近红外光谱法快速测定烟草中的常规化学成分含量

Determination of Routine Chemical Components in Tobacco with Fourier Transform Near-infrared

[目的]探讨近红外光谱法快速测定烟草中的常规化学成分含量。[方法]采用近红外光谱技术,选取单品种样品681个,结合偏最小二乘法(PLS),定量分析了烟草中总氯、烟碱、总钾、总糖、还原糖及总氮含量,并用实际样品对模型进行了验证。[结果]使用偏最小二乘法(PLS)为建模方法,建立了烟草中6种常规化学成分:总氯、烟碱、总钾,总糖、还原糖及总氮的近红外预测模型。6种组分最佳PLS预测模型的相关系数r分别为0.977 4、0.992 7、0.982 1、0.986 0、099 1和0.975 0。交叉检验的均方差(RMSECV)分别为0.057、0.126、0.160、1.170、0.994和0.127。[结论]所建模型精密度良好,近红外光谱法与行业标准方法所测值不存在显著差异,近红外光谱模型可以快速预测烟草中总氯、烟碱、总钾、总糖、还原糖及总氮的含量。

[ Objective ] To discuss routine chemical components in tobacco with fourier transform near-infrared method. [ Method] Selecting 681 samples, partial least squares （PLS） was used to forecast the contents of total chlorine, nicotine, total potassium, total sugar, reducing sugar and total nitrogen in tobacco. [Result] The correlation coefficients for these 6 components were 0.977 4, 0.992 7, 0.982 1,0.986 O, 0.991 1 and O. 975 O, respectively. The root mean square error of cross validation（RMSECV） were 0.057,0. 126, 0. 160, 1. 170, 0. 994 and 0. 127, respec- tively. The accuracy of the model was satisfactory. [ Conclusion] The results showed that the two methods have no significant difference. It could be used to determine the contents of total chlorine, nicotine, total potassium, total sugar, reducing sugar and total nitrogen in tobacco rapidly.