绞股蓝水分近红外定量模型的建立

Establishing the Quantitative Model of Moisture Content in Gynoste Pentaphylli Herba Seu Radix Mak by Near Infrared Spectroscopy

目的：运用近红外光谱技术对绞股蓝药材中水分进行快速测定.方法：采集不同产地不同批号绞股蓝样品的近红外光谱图,结合TQ Analyst 8.0软件,将其经一阶导数及Norris平滑（有效位数3,有效位间间隔3）预处理,在10 000 ～4 000cm-谱段内,选择前10个主成分,采用偏最小二乘法建立绞股蓝水分近红外光定量模型.结果：所建模型相关系数r=0.9688,交叉验证均方根偏差（RMSECV）=0.303,预测均方根偏差（RMSEP） =0.284,验证集预测回收率为101.14％（n=11）,RSD2.16％,并且仪器精密度良好,样品重复性佳,在2h内样品检测稳定.结论：该研究所建立的模型性能较好,绞股蓝水分含量测定较准确.可以应用于绞股蓝药材水分的快速测定.

Objective：This study aimed at establishing the quantitative model of moisture content in Gynoste Pentaphylli Herba Seu Radix by near infrared reflectance spectroscopy （NIRS）.Method：The near infrared spectra of Gynoste Pentaphylli Herba Seu Radix with different batch numbers from various areas were collected,and were subject to the analysis by TQ Analyst 8.0 software.Near infrared spectra of all the samples were pretreated with First Derivative and Norris filter （3 segment length and 3 gap between segments）.The wavebands containing 10 000-4 000 cm-1 were collected and the first 10 factors were used.The calibration model was built with partial least squares （PLS）.Result：The quantitative calibration model had good correlation coefficients （R） and low root mean square error of cross validation （RMSECV）,with a r value equal to 0.968 8,RMSECV value equal to 0.303 and root mean square error of prediction （RMSEP） value equal to 0.284.The average rate of recovery of validation was 101.14％ （n =11）,RSD 2.16％.At the same time,NIR instrument was good in precision.The samples had great performance in reproducibility and stability within 2 h.Conclusion：The quantitative calibration model of moisture content in Gynoste Pentaphylli Herba Seu Radix was proven to be credible and could be used to predict the moisture content in Gynoste Pentaphylli Herba Seu Radix accurately.The simple,fast and non-destructive advantages of NIRS can be used for quality control and exploitation of Gynoste Pentaphylli Herba Seu Radix.