Evaluation of the performance of the determination of anions in the water soluble fraction of atmospheric aerosols

The knowledge of the mass of particulate matter in air, its chemical composition and emission sources is of relevance for taking decisions concerning air quality management in urban areas. The interpretation of these data is a function of the quality of the measurement results expressed by their uncertainties. This study aimed at developing models of the performance of the determination of anions in the water-soluble fraction of atmospheric aerosols, capable of determining, separately, the contribution of aerosols sampling, extraction of water-soluble fraction of atmospheric aerosols and quantification, by ion chromatography, of anions in the extract. The sampling procedure was assessed from the dispersion of results of duplicate parallel sampling after subtracting the analytical component of this dispersion. These models are used to evaluate the adequacy of the measurement procedure for the determination of urban aerosol composition and to support strategies for reducing measurement uncertainty or cost of analysis. The method performance was studied for the following ranges considering extract dilution up to five times: 0.23-8 mu g m(-3) for chloride and nitrate, and 0.093-3.25 mu g m(-3) for sulphate. Measurements are fit for the analysis of urban aerosols since the relative expanded measurement uncertainty is smaller than a maximum value of 40%. The percentage contribution of the uncertainty components varies with the analyte and its mass concentration, the major components being 24-93% for the extraction, 43-59% for sampling, 0.2-28% for the interpolation of the sample signal in the calibration curve and 4-8% for air volume measurement. The typical composition of analysed air is: (1.12 +/- 0.26) mu g m(-3,) (1.02 +/- 030) mu g m(-3) and (0.76 +/- 0.22) mu g m(-3) of chloride, nitrate and sulphate in the water soluble fraction of aerosol, respectively, for a confidence level of approximately 95% considering a coverage factor of 2. (C) 2012 Elsevier B.V. All rights reserved.