Laboratory of Chemistry, Department of Science, Agricultural University of Athens, 75 Iera Odos, 11855 Athens, Greece.
J Agric Food Chem. 2010 May 26;58(10):6011-7. doi: 10.1021/jf100143n.
The feasibility of Raman spectroscopy for predicting the content of crocetin esters (crocins), and coloring strength was assessed. 114 samples from Greece, Iran, Italy and Spain were divided into two sets: a calibration set with 49 samples and a validation one with 65 samples. Calibration models for crocetin esters (r 0.97, RMSEC 0.92, RMSEP 0.97, RPD 3.46) and coloring strength (r 0.95, RMSEC 12.2, RMSEP 11.3, RPD 2.59) were built in the spectral region 1700-955 cm(-1) using partial least-squares (PLS) regression. The calibration models were validated using cross-validation, leaving one sample out (r 0.97, RMSECV 1.09 for crocetin esters and r 0.93, RMSECV 14.5 for coloring strength). The crocetin esters content as determined by liquid chromatography fluctuated between 18.8 and 31.7 mg/100 g saffron. The corresponding values, as calculated using the Raman method, fluctuated between 19.2 and 32.0 mg/100 g saffron. The coloring strength determined by the reference method ranged from 177.0 to 296.7 units, while with the Raman method the values were between 186.8 and 297.6 units. The results, as compared to the reference methods (liquid chromatography and UV-vis spectrophotometry), show that the proposed methodology gives data with acceptable accuracy. The proposed models can be used as a tool for rapid screening of quality in saffron samples.
本研究旨在评估拉曼光谱法预测藏红花中西红花酸酯(西红花苷)含量和色价的可行性。从希腊、伊朗、意大利和西班牙采集了 114 个样本,分为校准集(49 个样本)和验证集(65 个样本)。在 1700-955 cm(-1) 光谱范围内,采用偏最小二乘法(PLS)回归建立了西红花酸酯(r 0.97,RMSEC 0.92,RMSEP 0.97,RPD 3.46)和色价(r 0.95,RMSEC 12.2,RMSEP 11.3,RPD 2.59)的校准模型。通过交叉验证(每次留出一个样本)验证校准模型(r 0.97,RMSECV 1.09 用于西红花酸酯,r 0.93,RMSECV 14.5 用于色价)。用液相色谱法测定的西红花酸酯含量在 18.8-31.7 mg/100 g 藏红花之间波动,用拉曼法计算的相应值在 19.2-32.0 mg/100 g 藏红花之间波动。用参考方法(分光光度法和紫外可见分光光度法)测定的色价范围为 177.0-296.7 单位,而用拉曼法测定的色价范围为 186.8-297.6 单位。与参考方法(液相色谱法和紫外可见分光光度法)相比,结果表明该方法具有较高的准确性。所建立的模型可用于快速筛选藏红花样品的质量。