以QuEChERS萃取搭配極致液相層析/串聯式質譜儀分析食品中全氟碳化合物、鄰苯二甲酸酯、壬基酚及雙酚A

Abstract

全氟碳化合物(perfluoroalkyl substances, PFASs)、鄰苯二甲酸酯 (phthalate esters, PAEs)、壬基酚(nonylphenol, NP)以及雙酚A (bisphenol A, BPA)為環境中普遍存在之新興污染物，廣泛被應用於許多消費性和工業產品，例如食品包材、塑膠製品、個人保健用品、介面活性劑等。這些污染物對生殖和發育有不良影響，且可能干擾內分泌系統。日常生活中人們可能不斷且同時暴露於這些物質，其中水與食物是重要的暴露途徑。藉由分析不同種類食材，可了解上述汙染物進入食物鏈之潛勢，也可推估藉由攝食途徑之人體暴露；然而，迄今能於食材中同時檢測上述污染物之分析方法仍有限。因此，本研究開發一分析方法檢測六類食物中十種全氟碳化合物、六種鄰苯二甲酸酯、壬基酚及雙酚A；食材包括豬肉、豬肝、豬腎、魚肉、蛤蜊與牡蠣，為較容易累積上述物質之基質。樣本前處理技術使用QuEChERS（Quick, Easy, Cheap, Effective, Rugged and Safe），以乙腈作為萃取溶劑，並使用增強型基質去除脂質(EMR-Lipid)吸附劑進行樣本淨化，樣本濃縮後以極致液相層析串聯式質譜儀並搭配同位素稀釋技術定量分析，以多重離子監測模式獲取質荷比資訊。六種鄰苯二甲酸酯以正離子電灑游離法作為游離源，並利用Ascentis Express F5管柱於液相層析儀中搭配移動相：(A) 5mM醋酸胺水溶液(pH = 6.56)、(B)甲醇，進行梯度流析；十種全氟碳化合物、壬基酚和雙酚A以負離子電灑游離法作為游離源，並利用BEH C18管柱，於液相層析儀中搭配移動相：(A) 10mM 甲基嗎啡林(pH = 9.6)、(B)甲醇，進行梯度流析。 樣本前處理最佳化測試結果顯示，在液液萃取步驟中酸化水相或添加分散劑，並無提高大部分待測物之萃取效率，故在萃取時不需加酸和分散劑進行輔助。在分散式固相萃取（d-SPE）淨化步驟中，增強型基質去除脂質（EMR-Lipid）較一級二級胺 (Primary Secondary amine, PSA)，不減損太多待測物，且可去除可能干擾分析的食物基質，故選擇EMR-Lipid。樣本量與液液萃取上清液體積提取的部分，相較於兩克與提取全部體積(約九毫升)，一克與四毫升提供較低之離子抑制效應，故較適合用於分析。豬肉、豬肝、豬腎、魚肉、蛤蜊與牡蠣之基質效應因子分別為63.6-168%、43.0-147%、63.7-153%、60.9-198%、63.5 -149%與50.8-153%。萃取效率分別為14.7-96.8%、50.9-95.7%（除鄰苯二甲酸二乙酯與鄰苯二甲酸丁基苯酯萃取效率為10.6%與2.82%）、29.8-93.7%（除鄰苯二甲酸丁基苯酯萃取效率為1.99%）、20.0-104%、22.3-105%與17.6-106%。豬肉、豬肝、豬腎、魚肉、蛤蜊與牡蠣中待測物方法偵測極限分別為0.17-9.70 ng/g（濕重）、0.23-17.9 ng/g（濕重）、0.16-11.2 ng/g（濕重）、0.21-8.35 ng/g（濕重）、0.17-9.53 ng/g（濕重）與0.21-8.61 ng/g（濕重）。此方法測試五種添加濃度於六種食物之定量偏差與相對標準偏差；除鄰苯二甲酸二(2-乙基己基)酯、鄰苯二甲酸二異壬酯、鄰苯二甲酸二異癸酯及壬基酚受背景值影響較大，定量偏差大多均低於30%，相對標準差多數均低於20%。 本分析方法用於調查新竹地區三處大型傳統市場之六類食材中全氟碳化合物、鄰苯二甲酸酯、壬基酚與雙酚A之含量。除吳郭魚，所有食材樣本均測得鄰苯二甲酸二異壬酯，濃度範圍落在124-908 ng/g (濕重)。鄰苯二甲酸丁基苯酯見於豬肝與豬腎中，濃度範圍為33.1-177 ng/g (濕重)。長碳鏈全氟碳化合物在豬肝中偵測到低濃度。8碳以下之全氟羧酸化合物、雙酚A與鄰苯二甲酸二辛酯在所有食材中皆無發現。此外，豬肝為較受污染之基質，共檢測到8種待測物(全氟壬酸、全氟癸酸、全氟十一烷酸、全氟十二烷酸、全氟己烷磺酸、全氟辛烷磺酸及鄰苯二甲酸丁基苯酯、鄰苯二甲酸二異壬酯)。由本次調查可見部分待測物進入食物鏈之情勢，需要透過較大規模調查，加以瞭解這些污染物在食品中之分佈，以及人體藉由攝食途徑之可能暴露。

Perfluoroalkyl substances (PFASs), phthalate esters (PAEs), nonylphenol (NP), and bisphenol A (BPA) are emerging contaminants and ubiquitous in the environment. These compounds are widely used in many consumer and industrial products such as food container, plastics, personal care products and surfactants. They are reported to have adverse effects on reproduction and development and disrupt endocrine system. Moreover, the general population is continuously and simultaneously exposed to them in daily life; food is an important exposure route. Hence, it is crucial to investigate the levels of the above compounds in foods to realize the possible exposure for humans through food intake. However, limited methods are available to determine these contaminants together. Therefore, this study developed and validated a method for analyzing ten PFASs, six PAEs, NP and BPA in six types of foods, including pork, pork liver, pork kidney, fish, clams and oyster. The sample preparation technique QuEChERS (Quick, Easy, Cheap, Effective, Rugged and Safe) was used with acetonitrile as extraction solvent, and Enhanced Matrix Removal (EMR) – Lipid adsorbent was used for sample cleanup. After concentrations, the samples were injected onto ultra-performance liquid chromatography/tandem mass spectrometry (UPLC-MS/MS) with multiple reaction monitoring (MRM) and were quantified with isotope-dilution techniques. Six PAEs were separated on an Ascentis Express F5 column with mobile phases composed of (B) methanol and (A) 5 mM ammonium acetate(aq) (pH = 6.56), and were ionized with positive electrospray ionization (ESI+). The rest analytes were separated with a BEH C18 column with mobile phases composed of (B) methanol and (A) 10-mM N-methylmorpholine(aq) (pH = 9.6), and were ionized with ESI-. The results of optimization on QuEChERS sample preparation showed that acidification of aqueous phase and use of dispersant methanol during liquid-liquid extraction (LLE) didn’t improve extraction efficiencies of most analytes. Besides, EMR-Lipid was superior to the primary secondary amine (PSA) as cleanup sorbents since less analytes were lost during dispersive solid phase extraction (d-SPE) cleanup step. One-gram sample size and 4 mL of supernatant taken for cleanup after LLE offered lower ion suppression (IS%) compared with two-gram sample sizes and all the acetonitrile extract (about 9 mL). The matrix effect factors of pork, pork liver, pork kidney, fish, clams and oyster were 63.6-168%, 43.0-147%, 63.7-153%, 60.9-198%, 63.5-149% and 50.8-153%. The extraction efficiencies of six foods were as follows: 14.7-96.8%, 50.9-95.7% (except for DEP and BBP extraction efficiency 10.6% and 2.82%), 29.8-93.7% (except for BBP extraction 1.99%), 20.0-104%, 22.3-105%, and 17.6-106%. The limits of detection (LODs) of analytes in six foods were as follows: 0.17-9.70 ng/g wet weight (w.w.), 0.23-17.9 ng/g w.w., 0.16-11.2 ng/g w.w., 0.21-8.35 ng/g w.w., 0.17-9.53 ng/g w.w. and 0.21-8.61 ng/g w.w., respectively. This study tested the method accuracy and precision at five spiked levels in these foods, and most of the quantitative bias were lower than 30%, except DEHP, DINP, DIDP and NP, which were affected by the background levels; most relative standard deviations were below 20%. This method was applied to investigate analytes in foods collected from three major traditional markets in Hsinchu City. DINP was detected in most food samples with concentrations ranging from 124 to 908 ng/g w.w. BBP were found in pork liver and kidney with levels from 33.1 to 177 ng/g w.w. Most long-chain PFASs were detected in the pork liver with concentrations of sub- to few ng/g w.w.. Perfluoroalkyl carboxylates (PFACs) containing eight or less carbons, BPA and DNOP were not detected in all the food samples. Pork liver was much contaminated with eight analytes (PFNA, PFDA, PFUnDA, PFDoDA, PFHxS, PFOS, BBP and DINP). It is necessary to have additional survey in a large scale to realize the distribution of these contaminants in foods, and to get a better understanding of the possible exposure from food intake.