衰减全反射-中红外光谱法定量测定小麦粉中的石灰类添加物

摘要

In order to meet the need of on-site quick testing lime in wheat flour, the quantitative calibration models of CaO, Ca(OH)2, CaCO3, CaO+Ca(OH)2 and total lime were developed by using ATR-MIR spectroscopy combined with PLS algorithm, the external validations set were employed to validate the calibration models. The result showed that for the calibration models of CaO, Ca(OH)2, CaCO3, CaO+Ca(OH)2 and total lime, all of the determination coefficients (R2) were larger than 0.98, all of RMSEC (root mean square error of calibration) were less than 0.3, all of RMSECV (root mean square error of cross validation) were less than 0.5, all of the RMSEP (root mean square error of prediction) were less than 0.95, all of the RPD (ratio performance deviation) were more than 4.5, precision of the calibration models were higher and they could meet the needs of on-site quick testing lime in wheat flour. It can provide a reference for the rapid quality security screening of wheat flour, which is important for the quality control of wheat flour in the market. Meanwhile, it also will provide some references to the quantitative determination of other banned additives in wheat flour.%为了对市售小麦粉中的石灰类添加物进行现场快速检测,该文采用衰减全反射-中红外光谱(mid-infrared spectroscopy combined with attenuated total reflection,ATR-MIR)结合偏最小二乘算法建立了小麦粉中掺入氧化钙、氢氧化钙、碳酸钙、氧化钙+氢氧化钙以及石灰总量的定量校正模型,并采用外部验证集对各模型进行验证.结果表明,氧化钙、氢氧化钙、碳酸钙、氧化钙+氢氧化钙和石灰总量模型的决定系数均大于0.98;校正均方根误差(root mean square error of calibration,RMSEC)均小于0.3;交互验证均方根误差(root mean square error of cross validation,RMSECV)均小于0.5;外部验证的预测均方根误差(root mean square error of prediction,RMSEP)均小于0.95;各模型的相对预测性能(ratio performance deviation,RPD)均大于4.5,该模型具有较高的精度,可以满足小麦粉中石灰含量的现场快速检测要求.该研究可为市售小麦粉的快速质量安全筛查提供新的方法,对小麦粉市场质量监控具有重要意义,并且对小麦粉中其他违禁添加物的检测亦有参考价值.