Square wave cathodic (adsorptive) stripping voltammetry of selenium, arsenic, vanadium, and molybdenum.

The determination of selenium(IV), arsenic(III), vanadium(V), and molybdenum(VI) at sub-nanomolar to picomolar levels has been investigated with square wave cathodic stripping voltammetry (SWCSV) and square wave adsorptive stripping voltammetry (SWAdSV). Se(IV) and As(III) form intermetallic compounds with copper and are deposited at a hanging mercury drop electrode (HMDE). V(V) and Mo(V) form dihydroxynaphthalene (DHN) complexes which can be adsorbed at an HMDE. The reduction currents for the complexes are greatly enhanced by coupling a catalytic effect of bromate, and reaction mechanisms are reported. Generally, square wave voltammetry gives either better or similar sensitivity to linear scan and differential pulse voltammetry. High SW frequency can significantly increase the sensitivities for Se(IV) and As(III), but it has less of an effect for V(V) and Mo(VI). The latter is probably due to the slow reaction kinetics. Instrumental and chemical parameters have been thoroughly investigated and optimized to develop SW methods for the measurements of Se(IV), As(III), V(V), and Mo(VI). The methods developed give detection limits of 0.15, 0.2, 0.015, and 0.006 nM and sensitivities of 57, 410, 140, and 770 A/M for these elements, respectively. The linear concentration ranges are 0.15-1000, 0.2-8, 0.05-4, and 0.01-2 nM. Relative standard deviations corresponding to these methods are 10, 7, 1, and 1% at the 1 nM level. Various substances have also been tested as possible interferences and the effects of surface-active materials are exaggerated at longer adsorption times. All of methods are able to determine these elements in uncontaminated natural waters. A theoretical model based on ligand competition has been proposed to define the complexing capacity and metal speciation in natural waters. Extrapolation and linear regression can detect the natural ligands with large stability constants {dollar}\\rm (K\\sp\\prime\\sb{lcub}ML{rcub}C\\sb{lcub}L{rcub}/K\\sp\\prime\\sb{lcub}MA{rcub}C\\sb{lcub}A{rcub}\u3e1),{dollar} but nonlinear regression is suited to those ligands with mild stability constants {dollar}\\rm (K\\sp\\prime\\sb{lcub}ML{rcub}C\\sb{lcub}L{rcub}/K\\sp\\prime\\sb{lcub}MA{rcub}C\\sb{lcub}A{rcub}\u3e0.1).{dollar}