Comparison of Density Functional Theory Predictions with Experimentally Measured Lanthanum Strontium Ferrite Thin Film Oxygen Surface Exchange Coefficients

Introduction
Materials with high oxygen surface exchange coefficients at low temperature are needed for Solid Oxide Fuel Cell (SOFC) electrode applications. Perovskite (ABO_3−x) type oxides have been extensively studied as candidates for SOFC cathode materials¹. The objective of this research was to understand the role stress and crystal structure have on the oxygen surface exchange coefficient. To achieve this objective, the oxygen surface exchange coefficients of Lanthanum Strontium Ferrite (LSF) thin films of various compositions and crystal structures were experimentally measured with a new curvature relaxation technique. These measurements, and lattice constant measurements from the literature, were then compared to Density Functional Theory (DFT) calculations of these same LSF compositions.

Experimental Methods
    As a proof of concept, rhombohedral La_0.6Sr_0.4FeO_3-x (LSF64) thin films were sputtered onto one-side-polished, 1 inch diameter, 200 μm thick, (100)-oriented, (Y₂O₃)_0.13(ZrO₂)_0.87 (YSZ) single crystal wafers. These bilayers were subjected to mechano-chemical strain under different partial pressures of oxygen. Specifically, the equation shown in Figure 1 was used to measure the oxygen surface exchange coefficient (k_i) from the curvature of a bilayer responding to a step change in oxygen partial pressure^2-5, where K denotes the instantaneous bilayer curvature at a certain partial pressure of oxygen, K₀ denotes initial bilayer curvature, K_inf is the equilibrium bilayer curvature, k_i is the oxygen surface exchange coefficient for the fraction of a film with surface area A_i, t is time, and h_f is the film thickness.

Computational Methods
    Different crystal structures of LSF were modeled with DFT calculation using the Vienna Ab initio Simulation Package (VASP). The crystal lattices for orthorhombic, rhombohedral and cubic LSF were calculated using an energy minimization technique, as shown in Figure 2. The lattice for each of these structures was converted to a super-cell and extended in z-direction to create active surface. One oxygen atom was then deleted from the active surface to create an oxygen vacancy. The effect of lattice strain due to oxygen vacancy formation was calculated from comparisons between the original and relaxed structures.

Result & Discussion
    The DFT calculated lattice parameters (a=3.8905 A) for cubic, perovskite-structured LSF was in good agreement with experimental measurements from the literature (a=3.858 A).⁷ The calculated lattice lengths (a=5.662 A, b=7.945 A, c=5.601 A) for orthorhombic LSF also compared well with experimental measurements from the literature (a=5.556 A, b=7.885 A, c=5.653 A).⁸ Additional work is ongoing to study intermediate compositions, and the role of stress has on oxygen surface exchange.

Acknowledgements
This work was made possible through a Michigan State University faculty startup grant.

References
1.    M.S. Islam, J. Mater. Chem., 10, 1027 (2000).
2.    Q. Yang , T.E. Burye, R.R. Lunt and J.D. Nicholas, Solid State Ionics, 249–250, 123 (2013).
3.    J. Crank, The Mathematics of Diffusion, 2nd edition., p. 60, Oxford University Press, Oxford, UK (1995).
4.    G. Kim, S. Wang, A.J. Jacobson and  C.L. Chen, Solid State Ionics, 177, 1461 (2006).
5.    K. Kerman, C.H. Ko and S. Ramanathan, Phys. Chem. Chem. Phys., 14, 11953 (2012).
6.    A. Fossdal, M. Menon, I. Waernhus, K. Wiik, M.A. Einarsrud and T. Grande, J. Am. Ceram. Soc. , 87, 1952 (2004).
7.    Y. Takeda, K. Kanno, T. Takada, O. Yamamoto, M. Takano, N. Nakayama and Y. Bando, J. Solid State Chem., 63,  237 (1986).
8.    J. Luning, F. Nolting, A. Scholl, H. Ohldag, J.W. Seo, J. Fompeyrine et al., Phys. Rev. B, 67, 214433 (2003).