Feasibility of digital image colorimetry--application for water calcium hardness determination.

Interpretation and relevance of basic RGB colors in Digital Image-Based Colorimetry have been treated in this paper. The studies were carried out using the chromogenic model formed by the reaction between Ca(II) ions and glyoxal bis(2-hydroxyanil). It produced orange-red colored solutions in alkaline media. Individual basic color data (RGB) and also the total intensity of colors, I(tot), were the original variables treated by Factorial Analysis. Te evaluation evidenced that the highest variance of the system and the highest analytical sensitivity were associated to the G color. However, after the study by Fourier transform the basic R color was recognized as an important feature in the information. It was manifested as an intrinsic characteristic that appeared differentiated in terms of low frequency in Fourier transform. The Principal Components Analysis study showed that the variance of the system could be mostly retained in the first principal component, but was dependent on all basic colors. The colored complex was also applied and validated as a Digital Image Colorimetric method for the determination of Ca(II) ions. RGB intensities were linearly correlated with Ca(II) in the range 0.2-2.0 mg L(-1). In the best conditions, using green color, a simple and reliable method for Ca determination could be developed. Its detection limit was established (criterion 3s) as 0.07 mg L(-1). And the reproducibility was lower than 6%, for 1.0 mg L(-1) Ca. Other chromatic parameters were evaluated as dependent calibration variables. Their representativeness, variance and sensitivity were discussed in order to select the best analytical variable. The potentiality of the procedure as a field and ready-to-use method, susceptible to be applied 'in situ' with a minimum of experimental needs, was probed. Applications of the analysis of Ca in different real water samples were carried out. Water of the city net, mineral bottled, and natural-river were analyzed and results were compared and evaluated statistically. The validity was assessed by the alternative techniques of flame atomic absorption spectroscopy and titrimetry. Differences were appreciated but they were consistent with the applied methods.