Author: Naveen Agrawal
Introduction: Platinum is a model metal system for catalytic reactions. This computational study focuses to determine the preferred binding site for the atomic oxygen. In addition, the effect of increasing the coverage of oxygen on the adsorption energy of oxygen has been evaluated. Platinum has been also studied as a model electrocatalyst for a variety of electrochemical reactions, especially the Pt (111) facet which is the most stable facet thermodynamically. However, experiments and computational studies indicate the presence of atomic oxygen on the metal surface at oxidation potentials as low as 0.85 V (RHE) [1]. This study will show the trend in the adsorption affinity of atomic oxygen as we increase the coverage which could be considered as a consequence of higher oxidation potentials.
Methods: DFT (Density Functional Theory) calculations were performed within the Vienna Ab initio simulation package (VASP, version 5.4.4), using the periodic supercell approach [2]. The projector augmented wave (PAW) method was used for electron-ion interactions [3,4]. The Perdew-Burke-Ernzerhof functional described the electron-electron exchange and correlation energies [5]. A plane wave basis set is used with an energy cutoff of 450 eV. For geometry optimization, the convergence criteria of the forces acting on atoms were 0.05 eV Å-1, while the energy threshold-defining self-consistency of the electron density was 10-5 eV. 5-layer asymmetric atomic slab (with bottom 3-layer fixed to represent the bulk with lattice constant 3.97 Angstrom optimized with 11X11X11 k-points grid and 450 eV cutoff) with 3X3 unit cell and 15 Angstrom vacuum were considered to observe the effect of different coverage of atomic oxygen on the top layer. Correction to total electronic energy in VASP environment was performed using the method of Bengtsson to remove spurious periodic slab dipole interactions [6]. Previous studies have suggested that even 4 -layer atomic slab with bottom 2-layer fixed are enough to represent the bulk and surface features of the Pt (111) system [7].
The binding energy of atomic oxygen on Pt(111) was calculated w.r.t. a single oxygen molecule in vacuum as illustrated by equation 1.
\begin{equation}E_{binding}=\frac{E_{surf+O}-\frac{n*E_{O_2(g))}}{2}-E_{surf}}{n}\end{equation}
Where \(E_{binding}\) is the binding energy per atom of oxygen on Pt(111), \(E_{surf+O}\) is energy for the adsorbate surface system, \(E_{O_2(g)}\) is the energy of oxygen molecule in vacuum evaluated in a 20 A cubic unit cell in VASP with unrestricted spin, \(E_{surf}\) is the energy of Pt(111) slab as described above, and \({n}\) is number of atomic oxygen adsorbed onto Pt(111) slab.
Convergence tests: Convergence tests were performed to evaluate the optimal k-points grid, energy cutoff, and the vacuum thickness for the periodic supercell. The energy tolerance considered for these convergence tests was 0.015 eV. As the studied unit-cell was a 3X3 and 5-layer atomic slab, the k-points grids considered were (MxMx1), Where M ranges from 3 to 7. We need only unit discretization of the grid in Z direction because the Z-direction supercell dimension is almost 3 times higher than X and Y. The relative energy is plotted in Figure 1(a) across the irreducible k-points obtained through different grids. Similar convergence tests were performed to obtain the optimal energy cutoff and the vacuum thickness for the asymmetric 5-layer atomic slab as shown in the figure below. All the convergence tests were performed with a single oxygen present on the FCC site of Pt (111) 3X3 unit cell as shown in Figure 1(d). Based on the convergence tests k-points grid of 6X6X1 with 54 irreducible k-points, 450 eV energy cutoff, and 15 A of vacuum slab were determined to be suitable within the tolerance specified.
Figure 1: Convergence tests for k-points, energy cutoff, and Vacuum layer thickness for a 5-layer 3X3 unit cell of Pt (111)
Preferred binding site of atomic oxygen and coverage effect: Binding energy relative to atop (lowest binding energy site) is plotted against different binding configurations possible on Pt(111) for atomic oxygen in Figure 3. Based on the relative binding energy obtained, FCC site which is a three-fold hollow site on Pt (111) as shown in Figure 3 has been determined to be the optimal binding site for the atomic oxygen.
Figure 2: Different possible binding configurations for atomic oxygen on Pt(111)
Figure 3: Preferred binding for atomic oxygen
Figure 5 shows that the increased coverage on the fcc site reduces the per-atom binding energy of oxygen on Pt(111). This effect can be visualized from Figure 4, where increasing coverages can lead to metal atom sharing between the adsorbed oxygen atoms possibly reducing the binding per oxygen. In addition, the co-adsorbates of similar nature also have unfavorable dipole-dipole repulsion even without metal atom sharing which causes a reduction in overall binding at higher coverages.
Figure 4: 1/9 to 1/3 ML of oxygen on FCC sites for Pt(111)
Figure 5: Effect of coverage of atomic oxygen on binding energy
Conclusion: DFT calculations performed in this study agree with the earlier studies that preferred binding site of oxygen on Pt (111) is fcc (3-fold hollow site). The binding energy difference between hcp and fcc sites matches with earlier theoretical studies and experimental observations [8]. Adsorption energy of atomic oxygen decreases as we increase the coverage on the fcc sites which is consistent with experiments [9].
References:
- Kondo.et.al. The Journal of Physical Chemistry C, 29, 16118-16131.
- Kresse, G. and Furthmüller, J.Physical review B, 54(16), p.11169.
- P. E. Blochl, Phys Rev B, 1994, 50, 17953-17979
- G. Kresse.et.al Phys Rev B, 1999, 59, 1758-1775
- J. P. Perdew. et.al , Phys Rev B, 1992, 46, 6671-6687.
- Bengtsson, Physical Review B, 59(19), 12301.
- Kokalj.et.al. Journal of Physics: Condensed Matter, 39, 7463.
- Yeo, Y. Y., L. Vattuone.et.al, The Journal of chemical physics 106, no. 1 (1997): 392-401.
- Gu, Z.et.al, The Journal of Physical Chemistry C, 111(27), 9877-9883.
