LAUSR

Abstract Fluorescence excitation-emission matrix spectroscopy coupled to multi-way analysis has proved to be a powerful tool for the study of complex systems with analytical purposes. However, scattering phenomena that are usually present in fluorescence landscapes can significantly affect the performance of the chemometric modelling or lead to misinterpretations of the spectral information. In this work, an improved algorithm that collects the strengths of the reported approaches and enhances their performances was developed. The proposed algorithm, which is based on the signal-conservative principle, enables the fluorescence landscapes correction by preserving the inherent particularities of the original. Moreover, corrected second-order data were subjected to trilinear decomposition analysis to assess the performance of the scatter correction in terms of trilinearity and prediction ability. In light of the obtained results, this new strategy showed to be adequate for scattering correction in several experimental situations.