Nanoscale Surface Structure of Nanometer-Thick Ferroelectric BaTiO₃ Films Revealed by Synchrotron X-ray Scanning Tunneling Microscopy: Implications for Catalytic Adsorption Reactions

Abstract

Ferroelectric nanomaterials are of interest in catalysis, nonvolatile memory, and neuromorphic computing among other applications because of their switchable structure that can alter the electronic and interface properties of a single material. The investigation of the role of polarization on the surface structure and chemistry of ferroelectric nanomaterials is a longstanding challenge, as it ideally requires a combination of both nanoscale imaging and chemical spectroscopy. In this work, we study a model ferroelectric BaTiO 3 thin film by synchrotron Xray scanning tunneling microscopy (SX-STM), a unique method that integrates nanoscale surface imaging and chemically sensitive spectroscopy. We find that polarization switching from downward to upward in (001) single-crystalline BaTiO 3 thin films increases the intensity of X-ray absorption across Ba M, Ti L, and O K edges. Chemical mapping of nanometer-sized domains further demonstrates the modulation of surface structures upon polarization switching, as well as confirming the trends observed in singlepoint experiments across the surface. We complement these measurements with ab initio computational absorption spectroscopy to elucidate the effect of polarization switching on the core−hole excitations using the Bethe−Salpeter equation approach. Our experimental and theoretical results thus confirm a stronger binding strength for the upward-polarized surface with molecular O 2 as a model reactant, offering mechanistic evidence that supports previous reports. This work advances the understanding of the surface chemistry and electronic structure of ferroelectrics, which can ultimately aid strategies to design interfaces with tailored properties.

Investigation of local distortion effects on X-ray absorption of ferroelectric perovskites from first principles simulations

“…These are herein denoted the determinant and BSE approaches, respectively. By benchmarking against experimental data, 36 we show that while both approaches accurately predict the effect of ferroelectric modulation on the main edges and fine structures at the O–K edge, i.e. , low energy t 2g peaks, the consideration of multiplet and spin–orbit coupling effects is critical for accurately capturing the physics, and is necessary to model 2p → 3d transitions associated with Ti–L edges.…”