Data analysis of mass spectrum of aromatic secondary organic aerosol particles using Fuzzy clustering algorithms

HUANG Ming-Qiang,GUO Xiao-Yong,HAO Li-Qing,WANG Zhen-Ya,FANG Li,ZHANG Wei-Jun
DOI: https://doi.org/10.3969/j.issn.1000-0364.2010.01.027
2010-01-01
Journal of Atomic and Molecular Physics
Abstract:Secondary organic aerosol (SOA) formed from OH-initiated photooxidation of aromatic hydrocarbons may impact on the radiative balance associated with climate change, visibility degradation, and heath effects. An aerosol time-of-flight mass spectrometry (ATOFMS) designed and manufactured by our laboratory can be used to analyze SOA in real-time and generate thousands of mass spectra from typical measurement campaign. How to process these data and extract valuable information is one of the key problems for the ATOFMS. In this paper, the Fuzzy C-Means (FCM) algorithm is used to classify large numbers of mass spectrum of SOA particles formed from photooxidation of 1,3,5-trimethylbenzene analyzed by ATOFMS. Experimental results showed that SOA particles created by 1,3,5-trimethylbenzene photooxidation can be classify with five distinct chemical particle classes, the predominant components of these five particles are aromatic aldehyde, nitrophenol, phenolic and carboxylic acid compound. These also provide additional insight into the photooxidation mechanism for aromatic hydrocarbons.
What problem does this paper attempt to address?