Abstract:
Endocrine disrupting compounds, especially chemicals with estrogenic activity, have drawn increasing attention in recent years because of their eventual effects on human and wildlife development and reproduction. Regarding the toxicity of these substances the contamination state of the environment, especially surface waters, is not reached. Unfortunately the critical concentration for the estrogenic effect is extremely low. Therefore a monitoring of waste water is necessary. An immunosensor for fast and direct monitoring of surface water is presented. The multi-analyte application of this sensor uses a data acquisition called neural networks.
The first aspect of this work was the characterisation of the polyclonal antibodies with the reflectometric interference spectroscopy with regard to affinity constants, kinetics and cross-reactivity. Within these studies a mathematical model was developed for the description of the influence which a cross-reactant shows in the calibration of a standard analyte. With these information single-analyte calibration for estrone, estradiol, ethinylestradiol, bisphenol A, atrazine and simazine were developed on the total internal reflectance fluorescence (TIRF)- sensor and optimised in terms of limit of detection and stability of the biological compounds. Afterwards these analytes were used to develop a multi-analyte assay with cross-reactive antibodies and neural networks for data acquisition. The binary and ternary calibrations were optimised with respect to limit of detection, experimental design, and neural net design. The calibrations were validated with real samples. All limits of detection are in the sub ppb-range. Besides these cross-reactivity assays an assay for microcystin, a toxin of cyano bacteria, was developed. The main interest here was an online-monitoring of surface water to obtain an early-warning system that can immediately react e.g. on contamination of drinking water reservoirs or of lakes for bathing.