Establishment And Application Of The Rapid Method For Camellia Oil Quality Based On NMR Technology

As an object of study,the quality of camellia oil in our country was estimated systematically in this paper.Nuclear magnetic resonance?NMR?technique combined with powerful chemometric methods including principal component analysis?PCA?,orthogonal projection to latent structures discriminant analysis?OPLS-DA?,and partial least squares?PLS?were employed to evaluate the oxidative stability,oxidation level,storage history,nutritional quality and adulteration.Finally,a rapid analytical model and technical system were established to make an evaluation of camellia oil quality based on NMR technique.The main results are as follows:?1?The physical and chemical properties of camellia oil from Jiangxi Province were analyzed,including acid value,ultraviolet absorbance,induction time,14 kinds of fatty acids,?-tocophenol,polyphenol,squalene and 4 kinds of sterols.The acid value and fatty acid composition of all camellia oil were in the range of Chinese national standard.Oleic acid?C18:1?,linoleic acid?C18:2?and palmitic acid?C16:0?occupied a large proportion of fatty acids,with the content of 79.59%,7.99%and8.54%respectively.Besides triglycerides,there were a few minor components in camellia oil,such as squalen,sterol,?-tocopherol and polyphenol.As the greatest proportion of unsaponifiables section in camellia oil,squalene and sterol could be as a characteristic to distinguish different vegetable oils.The most abundant of squalene was presented in olive oil?4536.31 mg/kg?,and the highest content of?-sitosterol was observed in corn oil?5181.92 mg/kg?,then the highest content of brassicasterol and campesterol were found in rapeseed oil with the value of 4696.86 mg/kg and 704.57mg/kg respectively,and stigmasterol was showed in soybean oil with the highest content?405.98 mg/kg?.As a kind of natural antioxidant,?-tocopherol played an important role in improving the oxidation stability of camellia oil.In addition,according to the relationship between temperature and oil shelf life factor,the shelf life of camellia oil and blend camellia oil were predicted to be 2 months and 4 months respectively,on the basis of induction time,this was consistent with their actual storage time.Because most of the selected oil samples in this experiment were refined oils,their content of polyphenol was lower than the detection limit of LST6119-2007.?2?¹H NMR and?-tocopherol combined with PLS were employed to predicte the oxidation stability of camellia oil.The original data was preprocessed by four different methods:none processing?NO?,autoscaling?UV?,pareto scaling?Par?,and mean-centred?Ctr?.Then PLS model was established,and the optimal number of latent variable was selected based on the lowest root mean square error of cross validation?RMSECV?.Among them,the PLS model presented the best prediction ability by UV preprocessing method with the highest R²,the lowest root mean square error of estimation?RMSEE?and root mean square error of prediction?RMSEP?.Additionally,6 variables were screened out based on the variable importance in the projection?VIP??29?1in PLS model,and the a better prediction ability was presented for PLS model according to the selected 4 variables.?3?The oxidation level of camellia oil was detected by ¹H NMR.The results showed that the signals of oxidative product?10-8 ppm?,unsaturated fatty acid?USFA?signal?5.42-5.29 ppm?and monounsaturated fatty acid?MUSA?signal?2.79-2.70 ppm?could be used as the characteristic signals of camellia oil oxidation.The signals of oxidation products increased gradually with storage time,while the signal of unsaturated fatty acids decreased.Additionally,the variation of fatty acids in the oxidation process were different,with the following oxidation order:polyunsaturated fatty acids?PUFA??29?monounsaturated fatty acid?MUSA??29?saturated fatty acid?SFA?.?4?NMR combined with kinetic equation was employed to predicte the shelf life of camellia oil.1,2-diacylglycerols?1,2-DGs?and 1,3-diacylglycerols?1,3-DGs?could be observed in different vegetable oils based on ³¹P NMR and ¹H NMR.D value?1,2-DGs/total DGs?of virgin camellia oil and extra virgin olive oil decreased regularly with the increase of storage time.³¹P NMR combined with kinetic equation could be used to predict the shelf life of camellia oil accurately,and the correlation coefficient of kinetic equation?R²?reached 0.97782.Similarly,a high predictive ability could be obtained by ¹H NMR combined with kinetic equation with R²reaching 0.96223.?5?¹H NMR combined with PLS were employed to predict the nutritional quality of camellia oil.With the optimal number of latent variables selected based on the lowest RMSECV value,the influence of four different pretreatment methods?NO,UV,Par and Ctr?on PLS model were analyzed.According to the value of R²,RMSEE and RMSEP:NO was chosen as the best pretreatment method of C18:2,TFA,MUFA and PUFA with the lowest RMSEP value;while the model of C16:0 and SFA showed better with UV pretreatment method;C18:0 and C18:3 model presented the lowest RMSEE and RMSEP value by Par pretreatment method;Ctr were more suitable for C18:1 and USFA model with the highest R²,lowest RMSEE and RMSEP value.As for the predictive model of sterol content,the optimum pretreatment method of brassicasterol,campesterol,stigmasterol and?-sitosterol were Par,NO,UV and Par respectively.Besides,the influence of the number of input variables on the model was further analysed by VIP value.The PLS model could be simplified by screening variables,while the prediction ability presented more poor with lower R²,higher RMSEE and RMSEP value,further indicating the important influence of the selected 15 variables for PLS model.However,the PLS model remained to be improved with low R²?0.5945?,when predicting the content of?-tocopherol.?6?¹H NMR and chemometrics were employed to detect the adulteration of camellia oil with 3 different cheap vegetable oils.With the intensity of 15 selected ¹H NMR signals as input variables,PCA showed good group clustering results for pure and nonpure camellia oil,but unsatisfied identification accuracy for the adulterated oil types,indicating relatively small difference among those oils.Whereas these difference could be revealed by OPLS-DA,with identification accuracy higher than90%.PLS was further applied for the prediction of adulteration level in camellia oil.With less than 6 variables screened out by VIP scores as potential key markers,the developed PLS models showed better accuracy.The prediction results for 10 hold-out samples also confirmed that this method was accurate and fast for the detection of camellia oil adulteration.