結合地理資訊系統與擴散模式評估半導體廠無機酸排放對周遭空氣品質之影響

Abstract

半導體業對台灣經濟發展有重大貢獻，但是由於製程中大量使用各類無機酸原物料，經由工廠排放後，可能對於高科技產業園區周邊居民產生衝擊。本研究蒐集1999年起環保署規定半導體業需申報之無機酸原物料使用量，包括硫酸、硝酸、氫氟酸及鹽酸，以此了解近年來半導體業原物料使用趨勢，並利用地理資訊系統 (Geographic Information System，GIS)及空氣污染擴散模式（Industrial Source Complex Short Term 3rd edition，ISCST3）評估半導體業排放無機酸對當地村里潛在污染區域及影響程度，利用排放係數推估無機酸排放量進行空氣污染擴散模擬，並與周邊環境實測對照比較，提出大氣中無機酸隨季節變化之污染源濃度分佈等資訊。 資料顯示園區內污染源分佈集中，且硫酸在無機酸原物料使用屬最大量，其次為氫氟酸、鹽酸及硝酸，且本研究結果發現2004–2006年無機酸模擬季平均擴散濃度有逐年遞增趨勢。另污染分佈狀況受季節及環境因素影響變化大，以2006年為例，短時間內大氣中硫酸液滴於第二期及第三期區域會有較高模擬濃度，如雙溪村、高峰里、金山里各季最大日模擬濃度分別介於1.40–1.60 µg/m3、1.03–1.31 µg/m3和0.82–1.277 µg/m3。此外，周界六測站平均監測值皆遠高於模擬值，可能源於無法完全掌握其他逸散源的污染貢獻，故其他五測站歷年各無機酸平均模擬值與實測值之R2介於0.17–0.71 （排除污水廠測點），尤其硫酸液滴（R2=0.71）具有良好的相關性，推測園區內半導體業污染源排放，可能影響園區周界大氣環境中無機酸污染濃度。此外，園區下風區域有較高模擬濃度分佈，但該處僅設置一監測站可供比較，顯示目前對於高污染濃度分佈的資訊不足，有必要檢討監測站分佈位置。 年平均模擬值最高值出現於高峰里（0.114 µg/m3）和金山里（0.11 µg/m3），雖其值均未超過美國加州環境衛生健康評估中心列出之硫酸慢性暴露標準（1 µg/m3），但仍須留意相關半導體產業運作對周邊環境的衝擊，並注意四季無機酸於周邊大氣環境中污染濃度分佈，以降低污染物對易感族群的衝擊，本研究顯示運用大氣擴散模式及地理資訊系統可推估工廠污染排放對於週邊環境之影響。

Semiconductor manufacturing industries contribute a lot to the economic growth in Taiwan. However, the semiconductor manufacturing processes use a large amount of inorganic acids, such as sulfuric acid, nitric acid, hydrogen fluoride, and hydrogen chloride. The emission of these acids from the industries may impact the health of local residents near the high-tech park. The chemical usage of H2SO4, HNO3, HF and HCl by the semiconductor manufacturing industries was collected since 1999, which was the year that the industry has to report their use amount required by the regulations of air pollution. We evaluate the potential impact of inorganic acids emitted from the semiconductor manufacturing on air quality in the peripheral regions using air dispersion model (Industrial Source Complex Short Term, ISCST3) and geographic information system (GIS) in this study.the emission rates of the acids used in the ISCST3 were estimated based on the chemical use multiplied by the emission factor from literatures. Furthermore, the simulated values obtained from the air dispersion model were compared with measured concentrations from the air quality monitoring stations. The seasonal variations of acid concentrations in the air were discussed as well. There is a high density emitted stacks in the high-tech park. H2SO4 is the most used acid, followed by HF, HCl and HNO3. The simulated quarterly average concentrations of inorganic acids were increased from 2004 to 2006. There were seasonal variations in the concentrations of acids and their distribution patterns. For example, the simulated levels of H2SO4 in the year of 2006 were higher at the second and the third quarters than these at the first and the fourth quarter in the short-term simulation. The maximum average daily concentration of H2SO4 in the villages of Shuangsi, Gaofong, and Jinshan were 1.40–1.60 µg/m3, 1.03–1.31 µg/m3, 0.82–1.277 µg/m3, respectively. The simulated levels were lower than those measured at monitoring stations; the main cause would bethat not all emission sources were included., such as the waste water treatment plant The multiple regression correlation coefficients (R2) of the simulated and monitored data were ranged from 0.17 to 0.71, H2SO4 (R2=0.71) has the best correlation. The emission of these acids from the industries may more or less impact the air quality nearby the semiconductor manufacturing industries. Besides, only one monitoring site is located at downwind side, and the levels of acids on this site were higher either on measured data or simulated results, indicating that more monitoring sites would be needed at potenitally polluted area. The maximum yearly average concentrations of H2SO4 were at Gaofong and Jinshan, which were 0.114 µg/m3 and 0.11 µg/m3, respectively, at the year of 2006. Although the averaged simulated concentrations were lower than the chronic reference exposure levels (cRELs) set at 1 µg/m3 by the Office of Environmental Health Hazard Assessment (OEHHA),, state of California, USA, it is still a concern on the health impact on susceptible groups. This study demonstrated that the use of an air dispersion model coupled with the GIS can access the potential influences of acid emissions on the air quality from the industries.