Comparative parallel characterization of particle populations with two mass spectrometric systems LAMPAS 2 and SPASS

Two transportable laser mass spectrometers, Single Particle Analysis and Sizing System (SPASS) and Laser Mass Analyzer for Particles in the Airborne State (LAMPAS 2), have been applied to investigate the dependence of spectra patterns on instrumental parameters and data evaluation procedures in an inter-comparison experiment. Laboratory experiments showed the spectral response of both instruments for mineral particles before and after heterogeneous reactions. During a period of 47 h, both instruments determined size and chemical composition of several thousand single particles of an ambient particle population. Time-resolved evaluation (1-h resolution) of specific ion signals, which showed a characteristic temporal evolution, in combination with meteorological information, was used to select four periods for separate evaluation of particle spectra. Application of the two particle classification algorithms, fuzzy c-means clustering and k-means clustering, on the same data set (SPASS) showed only minor differences in spectral patterns and class abundances caused by the clustering method ("soft" or "hard" clustering). Spectral patterns determined for the data sets of two instruments (SPASS and LAMPAS 2) were similar for some particle types and could be compared directly (e.g., mineral or carbonaceous particles). For other types of particles, spectral patterns differed from each other and had to be interpreted using additional information on instrumental parameters (e.g., laser wavelengths or irradiance) and experimental conditions. The different response of SPASS and LAMPAS 2, as reflected in the different abundances of particle classes, indicates the necessity to determine adjustment factors for each instrument, for different particle classes, to enable a direct comparison of quantitative information from such online aerosol mass spectrometers and from bulk analysis. The reported results are an important basis for a general database of single particle spectra, spectral patterns of common and specific particle classes and abundances of these classes for atmospheric aerosols, showing their dependence on particle size, geographic location, meteorological conditions and time of analysis. (c) 2006 Elsevier B.V. All rights reserved.