以極致液相層析/串聯式質譜儀偵測水體中雌激性 化合物之方法開發與應用

Abstract

分布於環境水體的雌激性內分泌干擾物質 (estrogenic enocrine-disrupting chemicals, EDCs)，濃度雖低，卻可能足以改變水生生物的生殖表現。這些受汙染的表面水若進入飲用水源，可能干擾人類的內分泌系統。目前這類活性物質在各淨水處理單元的含量和移除所知有限，主因為其所含之EDCs濃度甚低，不易分析。因此，本研究目的是開發極致液相層析/串聯式質譜儀 (UPLC/MS/MS) 結合化學衍生，並透過層析技術降低真實樣品的基質干擾，以改進偵測在水環境中EDCs的靈敏度。 EDCs若具有酚類官能基，可與丹磺酰氯 (dansyl chloride) 或五氟溴甲苯(pentafluorobenzyl bromide, PFBBr) 化學試劑進行衍生反應，以增強待測物在液相層析/串聯式質譜儀之游離效率，改善偵測之靈敏度。本研究針對雌素酮 (estrone, E1)、動情激素 (17β-estradiol, E2)、雌素醇 (estriol, E3)、乙炔動情激素 (17α-ethinyl estradiol, EE2)、壬基酚 (4-nonylphenol, NP)、辛基酚 (4-tert-octylphenol, OP)、雙酚A (bisphenol A, BPA) 以及合成丹磺酰氯和五氟溴甲苯衍生之待測物，進行電灑游離法 (electrospray ionization, ESI)、大氣壓化學游離法 (atmospheric pressure chemical ionization, APCI)、大氣壓光游離法 (atmospheric pressure photoionization, APPI)，以及APCI/APPI雙重游離等四種游離源之靈敏度比較，同時評估極致液相層析管柱 (UPLC column)、管柱後分流 (post-column split)、複合式管柱 (mixed-mode column) 以及二維層析 (2D-LC) 四種方法之基質效應與方法表現。 本研究採集原水 (raw water) 基質，於固相萃取後添加標準品 (post-extraction addition)，以Waters 極致液相層析 (Acquity UPLC) 結合 Waters Quattro Premier XE三段式四極棒質譜儀 (triple-quadrupole mass spectrometer) 分析。結果顯示，未衍生之待測物在電灑游離法、丹磺酰氯衍生之待測物在三種游離法 (除大氣壓光游離法之外) 和五氟溴甲苯衍生之待測物在大氣壓化學游離法，於上述四種層析方法之間，其基質效應並無顯著差異。 丹磺酰氯衍生之待測物於電灑游離法具有優異之偵測靈敏度，其次是大氣壓光游離法，兩者訊號強度相較於未衍生之待測物均可提升三個數量級 (order of magnitude)。五氟溴甲苯衍生之待測物，其訊號強度亦可高於未衍生之待測物一個數量級。丹磺酰氯衍生之待測物在四種游離源下的訊號強度均優於五氟溴甲苯衍生之待測物。待測物在雙重APCI/APPI游離法和單一APCI游離法的訊號強度旗鼓相當，顯示雙重游離模式並未增強離子化效率。整體而言，以電灑游離法配合丹磺酰氯衍生技術並採用極致液相層析下，擁有最佳的靈敏度且相對上基質效應較低。 本研究進一步以UPLC/ESI(+)/MS/MS結合丹磺酰氯衍生技術，採集河水和汙水處理廠之放流水進行方法驗證 (method validation)；丹磺酰氯衍生之待測物 (E1, E2, E3 and EE2) 採用選擇性離子偵測 (selected ion monitoring, SIM) 和選擇性反應偵測 (selected reaction monitoring, SRM) 的on-column偵測極限，定義為三倍的訊雜比，分別為 0.44-1.5 pg以及0.05-0.20 pg，而污水處理廠之放流水和河水採選擇性反應偵測的偵測極限分別為0.23-0.52 ng/L 以及 0.56-0.91 ng/L。 本研究亦針對五個代表性飲用水淨水廠於不同單元之樣本進行檢測。類固醇雌激素於原水採選擇性離子偵測和選擇性反應偵測時之偵測極限分別為0.20-0.68 ng/L 和0.04-0.15 ng/L。原水中類固醇雌激素濃度範圍介於 < LOD-5.5 ng/L，透過多步驟淨水處理單元已明顯降低，清水中的濃度範圍則介於 < LOD-1.17 ng/L。壬基酚於各淨水單元皆能檢出，其濃度變異較大，介於25-378 ng/L; 而清水中的壬基酚濃度則低於83 ng/L。辛基酚和雙酚A於各淨水單元的濃度，則近於分析背景值。儘管各水廠水源不同，原水中類固醇雌激素濃度相似; 透過傳統處理或增設高級處理程序，清水中的濃度並無顯著差異。 開發偵測水體中雌激性化合物定性兼定量分析方法，加速樣本液相層析分離的速度 (< 5 min)，而且在搭配化學衍生後，環境樣本採用選擇性離子偵測模式 (單一質譜) 的偵測極限亦可達sub-ng/L，且其所定量之濃度與選擇性反應偵測模式 (串聯式質譜) 所量測者相符為本研究重要貢獻。此外，本研究發展之相關分析技術可拓展至其他基質或化合物，例如組織、血液和食品樣本，為微量有機分析另闢蹊徑。大規模的飲用水廠檢測證實EDCs確實存在飲用水中，特別是雌素酮、乙炔動情激素和壬基酚化合物。水處理技術成本高，對於這些微汙染物 (濃度接近偵測極限) 能再進一步改善移除效果的空間有限；相較之下，原水水源的保護和降低汙染來源可能是更為有效的作法。

Estrogenic endocrine-disrupting chemicals (EDCs) are widely distributed over the aquatic environment and may interfere with reproductive functions of aquatic creatures even at trace levels. EDCs may also disturb the endocrine system of human beings if they exist in drinking water. Currently little is known about the levels and removal of these bioactive substances in each step of drinking water treatment. One of the major obstacles to the relating studies is how to detect trace amount of EDCs in complex environmental matrixes. This study aimed at employing ultra-performance liquid chromatography/tandem mass spectrometry (UPLC/MS/MS) coupled with chemical derivatization and decreasing the matrix effects from real samples through chromatographic technique to improve the detection sensitivity of EDCs in water. Derivatization of phenolic EDCs using dansyl chloride and pentafluorobenzyl bromide (PFBBr) can enhance the ionization efficiency and improve the sensitivity on MS. Consequently, this study compared the sensitivities and matrix effects of four ionization methods combined with four liquid chromatographic systems on estrone (E1), 17β-estradiol (E2), estriol (E3), 17α-ethinylestradiol (EE2), 4-nonylphenol (NP), 4-tert-octylphenol (OP), bisphenol A (BPA) and their derivatives of dansyl chloride or PFBBr. The four ion sources were electrospray ionization (ESI), atmospheric pressure chemical ionization (APCI), atmospheric pressure photoionization (APPI) and APCI/APPI, respectively; the four liquid chromatography (LC) systems were UPLC method with and without post-column split (5:1), a mixed mode column and two-dimensional LC (2D-LC) method. This study evaluated the matrix effects and the performance of ionization methods using raw water with post-extraction spike. The analysis was carried out using a Waters Acquity UPLC coupled with a Waters Quattro Premier XE triple-quadrupole mass spectrometer. The results showed no significant differences in matrix effects among native analytes using ESI, dansylated analytes using each ion source except for APPI, and PFBBr derivatives using APCI for the above four LC methods. Dansylated compounds produced the most intense signals using ESI and then APPI, which were both up to three orders of magnitude than that without derivatization. PFBBr derivatization increased the signal intensity of analytes up to one order of magnitude than those of underivatized ones. The signal intensity of dansylated compounds was higher than those of PFBBr derivatives. The response of analytes using APCI/APPI mode was similar to that at APCI mode; the dual-source ionization did not improve the signals. Dansylated compounds using ESI under UPLC system yielded the strongest signals and were less susceptible to matrix effect; consequently, this combination was further evaluated and validated by sewage treatment plant effluents and river water. The on-column detection limit of dansylated compounds (E1, E2, E3 and EE2) using selected ion monitoring (SIM) and selected reaction monitoring (SRM) modes, defined as a signal-to-noise ratio of three, ranged from 0.44 to 1.5 pg and from 0.05 to 0.20 pg, respectively. The limits of detection (LODs) with SRM in sewage treatment plant effluents and river water were 0.23-0.52 and 0.56-0.91 ng/L, respectively. Using the developed method, the study surveyed the levels of these EDCs in different treatment steps of five representative drinking water treatment plants. The LODs of steroid estrogens in raw water with SIM and SRM were 0.20-0.68 ng/L and 0.04-015 ng/L, respectively. The four steroid estrogens in raw water and finished water ranged from < LOD to 5.5 ng/L and < LOD to 1.17 ng/L, respectively. NP survived in each treatment step and the concentrations ranged from 25 to 378 ng/L, and its levels in finished water were all below 83 ng/L. The levels of OP and BPA in most samples were close to the backgrounds in lab blanks. The levels of the steroid estrogens in raw water were similar among the five plants even though their locations and water sources are different. Most of the estrogenic chemicals can be removed effectively through drinking water treatment steps except for NP, and there were no differences in the removal efficiency between conventional steps only and those with additional advanced procedures. In conclusion, this study developed and validated a qualitative and quantitative method on detecting EDCs with dansyl chloride derivatization by UPLC/MS and UPLC/MS/MS. The new method provided a fast chromatographic separation (< 5 min) and the LODs of the EDCs in environmental waters using SIM reached low-ng/L, and the quantitative results were comparable with those of SRM. In addition, levels of the analytes were not significantly different in the finished water between conventional and advanced treatment processes; the four estrogens at levels of few ng/L cannot be further reduced after the treatment of activated carbon adsorption or reverse osmosis process. It is still essential to prevent water sources from contaminations of EDCs.