基于植物油三维荧光光谱的茶油鉴定模型

doi:10.3964/j.issn.1000-0593(2020)12-3854-06

Abstract
References
Related Citation (15)

Download: PDF (3320 KB)
Export: BibTeX | EndNote (RIS)

Abstract The fatty acid composition of camellia oil is similar to olive oil which is known as the “golden liquid”. The price of camellia oil is high. Some unscrupulous merchants add some cheap oils or waste oil in the high-priced oil to make huge profits, which seriously infringes on the rights and interests of consumers. To establish an efficient method for the identification of camellia oil is of great significance for preventing camellia oil from adulteration and safeguarding consumer rights. In this paper, three-dimensional fluorescence spectra of 54 samples including camellia oils, soybean oils and corn oils from different brands and processing techniques were analyzed. The analytical results show that the three-dimensional fluorescence spectra of camellia oils by different processing techniques are quite different; however, the difference between three-dimensional fluorescence spectra of camellia oils and soybean oils or camellia oils and corn oils is small. So, it is difficult to distinguish camellia oil from soybean oil or corn oil only by observing the three-dimensional fluorescence spectrum of camellia oil. A small program was written, and then the program was run on the 9.4 SAS software platform to transfer the two-dimensional function z=f(excitation, emission) into a one-dimensional function z=f(excitation-emission). The samples data matrix of the training set based on the one-dimensional function of the three-dimensional fluorescence spectra of plant oils was obtained. Orthogonal partial least squares discriminate analysis (OPSL-DA) on Simca 15. 0.2 software platform was used to construct the model for identification of camellia oil based on the training set. The parameter R² representing the fitting ability of the model is 0.84; the parameter Q² representing the predictive ability is 0.72, and the number of variables is 2. So the model is excellent. On this basis, a well-represented testing set is designed. There are 7 different brands of camellia oil, 18 adulterated camellia oils and 2 blind samples at the test set. The adulterants are soybean oil or corn oil or palm oil or cottonseed oil or cheap mixed oils, and the added levels are 4%, 10% and 16% respectively. Using the constructed model for identification of camellia oil, the 27 samples in the test set are predicted. The predicted results are all correct. The model can identify the adulterated camellia oil in the level of 4% adulterant, whose main components are very similar to camellia oil such as corn oil or soybean oil or mixed oil. Therefore, the model can quickly and effectively carry out the identification of camellia oil. The research of this paper provides a new way to apply three-dimensional fluorescence spectroscopy for rapid and effective identification of vegetable oil.

Key words：Camellia oil; Three-dimensional fluorescence spectroscopy; Orthogonal partial least squares discriminate analysis (OPSL-DA); Identification

Received: 2019-08-25 Accepted: 2019-12-13

ZTFLH:

O657.39

Corresponding Authors: HE Wen-xuan E-mail: 706828346@qq.com

Cite this article:

URL:

https://www.gpxygpfx.com/EN/10.3964/j.issn.1000-0593(2020)12-3854-06 OR https://www.gpxygpfx.com/EN/Y2020/V40/I12/3854