基于近红外光谱法快速评价功能香基质量稳定性研究

Study on Rapid Evaluation of Quality Stability of Functional Fragrance Base by Near Infrared Spectroscopy

为提高功能香基质量稳定性评价的时效性,基于近红外光谱数据建立功能香基的质量稳定性监控模型,并对监控模型的预测性能进行检验。结果表明:(1)利用方差光谱法优选出一阶导数和Norris平滑为功能香基GNA近红外光谱预处理方法;4000~4296、4443~5098、5268~7247 cm^(-1)为近红外特征光谱范围,经预处理后光谱信噪比增强。(2)功能香基GNA各合格样品与标准品的欧氏距离、全成分斜率相似度指标值分别服从正态分布N(0.755,0.114)、N(0.921,0.001)。(3)根据3σ控制原则,建立基于欧氏距离、全成分斜率相似度的近红外监控模型,监控模型对未知样品的判别结果与实际结果一致,对稀释、掺杂等不良样品的识别准确率达到100%。(4)相比较而言,全成分斜率相似度评价指标对稀释程度和掺杂比例的反映更加灵敏。基于近红外光谱的功能香基质量稳定性监控模型能够快速、准确地判别合格样品及稀释或掺杂的不良样品,适于大批量香基样品的快速监测。

In order to improve the timeliness of the quality stability evaluation of functional fragrance base,a quality stability monitoring model of functional fragrance base was established based on near-infrared(NIR)spectral data,and the predictive performance of the monitoring model was examined.The results showed that:(1)The first derivative and Norris smoothing were optimized by variance spectroscopy as the NIR spectral pretreatment method for functional GNA;4000~4296 cm^(-1),4443~5098 cm^(-1) and 5268~7247 cm^(-1) were the characteristic spectral ranges of NIR,and the spectral signal-to-noise ratio was enhanced after pretreatment.(2)The euclidean distance and full-component slope similarity index values of each qualified sample of functional fragrance base GNA and the standard obeyed the normal distribution N(0.755,0.114)and N(0.921,0.001),respectively.(3)According to the 3σcontrol principle,the NIR monitoring model based on euclidean distance and full-component slope similarity was established,and the discrimination results of the monitoring model on the unknown samples were consistent with that of the actual results,and the identification accuracy rates of dilute,doped and other undesirable samples were 100%.(4)In comparison,the similarity evaluation index of full-component slope value was more sensitive to dilution degree and doping ratio.The functional fragrance base quality stability monitoring model based on NIR spectroscopy can quickly and accurately discriminate qualified samples and diluted or adulterated defective samples,and is suitable for rapid monitoring of large quantities of functional fragrance base samples.