機車排放廢氣中微粒之碎形型態與化學組成特徵研究

Abstract

本研究以機車車體動力計模擬機車引擎運作，控制引擎轉速分別處於3600 rpm、4600 rpm和5800 rpm，並搭配使用92和95無鉛汽油兩種油品，對其排放微粒和聚集體，藉著高放大倍率之電子顯微攝影技術，建立起單一顆粒之圖譜資訊，並藉由運用碎形模式，對聚集體表面型態紋理進行碎形參數運算，另外再搭配化學分析方法，得到單一顆粒之元素重量與原子量百分比組成，建立起機車排放微粒物化特性資料，以便能提供鑑定微粒來源的有效證據。此外，為了解目前我國機車排放懸浮微粒實際情形，對八台機車採集排放微粒並進行相同分析，最後再與柴油車排放微粒進行微影圖像、碎形參數和元素組成三方面之比較。 在十萬倍SEM 之微影圖中顯示，機車車體動力計與實車排放微粒之型態非常相似，單元顆粒形狀多為圓球狀，粒徑介於10 ~ 50 nm之間，而其所形成的聚集體，粒徑則是介於100 ~ 1000 nm之間，常以團塊狀之聚集型態出現。挑選出清晰之影像進行碎形模式之分析，得到引擎在不同轉速及不同油品下之排放微粒其碎形參數非常接近，碎形參數值介於0.118 ~ 0.366，平均碎形參數為0.242。在實車方面，八輛機車之平均碎形參數範圍在0.222 ~ 0.292之間，為一合理之範圍。並且於不同機車碎形參數比較中，得到四行程之平均碎形參數較二行程高；排氣量125 c.c.與排氣量100 c.c.之平均碎形參數分佈無太大之差異；而惰轉(1800 rpm)之平均碎形參數較非惰轉(4600 rpm)為高。在元素組成方面，選取粒徑皆約為1 μm之微粒聚集體，得到機車排放微粒之主要成分為碳、氧、矽和鋁，重量百分比小於5%的元素大多為鉀和鐵，微量元素則有鈉、鎂、鈣、硫、錫和鉻，其中碳的重量百分比在11.37 ~ 53.41%之間。 在機車與柴油車排放微粒之比對研究中，十萬倍SEM之微影圖無太大之差異。碎形參數方面，柴油引擎動力計排放微粒之平均碎形參數範圍為0.194 ~ 0.374，而機車之平均碎形參數範圍則是0.118 ~ 0.366。在化學元素組成方面，依不同粒徑做碳含量之重量百分比圖來區別，有非常明顯之差異。機車之平均碳重量百分比為19.13 ~ 33.84%；柴油車之平均碳重量百分比為46.85 ~ 63.47%。若以碎形參數及化學元素組成來判別機車與柴油車排放微粒來源之依據，則在本研究中之準確率約為五成，無法辨識之微粒約為三成，此時以碳重量百分比組成來判別無法辨識之微粒時，則可以大幅的提高判別此兩種微粒之準確率。

This study investigated the hurst coefficient and chemical composition of suspended particulates emitted from motorcycles. The engines of the motorcycles were a dynamometer, two two-stroke motorcycles and six four-stroke motorcycles. Two unleaded gasolines (92 and 95) and three speeds of motorcycle engines (3600, 4600 and 5800 rpm) were conducted using the dynamometer. Size, morphology, hurst coefficient and elemental composition of the particles and aggregates generated from dynamometer were investigated with scanning electron microscopy (SEM) and fractal model analysis. Physical and chemical characteristic information of the motorcycle exhaust particles was built to provide valid evidence to identify the pollution sources. Besides, in order to know the present practical motorcycle exhaust condition, this study sampled and analyzed eight motorcycles, and then compared with diesel cars in morphology, hurst coefficient, and elemental composition respectively. According to the 105 magnification images, the morphology of aggregates generated from the dynamometer is very similar to that generated from the eight motorcycles. Most of the primary particles is spherical, and the size distribution was from 10 to 50 nm. The size distribution of the aggregates is from 100 to 1000 nm. The hurst coefficient of aggregates generated from the dynamometer is about 0.118∼0.366 at different engine speed and the mean hurst coefficient is 0.242. Besides, the mean hurst coefficient of aggregates generated from the eight motorcycles is about 0.222∼0.292. As to the comparison with the hurst coefficient of the aggregates from different motorcycles, we found that the mean hurst coefficient of the aggregates from four-stroke engines was higher than that from two-stroke engines. With the increases of exhaust, the hurst coefficient did not have a significant difference. In contrast, with the decrease of the rotation speed of the engines, the hurst coefficient would be enhanced. The elemental composition of the single particle were analyzed by EDS. This result indicated that the carbon(11.37~53.41%), oxygen, silicon and aluminum were the major elements in all of the chemical components. Besides, we also discovered that potassium and iron were the elements which weight percentage was below 5%. Moreover, there were some minor elements like sodium, magnesium, calcium, sulfur, tin and chromium. The hurst coefficient of aggregates generated from the diesel cars is very similar to that generated from the motorcycles. The mean hurst coefficient of aggregates generated from the diesel cars is about 0.194~0.374, and that generated from the motorcycles is about 0.118~0.366. There was a significant difference if we used the weight percentage of carbon component to distinguish the emitted sources of particles. The weight percentage of carbon component of particles emitted from motorcycles was 19.13~33.84%, and that emitted from diesel cars was 46.85~63.47%. If we used the hurst coefficient and elemental composition to judge the particle sources, the rate of accuracy in this research is about 50%, and the particles that couldn’t be distinguished is about 30%. We could improve the rate of accuracy if we used the carbon compostion to judge the particle sources.