Improved Colorimetric-Spectroscopic System Combined with Weighted Partial Least Squares for the Rapid Detection of Acephate In Grain

To enable the rapid detection of acephate residues, an improved colorimetric method was developed using palladium chloride (PdCl₂) as the chromogenic reagent. Sodium chloride was selected to replace hydrochloric acid as the optimized solvent, effectively circumventing the limitations of conventional PdCl₂ formulations. The resulting system yielded a discernible absorbance response for acephate at concentrations as low as 0.05 mg/kg, confirming the high sensitivity of the colorimetric reaction. To improve detection accuracy in the low-concentration range, a weighted partial least squares (wPLS) model coupled with leave-one-out cross-validation (LOOCV) was constructed using a reciprocal concentration weighting strategy. The effects of various combinations of three spectral ranges and five preprocessing methods on model performance were systematically evaluated. Optimal results for the prediction set were obtained within the 370–600 nm wavelength range, yielding a coefficient of determination (R2) of 0.9982 and a root mean square error of prediction (RMSEP) of 1.0963. Compared with the conventional PLS model, the proposed wPLS model exhibited distinct advantages in the low-concentration regime. Validation across five low-concentration groups revealed that the mean relative errors of the wPLS model were significantly lower than those of the PLS model. Spiked recovery tests performed in four grain matrices demonstrated recoveries ranging from 90.0% to 112.0%. This study establishes a reliable and safe tool for the on-site screening of acephate residues and provides valuable insights for the rapid detection of other sulfur-containing organophosphorus pesticides.