STUDY OF THE VARIABLES AFFECTING A FIA–CV–AAS SYSTEM FOR MERCURY DETERMINATION BY PHOTOCHEMICAL GENERATION

Photochemical vapor generation is an alternative to chemical reduction in cold vapor generation for the analytical determination of mercury. Gaseous mercury atoms are formed by UV irradiation in the presence of low molecular weight organic compounds (LMWOC) like organics acids and alcohols, and transferred by means of an inert gas to an atomic absorption cell (AAC) for its determination. The aim of this work was to study the different variables affecting the AAS response when photochemical generation is used for cold vapor generation of mercury. For that purpose a single line FIA system was mounted in order to obtain a transient signal. Air was used as the stream propelling the sample injected by the injection valve to the UV reactor and subsequently to the gas liquid separator. The gas phase was transferred to the AAC by an argon stream. The UV reactor was a commercial device, consisting of a 720µL tubular reactor inside a 19W low pressure mercury discharge lamp. For the cold-vapor photochemical generation, ethanol, formic and acetic acids were tried. For each of them, the variables considered were: LMWOC concentration, irradiation time, FIA carrier, argon flow rate and sample volume. The criterion for optimum conditions was to maximize the ratio S/s, S being the average signal (peak height) and s the estimation of the standard deviation of five replicates. Low LMWOC concentrations were found more suitable, acetic acid being the only one for which peak height was independent from its concentration. As FIA carriers, LMWOC solutions (at the tested concentrations), deionized water and air were tried, finding best results with air. Parameters that most influenced peak shape were irradiation time, that determined its rise portion, and the argon flow rate, that, in the rates assayed, only conditioned its width. As expected, sample volume affected the linear range and repeatability of the response. As air was used as carrier, efficiency could be calculated by signal attenuation in the waste solution from the gas liquid separator. Once the optimum conditions for the different variables were found, a calibration plot up to 50 μgL-1-Hg was constructed and the limits of detection and quantification calculated. In terms of optimum S/s ratio, efficiency, and method performance characteristics, formic acid gave the best results. Further investigation will be focused on attempting to improve the analytical variables of influence by reducing the volume of the AAC.