Direct analysis of calcium in liquid infant formula via laser-induced breakdown spectroscopy (LIBS)

The present work illustrates the potential of laser-induced breakdown spectroscopy (LIBS) for the direct analysis of liquid food products. The aim of the experiment was to predict calcium content in ready-to-feed infant formula. The analysis was performed by a LIBS system coupled to a liquid sample chamber with a rotatory wheel that presents the liquid to the laser beam as a thin film. Multivariate analysis with partial least squares regression (PLSR) was performed to correlate LIBS spectral data to reference calcium contents. The obtained PLSR model exhibited a good fit and linearity, as indicated by the coefficients of determination for calibration (R_c²) and cross-validation (R_cv²), with values of 0.96 and 0.89, respectively. The robustness of the calibration model was assessed by external validation showing a root-mean-square error of prediction of 6.45 mg 100 mL^-1. These results demonstrated the potential of LIBS for real-time analysis of liquid food products.