轻涂纸涂层中羧基丁苯胶乳含量的测定

摘要

In this paper,the styrene-butadiene latex (SB-latex) contents in the coating layer of light weight coated (LWC) paper were measured by using an attenuated total reflectance-Fourier transform infrared spectroscopy (ATR-FTIR).The peak areas under the specific absorption peaks at 2922 cm-1,1028 cm-1,756 cm-1,and 699 cm-1 were calculated,by using the relative software.So the relation between different contents of SB-latex and the peak areas under those four peaks was obtained.Three different models for predicting the SB-latex content in the coating layer of LWC paper were established by applying the polynomial curve fitting (PCF),partial least-squares regression (PLS regression),and generalized regression neural network (GRNN),respectively.The results showed that the attenuated total reflectance-Fourier transform infrared spectral information of the coating layer of LWC paper can be used to determine the SB-latex content in the coating layer.Among these three methods of predicting the content of SB-latex in the coating layer,the GRNN model has the best predicting effect since it has the minimum error,i.e.,the maximum error is-3.39％.%采用衰减全反射傅里叶变换红外光谱(ATR-FTIR)对涂有不同羧基丁苯胶乳含量的轻涂纸表面进行测定,选取羧基丁苯胶乳在波数为2922 cm-1、1028 cm-1、756 cm-和699 cm-1 4处特征吸收峰并使用红外光谱软件计算其峰面积,由此得到不同胶乳含量与这4处吸收峰峰面积之间的关系；分别采用多项式曲线拟合(PCF)、偏最小二乘回归(PLS regression)和广义回归神经网络(GRNN)进行建模,并对3种模型的预测效果进行比较.结果表明,轻涂纸涂层中的ATR-FTIR信息可以用来测定轻涂纸涂层中的羧基丁苯胶乳含量,3种预测模型中,GRNN方法具有最佳的预测效果,偏差最小(最大偏差为-3.39％).