Multi-channel anion sensing characteristics of a Ru(II) complex of the form [(bpy)2Ru (Hpzbzth)](ClO4)2 (1) {bpy = 2,2′-bipyridine and Hpzbzth = 3,5-bis(benzthiazol-2-yl)pyrazole} has been carried out in this work and analyzed through multiple machine learning tools. The complex possesses one pyrazole NH proton in its outer sphere which upon interaction with specific anions yields considerable change in its photo-redox behaviours. The anion sensing aspect of the complex is systematically investigated in acetonitrile through absorption, steady state and time-resolved emission spectroscopy as well as by square-wave (SWV) voltammetry. In absence of anions, the complex displays emission and represents the “on-state”, while complete quenching of emission takes place in presence of basic anions and designates the “off-state” of the system. The initial state of the complex could also be restored upon addition of acid and the process is reversible. Essentially, the complex acts as anion- and acid-induced molecular switches. The absorption, emission and electrochemical outputs of the complex in presence of anions and acid are employed to mimic different types of Boolean and Fuzzy logic (FL) functions. Additionally, we implemented herein different deep learning tools such as artificial neural networks (ANNs) and adaptive neuro-fuzzy inference system (ANFIS) to envisage the full anion sensing aspects of the complex. The outcomes of the three models are compared and also tallied with the experimental results.
Read full abstract