Enhancing ferroelectric characterization at nanoscale: A comprehensive approach for data processing in spectroscopic piezoresponse force microscopy

• Published in: Journal of applied physics Vol. 135; no. 19
• Main Authors: Valloire, H., Quéméré, P., Vaxelaire, N., Kuentz, H., Le Rhun, G., Borowik, Ł.
• Format: Journal Article
• Language: English
• Published: Melville American Institute of Physics 21.05.2024
• Subjects: Algorithms; Cross correlation; Data processing; Electric fields; Electronics; Electrostatics; Engineering Sciences; Ferroelectric materials; Ferroelectricity; Hysteresis loops; Machine learning; Material properties; Materials; Microscopy; Phase separation; Physics; Thin films
• ISSN: 0021-8979; 1089-7550; 1520-8850; 1089-7550
• DOI: 10.1063/5.0197226

Switching Spectroscopy Piezoresponse Force Microscopy (SSPFM) stands out as a powerful method for probing ferroelectric properties within materials subjected to incremental polarization induced by an external electric field. However, the dense data processing linked to this technique is a critical factor influencing the quality of obtained results. Furthermore, meticulous exploration of various artifacts, such as electrostatics, which may considerably influence the signal, is a key factor in obtaining quantitative results. In this paper, we present a global methodology for SSPFM data processing, accessible in open-source with a user-friendly Python application called PySSPFM. A ferroelectric thin film sample of potassium sodium niobate has been probed to illustrate the different aspects of our methodology. Our approach enables the reconstruction of hysteresis nano-loops by determining the PR as a function of applied electric field. These hysteresis loops are then fitted to extract characteristic parameters that serve as measures of the ferroelectric properties of the sample. Various artifact decorrelation methods are employed to enhance measurement accuracy, and additional material properties can be assessed. Performing this procedure on a grid of points across the surface of the sample enables the creation of spatial maps. Furthermore, different techniques have been proposed to facilitate post-treatment analysis, incorporating algorithms for machine learning (K-means), phase separation, and mapping cross correlation, among others. Additionally, PySSPFM enables a more in-depth investigation of the material by studying the nanomechanical properties during poling, through the measurement of the resonance properties of the cantilever–tip–sample surface system.