Speciation of selenium in natural waters and snow by DPCSV at the hanging mercury drop electrode

A procedure is described for the speciation of selenium in natural waters. According to this procedure the original concentrations of Se(IV), Se(IV) + Se(VI), and Se(IV) + Se(-II) in the sample can be determined in turn. Differential pulse cathodic stripping voltammetry (DPCSV) was used according to the method proposed by U. Baltensperger and J. Hertz (Anal. Chim. Acta, 1985, 172, 49-56) to measure the concentration of Se(IV), the only electrochemically active species. By properly selecting pH and reagent concentrations, the following reaction steps were accomplished under UV irradiation: (i) decomposition of organic compounds that are generally present in natural waters and hinder the DPCSV determination of Se(IV) (pH around 1.6, 50 mu l of 30% H2O2 in a 60-ml sample); (ii) quantitative reduction of Se(VI) to Se(IV) at pH 10; (iii) quantitative oxidation of Se(-II) compounds and organic interferents to Se(IV) in the presence of H2O2 (50 mu l was added to a 60-ml sample 2 mM in Na2B4O7). Various tests were performed in natural and synthetic seawater samples. In step 1, spikes of humic acid (up to 20 mg liter(-1)), benzene (80 mu g liter(-1)), and Triton X-100 (10 mg liter(-1)) were completely destroyed in 30 or 100 min depending on the irradiating device used. Trimethylselenonium ion and DL-selenomethionine, representative of selenocompounds in natural waters, also did not oxidize to Se(IV) at the nanogram per liter level. In step 2, no losses in Se(IV) spikes were observed after irradiation at pH 10, which means that Se(VI) reduction does not proceed toward oxidation states lower than +4. In step 3, spikes of standard solutions of trimethylselenonium and DL-selenomethionine were completely recovered as Se(IV), whereas Se(IV) spikes were not lost during the reaction step. Compared with other procedures reported in the literature, the present procedure requires a lower UV absorbed power per sample, a shorter time of irradiation, and a lower reagent concentration. This, in addition to the use of a 16-bit ADC sampling board and a suitable data processing program, allowed a Se(IV) concentration of 0.16 ng liter(-1) to be measured with an overall coefficient of variation as low as 16 (among parallel samples). Two different irradiation devices are described. (C) 1998 Academic Press.