红外光谱结合偏最小二乘法定量分析毛竹/聚丙烯复合材料的主成分

木塑复合材料(wood plastic composites, WPC)中生物质和塑料的比例影响其物理力学性能和价格。建立一种快速、准确的方法预测WPC中生物质和塑料的含量对于WPC市场的进一步发展具有重要作用。现有的检测方法主要为热分析法，然而，热分析法固有的缺陷(包括检检测时间长、测精度低、操作复杂等)严重限制了其应用范围。为此，本研究采用红外光谱(FTIR)结合偏最小二乘法(PLS)对毛竹/聚丙烯(PP)复合材料样品中毛竹及PP的含量进行了快速测定。以毛竹为生物质填料、PP为基体材料，同时加入一定量的添加剂，采用挤出成型法制备了42个不同毛竹/PP比例的WPC样品。采用KBr压片法收集42个WPC样品的红外光谱数据，利用PLS-2和完全交互验证方式建立样品中毛竹及PP含量和光谱数据间的相关性模型。内部交互验证结果表明，对原始光谱进行一阶导数和SNV预处理后，选择1 800～800 cm-1 波段建立的模型性能最佳。毛竹和PP含量的校正模型决定系数R2均为0.955，校正标准偏差SEC分别为1.827和1.848。毛竹和PP含量的预测模型决定系数R2均为0.950，交互验证标准偏差SECV分别为1.927和1.950，RPD值均为4.45。外部验证结果表明，毛竹和PP含量相对预测偏差均低于6%，FTIR结合PLS法可以同时快速、准确地预测毛竹/PP复合材料中毛竹及PP含量。

The Use of FTIR Coupled with Partial Least Square for Quantitative Analysis of the Main Composition of Bamboo/Polypropylene Composites

The biomass to plastic ratio in wood plastic composites (WPCs) greatly affects the physical and mechanical properties and price .Fast and accurate evaluation of the biomass to plastic ratio is important for the further development of WPCs .Quanti‐tative analysis of the WPC main composition currently relies primarily on thermo‐analytical methods .However ,these methods have some inherent disadvantages ,including time‐consuming ,high analytical errors and sophisticated ,which severely limits the applications of these techniques .Therefore ,in this study ,Fourier Transform Infrared (FTIR) spectroscopy in combination with partial least square (PLS) has been used for rapid prediction of bamboo and polypropylene (PP) content in bamboo/PP compos‐ites .The bamboo powders were used as filler after being dried at 105 ℃ for 24 h .PP was used as matrix materials ,and some chemical regents were used as additives .Then 42 WPC samples with different ratios of bamboo and PP were prepared by the methods of extrusion .FTIR spectral data of 42 WPC samples were collected by means of KBr pellets technique .The model for bamboo and PP content prediction was developed by PLS‐2 and full cross validation .Results of internal cross validation showed that the first derivative spectra in the range of 1 800~800 cm -1 corrected by standard normal variate (SNV) yielded the optimal model .For both bamboo and PP calibration ,the coefficients of determination (R2 ) were 0 .955 .The standard errors of calibra‐tion (SEC) were 1 .872 for bamboo content and 1 .848 for PP content ,respectively .For both bamboo and PP validation ,the R2 values were 0 .950 .The standard errors of cross validation (SECV) were 1 .927 for bamboo content and 1 .950 for PP content , respectively .And the ratios of performance to deviation (RPD) were 4 .45 for both biomass and PP examinations .The results of external validation showed that the relative prediction deviations for both biomass and PP contents were lower than ± 6% .FTIR combined with PLS can be used for rapid and accurate determination of bamboo and PP content in bamboo /PP composites .