基于显微多模态光谱融合技术的食用油中AFB1快速免标记检测机理研究

Mycotoxin contamination, currently, has become one of the major hidden dangers in food safety, therefore, developing a fast and sensitive detection method for mycotoxin is the urgent need. In this study, we propose a novel test method by fusing multimodal spectra for fast and label-free detecting Aflatoxin B1 (AFB1) in peanut oil. First, two spectral techniques of near infrared (NIR) spectroscopy and Raman spectroscopy will be in organically integrated in micro-scale, and a multimodal micro-spectra system will be designed and developed for data acquisition. Next, we will clarify the response law of the AFB1 molecule from complex background in NIR and Raman spectra, identify the corresponding spectral bands, and thus guidance to extract characteristic variables from massive spectral data. Then, we will illuminate the mechanism of spectral enhancement to AFB1 molecular, and accordingly, we will explore SERS substrate controlled synthesis, self-assembly method to realize the spectral enhancement corresponding to AFB1 molecular. Then, we will reveal the mechanism of mutual compensation of the information expressed by AFB1 molecules in the near infrared and Raman spectra, based on this, explore various data mining methods, including the characteristic variables extraction, spectral information fusion and nonlinear model constructing, and finally establish a AFB1 prediction model with a strong fault tolerance and high accuracy. This research provides a theoretical basis for fast and label-free detecting AFB1 in edible oil using constructing a multimodal micro-spectra detection system.

真菌毒素污染已成为食品安全的重大隐患之一，迫切需要研究一种快速、灵敏的检测方法。本研究拟以花生油中黄曲霉毒素B1（AFB1）为检测对象，提出一种显微多模态光谱融合检测新思路。研制一套显微多模态光谱融合检测系统，实现近红外光谱技术和拉曼光谱技术在显微尺度下的有机集成；解析复杂本底下AFB1分子在近红外光谱和拉曼光谱上的响应规律，明确AFB1分子所对应的谱带归属，以便从海量光谱数据中挖掘特征信息；阐明AFB1分子的拉曼光谱增强机制，探索表面增强拉曼光谱（SERS）基底的可控合成、自组装方法，实现AFB1分子光谱的特异性增强；揭示近红外和拉曼光谱在AFB1分子表达上的信息互补机制，在此基础上，探索特征信息提取、光谱特征信息融合和非线性模型构建等数据挖掘方法，建立一个容错性强、精度高的AFB1预测模型。项目研究将为建立一种显微多模态光谱融合检测新方法，实现AFB1的快速免标记检测提供理论基础。

真菌毒素污染已成为食品安全的重大隐患之一，迫切需要研究一种快速、灵敏的检测方法。本研究以花生油中黄曲霉毒素B1（AFB1）为检测对象，提出了一种显微多模态光谱融合检测新思路。研制了一套显微多模态光谱融合检测系统，实现了近红外光谱技术和拉曼光谱技术在显微尺度下的有机集成；解析了复杂本底下AFB1分子在近红外光谱和拉曼光谱上的响应规律，明确了AFB1分子所对应的谱带归属，以从海量光谱数据中挖掘特征信息；阐明了AFB1分子的拉曼光谱增强机制，探索了表面增强拉曼光谱（SERS）基底的可控合成、自组装方法，实现了AFB1分子光谱的特异性增强；揭示了近红外和拉曼光谱在AFB1分子表达上的信息互补机制，在此基础上，探索了特征信息提取、光谱特征信息融合和非线性模型构建等数据挖掘方法，建立了一个容错性强、精度高的AFB1预测模型。项目研究为建立一种显微多模态光谱融合检测新方法，实现了AFB1的快速免标记检测提供理论基础。