Developing a four-dimensional voltammetry as a powerful electroanalytical methodology for simultaneous determination of three colorants in the presence of an uncalibrated interference

Abstract In the present study, we have developed a novel electroanalytical method based on coupling of four-way multivariate calibration (MVC) models with third-order differential pulse voltammetric (DPV) data for simultaneous determination of amaranth (AM), tartrazine (TT) and quinoline yellow (QY) in the presence of sunset yellow (SY) as uncalibrated interference at a glassy carbon electrode (GCE). The third-order DPV data were recorded by changing pulse height and pulse time as instrumental parameters. After performing some preprocessing actions such as baseline correction by asymmetric least squares spline regression (AsLSSR) and potential shift correction by correlation optimized warping (COW) on the raw data, two well-known third-order algorithms including unfolded partial least squares/residual trilinearization (U-PLS/RTL) and multidimensional partial least squares/RTL (N-PLS/RTL) were used to build four-way calibration models and their performance was checked by predicting the concentrations of the validation and test sets as synthetic samples. The results confirmed more superiority of N-PLS/RTL than U-PLS/RTL in synthetic samples which encouraged us to apply the N-PLS/RTL to the analysis of real samples. The results of the analysis of real samples showed that the N-PLS/RTL had an acceptable potential for predicting concentrations of AM, TT and QY in real matrices.