Calibration of a Photometer for Quantification of Ethanol in Gasoline: Absorbance Determination

This paper presents a study about the multivariate calibration of a LED near-infrared photometer applied to the quantification of ethanol in gasoline. Based on the analysis of works of other authors, we implemented seven different methods for determination of absorbances from the raw optical data, which were used for calibration of the prototype. The methods either use different experimental conditions for the background reference measurement: with empty cuvettes, with cuvettes containing a reference sample, or without cuvettes; or are independent of a reference measurement. The results obtained by each method were compared in terms of the root-mean-square errors, coefficients of determination, and residual prediction deviations (RPDs). We discussed characteristics of each method and concepts regarding the modeling of the instrumentation system, spectral preprocessing, and calibration using multiple linear regression, which are applicable both to other photometers and to other types of substances. Experimental results showed that the accuracy of the equipment in quantifying the analyte is determined not only by the accuracy of the light measurement system, but also by the processing of the optical data.