| 摘要: | 大氣中細懸浮微粒(氣體動力直徑小於或等於2.5 μm，PM2.5) 對於氣候環境及人體健康有極大的影響，PM2.5濃度及化學組成隨時間及地點的不同會有明顯差異，必須固定時間在不同地點進行檢測，才能有系統地瞭解PM2.5的時間和空間變異。本文彙整2011年3月9日至11月16日，在環保署8個空品監測站進行41次的PM2.5質量濃度手動檢測及在新莊、忠明及小港站加入的PM2.5化學成分手動檢測，探討台灣都會地區PM2.5質量濃度及化學成分的變化，研討議題包括手動檢測及自動監測儀器間的差異及探討、自動監測轉換為手動檢測濃度、採樣及旅運空白分析、氣體吸附和微粒揮發干擾校正、質量閉合(mass closure)、都會區污染來源解析、微粒結合型態、都市霾害探討及能見度的影響。 研究結果顯示，8個空品站的R&P 2000標準方法檢測濃度與監測濃度變化趨勢一致，全數據相關性可達到0.87，但R&P2000檢測濃度大都低於監測濃度，差異百分比為32%。3個成分站的R&P2300檢測濃度與R&P2000檢測濃度，全數據相關性可達0.87，差異百分比約為21%。兩部手動檢測儀器間的PM2.5差異，主要受到R&P2300裝設蜂巢管及濾紙匣的影響。 成分空白分析顯示，碳成分的採樣空白及水溶性離子的二、三階空白濾紙容易超過儀器兩倍方法偵測極限。三個成分站在春季及秋季最主要成分為SO42-，夏季大都以有機碳(organic carbon, OC)為主要成分，春季揮發NO3-比例最低，夏季及秋季NO3-濃度雖然不高但揮發比例明顯。揮發NO3-及NH4+濃度，可藉由採集PM2.5質量濃度與環境最高溫度進行推估，相關性(R2)最佳達到0.84。以絕對主成分分析各成分站PM2.5質量濃度污染來源為光化學污染及汽油車排放影響，以光化學污染貢獻量較高。經由質量閉合估算及模式模擬，氣膠含水量在35% RH環境下對於PM2.5質量濃度仍佔有可觀比例。在氣膠結合型態方面，春季及夏季結合型態為(NH4)2SO4(S)及NH4NO3(S)，秋季則以(NH4)2SO4(S)、NH4NO3(S)及NH4Cl(S)為主。多元迴歸發現新莊及小港站影響能見度的PM2.5成分為NO3-及SO42-離子，忠明站為NH4+及Cl-離子，環境因子方面，各成分測站都選入相對濕度的影響。觀察採樣期間受到霾害影響時，各成分物種及氣狀污染物都有提升現象，其中NO3-濃度增加最明顯，造成霾害發生主要為機動車輛排放所致。霾害影響期間，影響小港站能見度的成分物種為NO3-及NH4+。 本文經由系統性檢測台灣PM2.5質量及成分濃度，獲得PM2.5地理及季節變化，建立手動檢測和自動監測濃度關係，評估易揮發成分影響，最後並解析PM2.5污染來源及影響能見度因子。Atmospheric fine particulate matter (with aerodynamic diameter less than or equal to 2.5 μm, PM2.5) have a great impact on both the environment and human health. PM2.5 mass concentration and chemical composition vary significantly in different times and places. To understand PM2.5 temporal and spatial variations systematically, aerosol samples collected at various places in a fixed time interval is preferred. This study collected 41 PM2.5 mass concentrations at the eight air quality monitoring stations of Taiwan Environmental Protection Administration (TEPA) and resolved PM2.5 chemical composition at the Xinzhuang, Chungming, and Siaogang stations from March 9 to November 16, 2011. The aim of this study is to investigate PM2.5 mass concentration and chemical composition in Taiwan metropolitan area. The subjects studied include deviations between manual collection and continuous monitoring, the conversion of PM2.5 from continuous monitoring to manual collection, field and trip blanks analysis, the artifact corrections from adsorbed gases and semi-volatilized particles, mass closure for analyzed and modeled aerosol components, source apportionment for urban aerosol, derivation for potential compound forms, and factors affecting urban haze and visibility. The results show that temporal variation of PM2.5 from R&P 2000 federal reference method (FRM) is consistent with continuous monitoring at the eight air quality stations. The coefficient of determination (R2) from linear regression for the whole data could reach at 0.87. However, most R&P 2000 FRM PM2.5 concentration are lower than that of continuous monitoring with an average of absolute deviation at 32%. By contrast, the manual collection of PM2.5 from R&P 2300 and R&P 2000 are correlated well with an R2 value at 0.87 and an average of absolute deviation at 21%. The deviations of PM2.5 mass concentration between two manual collectors are affected by honeycomb denuders and filter package installed in R&P2300. Blank analysis for chemical composition reveals that field blanks for carbonaceous content and the water-soluble ions in the second and the third filters frequently exceeded two times of method detection limit. SO42- was the major chemical component of PM2.5 at the three speciation sites in spring and fall. By contrast, organic carbon (OC) was the major chemical component in summer. The volatilized percentage of NO3- was lower in spring but higher in summer and fall although the volatilized concentration was not high. The volatilized NO3- and NH4+ concentrations can be estimated with the highest R2 at 0.84 by the collected PM2.5 mass concentration and the highest ambient temperature during collection. Absolute Principal Component Analysis showed that PM2.5 was mainly contributed by photochemical pollution followed by gasoline vehicle emissions. Aerosol water content occupied a considerable fraction of aerosol mass even under 35% RH from mass closure assessment and model simulation. For aerosol compound forms, (NH4)2SO4(S) and NH4NO3(S) were derived for spring and summer; while (NH4)2SO4(S), NH4NO3(S), and NH4Cl(S) were reasoned for fall. Multiple regression analysis (MRA) for atmospheric visibility showed that NO3- and SO42- were major PM2.5 components at the Xinzhuang and Siaogang stations and NH4+ and Cl- were significant at the Chungming station. Among the environmental factors, relative humidity was selected as a significant factor for visibility impairment in MRA. For the haze event during observation period, aerosol components and gaseous pollutants were significantly increased with the increase of NO3-. This indicates vehicle emissions were mainly responsible for haze event. The major chemical species affecting the visibility at the Siaogang station were NO3- and NH4+ during haze event period. In the present study, PM2.5 mass concentration and composition were investigated systematically to obtain PM2.5 geographical and seasonal variations. In addition, the relationship between manual measurement and continuous monitoring was established and volatilization of semi-volatiled species was evaluated. Finally, source apportionment for PM2.5 and factors affecting visibility degradation were resolved. |
