Abstract

The determination of molecular oxygen is of great interest in numerous fields ranging from biology, biotechnology, medicine, and chemistry. One of the most diffused optical methods is based on the quenching of luminescence of a luminophore, like Pt-TFPP, by the oxygen molecules. Since both the luminescence and the quenching are temperature dependent, the temperature of the indicator has to be continuously monitored and accounted for in the determination of the oxygen concentration. In this work, a new approach based on artificial neural networks is proposed. The neural network developed learns to predict the oxygen without any information about the temperature of the luminophore. The prediction of the neural network, in this case, the oxygen concentration, is, therefore, temperature immune. Additionally, the neural network learns from the cross-interference to predict also the temperature, making positionally accurate and fast temperature measurements not necessary anymore. This work shows how it is possible to extract a temperature-immune oxygen concentration using Pt-TFPP and a single optical channel for the measurement.

© 2021 The Author(s)