Quantitation of Saccharin and Cyclamate in Tabletop Formulations by Portable Raman and NIR Spectrometers in Combination with Partial Least Squares Regression

Abstract

Rapid, direct, and reagent-free screening tools using vibrational spectroscopy in combination with partial least squares regression (PLSR) were developed for the determination of sodium saccharin and sodium cyclamate in tabletop formulations. The four vibrational spectroscopic instruments employed were a portable Raman spectrometer, a NIR handheld device, and two FT-NIR benchtop spectrometers. Wavenumber ranges and type of spectral pretreatment were optimized for each PLSR calibration model using an independent validation set. Each sweetener model provided reliable predictions (low errors in validation and r ² above 0.90) for both saccharin and cyclamate samples. Optimized models were tested with four commercially available tabletop formulations in order to simulate the application of the developed models towards routine sweetener analysis. With the exception of one model, the sweetener concentration predictions in commercial tabletop formulations using the portable devices were not significantly different from those based on spectra collected on the benchtop spectrometers. PLSR-predicted mean sweetener concentrations were within 80–120% of the label declared values, while the sweetener without a label declaration had consistent concentrations across the analytical methods used. As shown by the good agreement between spectroscopic and chromatographic analyses, the portable spectrometers offer an alternative to traditional chromatographic methods. To our knowledge, this is the first time portable Raman and handheld NIR devices with PLSR calibration models have been employed to evaluate sweeteners, and these analytical methods hold potential to be used for rapid screening of tabletop formulations for quality assurance and for regulatory labeling verification.