一個PM2.5氣膠致冷採樣系統的設計及應用

Design and application of a PM2.5 aerosol chilling samping system

周侑霖(You-Lin Zhou)

指導教授 ： 蔡春進

國立交通大學/工學院/環境工程系所/碩士(2020年)

摘要

本研究目的在於探討氣膠採樣條件對PM2.5採樣精度的影響。過去一年的研究開發了一個4.17 L/min致冷鐵氟龍濾紙採樣器，可將氣膠流冷卻至4 ℃，結果顯示此採樣器可抑制微粒的揮發損失，但由於相對濕度（RH）只能控制在<35%以下，發現毛細管冷凝作用而在濾紙上產生冷凝水造成PM2.5質量濃度高估。本年度另外開發了一種新的5 L/min 致冷採樣器，使用了蜂巢式沸石乾燥管將RH降低至1-2%之間，且可以自動再生，經實驗室的測試結果發現100 nm-2.5 µm的微粒在致冷濾紙採樣器的微粒損失為6％-9％，為可接受範圍。在國立交通大學頂樓的採樣比對測試結果發現，新致冷採樣器的NH4+，NO3-，Cl-和SO42-的採樣偏差分別為-33.61 ± 6.99 %、-45.67 ± 8.47 %、-75.28 ± 10.91 %、0.29 ± 4.06 %，而常溫單一鐵氟龍採樣器採樣偏差分別為-39.05 ± 6.86 %、-55.11 ± 13.23 %、-84.35 ± 8.99 %、1.98 ± 4.25 %，結果顯示本致冷採樣器與常溫單一鐵氟龍採樣器相比可抑制一些揮發損失，但無法完全消除損失。去年的致冷採樣器的總離子的偏差僅為-2.44 ± 7.64 %，今年則較高為-22.86 ± 4.07 %，主要因為氣膠流太乾燥而使得氣膠的平衡氣體濃度昇高所致。新式致冷採樣器的PM2.5偏差為-8.13 ± 3.60 %，比先前因毛細管冷凝造成的正偏差15.56 ± 14.10 %低很多，只比常溫單一鐵氟龍採樣器的負偏差-5.88 ± 3.93%高一點。為了解決揮發嚴重的問題，將對此冷凝系統進行改良，將原先系統的乾燥管進行分流，混和大氣濕度與乾燥管濕度，混和出來的氣流濕度介於7~11 % RH，平均值在8.4 % RH左右，採樣結果顯示由於濕度提升，毛細冷凝水量增加，能捕捉及保留微粒中的揮發水溶性無機氣體，離子揮發程度較先前減少，總離子濃度的偏差由-22.86 ± 4.07 %變成-12.18 ± 2.55 %，毛細冷凝水的增加使質量濃度更接近真實值，PM2.5質量濃度偏差由-8.13 ± 3.60 %變成-0.17 ± 3.71 %。

關鍵字

手動採樣誤差 ； 半揮發性無機物種 ； 新致冷濾紙採樣器 ； 氣膠流除濕 ； 微利揮發損失

並列摘要

This project is to study the effect of aerosol sampling conditions on the PM2.5 sampling accuracy. The novel 4.17 L/min chilled Teflon filter sampler (CTFS) in which the air was chilled at 4 ℃ was developed in the previous year and the results showed that it can suppress the evaporation loss found in the conventional manual filter-based samplers. However, the existing CTFS was found to have the water vapor capillary condensation effect since the relative humidity (RH) was controlled to below 35% only. A new 5 L/min CTFS in which the zeolite dryer was used to reduce the RH to less than 2% was developed this year. An automatic desorption system was used to regenerate the system every 2 hours for reuse. It was found that the total particle loss of the new CTFS was 6%-9% for particles ranging from 100 nm 2.5 µm, which is acceptable. The field comparison test of the CTFS and the single Teflon filter sampler (STFS) and the porous metal denuder sampler as the reference was conducted on the top floor of the National Chiao Tung University. The results showed that the sampling biases of NH4+, NO3-, Cl- and SO42- for the CTFS were -33.61 ± 6.99 %, -45.67 ± 8.47 %, -75.28 ± 10.91 %, and 0.29 ± 4.06 %, respectively with the bias of the total ions of -22.86 ± 4.07%. In comparison, the sampling biases of these ions for the STFS were -39.05 ± 6.86 %, -55.11 ± 13.23 %, -84.35 ± 8.99 %, and 1.98 ± 4.25 %, respectively with the bias of the total ion of -26.14 ± 2.68%. It indicated that the present CTFS can suppress the evaporation loss as compared to the STFS but it does not eliminate the evaporation loss completely as compared to the previous CTFS which showed the bias of total ion of -2.44 ± 7.64%. It is because that the air is too dry and the equilibrium gas pressure of PM2.5 is too high. In addition, the bias of PM2.5 for the present CTFS was only -8.13 ± 3.60% while the previous CTFS over-measured PM2.5 due to water vapor condensation by as much as +15.56 ± 14.10%, implying that the present CTFS eliminated the water vapor condensation problem. On the other hand, the bias of PM2.5 for the STFS was -5.88 ± 3.93%. The slightly under-measurement of the present CTFS was due to the evaporation loss. In order to solve the problem of volatilization, the CTFS will be improved. The samping air of CTFS was divided to mix the ambient air and the air after dryer. The mixed air humidity was between 7~11 % RH, and the average value was 8.4 % RH. Around, the sampling results showed that due to the increase in humidity, the capillary condensed water increased, which could capture and retain the volatilized water-soluble inorganic gas from the particles. The degree of ion volatilization was reduced compared with the previous one. The bias of the total ion concentration changed from -22.86 ± 4.07% to -12.18 ± 2.55 %, the increase of capillary condensate makes the mass concentration closer to the true value, and the bias of PM2.5 mass concentration changes from -8.13 ± 3.60% to -0.17 ± 3.71%.。

並列關鍵字

FRM sampling error ； evaporation loss ； new chilled system ； water vapor condensation

參考文獻

Ahn, H., & Lee, C. H. (2004). Effects of capillary condensation on adsorption and thermal desorption dynamics of water in zeolite 13X and layered beds. Chemical engineering science, 59(13), 2727-2743.

Arhami, M., Kuhn, T., Fine, P. M., Delfino, R. J., & Sioutas, C. (2006). Effects of sampling artifacts and operating parameters on the performance of a semicontinuous particulate elemental carbon/organic carbon monitor. Environmental science & technology, 40(3), 945-954.

Chen, K.-C. (2014). 以 HTDMA 與 HT DMA-APM 系統探討無機鹽奈米微粒的吸溼行為. National Central University,

Cheng, Y. H., & Tsai, C. J. (1997). Evaporation loss of ammonium nitrate particles during filter sampling. Journal of Aerosol Science, 28(8), 1553-1567.

Collins, K. (2007). PLC programming for industrial automation: Exposure.