亞洲近年常發生不當來源油品混充正常油品的食品安全事件，因異常油品的情況多元且複雜，以致針對 特定檢測標的無法有效判別，故本研究以無特定標的分析 (non-target analysis) 概念建構可評估油品異常情況的 高解析質譜儀分析及資訊判讀平台。以6件市售豬油及13件有疑慮之豬油油品為範例，實驗結果顯示異丙醇 萃取層之質譜輪廓有顯著差異，針對差異性最大的疑慮豬油樣品探索主要的差異性訊號，其中荷質比 (mass to change ratio, m/z) 656.5821、684.6137、712.6445與740.6753經資料庫比對判斷為存在於植物中的三酸甘油酯，主 要由月桂酸 (12：0)、肉荳蔻酸 (14：0) 及棕櫚酸 (16：0) 脂肪酸所組成，透過標準品驗證確認656.5821為三月桂 酸甘油酯 (trilaurin)，顯示這些樣品的異常原因為受到植物油的污染，並由植物固醇的篩檢結果證實此推測； 此外亦發現到上述三酸甘油酯的二聚體化合物訊號，推測樣品亦有氧化的狀況。綜合以上結果，無特定標的 平台為判斷油品異常情況的有效工具，做為選擇後續確認試驗重點項目的依據，減少不必要的實驗項目，且 相較於目前檢測方法多針對單一種可能情境，本平台可藉由異常訊號的資訊取得多元的訊息，達到提升評估 效率的目的。

In recent years, the adulteration of edible oil with tainted caused increasing food safety issues in Asia. However, current analytical methods cannot effectively distinguish adulterated oils due to its complexity and diversity. In this study, high resolution mass spectrometer with non-target analysis platform was used to evaluate lard samples. Six commercially available lards and thirteen suspected tainted lards were used as experimented samples. Results showed that the mass spectrum profile of isopropanol extracted layer from suspected lard samples were significantly different from the others. The probable chemical compound of the mass to change ratio (m/z) 656.5821, 684.6137, 712.6445 and 740.6753 was triacylglycerol composed of lauric (12 : 0), myristic (14 : 0) and palmitic (16 : 0) acids. Furthermore, the 656.5821 m/z was confirmed to be a trilaurin, which is distributed mainly in vegetable oil. The results obtained were also verified by the phytosterol analysis and confirmed that the suspected lard samples were contaminated with vegetable oils. In addition, triacylglycerol dimers were also found and suggested that the lard samples were oxidized. Based on the above results, this study had developed an effective tool of determining tainted oils. From this method, characteristic signals information were obtained to determine the presence of tainted oil. Besides, this analytical platform may reduce unnecessary experiments and is more efficient compared to the current single target analysis.