New insights into O and OH adsorption on the Pt–Co alloy surface: effects of Pt/Co ratios and structures

Peng Zhao , Xiaoqian Qin , Haibo Li , Konggang Qu and Rui Li *
Department of Chemistry, Liaocheng University, Liaocheng 252000, China. E-mail: lirui@lcu.edu.cn

Received 20th May 2020 , Accepted 28th August 2020

First published on 28th August 2020


In this study, the electronic structure and adsorption properties of O and OH for a series of Pt–Co alloys with different Pt/Co ratios (5[thin space (1/6-em)]:[thin space (1/6-em)]1, 2[thin space (1/6-em)]:[thin space (1/6-em)]1, 1[thin space (1/6-em)]:[thin space (1/6-em)]1, 1[thin space (1/6-em)]:[thin space (1/6-em)]2, and 1[thin space (1/6-em)]:[thin space (1/6-em)]5) were systematically studied using density functional theory calculations. Our computational results demonstrated that the introduced Co atoms have multiple effects on the surface electronic structure in different atomic layers of the alloy, leading to the discrepancies in the electronic structure between Pt-skin structures and non-Pt-skin structures. Moreover, the influence of the surface electronic structure on the adsorption of O and OH slightly differs. Indeed, the adsorption of O is more remarkably affected by the Pt/Co ratio than the OH adsorption and better follows the d-band center theory. Due to the difference of the alloy structure and the effect of different layer Co atoms, the adsorption of O and OH on the alloy configurations with the same Pt/Co ratio has different outcomes. Our results suggested that the oxygen reduction reaction (ORR) activity is related not only to the Pt/Co ratio of alloy surfaces but also to the specific surface structure. Our research can provide theoretical insights into the development of ORR catalysts.


1. Introduction

Fuel cells, especially polymer-electrolyte membrane fuel cells (PEMFCs), are one of the most promising clean energy technologies applied in automobiles and electronic devices, due to their advantages such as high power density and environmental friendliness.1,2 In particular, anodic and cathodic reaction mechanisms have been more and more explored to further optimize the performance and increase the applicability of PEMFCs.3 Indeed, the oxygen reduction reaction (ORR) takes place on the cathode of the PEMFC single cell. Compared with the hydrogen oxidation reaction on the anode, the kinetic process of the ORR is much slower and, thus, limits the performance of the fuel cell.4 Therefore, a catalyst must be used to reduce the overpotential loss during the oxygen reduction process. Among the current fuel cell catalysts, Pt exhibits the best catalytic effect;5,6 however, the lack of Pt resources and the high cost directly restrict the large-scale commercial application of Pt in fuel cells.7

A possible approach to overcome this limitation is to introduce nonprecious metals into the pure Pt catalyst to form alloys;8 this strategy can significantly reduce the Pt content in the catalyst while maintaining the activity and stability of the catalyst. Pt-Based alloy catalysts, Pt–M (M = Fe, Co, Ni, V, and Cr), have been reported to be effective for the ORR.9–15 Among the Pt–M alloys, Pt–Co alloys have attracted increased attention due to their excellent performance. Previous studies demonstrated that the ORR activity of Pt-based alloys is closely related to their composition and surface structure.16 However, the mechanism of ORR strengthening by alloy structure is still unclear, and the relationship between the alloy composition and ORR activity is still controversial. Yue et al.17 studied the ORR activity using Pt–Co alloys in alkaline solutions, and their result indicated that the catalyst with a Pt/Co atomic ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1 had the highest ORR activity. Zhao et al.18 synthesized a carbon-supported Pt–Co alloy series with different compositions (Pt/Co atomic ratios = 81[thin space (1/6-em)]:[thin space (1/6-em)]19, 76[thin space (1/6-em)]:[thin space (1/6-em)]24, 59[thin space (1/6-em)]:[thin space (1/6-em)]41, 48[thin space (1/6-em)]:[thin space (1/6-em)]52, 40[thin space (1/6-em)]:[thin space (1/6-em)]60, and 26[thin space (1/6-em)]:[thin space (1/6-em)]74) in alkaline solution and found a strong correlation between the ORR activity and the alloy composition of the catalysts. They reported that the catalyst with a Pt/Co atomic ratio of 76[thin space (1/6-em)]:[thin space (1/6-em)]24 exhibited the best ORR performance. Polagani et al.19 investigated the catalytic activity of Pt–Co alloys for oxygen reduction in PEMFCs. They found that the fuel cell showed relatively inferior performance with an increase in the Co content, while the Pt83–Co17 composition performed better than Pt75–Co25 and Pt50–Co50.

In order to design more efficient catalysts, quantum chemical methods based on density functional theory (DFT) have been widely applied.20 In particular, the d-band model developed by Nørskov et al.21,22 has been used successfully to explain the trends in the reactivity of transition-metal and alloy surfaces. According to the d-band center theory, the adsorption capacity of molecules on metal surfaces is closely related to the d-band center of metals: the closer the d-band center of the outermost metal to the Fermi energy level is, the more stable the molecule adsorption is. Nørskov et al.23 also listed the characteristics that an ideal ORR catalyst should possess; a good catalyst should have a moderate adsorption capacity for the intermediate products of the reaction. However, due to the complexity of the ORR reaction and the diversity of alloy structures, the adsorption mechanism of ORR intermediates on alloy surfaces and the relationship between ORR reactivity and alloy composition are still unclear and require further research.

The surfaces of the nanoparticles usually consist of low index crystal facets such as the (111) facet which is thermodynamically the most stable surface,24 and thus a particularly important model for understanding the adsorption of various adsorbates. In this work, we have proposed a series of Pt–Co alloy surface models with different compositions and structures by replacing partial atoms of the Pt(111) surface with Co atoms. The DFT method has been applied to explore the relationship between the surface electronic structures, the molecular geometry, and the composition of the alloys. In addition, the adsorption of O and OH radicals on the surface of Pt–Co alloy was studied, and the effect of the electronic structure of the alloy on molecule adsorption and the relationship between the alloy composition and ORR activity were further assessed.

2. Computational details

In order to design the Pt–Co alloy surface models, we first established a pure Pt(111) surface, using a 3-layer slab model and a 2 × 2 supercell. In this model, the vacuum layer is set to 20 Å along the z direction to avoid interactions between the periodically repeated structures. Then, a series of Pt–Co alloy models were obtained by replacing partial Pt atoms with Co atoms. Here, we replace the Pt atoms in Pt(111) with 2, 4, 6, 8, and 10 Co atoms to obtain the alloy models with Pt/Co ratios of 5[thin space (1/6-em)]:[thin space (1/6-em)]1, 2[thin space (1/6-em)]:[thin space (1/6-em)]1, 1[thin space (1/6-em)]:[thin space (1/6-em)]1, 1[thin space (1/6-em)]:[thin space (1/6-em)]2, and 1[thin space (1/6-em)]:[thin space (1/6-em)]5, respectively.

For a fixed Pt/Co ratio, there are a number of possible alloy configurations. In this study, we have selected some structures for research based on pre-established sampling rules as follows. In our model, there are three atomic layers. For the Co atom replacing Pt, we only consider its distribution in different layers, not its specific position in the layer. We use PtmCon(x,y,z) to express a specific alloy structure, where m and n are the proportion of Pt and Co, and x, y, and z are the number of Co atoms in the first, second, and third layer, respectively. For example, in the Pt5Co1 structure, the Co atoms can be distributed in two ways. First, two Co atoms can be present in the same layer: in this case, there are three (C(3,1) = 3) kinds of alloy structures, namely Pt5Co1(2,0,0), Pt5Co1(0,2,0), and Pt5Co1(0,0,2). Second, two Co atoms can also be located in different layers: in this other case, there are also three (C(3,2) = 3) different alloy structures, namely Pt5Co1(1,1,0), Pt5Co1(1,0,1), and Pt5Co1(0,1,1). Therefore, six structures of Pt5Co1 can be generated (Fig. 1). Based on this rule, Pt2Co1, Pt1Co1, Pt1Co2, and Pt1Co5 have 15, 18, 15, and 6 structures, respectively. The optimized structures of all these alloy surfaces are shown in the ESI.


image file: d0cp02746d-f1.tif
Fig. 1 The alloy configurations with a Pt/Co atomic ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1. The colors of Pt and Co atoms are blue and pink, respectively.

The spin-polarized first-principle DFT calculations were performed using the Vienna Ab initio simulation package (VASP)25 with the projected augmented wave method.26 The exchange–correlation function is described by the parameterization scheme of Perdew–Burke–Ernzerhof (PBE)27 of the generalized gradient approximation. The PBE functional is placed as one of the most accurate functionals in describing the metal systems as reported in the previous study.28 A cutoff energy of 500 eV for the plane-wave basis set is used. The Brillouin zone was sampled by special k-points using the Monkhorst Pack scheme (4 × 4 × 1 for the structural optimization and 8 × 8 × 1 for the electronic structure calculations). The bottom layer was fixed during the structural optimization. The Grimme DFT-D329 scheme of dispersion correction with zero damping is adopted to account for the van der Waals interactions. The dipole correction was implemented only parallel to the Z direction to ensure vacuum level convergence. In this study, the adsorption energy Eads is defined as follows:

Eads = EtotalEslabEadsorbate
where Etotal is the total energy of the adsorbate and alloy surface, Eslab is the total energy of the alloy surface, and Eadsorbate is the total energy of the free adsorbate. Based on this equation, a negative value of Eads corresponds to high adsorption energy.

The d-band center (εd) in this work is defined as the average energy of electronic d-states projected onto surface atoms and calculated as

image file: d0cp02746d-t1.tif
where ρd(E) is the projected d-band density of states. The work function is calculated as W = EvEf, where Ev is the vacuum level and Ef is the Fermi level.

3. Results and discussion

3.1 Molecular geometries and electronic structures of Pt–Co alloys

Based on the composition of the first layer atoms, our model can be divided into two structural types: Pt-skin structure and non-Pt-skin structure. In the Pt-skin structure, the first layer atoms are all Pt atoms with the Co atoms only located in the second and third layers. Previous studies30 demonstrated that this structure could be easily formed in the acid medium when the outermost Co atoms were leached out. In the non-Pt-skin structure, the Co atom can be situated in the first layer, as reported in a previous study.31 Our computational results clearly indicate that, between the two structures, there are obvious differences in the molecular geometry and electronic structure. The detailed discussion is as follows.

Due to the short atomic radius of Co, the structure of the Pt–Co(111) surface changes when it is combined with Pt to form Pt–Co alloys compared with pure Pt(111). In addition to a smaller lattice parameter than that of Pt(111), in Pt–Co(111), the distance between the atomic layers is also significantly reduced. Fig. 2a displays the average interlayer distance of different components of Pt–Co alloys. In order to show the overall trend and individual differences, we combined the box plot with the scatter plot. The blue and orange data points represent the Pt-skin and non-Pt-skin structure, respectively, and are displayed on different sides of the box diagram to distinguish them. Besides, we add a horizontal jitter to the data points to avoid overlapping. The same settings are also applied to some of the following figures. It can be seen from Fig. 2a that with the increase of the number of Co atoms, the average interlayer distance decreases gradually, independent of the structure. This can have an impact on the electronic structure of the surface atoms; due to the decrease of the distance between the atoms, the electronic interaction is strengthened. In addition, due to the introduction of a heteroatom, the surface of the alloy is no longer flat. We calculated the standard deviation of the coordinates of the outermost atoms in the Z direction with the number of Co atoms on the surface (Fig. 2b). Our results show that the standard deviation of the Z coordinate is smaller when the first layer is composed of single species atoms than when Pt and Co atoms are mixed.


image file: d0cp02746d-f2.tif
Fig. 2 (a) The average interlayer distance of each structure with different Pt/Co ratios. (b) The standard deviation of the Z coordinate of the outermost atoms vs. the number of Co atoms on the surface.

To further understand the Pt–Co interaction and charge distribution in the alloy structure, the Bader charge32 population was calculated, and it shows that the Co atoms acted as an electron donor and transferred electrons to the Pt atoms (Table S1 in ESI). In addition, we calculated the d-band center of the outermost atoms for all alloy structures. We also calculated the d-band center of pure Pt and Co surface and found that the d-band center of Co (−1.33 eV) is much higher than that of Pt (−2.11 eV). This result is consistent with the previous studies.33,34 Moreover, the d-band center of the Pt–Co alloy surface calculated by Nørskov et al. is −1.72 eV,33 which is very similar to our result for the Pt5Co1(2,0,0) surface (−1.70 eV). The distribution of d-band centers calculated by us for different alloy compositions is shown in Fig. 3a. In general, the d-band center gradually moves towards the Fermi energy level with the increase in the content of Co atoms in the alloy; however, the data points are highly dispersed, and the span is relatively large. Especially in the Pt2Co1 and Pt1Co1 structures, the d-band center is distributed in the interval of ∼−1.3 to ∼−2.6 eV. Moreover, it can be seen that different alloy structures have different trends. For the Pt-skin structure, with the increase of the Co ratio, the d-band center gradually decreases, which means that it moves away from the Fermi energy level. This is because of the interaction between the first and second atomic layers in the Pt-skin structure. When the Co atoms are present in the second layer, electrons are easily transferred from Co in the second layer to the empty d band of the Pt atoms in the first layer, making the entire d-band of the outermost layer move closer to the Fermi energy level. Moreover, as the number of Co atoms in the second layer increases, the d-band center of the outermost layer decreases even more because more electrons transfer to the outermost layer. For example, the d-band centers of Pt5Co1(0,2,0) and Pt2Co1(0,4,0) are at −2.30 and −2.62 eV, respectively. Through the Bader charge analysis, we found 1.24 and 1.98 electron transferred from Co atoms to Pt atoms in Pt5Co1(0,2,0) and Pt2Co1(0,4,0), respectively. This effect can also be seen from the electron density difference. Fig. 4 shows that as the number of Co atoms in the second layer increases, the electron transfer between the two layers becomes more obvious. However, unlike the Pt-skin structure, for the non-Pt-skin structure, as the proportion of Co increases, the d-band center of the outermost layer tends to increase, moving towards the Fermi energy level. This is because the Co atom is present in the first layer, making the d-band center energy level of the whole atomic layer increase. The higher the number of the Co atoms in the first layer is, the closer to the Fermi level of the d-band center is. We calculated that the d-band centers of the outermost layer in Pt5Co1(2,0,0) and Pt2Co1(4,0,0) surfaces are −1.70 and −1.31 eV, respectively. Our results indicate that for the Co atoms in the Pt–Co alloys, there are two different effects on the d band center of the outermost layer. When the Co atoms are situated on the outermost layer, they cause the d-band center to move towards the Fermi energy level. When the Co atoms are located on the subsurface, they cause the d-band center to move away from the Fermi energy level. For non-Pt-skin structures, these two effects compete with each other; therefore, the overall result of the d-band center is more dispersed than in the Pt-skin structures. With the increase of Co content, no obvious changing trend can be observed.


image file: d0cp02746d-f3.tif
Fig. 3 (a) The d-band center of the outermost layer and (b) the work function of all the alloy configurations with different Pt/Co ratios.

image file: d0cp02746d-f4.tif
Fig. 4 The electron density difference between the first layer and the other layers of two alloy configurations. Cyan and yellow isosurfaces represent the charge depletion and accumulation, respectively. The isosurface value is 6 × 10−3 electrons per Å3.

Furthermore, we calculated the work function of all structures. Our previous research35 showed that the work function determines the degree of an electron's ability to transfer between interfaces; a smaller work function indicates that it is easier for an electron to leave the surface. Shen et al.36 proposed that, in addition to the d-band center, the work function can also be used as a catalyst descriptor for the adsorption of O and OH radicals onto the transition metals. In this study, the work function distribution of all structures is shown in Fig. 3b. It is found that, in general, the work function decreases with the increase of the Co ratio in both Pt-skin structures and non-Pt-skin structures, indicating that the Co atom plays a role in reducing the work function in both the first layer and other layers. Moreover, there are also obvious differences between different structures. The work function of Pt-skin structures is generally higher, which is because Pt has a higher electronegativity than Co. The work function of pure Pt(111) calculated by us is 5.75 eV, which is higher than the calculated value of the pure Co(111) (4.81 eV), and this conclusion is consistent with previous studies.28 However, there are exceptions. For example, the work function of Pt2Co1(1,0,3) (5.57 eV) is slightly higher than Pt2Co1(0,4,0) (5.51 eV) in the Pt-skin structure. This is because when Co is located in the second layer, the electron transfer to the first layer makes the surface rich in electrons, which reduces its work function. This effect decreases with the distance of Co from the first layer Pt. When Co is in the third layer, its work function is higher than that in the second layer; the work function of Pt2Co1(0,0,4) is 5.72 eV. This suggests that the work function of Pt–Co alloys is related to the positions and the number of Co atoms. In general, the farther the Co atoms are from the surface, the lower the number of Co atoms, and the larger the work function.

3.2 Adsorption of O and OH on Pt–Co alloys

Next, we investigated the adsorption properties of the O and OH radicals on different alloy structures. Both of them are intermediates in the ORR process where O is the product of oxygen dissociation and OH is the product of proton binding of the O atom. Previous studies37 reported that a good catalyst surface should have appropriate adsorption energy. For ORR catalysts, if the bound O is too weak, the dissociation process of O2 may be limited, while, if the bound OH is too strong, the active site may be occupied by OH, resulting in the poisoning of the catalyst.

First, we investigated the adsorption of O on Pt–Co alloy. Previous studies38,39 suggest that the preferential adsorption site of O on the Pt-based alloy (111) surface is the face-centered cubic (fcc) site. For the Pt–Co alloy, we only considered the adsorption of O on the fcc site. However, for a certain alloy surface, if the relative position and symmetry of the adsorbed O and alloy atoms are taken into account, the adsorption sites on the surface have a different environment, that is, the atoms nearest to the fcc positions may be rich in Pt or Co. Therefore, we considered the number of Co atoms near the fcc site in the first layer to be distributed from 0 to 3. Based on this, we took into account multiple fcc adsorption sites for each alloy structure to examine the effect of different adsorption environments.

For all the adsorption configurations, we analyzed the number of Co atoms in the nearest three atoms of the fcc sites as a function of the adsorption energy of O, as shown in Fig. 5a. We found that the adsorption energy of O increases with the number of Co atoms in the nearest three atoms of the fcc sites. Therefore, the advantageous adsorption site for O atoms is the Co-rich fcc site. Furthermore, the relationship between the content of Co atoms in each layer and the adsorption energy of O was also analyzed (Fig. 5b–d). We found that when Co is present in the first layer, the effect on the adsorption of O is the greatest followed by the second layer, while in the third layer, basically no effect on the adsorption energy was observed. For the first atomic layer, it is obvious that the adsorption energy for O increases with the increase of the Co content. For the second atomic layer, in general, with the increase of the number of Co atoms, the adsorption energy of O undergoes an obvious decreasing trend for the Pt-skin structure. This should be attributed to the role of the d-band center: indeed, the electron transfer from the Co of the second layer to the Pt of the first layer results in the d-band center of the first layer downwards. According to the d-band center theory, the downward shift of a d-band center leads to the increase of the number of antibonding electrons formed by O, thus weakening the adsorption.


image file: d0cp02746d-f5.tif
Fig. 5 The adsorption energy of O vs. the number of the Co atoms in (a) the nearest neighbor atoms, (b) the first atomic layer, (c) the second atomic layer, and (d) the third atomic layer.

However, this trend is less remarkable for the non-Pt-skin structure. Moreover, we analyzed the adsorption energy of O as a function of the Pt/Co atomic ratio (Fig. 6a). In general, with the increase of Co content in the alloy, the Pt-skin structure and non-Pt-skin structure exhibit different trends. For the Pt-skin structure, the adsorption energy of O tends to decrease with the increase in the Pt/Co ratio, which is mainly due to the contribution of the second layer of Co atoms. For the non-Pt-skin structure, we found that the adsorption energy of O increases with the increase of Co content. Moreover, for the same Pt/Co atomic ratio, the distribution of adsorption energy between different configurations is very wide, not as localized as in the Pt-skin structure. This is because the effect of Co in different layers on the adsorption of O is different. When Co is located in the first layer, the resulting adsorption of O is strengthened, but, when Co is present in the second layer, the adsorption of O is weakened. Therefore, we can state that for distinct configurations, with the increase of Co content, the effect of O adsorption is different. Nevertheless, both Pt-skin structures and non-Pt-skin structures have a good linear relationship between the average adsorption energy of O on each alloy configuration and the d-band center. Fig. 6b illustrates that the d-band center theory is applicable to both Pt-skin and non-Pt-skin structures. In addition, our results show that two different structures show a different relationship between the adsorption of O and the work function, with only the Pt-skin structure showing a good linear relationship (Fig. 6c).


image file: d0cp02746d-f6.tif
Fig. 6 (a) Comparison of the distributions of the adsorption energy of O in different Pt/Co ratios. The correlation between the average adsorption energy of O on each alloy structure and (b) the d-band center and (c) work functions.

Next, we investigated the adsorption properties of OH on the alloy surface. According to the previous DFT calculations,40,41 the optimal adsorption site for OH is the top site. Here, for different alloy surfaces, we have only studied the adsorption of OH at top sites. For the Pt–Co alloys, there could be two kinds of surface atoms, namely Pt and Co. In view of this, we investigated both possible adsorption sites for each structure. We found that the adsorption energy of OH on Co is greater than that on Pt for all structures; in other words, OH more easily adsorb on Co atoms on the surface than on Pt (Fig. 7a). By analyzing the change of the adsorption energy of OH with the Pt/Co atomic ratio (Fig. 7b), we found that the increase of Co atomic content does not significantly impact the adsorption energy except for the Pt-skin structure. In the Pt-skin structure, the adsorption energy of OH slightly decreases with the increase of Co contents, which should be influenced by the Co in the second layer, just like the O adsorption. The impact of the number of Co atoms in the first atomic layer and the second atomic layer on the adsorption energy of OH has also been assessed (Fig. 7c and d). In general, the effect of Co content in each layer on the OH adsorption is not as significant as that on O adsorption. This is because the adsorption of OH at the top position is less affected by the surrounding atoms. We can see from the electron density difference of O and OH adsorptions (Fig. 8) that the adsorption of O at the fcc site occurs with all three adjacent atoms, while the adsorption of OH occurs only with the top atoms. Therefore, the adsorption energy of OH on the alloy strongly depends on its adsorption position, while it is almost independent of the proportion of Pt/Co in the alloy's outermost layer. In addition, the relationship between the average adsorption energy of OH on different structures and their d-band centers is also a weak linear relationship (Fig. 9a). This may be due to the long distance of adsorption of OH from the surface. The average distance from the adsorbed O to the surface is 1.32 Å, while that for OH is 2.04 Å; therefore, the change of the surface electronic structure has a reduced effect on the OH adsorption.


image file: d0cp02746d-f7.tif
Fig. 7 (a) Comparison of the distributions of the adsorption energy of OH vs. (a) different top sites, (b) different Pt/Co ratios, (c) the number of the Co atoms in the first layer, and (d) the number of Co atoms in the second layer.

image file: d0cp02746d-f8.tif
Fig. 8 The electron density difference between Pt5Co1(1,1,0) and (a) adsorbed O and (b) adsorbed OH. The isosurface value is 8 × 10−3 electrons per Å3.

image file: d0cp02746d-f9.tif
Fig. 9 The correlation between the average adsorption energy of OH on each alloy structure and (a) the d-band center, and (b) the average adsorption energy of O.

In our work, to facilitate the establishment of models and calculations, the alloy surface was modeled using a three-layer model. In our model, each atomic layer had four atoms, and the surface coverage of O and OH adsorption is 1/4 ML. We also investigated the effect of larger alloy surface models on O and OH adsorption. Here, we have selected six structures of Pt5Co1, and expanded their initial unit cells by 2 × 2 in x and y directions, so that the coverage of O and OH is 1/16 ML. The result shows that when the coverage changed from 1/4 to 1/16 ML, the adsorption energy of O and OH did not change (Table S2 in the ESI). In addition, we also tested the influence of slab thickness on adsorption. Here, we have built a serial of four layer Pt–Co alloy structures (Fig. S5 in the ESI) and investigated the effect of Co in the fourth layer on the surface adsorption by changing the Pt/Co ratio. The result shows that the change of the Pt/Co ratio in the fourth layer did not affect the adsorption energy of O and OH (Table S3 in ESI). This suggests that our current model is suitable for studying the effect of Pt/Co ratio on surface adsorption.

Finally, we compared the adsorption properties of O and OH on the surface of the same alloy structures. As shown in Fig. 9b, it was found that whether it is a Pt-skin structure or a non-Pt-skin structure, the adsorption of O and OH has a weak linear relationship. In general, for most structures, when a structure can enhance the O adsorption, it is also favorable for OH adsorption. However, for these alloy surfaces, even if their Pt/Co atomic ratios are the same, their adsorption energies show a relatively dispersed arrangement. Besides, the Pt-skin structure exhibits weaker adsorption of O and OH than the non-Pt-skin structure. Therefore, in addition to the Pt/Co atomic ratio, the adsorption energy of adsorbates on Pt–Co alloys is also related to their specific structure. This may explain why there are different experimental results in the Pt/Co atomic ratio that produces the optimal ORR activity.

4. Conclusions

We used the DFT method to study the electronic structure of Pt–Co alloy surfaces with different compositions and the adsorption of O and OH on different surfaces. Our results show that the effects of the Co content on the adsorption of O and OH on Pt-skin structure and non-Pt-skin structure alloys are significantly different. This is mainly because the effect of Co atoms on the surface electronic structure is different in different atomic layers. When Co is located in the first layer, it will strengthen the adsorption, whereas, when Co is situated in the second layer, it will weaken the adsorption, in agreement with the d-band center theory. In addition, due to different adsorption positions and distances, the adsorption properties of O and OH are also significantly different. The linear relationship between the adsorption energy of O and the d-band center is better, which indicates that the O adsorption mechanism better follows the theory of the d-band center. Furthermore, even with the same compositions of Pt–Co alloys, due to the difference in the specific structure, the adsorption effect of ORR intermediates is different, which leads to different ORR activities. Therefore, our study provides new insights into the understanding of the effects of structure and composition on the catalytic activity of the alloy surface.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Natural Science Foundation of Shandong Province (Grant No. ZR2017QB010, ZR2019MB064), and the Development Project of Youth Innovation Team in Shandong Colleges and Universities (Grant No. 2019KJC031).

Notes and references

  1. S. Mekhilef, R. Saidur and A. Safari, Renewable Sustainable Energy Rev., 2012, 16, 981–989 CrossRef CAS .
  2. O. Z. Sharaf and M. F. Orhan, Renewable Sustainable Energy Rev., 2014, 32, 810–853 CrossRef CAS .
  3. E. H. Majlan, D. Rohendi, W. R. W. Daud, T. Husaini and M. A. Haque, Renewable Sustainable Energy Rev., 2018, 89, 117–134 CrossRef CAS .
  4. J. Stacy, Y. N. Regmi, B. Leonard and M. Fan, Renewable Sustainable Energy Rev., 2017, 69, 401–414 CrossRef CAS .
  5. N. Jung, D. Y. Chung, J. Ryu, S. J. Yoo and Y.-E. Sung, Nano Today, 2014, 9, 433–456 CrossRef CAS .
  6. C. Zhang, X. Shen, Y. Pan and Z. Peng, Front. Energy, 2017, 11, 268–285 CrossRef .
  7. R. Othman, A. L. Dicks and Z. Zhu, Int. J. Hydrogen Energy, 2012, 37, 357–372 CrossRef CAS .
  8. Z. Liu, L. Ma, J. Zhang, K. Hongsirikarn and J. G. Goodwin, Catal. Rev., 2013, 55, 255–288 CrossRef CAS .
  9. M. Ammam and E. B. Easton, J. Power Sources, 2013, 236, 311–320 CrossRef CAS .
  10. Y.-H. Cho, T.-Y. Jeon, S. J. Yoo, K.-S. Lee, M. Ahn, O.-H. Kim, Y.-H. Cho, J. W. Lim, N. Jung, W.-S. Yoon, H. Choe and Y.-E. Sung, Electrochim. Acta, 2012, 59, 264–269 CrossRef CAS .
  11. K. A. Kuttiyiel, Y. Choi, S.-M. Hwang, G.-G. Park, T.-H. Yang, D. Su, K. Sasaki, P. Liu and R. R. Adzic, Nano Energy, 2015, 13, 442–449 CrossRef CAS .
  12. U. A. Paulus, A. Wokaun, G. G. Scherer, T. J. Schmidt, V. Stamenkovic, V. Radmilovic, N. M. Markovic and P. N. Ross, J. Phys. Chem. B, 2002, 106, 4181–4191 CrossRef CAS .
  13. B. P. Vinayan, R. I. Jafri, R. Nagar, N. Rajalakshmi, K. Sethupathi and S. Ramaprabhu, Int. J. Hydrogen Energy, 2012, 37, 412–421 CrossRef CAS .
  14. H. Yano, M. Kataoka, H. Yamashita, H. Uchida and M. Watanabe, Langmuir, 2007, 23, 6438–6445 CrossRef CAS .
  15. K. Zhang, Q. Yue, G. Chen, Y. Zhai, L. Wang, H. Wang, J. Zhao, J. Liu, J. Jia and H. Li, J. Phys. Chem. C, 2011, 115, 379–389 CrossRef CAS .
  16. M. Liu, Z. Zhao, X. Duan and Y. Huang, Adv. Mater., 2019, 31, 1802234 CrossRef .
  17. Q. Yue, K. Zhang, X. Chen, L. Wang, J. Zhao, J. Liu and J. Jia, Chem. Commun., 2010, 46, 3369–3371 RSC .
  18. Y. Zhao, J. Liu, Y. Zhao and F. Wang, Phys. Chem. Chem. Phys., 2014, 16, 19298–19306 RSC .
  19. R. K. Polagani, P. L. Suryawanshi, S. P. Gumfekar, S. H. Sonawane and M. Ashokkumar, Sustainable Energy Fuels, 2018, 2, 1491–1499 RSC .
  20. C. Fu, C. Liu, T. Li, X. Zhang, F. Wang, J. Yang, Y. Jiang, P. Cui and H. Li, Comput. Mater. Sci., 2019, 170, 109202 CrossRef CAS .
  21. B. Hammer, Y. Morikawa and J. K. Nørskov, Phys. Rev. Lett., 1996, 76, 2141–2144 CrossRef CAS .
  22. J. Greeley, J. K. Nørskov and M. Mavrikakis, Annu. Rev. Phys. Chem., 2002, 53, 319–348 CrossRef CAS .
  23. J. K. Nørskov, J. Rossmeisl, A. Logadottir, L. Lindqvist, J. R. Kitchin, T. Bligaard and H. Jónsson, J. Phys. Chem. B, 2004, 108, 17886–17892 CrossRef .
  24. Y. Yang, M. Luo, W. Zhang, Y. Sun, X. Chen and S. Guo, Chem, 2018, 4, 2054–2083 CAS .
  25. G. Kresse and J. Furthmüller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186 CrossRef CAS .
  26. G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59, 1758–1775 CrossRef CAS .
  27. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868 CrossRef CAS .
  28. L. Vega, J. Ruvireta, F. Vines and F. Illas, J. Chem. Theory Comput., 2018, 14, 395–403 CrossRef CAS .
  29. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104 CrossRef .
  30. V. R. Stamenkovic, B. S. Mun, K. J. J. Mayrhofer, P. N. Ross and N. M. Markovic, J. Am. Chem. Soc., 2006, 128, 8813–8819 CrossRef CAS .
  31. V. Stamenkovic, T. J. Schmidt, N. M. Markovic and P. N. Ross, J. Phys. Chem. B, 2002, 106, 11970–11979 CrossRef CAS .
  32. G. Henkelman, A. Arnaldsson and H. Jónsson, Comput. Mater. Sci., 2006, 36, 354–360 CrossRef .
  33. B. Hammer and J. K. Nørskov, Advances in Catalysis, Academic Press, 2000, vol. 45, pp. 71–129 Search PubMed .
  34. L. Vega, B. Martinez, F. Vines and F. Illas, Phys. Chem. Chem. Phys., 2018, 20, 20548–20554 RSC .
  35. R. Li, W. Sun, C. Zhan, P. R. C. Kent and D.-e. Jiang, Phys. Rev. B, 2019, 99, 085429 CrossRef CAS .
  36. X. Shen, Y. Pan, B. Liu, J. Yang, J. Zeng and Z. Peng, Phys. Chem. Chem. Phys., 2017, 19, 12628–12632 RSC .
  37. V. Stamenkovic, B. S. Mun, K. J. J. Mayrhofer, P. N. Ross, N. M. Markovic, J. Rossmeisl, J. Greeley and J. K. Nørskov, Angew. Chem., Int. Ed., 2006, 45, 2897–2901 CrossRef CAS .
  38. Z. Duan and G. Wang, Phys. Chem. Chem. Phys., 2011, 13, 20178–20187 RSC .
  39. K. Li, Y. Li, Y. Wang, F. He, M. Jiao, H. Tang and Z. Wu, J. Mater. Chem. A, 2015, 3, 11444–11452 RSC .
  40. D. C. Ford, Y. Xu and M. Mavrikakis, Surf. Sci., 2005, 587, 159–174 CrossRef CAS .
  41. R. Jinnouchi, K. Kodama and Y. Morimoto, J. Electroanal. Chem., 2014, 716, 31–44 CrossRef CAS .

Footnote

Electronic supplementary information (ESI) available. See DOI: 10.1039/d0cp02746d

This journal is © the Owner Societies 2020