A comparative study of the CO oxidation reaction over pristine and C-doped boron nitride fullerene

Abstract

In this work, we employ density functional theory calculations to investigate the CO oxidation mechanisms over B₁₂N₁₂ and B₁₁N₁₂C nanocages. Two possible reaction pathways can be proposed for the oxidation of CO with O₂ molecule over these surfaces: CO + O₂ → OOCO → CO₂ + O_ads and CO + O_ads → CO₂. Both Eley–Rideal (ER) and Langmuir–Hinshelwood (LH) mechanisms are considered for these two reaction pathways. Our calculations indicate that the CO oxidation reaction over the B₁₁N₁₂C nanocage proceeds via the ER mechanism followed by LH mechanism with an activation energy (E_act) of 0.58 eV. In the case of the B₁₂N₁₂ nanocage, it can be estimated that both reaction pathways go through the LH mechanism. The E_act of the first reaction step is about 2.5 eV, while it is negligible for the second route. Based on the present theoretical results, the catalytic activity of the B₁₁N₁₂C nanocage toward the CO oxidation reaction is more than that of the B₁₂N₁₂ cluster. This can be related to the presence of a C atom that plays a significant role in the activation of the whole cluster. Meanwhile, the performance of the B₁₁N₁₂C nanocage as a catalyst used for the oxidation of CO with O₂ molecules may proceeded at near ambient temperatures. These results indicate that the B₁₁N₁₂C nanocage can be utilized as a favorable low-cost catalyst for the CO oxidation reaction.

---

1. Introduction

The catalytic oxidation of carbon monoxide (CO) to carbon dioxide (CO₂) has long been a major reaction in heterogeneous catalysis. CO is a colourless and odourless but extremely toxic gas that is mainly produced by incomplete burning of fuels in automobiles and industrial processes.^1–3 When inhaled, it combines irreversibly with hemoglobin to produce carboxyhemoglobin, the formation of which prevents the normal transfer of oxygen to body tissues. Thus, plentiful effort has been done to oxidize the CO molecule by various noble metals such as Au,⁴ Pd,^5,6 Pt^7,8 and Rh.⁶ For instance, Alavi et al.⁹ found that the CO oxidation over Pt(111) proceeds via an energy barrier of 1.00 eV, while it was reported by Stampfl and Scheffler to be about 1.02 eV over Ru(0001).¹⁰ In another work, Haruta et al.¹¹ reported that Au catalysts supported on a metal oxide surface exhibit high catalytic activity for CO oxidation at low temperature. However, despite their high activity and selectivity toward the CO oxidation, these noble metal catalysts are rare and very expensive. Therefore, decreasing or substituting of these expensive noble metal-based materials with low-cost catalysts is one of the main important steps for the large-scale practical application of the catalytic CO oxidation process.⁷

Recently, scientists have found that different kinds of nanostructures such as single-walled or multi-walled carbon nanotubes (CNTs),^12,13 graphene¹⁴ or boron-nitride nanotubes (BNNTs)¹⁵ have a great tendency toward the adsorption of toxic gas pollutants. These nanostructures are also a promising candidates to support metal atoms or clusters to realize new catalysts due to their large surface-to-volume ratio,¹⁴ and outstanding electrical, mechanical and thermal properties.¹⁶ Hence, many theoretical and experimental studies have been performed about the oxidation of CO on these nanostructures. For example, in one experimental study, Yoo et al. showed that the catalytic properties of graphene sheet enhances with the interaction of Pt nanoclusters.¹⁷ In another comparative theoretical investigation, Tang et al. reported that the surface activity of the Pt-embedded graphene is increased significantly toward the oxidation of CO molecule.¹⁸ Wang et al. found a strong chemical interaction between CO molecule and Si-doped BNNTs in comparison with pristine BNNTs.¹⁹

After the discovery of C₆₀,²⁰ numerous studies have been focused on this outstanding carbon-based material due to its unique electronic properties with potential applications in building nanodevices.^21,22 C₆₀ displays the promising ability for gas storage which makes it acts as a chemical sensor.²³ In recent years, many theoretical and experimental studies have focused on the possible fullerene-like structures.^24,25 They found that III–V compounds, especially the group III nitrides can be desirable nanoscaled materials due to their direct band gap which affords optical and magnetic characteristics. So, the III–V fullerene-like cages have been theoretically estimated and experimentally synthesized.^26–28 On the other hand, different III–V nanostructures such as BNNTs,^29,30 BN nanosheets^31,32 and BN clusters,^33,34 with the unique polarity of B–N bonds, have more improved surface reactivity than carbon nanostructures. In particular, (BN)_n (n = 4–30) nanocages have been extensively investigated because of their high temperature stability, large thermal conductivity, oxidation resistance and low electric constant.^35,36 For example, Nigam et al. have shown that a chemically inert (BN)₃₆ cluster can be activated by incorporating magnetic nanoparticles inside it.³⁷ Therefore, the adsorption of O₂ molecule on these clusters results in an O–O bond elongation and then improves the CO oxidation reaction. Among these BN nanocages, the fullerene-like cage B₁₂N₁₂ as the smallest stable cage cluster is extensively investigated. Oku et al. experimentally synthesized the B₁₂N₁₂ nanocage by laser desorption time-of-flight mass spectrometry and reported that these clusters consists of 4- and 6-membered rings of BN with the band gap energy of about 5.1 eV.³⁸ On the other hand, there are several interesting investigations about the adsorption of chemical toxic gases on this favorable nanocage. For instance, in our previous work,³³ the adsorption and decomposition of methanol on B₁₂N₁₂ nanocage have studied in details. We found that the electrical conductivity of the B₁₂N₁₂ enhances upon the adsorption of CH₃OH and this nanocage could be used as an efficient metal-free catalyst for the dehydrogenation of the methanol molecule.

Today, chemical doping with foreign atoms is a favorable approach to modify the properties of host materials. In comparison with the undoped B₁₂N₁₂ cluster, C-doped BN fullerene (B₁₁N₁₂C) displays more reactivity than B₁₂N₁₂.^39–41 For example, Wu et al. showed that the hydrogenation reaction on B₁₁N₁₂C cluster is both thermodynamically and kinetically favored under ambient conditions in which C atom acts as an activation center to dissociate H₂ molecule.²⁵ On the other hand, since the carbon atom has one less valence electron than nitrogen, the B₁₂N₁₁C cluster exhibits the electron acceptor property around the carbon atom. Although this leads to an increase in the hydrogen chemisorption energy, but it makes a worse dehydrogenation property for the corresponding hydrogenated B₁₂N₁₁C cage.⁴⁰ Furthermore, the adsorption and dehydrogenation of methanol on the B₁₁N₁₂C surface was recently studied.⁴² It is found that the B₁₁N₁₂C nanocage can effectively decompose the CH₃OH molecule with the carbon atom as an activation site. The results also indicated that in contrast to the B₁₂N₁₁C, more electrons can be transferred from the electron-rich B₁₁N₁₂C to the CH₃OH molecule and thus lower activation energy can be achieved for the dehydrogenation reaction of the methanol on this cluster.

In this study, the oxidation of CO by O₂ molecule on the B₁₂N₁₂ and B₁₁N₁₂C nanocages is investigated. Density functional theory (DFT) calculations are performed to show the strength and nature of the CO and O₂ interaction with the surface of these nanocages. The activation energies for the each reaction step on the B₁₂N₁₂ and B₁₁N₁₂C cages are discussed and the corresponding mechanisms are compared. The results of this study could be useful for designing and developing metal-free catalysts based on BN nanostructures. To the best of our knowledge, the oxidation of CO molecule over B₁₂N₁₂ and B₁₂N₁₁C nanocages is reported for the first time.

2. Computational details

All geometry optimizations and harmonic frequency calculations were performed at the M06-2X/6-31+G(2df, p) level of theory using the Gaussian 09 suite of programs.⁴³ Previous studies revealed that the M06-2X density functional is a reliable method for calculating the interaction energy and stability of noncovalent interactions.⁴⁴ Intrinsic reaction coordinate (IRC) calculations were performed in the forward and reverse directions to determine minimum-energy pathways. The adsorption energies (E_ads) of all the O₂/CO adsorbed nanocages were calculated as the difference between the total energy of the complexes and the energies of the corresponding isolated monomers at the M06-2X/6-31+G(2df, p) level. Both triplet and singlet spin states of O₂ molecule were considered. In order to check the thermodynamic feasibility of the adsorption process, the change of enthalpy (ΔH) and free-energy (ΔG) were obtained at 298.14 K and 1 atmosphere from the frequency calculations according to the following equations:where ε₀ is the total electronic energy, while H_corr and G_corr are thermal corrections which should be added to ε₀ to obtain the enthalpy and Gibbs free energy, respectively.

3. Results and discussions

3.1. Geometric structures of B₁₂N₁₂ and B₁₁N₁₂C clusters

First of all, the probable structure of fullerene-like nanocages B₁₂N₁₂ and B₁₁N₁₂C are analyzed. The optimized structure of pristine B₁₂N₁₂ cage and its corresponding geometric parameters are shown in Fig. 1. The B₁₂N₁₂ cage is the smallest stable BN fullerene³⁷ with T_h symmetry. The structural view of this small BN fullerene shows that this cluster contains from six 4-membered rings (4 MR) and eight 6-membered ring (6 MR) in which two individual B–N bonds exist: one is occupied by two 6 MRs and another by a 4 MR and a 6 MR. These bond lengths are about 1.44 and 1.48 Å, respectively, which are in good agreement with other related studies.^34,45,46 Considering the Mulliken charge density analysis indicates that the B and N atoms in B₁₂N₁₂ are positively and negatively charged with the same value of about 0.34e, respectively. Therefore, these electron-deficient (B) and electron-rich (N) sites can be regarded as a Lewis acid and bases, respectively. When one carbon atom is substituted with the B atom in the B₁₂N₁₂ cluster, the B₁₁N₁₂C nanocage is formed which its relaxed structure is depicted in Fig. 1. One can see that after C-doping, the total geometry of B₁₁N₁₂C is not significantly changed. The C–N bond length between two 6MRs decreases to 1.42 Å while the other one remains unchanged (1.48 Å). Upon C-doping, the charges redistribute on the BN cage, and the electron density is mainly accumulated around the C atom, which leads to the reduction of nearest B–N bonds length. Also, the ionic character of B–N bonds refers to a considerable electron transferring from B atoms to their adjacent N atoms. A great net charge of 0.84e is transferred from the nearest N atoms to the C atom. Moreover, spin density analysis indicates that the spin density is mainly localized over the C atom while the small pockets of opposite spin are placed on the nearest N atoms. This provides further supports for the activation of B₁₂N₁₂ cluster by C-doping.

Fig. 1 Optimized structures of pristine B₁₂N₁₂ and B₁₁N₁₂C clusters. All bond distances are in Å. Color code for each optimized structure: blue ball: N; brown ball: B; gray ball: C.

3.2. Adsorption of O₂ and CO molecules on the B₁₂N₁₂ and B₁₁N₁₂C clusters

Before exploring the oxidation of CO by O₂ molecule, it is necessary to consider the adsorption of O₂ and CO molecules over B₁₂N₁₂ and B₁₁N₁₂C cages, separately. Fig. S1 of the ESI† displays the geometric parameters of the most energetically favorable configurations of triplet (complex A) and singlet O₂ (complex B) adsorbed over the B₁₂N₁₂ nanocage. Besides, Table 1 lists the corresponding binding distances (R), adsorption energies (E_ads), net charge-transfer (q_CT), spin density, and thermodynamic parameters (ΔG₂₉₈ and ΔH₂₉₈) of these complexes. According to Fig. S1,† a triplet ground state of O₂ molecule is adsorbed in a tilted position over the B₁₂N₁₂ surface and forms the complex A. This is a common pattern for the adsorption of O₂ over different surfaces like Cu-⁴⁷ or Fe-doped⁴⁸ graphene and Si-doped BNNTs.⁴⁹ It is clear from the large distance of B–O bond (2.7 Å) that the O₂ molecule is physically adsorbed over the B₁₂N₁₂ surface with a small E_ads value of −0.09 eV. The adsorption process to form complex A is exothermic (ΔH₂₉₈ = −0.04 eV), but it has a positive ΔG₂₉₈ value of about 0.28 eV which confirms that this adsorption process is not favorable at ambient conditions (Table 1). It can be also found from analysis of the spin density distribution (reported in Table 1) that the spin density is chiefly localized over the O₂ molecule (1.98 au). Additionally, the electron density difference (EDD) map of this configuration is depicted in Fig. S1,† while the red and blue colors display the charge accumulation and depletion, respectively. The obtained EDD plot shows that the electrons are depleted around the B–O bond and are accumulated in one O atom with a small net charge of about 0.08e transfer from the gas molecule to the surface. These findings confirm that the interaction of O₂ molecule with the B₁₂N₁₂ surface is too weak to prevent desorption at room temperature. Moreover, it can be suggested that the O₂ adsorption has less effect on the electronic structure properties of the B₁₂N₁₂ due to the small E_ads and negligible charge transfer values.

Table 1 Calculated binding distances (R), adsorption energy (E_ads), net charge-transfer (q_CT), spin density, change of Gibbs free energy (ΔG₂₉₈) and change of enthalpy (ΔH₂₉₈) of triplet O₂ (A, D), singlet O₂ (B) and CO (C, E) adsorption over B₁₂N₁₂ and B₁₁N₁₂C

Surface	R (Å)	Eads (eV)	qCT (e)	Spin density (au)	ΔG298 (eV)	ΔH298 (eV)
a The estimated spin density over O2 molecule.b The estimated spin density over carbon atom of B11N12C.
B12N12				—
A	2.70	−0.09	0.08	1.98^a	0.28	−0.04
B	1.78	−0.21	0.30	—	0.24	−0.17
C	1.79	−0.10	0.17	—	0.37	−0.06
B11N12C				1.06^b
D	1.44	−1.78	0.19	0.97^a	−1.20	−1.71
E	1.82	−0.07	0.01	—	0.39	−0.03

---

On the other hand, complex B is achieved when a singlet spin state of O₂ molecule is considered. It is important to note that despite the larger E_ads (−0.21 eV) value of this complex, its total energy is smaller (less negative) than that of complex A. In addition, like complex A, this adsorption configuration is formed via an exothermic reaction while it is not a favorable reaction from the thermodynamic point of view (Table 1). In contrast with complex A, the EDD isosurface shows that the much more pronounced electron density is accumulated between the O₂ molecule and B₁₂N₁₂ (Fig. S1†). Also, a net charge of about 0.30e is transferred from the B₁₂N₁₂ surface to the O₂ molecule which shows a stronger interaction between the B and O atoms than that in complex A.

Furthermore, CO molecule is found to be weakly adsorbed over the B₁₂N₁₂ cluster while it is placed in a vertical (end-on) position right above the B atom of the surface with the small E_ads of −0.10 eV (complex C). The B–C bond length of this configuration is about 1.79 Å. Although the formation of this complex is almost exothermic (ΔH₂₉₈ = −0.06 eV), this is not a thermodynamically favorable reaction at room temperature (ΔG₂₉₈ > 0). There is also a small net charge (0.17e) transferred from the CO molecule to the B₁₂N₁₂ surface, indicated that the CO molecule can be easily desorbs from the nanocage. It is also clear from the EDD map that a small electron density exists between the B atom of the surface and C atom of the CO molecule which is originated from the large binding distance of B–C bond (Fig. S1†). Thus, it is predicted that the CO and triplet O₂ molecules can be simultaneously adsorbed over the B₁₂N₁₂ surface due to their similar adsorption energies.

The adsorption of individual O₂/CO molecules over the B₁₁N₁₂C cage is also studied in detail. It should be mentioned that both triplet and singlet spin states of O₂ molecule are considered for the O₂ adsorption configuration. However, in both cases a doublet spin state is achieved. Fig. S2† shows the most stable adsorption configuration of O₂ (configuration D) on the B₁₁N₁₂C surface in which the O–O molecule is in parallel position to the surface right above the C atom and chemisorbs with the E_ads of −1.78 eV (Table S1†). It is clear from the small C–O bond length (1.44 Å) and the large E_ads that there is a significant interaction between the O₂ molecule and the B₁₁N₁₂C surface. In contrast to the nearly unchanged O–O bond length after adsorption of triplet and singlet O₂ over B₁₂N₁₂ cluster (complexes A and B), here, the O–O bond length of the adsorbed O₂ molecule is elongated to 1.30 Å which is due to the net charge of 0.19e is transferred from the B₁₁N₁₂C to the 2π* orbitals of the O₂ molecule. It is clearly demonstrated in the EDD map in Fig. S2† that a large electron density is accumulated around the C–O bond which confirms that there is a significant chemical interaction between the gas molecule and the C atom of the B₁₁N₁₂C surface. These results indicate that the transferred electrons move from the C dopant to the adsorbed O₂ molecule. It should be noted that the adsorption of O₂ molecule on B₁₁N₁₂C cluster is exothermic (ΔH₂₉₈ = −1.71 eV). Also, it is a feasible reaction that can take place at room temperature (ΔG₂₉₈ = −1.20 eV). Besides, the spin density distribution analysis shows that a large spin density is accumulated over the O₂ molecule (0.98 au) rather than C atom (0.01 au) which provides an additional support for the activation of O₂ molecule over the B₁₁N₁₂C cluster.

In the next step, the adsorption of a single CO molecule on the B₁₁N₁₂C cluster is investigated. Complex E in Fig. S2† shows that this molecule is adsorbed in the same geometric configuration with complex C, right above the B atom of the cluster with almost the same E_ads and B–O bond length values (Table 1). Like complex C, the formation of configuration E is exothermic while it is thermodynamically impossible at room temperature (Table S1†). A net charge of about 0.01e is transferred from the CO molecule to the B₁₁N₁₂C cage. The EDD isosurface presented in Fig. S2† indicates that the electron density is depleted around the B–C bond which is caused from the weak interaction between the CO molecule and the B₁₁N₁₂C surface.

In summary, comparing the E_ads values of the adsorbed O₂ molecule over the B₁₂N₁₂ and B₁₁N₁₂C clusters clearly shows that the O₂ molecule is adsorbed more effectively over the B₁₂N₁₂C cluster and has a larger effect on it. Also, the CO molecule is considered to be adsorbed on the B₁₁N₁₂C surface later than the O₂ molecule due to its smaller (less negative) E_ads value. Therefore, in a mixture of O₂/CO molecules as the reaction gas, the B₁₁N₁₂C surface will be chiefly covered by adsorbed O₂ molecules. So, the configuration D can be used as an initial state for the CO oxidation reaction.

3.3. The CO oxidation mechanisms over the B₁₂N₁₂ cluster

There are two well-established mechanisms for CO oxidation with O₂, namely, the Langmuir–Hinshelwood (LH) mechanism and the Eley–Rideal (ER) mechanism.^50,51 In the LH mechanism, the reaction occurs between the co-adsorbed O₂ and CO species, while in the ER mechanism, the CO molecule reacts directly with the activated O₂. To investigate the reaction of CO with O₂ over the B₁₂N₁₂ cluster, the geometry optimizations are carried out considering both LH and ER mechanisms. In order to study the CO oxidation mechanism over the B₁₂N₁₂ surface, the co-adsorbed CO and triplet spin state of O₂ molecules is chosen as an initial state because of their similar adsorption energies. Previous studies demonstrated that the type of catalyst determines the exact reaction pathway. For instance, Li et al. have found that the CO oxidation proceeds via the favorable ER mechanism with a two-step route, while the LH mechanism is not kinetically favorable.⁵¹ Fig. S3† demonstrates the minimum energy pathway of the CO oxidation reactions over the B₁₂N₁₂ cluster in which the following reactions are occurred:

O₂ + CO → O_ads + CO₂ (3)

O_ads + CO → CO₂ (4)

The corresponding activation energies (E_act) and reaction energies are also reported in Table 2. The local configurations of the gas molecules over the B₁₂N₁₂ cluster and their initial state (IS), transition state (TS) and final state (P) are depicted in Fig. S4.† The co-adsorbed O₂/CO molecules over the B₁₂N₁₂ surface forms the IS-1 in which both O₂ and CO molecules are adsorbed over the surface with corresponding B–O1 and B–C bond lengths of 2.85 and 1.78 Å, respectively (Fig. S4†). It is important to note that the estimated E_ads value for CO and O₂ co-adsorption on B₁₂N₁₂ (−0.20 eV) is slightly more negative than that the sum of individual adsorption energies of these species (−0.19 eV). In addition, the relatively small difference between the adsorption energies of individual O₂ and CO molecules indicates that there is a certain probability of having O₂ and CO co-adsorbed on the B₁₂N₁₂. Then, the C atom of the CO molecule approaches to the O₂ atom and a peroxo-type structure O1–O2–C–O is produced (TS-1) over the B₁₂N₁₂ surface with the bond lengths of 1.34, 1.42 and 1.17 Å for O1–O2, O2–C and C–O, respectively (Fig. S4†). The activation energy (E_act) along the reaction pathway IS-1 → P-1 is 2.47 eV which is too high to proceeds rapidly at room temperature. Passing over the transition state TS-1, the formed CO₂ molecule with the E_ads value of −0.18 eV, desorbs easily from the surface, leaving an atomic oxygen (O_ads) adsorbed over the B atom of the B₁₂N₁₂ surface with the B–O_ads bond length and adsorption energy of 1.43 Å and −5.3 eV, respectively (Fig. S4†).

Table 2 Calculated activation energy (E_act) and reaction energy (ΔE) for different pathways of CO oxidation by O₂ molecule over B₁₂N₁₂ and B₁₁N₁₂C clusters

Reaction	Eact (eV)	ΔE (eV)
B12N12
IS-1 → P-1	2.47	−1.06
IS-2 → P-2	—	−1.40
B11N12C
IS-5 → P-5	0.58	−3.10
IS-6 → P-6	0.09	−1.51

---

Considering the high energy barrier of the CO₂ formation via the LH mechanism (Fig. S3†), the question arises whether this pathway (IS-1 → P-1) can be carried on via the ER mechanism. It is noteworthy that the E_ads of co-adsorbed O₂/CO molecules when they are proceeded via the ER mechanism in the triplet state of the O₂ molecule is about +1.26 eV which is not actually possible. Therefore, because of this positive E_ads value, the ER mechanism can't be considered for the first step.

In the following step, we check the activity of the atomic O_ads (O1) for CO oxidation via the LH mechanism. According to Fig. S4,† the configuration in which the C atom of the CO molecule is about 3.06 Å far away from the atomic O_ads is chosen as an IS-2. It is interesting to note that the P-2 is obtained via a barrier-less process. In TS-2, the O1–C–O complex is formed in a vertical position to the surface to facilitate the formation of CO₂ molecule (O1–C = 1.32 Å). Finally, in P-2, the second CO₂ molecule is formed which weakly interacted with the B₁₂N₁₂ (B–C = 2.99 Å) and hence can be easily released from the surface (E_ads = 0.21 eV). Generally, according to Table 2 and Fig. S3,† the reaction energy of these reaction steps (IS-1 → P-1 and IS-2 → P-2) are about −1.06 and −1.40 eV, respectively. It should be noted that since the second pathway with the LH mechanism proceeds via a barrier-less reaction, the ER mechanism does not consider for this reaction route.

Furthermore, the oxidation of CO molecule over the B₁₂N₁₂ surface is studied when the singlet O₂ is considered. The energy profile for the IS-3 → P-3 and IS-4 → P-4 reactions in which both LH and ER mechanisms are carefully studied is shown in Fig. S5.† Also, their local configurations along with some geometric values of the IS, TS and P are demonstrated in Fig. S6† and their corresponding thermodynamic values are listed in Table S1.† The stable configuration for IS-3 is achieved in which the CO and O₂ molecules are adsorbed over the surface via the ER mechanism. In this configuration, a five-member ring is formed and the carbon atom of the CO molecule approaches to the O2 atom with the O2–C bond length of 1.35 Å. It should be noted that the E_ads of CO + O₂ in this complex (−4.12 eV) is more negative than that in complex A due to the formation of the stable five-member ring. So, this can be used as a favorable configuration in CO oxidation reaction. Because of the great stability of this complex, the TS-3 with the O2–C bond length of 2.03 Å is difficultly formed via the high E_act of about 3.35 eV which cannot be overcome at room temperature (Fig. S5, Table S1†). Passing from TS-3, the P-3 is obtained with the reaction energy of 2.41 eV. In this state, the atomic O_ads (O1) remains in the surface with the B–O1 bond length of 1.38 Å while the first CO₂ molecule is weakly interacted to the surface (N–C = 3.04 Å) that can be released easily at ambient conditions. In the next pathway, another CO molecule is adsorbed over the B₁₂N₁₂ surface via the LH mechanism and forms the IS-4. Herein, the B–C bond length is about 1.76 Å and the carbon atom is about 2.83 Å far from the atomic O1. The CO₂ molecule is obtained passing from TS-4 with the small E_act of about 0.06 eV. It is clear from the P-4 configuration that the second CO₂ molecule can be easily desorbed from the B₁₂N₁₂ surface (Fig. S6†).

3.4. The CO oxidation mechanisms over the B₁₁N₁₂C cluster

Continually, the CO oxidation reaction is carefully investigated over the B₁₁N₁₂C surface with both LH and ER mechanisms. First, the ER mechanism is considered. The related energy profile for the CO oxidation reaction pathways IS-5 → P-5 and IS-6 → P-6 over the B₁₁N₁₂C nanocage are shown in Fig. S7.† The atomic configurations at various states along the reaction pathways are demonstrated at Fig. S8.† Beside, the corresponding E_act and reaction energy of these reaction pathways are summarized in Table 2. In the case of ER pathway, the gas-phase CO molecule is directly interacted with the adsorbed O₂ molecule. As it is clear from Fig. S8,† in the IS-5, CO molecule is physically adsorbed a little far away from the pre-adsorbed O₂ molecule with the O2–C bond length of 1.35 Å. Subsequently, the CO molecule approaches to the activated O₂ molecule in order to form TS-5 with the O2–C distance of 1.77 Å. Finally, a physisorbed CO₂ molecule is generated, with the E_ads value of −0.2 eV, which can be easily released from the cluster surface. Also, the O_ads (O1) is remained above the C atom of the B₁₁N₁₂C surface while its adsorption energy is about 2 eV more negative than that between the O and B atom of B₁₂N₁₂ cluster. This process is very close to the proposed ER pathway on the Si-doped BNNTs⁵² or N-⁵³ and Sn-doped graphene.⁵⁴ Although the reaction energy of this step is about −3.10 eV, but it is not a barrier-less pathway (E_act = 0.58 eV). Comparing with the moderate barrier of IS-7 → P-7 pathway (1.48 eV) followed by the LH mechanism (Fig. S9 and Table S1†), the ER mechanism is the favorable pathway for the CO oxidation over the B₁₁N₁₂C nanocage which is likely to take place at low temperature.

In the following, the reaction between the CO and the atomic oxygen is carefully studied based on the LH mechanism via the IS-6 → P-6 pathway (Fig. S7 and S8†). Once the first formed CO₂ molecule leaves from the surface of nanocage, the second CO is supposed to be co-adsorbed with the O1 at the reaction site. In the IS-6, the C–O1 bond is about 1.37 Å while the CO molecule is physically adsorbed over the B atom of the B₁₁N₁₂C cluster (B–C = 1.81 Å) while it is about 3.24 Å far away from the atomic O1. Attacking of O1 to the carbon atom of CO molecule generates TS-6 with the O1–C distance of 1.97 Å. The energy barrier equal to 0.09 eV is needed for the IS-6 → TS-6 reaction (Table 2). Like before, the CO₂ molecule is weakly interacted with the surface (E_ads = 0.1 eV and C–C = 3.04 Å), so, the CO₂ molecule can effectively desorbs from the B₁₁N₁₂C surface and the B₁₁N₁₂C catalyst can thus be recovered for a new round of CO oxidation.

For the ER mechanism, the process IS-6 → P-6 is completely different. As shown in Fig. S9 and S10,† the CO molecule is weakly adsorbed over the B₁₁N₁₂C surface (O1–C = 2.73 Å) and considered as the IS-8. In this state, the C atom of the CO molecule attacks the O1 to generate TS-8 while the distance between the atomic O1 and C atom of the CO molecule is decreased from 2.73 to 1.97 Å. The E_act of this reaction is calculated about 0.21 eV which is significantly larger than that in LH mechanism (0.09 eV). At the end, the final product CO₂ is formed over the B₁₁N₁₂C cluster and released from the surface at room temperature. Interestingly, it can be suggested that the oxidation of CO molecule over the B₁₁N₁₂C is preferred to proceed via the ER mechanism followed by LH mechanism.

Based on the present theoretical results, the catalytic activity of B₁₁N₁₂C nanocage toward the oxidation of CO molecule is more than that in B₁₂N₁₂ cluster due to the existence of C atom as an activation site. Moreover, the corresponding activation energy barriers of these mechanisms are significantly lowered compared with the undoped B₁₂N₁₂ case. Meanwhile, the performance of B₁₁N₁₂C nanocage as a catalyst used for the oxidation of CO with O₂ molecule may be proceed at near ambient temperatures.

4. Conclusion

In summary, the catalytic properties of pristine and C-doped B₁₂N₁₂ fullerene-like nanocages for the CO oxidation reaction were studied by means of DFT calculations. The equilibrium geometries and adsorption behavior of O₂ and CO molecules adsorbed on these nanocages were also identified. Both triplet and singlet spin states of the O₂ molecule were considered. Moreover, the CO oxidation reaction over these two different nanocages was studied via the ER and LH mechanisms, separately. Comparing the O₂/CO adsorption configurations over the B₁₁N₁₂C cluster revealed, the adsorption of O₂ molecule is more favorable than the CO molecule. It should be noted that in contrast to B₁₂N₁₂, the O₂ adsorption over the B₁₁N₁₂C cluster has a larger effect on electronic structure properties of the nanocage due to its larger E_ads and charge-transfer. In addition, a chemisorbed triplet O₂ molecule is more stable than the singlet one. Therefore, the triplet ground state of O₂ molecule was considered for the investigation of the CO oxidation reaction. Furthermore, it is found that the CO oxidation reaction over the B₁₁N₁₂C nanocage is favorably proceeded via the ER mechanism followed by LH mechanism with the E_act of about 0.58 eV, while it proceeds through the LH mechanism in both pathways over the B₁₂N₁₂ cage (E_act = 2.47 eV). According to obtained results, it can be concluded that the B₁₁N₁₂C cluster may be utilized as a favorable low-cost catalyst for CO oxidation reaction.

References

1. E. J. Peterson, A. T. De La Riva, S. Lin, R. S. Johnson, H. Guo, J. T. Miller, J. H. Kwak, C. H. F. Peden, B. Kiefer, L. F. Allard, F. H. Ribeiro and A. K. Datye, Nat. Commun., 2014, 5, 1–11.
2. S. Royer and D. Duprez, ChemCatChem, 2011, 3, 24–65.
3. H. J. Freund, G. Meijer, M. Scheffler, R. SchlÖgl and M. Wolf, Angew. Chem., Int. Ed., 2011, 50, 10064–10094.
4. N. Lopez, T. V. W. Janssens, B. S. Clausen, Y. Xu, M. Mavrikakis, T. Bligaard and J. K. Nørskov, J. Catal., 2004, 223, 232–235.
5. X. Q. Gong, Z. P. Liu, R. Raval and P. Hu, J. Am. Chem. Soc., 2004, 126, 8–9.
6. I. Nakai, H. Kondoh, T. Shimada, A. Resta, J. N. Andersen and T. Ohta, J. Chem. Phys., 2006, 124, 224712.
7. K. Bleakley and P. Hu, J. Am. Chem. Soc., 1999, 121, 7644–7652.
8. M. S. Chen, Y. Cai, Z. Yan, K. K. Gath, S. Axnanda and D. W. Goodman, Surf. Sci., 2007, 601, 5326–5331.
9. A. Alavi, P. Hu, T. Deutsch, P. L. Silvestrelli and J. Hutter, Phys. Rev. Lett., 1998, 80, 3650–3653.
10. C. Stampfl and M. Scheffler, Phys. Rev. Lett., 1997, 78, 1500–1503.
11. M. Haruta, T. Kobayashi, H. Sano and N. Yamada, Chem. Lett., 1987, 2, 405–408.
12. S. Furmaniak, A. P. Terzyk, P. A. Gauden, R. P. Wesołowski and P. Kowalczyk, Phys. Chem. Chem. Phys., 2009, 11, 4982–4995.
13. J. Gao and Y. L. Loo, Phys. Chem. Chem. Phys., 2014, 16, 10861–10865.
14. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666–669.
15. S. Lin, X. Ye and J. Huang, Phys. Chem. Chem. Phys., 2015, 17, 888–895.
16. M. Cinke, J. Li, C. W. Bauschlicher, A. Ricca and M. Meyyappan, Chem. Phys. Lett., 2003, 376, 761–766.
17. E. Yoo, T. Okata, T. Akita, M. Kohyama, J. Nakamura and I. Honma, Nano Lett., 2009, 9, 2255–2259.
18. Y. Tang, Z. Yang and X. Dai, Phys. Chem. Chem. Phys., 2012, 14, 16566–16572.
19. R. Wang and D. Zhang, Aust. J. Chem., 2009, 61, 941–945.
20. H. W. Kroto, J. R. Heath, S. C. O'Brien, R. F. Curl and R. E. Smalley, Nature, 1985, 318, 162–163.
21. R. Czochara, P. Ziaja, P. Piotrowski, R. Pokrop and G. Litwinienko, Carbon, 2012, 50, 3943–3946.
22. É. Tóth, R. D. Bolskar, A. Borel, G. González, L. Helm, A. E. Merbach and L. J. Wilson, J. Am. Chem. Soc., 2005, 127, 799–805.
23. M. K. Shukla, M. Dubey, E. Zakar, R. Namburu and J. Leszczynski, Chem. Phys. Lett., 2010, 493, 130–134.
24. Q. Sun, Q. Wang and P. Jena, Nano Lett., 2005, 5, 1273–1277.
25. H. Wu, X. Fan and J. Kuo, Int. J. Hydrogen Energy, 2012, 37, 14336–14342.
26. C. Balasubramanian, S. Bellucci, P. Castrucci, M. De Crescenzi and S. V. Bhoraskar, Chem. Phys. Lett., 2004, 383, 188–191.
27. H. S. Wu, F. Q. Zhang, X. H. Xu, C. J. Zhang and H. Jiao, J. Phys. Chem. A, 2003, 107, 204–209.
28. A. Soltani, S. G. Raz, M. R. Taghartapeh, A. V. Moradi and R. Z. Mehrabian, Comput. Mater. Sci., 2013, 79, 795–803.
29. H. Choi, Y. C. Park, Y. H. Kim and Y. S. Lee, J. Am. Chem. Soc., 2011, 133, 2084–2087.
30. M. D. Esrafili and R. Nurazar, Appl. Surf. Sci., 2014, 314, 90–96.
31. E. C. Anota, A. R. Juárez, M. Castro and H. H. Cocoletzi, J. Mol. Model., 2013, 19, 321–328.
32. J. Beheshtian, H. Soleymanabadi, A. A. Peyghan and Z. Bagheri, Appl. Surf. Sci., 2013, 268, 436–441.
33. M. D. Esrafili and R. Nurazar, Superlattices Microstruct., 2014, 67, 54–60.
34. M. D. Esrafili and R. Nurazar, Surf. Sci., 2014, 626, 44–48.
35. D. B. Zhang, E. Akatyeva and T. Dumitrică, Phys. Rev. B: Condens. Matter Mater. Phys., 2011, 84, 115431.
36. D. L. Strout, J. Phys. Chem. A, 2001, 105, 261–263.
37. S. Nigam and C. Majumder, ACS Nano, 2008, 2, 1422–1428.
38. T. Oku, A. Nishiwaki and I. Narita, Sci. Technol. Adv. Mater., 2004, 5, 635–638.
39. A. W. van den Berg and C. O. Areán, Chem. Commun., 2008, 6, 668–681.
40. H. Y. Wu, X. F. Fan, J. L. Kuo and W. Q. Deng, Chem. Commun., 2010, 46, 883–885.
41. M. D. Esrafili and R. Nurazar, J. Cluster Sci., 2015, 26, 595–608.
42. M. D. Esrafili and R. Nurazar, Comput. Mater. Sci., 2014, 92, 172–177.
43. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ö. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian, Inc, Wallingford CT, 2009.
44. Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215–241.
45. S. Yourdkhani, T. Korona and N. L. Hadipour, J. Phys. Chem. A, 2015, 119, 6446–6467.
46. F. Zhao, Y. Wang, M. Zhu and L. Kang, RSC Adv., 2015, 5, 56348–56355.
47. E. H. Song, Z. Wen and Q. Jiang, J. Phys. Chem. C, 2011, 115, 3678–3683.
48. Y. F. Li, Z. Zhou, G. T. Yu, W. Chen and Z. F. Chen, J. Phys. Chem. C, 2010, 114, 6250–6254.
49. M. D. Esrafili and N. Saeidi, Struct. Chem. DOI:10.1007/s11224-015-0590-z.
50. T. Engel and G. Ertl, J. Chem. Phys., 1987, 69, 1267–1281.
51. F. Li, J. Zhao and Z. Chen, J. Phys. Chem. C, 2012, 116, 2507–2514.
52. M. D. Esrafili, N. Saeidi and P. Nematollahi, RSC Adv., 2015, 5, 100290.
53. M. D. Esrafili, R. Mohammad-Valipour, S. M. Mousavi-Khoshdel and P. Nematollahi, ChemPhysChem, 2015, 16, 3719–3727.
54. M. D. Esrafili and N. Saeidi, Phys. E, 2015, 74, 382–387.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra25069b

---