- 學位論文
改善PM2.5採樣儀器的準確性
The improvement of the accuracy of the PM2.5 sampling system
摘要
流行病學研究顯示空氣中氣動粒徑<2.5 µm (PM2.5) 的懸浮微粒濃度與死亡率有顯著關係，許多國家均會訂定PM2.5空氣品質標準，以評估危害暴露的風險等級，因此就需要即時及準確量測懸浮微粒濃度的方法。而美國環保署 (USEPA) 所認證的FRM或FEM的PM2.5採樣或監測系統均會在採樣進口使用Well Impactor Ninety-Six (WINS) 或Very Sharp-Cut Cyclone (VSCC) 等微粒分徑器作為PM2.5的粒徑篩選，因此PM2.5微粒分徑器的效能變化對於量測PM2.5的準確性具有關鍵性的影響。 然而，由實驗結果顯示當微粒負載由0-6 mg，WINS的dpa50由2.44 μm下降至2.05 μm，且在17天連續大氣環境採樣比對，WINS在不更換玻璃纖維濾紙及真空油情況下，其dpa50下降至2.12 μm且低估20%的質量濃度，原因在於所收集的微粒會累積於衝擊表面上並形成一堆積層，使後續收集的微粒撞擊於堆積層導致小於氣動粒徑2.5 μm的微粒會被堆積層所收集，導致PM2.5濃度被低估；VSCC則在不清潔內部微粒負載的情況下環境比對結果發現在連續採樣6天後的採樣誤差則呈現不穩定變化。 因此本研究所開發的Modified WINS (M-WINS)，目的為消除微粒彈跳及微粒負載影響，以達到在長時間的採樣監測過程無須維護衝擊面並維持微粒分徑的效能。M-WINS1以濕潤的玻璃纖維濾紙作為衝擊面，並以連續水流沖洗玻璃纖維濾紙上所累積的微粒，M-WINS2則以矽油浸泡玻璃纖維濾紙作為衝擊面，並以間歇式注油沖洗衝擊面取代連續水流，以消除微粒過量負載問題並延長微粒分徑器的使用時間。由結果顯示當微粒負載由0-6 mg，M-WINS1的dpa50介於2.44 μm至2.49 μm，顯示微粒負載量增加並沒有明顯的改變dpa50，且在17天連續大氣環境採樣比對，結果亦符合USEPA規範 (dpa50 = 2.5 ± 0.2 μm及採樣誤差< ±5%)。M-WINS2 則與 WINS 相比較具有相近的截取直徑 (dpa50= 2.43 ± 0.02 μm) 及陡峭度 (GSD = 1.22) ，且當微粒負載質量由0-3.6 mg，M-WINS2的dpa50由2.48 μm 下降至2.18 μm，經由3mL/min矽油連續沖洗1分鐘後，其dpa50回到2.46 μm，結果顯示亦可以消除微粒負載影響，後續沖洗後的矽油亦可經由過濾後再利用，可減少矽油使用量。由結果顯示M-WINS的設計在使用連續水流或是間歇性的注油均可應用於長時間的採樣或監測系統，並且無須維護衝擊面亦可維持微粒分徑的效能。 此外，M-WINS1以連續水流沖洗累積於濕潤玻璃纖維上的微粒，沖洗的水樣可用來即時分析水溶性離子。為了準確量測水溶性離子的濃度，因此針對M-WINS1的氣體入口的微粒損失進行探討，結果顯示2.1至10.3 µm的微粒，在M-WINS1入口管的微粒損失對於油酸微粒(oleic acid)為0至77%，螢光胺微粒(ammonium fluorescein)則為0至67%，發現大微粒會因為微粒慣性或攔截作用沉積於收縮管及內壁，導致原因在於進氣入口管（內徑= 28.4 mm）在噴嘴前(inner diameter = 28.4 mm)有兩個邊緣尖銳的收縮管（第一層由28.4至12.7 mm，第二層由12.7至3.91 mm），因此本研究設計內部漸縮管為平滑邊緣的氣體入口管，以減少微粒的損失，結果顯示dpa小於4 µm的微粒，微粒損失由<11.8% 減少至<4.3%，dpa小於5.8 µm的微粒，微粒損失由33.4% 減少至21.5%，dpa小於10.3 µm的微粒，微粒損失由68.2% 減少至49.9%，平均粗微粒的損失由28.7 ± 11%下降至19.5 ± 7%，此微粒損失量可用於校正粗微粒的水溶性離子的濃度，因此本研究設計的新PM2.5進口可用於未來粗微粒水溶性離子的即時分析系統。
並列摘要
PM2.5 air quality standard was promulgated in many countries to protect public health since epidemiology studies have found association between particles with aerodynamic diameter <2.5 µm (PM2.5) and increased mortality, both daily and over time. FRM (Federal Reference Method) PM2.5 samplers and FEM (Federal Equivalent Method) monitors are mainly used as compliance monitors, which make use of size-selective inlets followed by a filter to collect particles for further analysis. The PM2.5 size-selective inlets which are WINS (Well Impactor Ninety-Six) or PM2.5 VSCC (Very Sharp-cut Cyclone) inlets play an important role in the sampling accuracy of the sampling systems. However, the field test of the WINS without regular replacement of the well shows the decrease in the cutoff diameter (dpa50) down to 2.12 µm and the increase in the negative sampling bias to 20% after 17 continuous sampling days. It is due to particles pile up the mound on the substrate, resulting in the particle loading effect. The performance of the uncleaned PM2.5 VSCC is claimed not to be affected by the particle loading effect. However, when the sampling period is over 6 days, we found fluctuations in measured PM2.5 concentrations as compared to the concentrations of the VSCC which is cleaned daily. In this study, a modified WINS (M-WINS) were developed to eliminate the particle bounce and overloading effects for unattended operation over a long sampling period. The M-WINS¬1 which uses water to wash the wetted glass fiber filter substrate (GFFS) clean continuously to eliminate the particle overloading effect and extend the service life of the impactor without the need of impaction well cleaning and replacement of the oil-soaked GFF. The laboratory test shows that dpa50 of the M-WINS1 only varies slightly from 2.44 to 2.49 µm when the loaded particle mass increases from 0–6 mg. The field test shows that dpa50 and the sampling bias of the M-WINS1 meet the USEPA requirements (dpa50 = 2.5 ± 0.2 µm and sampling bias < ±5%) after 17 continuous sampling days. Instead of using continuous water flow, the M-WINS2 uses vacuum oil which is injected periodically to wash off the deposited particles. The laboratory test results show that the M-WINS2 has a similar cut-size (2.43 ± 0.02 µm) and sharpness (1.22) as the WINS. Without cleaning the substrate, the particle loading test at the laboratory for M-WINS2 shows that the cut-size decreases from 2.49 to 2.18 µm as the particle mass loading increases from 0 to 3.5 mg with a particle mound formed on the substrate. The cut-size of M-WINS2 is able to recover to 2.46 µm when the oil flow is injected for just 1-min at the flow rate of 3 mL min-1 to wash off the mound. This saves oil consumption. Therefore, the current M-WINS can be used as the PM2.5 size-selective inlet for unattended, continuous sampling for a long time with a very good sampling accuracy. In addition, the liquid sample collected from the M-WINS¬1 can be analyzed for measurement of the water-soluble ions (WSI) in deposited particles. In order to determine the accurate ion concentration of WSIs, the particle loss in the inlet tube of the M-WINS1 was studied. The results show that the particle loss in the inlet and outlet tubes of the M-WINS1 increases from 0 to 77% for oleic acid (OA) particles and 0 to 67% for ammonium fluorescein (AF) particles when the particle sizes are increased from 2.3-10.3 µm. Particle loss is severe for large particles since the original inlet tube (inner diameter = 28.4 mm) have two sharp-edged contractions (28.4 to 12.7 mm, then 12.7 to 3.91 mm) before the nozzle of 3.91 mm in diameter, large particles deposit on the contractions and inner walls due to inertial impaction and interception. A new inlet tube was made by smoothing the sharp edges to eliminate the particle loss. The results show that the particle loss is reduced from <11.8% to <4.3% for dpa less than 4 µm, <33.4% to <21.5% for dpa less than 5.8 µm and <68.2% to <49.9% for dpa less than 10.3 µm. The total estimated particle loss in coarse particles in the M-WINS is reduced from 28.7 ± 11% for the original inlet tube to 19.5 ± 7% for the new inlet tube, which is reasonably small enough for correcting the measured ion concentrations in coarse particles. Therefore, the new PM2.5 inlet developed in this study is useful for future automatic monitoring of ion concentrations in coarse particles.