Development of a Voltammetric Methodology Based on an Artificial Neural Network for the Quantification of Sodium Diclofenac in Pharmaceutical Samples

Abstract

Sodium diclofenac is a widely used anti-inflammatory drug that can cause heart diseases if consumed constantly in high doses. Consequently, it is essential to have strict control of the amounts of this active principle in pharmaceutical products. The combination of electroanalytical techniques with advanced chemometrics has risen as a viable alternative for the exact and precise determination of active principles even in the presence of chemical interferences. In this research, an artificial neural network (ANN) for the voltammetric quantification of diclofenac in the presence of paracetamol, pyridoxine, and caffeine is presented, using a carbon paste electrode modified with multilayer carbon nanotubes and titanium dioxide nanoparticles. Cyclic voltammetry is performed to study the effect of the interferences on diclofenac response. Subsequently, a set of diclofenac standards and interferents was prepared using a fractional factorial design to build the response model and perform differential pulse voltammetry to produce the data of the input layer of the ANN. The ANN developed was able to predict the concentration of diclofenac even in the presence of the interferences, since multiple correlation coefficients of 0.9917 and 0.8387 were obtained for training and test data in the analysis of pharmaceutical samples with a recovery percentage of 95.9%.