Guolong Yang

Detection and quantification of adulteration of sesame oils with vegetable oils using gas chromatography and multivariate data analysis

This study was performed to develop a hierarchical approach for detection and quantification of adulteration of sesame oil with vegetable oils using gas chromatography (GC). At first, a model was constructed to discriminate the difference between authentic sesame oils and adulterated sesame oils using support vector machine (SVM) algorithm. Then, another SVM-based model is developed to identify the type of adulterant in the mixed oil. At last, prediction models for sesame oil were built for each kind of oil using partial least square method. To validate this approach, 746 samples were prepared by mixing authentic sesame oils with five types of vegetable oil. The prediction results show that the detection limit for authentication is as low as 5% in mixing ratio and the root-mean-square errors for prediction range from 1.19% to 4.29%, meaning that this approach is a valuable tool to detect and quantify the adulteration of sesame oil.

Solvent-free enzymatic transesterification of ethyl ferulate and monostearin: Optimized by response surface methodology

In this study, enzymatic transesterification of ethyl ferulate (EF) and monostearin for feruloylated lipids production was investigated. Enzyme screening and the effect of feruloyl acceptors on the transesterification were also studied. Effects of reaction variables (reaction temperatures, enzyme load, and reaction time) on the transesterification were optimized using response surface methodology (RSM). The optimum conditions were as follows: reaction temperature 74°C, reaction time 23h, and enzyme load 20% (w/w, relative to the weight of substrates). Under these conditions, EF conversion was 98.3±1.1%, and the transesterification product was consisted of 19.2±2.1% glyceryl ferulate (FG), 32.9±1.9% diferuloylated glycerols (DFG), 36.6±2.2% feruloylated monoacylglycerols (FMAG), 9.1±2.0% feruloylated diacylglycerols (FDAG), and 0.5% ferulic acid (FA). Analysis of variance (ANOVA) showed that the regression equation was adequate for predicting EF conversion. The activation energies for hydrolysis to form FG+DFG and transesterification to form FMAG+FDAG were calculated as 22.45 and 51.05kJ/mol, respectively, based on Arrhenius law.

Enzymatic epoxidation of biodiesel optimized by response surface methodology

During the enzymatic epoxidation of biodiesel, stearic acid was selected as oxygen carrier. Enzyme screening and the load of stearic acid were investigated. The effects of four main reaction conditions including reaction time, temperature, enzyme load, and mole ratio of H2O2/C=C-bonds on the epoxy oxygen group content (EOC) of epoxidized biodiesel were analyzed. Response surface methodology (RSM) was employed to study and optimize the reaction variables in the enzymatic epoxidation of biodiesel. A second-order model satisfactorily fitted the data (R 2 = 0.9893), with non-significant lack of fit. A higher EOC (6.40 ± 0.20%) of epoxidized biodiesel was obtained at the optimum conditions of 55.4°C, 3.91% enzyme load (relative to the weight of biodiesel), 2.93:1 mole ratio of H2O2/C=C-bonds, and 7.3 h. The relationship between initial reaction rate and temperature was also established, and the activation energy (Ea) of the enzymatic epoxidation is 16.03 KJ/mol.

Thermal Losses of Tertiary Butylhydroquinone (TBHQ) and Its Effect on the Qualities of Palm Oil

The rules and patterns of thermal losses of tertiary butylhydroquinone (TBHQ) in palm oil (PO) and its effect on the qualities of PO were investigated by oven heating method. Volatilization and transformation products of TBHQ in PO were also studied in detail under heating treatment. Results showed that at low temperature (< 135°C), TBHQ had better antioxidative properties, while its antioxidative potency to PO was significantly weakened at high temperature (≥ 135°C). In addition, as heating temperatures increased and heating time prolonged, losses of TBHQ significantly increased in PO. Volatilization was the major pathway for losses of TBHQ in PO under heating treatment. Meanwhile, a small portion of TBHQ was transformed and the major transformation product was 2-tertbutyl-1,4- benzoquinone (TQ). Moreover, TQ and several decomposition products of PO were also observed in the volatilization products of TBHQ.

Identification of Adulteration of Sesame Oils Using Least Squares Support Vector Machine Coupled with Particle Swarm Optimization and Partial Least Squares

Through particle swarm optimization(PSO),least squares support vector machine(LSSVM) and partial least squares(PLS) regression,this study was performed to the development of a new method for detection and quantification of adulteration of sesame oil with vegetable oils using gas chromatographic(GC) technique.Based on principal component analysis(PCA),the GC data of total 857 samples including 117 authentic sesame oils and 740 adulterated sesame oils were firstly analyzed for dimension reduction.Using the PCA-filtered GC data,a hierarchical approach including two steps was established for the detection and the quantification of oil samples.At the first step,a model was constructed to discriminate between authentic sesame oils and adulterated sesame oils using least squares support vector machine(LSSVM) algorithm.Then,another LSSVM-based model was developed to identify the type of adulterant in the mixed oil.At last,the PLS models were built to quantification of the adulterated oils.The prediction results showed that the classification model could achieve correct rate 100.0% and 98.7%,and the root-mean-square errors of PLS model were 3.91%,meaning that this approach is a valuable tool to detect and quantify the adulteration of sesame oil compared with other methods such as BP neural network and support vector machine.

Using Consensus Strategy and Interval Partial Least Square Algorithm in Wavelet Domain for Analysis of Near-infrared Spectroscopy

To improve the stability and precision performance of partial least square regression (PLS) model in near-infrared analysis application, the consensus strategy was applied in the wavelet domain. Taking the advantage of multiscale property of wavelet packet analysis, a new modelling method was developed based on the idea of the interval PLS algorithm and named as WpCo-iPLS algorithm. In WpCo-iPLS model, wavelet packet transform (WPT) algorithm was firstly adopted to split the raw spectra into a series of frequency components in wavelet domain. Then, coupled with the consensus strategy, multiple members of PLS models were established on the interval frequency components. To reduce the dependence on single model, an optimization of the weight parameters of member models was conducted. At last, a consensus model was achieved by effectively combining all the member models. To validate the WpCo-iPLS algorithm, it was applied to measure the six kinds of contents concentration of diesel samples using NIR spectra. The experimental results showed that the prediction ability and robustness of WpCo-iPLS model was stronger than that of conventional consensus algorithms, indicating that it is a promising consensus strategy for modelling using NIR spectra.