Combining Raman microspectrometry and chemometrics for determining quantitative molecular composition and mixing state of atmospheric aerosol particles

Determining quantitative molecular composition of atmospheric particles is required for assessing their environmental and health impacts. The presented algorithm was designed to analyse numerous Raman spectra of metal-rich atmospheric particles. Multivariate curve resolution-alternating least squares procedure (MCR-ALS) has been applied to resolve complex data from Raman microanalysis by means of a computer-assisted analytical procedure called Single Particle Analysis (SPA). The SPA - contrary to Raman mapping - provides data in which each single particle is assigned to a single spectrum, in the group with a statistically significant size. During the procedure, the relative contributions of individual compounds in the recorded Raman spectra have been specified. Grouping and relationship determination of the collected data have been performed by hierarchical cluster analysis (HCA) and principal component analysis (PCA). A new methodology is proposed to quantitatively determine the molecular composition and chemical mixing of single airborne particles based on the data from the automated Raman microspectrometry measurements.