Cation segregation on perovskite oxide surfaces affects vastly the oxygen reduction activity and stability of solid oxide fuel cell (SOFC) cathodes. A unified theory that explains the physical origins of this phenomenon is therefore needed for designing cathode materials with optimal surface chemistry. We quantitatively assessed the elastic and electrostatic interactions of the dopant with the surrounding lattice as the key driving forces for segregation on model perovskite compounds, LnMnO3 (host cation Ln = La, Sm). Our approach combines surface chemical analysis with X-ray photoelectron and Auger electron spectroscopy on model dense thin films and computational analysis with density functional theory (DFT) calculations and analytical models. Elastic energy differences were systematically induced in the system by varying the radius of the selected dopants (Ca, Sr, Ba) with respect to the host cations (La, Sm) while retaining the same charge state. Electrostatic energy differences were introduced by varying the distribution of charged oxygen and cation vacancies in our models. Varying the oxygen chemical potential in our experiments induced changes in both the elastic energy and electrostatic interactions. Our results quantitatively demonstrate that the mechanism of dopant segregation on perovskite oxides includes both the elastic and electrostatic energy contributions. A smaller size mismatch between the host and dopant cations and a chemically expanded lattice were found to reduce the segregation level of the dopant and to enable more stable cathode surfaces. Ca-doped LaMnO3 was found to have the most stable surface composition with the least cation segregation among the compositions surveyed. The diffusion kinetics of the larger dopants, Ba and Sr, was found to be slower and can kinetically trap the segregation at reduced temperatures despite the larger elastic energy driving force. Lastly, scanning probe image contrast showed that the surface chemical heterogeneities made of dopant oxides upon segregation were electronically insulating. The consistency between the results obtained from experiments, DFT calculations, and analytical theory in this work provides a predictive capability to tailor the cathode surface compositions for high-performance SOFCs.
