The development of the first microfluidic paper-based analytical device (µPAD) for the speciation of inorganic arsenic in environmental aqueous samples as arsenite (As(III)) and arsenate (As(V)) which implements hydride generation on a paper platform is described. The newly developed µPAD has a 3D configuration and uses Au(III) chloride as the detection reagent. Sodium borohydride is used to generate arsine in the device’s sample zone by reducing As(III) in the presence of hydrochloric acid or both As(III) and As(V) (total inorganic As) in the presence of sulfuric acid. Arsine then diffuses across a hydrophobic porous polytetrafluoroethylene membrane into the device’s detection zone where it reduces Au(III) to Au nanoparticles. This results in a color change which can be related to the concentration of As(III) or total inorganic As (i.e., As(III) and As(V)) concentration. Under optimal conditions, the µPAD is characterized by a limit of detection of 0.43 mg L⁻¹ for total inorganic As (As(III) + As(V)) and 0.41 mg L⁻¹ for As(III) and a linear calibration range in both cases of 1.2–8.0 mg As L⁻¹. The newly developed µPAD-based method was validated by applying it to groundwater and freshwater samples and comparing the results with those obtained by conventional atomic spectrometric techniques. Graphical abstract

The present work explores the slurry sampling approach for automatic, flow-based plant analysis. For this purpose, pinch valves were introduced into a multi-syringe flow injection analysis manifold to provide the repeatable aspiration of a few microliters of plant suspension before the material was further processed through the flow system. For validation of the proposed approach, the determination of potassium by flame emission spectrometry was implemented. Several parameters were studied: the concentration of plant particles in the sample suspension and the utilization of matrix modifiers. Microwave digestion was also implemented; no significant difference was found when certified reference material was analyzed with or without the in-line digestion step. The system was successfully applied to 13 samples within a concentration range of 2.5 to 100 mg g-1. A determination frequency of 28 h-1 was achieved and the precision was better than 4.0% (n = 12).