以揮發性有機化合物及其光化學損失探討鄰近特殊工業區污染源:受體模式及PM2.5有機特徵物種

Abstract

本研究的目的在於探討鄰近特殊工業區的揮發性有機化合物（Volatile Organic Compounds, VOCs）及其光化學損失對污染源解析的影響，以及基於推估初始濃度的結果去探討特定VOC與其衍生的二次有機氣膠（Secondary Organic Aerosols, SOA）的日夜濃度。研究資料的來源有兩種，分別為2023年4月1日至6月30日位於高雄林園特殊工業區光化學評估監測站的VOCs連續監測資料，以及2024年5月11日至6月4日與前者同址的PM2.5中有機特徵成分手動採樣。首先使用正矩陣因子法（Positive Matrix Factorization, PMF）對VOCs的量測濃度與初始濃度進行來源解析，並結合風玫瑰圖及機率密度函數，分析污染源的空間分佈特徵。後續並進一步基於量測濃度的受體模式源解析結果，應用光化學年齡參數方法(Photochemical-age-base parameterization method)估算VOCs的初始濃度，修正揮發性有機物的光化學損失。最後針對PM2.5中的二次有機特徵物種及其VOCs前驅物，進行日夜間變化趨勢分析，探討二次氣膠生成與前驅物的關聯性。 本研究在使用新的光化學年齡參數方法執行架構後，能夠推估出與文獻相似的結果。光化學損失濃度最高物種為異戊二烯（isoprene）和丙烯（propylene），其光化學損失分別為0.91 μg/m3及0.83 μg/m3，佔初始濃度的71%及21%。VOCs量測濃度的PMF源解析結果為5個來源，分別是交通排放與老化氣團、石化業、汽油蒸發及液化石油氣、溶劑使用及生物源及工業來源。而VOCs初始濃度的PMF來源解析結果為7個來源，將量測濃度PMF源解析結果中汽油蒸發及液化石油氣、生物源及工業來源分別獨立出來。本研究發現VOC(異戊二烯及甲苯)與其PM2.5中的衍生物(二甲基赤藻糖醇及2,3-二羥基-4-氧基戊酸)日夜趨勢不一致，可能原因與來源特性、反映條件及稀釋效應有關。 本研究提出結合光化學年齡參數方法與受體模式的新架構，改善光化學損失對VOCs來源辨識的影響，進一步提升模型對VOCs的解析能力，能夠更準確地探討污染來源以改善空氣品質。

This study investigates the volatile organic compounds (VOCs) near a petrochemical industrial park and the impact of their photochemical losses on source apportionment. It further examines diurnal variations in specific VOCs and their derived secondary organic aerosols (SOAs) based on the estimated initial concentrations. Two data sources were employed: continuous VOC monitoring from a photochemical assessment station in the Linyuan industrial area of Kaohsiung from April 1 to June 30, 2023, and manual sampling of organic constituents in PM2.5 from May 11 to June 4, 2024, at the same location. Positive Matrix Factorization (PMF) was applied to both observed and initial VOC concentrations to identify source contributions, complemented by wind rose diagrams and probability density functions to analyze the spatial distribution characteristics of pollution sources. Diurnal variations in PM2.5 tracers and VOC precursors were explored to link secondary aerosol formation with its precursors. After implementing the new photochemical-age-based parameterization framework, this study was able to estimate results comparable to those reported in the literature. The photochemical losses were most pronounced for isoprene (0.91 μg/m³, 71% of the initial concentration) and propylene (0.83 μg/m³, 21% of the initial concentration). PMF source apportionment of observed VOC concentrations identified five sources: traffic emissions and aged air masses, petrochemical industries, gasoline evaporation and liquefied petroleum gas, solvent usage, and biogenic/industrial sources. In contrast, apportionment of initial VOC concentrations revealed seven sources, separating gasoline evaporation, liquefied petroleum gas, biogenic, and industrial emissions into distinct categories.The study identified discrepancies in diurnal trends between VOCs (such as isoprene and toluene) and their PM2.5-derived secondary products (including dimethyl erythritol and 2,3-dihydroxy-4-oxopentanoic acid), potentially due to source characteristics, reaction conditions, and dilution effects. By integrating the photochemical-age-based parameterization method with receptor models, this research improves the identification of VOC sources and enhances model resolution, providing a more accurate basis for air quality management.