Ab initio modeling and experimental analysis of contrast in broadband coherent anti-Stokes Raman spectroscopy: A case study on ferroelectric domain walls

Raman imaging, based on incoherent spontaneous Raman scattering (SR) or coherent techniques such as broadband coherent anti-Stokes Raman spectroscopy (BCARS), is a powerful tool for visualizing local lattice distortions. It can thus be employed, e.g., to identify and visualize ferroelectric (FE) domain walls (DWs). While SR-based DW imaging relies on small changes in the intensity or position of Raman peaks at the DW, BCARS measurements may detect DW signatures with a drastically increased contrast and orders of magnitude faster imaging speed. However, while SR-based contrasts can be predicted from existing models, the increased DW contrasts in BCARS remain poorly understood. In this work, we develop a first-principles approach for the simulation of CARS spectra to investigate the origin of this BCARS signal contrast at periodically poled stoichiometric lithium niobate (sPPLN) DWs. Combining these atomistic simulations with strain modeling allows us to explore how strain affects the CARS signal generation at FE DWs. We then apply polarization-sensitive BCARS to image sPPLN DWs for different polarization configurations. Our results reveal additional Raman peaks specific to the DWs across all investigated polarization combinations, yielding a strong DW-to-bulk contrast that enables in situ identification and localization of DWs in two-dimensional BCARS maps. This study demonstrates the practical utility of polarization-sensitive BCARS for fast DW imaging and provides fundamental insights into the BCARS contrast mechanism in FE materials, both in strained regions and in the unstrained bulk.