| 摘要: | 有害空氣污染物 (Hazardous Air Pollutants, HAPs) 對民眾健康危害影響為近年來備受關注的議題，行政院環保署於民國106年預告「固定污染源有害空氣污染物排放標準」草案，優先管制72項空氣污染物。
為了即時掌握HAPs排放來源及有效掌握大氣中HAPs濃度藉以評估健康風險，本研究嘗試開發一套熱脫附氣相層析質譜儀在線分析方法，可獲得28種HAPs濃度小時連續數據。
本研究利用前濃縮儀 (Thermal Desorption, TD) 串連GC/MS成為一在線系統 (on-line TD-GC/MS) ，並因應台灣高濕度環境，於前端加裝除水系統，以環檢所NIEA A715.15B標準方法為基礎，建立相關品保品管規範，其中包含檢量線建立、偵測極限、準確度、精密度等。檢量線結果顯示，RSD值介於2.59% ~ 27.3%之間，R2值介於0.993 ~ 1.000，經預估確認得方法偵測極限為0.11 ~ 1.76 ppb，而準確度與精確度結果，其回收率介於72.40% ~ 130.50%之間，RSD值介於1.42% ~ 16.94%。
建立完成相關條件後，本研究分別於北部工業區及南部工業區附近進行兩場次實地監測，時間分別為2019年10月5日至11月7日及2020年03月17日至04月29日，兩個實場測試結果皆有測到地區性特殊事件並與其他分析方法進行趨勢比較以為驗證。在平均濃度上的比較，第一場次偵測到的總濃度為10.05 ppb，而第二場次濃度則為71.12 ppb，由於第一場次為東北風季節，工業區位於下風處，導致於第一場次的監測濃度較低；而第二場次監測期間發生多次高值現象平均後造成污染物濃度偏高。
在驗證on-line TD-GC/MS方法之可靠度上，我們使用採樣罐進行隨機採樣。本研究於第一場次採樣結果發現2019/10/28中午11點左右on-line TD-GC/MS偵測到CCl4之瞬間高值，約為2.74 ppb，而採樣罐亦在當下恰巧採集到CCl4之濃度約為2.3 ppb，連續式監測結果與隨機式採樣結果比對濃度相近，可證實本研究Online GC/MS於連續監測上的表現優良，相較於離線式隨機採樣，不僅可得知環境濃度趨勢及未知成分，亦可以捕捉到突發高值現象，有效彌補離線式採樣盲目、隨機的缺點。
質譜儀雖然具有強大的定性及定量能力，但經由兩次實場測試中發現，離子源劣化的速度相當快，本研究利用內標準品進行長期的比對，結果顯示，當每次更換新的離子源後，內標的感度約在四至五天會下降80%，而本次分析物中含有氟氯碳化物，其濃度於大氣中穩定存在，故常被作為環境內標，而本研究將真實採樣時偵測到的大氣內標CFC-11與原始內標進行趨勢比對，可發現下降趨勢與原始已知濃度之內標相同，可以佐證感度下降的主因與儀器本身有關。然而在尚未開發出可以更持久性使用的離子源前，離子源應例行性7-10天清洗一次。未來將繼續改善離子源相關問題，提出更有效的連續分析方法，達到周界連續監測並改善維護保養的難度。
;Hazardous Air Pollutants (HAPs) are closely associated with public health. This issue has received much attention in recent years. In 2017, the Environmental Protection Administration (EPA) of Taiwan announced the draft of “Emission Standards for Hazardous Air Pollutants from Stationary Pollution” to control 72 HAPs. In order to immediately control the source of HAPs emissions and effectively access ambient concentrations of HAPs, the development of a method to continuously monitor HAPs in atmosphere with gas chromatography mass spectrometry (GC/MS) to provide hourly continuous data of HAPs has undergone intensive research by our laboratory funded by Taiwan EPA.
In this research, we used the thermal desorption (TD) technique for pre-concentration to couple with GC/MS. Due to the perennial high humidity on the island, water removal prior to TD-GC/MS analysis is critical. Our under-developed on-line TD-GC/MS method for HAPs is mainly based on the existing NIEA A715.15B method in terms of the requirements of quality control/assurance for target HAPs. These requirements include concentration calibration, as well as the assessment of accuracy (recovery), precision and method detection limits (MDL). Our QA results show that the relative standard deviations (RSD) range from 2.59% to 27.3% The linearity (R2) ranges 0.993 ~ 1.000. The MDL range 0.11 ~ 1.76 ppb. The accuracy (recovery) precision (relative standard deviations %) is 72.40% ~ 130.50% and 1.42% ~ 16.94%.
After the completion of QA assessment in the laboratory, a month-long field measurement was conducted from October 5 to November 7, 2019 in Taoyuan city and March 17 to April 29, 2020 in Kaohsiung city. Both were targeting industrial emissions and had detected industrial related species on several occasions. The field measurement was inter-compared with other methods of PTR-QMS and GC-FID with comparable results observed. The average concentration of all species summed during the first field measurement period was 10.0 ppb compared to 71.1 ppb for the second. The much lower level in the first period was due to the fact that measurement was conducted during the northeasterly monsoon season and the monitoring site was at the upwind of the target industrial zone. The second period observed higher levels of pollutants as well as more local emission events due to the favorable location and meteorology.
In order to ensure the credibility of the on-line TD-GC/MS data, we also conducted canister sampling at both sites. It was found that carbon tetrachloride which was banned by the Montreal Protocol, but detected by both the canister sampling and TD-GC/MS on October 28, 2019 with comparable results of 2.3 ppb vs. 2.7 ppb, respectively, which effectively validated the online TD-GC/MS method. This online method not only can provide the temporal trends and identify unknown compounds, but also capture the unexpected emission episodes.
Although the mass spectrometer has strong qualitative and quantitative capabilities, a major problem with the GC/MS method was found in the on-line mode. The rapid formation of oxides on the filament of the ion source in MS deteriorates the sensitivity over time. Based on the signals of the internal standards the sensitivity dropped 80% within the first 4-5 days. In the study, we also used chlorofluorocarbons, whose concentration are stable in the atmosphere as the "intrinsic" internal standards. The signal of CFC-11 (CCl3F) was compared with those of the originally added internal standards, and the same downward trend as those of the original internal standards also appeared, confirming that the decrease in sensitivity is owing to the MS detection. Frequent cleaning the ion source every 7-10 days was adopted as a tentative solution before a more permanent one developed by the MS industry becomes available. |
