Diferentes estrategias para la generación y modelado de datos de segundo orden. Aplicaciones para la resolución de diferentes problemas analíticos

LUCIANA VERA-CANDIOTI; YAMILE S. CARO; MARÍA M. DE ZAN; ROMINA BRASCA; MIRTA ALCARÁZ; MATÍAS MARCHISIO; FLORENCIA PICEH; MARÍA CÁMARA; AGUSTINA V. SCHENONE; MARÍA J. CULZONI; HÉCTOR C. GOICOECHEA

Santa Fe

Feria; 40 años FBCB. Vení a conocerla a la Estación; 2012

Facultad de Bioquímica y Cs. Biológicas (UNL)

Second-order data enclose the so-called second-order advantage, which allows predicting the concentration of the analyte of interest even in the presence of unknown interferents, as well as enabling several analytes to be determined simultaneously [1].In this report, results for several experimental data sets are presented in order to show the great potentiality of the second order data modeled with convenient algorithms to solve different analytical problems. They present the following challenges to second-order algorithms: 1) linear dependency due to a kinetic reaction in one mode, 2) peak shifts (CE data), and 3) non-linearity. In all of these cases, deviations from the ideal trilinearity are likely to occur due to changes in component profiles from sample to sample. Data set 1 involves five fluoroquinolones which are determined in environmental samples (i.e., river water) by using capillary electrophoresis with diode array detection. Multivariate curve resolution with alternating least squares (MCR-ALS) without pretreatment outperformed parallel factor analysis (PARAFAC) and partial least squares followed by residual bilinearization (PLS/RBL) in profiles extraction and quantitation of the five analytes. Data set 2 includes fluorescence-time measurements made by creating a gradient within a flow injection system. These second order data were applied to resolve mixtures of two antihistaminic drugs (loratadine and desloratadine) in serum samples. The use of surfactant was necessary to obtain the selectivity to differentiate both spectra. This kind of data present the problem of complete overlapping of profiles in one data dimension, which can be regarded as a special and serious case of linear dependency [2]. The strategy involves the building of a MCR-ALS model composed of matrices augmented in the temporal mode, i.e. spectra remain invariant while time profiles may change from sample to sample.In Data set 3, fluorescence-time data were studied for the oxidation reaction of loratadine and desloratadine with potassium bromate, in order to determine both drugs in human serum samples. Linear dependency in both temporal and spectral modes precluded the use of MCR-ALS and PLS/RBL, providing better results than PARAFAC. Data set 4 consists of fluorescence-time data obtained for the oxidation reaction of three dyes with potassium bromate. The possibility of exploiting the second-order advantage from these non-linear second-order data could be reached by the application of two successive methods: the first one modeled the calibration and validation data removing the contribution of unexpected components, and the second one models the non-linear relationship. MCRALS was the only strategy that retrieved reasonably accurate predictions.