Ultraviolet-visible absorption spectrometry is used for the determination of many inorganic, organic, and biological species. However, poor selectivity limits its applications for quantitative analysis in complex mixtures. In this work, its mathematical selectivity was explored using trilinear decomposition on spectral-pH measurements for the selective determination of two flavonoids in three Chinese herbal medicines. Through introducing a pH mode, second-order spectral pH data were constructed per sample. By using trilinear decomposition on the calibration and prediction samples, the rutin signals were extracted from lonicerae japonicae flos, perillae folium, and perillae fructus samples, respectively. Novel second-order calibration was developed for interference-free determination of rutin in these medicines followed by univariate regression of the decomposed intensity versus concentration. The predicted 95% confidence intervals for the concentration of rutin in these samples were 5.19 ± 0.16, 5.81 ± 0.10, and 7.00 ± 0.36 μg mL⁻¹, respectively. These results were validated by high performance liquid chromatography (HPLC). Similarly, interference-free determination of isorhamnetin was achieved in these materials. By increasing the selectivity of ultraviolet-visible spectroscopy with trilinear decomposition, the requirements for sample preparation and chemical separation are reduced. The developed spectral pH second-order calibration allows direct analysis of complex samples by ultraviolet-visible absorption spectrometry due to the mathematical selectivity.

Ultraviolet-visible absorption spectrometry is used for the determination of many inorganic, organic, and biological species. However, poor selectivity limits its applications for quantitative analysis in complex mixtures. In this work, its mathematical selectivity was explored using trilinear decomposition on spectral-pH measurements for the selective determination of two flavonoids in three Chinese herbal medicines. Through introducing a pH mode, second-order spectral pH data were constructed per sample. By using trilinear decomposition on the calibration and prediction samples, the rutin signals were extracted from lonicerae japonicae flos, perillae folium, and perillae fructus samples, respectively. Novel second-order calibration was developed for interference-free determination of rutin in these medicines followed by univariate regression of the decomposed intensity versus concentration. The predicted 95% confidence intervals for the concentration of rutin in these samples were 5.19 ± 0.16, 5.81 ± 0.10, and 7.00 ± 0.36 μg mL⁻¹, respectively. These results were validated by high performance liquid chromatography (HPLC). Similarly, interference-free determination of isorhamnetin was achieved in these materials. By increasing the selectivity of ultraviolet-visible spectroscopy with trilinear decomposition, the requirements for sample preparation and chemical separation are reduced. The developed spectral pH second-order calibration allows direct analysis of complex samples by ultraviolet-visible absorption spectrometry due to the mathematical selectivity.

Supplemental Table 1. Qualitative compounds using commercial standards and databases; Supplemental table 2 Differential metabolites of tobacco samples beteween different cultivars and geographic origins.; Supplemental table 3 Correlation analysis between the identified differential metabolites.

Supplemental Table 1. Qualitative compounds using commercial standards and databases; Supplemental table 2 Differential metabolites of tobacco samples beteween different cultivars and geographic origins.; Supplemental table 3 Correlation analysis between the identified differential metabolites.