多元曲线分辨-交替最小二乘方法在联用数据分辨中的应用

Application of Multivariate Curve Resolution-Alternating Least Square on Analysis of Hyphenated-Instrument Data

通过对模拟数据和高效毛细管电泳实验数据的分析,讨论了多元曲线分辨-交替最小二乘方法(MCR-ALS)在毛细管电泳-二极管阵列检测(CE-DAD)联用数据分辨中的应用。讨论了几种因素对MCR-ALS单个数据矩阵分辨结果的影响,包括待分析物光谱间的相似程度、浓度曲线的重叠程度以及由渐进因子分析(EFA)所得到的浓度初始值等。MCR-ALS还可用于多个数据矩阵的同时分析,即二阶MCR-ALS。结果表明,与一阶MCR-ALS相比,二阶MCR-ALS方法能够更好地解决各种分辨问题,得到合理和满意的分辨结果。

Study is concerned with application of multivariate curve resolution with alternating least squares （MCR-ALS） methods to both simulated and experimental data. The experimental data are from capillary electrophoresis with diode array detector （CE-DAD）. Factors that affect resolution when MCR-ALS is applied to single matrix are discussed, including spectra similarities of components, type of overlapping peaks and initials from evolving factor analysis （EFA）. MCR-ALS is also used when several matrices are analyzed simultaneously and in this case, the problem such rank-deficiency can be overcome.