Source apportionment of PM2.5 in Beijing using positive matrix factorization

Abstract

In order to do source apportionment of PM2.5 and source trajectories, particle induced X-ray emission and energy disperse-X ray fluorescence were used to analyze chemical composition of samples. Five major sources have been identified by using positive matrix factorization. Source trajectories have been determined by conditional probability function and the potential source contribution function. Extreme events, such as sandstorms, have been identified and traced.