Search

This study exhibits that size fractionation of humic substances (HS) and their metal complexes by ultrafiltration is an efficient procedure for simultaneous determination of stability constants. Using sequential-stage ultrafiltration and a radiotracer technique the HS–Cu and HS–Zn complexes studied can gently be size-fractionated and their free metal fractions simply be discriminated. The conditional stability constants Ki obtained for size fractions of these HS metal complexes exhibit a clear molecular size dependence. Accordingly, the highest Ki values (6.6 for Zn and 6.4 for Cu) are found in the HS fractions of >105 kDa. Moreover, the overall stability constants K found for Cu (log K=5.5) and Zn complexes (log K=4.5) of the aquatic HS complexes studied are quite comparable to those reported in the literature.

A concise overview (75 references) of the analytical fractionation of aquatic humic substances using sequential-stage ultrafiltration is presented. First, humic substances in aquatic environments and actual problems connected with their fractionation and analysis are briefly considered. The molecular size classification of dissolved humic substances by means of multistage ultrafiltration, with special emphasis on on-line techniques, is the focal point of the discussion. In particular, the capabilities of ultrafiltration for the size fractionation and characterization of species formed between colloidal humic substances and pollutants (e.g. metals) are stressed.

A new method for the determination of metals in waters by flame atomic-absorption spectrometry is described. The metals are retained by water-soluble polymers in a membrane filtration cell and the factors which influence their determination are discussed. The method has been applied to the determination of Ni, Cu, Zn, Hg and Cd in drinking and river water with poly(ethyleneimine) and its thiourea derivative as complexing polymers. The metals were determined in the aqueous concentrate after a 250-fold preconcentration by 2% polymer solution at pH 7. The metal recoveries were at least 92%, and the limits of detection (ng/mg) were 0.012 for Cu, 0.006 for Zn, 0.03 for Ni, 0.004 for Cd and 0.0001 for Hg (cold vapour method). When a new type of membrane filtration cell is used even higher preconcentration factors can be achieved and lower concentrations can be determined.

The determination of traces of chloride by means of the molecular absorption of AlCl in graphite cuvettes is described. An extraction method for separation and preconcentration of the chloride has been developed, to avoid matrix effects. Extraction was done with 0.03M triphenyltin hydroxide in o-xylene, and stripping with 0.025M barium hydroxide. The method is sensitive and specific for chloride. The detection limit is about 1.5 ng of chloride, and the chloride content of 20 ml of 10−7M solution can be determined. The determination is possible in presence of excess of bromide and iodide.

The review discusses the problems of rational combination of solvent-extraction separation and preconcentration of trace elements with their determination by electrothermal atomic-absorption spectrometry.

Extraction—chromatographic separation of ortho- and pyrophosphate anions on an inert support modified with an organotin extractant was studied and used for their subsequent determination in a flow-injection system. The proposed FIA manifold includes an extraction—chromatographic mini-column, on which the phosphate anions are separated and preconcentrated, and a post-column spectrophotometric detector. For the determination of orthophosphate, the absorbance of the reduced 12-molybdophosphoric acid is monitored at 660 nm. The sum of ortho- and pyrophosphate can be determined after preliminary hydrolysis of pyrophosphate to orthophosphate in neutral solution at 50° by use of inorganic pyrophosphatase. For a sample volume of 6 ml, the calibration graph is linear within a range of 5.0–100.0 ng/ml P. The limit of detection is 0.3 ng/ml P. The recovery of the ions to be determined is not less than 96%, the relative error is not worse than 4%. The proposed method was used for the analysis of river water samples.