Qualitative Investigation of Sunflower Oil Peroxide Value Based on LF-NMR and Discriminant Analysis

Yi-miao XIA,Xin WANG,Hai-yan LU,Bao-lin LIU
DOI: https://doi.org/10.13982/j.mfst.1673-9078.2015.6.049
2015-01-01
Abstract:A rapid, qualitative method was developed to determine whether the peroxide value of sunflower oil exceeded the limit using low-field nuclear magnetic resonance(LF-NMR). One hundred and sixty-two samples oxidized to different degrees, including 74 oil samples exceeding the limit, were collected from 8 brands(LH, JLY, FLM, RQ, HS, KW, DL, and RS) and were studied. A direct discrimination method was used to determine the peroxide values using LF-NMR relaxation characteristics(T21, S21, T22, S22, T23, S23, and T2W) as independent variables. The results indicated that consistency could be found between the relaxation characteristics and oxidation time. The overall accuracy of the discrimination model was 85.8%, among which those of RQ and RS sunflower oil were above 90%, while the rest were above 80%, except for that of HS(77.27%)。These results indicate that there was a good discriminatory effect. Therefore, it is feasible to rapidly assess peroxide values using LF-NMR relaxation characteristics of sunflower oil, which is a time and energy-saving method and has important practical implications.
What problem does this paper attempt to address?