Nature of the Ba 4d Splitting in BaTiO3Unraveled by a Combined Experimental and Theoretical Study

In this paper, synchrotron radiation photoemission techniques are used to unravel the pending nature of the two Ba 4d components, also observed on Ba 3d lines, and commonly named Ba(α) and Ba(β). Investigations were carried out on the (001) surface of epitaxial very thin (≤20 nm) films combined with DFT calculations. Photoelectron diffraction experiments reveal different behaviors for Ba(α) and Ba(β) components. While six well-defined diffraction features corresponding to the BaTiO3 bulk structure are observed for the Ba(α) component, no feature really appears in the diffraction pattern of Ba(β). This observations reveal the absence of forward scattering processes, hence this component comes mostly from surface BaO plane. Supporting these results, DFT calculations demonstrate a shift to lower binding energy (BE) only for the Ba 4d peak from the topmost BaO layer with respect to the photoelectron peak stemming from bulk layers. Hence, only one peak (the Ba(α) one) should be observed for a sample with a complete TiO2 termination. This hypothesis is supported by photoemission measurements carried out at low photon energy in grazing angle emission, in order to enhance the signal from topmost layers with respect to the bulk. For each sample, the Ba(β)/Ba(α) intensity ratio is well reproduced by a model where the Ba(β) signal is only coming from the topmost surface BaO plane.