建立台灣光電半導體產業之員工於製程中受化學品暴露之健康風險評估流程

Developing a Health Risk Assessment Process of Workers Exposed to Chemicals in the Taiwanese Optoelectronic Semiconductor Industry

黃紹祖(Shao Zu Huang)

指導教授 ： 吳焜裕

國立臺灣大學/公共衛生學院/公共衛生碩士學位學程/碩士(2015年)

https://doi.org/10.6342/NTU.2015.01092

摘要

本研究的目的是建立並操作一套於光電半導體產業適用之健康風險評估流程，以評估長期暴露於黃光製程中所使用的化學品可能造成的健康風險。研究最初先以化學品分級管理來篩選出高風險的化學物質。健康風險評估則採用美國環保署(US EPA)的人類健康風險評估流程進行。暴露評估使用了Stoffenmanager暴露模式工具以預測在該暴露情境下員工受暴露的濃度。如果工廠已有收集作業環境測定資料，該數據則可用於建立機率分佈，並將暴露模式的估計濃度作為前提假設分佈，利用馬爾可夫鏈蒙地卡羅法進行貝氏統計更新。最後，再以更新後的分佈計算出每個化學品的風險係數（Hazard Index, HI）。篩選過程中選出常用於黃光製程的有機混合物光阻劑作為主要評估的對象。光阻劑的確切化學組成因為商業機密的關係而資訊有限。本研究根據現有文獻和專利中的描述推測光阻劑裡的潛在化合物以及揮發物。該光阻劑的主要成分為丙二醇單甲醚乙酸酯 (PGMEA)、酚醛樹脂和光活性化合物 (DNQ)。潛在的揮發物包括苯酚，甲酚，苯，甲苯，和二甲苯。個化合物的劑量反應關係經重新評估後依據是否有新的文獻資料來判斷是否需重新評估。各個揮發物的HI 以及其第95個百分位值為苯酚: 2.098 (95%: 6.59)、苯11.08 (95%: 34.75)、甲酚: 0.2868 (95%: 0.9025)、PGMEA: 0.13 (95%: 0.2628), 甲苯: 0.02764 (95%: 0.087)、以及二甲苯: 0.01021 (95%: 0.0.032)。因為PGMEA是唯一有作業環境測定數據的化合物，其他化合物的HI只依排名來評估應優先關注的化合物。其中又以苯酚和苯的HI和致癌風險最高，顯示風險管理者應考慮優先將分配資源來驗證該化合物在工作環境中的暴露情況。

關鍵字

黃光製程 ； 光阻劑 ； 健康風險評估 ； Stoffenmanager 暴露模式 ； 貝氏統計

並列摘要

The objective of this study was to develop and demonstrate a health risk assessment process for use in the optoelectronic semiconductor industry to assess the potential health risks of chronic exposure to chemicals used in photolithography. Chemical control banding tool was first used to screen out high risk chemicals to be assessed. Human health risk assessment was conducted using the United States Environmental Protection Agency (U.S. EPA) risk assessment process. For exposure assessment, exposure modeling tool Stoffenmanager was used to predict an estimate of worker’s exposure given an exposure scenario. If the factory has collected personal sampling data, the information was used, along with the exposure model estimation as the prior distribution, to set up a Bayesian statistical update using Markov Chain Monte Carlo method. The updated distribution of the estimated exposure was then used to determine the hazard index (HI) of each compound. A photoresist had the greatest risk according to control banding. Its composition and by-products were determined by existing literatures and patents descriptions. Its main composition include Propylene glyco methyl ether acetate (also known as 1-Methoxy-2-propanyl acetate, or PGMEA), novolac resin, and photoactive compound (DNQ). Potential by-products of the photoresist include Phenol, Cresol, Benzene, Toluene, and Xylene. The dose response of each compound was re-assessed and updated when possible. Their reference concentrations and hazard indexes were determined. The resulting HIs and their 95th percentiles for each compound was 2.098 (95%: 6.59) for phenol, 11.08 (95%: 34.75) for benzene, 0.2868 (95%: 0.9025) for cresol, 0.13 (95%: 0.2628) for PGMEA, 0.02764 (95%: 0.087) for xylene, and 0.01021 (95%: 0.0.032) for toluene. Because PGMEA was the only compound that had sampling data, the HI of the other compounds were only ranked to determine compounds that should take precedence validation. Phenol and benzene in particular had the highest scores of HI and cancer risk, suggesting call for further validations.

並列關鍵字

Photolithography ； Photoresist ； Risk Assessment ； Stoffenmanager ； Bayesian Statistics

參考文獻

9. 勞動部職業安全衛生署, 危害性化學品評估及分級管理辦法總說明. 2014.

1. Chen, H.W., Exposure and health risk of gallium, indium, and arsenic from semiconductor manufacturing industry workers. Bull Environ Contam Toxicol, 2007. 78(1): p. 5-9.

2. Chepesiuk, R., Where the chips fall: Environmental health in the semiconductor industry. Environmental Health Perspectives, 1999. 107(9): p. A452-A457.

3. Kinoulty, M., & Williams, N. , Occupational health provision and health surveillance in the semiconductor industry. Occupational Medicine (Oxford, England), 2006. 56(2): p. 100-101.

5. Chelton, C.F., Glowatz, M., & Mosovsky, J. A. , Chemical hazards in the semiconductor industry. IEEE Transactions on Educations, 1991(34(3)): p. 269-288.