The application of Raman spectroscopy combined with multivariable analysis on source apportionment of atmospheric black carbon aerosols

Abstract

Source apportionment studies become increasingly crucial for black carbon (BC) in atmospheric particulate matters given its linkage with adverse public health and climate impacts. In this work, a facile and rapid method using Raman spectra combined with stepwise discriminant analysis (SDA) was proposed to identify and quantify the contributions of atmospheric BC sources. Four BC samples from biomass burning, coal combustion, gasoline and diesel vehicle emission were characterized by Raman spectra. The SDA model was established based on 10 parameters with significant differences (p < 0.05), giving an accuracy of 83% with a cross-validation rate of 80%. Utilizing four suggested discriminant variables from SDA model, vehicle emission was predicted as the dominant contributor to ambient BC particles, among which gasoline contributed much higher than diesel at an urban road intersection in Shanghai, China. This new method shows great potential to classify and investigate the sources of atmospheric BC aerosols and provide more effective information on air pollution control measures.