Rapid quantitative analysis of natural indigo dye content using near-infrared spectroscopy

Natural indigo, the most widely produced and utilized natural dye, encounters quality challenges due to the lack of standardization in the natural dye industry. Rapid determination of natural indigo dye contents before the dyeing process appears extremely important. In this study, two prediction models for different concentrations were established using partial least squares in conjunction with near-infrared analysis quantitatively to analyze the natural indigo dye content. A total of 228 indigo samples were collected from 14 different dyestuffs across various regions, with concentrations ranging from 100 to 1000 mg/L and 10 to 100 mg/L, respectively. The spectral pre-processing methods of multiplicative scatter correction plus first-order derivative and Savitzky–Golay smoothing plus band normalization plus first-order derivative were selected to enhance the model prediction accuracy. The optimized model exhibited excellent prediction accuracy. Within the concentration range of 100–1000 mg/L, the model has an R2 value of 0.9994, and a root mean square error of prediction value of 6.36 mg/L. In the concentration range of 10–100 mg/L, the model returned an R2 value of 0.9907, and a root mean square error of prediction value of 2.80 mg/L. The model's detection limit stands at 49.2 mg/L. The results demonstrated that the near-infrared models developed in this study can be used rapidly and accurately for the quantitative determination of natural indigo dyes.