ABSTRACT
The development of analytical chemistry is omnipresent in all fields, this leads to considerable consumption of organic solvents and hazardous reagents with an increase in the production of waste to be treated. In this work, we developed simple, fast, cost-effective and above all environmentally friendly methods for the analysis of Acetaminophen (ACT) and Ascorbic acid (ASC) in synthetic mixtures and pharmaceutical formulation, using UV spectroscopy. Four chemometric methods were studied, including PLS-1 with full-spectrum (Full-PLS) and PLS-1 using three variable selection methods, namely subset selection through a genetic algorithm (GA), uninformative variable elimination using iterative predictor weighting (IPW), and variable selection by sub-window permutation analysis (SwPA). The accuracy of the developed methods was evaluated through the root mean square error of prediction (RMSEP), the mean absolute percentage error (MAPE) and the recovery values. All methods showed more accurate prediction results in comparison with full-PLS calibration. Furthermore, the results indicate that the GA-PLS models showed the highest prediction accuracy among all other models with RMSEP and MAPE values of (0.0494 and 0.610) and (0.0163 and 0.321) for the estimation of ACT and ASC, respectively. The proposed methods were successfully applied to the determination of ACT and ASC in their combined dosage form. In addition, the results obtained were statistically compared to those of the conventionally used HPLC method and were found to be in good agreement. The main advantages of the developed methods over HPLC during routine analysis are that they are faster, inexpensive, simple to perform, without the need for major pretreatment of samples. Besides, no organic solvents are used, and thus toxicity and pollution are avoided.
