Pd₁/BN as a promising single atom catalyst of CO oxidation: a dispersion-corrected density functional theory study

Abstract

Single metal atom catalysts exhibit extraordinary activity in a large number of reactions, and some two-dimensional materials (such as graphene and h-BN) are found to be prominent supports to stabilize single metal atoms. The CO oxidation reaction on single Pd atoms supported by two-dimensional h-BN is investigated systematically by using dispersion-corrected density functional theory study. The great stability of the h-BN supported single Pd atoms is revealed, and the single Pd atom prefers to reside at boron vacancies. Three proposed mechanisms (Eley–Rideal, Langmuir–Hinshelwood, and a “new” termolecular Eley–Rideal) of the CO oxidation were investigated, and two of them (the traditional Langmuir–Hinshelwood mechanism and the new termolecular Eley–Rideal mechanism) are found to have rather small reaction barriers of 0.66 and 0.39 eV for their rate-limiting steps, respectively, which suggests that the CO oxidation could proceed at low temperature on single Pd atom doped h-BN. The current study will help to understand the various mechanisms of the CO oxidation and shed light on the design of CO oxidation catalysts, especially based on the concept of single metal atoms.

---

1. Introduction

As a poisonous and harmful gas in the air, the carbon monoxide (CO) catalytic oxidation reaction is of practical importance to reduce or eliminate its emission from automobiles and industry. Therefore, many efforts and studies have been devoted to design more efficient catalysts for this important reaction. As typical catalysts for CO oxidation, noble metals, such as Au,¹ Pt^2–4 and Pd^3,4 have been widely investigated. Given the low catalytic activities, the poor stabilities and the high cost of the noble metal catalysts, “single metal atom” catalysts proposed by Thomas et al.⁵ have sparked worldwide interest in very recent years owing to their great activity and the reduction of cost.

Experimentally, there has been some progress on the concept of “single atom” catalysts (SAC) made through the use of mass-selected soft-landing technique, improved wet chemistry method, or atomic layer deposition methods to synthesize the single-atom species on various supports.^6–8 The stability of the single atom catalysts is a crucial challenge due to the mobility of single atoms.⁹ A practical Pt single atom catalyst supported by iron oxide was experimentally prepared for the first time by Qiao et al. (denoted Pt₁/FeO_x), and the excellent performance of Pt₁/FeO_x demonstrates that heterogeneous catalysis with single atom catalysts is no longer an unattainable goal.¹⁰ Moreover, two-dimensional (2D) materials with large specific surface area and great stability, such as graphene, and freestanding hexagonal BN monolayers (h-BN), are found to be prominent catalyst supports to stabilize single metal atoms. For example, single atomic Au on h-BN with boron vacancies exhibits great thermal stability even during the CO oxidation process,¹¹ while single Ru atoms embedded on h-BN are extraordinarily stable because of their dramatically huge diffusion energy barrier.¹²

Graphene supported individual, isolated Fe,¹³ Au,¹⁴ Pt^15,16 and Cu¹⁷ atoms present great activity and selectivity to the CO oxidation reaction and h-BN supported single atomic Au,¹¹ Ru,¹² Co,¹⁸ Cu,¹⁹ Pt²⁰ and Fe²¹ promote CO oxidation at the low temperature. Pd and its alloys, as typical catalysts for CO oxidation, present high activity.^22,23

In the current study, considering the joint system of Pd and 2D freestanding material, we focus on the stability and activity of the single atomic Pd supported by h-BN (Pd₁-BN) for CO oxidation. The stability of Pd₁-BN will be revealed by the mobility of the single Pd adatom from the calculated diffusion barrier. The entire CO oxidation process on Pd₁-BN is investigated with various adsorption configurations of CO and O₂ and the different reaction mechanisms of CO oxidation, via a new mechanism over the normal bi-molecule Eley–Rideal (ER) and Langmuir–Hinshelwood (LH) mechanisms. Whether and how single Pd atoms could promote the CO oxidation will be focused on. The current study will help to understand the various mechanisms of CO oxidation and shed light on the design of the CO oxidation catalyst based on the single metal atom concept, and various catalysts for other reactions based on the single metal atom concept.

2. Methods

All spin-unrestricted calculations were performed within dispersion-corrected density functional theory DFT (DFT-D) computations as implemented in DMol³ code embedded in Materials Studio (Accelrys, San Diego, CA), using a DFT semi-core pseudopotential²⁴ with GGA-PBE functional²⁵ with long-range dispersion correction via Tkatchenko and Scheffler’s scheme.²⁶ During geometrical optimization, the basis set cut-off was chosen to be 3.5 Å. The convergence tolerances for the geometry optimization were set to 10⁻⁵ Ha (1 Ha = 27.21 eV) for the energy, 0.002 Ha Å⁻¹ for the force, and 0.005 Å for the displacement. The electronic SCF tolerance was set to 10⁻⁶ Ha. In order to achieve accurate electronic convergence, we apply a smearing of 0.005 Ha to the orbital occupation. The reciprocal space was sampled with a (5 × 5 × 1) k-points grid generated automatically using the Monkhorst–Pack method²⁷ for the relaxation calculations. Complete linear synchronous transit (LST)/quadratic synchronous transit (QST) calculations were performed to locate transition states (TS) in DMol³ code. It is found that all the minima possess real frequencies, and the transition states have only one imaginary frequency. Additionally, the contributions of the long-range dispersion correction via Tkatchenko and Scheffler’s scheme are also tested, and we found that the long-range dispersion correction yields essentially different adsorption energies (increased by 0.05–0.18 eV), while its influence on the energy barriers and the reaction energies are negligible.

The calculated lattice constant of h-BN is 2.51 Å with the B–N bond of 1.45 Å, which is in good agreement with the previous study.²¹ The BN support is modeled by a 4 × 4 h-BN supercell with a vacuum layer of 15 Å. The test calculations suggest that the current model is suitable to reveal the CO oxidation process. The adsorption energy (E_ads) is defined as E_ads = E_adsorbate + E_support − E_{adsorbate/support}, where E_adsorbate, E_support and E_{adsorbate/support} are the total energies of the free adsorbate, the corresponding support and the support with the adsorbate, respectively. All three types of energy were derived from the calculations using the same periodic box dimensions and the same calculated setting. With this definition, a positive value indicates an exothermic adsorption.

3. Results and discussion

3.1 Single Pd atoms on h-BN

For the h-BN support, there are three typical configurations, defect-free h-BN, N-defect h-BN (N_V/BN) and B-defect h-BN (B_V/BN). First, the adsorption of single Pd atom on defect-free h-BN has been explored with the various initial structures, and there are two stable adsorption configurations found with Pd atom on the top of nitride atom and on the hollow site of hexagonal ring, respectively, as shown in Fig. 1a and b. The distances of the Pd adatom to the BN support are about 2.16 and 2.12 Å, respectively, and negligible distortion has been found for the BN support by the Pd adsorption. The calculated adsorption energies are 0.79 and 0.59 eV, respectively, both of which are much smaller than the binding energy of the Pd dimer (1.19 eV) and bulk Pd (3.89 eV). This indicates that the Pd atoms prefer to cluster on the defect-free h-BN monolayer.

Fig. 1 Top and side views of the most stable configurations of Pd adsorbed on (a) nitride site, (b) hollow site, (c) nitride and (d) boron defect supports, respectively. (e) The spin-polarized partial density of states (PDOS) projected on Pd-4d (red) and N-2p (blue) states. The Fermi level is set to zero. (f) The charge difference density between the single Pd atom and B_V-BN. For the contour plots, the charge accumulation regions are rendered in yellow while the charge deplete regions are shown in red. The contour value of the charge difference density is ±0.05 a.u. Hereafter, the pink, blue and turquoise spheres represent B, N and Pd, respectively. Bond distances are in angstroms.

For the defect h-BN, we consider both the N-defect and B-defect h-BN as supports to anchor single Pd atoms from the various initial structures.²⁸ Their most stable configurations are presented in Fig. 1c and d, and marked as Pd₁-N_V/BN and Pd₁-B_V/BN, respectively. Upon binding on the defect-BN, the formed Pd–B and Pd–N bonds secure the single Pd atom with bond lengths of 2.04 and 1.98 Å, respectively. The corresponding adsorption energies are 4.51 and 5.33 eV, respectively, both of which are larger than the binding energy of Pd dimer (1.19 eV) and Pd bulk (3.89 eV). This indicates that both defect-BN configurations provide stable anchoring sites for the single Pd atoms. The B-defect is more strongly binding than the N-defect in BN, according to the calculated formation energies²⁹ and fortuitously boron mono-vacancies are preferably fabricated experimentally by high energy charged electrons.³⁰ Thus, the Pd₁-B_V/BN configuration would be more preferable than the Pd₁-N_V/BN configuration, and the former will be focused on in the following.

To gain a deeper insight into the properties of the Pd₁-B_V/BN configuration, we perform the electronic structure analysis of the partial density of states (PDOS) and the Mulliken charge population analysis. The DOS curves are presented in Fig. 1e, and there are many overlapped peaks between the Pd-4d states and the N-2p states, indicating the hybridization between Pd-4d states and N-2p states, resulting in the strong interaction of the Pd and its three neighbor N atoms. Moreover, there are four sharp peaks (two from Pd-4d states and the other two from N-2p states) across the Fermi level (E_F), suggesting the high activity of the Pd₁-B_V/BN system. As indicated by the charge density difference (Fig. 1f), there is charge transfer from the Pd atom to the B_V/BN support, and the transferred charge is mainly localized on the three neighboring N atoms of the Pd adatom. The Mulliken charge analysis shows the charge transfer from Pd atom to B_V/BN support is 0.19|e|.

As a catalyst, the stability of single atomic Pd adsorbed on B_V/BN is highly important. Here, we investigate the mobility of the single Pd atom to reveal the stability of the Pd₁-B_V/BN system, and we take the most stable adsorption site as the initial state (IS), and the other two stable ones as the final states (FS1 and FS2, respectively). The calculated mobility barrier of the Pd atom from IS to FS1 and FS2 are 3.77 and 5.26 eV, respectively (see Fig. 2). The large diffusion barriers indicate the immobility of the single Pd atom on the B-defect h-BN. The immobility of the single Pd atom together with its strong adsorption energy (compared to the binding energy of Pd dimer and bulk) implies that the metal clustering of Pd atom would be rather difficult and the Pd₁-B_V/BN configuration is extraordinarily stable, which might allow higher temperature conditions of operation.

Fig. 2 The reaction profiles of diffusion of single Pd atom from its most stable adsorption site to the other stable adsorption sites.

3.2 Adsorption of O₂, CO, O and CO₂ on Pd₁-B_V/BN

Before investigating the CO oxidation reaction, we first study the adsorption of various species involved in CO oxidation on Pd₁-B_V/BN, including O₂, CO, O and CO₂. Various initial adsorption configurations have been considered to obtain the most stable ones, which are summarized in Fig. 3. It is found that all the species prefer to reside on the Pd site but not the BN support.

Fig. 3 Top and side views of the most stable adsorption configurations of (a) O₂, (b) CO, (c) O and (d) CO₂ on Pd₁-B_V/BN. Hereafter, the red and gray spheres represent O and C atoms, respectively.

The adsorbed O₂ molecule anchors on the Pd adatom and prefers to lie parallel to the h-BN support, and the O–O bond is enlarged from 1.23 to 1.33 Å with two Pd–O bonds of 2.12 Å being formed, and the calculated adsorption energy is 1.01 eV. According to the Mulliken charge analysis, the adsorbed O₂ is negatively charged by 0.34|e|, and the single Pd adatom is positively charged by 0.28|e|, indicating the key role of the Pd atom to activating the adsorbed O₂. Fig. 4 presents the PDOS curves of the related systems of O₂ and CO adsorption on Pd₁-B_V/BN. Upon O₂ adsorption, the curve of Pd-4d states is broadened, and there are two new hole peaks appearing above the Fermi level to overlap with the 2π* orbitals of O₂. At the energy region of −7.5 to −5.0 eV, there are three peaks of the 1π orbitals of adsorbed O₂, and the three peaks well overlap with the three peaks of the Pd-4d states, implying the strong hybridization between the Pd-4d states and the 2p states of the adsorbed O₂. The elongation of the O–O bond, the substantial charge transfer and the strong hybridization of the Pd-4d states and O-2p states suggest that the O₂ is significantly activated.

Fig. 4 PDOS for (a) O₂ and (b) CO on Pd₁-B_V/BN. The red (blue) curves represent the PDOS of Pd-4d states without (with) O₂ or CO adsorption and the magenta dashed curves represent the PDOS of the 2p states of the adsorbate (O₂ or CO). The Fermi level was set to zero.

The most stable CO adsorption configuration is presented in Fig. 3b and the CO is tilted to the Pd₁-B_V/BN plane with a C–O bond of 1.15 Å (similar to gaseous CO), and C–Pd bond of 1.91 Å. The calculated adsorption energy is 1.07 eV, which is only 0.06 eV more favorable than that for O₂ adsorption, suggesting that the adsorption of CO is slightly preferred. The Mulliken charge analysis shows that the adsorbed CO is positively charged by 0.26|e| and the Pd accepts 0.25|e| while the changes of boron and nitride atoms charges are negligible, indicating the CO and the Pd act as a charge donor and acceptor, respectively. As shown in Fig. 4b, upon CO adsorption, the Pd-4d states split into two peaks near the Fermi level, and the two peaks strongly hybridize with the 2π* orbitals of the absorbed CO. In the energy area of −12 to −7 eV, the 4σ and 1π orbitals of CO are also hybridized with the Pd-4d states. The strong hybridization between CO and Pd is in line with the strong interaction between Pd and CO.

The most stable adsorption structure of a single O atom on Pd₁-B_V/BN is presented in Fig. 3c. Compared with that of O₂ adsorption (Fig. 3a), the adsorption energy of the O atom is reduced to 0.34 eV and the O–Pd bond is reduced to 1.81 Å. The adsorbed O atom is negatively charged by 0.48|e| [about half of q(O²⁻)],³¹ indicating the formation of the O⁻ species. The O⁻ species is believed to be more reactive than O²⁻.³² As presented below, the O⁻ species shows high activity to oxidize the CO.

As the final product of CO oxidation, only the physical adsorption configuration is found for the CO₂ adsorption (Fig. 3d), with a rather small adsorption energy of 0.06 eV and a long Pd–O distance of 2.93 Å. The rather weak interaction (the small adsorption energy) together with the negligible charge transfer between the CO₂ and support indicates that the formed CO₂ species would be spontaneously released at room temperature.

3.3 CO oxidation on Pd₁-B_V/BN

Generally, the normal reaction mechanisms of the CO oxidation are the ER mechanism and LH mechanism.³³ The carbonate-like (CO₃) and peroxide-like complex (OCOO) are the essential intermediates for CO oxidation via ER and LH mechanisms, respectively. For the ER mechanism, the gas-phase CO molecules directly react with activated O₂ to form the carbonate-like intermediate or CO₂. For the LH mechanism, this involves the co-adsorption of O₂ and CO molecules, the formation of a peroxide-like intermediate complex, and desorption of a CO₂ molecule.

Interestingly, two new CO oxidation mechanisms have been proposed, denoted as termolecular Eley–Rideal mechanism (TER) by Yang et al.¹¹ and termolecular Langmuir–Hinshelwood mechanism (TLH) by Zeng et al.³⁴ In the TER mechanism, a free O₂ molecule is activated by two co-adsorbed CO molecules to form an OCO–metal–OCO intermediate, which is a pentagonal ring structure, and the intermediate configuration would dissociate into two CO₂ molecules synchronously. In the TLH mechanism, the initial state is the configuration whereby one O₂ molecule and two CO molecules are co-adsorbed on catalyst. The presence of the OCOPdOCO intermediate and the co-adsorption configuration of two CO molecules (as presented below) on the Pd₁-B_V/BN support implies the probability of the TER mechanism for CO oxidation. However, the absence of the co-adsorption configuration of one O₂ and two CO molecules suggests that the TLH mechanism is not applicable for CO oxidation on the Pd₁-B_V/BN system. Furthermore, the co-adsorption configuration of three CO molecules is also taken into consideration, and it is found that the Pd₁-B_V/BN is unable to chemisorb a third CO, which avoids the CO poisoning problem²⁰ and leaves the possibility of CO oxidation through the ER, LH and TER mechanisms.

3.3.1 ER mechanism. The reaction profiles of the ER mechanism are presented in Fig. 5, including the involved co-adsorption configurations and the transition states (TS). In Fig. 5, the physisorption configuration of CO above the pre-adsorbed O₂ on Pd₁-B_V/BN is selected as the initial state (IS). When approaching the activated O₂, CO binds to O₂ with the breakage of the O–O bond and the formation of new C–O bonds to form a carbonate-like intermediate state (MS) via an energy barrier of 1.21 eV and exothermic reaction energy of 3.70 eV (the reaction energy is defined as the change in the total energies between reactants and products). Subsequently, the carbonate-like structure dissociates by cleavage of one of C–O bonds via the reaction energy barrier of 1.29 eV and endothermic reaction energy of 0.98 eV, resulting in the formation of a molecule of CO₂ and an atomic O atom on the Pd₁-B_V/BN. The calculated adsorption energy of CO₂ on the Pd₁-B_V/BN with the adsorption of atomic O is 0.10 eV, and the CO₂ is as far as 2.93 Å from the adsorbed atomic O. Due to the rather weak adsorption and the long distance to the support, the formed CO₂ will be spontaneously released as free CO₂ at room temperature. The rate-limiting step of the current ER mechanism is the formation of the CO₂ from carbonate-like structure with the cleavage of one of the C–O bonds, with the corresponding reaction energy barrier of 1.29 eV and endothermic reaction energy of 0.98 eV.

Fig. 5 The reaction profiles of CO oxidation via the ER mechanism. Hereafter, the local configurations of the adsorbates on Pd₁-B_V/BN at initial states (IS), transition states (TS), intermediate states (MS) and final states (FS) along the minimum-energy pathway are shown in the insets, respectively.

Given that the adsorption energies of CO and O₂ are both lower (∼0.2 eV) than the reaction barrier of the formation of the carbonate-like structure, the desorption of adsorbed CO and O₂ would be slightly preferable relative to the formation of the carbonate-like structure. This indicates that the CO oxidation over Pd₁-B_V/BN via the ER mechanism is not preferable, which is similar to that of CO oxidation on Cu embedded in h-BN via the ER mechanism.¹⁹ Therefore, the LH and TER mechanism will be emphasized in the following. The oxidation of the second CO by the remaining atomic O on Pd₁-B_V/BN will be presented in the next subsection.

3.3.2 LH mechanism. Fig. 6 displays the reaction pathway of CO oxidation via the LH mechanism. Different from the ER mechanism, the co-adsorbed CO and O₂ will react to form a peroxide-like intermediate. For the initial state (IS1), the co-adsorption energy of CO and O₂ is 1.34 eV, in which only one oxygen atom (O2) of the O₂ molecule binds to the Pd atom. The co-adsorption energy is larger than those of the isolated CO (1.07 eV) and O₂ (1.01 eV), indicating the enhancement of the co-adsorption of CO and O₂. When CO and O₂ are co-adsorbed on Pd₁-B_V/BN, one of the O atoms (O1) in the adsorbed O₂ molecule starts to approach the Pd atom to form a more stable co-adsorption configuration (MS1), in which the two oxygen atoms of the O₂ molecule both bind to the Pd atom with Pd–O bonds of 2.16 Å, and the co-adsorption energy of CO and O₂ is 1.73 eV. The corresponding energy barrier and exothermic reaction energy are 0.13 and 0.39 eV, respectively. Subsequently, the O1 atom will approach the carbon atom of CO to form an OCOO intermediate (MS2), in which the O1–O2 bond is enlarged to 1.47 Å, while the O1–C bond is decreased to 1.35 Å. The corresponding reaction barrier is 0.66 eV, whereas the reaction is endothermic (0.13 eV).

Fig. 6 The reaction profiles of CO oxidation via the LH mechanism. The two O atoms of the O₂ molecule are labelled as O1 and O2, respectively.

Upon the cleavage of the O–O bond of the oxygen molecule a CO₂ molecule is formed, and the corresponding reaction energy barrier and exothermic reaction energy is 0.61 and 2.11 eV, respectively. Similarly, due to its rather weak adsorption and the long distance to the support, the formed CO₂ will be spontaneously released as free CO₂ at room temperature.

Upon the release of the formed CO₂, there is an atomic O left on the Pd₁-B_V/BN with a Pd–O bond of 1.81 Å. It is found that this atomic O is negatively charged by 0.48|e| [about half of q(O²⁻)], implying the formation of the O⁻ species.³¹ The O⁻ species is believed to be more reactive than O²⁻,³² and will readily oxidize a second CO. The 2nd CO is first physically adsorbed at the atomic O with a C–O distance of 3.03 Å. When the CO approaches the atomic O to form the second CO₂ molecule the reaction energy barrier and exothermic reaction energy are 0.23 eV and 2.94 eV, respectively. The adsorption of the formed CO₂ is rather weak (0.06 eV), suggesting that the formed CO₂ species would be released spontaneously and Pd₁-B_V/BN system is renewed and ready for a new cycle of CO oxidation.

To summarise, the corresponding rate-limiting step is TS2 (0.66 eV) with the formation of the OCOO intermediate, which is endothermic (0.13 eV). In the LH mechanism, the reaction energy of the entire reaction process from the co-adsorption of O₂ and CO to the formation of the first CO₂ is exothermic (2.37 eV), and it can compensate for the small endothermic reaction energy of the rate-limiting step (0.13 eV). In addition, the final product in FS2, CO₂, shows weak adsorption on Pd₁-B_V/BN catalyst and it will be easily desorbed, suggesting the concentration of reactants (CO and O₂) is higher on the catalyst than the product (CO₂). All these results indicate that the CO oxidation reaction via the LH mechanism is hardly inhibited by the step from MS1 to MS2 and the chemical equilibrium will shift to the formation of the product CO₂.

3.3.3 TER mechanism. As mentioned above, a new mechanism for CO oxidation, denoted as TER mechanism¹¹ is reported for single Au atoms supported by h-BN. In the TER mechanism, a free O₂ molecule is activated by two co-adsorbed CO molecules to form an OCO–metal–OCO intermediate. Here, we take the co-adsorption configuration of two CO molecules as the first stage of the TER mechanism. The reaction profiles of the TER mechanism are presented in Fig. 7, including the involved co-adsorption configurations, the transition states (TS) and the important intermediate (OCOPdOCO).

Fig. 7 The reaction profiles of CO oxidation via the TER mechanism. The two O atoms of O₂ molecule are labelled as O1 and O2, and the C and O atom of two CO molecules are labelled as C1, C2, O3 and O4, respectively.

For the configuration of two CO molecules co-adsorbed on Pd₁-B_V/BN, the adsorption energy of the second CO is 0.77 eV and the co-adsorption energy is 1.84 eV. In the reported TER mechanism of the CO oxidation on the Au–BN system, the two adsorbed CO molecules would promote the activation of the O₂ molecule, i.e. the O₂ molecule is activated by CO molecules directly.¹¹ Indeed, we also found the activation of the O₂ is promoted by the two pre-adsorbed CO molecules on the Pd₁-B_V/BN. In the initial state, two CO molecules are chemically co-adsorbed on a Pd atom with one physically adsorbed O₂ molecule, where the distances of O1–C1 and O2–C2 are 3.46 and 3.78 Å, respectively, and the angle of C1–Pd–C2 is 95.08°. The physically adsorbed O₂ molecule first approaches the two carbon atoms simultaneously via a rather small reaction energy barrier of 0.03 eV to form the OCOPdOCO intermediate, and the corresponding (exothermic) reaction energy is 1.23 eV. The formed OCOPdOCO intermediate has a pentagonal ring structure (see the “MS” of Fig. 7), in which the O–O bond is enlarged to 1.49 Å. Subsequently, the OCOPdOCO intermediate dissociates to form two adsorbed CO₂ molecules with the cleavage of the O–O bond via TS2, and the corresponding reaction energy barrier and (exothermic) reaction energy are 0.39 and 4.32 eV, respectively. Moreover, the adsorption energies of the formed CO₂ molecules are rather small (∼0.16 eV), suggesting that the formed CO₂ molecules would be released spontaneously and the Pd₁-B_V/BN system is renewed and ready for a new cycle of CO oxidation.

To summarise, in the current TER mechanism, the rate-limiting step is the formation of the two CO₂ molecules from the dissociation of the OCOPdOCO intermediate with a small reaction energy barrier of 0.39 eV. This suggests high activity of the Pd₁-B_V/BN for CO oxidation, which is mainly attributed to the “CO-promoted O₂ activation”.¹¹ The positively charged carbon atoms of the CO molecules would be a good anchor for the negatively charged O₂. The energy barrier of the rate-limiting step for TER is dramatically lower than that for LH mechanism (0.39 vs. 0.66 eV), indicating that the TER mechanism (CO-promoted O₂ activation) would be more preferable than the LH mechanism. According to the initial stage of LH (CO and O₂) and TER mechanism (2CO), when first loading one reactant (O₂) at a time, beginning with the co-adsorption of O₂ and CO followed by the formation of OCOO intermediate, the LH mechanism will take place. On the contrary, when first loading another reactant (CO) at a time, beginning with the co-adsorption of two CO followed by the formation of OCOPdOCO intermediate, the TER mechanism will take place. However, it is quite impractical to feed the two reactants sequentially. In order to improve the reaction activity by the most preferable mechanism (TER), the concentration of the CO should dominate over that of O₂ in the mixed feed. In order to improve the reaction activity by the most preferable mechanism (TER), the concentration of the CO should dominate over the O₂ in their mixed feed. CO will firstly occupy the Pd site and be oxidized by the activated O₂ via the most preferable TER mechanism, which also shrinks the possibility of the co-adsorption of CO and O₂ on Pd₁-B_V/BN, i.e. the LH mechanism.

4. Conclusion

The electronic structure analysis and the mechanisms of CO catalytic oxidation to CO₂ on Pd₁-B_V/BN are studied by DFT-D calculations, systematically. Specifically, we calculate the diffusion barrier of single Pd atom on B_V/BN to reveal the stability of the Pd₁-B_V/BN system. It is found that the metal clustering of Pd atom is disfavored and the Pd₁-B_V/BN system is stable.

Summarizing our calculated reaction energy barriers for all of the CO oxidation mechanisms (ER, LH, TER), it is concluded that the traditional mechanism (LH) and the new mechanism (TER) are both preferable for CO oxidation on Pd₁-B_V/BN under experimental conditions because of the high reaction energy barrier of the rate-limiting step in ER mechanism. The rate-limiting steps for the preferable LH and TER mechanisms are the formation of the OCOO intermediate and the formation of the two CO₂ molecules from the dissociation of the OCOPdOCO intermediate, respectively, and the corresponding reaction barriers of the rate-limiting steps for the two mechanisms are 0.66 and 0.39 eV, respectively. Given the relatively small energy barriers for LH and TER (0.66 and 0.39 eV, respectively), we believe that both two mechanisms are possible and could occur at low temperature. Also, the energy barrier of the rate-limiting step for TER is dramatically lower than that for LH mechanism (0.39 vs. 0.66 eV), indicating that the TER mechanism (CO-promoted O₂ activation) would be more preferable than the LH mechanism.

Utilization of the novel properties of SACs seems to provide vast opportunities for their applications in heterogeneous catalysis. The current study will help to understand the various mechanisms of the CO oxidation and shed light on the design of CO oxidation catalysts based on the single metal atom concept, to lower CO emissions, and solve the growing environmental problem. With the establishment of SACs as a new concept and a thorough elucidation of the nature of single atom catalysis, we believe that the development and understanding of SACs might be a new frontier in heterogeneous catalysis.

Acknowledgements

This work is supported by the National Natural Science Foundation of China (Grant No. 51401078, and 11147006), and sponsored by Program for Science & Technology Innovation Talents in Universities of Henan Province (Grant No. 15HASTIT016). Supported by the High Performance Computing Center of Henan Normal University and the Science Foundation for the Excellent Youth Scholars of Henan Normal University. Doctor C. He also acknowledges the support from the Henan Joint Funds of the National Natural Science Foundation of China (Grant No. U1404216, and U1504108), and the Natural Science Foundation of Nanyang Normal University (No. ZX2014088 and QN2015020). We are grateful to Computing Center of Jilin Province for essential support.

References

1. Z.-P. Liu, P. Hu and A. Alavi, J. Am. Chem. Soc., 2002, 124, 14770.
2. K. Bleakley and P. Hu, J. Am. Chem. Soc., 1999, 121, 7644.
3. A. Eichler, Surf. Sci., 2002, 498, 314.
4. X.-Q. Gong, Z.-P. Liu, R. Raval and P. Hu, J. Am. Chem. Soc., 2004, 126, 8.
5. J. M. Thomas, R. Raja and D. W. Lewis, Angew. Chem., Int. Ed., 2005, 44, 6456.
6. B. Qiao, A. Wang, X. Yang, L. F. Allard, Z. Jiang, Y. Cui, J. Liu, J. Li and T. Zhang, Nat. Chem., 2011, 3, 634.
7. S. Sun, G. Zhang, N. Gauquelin, N. Chen, J. Zhou, S. Yang, W. Chen, X. Meng, D. Geng and M. N. Banis, Sci. Rep., 2013, 3, 1775.
8. X. Zhang, J. Guo, P. Guan, C. Liu, H. Huang, F. Xue, X. Dong, S. J. Pennycook and M. F. Chisholm, Nat. Commun., 2013, 4, 1924.
9. A. Uzun, V. Ortalan, N. D. Browning and B. C. Gates, J. Catal., 2010, 269, 318.
10. X.-F. Yang, A. Wang, B. Qiao, J. Li, J. Liu and T. Zhang, Acc. Chem. Res., 2013, 46, 1740.
11. K. Mao, L. Li, W. Zhang, Y. Pei, X. C. Zeng, X. Wu and J. Yang, Sci. Rep., 2014, 4, 5441.
12. C. Huang, X. Ye, C. Chen, S. Lin and D. Xie, Comput. Theor. Chem., 2013, 1011, 5.
13. Y. Li, Z. Zhou, G. Yu, W. Chen and Z. Chen, J. Phys. Chem. C, 2010, 114, 6250.
14. Y.-H. Lu, M. Zhou, C. Zhang and Y.-P. Feng, J. Phys. Chem. C, 2009, 113, 20156.
15. X. Liu, Y. Sui, T. Duan, C. Meng and Y. Han, Phys. Chem. Chem. Phys., 2014, 16, 23584.
16. Y. Tang, Z. Yang and X. Dai, Phys. Chem. Chem. Phys., 2012, 14, 16566.
17. E. H. Song, Z. Wen and Q. Jiang, J. Phys. Chem. C, 2011, 115, 3678.
18. S. Lin, X. Ye, R. S. Johnson and H. Guo, J. Phys. Chem. C, 2013, 117, 17319.
19. X. Liu, T. Duan, Y. Sui, C. Meng and Y. Han, RSC Adv., 2014, 4, 38750.
20. X. Liu, T. Duan, C. Meng and Y. Han, RSC Adv., 2015, 5, 10452.
21. P. Zhao, Y. Su, Y. Zhang, S.-J. Li and G. Chen, Chem. Phys. Lett., 2011, 515, 159.
22. D. Palagin and J. P. K. Doye, Phys. Chem. Chem. Phys., 2015 10.1039/C5CP00889A.
23. Y. Zhou, Z. Wang and C. Liu, Catal. Sci. Technol., 2015, 5, 69.
24. B. Delley, Phys. Rev. B: Condens. Matter, 2002, 66, 155125.
25. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
26. A. Tkatchenko and M. Scheffler, Phys. Rev. Lett., 2009, 102, 073005.
27. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1976, 13, 5188.
28. A. Zobelli, A. Gloter, C. Ewels, G. Seifert and C. Colliex, Phys. Rev. B: Condens. Matter Mater. Phys., 2007, 75, 245402.
29. A. Du, Y. Chen, Z. Zhu, R. Amal, G. Q. Lu and S. C. Smith, J. Am. Chem. Soc., 2009, 131, 17354.
30. C. Jin, F. Lin, K. Suenaga and S. Iijima, Phys. Rev. Lett., 2009, 102, 195505.
31. Z. Lu, D. Ma, L. Yang, X. Wang, G. Xu and Z. Yang, Phys. Chem. Chem. Phys., 2014, 16, 12488.
32. I. Kocemba and J. M. Rynkowski, Catal. Today, 2011, 169, 192.
33. K. J. Laidler, Pure Appl. Chem., 1996, 68, 149.
34. C. Liu, Y. Tan, S. Lin, H. Li, X. Wu, L. Li, Y. Pei and X. C. Zeng, J. Am. Chem. Soc., 2013, 135, 2583.

---