Abstract
The rising incidence of adulteration in high-value foods requires rapid, reliable, and cost-effective analytical methods. This study presents a novel chemometric approach using ATR-FTIR spectroscopy to detect fake pistachio powder (FPP) in pistachio powder (PP) and to evaluate its potential for semi-quantitative estimation of the adulteration level. Two-dimensional correlation spectroscopy (2D-COS) identified key spectral regions associated with compositional changes due to adulteration. Principal component analysis (PCA) enabled visualization of sample variability and clustering. PLS regression was additionally investigated for estimating FPP content. The best PLS model yielded an RMSEP of 24.0971 across the complete 0-100% prediction range, indicating that the regression model is more appropriate for semi-quantitative assessment than for precise quantification. For classification, partial least squares discriminant analysis (PLS-DA) achieved high sensitivity (≈99-100%), accuracy (≈99%), and specificity (up to 96%) in the 1577-1110 cm-1 region. Overall, combining 2D-COS-guided spectral selection with chemometric modeling shows potential as a rapid and non-destructive screening approach for assessing PP authenticity under controlled conditions.