关键词: 红外光谱, 多元线性回归, 木塑复合材料(WPC), 定量分析, 木粉

Abstract: The Chinese fir/polypropylene(PP) composites samples were analyzed by FT-IR with KBr pellets method. The characteristic adsorption peaks of Chinese fir were assigned to 1740-1730 cm^-1, 1610-1590 cm^-1, 1270-1260 cm^-1, 1060-1050 cm^-1, and 1040-1030 cm^-1. The peak of PP at 1377 cm^-1 was taken as reference. Correlation analysis was performed between wood flour content and the relative intensities of different spectral peaks and the stepwise multiple linear regression was used to establish the equations. The results showed that the binary linear regression equation with I_(1060-1050)/I₁₃₇₇ and I_(1270-1260)/I₁₃₇₇ as variables and the ternary linear regression equation based on the variables I_(1060-1050)/I₁₃₇₇, I_(1040-1030)/I₁₃₇₇ and I_(1270-1260)/I₁₃₇₇ possessed better prediction accuracy. There is a strong correlation between FT-IR-predicted wood flour content and referenced wood flour content. The coefficients of determination (R_c²) of calibration exceed 0.98, and the R_p² of cross validation were above 0.96. The results of external validation showed that linear regression equations had good predictabilities, and the relative deviations of prediction ranged from -7.4% to -0.9%. The results also showed that the prediction accuracy of the ternary linear regression equation was slightly better than that of the binary linear regression equation.

Key words: FT-IR, multiple linear regression, wood plastic composites (WPC), quantitative analysis, wood flour