On the calibration of chemical sensors based on photoluminescence: Selecting the appropriate optimization criterion

In this work, we demonstrate the importance of the optimization criterion used for calibrating photoluminescence-based chemical sensors. To demonstrate this claim, we have considered three different calibration criteria (criterion (a): minimization of the square error in the analytical signal, criterion (b): minimization of the square error in the concentration, and criterion (c): minimization of the square relative error in the concentration) as well as three classical calibration models (Stern-Volmer, Lehrer and Demas). We demonstrate that different criteria provide different calibration parameters, and therefore the criterion has to be properly selected for accurately calibrating an optical sensor based on photoluminescence. These claims have been demonstrated by using experimental results from three real experiments, involving three different oxygen-sensitive sensing phases characterized by luminescence intensity and luminescence lifetime measurements with three different instruments (a luminescence spectrometer, a dual-phase lock-in amplifier, and a measuring system based on the I/Q method). In addition, we provide details for the implementation of the calibration procedures as well as code (compatible with matlab and Octave) that performs the calibration for each model and criterion. In our opinion, the supplementary material provided in this work should be of great help to scientists who perform calibration procedures of photoluminescent chemical sensors.