A Study of Sampling Artifacts of Manual PM2.5 Samples
|關鍵字:||PM2.5微粒;採樣誤差;多孔金屬固氣分離器;FDMS自動監測器;Filter Dynamic Measurement System;sampling artifacts;volatilization|
|摘要:||本研究利用一部由本團隊自行開發且可同時採集4組PM10及4組PM2.5樣本的多濾紙PM10-PM2.5採樣器(Multi-Filter PM10-PM2.5 Sampler, MFPPS)與一部WINS PM2.5採樣器(Partisol* Ambient Particulate Sampler, Model 2000-FRM)及一部雙道採樣器(Dichotomous, Andersen Model SA-241, Andersen Inc., Georgia,USA)和一部可即時監測器-濾紙動態量測系統(Filter Dynamic Measurement System, FDMS, Model 1405-DF, Thermo)於交大走廊進行24小時並列採樣。其中MFPPS四個PM2.5頻道分別裝有三張鐵氟龍濾紙及一個多孔金屬片固氣分離器(Porous Metal Denuder, PMD)。本研究將MFPPS之樣本進行不同分析方式，其結果和FRM手動採樣器及FDMS比對，以探討手動採樣器於採樣過程或是濾紙調理過程可能產生的採樣誤差對PM2.5質量濃度的影響。
研究結果顯示，手動採樣器之PM2.5採樣結果會低於FDMS的修正採樣誤差之測值(PM2.5, F(b-r))，平均相對誤差為-25□10%，但卻和FDMS中未作採樣過程誤差修正之結果(PM2.5, Fb)接近，平均相對誤差為0.97□13.7%。由此可知，手動採樣會因部分揮發性物質在採樣過程中揮發而有低估PM2.5濃度的情況。本研究以PMD修正採樣過程誤差(PM2.5,Md)，該結果和PM2.5, F(b-r)的測值接近，顯示採樣過程中無機鹽類有顯著揮發情形。在20次的採樣實驗中，以手動採樣器之測值計算無機鹽類揮發量佔總PM2.5質量的比例為6至30 %，平均為16□6.8%，以FDMS之測值計算該比例為5至33 %，平均為20.1□9.7%，揮發量會隨著濾紙上的微粒質量濃度增加而降低。其中NH4+、NO3-及Cl-的平均揮發量分別為18-80% (平均為48□18.1%)、22-94% (平均為60□20.8%)及49-91% (平均為78□7.3%)，而這些揮發物質主要為NH4NO3及NH4Cl微粒。
評估調理過程對樣本質量濃度及各離子濃度的影響結果顯示，48-h後僅有些微變化，該變化量佔總PM2.5質量的比例介於0至3%，平均為1□1.3 %，經過120小時調理後，質量濃度才有略微下降的情形，下降量佔總PM2.5質量的比例介於1至7 %，平均為5□2.4%。由此結果可知，調理過程中因微粒揮發所產生的負向誤差對樣本質量濃度不會有太大的影響。|
In this study, the Multi-Filter PM10-PM2.5 Sampler (MFPPS), two reference FRM samplers including Dichotomous and EPA WINS PM2.5 sampler, and a real-time monitor FDMS (Filter Dynamic Measurement System, FDMS, Model 1405-DF, Thermo) were collocated in National Chiao Tung University to conduct the 24-h comparison tests. Three Teflon filters and a pore metal denuder (PMD) were assembled in MFPPS. In order to evaluate the effect of sampling artifacts on the mass concentrations measured by the PM2.5 manual samplers during sampling or conditioning process, the samples in MFPPS were analyzed by different subsequent analysis methods, and the mass and ions concentrations of PM¬2.5 were compared with those measured by FRM and FDMS samplers. The results showed that the PM2.5 mass concentrations measured by FRM sampler agreed well with those measured by base flow in FDMS(PM2.5, Fb). However, the PM2.5 mass concentrations were lower than those measured by base flow-reference flow in FDMS (PM2.5, F(b-r)). This indicated that the manual samplers underestimated the PM2.5 mass concentrations due to the evaporation of semi-volatile species of PM2.5 samples in the duration. In this study, the PMD was fixed in the MFPPS to modify the sampling artifacts. The mass concentrations of PM2.5,Md and FDMS(b-r) were very close. This implied that the evaporation of inorganic species in the PM2.5 samples during sampling were significant. During 20 times of sample tests, the ratio of NH4NO3 and NH4Cl volatilized mass concentrations to total PM2.5 mass concentrations ranged from 6 to 30% with an average of 16%. It was observed that the volatilization amount tended to decrease with the increase of PM2.5 mass concentrations. Compared with total NH4+, NO3- and Cl- in particulate, evaporative NH4+, NO3- and Cl- loss ranged from 18-80 % (ave. 48 %), 22-94 % (ave. 60 %), and 49-91 % (ave. 78 %), respectively. In this study, the artifacts of conditioning process after sampling were also evaluated. It was found that the ratio of decrease amount to total PM2.5 mass concentrations with 48-h conditioning process were less than 3% (ave. 1%). After 120 hours conditioning process, the ratio ranged from 1 to 7% (ave. 5%). This implied that the decrease amount due to conditioning process was insignificant compared with the total particle mass.
|Appears in Collections:||Thesis|