摘要
烟用香精香料成分较多、来源复杂,有必要对其质量进行评价以确保不同批次产品质量稳定。现行标准规定采用气相色谱法对乙醇、1,2-丙二醇进行测量,但方法缺陷较多。近红外光谱法(NIR)可有效克服气相色谱法所存在的缺陷。采用NIR获取烟用香精香料的光谱数据,使用偏最小二乘法和支持向量机分别建立乙醇和1,2-丙二醇的定量模型。研究结果发现:四个定量模型的Q_(cv)^(2)和Q_(pre)^(2)值都大于0.98;对于乙醇而言,两种多元校正算法的RMSECV和RMSEP值都不超过0.02;模型参数验证了NIR可快速测定烟用香精香料中乙醇、1,2-丙二醇含量具有可行性,这为以后一张光谱图预测多项理化指标数值打下坚实基础。
Tobacco flavors and fragrances have various components and their sources are complex.It is necessary to evaluate their quality to ensure quality stability between different batches.The current standard stipulates the use of gas chromatography for determination of ethanol and 1,2-propylene glycol.However,the method has various defects.Near-infrared spectroscopy(NIR)can effectively overcome the defects of gas chromatography.In this work,NIR was used to obtain spectral data of tobacco flavors and fragrances.Partial least square and support vector machine were used to establish quantitative models of ethanol and 1,2-propylene glycol,respectively.The results show that Q_(cv)^(2)and Q_(pre)^(2)values of the four quantitative models are greater than 0.98.In terms of ethanol,RMSECV and RMSEP values of the two multivariate calibration algorithms are no more than 0.02.Models performances verify the feasibility of rapid determination of ethanol and 1,2-propylene glycol by NIR and it can be realized that one spectrum is for rapid measurement of various physical-chemical indicators.
作者
黄扬明
王瑶
何媛
彭军仓
闵顺耕
HUANG Yang-ming;WANG Yao;HE Yuan;PENG Jun-cang;MIN Shun-geng(College of Science,China Agricultural University,Beijing 100193,China;Technology Center,Shaanxi Tobacco Industry Co.,Ltd.,Baoji 721013,China)
出处
《光谱学与光谱分析》
SCIE
EI
CAS
CSCD
北大核心
2023年第S01期89-90,共2页
Spectroscopy and Spectral Analysis
基金
国家自然科学基金项目(31301685)资助
关键词
近红外光谱法
烟用香精香料
乙醇
1
2-丙二醇
Near-infrared spectroscopy
Tobacco flavors and fragrances
Ethanol
1,2-Propylene glycol
