Performance critical comparison of offline SPE, online SPE, and direct injection for the determination of CECs in complex liquid environmental matrices

Abstract

Sample preparation for the analysis of organic micropollutants in wastewater samples is commonly carried out by solid-phase extraction (SPE) procedures, which involve different manual laboratory operations. This conventional approach requires several hours of counter labour and entail the use of a lot of disposable material, and the subsequent contaminated non-recyclable plastic-residue production. In contrast, by coupling and automatizing the pre-treatment to the instrumental analysis most of that burden erases, sample size gets miniaturized and, thus, storage becomes freed-up. Even lab counters get cleared off from sample pre-treatment apparatus. However, method performance could get alter as a trade-off. This paper presents the results from a study in which methodology, including SPE online-coupled to UHPLC-MS/MS chromatography, was developed for multi-residue (58) determination of veterinary and pharmaceutical drugs in urban and piggery wastewater (influent and effluent to wastewater treatment plants (WWTPs)). Similarly, the direct injection (DI) of large volumes (hundreds of µL) of same matrix samples into the chromatographic system was optimized too. The performance of both automated methods was statistically compared with the classical off-line SPE. As dealing with trace analysis, suitable injection volumes for the alternative approaches were selected on the premise of low limits of quantification (MLQs). Under the selected conditions, validation parameters such as linearity range, method quantification limits, peak shape and carry over were determined. Usually more than 50 % of the analytes showed MLQs below 50 ng/L, for all matrices and methodologies, especially for DI. Real wastewater samples from a local urban WWTP and farm were analysed with all three tested methodologies. Determined concentrations and removal rates were statistically compared and turned out being quite similar. However, analysis under offline SPE and DI approaches provided a larger amount of information as they reached lower MLQs. Offline-SPE provided the worst precision among all.