Raman signatures of ferroic domain walls captured by principal component analysis

TL;DR

This paper demonstrates that Principal Component Analysis (PCA) applied to Raman spectroscopy maps can effectively identify and analyze subtle spectral variations at ferroic domain walls, offering a new statistical approach for studying these features.

Contribution

The study introduces PCA as a rapid, reliable method to detect small Raman peak shifts at ferroelectric and ferroelastic domain walls without prior assumptions.

Findings

PCA can accurately identify small Raman peak variations at domain walls.

Peak shifts can be deduced from spectral data using a Taylor expansion.

PCA separates contributions of domains and domain walls in Raman spectra.

Abstract

Ferroic domain walls are currently investigated by several state-of-the art techniques in order to get a better understanding of their distinct, functional properties. Here, Principal Component Analysis (PCA) of Raman maps is used to study ferroelectric domain walls (DWs) in LiNbO3 and ferroelastic DWs in NdGaO3. It is shown that PCA allows to quickly and reliably identify small Raman peak variations at ferroelectric DWs and that the value of a peak shift can be deduced - accurately and without a-priori - from a first order Taylor expansion of the spectra. The ability of PCA to separate the contribution of ferroelastic domains and DWs to Raman spectra is emphasized. More generally, our results provide a novel route for the statistical analysis of any property mapped across a DW.