以極致液相層析串聯式質譜術分析食品中鄰苯二甲酸酯類與雙酚類化合物—樣品前處理方法優化

Abstract

鄰苯二甲酸酯 (phthalate esters, PAEs) 為目前廣泛使用之塑化劑 (plasticizer)，常用於軟化聚氯乙烯 (polyvinyl chloride, PVC) 產品。雙酚類 (bisphenols, BPs) 經常作為聚碳酸酯 (polycarbonate, PC) 與環氧樹脂 (epoxy resins) 之合成原料，並用於食品接觸容器具與金屬罐頭之內部塗層中。鄰苯二甲酸酯與雙酚類皆為已知的內分泌干擾物質 (endocrine-disrupting chemicals, EDCs)，其不僅會由農作物栽種環境中吸收，還會因食品包材之滲漏而污染食品，最終人體透過攝食方式暴露這些化合物，進而造成健康危害。本實驗室先前開發了以極致液相層析搭配串聯式質譜儀同時檢測不同食品中鄰苯二甲酸酯與雙酚類化合物之方法，但因鄰苯二甲酸酯與雙酚類之背景濃度干擾，以及不佳的淨化效率，影響了待測物的定量，故本研究欲針對先前開發之樣品前處理方法進行優化。 1 g 之均質食品樣本於自動加壓流體萃取系統 (Energized Dispersive Guided Extraction, EDGE) 中進行兩個循環之萃取，每個循環使用 5 mL 的丙酮，於 65℃ (30–34 psi) 萃取 5 分鐘。萃取完成後，取 4 mL 萃取液混合 1 mL 去離子水，通過經甲醇與丙酮預洗過之 Captiva EMR-Lipid 固相萃取匣進行淨化。將淨化液濃縮至 2 mL，並經0.2-μm PTFE 濾膜後，以 Waters I-Class PLUS極致液相層析搭配Waters Xevo TQ-XS 串聯式質譜儀分析食品中 12 種鄰苯二甲酸酯與 6 種雙酚類。本研究使用 Ascentis Express F5 (30 × 2.1 mm, 2.0 μm) 管柱進行層析， UniSpray 為游離源。12 種鄰苯二甲酸酯與雙酚A二環氧甘油醚 (bisphenol A diglycidyl ether, BADGE) 於正電模式下進行梯度流析，管柱溫度 35℃，有機移動相為甲醇，水性移動相為 5 mM 醋酸胺水溶液/0.1% 醋酸水溶液 (pH 4.19)，流速為 0.5 mL/min，總層析時間需 8.5 分鐘。其餘 5 種雙酚類於負電模式下進行檢測，管柱溫度為 40℃，有機移動相為甲醇，水性移動相為 Milli-Q water，流速為 0.5 mL/min，層析時間共需 7.35 分鐘。標準品配製於甲醇中，進樣體積為 2 μL，而樣品溶液丙酮/水 (1:1, v/v) 混合溶液，進樣體積為 4 μL。標準品檢量線之線性範圍介於 0.1–500 ng/mL，決定係數 r2 皆達到 0.994 以上，儀器偵測極限介於 1.04–2,658 fg，儀器定量極限則在 3.33–8,861 fg 之間。 在 Captiva EMR-Lipid 淨化過程中，選用丙酮/水 (80:20, v/v) 作為溶劑，可保留較乙腈/水 (80:20, v/v) (9.0–70.3%) 多的高分子量鄰苯二甲酸酯 (high molecular-weight PAEs, HMW PAEs) 於淨化液中，介於 47.1–81.4% 之間。 Captiva EMR-Lipid 固相萃取匣本身具高鄰苯二甲酸酯背景濃度，本研究測試了三種預洗條件： (1) 4 mL 丙酮、(2) 4 mL甲醇與 4 mL 丙酮、以及 (3) 4 mL 二氯甲烷與 4 mL 丙酮；其中利用 4 mL 甲醇與 4 mL丙酮進行預洗能最有效的降低其背景干擾。為提升樣品中待測物的濃度，將淨化液濃縮至 2 mL (含50% 的水) 並進樣 4 μL可取得良好的對稱峰型，未有管柱過載之問題。 12 種鄰苯二甲酸酯及 6 種雙酚類於地瓜葉、空心菜、雞腿肉與鮭魚的基質效應因子分別為 43.4–139%、9.4–57.9%、55.7–96.7%、以及69.4–109%，相較於富含色素的空心菜，油脂含量高的雞腿肉與鮭魚中之待測物受到的基質效應較少，故 Captiva EMR-Lipid 固相萃取匣可能較適用於淨化動物性樣本，植物性樣本之淨化程序需再重新設計。18 種目標分析物於地瓜葉與鮭魚的萃取效率分別介於 0.1–34.6% 與 0.9–163% 之間，其中高分子量鄰苯二甲酸酯皆低於 20%。由於前處理過程中高分子量鄰苯二甲酸酯與其內標的損失會對定量造成影響，未來需設法提升其萃取效率。

Phthalate esters (PAEs) are widely used as plasticizers for softening rigid plastics, such as polyvinyl chloride (PVC). Bisphenols (BPs) are the monomers used to produce polycarbonate plastics and epoxy resins, which are used in food and beverage contact appliances. PAEs and BPs are endocrine disruptors that can be absorbed from the environment into cultivated plants and leach from food contact materials to foodstuffs, leading to human exposure via ingestion and causing adverse health effects. Our laboratory developed a method for determining 12 PAEs and six BPs in food simultaneously using ultra-performance liquid chromatography coupled with tandem mass spectrometry (UPLC-MS/MS); however, the deficient cleanup efficiency and the backgrounds of PAEs and BPs from sample preparation processes affected the quantification of these compounds. Therefore, this study aimed to optimize the sample preparation steps. One gram of homogenized food samples were extracted twice with 5 mL of acetone at 65℃ (30–34 psi) for 5 minutes using energized dispersive guided extraction (EDGE) system. After the extraction, 4 mL of the extracts were mixed with 1 mL of Milli-Q water, then were cleaned up by a methanol/acetone-prewashed Captiva EMR-Lipid cartridge. The cleanup eluents were concentrated to 2 mL and were filtered by 0.2-μm PTFE syringe filters, and 12 PAEs and six BPs in the extract were analyzed with UPLC-MS/MS. Ascentis Express F5 (30 × 2.1 mm, 2.0 μm) column and UniSpray were used for chromatographic separation and ionization of the analytes, respectively. 12 PAEs and BADGE were detected at positive mode; the column temperature was 35℃, and the mobile phases were composed of 5-mM ammonium acetate/0.1% acetic acid(aq) (pH 4.19) and methanol; the flow rate was 0.5 mL/min, and the total chromatographic time was 8.5 minutes. Five BPs were analyzed at negative mode; the column temperature was 40℃, and the mobile phases were composed of Milli-Q water and methanol; the flow rate was 0.5 mL/min, and the total chromatographic time was 7.35 minutes. The injection volumes of standard solutions and samples were 2 μL and 4 μL, respectively. The linear ranges of calibration curves were 0.1−500 ng/mL with the coefficient of determination (r2) greater than 0.994. The instrumental detection limits (IDLs) were 1.04–2,658 fg, and the instrumental quantitation limits (IQLs) ranged from 3.33–8,861 fg. After passing through the Captiva EMR-Lipid cartridge, acetone/water (80:20, v/v) kept more high molecular-weight (HMW) PAEs (47.1–81.4%) in the cleanup eluents than those with acetonitrile/water (80:20, v/v) (9.0–70.3%). Owing to the high background levels of PAEs in Captiva EMR-Lipid cartridge, prewashing the cartridges with 4 mL of methanol followed by 4 mL of acetone was the most efficient method for reducing the background levels compared with other prewash combinations, which were (1) 4 mL of acetone and (2) 4 mL of dichloromethane followed by 4 mL of acetone. To increase the concentrations of analytes, the eluents were concentrated to 2 mL (with 50% water) and injected 4 μL into UPLC-MS/MS, which provided good peak shapes of all analytes (no fronting peaks). The matrix effect factors of 12 PAEs and six BPs in sweet potato leaves, water spinach, chicken, and salmon were 43.4–139%, 9.4–57.9%, 55.7–96.7%, and 69.4–109%, respectively; PAEs and BPs in lipid-rich chicken and salmon samples were subjected to less ion suppression than those in pigment-rich water spinach, which meant Captiva EMR-Lipid cartridges were suitable for the cleanup of animal foods, and the cleanup procedure for plant foods needs to be improved. The extraction efficiencies of the 18 analytes in sweet potato leaves and salmon ranged between 0.1–34.6% and 0.9–163%, respectively; among these, the extraction efficiencies of HMW PAEs were lower than 20%. Due to the loss of HMW PAEs and their ISTDs during sample preparation processes, the sample pretreatment steps required further optimization to increase the extraction efficiencies of these compounds.