| dc.description.abstract | At the end of 2016, our laboratory set up a gas chromatographic (GC) system at the northern tip of Taiwan, Fuguijiaoto, to measure real-time mixing ratios of ethane, propane, ethylene and propylene. We found that the ratio of ethane to ethylene can be used as an effective indicator of the long-range transported (LRT) polluted air masses. However, due to high air humidity, the pre-concentration using the sorbent trap at sub-ambient temperatures was frequently clogged due to icing, resulting in insufficient trapping or even damage of the sorbent packing. As a result, the objective of this study is to solve the problem of ice clogging during trapping to facilitate continuous monitoring without interruption.
The solution is to add an empty glass tube in front of the sorbent trap in the same cooling module to be cooled at -30℃ when trapping. Excess water vapor will condense onto the glass tube, while the air sample that has been mostly dried can pass through to the trap. Compared with the Nafion dryer, our solution is more rugged and involves no consumables, and is almost maintenance free.
Laboratory test showed that the linearity of the four target species is higher than 0.99 (R2). Precision (RSD) is between 0.66 and 0.94%, and the measurement detection limits (MDL) is between 0.16 and 0.38 ppb. From December 5, 2017 to January 7, 2018, the improved system was placed in the Fuguijiao station again with 15 min GC cycle time, and no interruption due to ice clogging of the trap had occurred, suggesting the water removal solution was successful.
Continuous data of PM2.5, PM10, O3, NOX, CO, SO2 and wind parameters were used to support the interpretation of the observed ethane/ethylene ratios at the Fuguijiao station. We found that the high PM2.5 peak of 12/24 was a local event as suggested by the relatively low values of ethane/ethylene, while the PM2.5 peak of 12/25 was a LRT event as indicated by the relatively high values of ethane/ethylene. This finding was consistent with our presumption of the ethane/ethylene ratio as an indicator of LRT. However, the PM2.5 peaks at 12/8 and 12/11 are more intriguing. The ethane/ethylene ratios showed relatively low values which should have been indicative of local events. However, both the wind field and model simulation suggested otherwise. We found that the SO2 and NOX levels during these two time periods also elevated, consistent with the relatively low ethane/ethylene ratios. It turns out that the contradiction was most likely caused by the ship emissions off-shore, which explains the elevated values of NOx, SO2, ethane and ethylene, as well as the relatively low ratios of ethane/ethylene. | en_US |