利用近紅外線光譜儀快速偵測落花生黃麴毒素含量

Abstract

落花生(Arachis hypogaea)原生於南美洲，富含油脂與蛋白質，其用途廣泛可供為食用、加工及油用，加上落花生的適應性大而栽培甚為普遍，是世界上重要的食品與油料作物，但落花生及其產品易受黃麴菌感染而產生黃麴毒素，危害人體健康。因此本研究利用近紅外光譜分析技術測定落花生顆粒、落花生碎粒、花生醬以及帶莢落花生等不同類型的落花生及其產品之黃麴毒素含量，在近紅外光譜尋找特定波長以建立測定黃麴毒素檢測其含量的迴歸方程式及線性區別方程式，本研究設定High performance liquid chromatography之分析值為真值。 研究結果顯示，在反射波長806 nm及1508 nm位置所建立的迴歸方程式(A)能夠測定落花生碎粒的黃麴毒素含量，其R2達0.638；在反射波長539 nm及806 nm位置所建立的迴歸方程式(B)能夠測定落花生顆粒的黃麴毒素含量，其R2則為0.581；在反射波長539 nm、806 nm及1508 nm位置所建立的迴歸方程式(C)能夠測定落花生碎粒及落花生顆粒的黃麴毒素含量，其R2高達0.905；在反射波長600 nm至700 nm位置之間所建立的迴歸方程式(D)能夠測定帶莢落花生的黃麴毒素含量，其R2為0.842；在反射波長1712 nm位置所建立的迴歸方程式(E)能夠測定花生醬的黃麴毒素含量，其R2為0.704。在區別分析模式若以15 ppb作為區分是否超標的標準，則結果顯示上述A、B、C及D四個回歸方程式分別應用於落花生碎粒、顆粒、碎粒合併顆粒及帶莢落花生之黃麴毒素含量超標之區分均可達100%之正確率；而在反射波長1712 nm位置所建立的區別分析模式能夠檢測花生醬的黃麴毒素含量，其判別率為88.9%。 本研究推薦花生顆粒與落花生碎粒樣品的檢測可以反射波長539 nm、806 nm及1508 nm所建立的迴歸方程式及區別分析模式進行檢測；而帶莢落花生的黃麴毒素含量的檢測則可以反射波長600 nm至700 nm之間檢測與區分；花生醬則可以反射波長1712 nm所建立的方程式檢測與區分。 本研究另以ELISA方法做為NIR分析之比對，研究結果顯示其在顆粒落花生、碎粒落花生及帶莢落花生部份均有極優良之測定效果，但是在花生醬部份測定效果較差。綜合言之，NIR應用在測定顆粒落花生、碎粒落花生及帶莢花生之效果較佳，而應用在花生醬之黃麴毒素檢測則欠佳。本研究的結果可應用在初步大量快速篩檢海關及市場不同類型落花生及其產品之黃麴毒素含項是否超過標準。

Peanut (Arachis hypogaea) originated in South America. It is full of oils and proteins for food, processing and oil and has a wide range of uses. The wide range of adaptability makes it become an important crop in the world. But peanuts and its products are easy infected by Aspergillus and occur aflatoxins. There in this study, near-infrared spectroscopy (NIR) was used to analyze the content aflatoxins in peanut kernel, peanut crumbs, peanut paste, and peanut pods, and to find out the regression and discrimination model of aflatoxins in different types of peanuts. The results showed that the regression model (A) with reflectance wavelengths at 806 nm and 1508 nm can detect the content of aflatoxins in peanut crumbs, and the R2 is 0.638. The regression model (B) with reflectance wavelengths at 539 nm and 806 nm can detect the content of aflatoxins in peanut kernel, and the R2 is 0.581. The regression model (C) with reflectance wavelengths at 539 nm, 806 nm, and 1508 nm can detect the content of aflatoxins in peanut crumbs and whole peanut kernel, and the R2 is 0. 905. The regression model (D) with reflectance wavelengths at 600 nm to 700 nm can detect the content of aflatoxins in peanut paste, and the R2 is 0.842. The regression model (E) with reflectance wavelengths at 1712 nm can detect the content of aflatoxins in peanut pod, and the R2 is 0. 704. In discriminant analysis, 15 ppb was used as a threshold of discriminative examination to determine if the target exceeded or not. The results showed that models of A, B, C, and D can 100% detect with the peanut kernel, peanut crumbs, and peanut paste samples. The discriminant analysis model can 88.9% detect with reflectance at 1712 nm in peanut pod. This study recommends that the detection of peanut crumbs, and peanut kernel can use the reflectance in 539 nm, 806 nm, and 1508 nm, it is still quite accurate under 15 ppb of aflatoxins. The aflatoxins content of peanut paste can be detected and distinguished between the reflection wavelengths from 600 nm to 700 nm. The peanut pods can be used to detect and distinguish the equation established by reflecting the wavelength of 1712 nm. In ELISA method, there is a good detection for aflatoxins in whole kernels, peanut crumbs, and peanut pods, but the detection was poor in peanut paste. NIR is a useful method of detection aflatoxins for peanut kernel, peanut crumbs, and peanut pod. The application for peanut paste to detect aflatoxins is poor. The results of this study can be used to analyze the contents of aflatoxins in various types of peanut and its products in import border and market examinations. It can screen out the aflatoxins contamination products quickly, to inspect the contents more than the acceptable threshold or not in peanuts.