关键词: 近红外光谱, 温度效应, 直链烷烃, 交替三线性分解, 定量光谱温度关系(QSTR)

Effect of temperature on near-infrared (NIR) spectra has been studied and applied to structural and quantitative analyses. To investigate the effect of temperature on NIR spectra of alkyl organic system, n-alkanes were studied in this work. NIR spectra of pure n-alkanes (hexane to decane), binary (hexane and octane) and ternary (octane, nonane and decane) mixtures were measured. In the experiments, temperature was controlled to change from 60 to 20℃ with a step of ca. 5℃. Comparing the spectra at different temperatures, only a little difference in peak intensity of some bands can be found. Therefore, alternating trilinear decomposition (ATLD) algorithm was adopted to analyze the three-order data matrix. The results show that two spectral loadings are obtained because the influence of temperature on the spectra of terminal ethyl (C₂H₅) groups differs from that of mid-chain methylene (CH₂) groups. Furthermore, the temperature scores of CH₂ and C₂H₅ groups decrease linearly with temperature, implying that the temperature effect can be quantitatively described by a quantitative spectra-temperature relationship (QSTR) model. The QSTR model provides an efficient way to predict the temperature of n-alkane solutions. Good linearity also exists between sample scores and carbon number or the relative content of CH₂ and C₂H₅ groups in the molecules of the n-alkanes. Linear models between the two scores and the relative content of CH₂ and C₂H₅ groups are obtained, respectively, using the least square fitting of the score and the relative contents. The model can be used for prediction of the relative content of CH₂ and C₂H₅ groups in mixtures, which can further be used to estimate the composition of the mixtures. Furthermore, the relationship between the scores and the carbon atom numbers is modeled using multivariate linear regression (MLR). The composition of n-alkane mixtures can also be estimated through the predicted carbon number using the MLR model. These models are validated by binary and ternary mixtures of the n-alkanes. It was indicated that the relative contents of CH₂ and C₂H₅ groups or the carbon atom number can be predicted using the models. Therefore, a new way for quantitative estimation of the composition in n-alkane mixtures was developed using the temperature effect of the near-infrared spectra.

Key words: near-infrared spectra, temperature effect, n-alkane, alternating trilinear decomposition, quantitative spectratemperature relationship