Intercomparison and characterization of 23 Aethalometers under laboratory and ambient air conditions: procedures and unit-to-unit variabilities

Aerosolized black carbon is monitored worldwide to quantify its impact on air quality and climate. Given its importance, measurements of black carbon mass concentrations must be conducted with instruments operating in quality-checked and ensured conditions to generate data which are reliable and comparable temporally and geographically. In this study, we report the results from the largest characterization and intercomparison of filter-based absorption photometers, the Aethalometer model AE33, belonging to several European monitoring networks. Under controlled laboratory conditions, a total of 23 instruments measured mass concentrations of black carbon from three well-characterized aerosol sources: synthetic soot, nigrosin particles, and ambient air from the urban background of Leipzig, Germany. The objective was to investigate the individual performance of the instruments and their comparability; we analyzed the response of the instruments to the different aerosol sources and the impact caused by the use of obsolete filter materials and the application of maintenance activities. Differences in the instrument-to-instrument variabilities from equivalent black carbon (eBC) concentrations reported at 880 nm were determined before maintenance activities (for soot measurements, average deviation from total least square regression was −2.0 % and the range −16 % to 7 %; for nigrosin measurements, average deviation was 0.4 % and the range −15 % to 17 %), and after they were carried out (for soot measurements, average deviation was −1.0 % and the range −14 % to 8 %; for nigrosin measurements, the average deviation was 0.5 % and the range −12 % to 15 %). The deviations are in most of the cases explained by the type of filter material employed by the instruments, the total particle load on the filter, and the flow calibration. The results of this intercomparison activity show that relatively small unit-to-unit variability of AE33-based particle light absorbing measurements is