Cu₅₄, Cu₅₅ and Ru₅₅ nanoclusters.

Conclusion

A systematic study of ammonia adsorption and dehydrogenation on small Ru@Cu₅₄ nanoclusters has been carried out using DFT calculations. The most stable adsorption structures for the NH₃, NH₂, NH, N, and H species were determined, indicating that NH₃, NH₂ and NH prefer to adsorb on the top site while N and H atoms preferentially adsorb on the fcc and bridge sites, respectively. In order to determine the catalytic activity of the Ru@Cu₅₄ NC for NH₃ dehydrogenation, the adsorption energy of the N intermediate was compared with the optimum value of the famous volcano curve (−5.81 eV), and it was found that the adsorption energy of N on the Ru@Cu₅₄ NC fcc site (−5.42 eV) is very close to that value. The overall reaction energy for stepwise dehydrogenation on Ru@Cu₅₄ is found to be lower than that of Cu₅₅. Also, the energy required for the last step of the dehydrogenation of NH₃ in the Cu₅₅ NC is reduced abruptly when an apex Cu atom is replaced by Ru. The reduced energy requirement for NH₃ dehydrogenation in the case of Ru@Cu₅₄ compared to that of Cu₅₅ can make it a better catalyst for this purpose. Also, Ru@Cu₅₄ is found to be advantageous for catalyst regeneration by the removal of residual N atoms. Additionally, the d-band centre calculation of the three nanoclusters also confirms the calculated N atom adsorption energy trend.

Data availability

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to ethical restrictions.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors are thankful to Dr P. K. Mohapatra, Associate Director, Radiochemistry and Isotope Group, Bhabha Atomic Research Centre (BARC) and Dr S. C. Parida, Head, Product Development Division, BARC for their interest and encouragement during progress of this work. The authors are also thankful to the members of the Computer Division, BARC for their kind cooperation during the work.

References

1. L. Schlapbach, Nature, 2009, 460, 809–811.
2. T. N. Veziroglu, Chem. Ind., 1999, 53, 383–393.
3. M. Momirlan and T. N. Veziroglu, Sustainable Energy Rev., 2002, 6(1–2), 141–179.
4. P. Parthasarathy and K. S. Narayanan, Renew. Energy, 2014, 66, 570–579.
5. R. Z. Sørensen, L. J. E. Nielsen, S. Jensen, O. Hansen, T. Johannessen, U. Quaade and C. H. Christensen, Catal. Commun., 2005, 6, 229–232.
6. J. Zhang, H. Y. Xu, X. L. Q. J. Jin and W. Z. Li, Catal. Commun., 2006, 7, 148–152.
7. S. F. Yin, B. Q. Xu, X. P. Zhou and C. T. Au, Appl. Catal., A, 2004, 277, 1–9.
8. S. Maier, I. Stass, J. I. Cerda and J. Salmeron, J. Phys. Chem. C, 2012, 116(48), 25395–25400.
9. E. Salli, V. Hanninen and L. Halonen, J. Phys. Chem. C, 2010, 114, 4550–4556.
10. S. Stolbov and T. S. Rahman, J. Chem. Phys., 2005, 123, 204716.
11. M. Grunze, P. A. Dowben and C. R. Brundle, Surf. Sci., 1983, 128, 311–324.
12. X. Duan, G. Qian, C. Fan, Y. Zhu, X. Zhou, D. Chen and W. Yuan, Surf. Sci., 2012, 606, 549–553.
13. X. Duan, J. Ji, G. Qian, C. Fan, Y. Zhu, X. Zhou, D. Chen and W. Yuan, J. Mol. Catal. A: Chem., 2012, 357, 81–86.
14. R. J. Lin, F. Y. Li and H. L. Chen, J. Phys. Chem. C, 2011, 115, 521–528.
15. J. Robinson and D. P. Woodruff, Surf. Sci., 2002, 498, 203–211.
16. B. Xing and G. C. Wang, Phys. Chem. Chem. Phys., 2014, 16, 2621–2629.
17. J. A. Herron, S. Tonelli and M. Mavrikakis, Surf. Sci., 2012, 606, 1670–1679.
18. G. Novell-Leruth, A. Valcarcel, J. Perez-Ramirez and J. M. Ricart, J. Phys. Chem. C, 2007, 111, 860–868.
19. Z. Jiang, Q. Pan, M. Li, T. Yan and T. Fang, Appl. Surf. Sci., 2014, 292, 494–499.
20. L. Bartels, M. Wolf, T. Klamroth, P. Saalfrank, A. Kuhnle, G. Meyer and K. H. Rieder, Chem. Phys. Lett., 1999, 313, 544–552.
21. P. Baumgartel, R. Lindsay, T. Giessel, O. Schaff, A. M. Bradshaw and D. P. Woodruff, J. Phys. Chem. B, 2000, 104, 3044–3049.
22. Z. Jiang, P. Qin and T. Fang, Chem. Phys., 2014, 445, 59–67.
23. J. Robinson and D. P. Woodruff, Surf. Sci., 2002, 498, 203–211.
24. B. Xing and G.-C. Wang, Phys. Chem. Chem. Phys., 2014, 16, 2621–2629.
25. Z. Jiang, P. Qin and T. Fang, Chem. Phys., 2014, 445, 59–67.
26. Z. Jiang, P. Qin and T. Fang, Appl. Surf. Sci., 2016, 371, 337–342.
27. S. Chen, X. Chen and H. Zhang, J. Mat. Sci., 2017, 52, 3162–3168.
28. X. Chen, J. Zhou, S. Chen and H. Zhang, J. Nanopart. Res., 2018, 20, 148–156.
29. G. Lanzani and K. Laasonen, Int. J. Hydrogen Energy, 2010, 35, 6571–6577.
30. W. Eberhardt, Surf. Sci., 2002, 500, 242–270.
31. F. Baletto and R. Ferrando, Rev. Mod. Phys., 2005, 77, 371–423.
32. P. Cheng, Y. Yang, N. Ahmed, S. Zhang and S. Huang, Int. J. Hydrogen Energy, 2016, 41, 3844–3855.
33. P. Su, W. Pei, X. Wamg, Y. Ma, Q. Jiang, J. Liang, S. Zhao, J. Zhao, J. Liu and G. Q. Lu, Angew. Chem., 2021, 133(29), 16180–16186.
34. S. Ajmal, H. T. D. Bui, V. Q. Bui, T. Yang, X. Shao, A. Kumar, S. G. Kim and H. Lee, Chem. Eng. J., 2022, 429, 132282.
35. M. J. Ungerer and N. H. de Leeuw, Nanomaterials, 2023, 13, 1118–1131.
36. B. Zhao, R. Zhang, Z. Huang and B. Wang, Appl. Catal., A, 2017, 546, 111–121.
37. Z. Lin, Z. Cai-Bei and Q. Yang, Chin. Phys., 2007, 16, 77–82.
38. R. Zhang, M. Peng, T. Duan and B. Wang, Appl. Surf. Sci., 2017, 407, 282–296.
39. D. H. Lim, J. H. Jo, D. Y. Shin, J. Wilcox, H. C. Ham and S. W. Nam, Nanoscale, 2014, 6, 5087–5092.
40. L. Ma and J. Akola, Phys. Chem. Chem. Phys., 2019, 21, 11351–11358.
41. S. M. Sintilo, A. B. Hernandez, A. A. P. Cid, W. I. Hernandez and M. S. Villanueva, ACS Omega, 2022, 7, 34401–34411.
42. M. B. Gawande, A. Goswami, F. X. Felpin, T. Asefa, X. Huang, R. Silva, X. Zou, R. Zboril and R. S. Varma, Chem. Rev., 2016, 116(6), 3722–3811.
43. G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 49, 14251–14269.
44. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6, 15–50.
45. W. Kohn and L. Sham, Phys. Rev. A, 1965, 140, 1133–1138.
46. P. E. Blöchl, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 50, 17953–17979.
47. G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758–1775.
48. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
49. R. P. Feynman, Phys. Rev., 1939, 56, 340–343.
50. H. Hellman, Introduction to Quantum Chemistry; Deuticke, Leipzig, 1937.
51. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1976, 13, 5188–5192.
52. P. E. Blöchl and O. Jepsen, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 49, 16223–16233.
53. D. A. Hangsen, D. G. Vlachos and J. G. Chen, Nat. Chem., 2010, 2(6), 484–489.
54. A. Logadottir, T. H. Rod, J. K. Nørskov, B. Hammer, S. Dahl and C. J. H. Jacobsen, J. Catal., 2001, 197, 229–231.
55. T. Bligaard, J. K. Nørskov, S. Dahl, J. Matthiesen, C. H. Christensen and J. Sehested, J. Catal., 2004, 224, 206–217.
56. B. Hammer and J. K. Norskov, Adv. Catal., 2000, 45, 71–129.

---