MODELOS DE CALIBRACIÓN PARA LA CUANTIFICACIÓN ESPECTROFOTOMÉTRICA DE HIDROCARBUROS TOTALES DEL PETRÓLEO EN SUELO

Abstract

Calibration models used in analytical chemistry must be previously validated to achieve their aim. Generally, this is achieved by means of regression analysis. This study proposes strategies of internal and external validation of simple linear calibration models (LCM) to determine total petroleum hydrocarbons (TPH) by means of ultraviolet visible spectrometry (UV-Vis). In the crossed validation, the model is used to estimate one by one the reference concentrations with which it was constructed/calibrated, whereas in the external validation the model estimates reference concentrations that were not used in its calibration, which demonstrates its robustness. Four models were constructed and validated, demonstrating in all cases the linearity and robustness of the MCLS in the studied range. All the models were adequate for the spectrometric quantification of TPH in soils or in water extracted by different solvents, in quantities that comply with the maximum permissible limits established by Mexican regulations. In the calibrated models, the matrix effect was identified due to the type of solvent used to dilute the hydrocarbons and the type of hydrocarbon, which is influenced by the limits of detection and quantification.