Author : Junseok Kim
Project description
The aim of this project is to predict the most preferred binding site of atomic O on Pt (111) surface and investigate the effect of surface coverage using density functional theory (DFT). DFT is a powerful tool to calculate the adsorption energies of different atoms or molecules on a metal surface [1]. The adsorption energy is the energy required to separate an atom or molecule far away from a surface. There are 4 high symmetric binding sites on fcc Pt (111) surface for O atom; atop, bridge, fcc hollow and hcp hollow (Figure 1). By comparing the adsorption energy for all the binding sites of O atom on Pt (111) surface, we determined energetically the most favorable site that O atom may bind. In addition, since we are using periodic boundary condition in DFT calculation, we must consider the interference by atoms or molecules of the neighboring cell on the adsorption energy. Thus, this interference was also investigated by changing the surface coverage.
Figure 1. All possible binding sites of atomic O on (2 x 2) Pt (111) supercell; atop, bridge, fcc hollow and hcp hollow (Blue atom – Pt, Red atom – O).
Method
All DFT energy calculations were performed using material studio with CASTEP calculation package that is based on a plane wave basis set [2]. The generalized gradient approximation (GGA) – Perdew Burke Ernzerhof (PBE) was used as an exchange-correlation functional [3]. On-the-fly generated (OTFG) ultrasoft pseudopotential was also employed to describe the interactions of ionic core and valance electrons. To generate the pseudopotential, 32 valance electrons (4f14 5s2 5p6 5d9 6s1 of the electron configuration) with 2.403 Bohr cutoff radius was used for Pt, while 6 valence electrons (2s2 2p4 of the electron configuration) with 1.1 Bohr cutoff radius was used for O [4].
We used 3.972 Å as the optimized lattice constant, which was obtained from the geometry optimization of bulk fcc Pt using CASTEP, with a 12 x 12 x 12 k-point grid and 450 eV of cutoff energy. The k-point sampling and cutoff energy for bulk Pt was already discussed on the previous post. Thus, we tested the convergence of the total energy for bare Pt (111) surface slab with respect to k-point sampling and cut off energy in this project. For the convergence test, we used a slab of 4-layer unrelaxed (1 x 1) Pt (111) surface with 12 Å vacuum spacing. The tolerance for the energy convergence was set as 0.01 eV/atom.
To calculate the adsorption energy, we used the following equation [5],
Eads = EO-Pt surf – 0.5 EO2 – EPt surf
Where Eads is the adsorption energy, EO-Pt surf is the total energy of O atom absorbed on Pt(111) surface, EO2 is the total energy of O2 molecule in the gas phase, and EPt surf is the total energy of bare Pt (111) surface.
For the total energy calculation of O2 molecule in the gas phase, a cubic supercell with side length of 10 Å, which was arbitrarily chosen to model the gas phase state, and 1 x 1 x 1 k-point grid were used. For Pt (111) surface total energy with or without O atom, we employed the same slab used for the convergence test, but two layers from top surface were relaxed. Since our slab is asymmetric, two top layers relaxed and two bottom layers fixed, self-consistent dipole correction was included for all surface energy calculations. In particular, we kept 2.11 Å of Pt-O bond length that is experimentally observed for atop site during the energy calculations [6]. O atom was placed on each site of Pt (111) maintaining the bond length and we calculated single point energy for the binding state.
1 monolayer (ML) surface coverage can be defined as 1 adsorbate per 1 surface atom. For the investigation of the surface coverage effect on the adsorption energy, (1 x 1) and (2 x 2) Pt (111) surfaces with 1 O atom were used for 1 ML and 0.25 ML, respectively (Figure 2).
Figure 2. 1 ML (Left) and 0.25 ML (Right) coverage of O atom binding on hcp hollow site of Pt (111) surface (Blue atom – Pt, Red atom – O).
Result and discussion
Convergence test for cutoff energy and k-point sampling
Figure 3. The total energy plot as a function of a) cutoff energy and b) k-point grid with using bare (1 x 1) Pt (111) supercell.
Figure 3 shows the total energy convergence with varying cutoff energy and M values of M x M x 1 k-point. Since the used supercell has a larger value for z-direction (c coordinate), we set 1 for z-direction k space. From the convergence test, we obtained 450 eV of cutoff energy and 10 x 10 x 1 k-point grid for the total energy calculations. In particular, 5 x 5 x 1 k-point grid was used for (2 x 2) Pt (111) supercell.
Adsorption energy and the coverage effect
Table 1. Adsorption energy comparison for all binding sites with 1 and 0.25 ML surface coverage.
For both 1 and 0.25 ML coverage, it is confirmed that the adsorption energy is the lowest at fcc hollow site of Pt (111) as shown in Table 1, which means that fcc hollow site is the most favorable binding place for atomic O. Hcp hollow site is also energetically favorable compared to two other sites, atop and bridge. On the other hand, it is expected that O atom may rarely bind to atop site of Pt (111) since the adsorption energy is the highest at atop site. We also confirmed that this preference is consistent with those from other DFT calculation results [7, 8] and an experimental study [9]. Those studies also indicated fcc hollow site as the most preferred binding site of O atom on Pt (111) surface.
As the coverage decrease from 1 ML to 0.25 ML, the adsorption energies of 0.25 ML become much lower than those of 1 ML except atop site (Table 1). Regarding to the binding state of Pt and O atoms at 1 ML and 0.25 ML, the result is reasonable. At 1 ML coverage, only one Pt atom on Pt surface is quantitatively affordable to bind one O atom for all binding sites, since one O atom per one Pt atom is placed as shown in Figure 2. However, at 0.25 ML coverage, 2 and 3 Pt atoms on Pt surface are available to bind one O atom for bridge and hollow sites, respectively, while only one Pt atom is still available for atop site. Even if 2 and 3 Pt atoms could participate in binding O atom for bridge and hollow site at 1 ML coverage, the binding would be much weaker than that of 0.25 ML coverage. As a result, this makes the adsorption energy significantly increase for bridge and hollow sites, but unchanged for atop site when decreasing the surface coverage.
Limitation
There is a possible limitation of this project. We used only 4 layers and 12 Å of vacuum spacing for our slab, which were arbitrarily chosen, to reduce the computation time. However, since the total slab energy can be also affected by the number of layers and vacuum spacing, it is highly required to optimize those parameters in the future study, even though it may not much affect the result for the binding preference of atomic O on Pt (111) surface.
Conclusion
By calculating and comparing the adsorption energy for all possible binding sites; atop, bridge, fcc hollow and hcp hollow, we could conclude that the most preferred binding site of atomic O on Pt (111) surface is fcc hollow site. And, we also could observe that lowering the surface coverage can lead to a stronger adsorption of atomic O on Pt (111) surface, especially for bridge and hollow sites. Even though a limitation for calculation exists, the preference we observed here is well matched to those of other computational and experimental studies.
References
[1] “Density-Functional Theory of Atoms and Molecules”, R.G. Parr, W. Yang, Y. Weitao (1994), Oxford University Press.
[2] “First principles methods using CASTEP”, Zeitschrift fuer Kristallographie 220(5-6) pp. 567-570 (2005) S. J. Clark, M. D. Segall, C. J. Pickard, P. J. Hasnip, M. J. Probert, K. Refson, M. C. Payne
[3] Perdew, J. P; Burke, K; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865-3868
[4] CASTEP GUIDE, BIOVIA, UK, 2014. URL : http://www.tcm.phy.cam.ac.uk/castep/documentation/WebHelp/content/pdfs/castep.htm.
[5] Density functional theory : a practical introduction / Section 4.9, David S. Sholl and Jan Stekel
[6] “Bonding Mechanism and Atomic Geometry of an Ordered Hydroxyl Overlayer on Pt(111)”, Seitsonen et al., J. Am. Chem. Soc. 2001, 123, 7347-7351
[7] “Atomic and molecular adsorption on Pt(111)”, D.C. Ford et al., Surface Science 587 (2005) 159–174
[8] “Chemisorption of atomic oxygen on Pt(1 1 1) and Pt/Ni(1 1 1) surfaces”, T. Jacob et al. / Chemical Physics Letters 385 (2004) 374–377
[9] “Oxygen interactions with the Pt(111) surface”, J.L. Gland, B.A. Sexton, G.B. Fisher, Surf. Sci. 95 (1980) 587.
