SANHUEZA, M.I., MELÉNDREZ, M.F., VON PLESSING, C., BASTIDAS, C.Y., NEIRA, J.Y., FREER, J., CASTILLO, R.D.P.:
J Raman Spectrosc. 2020; 51: 649– 659.
Raman microimaging, as a product of Raman microspectroscopy mapping and multivariate analysis, was used for the localization and quantification of active pharmaceutical ingredients (APIs) in tablets made in laboratory. This was done to develop an analytical strategy to simultaneously recover qualitative and quantitative information of solid dosage forms at a microscopic level by using a nondestructive method. A chemical system, composed of acetaminophen (AMP), caffeine, and one excipient (microcrystalline cellulose), was subjected to chemometric analysis through principal component analysis (PCA) and multivariate curve resolution with alternating least squares (MCR–ALS). This was done by using Raman spectra obtained from microscopic images with pixel sizes of 15 × 15 μm to localize the APIs in the tablets. Partial least squares (PLS) was applied as a calibration method to obtain bulk and single‐pixel concentrations of APIs in the samples. MCR–ALS provided better results than PCA for the localization of APIs. PLS achieved satisfactory root mean standard error values in the external validation set (<4% w/w) in bulk concentration determinations of AMP. This method also achieved concentrations for each pixel of the images, reconstructing images very similar to those obtained by MCR–ALS. Consequently, simultaneous localization and quantification of AMP was possible. Finally, the performance of Raman microimaging was evaluated through estimation of analytical figures of merit (AFOMs) of the technique used to assess the quantification of APIs. This included different calculus of uncertainty in the signal in a technique where the signal/noise ratio is low, and AFOMs for multivariate quantification are not often reported.