DFT insight into the effect of Cu atoms on adsorption and dissociation of CO₂ over a Pd₈/ TiO₂(101) surface

Abstract

In order to improve the photocatalytic activity of a bimetallic cocatalyst, understanding its mechanism is very important for the development of a CO₂ photocatalyst. In this study, density functional theory (DFT) calculations were performed to investigate CO₂ adsorption and dissociation over Pd–Cu bimetallic clusters loaded on a TiO₂(101) surface, aiming at understanding the origin of the effect caused by the presence of Cu. The results demonstrated that the introduction of a Cu atom has a dual effect on the adsorption and dissociation of CO₂: (1) it provides the positive polarization charge center to enhance CO₂ adsorption, and (2) it up-shifts the d-band center of the Cu atom to improve the activation of CO₂. Thus, the activity of the Pd₇Cu₁/TiO₂(101) surface, as compared with that of the Pd₈/TiO₂(101) surface, can be significantly improved, and the active center is the introduced Cu atom. This result is not only helpful for the development of effective CO₂ photocatalysts but also crucial to understand the basic mechanism of bimetallic catalysis.

---

1. Introduction

In recent decades, the significant increase in the amount of carbon dioxide (CO₂) in the atmosphere has been widely regarded as a global environmental problem.¹ Therefore, extensive research has been carried out to reduce CO₂ levels. In particular, the photocatalytic reduction of CO₂ into hydrocarbon fuels and chemicals is considered an effective method of controlling and utilizing CO₂.^2–4 Photocatalytic CO₂ reduction by H₂O was first reported more than 30 years ago and is currently gaining more attention, because sunlight can be used as the primary energy source for this reduction. There have been many studies on CO₂ photoreduction, and various materials (especially semiconductors) have been evaluated for this purpose.^2–7 Among all types of semiconductors, TiO₂ has been the most commonly used photocatalyst because of its many advantages, including chemical and thermal stability, abundance, low toxicity, low cost, and high UV photoactivity.^8–10

Although TiO₂ is the most explored semiconductor for the photocatalytic reduction of CO₂, its efficiency is still far from optimal due to the lack of visible light photoreaction, rapid electron–hole recombination of photogenerated charges, and inefficient CO₂ capture.^11,12 The combination of TiO₂ and other active ingredients is expected to overcome these limitations. In the past few decades, noble metals, such as Au, Pt, Ag, Cu, and Pd, have been studied as cocatalysts to promote the capability of TiO₂ semiconductors in photocatalytic CO₂ activation and conversion.^11–18 In addition, bimetallic cocatalysts (such as PdCu, AuCu, CuPt, and AgPd, etc.) have higher CO₂ conversion activity than single metal nanoparticles (NPs).^{11,12,16,19–25} For example, Garcia's group reported that Au and Cu loaded TiO₂ photocatalyst in the appropriate Au/Cu ratio is an extremely efficient material for the solar-light reduction of CO₂ to CH₄.¹¹ Tan et al. reported that the Ag/Pd bimetals supported on N-doped TiO₂ nanosheet exhibit high selectivity for CO₂ conversion.¹² Recently, Huang et al. reported that the efficiency of CO₂ hydrogenation to C₂H₅OH can be optimized by adjusting the composition of Pd–Cu NPs and catalyst support.²⁵ Long et al.¹⁶ also indicated that isolated Cu atoms in a Pd lattice can form highly selective active sites for photocatalytic conversion of CO₂ to CH₄.

It has been established that bimetallic cocatalysts often exhibit better catalytic performance than their corresponding elemental metal counterparts due to their composition and synergic effects on the catalytic properties.²⁶ However, the mechanism for improving photocatalytic activity and selectivity of bimetallic cocatalysts remains unclear and needs further study. Furthermore, the photoreduction of CO₂ is a complex reaction process, mainly encompasses the following elementary steps: (i) photon absorption and excited carrier generation; (ii) activation of CO₂ to form CO₂^δ−; (iii) dissociation of the C–O bond; and (iv) desorption of reduced products from the active sites.⁴ Among them, some of the main challenges have not been completely resolved, such as the complicated activation and adsorption mechanisms of CO₂, the mechanisms and pathways of photocatalytic reaction, and low efficiency and selectivity of different products.¹³ Therefore, it is essential to use theoretical calculations to explain the role of bimetallic cocatalysts in enhancing the activity, selectivity and stability of semiconductor photocatalysts, and to provide theoretical basis for design of new, high-performance CO₂ reduction cocatalysts. In this study, we chose TiO₂(101) loaded with Pd–Cu bimetallic NPs, which is reported to be an effective catalyst for the photocatalytic reduction of CO₂,¹⁶ to explore the catalysis of bimetallic nanoparticles, and to clarify how the presence of Cu atom improves the catalytic activity of Pd-loaded TiO₂(101) surface. The (101) surface of anatase was selected because this surface is the most stable. The calculation results indicates that the electronic structure of supported Pd–Cu clusters is very important to the catalytic performance. The strong positive polarization potential and the elevated Cu d-band center make Pd₇Cu₁ have higher activity on TiO₂(101), which promotes adsorption and dissociation of CO₂.

2. Computational methods

All of the DFT²⁷ calculations were carried out with the GGA-PBE^28–31 functional using the CASTEP package.³² Electronic wave functions were expanded in a plane-wave basis set, and ultrasoft pseudopotentials were used to describe the ionic cores.³³ The cutoff energy was set to 400 eV, and the Monkhorst–Pack³⁴ k-points sampling was generated with a 2 × 2 × 1 grid. In order to determine the activation barriers of CO₂ dissociation, the complete LST/QST method was used to search for the transition states.³⁵ To confirm the transition states, we performed nudged elastic band (NEB) calculations using TS confirmation.

It is well-known that the standard DFT calculation will greatly underestimate the band gap due to the insufficient cancellation of the self-interaction energy inherent in the DFT functional, which may affect the calculation results.^36–38 To investigate this possibility, we did the test calculations by using the DFT+U method for Pd₈ cluster on TiO₂(101) surface and CO₂ on Pd₈/TiO₂ (101) surface. U = 3.5 eV was chosen according to the literature.^36,39 It can be seen from Table S1 in ESI† that the charge distribution on the Pd₈ cluster calculated by DFT+U is similar to that calculated by DFT, and the adsorption stability trend of Pd₈/TiO₂(101) to CO₂ is consistent (see Fig. S1 in ESI†). Because we only focus on qualitative trend analysis, the DFT is used in the following calculation.

A six layer slab with a (3 × 3) supercell was adopted to simulate the TiO₂(101) surface. The Pd₈ cluster has been studied for a long time and widely reported in the literature. Previous DFT calculations have shown that the lowest energy structure of the free Pd₈ cluster has a bicapped octahedral geometry with D_2d symmetry.^40,41 Therefore, Pd₈ clusters with D_2d symmetry were used in this work, as shown in Fig. 1a. During the calculation, the three top layers and the adsorbates relaxed. A vacuum region of 15 Å was chosen for TiO₂(101). The free CO₂ molecule was optimized in a 10 Å × 10 Å × 10 Å unit cell. The optimized C–O bond length and O–C–O angle are 1.178 Å and 180.0°, respectively, which are consistent with the experimental⁴² and theoretically⁴³ values reported previously.

Fig. 1 (a) Pd₈ cluster, (b) side view of the Pd₈/TiO₂(101) surface, and (c) top view of the Pd₈/TiO₂(101) surface.

The adsorption energy of CO₂ is calculated as follows:

E_ads = E_{CO2 + surface} − [E_CO2 + E_surface]

where E_{CO2 + surface} represents the energy of the Pd–Cu/TiO₂(101) surface with the adsorbed CO₂, E_CO2 represents the energy of free CO₂, and E_surface represents the energy of the Pd–Cu/TiO₂(101) surface.

3. Results and discussion

3.1. Pd₈ and Pd₇Cu₁ clusters on TiO₂(101)

In order to find the most stable configuration of Pd₈ on TiO₂(101) surface, we constructed various possible configurations (see Fig. S2 in ESI†). Through optimization, we found that the structure shown in Fig. 1 is the most stable. As for Pd₇Cu₁ on TiO₂(101) surface, we tested eight possible replacing modes in order to find the most stable configuration, as shown in Fig. S3.† We chose the most stable configuration for further research, as shown in Fig. 2. A comparison to the optimized Pd₈ cluster shows that the Pd–TiO₂(101) interactions result in a completely different structure. On TiO₂(101), the cluster adopts a bilayer structure. This is due to the fluxionality of the metal cluster.

Fig. 2 The side (a) and top view (b) of the most stable model of the Pd₇Cu₁ cluster on the TiO₂(101) surface.

The charge density difference and Mulliken charge analysis for the Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces were calculated to further understand the interaction between the Pd₈ (or Pd₇Cu₁) cluster and the TiO₂ surface from the perspective of electronic structure. As shown in Fig. 3a, when Pd₈ or Pd₇Cu₁ clusters supported on the surface of TiO₂(101), obvious charge transfer occurs at the interface, which is a long-range interaction that affects the charge distribution of the Pd or Cu atoms in the upper layer, and can lead to charge polarization between the clusters and the TiO₂(101) surface. It can be seen from the electrostatic potential of Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) in Fig. S4† that the polarization potentials of Pd₈ and Pd₇Cu₁ are positive, while the polarization potentials of TiO₂(101) surface are negative. Therefore, when the interface forms and reaches an equilibrium, the final polarized electric field is from the clusters to the TiO₂(101) surface.

Fig. 3 Charge density difference for the Pd₈/TiO₂(101) (left) and Pd₇Cu₁/TiO₂(101) (right) surfaces.

As shown in Fig. 3b, although both the Pd₈ and Pd₇Cu₁ clusters can provide electrons to the TiO₂(101) surface, the charge transfer of each atom is different. Mulliken charge analysis also confirmed this conclusion. The results in Table 1 show that all interfacial Pd atoms (from Pd-1 to Pd-4) of the Pd₈ cluster are positively charged. However, all of the top-layered Pd atoms (from Pd-5 to Pd-8) are negatively charged. For the Pd₇Cu₁ cluster, because the electronegativity of Cu is lower than that of Pd, the Cu-6 can provide its electrons to the neighboring Pd atoms and become positively charged.

Table 1 The Mulliken charges carried by Pd and Cu atoms on the Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces

Pd8/TiO2(101)	Pd-1	Pd-2	Pd-3	Pd-4	Pd-5	Pd-6	Pd-7	Pd-8	Total
Charge/|e|	0.19	0.13	0.08	0.09	−0.10	−0.01	−0.08	−0.01	0.29
Pd7Cu1/TiO2(101)	Pd-1	Pd-2	Pd-3	Pd-4	Pd-5	Cu-6	Pd-7	Pd-8	Total
Charge/|e|	0.17	0.16	0.14	0.04	−0.18	0.14	−0.10	−0.05	0.32

---

3.2. Adsorption and activation of CO₂ over Pd₈/TiO₂(101)

For the Pd₈/TiO₂(101) surface, the most stable CO₂ adsorption configuration is shown in Fig. 4a, labeled as Pd₈–CO₂. The other unstable adsorption models are displayed in Fig. S5.† We summarized our calculated results in Table 2. In order to describe conveniently, the two oxygen atoms of the CO₂ are labeled Oa and Ob, respectively.

Fig. 4 Top and side views of CO₂ adsorption configuration on (a) the Pd₈/TiO₂(101) surface and (b) the Pd₇Cu₁/TiO₂(101) surface. Distances are in Å.

Table 2 Adsorption energies and structural parameters of CO₂ adsorbed on Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces

Configurations	C–Oa bond (Å)	C–Ob bond (Å)	Oa–C–Ob angle (°)	Eads (eV)
Pd8–CO2	1.249	1.233	141.7	−0.58
Pd7Cu1–CO2	1.281	1.249	135.9	−1.06
CO2 molecule	1.180	1.180	180.0

---

In configuration Pd₈–CO₂ (Fig. 4a), CO₂ is adsorbed on Pd-6 and Pd-7 atoms in form of a bidentate carbonate, with an adsorption energy of −0.58 eV. The Oa bonds to the Pd-6 atom, and the C atom bonds to the Pd-7 atom. The bond lengths of Oa–Pd-6 and C–Pd-7 bonds are 2.175 Å and 2.004 Å, respectively. As shown in Table 2, compared with the free CO₂ molecule, the adsorbed CO₂ in configuration Pd₈–CO₂ has been heavily distorted. The length of the C–Oa and C–Ob bonds are both elongated, and the Oa–C–Ob angle is reduced to 141.7°, which indicates that CO₂ is activated. In addition, it can be seen from Table S2† that the adsorbed CO₂ has a negative charge of −0.43|e|, as a consequence of electron transfer from surface to CO₂, and leading to a bent anionic CO₂^δ− species. The CO₂^δ− can increase the reactivity of CO₂, including the breaking of C–O bond. Local density of states (LDOS) analysis was also conducted to further understand the interaction between the Pd₈/TiO₂(101) surface and adsorbed CO₂. Fig. S6a† shows the LDOS of free CO₂; Fig. S6b† shows the LDOS of Pd₈–CO₂ configuration. From Fig. S6b,† for C and Pd-7 (or Oa and Pd-6) atom in configuration Pd₈–CO₂, it can be clearly seen that the orbital overlap is more obvious in the range of −10 to 0 eV. This indicates that the interaction between the CO₂ and the Pd₈/TiO₂(101) is strong.

3.3. Adsorption and activation of CO₂ over Pd₇Cu₁/TiO₂(101)

For CO₂ adsorption on Pd₇Cu₁/TiO₂(101), the most stable adsorption configuration is shown in Fig. 4b, labeled as Pd₇Cu₁–CO₂. Other unstable adsorption configurations are shown in Fig. S5.† The most stable adsorption configuration Pd₇Cu₁–CO₂ appears at the Cu-6 site, which is assisted by nearby Pd-7 and Pd-8 atoms. In configuration Pd₇Cu₁–CO₂, the C atom of CO₂ bridges two Pd atoms (Pd-7 and Pd-8), and the Oa atom bonds to the Cu-6 atom. The bond lengths of the Oa–Cu-6, C–Pd-7, and C–Pd-8 bonds are 1.975 Å, 2.093 Å, and 2.127 Å, respectively. This configuration has a high exothermic adsorption energy of −1.06 eV. According to Table 2, the bond lengths of C–O_a and C–O_b bonds are elongated to 1.281 Å and 1.249 Å, respectively, which are longer than those of gas-phase CO₂ molecule, and the Oa–C–Ob angles are reduced to 135.9°, indicating CO₂ is activated upon adsorption on Pd₇Cu₁/TiO₂(101) surface. The Mulliken charge analysis in Table S2† shows that the adsorbed CO₂ on Pd₇Cu₁/TiO₂(101) surface has a negative charge of −0.61|e|. This indicates that electron transfer occurs from surface to CO₂, and forms a negatively charged CO₂^δ− species. The LDOS in Fig. S6c† shows that the resonant peaks lie mainly in the range of −10 to 0 eV, indicating that the adsorbed CO₂ has a strong tendency to hybridize with the Pd₇Cu₁/TiO₂(101) surface at lower energy levels. Moreover, it can be seen from Table 2 that the presence of monatomic Cu significantly increases the adsorption stability of CO₂. Structurally, Pd₇Cu₁–CO₂ is similar to Pd₈–CO₂ on the Pd₈/TiO₂(101) surface without Cu (Fig. 4a), but its adsorption energy is more negative (the adsorption energy increases by 0.48 eV). In addition, when CO₂ adsorbs on the Pd₈/TiO₂(101) surface, the Oa–Pd-6 bond distance is 2.175 Å. However, when CO₂ adsorbs on the Pd₇Cu₁/TiO₂(101) surface, the bond distance of the Oa–Cu-6 is shortened to 1.975 Å. This indicates that the interaction between CO₂ and surface becomes stronger. The Mulliken charge analysis of CO₂ on Pd₇Cu₁/TiO₂(101) (Table S2†) showed the largest gain of electrons, suggesting stronger formation of CO₂^δ− anions. This suggests that the introduction of Cu atom enhances the activation of CO₂. Thus, Pd₇Cu₁/TiO₂(101) is more favorable for CO₂ adsorption and activation compared to Pd₈/TiO₂(101). This conclusion is similar to our previous work, where we also found the presence of Cu atoms to promote adsorption and activation of CO₂ on Au/TiO₂(101).⁴³ In addition, this result is in good agreement with a recent experimental report by Long et al.¹⁶ who studied photocatalytic CO₂ reduction by H₂O on clean TiO₂, Pd_xCu₁–TiO₂ (x = 1, 3, 5, 7, 9, 11), and Pd–TiO₂. They found that the Pd_xCu₁ alloys with isolated Cu atoms in Pd lattice can provide highly active sites to enhance the adsorption and activation of CO₂.

3.4. Dissociation of CO₂ over Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101)

Having elucidated the adsorption and activation of CO₂, we now discuss the effect of bimetallic cocatalyst Pd–Cu on the dissociation of CO₂. There are two possible pathways for CO₂ dissociation: direct dissociation or H-assisted dissociation. Here, H comes from the dissociation of H₂O in the system, see Fig. S7.† We first study the direct dissociation path. The potential energy profiles are displayed in Fig. 5. Optimized structures of the transition states and products involved in this pathway are also included in Fig. 5. Starting from Pd₈–CO₂, CO₂ subsequently dissociates into adsorbed CO* and O* on the Pd₈/TiO₂(101) surface through the transition state TS1. Here the asterisk (*) indicates an adsorbed species. As represented by the black line in Fig. 5, this process is endothermic by 1.44 eV and has an activation barrier of 2.80 eV. Finally, the CO* can be hydrogenated to HCO* via a C–H bond formation (see Fig. S8†) or desorbs from the surface into the gas phase with a heat of 0.28 eV needed to fulfill this process (see Fig. 5). And at the same time, the O* may migrate to surface oxygen vacancy (one of the most common defects in metal oxides) and becomes lattice oxygen or combine with H* in the system to form surface hydroxyl groups (OH*). For CO₂ dissociation on the Pd₇Cu₁/TiO₂(101) surface (the red line in Fig. 5), beginning with Pd₇Cu₁–CO₂, the adsorbed CO₂ dissociated into CO* and O* crosses a barrier of 2.18 eV and is endothermic by 0.59 eV. Then, CO* desorption or further hydrogenation to generate HCO*. Therefore, from the viewpoint of activation barrier and reaction energy, the Pd₇Cu₁/TiO₂(101) surface is more favorable for CO₂ direct dissociation than the Pd₈/TiO₂(101) surface, both kinetically and thermodynamically.

Fig. 5 Potential energy profile for the direct dissociation of CO₂ on the Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces. Distances are in Å.

Next, we study the H-assisted CO₂ dissociation. The existence of the H atom may activate CO₂ via formation of COOH* or HCOO* species, which will alter the potential of CO₂ dissociation. We started with the most stable adsorption configurations Pd₈–CO₂ and Pd₇Cu₁–CO₂ and placed an H atom on the Pd-5 and Pd-8 site. For COOH* formation, the structural tends to form the cis-COOH* (the formation of trans-COOH* is shown in Fig. S9†), in which the H atom attack the Oa atom of CO₂ to form an Oa–H bond, as shown in Fig. 6. On Pd₈/TiO₂(101) surface, the potential energy profile in Fig. 6 (black line) shows that the energy barrier for the first step of the cis-COOH* formation is 1.62 eV (TS3). This process is endothermic by 0.82 eV. Then the C–Oa band breaks and forms adsorbed CO* and OH* through the transition state TS4, with an activation barrier of 0.78 eV. This process is exothermic by 0.21 eV. Thus, the rate-determining step of this route is the process of forming cis-COOH*. On the Pd₇Cu₁/TiO₂(101) surface, the energy profile in Fig. 6 (red line) shows that the first step of the cis-COOH* formation is endothermic by 0.34 eV and has an activation barrier of 1.49 eV (TS5). Then the C–Oa band breaks and forms CO* and OH* through the transition state TS6, with an activation barrier of 0.30 eV. This process is exothermic by 0.60 eV. The activation barrier of TS5 is higher than TS6, indicating that the formation of cis-COOH* is also the rate-determining step in this reaction. The above results for H-assisted CO₂ dissociation via the cis-COOH* path on Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces show that the activation barrier of the rate-determining step (TS5, 1.49 eV) on the Pd₇Cu₁/TiO₂(101) surface is lower than on the Pd₈/TiO₂(101) surface (TS3, 1.62 eV). Moreover, the corresponding reaction energies are 0.97 and 0.34 eV, indicating that Pd₇Cu₁/TiO₂(101) surface is more favorable for CO₂ dissociation than Pd₈/TiO₂(101) surface. Therefore, compared to the Pd₈/TiO₂(101) surface, the H-assisted CO₂ dissociation via the cis-COOH* path on the Pd₇Cu₁/TiO₂(101) surface is more favorable both kinetically and thermodynamically. This result indicates that the presence of Cu-6 atom promotes the H-assisted CO₂ dissociation process.

Fig. 6 Potential energy profile for the H-assisted dissociation of CO₂ via the cis-COOH* path on the Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces. Distances are in Å.

As for HCOO* formation, the H atom attack the C atom of CO₂ to form an C–H bond, as shown in Fig. 7. On Pd₈/TiO₂(101) surface, a stable bidentate HCOO* configuration was found, in which the two O atoms bond to surface Pd-6 and Pd-7, respectively. CO₂ hydrogenation to HCOO* requires surmounting a barrier of 1.45 eV, and the reaction is 0.13 eV endothermic. Next, the HCOO* cleaves to form HCO* and O* in a 1.27 eV endothermic step with 2.84 eV energy barrier, limiting the rate of reaction. Then the H atom of HCO* combine with O* and C–H band breaks to form CO* and OH* with an energy barrier of 1.85 eV and reaction energy of 0.99 eV. Finally, CO* desorption or further hydrogenation to generate HCO*. On the Pd₇Cu₁/TiO₂(101) surface, the energy profile in Fig. 7 (red line) shows that the first step of the HCOO* formation is endothermic by 0.08 eV and has an activation barrier of 1.24 eV. Subsequently, the C–Oa band breaks and forms HCO* and O* through the transition state TS11, with an activation barrier of 2.50 eV. This process is endothermic by 1.13 eV. Then the H atom of HCO* combine with O* and C–H band breaks to form CO* and OH* with an energy barrier of 1.57 eV and reaction energy of 0.92 eV. Finally, CO* desorption or further hydrogenation to generate HCO*. This result for H-assisted CO₂ dissociation via the HCOO* path on Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces show that the activation barrier of the rate-determining step on the Pd₇Cu₁/TiO₂(101) surface is lower than on the Pd₈/TiO₂(101) surface (2.50 vs. 2.84 eV). Moreover, the corresponding reaction energies are 1.09 and 0.82 eV, indicating that Pd₇Cu₁/TiO₂(101) surface is more favorable for CO₂ dissociation than Pd₈/TiO₂(101) surface. Therefore, compared to the Pd₈/TiO₂(101) surface, the H-assisted CO₂ dissociation via the HCOO* path on the Pd₇Cu₁/TiO₂(101) surface is also more favorable both kinetically and thermodynamically, indicating that the presence of Cu-6 atom promotes the H-assisted CO₂ dissociation process. In addition, compared to the COOH*-mediated route, HCOO*-mediated path is less competitive because of its higher barrier (2.84 and 2.50 vs. 1.62 and 1.49 eV). Combined with CO₂ adsorption and activation over Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces, we can conclude that the catalytic activity of the Pd₇Cu₁/TiO₂(101) is higher than that of Pd₈/TiO₂(101).

Fig. 7 Potential energy profile for the H-assisted dissociation of CO₂ via the HCOO* path on the Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces. Distances are in Å.

3.5. Discussion

We have studied adsorption, activation, and dissociation of CO₂ over Pd–Cu supported on TiO₂(101). In this section, the enhancing effect of the Cu on the catalytic activity of the Pd supported TiO₂(101) is discussed. It is a well-known fact that in addition to the surface atomic structure, the surface electronic states are also closely related to the catalytic activity. Therefore, in order to determine the intrinsic factors that promote the catalytic activity of Pd₇Cu₁/TiO₂(101), a detailed comparison of the surface electronic states between the surfaces of Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) was conducted.

As shown in Fig. 8, the 2D electric potential maps of Pd₇Cu₁/TiO₂(101) shows that the electric potential distribution on the Cu-6 and Pd-8 atoms is no longer uniform due to the large charge polarization between Cu-6 and Pd atoms. The positive polarization potential of the Cu-6 atom is higher than that of the Pd-8 atom. This means that the electric field on the Cu-6 atom is higher than that on the Pd-8 atom. This response can promote the adsorption of CO₂ (in which the O atoms are negatively charged) by increasing the electrostatic attraction between the CO₂ and the Cu-6 atom. Therefore, the surface activity of the Pd₇Cu₁/TiO₂(101) is improved compared to the Pd₈/TiO₂(101), and the active center is the introduced Cu atom. It is precisely because of the stable adsorption of CO₂ on Cu-6 atoms and the special electric field on Cu-6 atoms that the reactivity of CO₂ is enhanced. This result is consistent with previous results. Jia et al.⁴⁴ investigated the CO oxidation on Ru–Pt supported on TiO₂(101) and found that the special electric field on the top-layered Ru atom can enhance the stability and reactivity of CO + O₂. Luo et al.⁴⁵ reported that the optimal performance shows a strong dependence on the interaction between CO₂ and the local electric field.

Fig. 8 The 2D electric potential maps for the (a) Pd₈/TiO₂(101) and (b)Pd₇Cu₁/TiO₂(101) surfaces.

Fig. 9 is DOS projected onto the d orbitals of Pd-6 (or Cu-6) and Pd-8 on the Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces. It can be seen from the Fig. 9 that the d-band center (ε_d) of the Cu-6 atom on Pd₇Cu₁/TiO₂(101) is higher than that of the Pd-6 atom on Pd₈/TiO₂(101) (−1.31 versus −1.46 eV). It is a well-known fact that, relative to the Fermi level, the higher the energy of the d-band states, the stronger the interaction with the adsorbate.⁴⁶ This is because when the d-band states is close to the Fermi level, the antibonding states can be pushed above the Fermi level.⁴⁶ Therefore, Pauli repulsion decreases and the binding strength between the adsorption site and adsorbate increases.⁴⁶ Thus, the adsorption strength of CO₂ on Pd₇Cu₁/TiO₂(101) surface is higher than that of Pd₈/TiO₂(101) surface. Our results are consistent with previous results that show the adsorption strength of CO₂ is controlled by the d-band center of the metal surfaces. For example, Long et al.¹⁶ calculated the d-band centers of Cu atom in Pd₇Cu₁ and Pd₁Cu₁ lattices, and found that the Cu d-band center of Pd₇Cu₁ lattice is significantly higher than that of Pd₁Cu₁ lattice (−1.161 versus −1.452 eV), and the adsorption energy of CO₂ on Pd₇Cu₁ lattice is stronger than that on Pd₁Cu₁ lattice (−0.463 versus −0.308 eV).

Fig. 9 Projected d-band states of Pd and Cu on the Pd₈/TiO₂(101) and Pd₇Cu₁/TiO₂(101) surfaces. The d-band center is marked by black arrow. The Fermi level is set to 0 eV.

Based on the above discussion, the following conclusions can be drawn: due to the strong positive polarization potential and the elevated Cu d-band center, the Pd₇Cu₁/TiO₂(101) has high activity, which promotes CO₂ adsorption and dissociation.

4. Conclusions

In order to understand the origin of the activity of Pd–Cu bimetallic nanoclusters supported on TiO₂(101), adsorption, activation, and dissociation of CO₂ on TiO₂-supported Pd and Pd–Cu clusters were investigated using DFT calculation. The results show that the activity of Pd₇Cu₁/TiO₂(101) can be greatly improved by introducing Cu atom. Because of the significant charge polarization between the Cu atom and the neighboring Pd atoms on the Pd₇Cu₁/TiO₂(101) surface, the positive polarization potential of the Cu atom is much higher than that of the Pd atom. This response can promote the adsorption of CO₂ by increasing the electrostatic attraction between CO₂ and the Cu site. Meanwhile, the elevated Cu d-band center on the surface of Pd₇Cu₁/TiO₂(101) would increase its surface catalytic activity. Therefore, compared to Pd₈/TiO₂(101) surface, the activity of Pd₇Cu₁/TiO₂(101) surface is higher, and the active center is the introduced Cu atom. This result is not only immensely helpful for the development of effective CO₂ photocatalysts, but also essential for understanding the basic mechanisms of bimetallic catalysis.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This research was supported by the Foundation (No. ZZ20200133) of State Key Laboratory of Biobased Material and Green Papermaking, Qilu University of Technology (Shandong Academy of Sciences) and the Natural Science Foundation of Shandong Province of China (Grant No. ZR2018LA013).

References

1. K. Li, B. Peng and T. Peng, ACS Catal., 2016, 6, 7485–7527.
2. V. P. Indrakanti, J. D. Kubicki and H. H. Schobert, Energy Environ. Sci., 2009, 2, 745–758.
3. A. Dhakshinamoorthy, S. Navalon, A. Corma and H. Garcia, Energy Environ. Sci., 2012, 5, 9217–9233.
4. W. Chu, Q. Zheng, O. V. Prezhdo and J. Zhao, J. Am. Chem. Soc., 2020, 142, 3214–3221.
5. S. Navalón, A. Dhakshinamoorthy, M. Álvaro and H. Garcia, ChemSusChem, 2013, 6, 562–577.
6. Y. Izumi, Coord. Chem. Rev., 2013, 257, 171–186.
7. S. N. Habisreutinger, L. Schmidt-Mende and J. K. Stolarczyk, Angew. Chem., Int. Ed., 2013, 52, 7372–7408.
8. S. C. Roy, O. K. Varghese, M. Paulose and C. A. Grimes, ACS Nano, 2010, 4, 1259–1278.
9. X. Chen and S. S. Mao, Chem. Rev., 2007, 107, 2891–2959.
10. A. Kubacka, M. Fernandez-Garcia and G. Colon, Chem. Rev., 2012, 112, 1555–1614.
11. Ş. Neaţu, J. A. Maciá-Agulló, P. Concepción and H. Garcia, J. Am. Chem. Soc., 2014, 136, 15969–15976.
12. D. Tan, J. Zhang, J. Shi, S. Li, B. Zhang, X. Tan, F. Zhang, L. Liu, D. Shao and B. Han, ACS Appl. Mater. Interfaces, 2018, 10, 24516–24522.
13. X. Li, J. Yu, M. Jaroniec and X. Chen, Chem. Rev., 2019, 119, 3962–4179.
14. V. Vaiano, D. Sannino and P. Ciambelli, Photochem. Photobiol. Sci., 2015, 14, 550–555.
15. Y. Zhang, X. Wang, P. Dong, Z. Huang, X. Nie and X. Zhang, RSC Adv., 2018, 8, 15991–15998.
16. R. Long, Y. Li, Y. Liu, S. Chen, X. Zheng, C. Gao, C. He, N. Chen, Z. Qi, L. Song, J. Jiang, J. Zhu and Y. Xiong, J. Am. Chem. Soc., 2017, 139, 4486–4492.
17. Y. Wang, Q. Lai, F. Zhang, X. Shen, M. Fan, Y. He and S. Ren, RSC Adv., 2014, 4, 44442–44451.
18. L. Liu, Z. Liu, H. Sun and X. Zhao, Appl. Surf. Sci., 2017, 399, 469–479.
19. X. Zhang, F. Han, B. Shi, S. Farsinezhad, G. P. Dechaine and K. Shankar, Angew. Chem., Int. Ed., 2012, 51, 12732–12735.
20. S. Lee, S. Jeong, W. D. Kim, S. Lee, K. Lee, W. K. Bae, J. H. Moon, S. Lee and D. C. Lee, Nanoscale, 2016, 8, 10043–10048.
21. Q. Kang, T. Wang, P. Li, L. Liu, K. Chang, M. Li and J. Ye, Angew. Chem., Int. Ed., 2015, 54, 841–845.
22. M. Tahir, B. Tahir and N. A. S. Amin, Appl. Catal., B, 2017, 204, 548–560.
23. Q. Chen, X. Chen, M. Fang, J. Chen, Y. Li, Z. Xie, Q. Kuang and L. Zheng, J. Mater. Chem. A, 2019, 7, 1334–1340.
24. J. Jiao, Y. Wei, Y. Zhao, Z. Zhao, A. Duan, J. Liu, Y. Pang, J. Li, G. Jiang and Y. Wang, Appl. Catal., B, 2017, 209, 228–239.
25. S. Bai, Q. Shao, P. Wang, Q. Dai, X. Wang and X. Huang, J. Am. Chem. Soc., 2017, 139, 6827–6830.
26. Y. Yang and D. Cheng, J. Phys. Chem. C, 2014, 118, 250–258.
27. M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias and J. D. Joannopoulos, Rev. Mod. Phys., 1992, 64, 1045–1097.
28. J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Perderson, D. J. Singh and C. Fiolhais, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 46, 6671–6687.
29. J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Perderson, D. J. Singh and C. Fiolhais, Phys. Rev. B: Condens. Matter Mater. Phys., 1993, 48, 4978.
30. J. P. Perdew, K. Burke and Y. Wang, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 16533–16539.
31. J. P. Perdew, K. Burke and Y. Wang, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 57, 14999.
32. M. Segall, P. Lindan, M. Probert, C. Pickard, P. Hasnip, S. Clark and M. Payne, J. Phys.: Condens. Matter, 2002, 14, 2717–2744.
33. D. Vanderbilt, Phys. Rev. B: Condens. Matter Mater. Phys., 1990, 41, 7892–7895.
34. H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1976, 13, 5188–5192.
35. T. A. Halgren and W. N. Lipscomb, Chem. Phys. Lett., 1977, 49, 225–232.
36. C.-T. Yang, B. C. Wood, V. R. Bhethanabotla and B. Joseph, J. Phys. Chem. C, 2014, 118, 26236–26248.
37. Y. Han, M. Zhang, W. Li and J. Zhang, Phys. Chem. Chem. Phys., 2012, 14, 8683–8692.
38. C. T. Yang, B. C. Wood, V. R. Bhethanabotla and B. Joseph, Phys. Chem. Chem. Phys., 2015, 17, 25379–25392.
39. D. C. Sorescu, W. A. Al-Saidi and K. D. Jordan, J. Chem. Phys., 2011, 135, 124701.
40. J. X. Liu, Y. Su, I. A. W. Filot and E. J. M. Hensen, J. Am. Chem. Soc., 2018, 140, 4580–4587.
41. Y. Li, C. Liu, Z. Pu, L. Bao, Y. Zhu and J. Ma, J. Phys. Chem. C, 2019, 123, 13739–13747.
42. H. J. Freund and M. W. Roberts, Surf. Sci. Rep., 1996, 25, 225–273.
43. L. Liu and P. Lv, New J. Chem., 2020, 44, 14662–14669.
44. C. Jia, W. Zhong, M. Deng and J. Jiang, J. Chem. Phys., 2018, 148, 124701.
45. H. Jiang, Z. Hou and Y. Luo, Angew. Chem., Int. Ed., 2017, 56, 15617–15621.
46. J. Xiao and T. Frauenheim, J. Phys. Chem. C, 2013, 117, 1804–1808.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1ra01724a

---