根据58个丙烯腈-丁二烯-苯乙烯(ABS)树脂样品的近红外漫反射光谱数据结合偏最小二乘法(PLS)建立了ABS树脂中氮含量定量分析的校正模型。结果表明:在近红外光谱区(7 501.8~5 449.9 cm-1)内,主因子数为9时,定量模型的准确度最好。通过对模型进行t-检验,在显著性水平为0.05的条件下,其测定结果与标准方法的测定结果对比,两者无显著性差异。预测相关系数(r)和相对标准偏差(RSD)分别为0.992 4和3.3%。

Base on the spectral data of IR diffuse reflection of 58 samples of ABS resin (copolymer of acrylonitrile, butadiene and styrene) and on the application of calibration method of partial least square, a calibration model for determination of nitrogen in the resin was established. As shown by experimental results, highest accuracy was attained in the NIR spectral region of 7 501.8 to 5 449.9 cm-1 when taking 9 as the principal factor number. In the significance test of the model by t-test at the significance level of 0.05, no significant difference was found between the results found by the present method and the results found by the standard method. Values of predictive correlation coefficient (r) and relative standard deviation (RSD, n=7), found were 0.992 4 and 3.3% respectively.