受體模式PMF結合PM2.5碳成分之探討

Abstract

本研究使用受體模式正矩陣因子法（Positive matrix factorization, PMF）來分析採樣點周遭之污染源對於PM2.5之貢獻量，同時還探討粒狀物之碳成分在源解析中可扮演之角色，試了解粒狀物之碳成分對於PMF來源解析的影響為何。本研究先針對台灣南部一鋼鐵工業區進行PMF分析，總共解析出5個因子，其類別為重油燃料、海洋性氣膠與交通源、地殼土壤揚塵、煤炭燃燒、金屬工業源。其中重油燃料和金屬工業源佔總PM2.5貢獻量之比例最高，分別為29.1 %和27.2 %。同時利用此實驗數據，另增加了粒狀物之碳成分數據後，再進行一次PMF分析，結果僅再次佐證各因子的類別。 本研究還運用了台中市區之每小時大氣連續監測資料來進行PMF源解析，且使用了三種不同分析或是計算碳成分的數據，並與沒有使用碳成分之分析結果進行比較。例如數據中粒狀物碳成分的分析方法，其中一種是使用Sunset OC/EC的碳成分數據，另外兩種是以Aethalometer吸光係數換算。而在不使用碳成分之情況下，仍然可以得到一交通源因子，但無法進一步找到其他碳排放源。其中兩種增加碳成分的PMF解析結果可得到六個因子，因子類別為土壤地殼揚塵、二次氣膠源、兩種工業源、汽油生物質燃燒排放以及柴油排放源，其中二次氣膠源對PM2.5的貢獻量在所有結果中都是最高的。相較於利用Sunset OC/EC碳成分的PMF來源解析，以Aethalometer吸光數據換算之碳成分的結果具有較清晰之排放因子，且該模式之誤差評估結果也比較完善。

In recent years, fine suspended particles PM2.5 (particles with aerodynamic diameter less than or equal to 2.5 μm) has been one of the most popular study categories in air pollution. This research intends to use the positive matrix factorization (PMF) method to analyze the PM2.5 contribution of the pollution sources around the sampling area, and to discuss the role carbonaceous component of the particulate matter plays in the source apportionment. In the PMF analysis of a steel industry zone in southern Taiwan, five factors were chosen. The categories are heavy oil fuel, marine aerosol and transportation sources, crustal soil dust, coal combustion, and metal industry source, of which heavy oil fuel and metal industry source’s contributions were the highest, account for 29.1% and 27.2% of the total PM2.5 contribution, respectively. In this study, the carbonaceous composition was added to perform another PMF analysis. And its concentration distribution also confirmed the above result. Since there are some limitations in the data of the steel industry zone, this study also used a hourly atmospheric monitoring data from Taichung City to perform PMF analysis. Four types of results were resolved from this data set. Two of which have 6 factors and the other two have 5 factors. The categories of the factors are soil crustal dust, secondary aerosol source, two industrial sources, gasoline combustion and diesel combustion sources, of which the secondary aerosol source has the highest contribution in all four results. The difference among the four results are mainly about their carbonaceous composition data. The first result utilized carbon aerosol measured by the Sunset OC/EC analyzer, the other two used Aethalometer data, and the last one did not contain carbon data. The carbonaceous aerosol measured by different analytical instruments had different effects on the PMF results. By using PM2.5 carbon fraction analyzed with Sunset OC/EC as input data, PMF could divide the traffic source into different categories. The first type of Aethalometer result could not further divide the carbon source, while the second type could divide it into gasoline and diesel factor. According to the error estimates of the PMF results, the second types of the Aethalometer results is more robust. Without carbonaceous aerosol, the traffic source cannot be separated. From the results of this analysis, the carbonaceous aerosol can help the PMF model to have a more in-depth analysis of the carbon source.