10th International Aerosol Conference September 2 - September 7, 2018 America's Center Convention Complex St. Louis, Missouri, USA
Abstract View
A Method for Testing the Accuracy of the Hourly-Averaged PM2.5 Mass Concentration by Automated Measuring Systems for Continuous Monitoring with Laboratory Generated Aerosols
Yoshiko Murashima, HIROMU SAKURAI, AIST
Abstract Number: 1175 Working Group: Instrumentation
Abstract We evaluated a method for testing PM2.5 automated measuring systems for continuous monitoring (PM2.5 systems hereafter) in the laboratory, with respect to the accuracy of the hourly-averaged mass concentration. While the accuracy of the daily-averaged mass concentration of PM2.5 systems is tested and must be within a prescribed tolerance for qualifying for certification, there is no tolerance set for the accuracy of the hourly concentration. The accuracy of hourly concentrations is important in Japan, however, since hourly concentrations are used for issuing alerts by local governments when high PM2.5 concentrations are expected. Testing the accuracy of hourly concentrations is not straightforward, though, since the method for testing the accuracy of daily concentrations, i.e., the comparison with the daily concentration determined by the reference filter/gravimetric method, is not applicable to hourly concentrations because of the lack of sufficient sensitivity. We propose a method for testing the accuracy of hourly concentrations that is to be performed in the laboratory, which uses laboratory-generated aerosols and a reference instrument that has a time resolution better than an hour. This method is reproducible and well-defined in comparison with methods with ambient aerosols. In this presentation, results from our experiments for validating the proposed testing method are reported.
Three types of PM2.5 systems were tested. Two of them were beta attenuation monitors with Japanese certification (BAM-A and -B), and the other was a tapered element oscillating microbalance (TEOM). For reference measurement, we used two methods. One was a traditional filter/gravimetric method with two samplers as the primary standard, and the other was a scanning mobility particle sizer (SMPS) spectrometer as the secondary standard. The mass concentration by the SMPS was calibrated by the filter method. As test aerosols, we used two types of particles generated by the spray-drying method. One was an aerosol of 300-nm PSL spheres, and the other was a polydispersed aerosol of ammonium sulfate particles with mass mode diameter of about 120 nm. The particles were conditioned for concentration, humidity, and electrical charge, and then delivered to a manifold from which the tested PM2.5 systems and references sampled.
We carried out two tests in the experiments. One was a test at a constant particle concentration of about 100 μg/m3 for 3 hours. Two filter samplers gave reference 3-h average concentrations, which were compared to 3-h concentrations by the SMPS to determine correction factors for the SMPS for each particle type, i.e., PSL and ammonium sulfate. Hourly concentrations by the SMPS after correction with an appropriate correction factor thus determined were used as reference for evaluating the accuracy of hourly concentrations by the PM2.5 systems. The agreement between the PM2.5 systems and reference was generally good. BAM-B, however, showed large disagreements for the first hour of the 3-h measurement periods. The disagreement seemed to occur due to slow response of BAM-B. The hourly concentrations for the first hour by BAM-B seemed to be affected by the concentration before the 3-h periods, probably because the averaging time length that was set in the instrument was longer than an hour.
The other test was done at five concentration levels changed stepwise to evaluate the linearity of the PM2.5 systems. The concentration levels were 10, 20, 40, 80, and 160 μg/m3. Each concentration level was kept for 2 hours and the hourly concentrations of the second hour at each concentration level were compared with the hourly concentrations by the reference SMPS. The linearity was found fairly good for all of the PM2.5 systems in the concentration range evaluated.