Simultaneous Spectrophotometric Determination of Chlordiazepoxide and Clidinium Bromide in Sugar Coated Tablets by Partial Least Squares, Principal Component Regression and Artificial Neutral Network

Abstract

Multivariate spectrophotometric approaches based on namely partial least squares (PLS), principal component regression (PCR) and artificial neural network (ANN) were proposed for the simultaneous determination of chlordiazepoxide (CHP) and clidinium bromide (CLB) in sugar coated tablets without any chemical separation step. As an alternative approach, ANN was applied to obtain the calibration function between the training data matrix and its corresponding absorbance data matrix. A training set consisting of 24 mixture solutions in the concentration range of 0-25 mu g/mL CHP and 0-30 mg/mL was prepared in methanol and 0.1 M NaOH (60:40, v/v). PLS and PCR calibrations were constructed by using their chemometric algorithms dependent on the relationships training set and absorbance data matrix obtained in the spectral range of 210-400 nm. As a comparison approach, ANN was applied to the determination of the investigated drugs in their mixtures and the commercial pharmaceutical preparation to clarify interactions between the analyzed compounds and excipients in samples. The validation of these chemometric methods was carried out by analyzing the artificial mixtures containing CHP and CLB drugs. The mathematical treatments were performed by using Microsoft Excel, a neural network algorithm written in Matlab 7.0 and the PLS toolbox 3.5. These multivariate approaches were successfully applied to the real CHP and CLB samples. As a result, the proposed chemometric methods were found suitable for the quality control and routine analysis of two drugs in sugar coated tablets.