Quantitation of volatile aldehydes using chemoselective response dyes combined with multivariable data analysis

Current work proposed a novel quantitative method of volatile aldehydes (VAs) using chemoselective response dyes (CRDs) combined with multivariate data analysis. Multivariate spectral data of selected CRDs was obtained by visible near-infrared spectroscopy. The Synergy-interval Partial Least Squares (Si-PLS) algorithm processed multivariate spectral data to establish VAs quantitative prediction models at the level of 0.0002 v/v to 0.18 v/v. The prediction coefficient (Rp) values of models ranged from 0.8399 to 0.9886, and the Root Mean Square Error of Prediction (RMSEP) values were less than 0.01. These models were verified by classification of aging rice samples, and 93% samples were correctly identified in prediction set. In addition, Density Functional Theory (DFT) calculations explored the interaction mechanism between selected CRDs and VAs. The optimized Highest Occupied Molecular Orbital-Lowest Unoccupied Molecular Orbital (HOMO-LUMO) energy levels, dipole moment, distance between molecules were found to have strong correlations with the interaction.