On the performance of a semi-continuous PM2.5 sulphate and nitrate instrument under high loadings of particulate and sulphur dioxide

The need for highly time-resolved data on atmospheric aerosol composition has prompted the development of several semi-continuous techniques in recent years. It is necessary to evaluate the performance of these new techniques under different meteorological and chemical conditions. We report on the results of intercomparison between a new commercial PM2.5 semi-continuous sulphate and nitrate instrument (AIM, URG9000B) and filter-based technique. The study was carried out in the summer of 2005 in two polluted suburban/rural areas near Shanghai and Beijing, China, yielding unique results on the performance of this instrument in more polluted environments compared to most of the previously reported studies. The two methods had a good (R-2 > 0.67) overall correlation for sulphate and nitrate, but showed discrepancies in absolute concentrations. At low levels of sulphate (<20 mu g m(-3)) and SO2 (<30 ppbv), the AIM data agreed with the filter result in Beijing (AIM = 0.97 x filter + 0.60, R-2 = 0.94). However, the semi-continuous instrument gave much higher sulphate values at high SO2 concentrations (in Shanghai), which was due to a positive interference from SO2 that broke through the denuder. At high sulphate loading, on the other hand, the AIM gave a much smaller sulphate reading, which was apparently due to an insufficient supply of water vapor to dissolve all the soluble ions under high aerosol loadings. For nitrate, while the two methods showed a good agreement in Shanghai (AIM = 0.83 x filter + 0.00, R-2 = 0.96), evaporative loss from the filters at high temperatures and negative bias in the AIM at high particle loading complicated the correlation in Beijing. Our study points out the need for careful evaluation and modification of the AIM for use in highly polluted environments. Under clean to moderately polluted conditions, the AIM can provide useful, highly time-resolved sulphate and nitrate data. (C) 2007 Elsevier Ltd. All rights reserved.