Guided episodic sampling for capturing and characterizing industrial plumes

Publication date: February 2018
Source:Atmospheric Environment, Volume 174
Author(s): Chang-Feng Ou-Yang, Wei-Cheng Liao, Chih-Chung Chang, Hsin-Cheng Hsieh, Jia-Lin Wang
An integrated sampling technique, dubbed trigger sampling, was developed to capture characteristic industrial emissions or plumes. In the field experiment, a hydrogen sulfide (H2S) analyzer was used as the triggering instrument at the boundary of a refinery plant due to frequent complaints of foul smell from local residents. Ten episodic samples were captured when the H2S level surpassed the prescribed trigger level of 8.5 ppbv over a three-day period. Three non-episodic (blank) samples and 23 road-side samples were also collected for comparison. All the 36 flask samples were analyzed by gas chromatography–mass spectrometry/flame ionization detection (GC–MS/FID) for 108 volatile organic compounds (VOCs). The total VOC abundance of the event samples was exceedingly higher than the non-episodic samples by over 80 times in the extreme case. Alkanes were found to be the dominant constituents in the event samples, amounting to over 90% of the total VOC concentrations vs. only 30–40% for the blank and metropolitan samples. In addition, light alkanes in the event samples were highly correlated with the trigger species H2S (R² = 0.82), implying their common origin. The matrix of chemical composition vs. sample types permitted easy visualization of the dominance of light alkanes for the event samples compared to other types of samples. Principle component analysis (PCA) identified two major contributors to cover 93% of the total variance arising from the 36 samples, further quantifying the distinction of the triggered episodic samples from the contrast samples. The proposed trigger sampling is a coupling of fast-and-slow measurement techniques. In this example, the fast-response H2S analyzer served to "guide" sampling to capture industrial plumes which were then characterized by a relatively slow method of GC–MS/FID for detailed chemical composition representative of the prominent sources.

Graphical abstract

from #ORL-AlexandrosSfakianakis via ola Kala on Inoreader http://ift.tt/2BHGxju