Second-order calibration for simultaneous determination of pharmaceuticals in water samples by solid-phase extraction and fast high-performance liquid chromatography with diode array detector

• Fast SPE-HPLC-DAD method for simultaneous analysis of pharmaceuticals in water
• Optimization of extraction procedure using face-centered central composite design
• MCR followed by COW for efficient chromatographic peak alignment
• Multivariate analytical figures of merit for evaluation of MCR-COW-PARAFAC and MCR-COW-MCRtril methods

A fast high-performance liquid chromatography–diode array detection (HPLC–DAD) approach combined to solid phase extraction (SPE) as a pre-concentration step is developed for simultaneous determination of five selected pharmaceuticals (carbamazepine, naproxen, diclofenac, gemfibrozil and mefenamic acid) in water samples. The effective factors on the efficiency of SPE procedure are optimized using faced-centered central composite design (FCCD). In addition, multi-response optimization by using Derringer's desirability function is used to find the optimum experimental conditions for extraction of analytes from well and river waters. Due to the complexity of water matrices and the presence of different chromatographic issues, new combination of multivariate curve resolution–correlation optimized warping to parallel factor analysis (MCR–COW–PARAFAC) and multivariate curve resolution-alternating least squares with trilinearity constraint (MCR–COW–MCRtril) is proposed for simultaneous determination of five target pharmaceuticals in these samples. This strategy allowed us to overcome matrix effects and to exploit second-order advantage. Recoveries ranging from 80.12% to 102.71% and relative standard deviations below 8.0% for all pharmaceuticals proved the accuracy and precision of the proposed method. In addition, the values of relative error in calibration curves (RE, %) were below 11.24% and 12.19% for MCR–COW–PARAFAC and MCR–COW–MCR, respectively. Furthermore, the values of detection limits (LODs) and quantification limits (LOQs) were between 0.02 to 0.32 ng mL− 1 and 0.07 to 1.07 ng mL− 1, respectively. In addition, analytical figures of merit using multivariate approaches were calculated for both methods. In this regard, the values of sensitivity, LOD and LOQ showed an improvement compared to univariate techniques. Inspection of the results showed that due to the efficient correction of elution time shifts using MCR–COW method and preserving trilinear data structure, both MCR–COW–PARAFAC and MCR–COW–MCRtril give similar results. These results confirmed that coupling optimized SPE–HPLC–DAD method with second-order calibration algorithms can be considered as an efficient method for fast, simple and cost effective quantification of pharmaceuticals in highly contaminated samples, such as river and well waters.