氯盐近红外光谱分析在生物医学上有着十分重要的意义.在室温下对氯化钠,氯化钾,氯化钙水溶液的近红外光谱采集分析表明,氯盐溶液浓度的变化会影响水分子的氢键而使得水溶液的近红外光谱发生变化.通过选择合理波长区间以及温度变化影响为零的波长点(等吸光点),减小温度的干扰对溶液近红外光谱影响,采用偏最小二乘法建立了氯盐溶液的浓度模型用于预测氯盐的离子浓度.分析氯盐溶液中氯盐阳离子的离子半径大小、离子电荷数和离子在水中络合效应等因素对水的近红外光谱所造成的影响以及产生影响因素的原因.实验结果表明,温度和浓度都会影响溶液的近红外光谱,氯盐溶液浓度较低时,温度的影响占主导地位;氯盐浓度高时,浓度的影响占主导地位.氯盐在水溶液中形成络合物,并与氯盐阳离子共同作用对水的氢键产生影响,对于浓度相同种类不同的氯盐溶液,形成的络合物和阳离子破坏效应对水的氢键的破坏作用为:CaCl2>NaCl>KCl.最终建立的样品浓度模型校正集的决定系数(R2)=99.97%,交叉验证均方误差(RMSECV)=4.51,剩余预测偏差(RPD)=62.7,满足日常生化检测精度要求。

Determination of chloride salt solution by near infrared spectrum plays a very important role in Biomedicine.The near infrared spectrum analysis of Sodium chloride,potassium chloride,calcium chloride aqueous solution shows that the concentration change of chloride salt can affect hydrogen bond,resulting in the variation of near infrared spectrum of water.The temperature influence on NIR spectrum has been decreased by choosing reasonable wavelength range and the wavelength where the temperature effects are zero(isosbestic point).Chlorine salt prediction model was established based on partial least squares method and used for predicting the concentration of the chlorine ion.The impact on near infrared spectrum of the cation ionic radius,the number of ionic charge,the complex effect of ionic in water has also discussed in this article and the reason of every factor are analysed.Experimental results show that the temperature and concentration will affect the near-infrared spectrum of the solution,It is found that the effect of temperature plays the dominant role at low concentrations of chlorine salt;rather,the ionic dominates at high concentration.Chloride complexes are formed in aqueous solution,It has an effect on hydrogen bond of water combining with the cations in chlorine salt solution,Comparing different chloride solutions at the same concentration,the destruction effects of chloride complexes and catnions on the hydrogen bond of water increases in the sequences:CaCl2>NaCl>KCl.The modeling result shows that the determination coefficients(R2)=99.97%,the root mean square error of cross validation(RMSECV)=4.51,and the residual prediction deviation(RPD)=62.7,it meets the daily requirements of biochemical detection accuracy.