Metal oxide surface dynamics from molecular dynamics simulations: the α-Al2O3(0 0 0 1) surface

doi:10.1016/S0039-6028(00)00988-2

Surface Science

Volume 474, Issues 1–3, 1 March 2001, Pages 107-113

https://doi.org/10.1016/S0039-6028(00)00988-2 Get rights and content

Abstract

Constant-stress, constant-temperature (10, 300 and 700 K) molecular dynamics simulations were carried out with shell-model potentials for an infinite crystalline Al-terminated α-Al₂O₃(0 0 0 1) slab of ∼25 Å thickness. The surface undergoes considerable relaxation at 10 K, but exhibits ordered surface structures at all three temperatures. The relaxation causes the (0 0 0 1) surface at 10 K to appear oxygen-terminated. The ionic motion within the central and surface regions of the slab system has been analysed in terms of mean-square displacements. At room temperature the 〈u²〉_surface/〈u²〉_bulk ratio for the Al ions is ≈2.5 and when only the out-of-plane surface motion is considered the ratio is as large as ≈3.5. The O ion motion at the surface is slightly smaller than that of the cations.

Introduction

This paper forms part of a series where we are exploring the surface dynamics of simple metal oxides, pure and defective, using molecular dynamics (MD) simulations. Other papers in this series treat the low-index faces of MgO [1] and ceria [2].

For metals, extensive experimental investigations of single crystal surfaces have helped create a rather solid foundation for the understanding of surface structure and functionality. In contrast, for metal oxides, experimental research on single crystals has developed much less rapidly, in part because the metal oxides are poor electrical conductors and build up charge when bombarded with electrons, which hinders ultrahigh-vacuum electron spectroscopy and diffraction experiments. Thus, it has then often been difficult to obtain reliable experimental atomic-level information concerning the average surface structure of metal oxides. Structural details about individual surface ions are of course even more complex and difficult to retrieve. Here the MD method emerges as a powerful alternative to surface structure experiments at functionally interesting temperatures.

Concerning surface dynamics, the experimental data is even more scarce. In a series of cumbersome measurements, Somorjai and coworkers as well as other research groups (see references in the monograph “Introduction to Surface Chemistry and Catalysis” by Somorjai [3]) have measured the mean-square displacement (msd) component perpendicular to the surface plane, 〈u²_⊥〉, for several cubic metal surfaces (often using low-energy electron-diffraction experiments). The 〈u²_⊥〉_surface/〈u²〉_bulk values found at room temperature typically lie in the range 1.8–7.0. For metal oxide surfaces on the other hand, there exist much fewer experimental determinations of the ionic dynamics. On the theoretical side, many quantum-mechanical and atomistic simulation papers have been published describing metal oxide surfaces under static conditions, with or without adsorbed molecules on top, but there are virtually no investigations of the surface dynamics at non-zero temperatures. In addition to our own work cited above, there exist in the literature only a handful reports of MD simulations of metal oxides with surfaces present, but these have all mainly been concerned with how the average overall structure changes at non-zero temperatures, and not with the atomic motion at the surface.

The oxide system under study in this paper is α-Al₂O₃ and its (0 0 0 1) surface. The surface structure of α-Al₂O₃(0 0 0 1) has been the subject of several atomistic and quantum-chemical studies in the past (see references in Table 2). As stated, to the best of our knowledge, there exist no previous theoretical reports concerning the ion dynamics for α-Al₂O₃(0 0 0 1) surfaces, nor for bulk α-Al₂O₃ actually. Blonski and Garofalini [4] have performed MD cluster simulations for a number of α-Al₂O₃ faces, but mainly for the purpose of comparing their relative surface energies, and dynamical properties were not discussed.

In addition to the ionic surface dynamics, in this paper we have also investigated the properties of the oxide films below the very surface layer, and how structure and dynamics change with depth into the slab. In this paper we will use the notations msd and 〈u²〉 interchangeably.

Section snippets

The molecular dynamics strategy and the interatomic potentials

The MD simulations were performed with a constant-stress, constant-temperature MD program for systems periodic in either three or two dimensions, with dynamically variable lattice vectors (lengths and angles) [5]. The Ewald summation technique for both two- and three- dimensionally periodic systems [6], [7], [8] is implemented in the code. All ions were allowed to move in the simulations. The Nosé–Hoover formalism [9], [10] was used for the 300 and 700 K simulations, but our 10 K MD simulations

Surface structure

Our simulation gives room-temperature cell parameters of 4.83 and 12.69 Å, in reasonable agreement with experiment (they differ by 1.5% and 2.3%, respectively, from experiment). The interatomic distances in the center of the slab remain virtually uneffected by the presence of the surface, but, as is well known (see Table 2 and references therein), the surface layers undergo significant rearrangement with the Al ions sinking down into the space between the upper oxygens (see the 10 K structure

Concluding remarks

We have calculated mean-square (m.s.) amplitudes for the cations and anions in bulk alumina and at the (0 0 0 1) surface for different temperatures using classical MD simulations. For the very stable (0 0 0 1) surface the computed msds can be interpreted in a straightforward manner as m.s. amplitudes arising from ionic vibrations around equilibrium positions, and we make the following conclusions:

•
〈u²〉_surface for O and Al at the surface are approximately the same.
•
〈u²_⊥〉_surface is approximately twice as

Acknowledgements

Valuable discussions with Dr. Mark Wojcik and financial support from The Swedish Natural Science Research Council are gratefully acknowledged.

References (30)

S Blonski et al.
Surf. Sci.
(1993)
M Baudin et al.
Surf. Sci.
(1997)
W.C Mackrodt et al.
J. Cryst. Growth
(1987)
M Causa et al.
Surf. Sci.
(1989)
I Manassidis et al.
Surf. Sci. Lett.
(1993)
V.E Puchin et al.
Surf. Sci.
(1997)
M. Baudin, M. Wojcik, K. Hermansson, submitted for...
M. Baudin, M. Wojcik, K. Hermansson, Surf. Sci., accepted for...
G. Somorjai, Introduction to Surface Chemistry and Catalysis, Wiley, New York, 1994; see especially references in...
P.P Ewald
Ann. Phys.
(1921)

D.E. Parry, Surf. Sci. 49 (1975) 433; Erratum 54 (1976)...

D.M Heyes et al.

J. Chem. Soc. Faraday Trans. II

(1977)

S Nosé

J. Chem. Phys.

(1984)

W.G Hoover

Phys. Rev. A

(1985)

B.G Dick et al.

Phys. Rev.

(1958)

Cited by (34)

Polybenzopyrrole/nano-alumina composite blend with zirconium silicate reinforced epoxy as protective coating to subside corrosion of carbon steel within a dilute NaCl solution
2024, Journal of Molecular Structure
The present work inspects the corrosion retardation of carbon steel by zirconium silicate reinforced epoxy blended with fabricated (PB/AN) polybenzopyrrole/nano-alumina composite, termed as PB/AN@ZE coating. Nano-alumina (AN), polymer polybenzopyrrole (PB), and polymer nanocomposite (PB/AN) were conveniently synthesized and their structural, thermal and morphological information were confirmed through Fourier transform infrared spectroscopy, X-ray diffraction, Thermogravimetric analysis, Atomic Force Microscopy and Field Emission Scanning Electron Microscopy. The molecular dynamics computation data evidence that the designed composite PB/AN is highly stable and found to have the interaction energy of 67.63 eV. Anti-wettability behaviour of the formulated coating was examined using Contact-angle Goniometer. The corrosion-resistant behaviour with different weight% loading of PB/AN in the zirconium silicate reinforced epoxy coating on low-carbon steel was examined in 3.5 % NaCl by using EIS or Electrochemical Impedance Spectroscopy and PDP or Potentiodynamic polarisation studies. The EIS and PDP results of 0.25% PB/AN@ZE coating indicated the excellent barrier to control the corrosion kinetics with high protection efficiency of 97.14 % even after 24 days immersion in NaCl solution and damage induced coating (kept immersed for 24 h in NaCl) was evidenced with a high impedance value of 1.48 × 10⁹ Ω.cm².
Creation of models and parametrization of a classical force field for amorphous Al<inf>2</inf>O<inf>3</inf>/water interfaces based on Density Functional Theory
2017, Computational Materials Science
Citation Excerpt :
In order to achieve this goals, classical force-field potentials enabling the dynamical simulation of systems containing several hundred thousand atoms for at least several hundreds of nanoseconds are needed. Classical force fields describing the interactions in alumina bulk and dry surfaces are already available, either based on fixed charges [10,11] or including polarizability [12] and charge transfer schemes [13]. Potentials mapping the interactions between the surface and its surrounding environment have also been presented, for instance general oxide force fields [14], or parametrized for special cases such as crystalline alumina surfaces [7] or small molecule/small cluster interactions [15].
We present realistic models of amorphous alumina/water interfaces taking into account the chemical termination with hydroxyl groups and chemisorbed water molecules, as predicted by DFT-based MD simulations. On the basis of these models we parametrize a fixed-charges force field to describe the interactions within alumina and across the alumina/water interface, which can be seamlessly combined with biomolecular force fields via straightforward combination rules. The potential parameters are fitted to DFT reference calculations of water/surface interactions and are shown to be very well transferable to organic molecules adsorbing on alumina, in particular amino acid analogues. The correct protonation/deprotonation equilibria of the surface sites in contact with water solutions at different pH values is taken into account and implemented in the surface models and force field parameters via adjustments of the atomic point charges. The ability of our models and force field to predict experimental observables is demonstrated by unbiased calculations of the heat of immersion of alumina powders and of the adsorption free energy of glutathione disulfide on alumina in comparison with experimental data.
Atomistic simulation and virtual diffraction characterization of stable and metastable alumina surfaces
2014, Acta Materialia
The structures of select alumina surfaces are studied using molecular statics and molecular dynamics simulations and are characterized using virtual diffraction methods. First, bulk alumina simulations are performed to validate the transferability of the ReaxFF potential to model different alumina phases. Bulk alumina simulations accurately predict α-Al₂O₃ as the lowest energy crystalline phase; however, they unexpectedly predict an even lower-energy amorphous phase. At 0 K, virtual X-ray diffraction patterns of the bulk crystalline phases and select alumina surfaces are validated by experimental studies. Molecular statics simulations of select alumina surfaces are consistent with prior first-principles studies. However, molecular dynamics simulations show that many surfaces experience significant reconstructions at temperatures below what is expected from experiments. It is believed that premature surface reconstructions are biased by the predicted lower-energy amorphous phase and occur due to the extra degrees of freedom allowed by the free surfaces as well as the available thermal energy during dynamics. Discrete peaks appearing in virtual selected-area electron diffraction patterns indicate that the reconstructions are not fully amorphous due to lattice constraints imposed by the internal bulk structure. Bulk and surface energies are tabulated for each simulation to be used in future predictive mesoscale models of polymorphic alumina.
Simulation of structural features on mechanochemical synthesis of Al <inf>2</inf>O <inf>3</inf>-TiB <inf>2</inf> nanocomposite by optimized artificial neural network
2012, Advanced Powder Technology
Citation Excerpt :
In both cases, TiB2 and Al2O3, mean peaks for the simulated data are higher than experimental ones leading the simulated interplanar distances to be lower than experimental results. However, herein the obtained value of MAE was more reasonable compare with other reported data, for example according to Baudin and Hermansson [29] for Al2O3 their obtained results differ by 1.5–2.3% from the experiment. In this fact, nearly good agreement between the neural predictions and experimental verifications has been demonstrated but as it can be seen in these figures the average errors are still varied.
In this study, structural features of alumina–titanium diboride nanocomposite (Al₂O₃–TiB₂) were simulated from the mixture of titanium dioxide, boric acid and pure aluminum as raw materials via mechanochemical process using the optimized artificial neural network. The phase transformation and structural evolutions during the mechanochemical process were characterized using X-ray powder diffractometry (XRD). For better understanding the refining crystallite size and amorphization phenomena during the milling, XRD data were modeled and simulated by artificial neural network (ANN). An ANN consisting of three layers of neurons was trained using a back-propagation learning rule. Also, the ANN was optimized by Taguchi method. Additionally, the crystallite size, interplanar distance, amorphization degree and lattice strain were compared for the simulated values and experimental results.
Atomistic simulation study of atomic size effects on B1 (NaCl), B2 (CsCl), and B3 (zinc-blende) crystal stability of binary ionic compounds
2011, Computational Materials Science
Citation Excerpt :
The EIM format is chosen because it uses an environment dependent many-body term to analytically address the variable charge interaction caused by electronegativity difference between different species. The fixed charge models commonly used for ionic compounds [18–26] are most accurate for defect-free stoichiometric compounds of a specific crystal structure as they do not allow environment-dependent charge variation. Because atoms remain ionized even in the vapor phase, the fixed charge models also typically overestimate cohesive energies of compounds by 4–5 times [27].
Ionic compounds exhibit a variety of crystal structures that can critically affect their applications. Traditionally, relative sizes of cations and anions have been used to explain coordination of ions within the crystals. Such approaches assume atoms to be hard spheres and they cannot explain the observed structures of some crystals. Here we develop an atomistic method and use it to explore the structure-determining factors beyond the limitations of the hard sphere approach. Our approach is based upon a calibrated interatomic potential database that uses independent intrinsic bond lengths to measure atomic sizes. By carrying out extensive atomistic simulations, striking relationships among intrinsic bond lengths are discovered to determine the B1 (NaCl), B2 (CsCl), and B3 (zinc-blende) structure of binary ionic compounds.
Oxygen vacancy diffusion in alumina: New atomistic simulation methods applied to an old problem
2009, Acta Materialia
Understanding diffusion in alumina is a long-standing challenge in ceramic science. The present article applies a novel combination of metadynamics and kinetic Monte Carlo simulation approaches to the investigation of oxygen vacancy diffusion in alumina. Three classes of diffusive jumps with different activation energies were identified, the resulting diffusion coefficient being best fitted by an Arrhenius equation having a pre-exponential factor of 7.88 × 10⁻² m² s⁻¹ and an activation energy of 510.83 kJ mol⁻¹. This activation energy is very close to values for the most pure aluminas studied experimentally (activation energy 531 kJ mol⁻¹). The good agreement indicates that the dominating atomic-scale diffusion mechanism in alumina is vacancy diffusion.

View all citing articles on Scopus

View full text

Metal oxide surface dynamics from molecular dynamics simulations: the α-Al2O3(0 0 0 1) surface

Abstract

Introduction

Section snippets

The molecular dynamics strategy and the interatomic potentials

Surface structure

Concluding remarks

Acknowledgements

Surf. Sci.

Surf. Sci.

J. Cryst. Growth

Surf. Sci.

Surf. Sci. Lett.

Surf. Sci.

Ann. Phys.

J. Chem. Soc. Faraday Trans. II

J. Chem. Phys.

Phys. Rev. A

Phys. Rev.

Metal oxide surface dynamics from molecular dynamics simulations: the α-Al₂O₃(0 0 0 1) surface