【摘要】 用聚氨葡糖膜被动采集微量甲醛气体和高灵敏红外光谱分析仪,检测其中气体的光谱。再用多元散射校正（MSC）法,校正因膜厚薄不均导致光程异常,以了解光谱与甲醛吸附量（从10μg～105.2μg）与不同波数的关系。再用偏最小二乘（PLS）-去一交互检验法,在相关性好的光谱区间建模。结果表明:根据不同波数的关系,了解与甲醛含量变化密切相关的光谱区间,与聚氨葡糖膜的红外吸收谱带相对应。含有氨基吸收谱峰特征区间1315 cm^-1～1632 cm^-1建模优势明显,当隐变量为6时,最小预测误差为6.44μg,相关系数为0.9847。利用聚氨葡糖膜吸附甲醛气体的特性,建立的甲醛自动采样技术,有效采集微量甲醛样品,测定聚氨葡糖膜的红外光谱,了解甲醛浓度的变化,为快速测量室内微量甲醛气体奠定了基础。更多还原

【Abstract】 A well-prepared chitosan film was utilized in the passive sampling of formaldehyde.The change of the film after collecting formaldehyde was recorded by Fourier Transform Infrared（FTIR） spectroscopy.To reduce non-linearity caused by differences in path length between films,the IR spectra were pretreated by multiplicative scatter correction（MSC） method.The whole spectra were investigated to find special regions correlated with the amount of absorbed HCHO（from 10μg to 105.2μg） in the film.Leaving-one-out cross validation of partial least squares（PLS） was used to build model between spectra and concentrations of formaldehyde.Results indicate that the spectral region with the biggest correlation coefficient and the lowest errors of root mean square error of prediction（RMSEP） was located at 1315～1632 cm^-1,which covered the N-H group vibrations.When the number of latent variables was 6,the minimum of RMSEP was 6.44μg and the correlation coefficient was 0.9847.The method developed supplied a technique for simple and rapid passive sampling and IR determination of HCHO,which has potential application to quantitative analysis of HCHO by FTIR.更多还原