An O-vanillin scaffold as a selective chemosensor of PO43− and the application of neural network based soft computing to predict machine learning outcomes

Abstract

O-Vanillin derived Schiff base 1-[(E)-(2-hydroxy-3-methoxybenzylidene) amino]-4-methylthiosemicarbazone (VCOH) has been synthesized for colorimetric and fluorescence chemosensors towards PO₄³⁻ ions. A fluorescence ‘turn-on’ sensing mechanism of VCOH towards PO₄³⁻ ions has been explained due to emission from the VCO⁻ ion formed upon transfer of the phenolic proton of VCOH to a PO₄³⁻ ion. The 1 : 1 stoichiometry between the VCOH probe and PO₄³⁻ ion is confirmed by Job's plot based on UV-vis titration. The limit of detection (LOD) of VCOH towards PO₄³⁻ ions is found to be 0.49 nM. The PO₄³⁻ ion sensing property of probe VCOH has been applied to prepare portable paper strips and for the analysis of real water samples. Fluorescence ‘turn-on’ and ‘turn-off’ responses of VCOH towards PO₄³⁻and H⁺ respectively have been used to construct a molecular logic gate. Fluorescence based sensing studies in which the concentration of analytes is adjusted over a broad range can be both laborious and expensive. In order to address these challenges, we have utilized various soft computing methods, including artificial neural networks (ANN), fuzzy logic (FL), and adaptive neuro-fuzzy inference systems (ANFIS), to appropriately model the ‘turn-on’ and ‘turn-off’ behaviors of the VCOH probe upon addition of PO₄³⁻ and H⁺ respectively as well as to predict the experimental sensing data.