LAUSR

Abstract Time-resolved quantitative and qualitative measurement of the particulate matter (PM) - composition is an ongoing challenge for scientists and instrument developers. Raman spectroscopy is especially useful for PM analysis, however, its application has been restricted mainly to qualitative laboratory investigation. A low particle load on the filter and the resulting high elastic backscattering make Raman spectroscopic quantitative monitoring a demanding task with regard to optical design and the mathematical tools needed for data evaluation. In this paper, we present environmental measurements taken over several weeks with the fully automatized Raman Soot Spectrometer (RaSoS). The hourly recorded Raman spectra are characterized using factor analytic methods and a linear model for quantitative determination of the apparent elemental carbon (EC) filter loading is derived. Its predictions are compared with the measurements of an aethalometer which utilizes light absorption to measure the black carbon (BC) - concentration. Raman spectra taken a year earlier at the same location yield plausible predictions using the same model. The EC-scores of the RaSoS generally agree well with the BC measurements of the aethalometer ( σ = 0.82 – 0.92 ). However, for very high concentrations of BC and for several short time periods, there are deviations which may be connected to varying PM-composition.