Theoretical study of CO catalytic oxidation on free and defective graphene-supported Au–Pd bimetallic clusters

Abstract

In this work, CO adsorption and oxidation on free and defective graphene-supported Au_mPd_n (m + n = 13) clusters with either icosahedral (ICO) or truncated octahedral (TO) structures are investigated using density functional theory calculations. It is found that CO adsorbs more strongly than molecular O₂ on the surface of these clusters, which is attributed to the strong hybridization between metal-d and C-p orbitals. In addition, the structural transformation from ICO to TO is observed for the Au-rich clusters upon CO and O₂ adsorption because the stability of the ICO cluster is lower than that of the TO one. It is also found that the free Au₁₂Pd₁ cluster with TO structure exhibits the lowest energy barrier for CO oxidation among the free clusters (0.17 eV). When the Au₁₂Pd₁ cluster is supported on the single vacancy defective graphene, an increase in stability, a decrease in the adsorption strength of CO and O₂, and a slight increase in the energy barrier (0.41 eV) for CO oxidation compared to the corresponding free cluster are observed. Our results are expected to be useful for future applications of graphene-supported bimetallic nanocatalysts.

---

I. Introduction

The fascinating catalytic properties of metal clusters stem from their non-crystalline configurations, large surface-to-volume ratio, and finite size effect.¹ One of the most famous examples is gold; although gold has long been regarded as an inert metal, the pioneering work of G. J. Hutchings and Haruta et al. demonstrated that gold nanoparticles are exceptionally active catalysts towards a number of reactions.^2–5 The activity of Au clusters for CO oxidation is strongly dependent on the cluster size,⁶ cluster shape,⁷ surface coating,⁸ and oxide support.⁹ In recent years, bimetallic clusters have attracted considerable attention since the introduction of a second metal into the single metal cluster can improve the catalytic activities of the clusters by the synergetic effect.¹⁰ Notably, Au-based bimetallic clusters prepared by the mixing of Au with other metals (Pd, Pt, or Cu)^11–14 have been used to improve the catalytic activities, stabilities, and selectivities of pure Au clusters for CO oxidation.^15–18 For example, Scott and co-workers found that the supported Au–Pd clusters exhibit higher catalytic activities towards CO oxidation compared with the supported Au-only or Pd-only catalysts.¹⁹

Graphene, a one-atom-thick carbon sheet with unique electronic and geometric properties, is regarded as one of the most promising candidates for the next generation of electronic materials.^20–22 Recently, graphene has been extensively studied as a support for heterogeneous catalysts due to its huge surface to-volume ratio for the catalytic reaction.^23–27 The presence of carbon vacancies on graphene can be used as anchoring points for the binding of metal clusters. For example, the binding of Pt, Fe, Pt, and Al metal clusters with graphene^28–30 can be enhanced by the dangling sp² carbon bonds at the monovacancy sites of defective graphene, significantly enhancing the stabilities of metal clusters.³¹ In addition, the defects in graphene can promote the adsorption of small molecules on metal clusters.²⁸ Therefore, the defective graphene-supported Au–Pd bimetallic clusters are anticipated to be very active catalysts for CO oxidation, and further studies are needed to verify this expectation.

Density functional theory (DFT)³² calculations have been used to evaluate the catalytic properties of metal clusters for CO oxidation. In a previous work, DFT calculations were performed to study CO oxidation on Au–Pd bimetallic surfaces. Zhang et al.³³ used DFT to study the catalytic properties of CO oxidation on Au/Pd (111) surfaces and found that the system with Au on the surface of a Pd₁₆Au₄ slab exhibited the lowest energy barrier for CO catalytic oxidation. On the other hand, Yuan et al.³⁴ investigated CO oxidation on Pd-decorated Au (111) surfaces, finding that the Pd₂-decorated Au (111) surface was highly reactive for CO oxidation. In addition, free Au/Pd core–shell clusters have been found to be highly efficient catalysts for CO oxidation by DFT calculations.^35,36 However, to the best of our knowledge, a theoretical understanding of the reaction mechanism of CO oxidation on graphene-supported Au–Pd bimetallic clusters is still lacking,³⁷ and the details of how the support and cluster shape affect the reaction mechanism are also unknown.

In this paper, we apply DFT calculations to investigate CO adsorption and oxidation on free and defective graphene-supported Au_mPd_n (m + n = 13) clusters with either icosahedral (ICO) or truncated octahedral (TO) structures. For the free Au_mPd_n (m + n = 13) clusters, we study the relatively stable structures of 13-atom Au–Pd clusters as well as O₂ and CO adsorption, the co-adsorption of O₂ and CO, and CO oxidation on these clusters. We compare the energy barrier for CO oxidation on the defective graphene-supported Au₁₂Pd₁ cluster with that of CO oxidation on the corresponding free cluster.

II. Computational details

In this work, 13-atom Au–Pd bimetallic clusters with icosahedral (ICO) and truncated octahedral (TO) structures are studied. From the structural point of view, attention has usually been concentrated on high-symmetry configurations,³⁸ which are often found to be the lowest-energy structures of Au–Pd bimetallic clusters at the DFT level. The ICO or TO structures are good candidates for smaller clusters due to their more favourable surface energies at the expense of introducing internal strain.³⁹ The lowest-energy atomic ordering of such clusters is calculated at the empirical potential level^40–42 and then subjected to local relaxation by DFT. In the calculations, an EAM empirical potential^43–45 is used to model the metal–metal interactions; its explicit form^43,44,46 and parameters for the Au–Pd systems^46,47 can be found elsewhere. The resulting configurations are used to study the adsorption and reaction properties. Fig. 1 shows the equilibrium geometries of the four free metal clusters with ICO and TO structures used in this work.

Fig. 1 The configurations of Pd₁₃, Pd₁₂Au₁, Au₁₂Pd₁, and Au₁₃ with icosahedral (ICO) and truncated octahedral (TO) structures. Au and Pd atoms are represented by yellow and blue spheres, respectively.

DFT calculations were performed using the PWSCF (plane-wave self-consistent field) plane wave code in the Quantum ESPRESSO package.⁴⁸ All the calculations were carried out by the spin-polarized generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE)⁴⁹ xc-functionals and ultrasoft pseudopotentials.⁵⁰ It is noted that ultrasoft pseudopotentials have been used successfully to investigate the chemical properties of many transition metal systems including Au–Pd systems.^50–53 Spin-polarized calculations were performed using values of 40 and 400 Ry as the energy cut-offs for the selection of the plane waves describing the wave function and the electronic density, respectively. For free metal clusters, a rectangular supercell with a size of 30 × 30 × 30 ų was employed in the calculations. The first Brillouin zone was sampled at the Gamma-point, and the electronic levels were broadened using a Gaussian smearing technique with a smearing parameter of 0.002 Ry. The geometry of the cluster upon or without adsorption is optimized until the total energy converges to 10⁻⁶ eV. For the Au₁₂Pd₁ cluster supported on defective graphene, an orthorhombic supercell of 14.76 × 14.76 × 31.51 ų with periodic boundary conditions was used. After checking for the possibility of one, two, three and four atom conformations with defective graphene, only the configuration with one metal atom interacting with the graphene cluster was found to be stable. The Au₁₂Pd₁–graphene system was separated from its periodic images in the z direction by a vacuum space of 25 Å. The integration of the Brillouin zone was conducted using a 2 × 2 × 1 Monkhorst–Pack grid with the K-points for the Au₁₂Pd₁–graphene system. In this work, the van der Waals (vdW) interaction⁵⁴ is not included since the vdW interaction has only a slight effect on the adsorption and energy barrier energies for CO oxidation on metal systems.^24,26,27

We calculated the adsorption energies according to the equation:

ΔE_ads = E_total − E_cluster − E_adsorbate (1)

Furthermore, the co-adsorption energies of CO with molecular or atomic oxygen were calculated as:

ΔE_coads = E_total − E_{cluster(or cluster+graphene)} − E_CO − E_{O2(or O)} (2)

In the above equations, E_total, E_{cluster (or cluster+graphene)}, E_adsorbate, E_CO and E_{O2 (or O)} correspond to the electronic energies of the adsorbed species on the cluster, the bare cluster (or the bare cluster on graphene), a gas-phase adsorbate, gaseous CO and gaseous O₂ (or O), respectively. The climbing-image nudged elastic band (NEB) method⁵⁵ was adopted here to determine the reaction pathway for CO oxidation on Au–Pd clusters. The reaction paths are interpolated by fitting a cubic polynomial through all images on each path guided by the force tangent to the reaction coordinate at each image. Once a minimum-energy path is determined, the transition state is located.⁵⁶

III. Results and discussion

A. CO and O₂ adsorption on free clusters

The calculated average bond lengths, average binding energies, and d-band centres of Pd₁₃, Pd₁₂Au₁, Au₁₂Pd₁, and Au₁₃ with ICO and TO structures are shown in Table 1. The average bond length increases with increasing Au concentration for clusters with the same morphology because Au is larger than Pd. In addition, the average bond lengths of the TO structures are smaller than those of the ICO structures for free clusters with the same composition. The average binding energy corresponds to the stability of the cluster. As listed in Table 1, the absolute value of the average bonding energy for the Pd-rich clusters follows the order TO < ICO, while the absolute value of the average bonding energy of TO is bigger than that of ICO for the Au-rich clusters. Thus, the ICO structure is more stable than the TO structure for Pd-rich clusters, while the opposite order is found for Au-rich clusters. Relatively stable structures should be used to study the catalytic properties of the clusters due to the structural interconversion between isomers in the working environment. For example, the ICO to TO structural transformation in Au-rich clusters upon O₂ adsorption is observed because the stability of the ICO cluster is lower than that of the TO structure for Au-rich clusters (Fig. 2). Thus, the following paragraphs only discuss the adsorption, co-adsorption and CO oxidation reaction on ICO Pd₁₂Au₁ and Pd₁₃ clusters and TO Au₁₂Pd₁ and Au₁₃ clusters.

Table 1 Calculated average bond lengths R (in Å), average binding energies E_b (in eV per atom), and d-band centres E_d (eV) for Pd₁₃, Pd₁₂Au₁, Au₁₂Pd₁, and Au₁₃ with icosahedral (ICO) and truncated octahedral (TO) structures

Cluster	Pd13		Pd12Au1 (Au in the core)		Au12Pd1 (Pd in the core)		Au13
Isomer	ICO	TO	ICO	TO	ICO	TO	ICO	TO
Length	3.22	2.72	3.25	2.73	3.34	2.82	3.37	2.84
Eb (eV per atom)	−2.29	−2.24	−2.23	−2.22	−1.99	−2.07	−1.89	−1.96
Ed (eV)	−2.28	—	−1.51	—	—	−2.73	—	−2.74

---

Fig. 2 The structural transformation from icosahedron to truncated octahedron for the free Au₁₂Pd₁ cluster upon O₂ adsorption.

In the ICO structure, there are 20 equivalent triangular fcc (111)-like facets (Fig. 3). In the TO structure, the facets consist of both triangles and quadrangles (see Fig. 3). It is found that CO and O₂ cannot be adsorbed on the (100) hollow site in the quadrangle of the TO structure. To compare the Pd-rich clusters with the ICO structure, only the triangular facet was chosen to study CO adsorption and oxidation on Au-rich clusters. For CO adsorption, the molecule prefers the “end-on” configuration with the C atom binding to the Pd atoms (see Fig. 3).⁵⁷ Three possible sites were calculated: on the top (T), on the bridge (B), and on the hollow (H). We have calculated the energies of CO adsorption at these sites on the Pd₁₃, Pd₁₂Au₁ (Au in the centre), Au₁₂Pd₁ (Pd in the centre), and Au₁₃ clusters. The results are summarized in Table 2. H is the most stable adsorption site for Pd₁₃ and Pd₁₂Au₁ clusters, with adsorption energies of −2.56 and −2.44 eV, respectively, while T (vertex) is the most stable site for Au₁₂Pd₁ and Au₁₃ clusters, with adsorption energies of −1.41 and −1.36 eV, respectively.

Fig. 3 The possible adsorption sites for CO and O₂ molecules on the icosahedral (ICO) and truncated octahedral (TO) clusters.

Table 2 Calculated adsorption energies (in eV) of O₂ and CO at different adsorption sites on icosahedral Pd₁₂Au₁ (Au in the core) and Pd₁₃ clusters and truncated octahedral Au₁₂Pd₁ (Pd in the core) and Au₁₃ clusters

Cluster		Adsorption site
		T	B	H	B1	T-H
Pd12Au1	O2	−0.71	−0.73	−0.64	−1.44	−1.58
	CO	−1.84	−2.33	−2.56	U	U
Pd13	O2	−0.78	−0.57	−0.55	−1.34	−1.43
	CO	−1.35	−2.23	−2.44	U	U
Au12Pd1	O2	−0.37	U	U	−0.63	U
	CO	−1.41	−1.18	U	U	U
Au13	O2	−0.44	U	U	U	U
	CO	−1.36	−1.29	U	U	U

---

For O₂ adsorption, we calculated five possible configurations (see Fig. 3): (1) on the top (T), (2) on the bridge (B), (3) on the hollow (H), (4) between two top sites (B1), and (5) between the hollow site and the top site (H–T). For O₂ adsorption on B1 and H–T sites, the molecule prefers a “side-on” configuration with both oxygen atoms bound to the Au or Pd atoms.⁵⁷ At the T, B, and H sites, however, O₂ prefers an “end-on” configuration with one oxygen atom binding to the Pd atom. The corresponding adsorption energies are summarized in Table 2. The H–T configuration is the most stable O₂ adsorption site on Pd₁₃ and Pd₁₂Au₁ clusters, with adsorption energies of −1.43 and −1.58 eV, respectively. In contrast, the B1 site is the most stable O₂ adsorption site for Au₁₂Pd₄₃, with an adsorption energy of −0.63 eV. For the Au₁₃ cluster, the T site is the only stable configuration, with an adsorption energy of −0.44 eV.

The adsorption energies of CO at the most favourable adsorption site for the Pd₁₃, Pd₁₂Au₁, Au₁₂Pd₁, and Au₁₃ clusters are −2.44, −2.56, −1.41, and −1.36 eV, respectively. The adsorption energies of O₂ at the most favourable adsorption site for the Pd₁₃, Pd₁₂Au₁, Au₁₂Pd₁, and Au₁₃ clusters are −1.43, −1.58, −0.63, and −0.44 eV, respectively. The adsorption strength of both CO and O₂ follows the order Au₁₃ < Au₁₂Pd₁ < Pd₁₃ < Pd₁₂Au₁. To understand the differences in the adsorption strength among the four selected clusters, we investigated the electronic structures of these clusters by calculating their local density of states (LDOS; Fig. 4). The d orbitals of metals play an extremely important role in catalysing the oxidation of CO. To quantitatively investigate the changing trend of the cluster d orbitals in the reaction processes, we calculated the d-band centres of the clusters as:

Fig. 4 The electronic local density of states (LDOS) of the system projected on the d-states of icosahedral (ICO) Pd₁₂Au₁ and Pd₁₃ clusters and truncated octahedral (TO) Au₁₂Pd₁ and Au₁₃ clusters.

The determined E_d of these clusters, which are listed in Table 1, follow the order: Au₁₃ (−2.74 eV) < Au₁₂Pd₁ (−2.73 eV) < Pd₁₃ (−2.28 eV) < Pd₁₂Au₁ (−1.51 eV). Accordingly, the more negative the d-band centre, the weaker the adsorption strength of CO and O₂. This suggests that the position of the d-band centre is the determining factor for the adsorption strength of CO and O₂.

We also find that the adsorption of CO on the clusters is much stronger than that of O₂. This may be due to the stronger hybridization between the Au (Pd)-d and C-p states compared with the hybridization between the Au (Pd)-d and O-p states. As shown in Fig. 5A and B, the Au-d states for the Au₁₂Pd₁ cluster indicated by the peaks at −7 to −5 eV decrease upon CO and O₂ adsorption due to electron depletion. This result supports the idea that the extent of the decrease in the electronic population of the d-orbital is larger for CO adsorption than for O₂ adsorption. In addition, the Au-d states increase more obviously at the energy level from −12 to −7 eV for the Au₁₂Pd₁ cluster upon CO adsorption compared with O₂ adsorption. This results from the electron accumulation and the increased hybridization region between the Au-d and C-p states. The charge density difference indicates a more distinct charge transfer phenomenon with CO adsorption (see Fig. 5B), again suggesting that the adsorption of CO on the clusters is much stronger than that of O₂.

Fig. 5 The partial density of states projected on the Au atom and the adatom (C or O) and the charge density difference plot of the free Au₁₂Pd₁ cluster upon O₂ (A) and CO (B) adsorption.

B. CO oxidation on free clusters

For CO oxidation, a bimolecular Langmuir–Hinshelwood (LH) mechanism is investigated since the adsorption of CO is much stronger than that of O₂ on these clusters. The potential energy profiles and configurations for CO oxidation on Pd₁₃, Pd₁₂Au₁, Au₁₂Pd₁, and Au₁₃ clusters are shown in Fig. 6A–D, respectively. For clarity, the corresponding reaction energies and energy barriers for adsorption on these clusters are also listed in Table 3. To search for the minimum energy paths for CO oxidation, we select the most stable co-adsorption configuration as an initial state, where CO and O₂ molecules are co-adsorbed at their most preferable sites according to the data in Table 2. For the Pd-rich clusters in this work, CO and O₂ prefer to adsorb on the H and H–T sites of Pd₁₃ and Pd₁₂Au₁, respectively; thus, the co-adsorption of the two molecules are too close. The choice of H (for CO) and the suboptimum B₁ (for O₂) is also too close. Alternatively, we absorb CO on the suboptimum site (B site) and O₂ on the H–T site. For Au-rich clusters, CO and O₂ are co-adsorbed at two T sites.

Fig. 6 The minimum energy path (MEP) for CO adsorption, co-adsorption, oxidation, and desorption on the free Pd₁₂Au₁ (A), Pd₁₃ (B), Au₁₂Pd₁ (C) and Au₁₃ (D) clusters. The yellow, blue, grey, and red spheres represent Au, Pd, C, and O atoms, respectively.

Table 3 Calculated relative adsorption energies (E_ad in eV) of the initial adsorption (CO), co-adsorption (CO + O₂), intermediate (OCOO), and final (CO₂ + O) states and energy barriers (E_act in eV) of icosahedral Pd₁₂Au₁ and Pd₁₃ clusters, truncated octahedral Au₁₂Pd₁ and Au₁₃ clusters, and the defective graphene-supported Au₁₂Pd₁ cluster

Cluster	Adsorption energy
	Ead/CO	Ead/CO–O2	Ead/OOCO	Ead/O+CO2	Eact
Pd12Au1	−2.33	−3.57	−2.41	−4.45	0.79
Pd13	−2.23	−3.46	−2.24	−4.04	0.81
Au13	−1.36	−3.12	−2.09	−2.98	0.42
Au12Pd1	−1.41	−2.46	−1.94	−2.87	0.17
Supported-Au12Pd1	−1.12	−1.76	−1.68	−3.20	0.41

---

Next, CO and O₂ move closer to form the intermediate state OCOO* with a C–O single bond and a peroxide O–O bond, followed by the crossing of the first energy barrier (TS1). The corresponding energy barriers (TS1) are found to be 0.81, 0.79, 0.17, and 0.42 eV for Pd₁₃, Pd₁₂Au₁, Au₁₂Pd₁, and Au₁₃ clusters, respectively. Notably, the energy barriers for the formation of the intermediate state OCOO* (TS1) on Au-rich clusters are lower than those on the Pd-rich clusters. Among the free clusters, the Au₁₂Pd₁ cluster with TO structure exhibits the lowest energy barrier (0.17 eV) for CO oxidation.

In the third step, the peroxide O–O bond dissociates into two O atoms by crossing the second energy barrier (TS2). The calculated TS2 energy barriers are always zero for all the clusters considered in this work, meaning that CO₂ is easily desorbed from the four free clusters without any barrier. To fully restore the catalytic system, the oxygen adatom left by the initially investigated reaction step needs to react with another CO molecule to produce another CO₂ molecule. This often occurs as CO comes from the gas phase to form a second CO₂ according to an Eley-Rideal (ER) mechanism; this reaction is generally barrier-less. One of the most important factors that can affect the reaction mechanism is the capture of oxygen. The low adsorption energies of O₂ on CO-preadsorbed clusters (E_ad(CO, O₂) − E_ad(CO)) significantly reduce the possibility of the CO + O₂ reaction. In this work, the absolute values of E_ad(CO, O₂) − E_ad(CO) are greater than 0.50 eV, enhancing the possibility of the CO + O₂ reaction. From the analysis mentioned above, it is found that the Au₁₂Pd₁ cluster is the best candidate for CO oxidation among the free clusters studied herein.

C. CO oxidation on supported cluster

When the Au₁₂Pd₁ cluster is supported on single vacancy defective graphene, no structural change of the metal cluster is observed. To understand the structural stability of the supported metal cluster, the distortion energy (E_dis) of the metal cluster is defined as the difference between the energy of the metal cluster in the configuration interacting with graphene and the energy of the bare cluster. The absolute value of E_dis represents the distortion level; a negative value means that the cluster is changed into a more stable geometric configuration, while a positive value means that the cluster retains the original structure with relaxation upon adsorption. In this work, the E_dis of the Au₁₂Pd₁ cluster is −0.46 eV, indicating an increase in stability when the Au₁₂Pd₁ cluster is supported on single vacancy defective graphene. In addition, the energy difference between the supported and free graphene is about −0.17 eV, indicating a slight structural change for the single vacancy defective graphene due to the addition of the Au₁₂Pd₁ cluster.

The adsorption strengths of CO (−1.12 eV) and O₂ (−0.30 eV) on the T sites of the single vacancy defective graphene-supported Au₁₂Pd₁ cluster are smaller than those on the corresponding free Au₁₂Pd₁ cluster at the same adsorption sites. This suggests that the monovacancy defective graphene support stably anchors the Au₁₂Pd₁ cluster at the vacancy site, consequently preventing the Au₁₂Pd₁ cluster from being significantly altered upon O₂ or CO adsorption. To understand the difference in the adsorption strengths of the free and defective graphene-supported Au₁₂Pd₁ clusters, the LDOS of these clusters are calculated (Fig. 7A). Accordingly, the d-band centres are determined to be −2.73 and −2.81 eV for the free and defective graphene-supported Au₁₂Pd₁ clusters, respectively. These results agree with the point that the more negative the d-band centre, the weaker the adsorption strengths of CO and O₂. In addition, a band gap of ∼0.5 eV between the highest occupied molecular orbital and the lowest unoccupied molecular orbital is found for the defective graphene-supported Au₁₂Pd₁ cluster, which can contribute to the enhanced stability⁵⁸ (Fig. 7A).

Fig. 7 (A) LDOS of the system projected on the d-states of the free and graphene-supported Au₁₂Pd₁ clusters. (B) PDOS of the s, p and d states of bound Au atoms at the monovacancy site of graphene. (C) PDOS of the p and s states of carbon atoms nearest the monovacancy upon adsorption. Spin-up (↑) and spin-down (↓) states are marked as positive and negative values, respectively.

The projected density of states (PDOS) has been analysed for the valence electrons of the defective graphene-supported Au₁₂Pd₁ cluster (Fig. 7B and C). As shown from the PDOS of the Au (binding at the monovacancy site of graphene) and C (nearest to the monovacancy) atoms, the strong hybridization between the Au-5d and C-2p states of the defective graphene occurs through almost the entire energy region, indicating a covalent bonding interaction between Au and C atoms. Note that the Au-5d orbital is important in the adsorption of reactants. The Löwdin charge analysis reveals that the negative excess charge on the bound Au atom is −0.20 e, and the total positive excess charge on the defective graphene is 0.18 e. This indicates that charge is transferred from the bound Au atom to defective graphene, possibly explaining the weaker O₂ and CO adsorption energies on the defective graphene-supported Au₁₂Pd₁ system compared to the free Au₁₂Pd₁ cluster.⁵⁹ Notably, the work-function can explain the charge transfer between the adsorbate and the substrate because both terms refer to energy required to remove electrons from an adsorbate-substrate system.⁵⁹

Due to the larger adsorption strength of CO compared to O₂, a bimolecular Langmuir–Hinshelwood (LH) mechanism is investigated for CO oxidation on the defective graphene-supported Au₁₂Pd₁ cluster. The adsorption energies and reaction energy barriers are listed in Table 3. The reaction path for CO catalytic oxidation on the defective graphene-supported Au₁₂Pd₁ cluster is also shown in Fig. 8. The energy barrier for the formation of the intermediate state OCOO* is 0.41 eV; this value is a little higher than that of the corresponding free cluster, but it is still a very low barrier (<0.5 eV). Notably, the energy barrier for CO oxidation on the Au₁₂Pd₁ cluster on graphene (0.42 eV) is lower than those for graphene-supported single Cu (0.54 eV),⁶⁰ Si (0.57 eV),⁶¹ Fe (0.58 eV),²⁶ and Pt (0.59 eV)²⁷ atom catalysts. In conclusion, the single vacancy defective graphene-supported Au₁₂Pd₁ cluster exhibits an increase in stability, a decrease in the adsorption strength of CO and O₂, and a slight increase in the energy barrier (0.41 eV) for CO oxidation compared to the corresponding free cluster.

Fig. 8 The minimum energy path (MEP) for CO adsorption, co-adsorption and oxidation on the defective graphene-supported Au₁₂Pd₁ cluster.

IV. Conclusions

In summary, we investigated CO adsorption and oxidation on free and defective graphene-supported Au_mPd_n (m + n = 13) clusters with either icosahedral (ICO) or truncated octahedral (TO) structures using density functional theory calculations. For the free Au_mPd_n (m + n = 13) clusters, we studied the relatively stable structures of 13-atom Au–Pd clusters as well as the O₂ and CO adsorption, the co-adsorption of CO and O₂, and CO oxidation on these clusters. We found that the adsorption of CO is much stronger than that of O₂ on the cluster surfaces, which is attributed to the strong hybridization between the metal-d and C-p orbitals. In addition, the structural transformation from ICO to TO is observed in the Au-rich clusters upon CO and O₂ adsorption because the stability of the ICO cluster is lower than that of the TO structure. We also found that the free Au₁₂Pd₁ cluster with the TO structure exhibits the lowest energy barrier (0.17 eV) for CO oxidation among these free clusters. We comparatively studied the energy barriers for CO oxidation on the defective graphene-supported Au₁₂Pd₁ cluster and the corresponding free cluster. The results indicate that the single vacancy defective graphene-supported Au₁₂Pd₁ cluster possesses an increased stability, a decreased adsorption strength for CO and O₂, and a slightly increased energy barrier for CO oxidation compared with the corresponding free cluster.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (21106003, 21176010, 91334203), the Beijing Novel Program (Z12111000250000), the “Chemical Grid Project” of BUCT and the Supercomputing Center of the Chinese Academy of Sciences (SCCAS).

References

1. Y. Z. Lu and W. Chen, Chem. Soc. Rev., 2012, 41, 3594.
2. B. Nkosi, N. J. Coville and G. J. Hutching, J. Chem. Soc., Chem. Commun., 1988, 1, 71.
3. M. Haruta, Nature, 2005, 437, 1098.
4. M. Haruta, T. Kobayashi, H. Sano and N. Yamada, Chem. Lett., 1987, 16, 405.
5. A. S. K. Hashmi and G. J. Hutchings, Angew. Chem., Int. Ed., 2006, 45, 7896.
6. Y. Gao, N. Shao, Y. Pei, Z. F. Chen and X. C. Zeng, ACS Nano, 2011, 5, 7818.
7. W. Liu, Y. F. Zhu and Q. Jiang, J. Phys. Chem. C, 2010, 114, 21094.
8. O. Lopez-Acevedo, K. A. Kacprzak, J. Akola and H. Hakkinen, Nat. Chem., 2010, 2, 329.
9. B. Yoon, H. Hakkinen, U. Landman, A. S. Worz, J. M. Antonietti, S. Abbet, K. Judai and U. Heiz, Science, 2005, 307, 403.
10. B. Hammer and J. K. Norskov, Adv. Catal., 2000, 45, 71.
11. X. Bokhimi, R. Zanella and C. Angeles-Chavez, J. Phys. Chem. C, 2010, 114, 14101.
12. Y. Gao, N. Shao, Y. Pei and X. C. Zeng, Nano Lett., 2010, 10, 1055.
13. Y. Luo, H. O. Seo, K. D. Kim, M. J. Kim, W. S. Tai, M. Burkhart and Y. D. Kim, Catal. Lett., 2010, 134, 45.
14. J. Xu, T. White, P. Li, C. He, W. Y. Jianguo Yu and Y.-F. Han, J. Am. Chem. Soc., 2010, 132, 10398.
15. H. C. Ham, J. A. Stephens, G. S. Hwang, J. Han, S. W. Nam and T. H. Lim, J. Phys. Chem. Lett., 2012, 3, 566.
16. F. Gao, Y. L. Wang and D. W. Goodman, J. Phys. Chem. C, 2009, 113, 14993.
17. K. Persson, A. Ersson, K. Jansson, N. Iverlund and S. Jaras, J. Catal., 2005, 231, 139.
18. J. B. Xu, T. S. Zhao, S. Y. Shen and Y. S. Li, Int. J. Hydrogen Energy, 2010, 35, 6490.
19. R. W. J. Scott, C. Sivadinarayana, O. Wilson and R. M. Crooks, J. Am. Chem. Soc., 2005, 127, 1380.
20. K. R. Ratinac, W. R. Yang, S. P. Ringer and F. Braet, Environ. Sci. Technol., 2010, 44, 1167.
21. A. K. Geim, Science, 2009, 324, 1530.
22. C. N. R. Rao, A. K. Sood, K. S. Subrahmanyam and A. Govindaraj, Angew. Chem., Int. Ed., 2009, 48, 7752.
23. A. Prestianni, A. Martorana, F. Labat, I. Ciofini and C. Adamo, J. Mol. Struct.: THEOCHEM, 2009, 903, 34.
24. Q. G. Jiang, Z. M. Ao, S. Li and Z. Wen, RSC Adv., 2014, 4, 20290.
25. D. W. Boukhvalov and M. I. Katsnelson, J. Phys. Chem. C, 2009, 113, 14176.
26. Y. F. Li, Z. Zhou, P. Jin, Y. S. Chen, S. B. B. Zhang and Z. F. Chen, J. Phys. Chem. C, 2010, 114, 6250.
27. Y. Tang, Z. Yang and X. Dai, Phys. Chem. Chem. Phys., 2012, 14, 16566.
28. D. H. Lim and J. Wilcox, J. Phys. Chem. C, 2011, 115, 22742.
29. D. H. Lim, A. S. Negreira and J. Wilcox, J. Phys. Chem. C, 2011, 115, 8961.
30. I. Fampiou and A. Ramasubramaniam, J. Phys. Chem. C, 2012, 116, 6543.
31. A. J. Logsdail and J. Akola, J. Phys. Chem. C, 2011, 115, 15240.
32. E. Aprà, F. Baletto, R. Ferrando and A. Fortunelli, Phys. Rev. Lett., 2004, 93, 065502.
33. J. Zhang, H. M. Jin, M. B. Sullivan, F. Chiang, H. Lim and P. Wu, Phys. Chem. Chem. Phys., 2009, 11, 1441.
34. D. W. Yuan, Z. R. Liu and J. H. Chen, J. Chem. Phys., 2011, 134, 054704.
35. H. L. Chen, C. H. Su and H. T. Chen, J. Chem. Phys., 2012, 536, 100.
36. S. L. Peng, L. Y. Gan, R. Y. Tian and Y. J. Zhao, Comput. Theor. Chem., 2011, 977, 62.
37. B. Zhu, G. Thrimurthulu, L. Delannoy, C. Louis, C. Mottet, J. Creuze, B. Legrand and H. Guesmi, J. Catal., 2013, 308, 272.
38. E. Apra and A. Fortunelli, J. Phys. Chem. A, 2003, 107, 2934.
39. F. Baletto, R. Ferrando, A. Fortunelli, F. Montalenti and C. Mottet, J. Chem. Phys., 2002, 116, 3856.
40. D. Cheng, S. Huang and W. Wang, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74, 064117.
41. D. Cheng, X. Liu, D. Cao, W. Wang and S. Huang, Nanotechnology, 2007, 18, 475702.
42. D. Cheng and W. Wang, Nanoscale, 2012, 4, 2048.
43. R. A. Johnson, Phys. Rev. B: Condens. Matter Mater. Phys., 1989, 39, 12554.
44. R. A. Johnson, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 41, 9717.
45. D. Cheng, I. S. Atanasov and M. Hou, Eur. Phys. J. D, 2011, 64, 37.
46. I. S. Atanasov and M. Hou, Eur. Phys. J. D, 2009, 52, 51.
47. I. Atanasov and M. Hou, Surf. Sci., 2009, 603, 2639.
48. P. Giannozzi, S. Baroni, N. Bonini, M. Calandra, R. Car, C. Cavazzoni, D. Ceresoli, G. L. Chiarotti, M. Cococcioni, I. Dabo, A. Dal Corso, S. de Gironcoli, S. Fabris, G. Fratesi, R. Gebauer, U. Gerstmann, C. Gougoussis, A. Kokalj, M. Lazzeri, L. Martin-Samos, N. Marzari, F. Mauri, R. Mazzarello, S. Paolini, A. Pasquarello, L. Paulatto, C. Sbraccia, S. Scandolo, G. Sclauzero, A. P. Seitsonen, A. Smogunov, P. Umari and R. M. Wentzcovitch, J. Phys.: Condens. Matter, 2009, 21, 359902.
49. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
50. D. Vanderbilt, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 41, 7892.
51. K. Miwa and A. Fukumoto, Phys. Rev. B: Condens. Matter Mater. Phys., 2002, 65, 155114.
52. B. S. Pujari, S. Gusarov, M. Brett and A. Kovalenko, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 84, 041402.
53. D. Cheng, H. Xu and A. Fortunelli, J. Catal., 2014, 314, 47.
54. M. K. Srivastava, Y. Wang, A. F. Kemper and H. P. Cheng, Phys. Rev. B: Condens. Matter Mater. Phys., 2012, 85, 165444.
55. G. Henkelman, B. P. Uberuaga and H. Jonsson, J. Chem. Phys., 2000, 113, 9901.
56. Y. Xu, A. V. Ruban and M. Mavrikakis, J. Am. Chem. Soc., 2004, 126, 4717.
57. C. M. Chang, C. Cheng and C. M. Wei, J. Chem. Phys., 2008, 128, 124710.
58. G. Chen, S. J. Li, Y. Su, V. Wang, H. Mizuseki and Y. Kawazoe, J. Phys. Chem. C, 2011, 115, 20168.
59. D. H. Lim and J. Wilcox, J. Phys. Chem. C, 2011, 115, 22742.
60. E. H. Song, Z. Wen and Q. Jiang, J. Phys. Chem. C, 2011, 115, 3678.
61. J. X. Zhao, Y. Chen and H. G. Fu, Theor. Chem. Acc., 2012, 131, 1.

---