|Abstract: ||本研究以質子轉移反應質譜儀(proton-transfer-reaction mass|
spectrometry, PTR-MS)為技術核心建立具快速、高靈敏空氣污染鑑識 方法,應用於工業區空汙問題診斷。PTR-MS 屬於直接進樣式質譜設 備(direct injection mass spectrometry, DIMS),有別於一般傳統層析質 譜法,無需透過前處理與管柱分離機制,可直接快速分析大氣中揮發 性有機化合物(volatile organic compounds, VOCs),質譜性能具有高時 間解析與低偵測極限的優點。本研究將使用兩種類型 PTR-MS,包含 四極柱(PTR-QMS)與飛行時間質譜儀(PTR-ToF/MS)應用敏感性工業 區排放之 VOC 鑑定。
本論文結構分為「質譜儀分析方法與應用技術開發」與「實場觀 測與模式整合」兩大主題;前者針對 PTR-MS 性能建立實驗室測試程 序,後者則探討排放傳送至受體之關係。研究成果可分為三個部分: (1) 實驗室 PTR-MS 測試包含化學物質分析與資料庫建立;(2) 污染 物觸發技術開發與現地實測 (3) 現地污染物觀測結合模式模擬進行 個案分析。
在第一部分研究成果中,實驗選定工業區製程常用之 25 種原物 料(6 種醇類、2 種醛類、5 種酯類、2 種醯胺類、6 種鹵碳化合物、1 種呋喃、3 種芳香烴)為測試對象,以滲透法製備標準氣體,配合階段
稀釋測試 QMS 分析的即時性、穩定性、線性與準確度。QMS 實驗結 果顯示,QMS 對於 25 種化學物質之整體分析變異 RSD 小於 10%、 線性表現(R2)大於 0.99,準確度部分透過計算之理論反應速率常數(k 值)校正後,顯示 QMS 不透過繁瑣的分析校正的半定量分析方法,於 ppb 濃度等級之分析範圍,相對誤差可在 2 倍以內。ToF/MS 同重物 分析實驗部分,以 ToF/MS 針對 8 組同重物進行測試(60 g/mole、62 g/mole、73 g/mole、74 g/mole、90 g/mole、106 g/mole、120 g/mole、 179 g/mole),測試結果發現質量解析度於 5000 m/Δm 下,各組同重物 成功由 ToF/MS 分離鑑別,依據監測質量與質量誤差數據評估,其質 量分辨能力可至小數第二位或至第三位。
在第二部分研究成果中,實驗成功建立 H2S 與 QMS 觸發觀測技 術,並且於選定之三處工業區(桃園煉油廠、新竹工業區、中壢工業 區)執行現地觀測。觸發系統能在突發性高值事件當下立即觸動採樣 裝置保存事件空氣樣本,樣本再攜回實驗室以 GC/MS/FID (gas chromatography/mass spectrometry/flame ionization detection)分析 108 種 VOCs,以鑑定空汙事件之 VOC 組成。在桃園煉油廠 H2S 觸發實 測中,監測期間在採樣閾值 8.5 ppb 下成功觸發 10 個事件樣本,經由 GC/MS/FID 的分析,事件樣本中 C2-C5 烷類占總量之 90%,顯示來工 業區存在逸散量大之排放源。再以同樣概念將 QMS 作為觸發系統,
於新竹工業區與中壢工業區內測試,QMS 發現產業組成混雜的工業 區周界中存在幾種主要 VOCs,包含 methanol、acetaldehyde、acetone、 methyl ethyl ketone (MEK)等常見之有機溶劑,濃度可達數百個 ppb; 再以這些濃度顯著的 VOCs 作為觸發對象,鎖定周界高值事件進行採 樣。新竹工業區部分共採集 17 個事件樣本(觸發閾值:methanol = 250 ppb、acetaldehyde = 60 ppb、MEK = 250 ppb、N,N-dimethylformamide (DMF) = 50 ppb);中壢工業區部分共 13 個事件樣本(觸發閾值: methanol = 160 ppb、acetaldehyde = 100 ppb、acetone = 1000 ppb、MEK = 250 ppb、toluene = 150 ppb);事件樣本顯示含氧有機物質 OVOCs (acetone、MEK、methyl isobutyl ketone (MIBK)、methyl butyl ketone (MBK)、ethyl acetate (EA)、vinyl acetate (VA)、isopropanol (IPA)、methyl tert-butyl ether (MTBE))之比例亦相當高,新竹工業區占 108 種 VOCs 總濃度的 15% - 60%,而中壢工業區則占 108 種 VOCs 總濃度的 32% - 84%,顯示工業區大量 OVOC 排放可能是造成異味事件與空氣品質 不良的元兇之一。
第三部分研究成果中,以雙機監測模式將 QMS 與 ToF/MS 架設 於工業區現址進行同步觀測,執行污染物調查工作,ToF/MS 設置於 排放源中心,作為鑑定周界化學物質之角色;此外,為了快速針對觀 測資料進行污染物鑑定,本研究建立雪球抽樣法(snowball sampling),
用以從龐大的觀測資料快速的篩選出主要 VOC 成分。QMS 則設置於 排放源之下風處,作為攔截廠區污染排放之角色。雙機監測的方法針 對六輕工業區(長春大連)與龍德工業區進行表現。六輕工業區部分, 發現 37 種主要成分,包含醇、醛、酮、酯、有機酸、烯類、芳香烴, 但以 methanol 為主要的排放成分;龍德工業區發現 15 種主要成分, 包含醇、酯、酮、有機酸、芳香烴,以 methyl acetate、ethyl acetate 為主要的排放成分。
PTR-MS 觀測結果會進一步與 PAMS-AQM 大氣模式模擬結合, 解釋污染物之排放源與受體的關係;PAMS-AQM 能於最小 3 x 3 公 里網格解析下,模擬氣象場中污染物之傳輸與擴散行為。六輕長春大 連主要排放物 methanol 之模擬結果顯示,現場污染物濃度明顯受到 當地風場影響,可使 methanol 濃度由 5.5 ppm 瞬間降至數個 ppb。在 龍德工業區部分,主要排放物 methyl acetate 與 ethyl acetate 會因海陸 風因素,每日正午之海風會將污染物往西吹向下風處造成高值事件。
PTR-MS 與傳統線上 GC/MS 於工業區場址執行平行分析工作, 選擇空氣中主要的 VOCs (methanol、acetaldehyde、benzene、toluene、 C2-benzene、MEK、acetone、ethyl acetate (EA))比對,結果顯示 QMS 量測之趨勢變化與 GC/MS 趨勢相當吻合,亦說明 PTR-MS 對於空氣 中之高反應性物質(如 OVOCs)之分析能力優於傳統 GC/MS 方法。;This study is to develop rapid, sensitivity and diagnostic methodology involving the use of proton-transfer reaction mass spectrometry (PTR-MS) to investigate chemical pollutants around industrial parks. PTR-MS is one of direct injection mass spectrometry (DIMS), which can measure trace levels of toxic and odorous volatile organic compounds (VOCs) at the speed of few seconds to minutes without prior chemical separation as in gas chromatography mass spectrometry (GC-MS). Two types of mass spectrometry were involved: quadruple (called PTR-QMS) and time of flight (called PTR-ToF/MS).
This thesis is structured in two themes: 1. development of in-lab VOC testing procedures for PTR-MS; 2. integrating PTR-MS measurements with model simulations to establish source-and-receptor relationship. Major achievements are classified into: (1) in-lab testing of PTR-MS, (2) developing triggered sampling techniques to capture plume events in industrial parks, (3) field measurements with model simulations to diagnose chemical pollution near industrial parks.
Achievement (1): Twenty-five odorous VOC gas streams (6 alcohols, 2 aldehydes, 5 esters, 2 amides, 6 halocarbons, 1 furan and 3 aromatics) were generated via the method of permeation to test PTR- QMS with stepwise dilution for its response to the rapid concentration change. The results showed relative standard deviations (RSDs) smaller than 10% and linearity expressed as R2 better than 0.99 for most of the 25 compounds. If using theoretical k values for estimating concentrations without calibration, the estimated concentrations at ppb level can be accurate within a factor of two, which can qualify PTR-MS as a semi- quantitative method. The test of 8 pairs of isobaric compounds (i.e., 60 g/mole, 62 g/mole, 73 g/mole, 74 g/mole, 90 g/mole, 106 g/mole, 120
g/mole, 179 g/mole) showed that PTR-ToF/MS can successfully separate all these 8 pairs of compounds with mass accuracy down to the second or third decimal point under the 5000 m/Δm mass resolution as claimed by the manufacturer.
Achievement (2): H2S and PTR-QMS triggered sampling apparatuses were constructed in laboratory and tested in three industrial parks (Taoyuan refinery plant, Hsinchu industrial park, Chungli industrial park) to capture event samples using H2S or individual VOCs as the trigger compound. The triggered samples were then analyzed with in-lab GC/MS/FID for 108 VOCs to characterize the chemical composition of the events. Ten event samples were triggered by H2S analyzer at the 8.5 ppb threshold value at Taoyuan refinery plant to address the foul smell problem. The result shows that C2-C5 alkanes are the major component for the event samples (90% of total l08 VOCs), which is consistent with the notion that the volatile part of hydrocarbons are released as fugitive emissions from refinery. The same concept was tested in field at both Hsinchu and Chungli industrial parks using PTR-QMS as the trigger device. Several prominent compounds (such as methanol, acetaldehyde, acetone, methyl ethyl ketone (MEK), etc.) were found and thus used as the trigger gases. Seventeen event samples were triggered at Hsinchu (threshold value: methanol = 250 ppb, acetaldehyde = 60 ppb, MEK = 250 ppb, N,N-dimethylformamide (DMF) = 50 ppb). Thirteen event samples were triggered at Chungli (methanol = 160 ppb, acetaldehyde = 100 ppb, acetone = 1000 ppb, MEK = 250 ppb, toluene = 150 ppb). It were found that OVOCs (acetone, MEK, methyl isobutyl ketone (MIBK), methyl butyl ketone (MBK), ethyl acetate (EA), vinyl acetate (VA), isopropanol (IPA), methyl tert-butyl ether (MTBE)) accounted for 15% - 60% of the total amount of 108 VOCs in Hsinchu, and 32% - 84% in Chungli. Such large fractions of OVOCs in the air could explain the frequent odor complains reported in the vicinity of the industrial parks.
Achievement (3): Both QMS and ToF/MS were employed simultaneously to form a dual-instrument deploying approach in the field. With such an approach, the ToF/MS with its superior mass resolution was placed in the industrial park to identify characteristic pollutants. To facilitate rapid compound identification, a searching method called “snowball sampling” was designed to rapidly find prominent compounds buried in a large mass dataset. Simultaneously, QMS was deployed downwind to register pollution events. The dual-instrument approach was conducted in Mailiao and Longde industrial parks. In Mailiao, 37 prominent compounds (including alcohols, aldehydes, ketones, organic acids, alkenes and aromatics) were found, and the methanol’s level topped the compound list. In Longde, 15 prominent compounds (including alcohols, aldehydes, ketones, organic acids and aromatics) were found, and methyl acetate and ethyl acetate had the highest levels.
Atmospheric modeling was coupled with chemical measurements with PTR-MS to establish the source-to-receptor transport. The PAMS- AQM can provide overall transport and dispersion of pollutants due to meteorology with a fairly coarse resolution of 3 x 3 km. This approach was employed in industrial parks to explain the observed events of methanol (Max. = 5.5 ppm) in Mailiao in the local wind field. In another experiment, ethyl acetate and methyl acetate were the two major pollutants in Longde. The simulations successfully explained the events of the two esters observed at the downwind site due to the land-sea- breeze.
PTR-MS was validated in the field by conventional in-situ GC/MS. The concentration variations of prominent VOCs (methanol, acetaldehyde, MEK, EA, acetone, benzene, toluene, C2-benzene) in QMS
measurements agreed well with the variations of GC/MS method at sub- ppb level. The validation by GC-MS gave confidence to PTR-MS measurements of more reactive VOCs, such as OVOCs, which the conventional GC-MS method is more difficult of measuring.