Analysis of Chemical Heterogeneity in Electrospun Fibers Through Hyperspectral Raman Imaging Using Open-Source Software.

Electrospinning is a versatile technique for producing porous nanofibers with a high specific surface area, making them ideal for several tissue engineering applications. Although Raman spectroscopy has been widely employed to characterize electrospun materials, but most studies report bulk-averaged properties without addressing the spatial heterogeneity of their chemical composition. Raman imaging has emerged as a promising tool to overcome this limitation; however, challenges remain, including limited sensitivity for detecting minor components, reliance on distinctive high-intensity bands, and the frequent use of commercial software. In this study, we present a methodology based on Raman hyperspectral image processing using open-source software (Python), capable of identifying components present at concentrations as low as 2% and 5%, even in the absence of exclusive bands of high or medium intensity, respectively. The proposed approach integrates spectral segmentation, end member extraction via the N-FINDR algorithm, and analysis of average spectra to map and characterize the chemical heterogeneity within electrospun fibers. Finally, its performance is compared with the traditional approach based on band intensities, highlighting improvements in sensitivity and the detection of weak signals.