[Quantitative Detection of Iron in Soybean Oils with Laser Induced Breakdown Spectroscopy and Simple Regression Methods].

LIBS (laser-induced breakdown spectroscopy) was used to detect Fe element content in soybean oil quantitatively. In this experiment, a series of soybean oil samples with different concentrations of Fe were used; LIBS spectra were collected with a two-channel high precision spectrometer. According to the LIBS spectrum of samples, two characteristic wavelength of Fe (404.58 and 406.36 nm) were determined, and different simple regression methods (exponential regression, linear regression and quadratic regression) were used to establish the quantitative analysis models of Fe content using each characteristic spectral line. The results indicate that the average relative error of Fe I 404.58 and Fe I 406.36 in simple exponential regression, linear regression and quadratic regression models were 29.49%, 8.93%, 8.70% and 28.95%, 8.63%, 8.44%, respectively. The results of Fe I 406.36 regression models is better than that of Fe I 404.58, and the quadratic regression model is optimal among the three regression models. According to these results, LIBS technology has certain feasibility for detecting Fe in soybean oil; the quadratic linear regression model can improve the prediction accuracy of Fe element effectively.

[Model Optimization of Ternary System Adulteration Detection in Camellia Oil Based on Visible/Near Infrared Spectroscopy].

Visible/near infrared spectroscopy combined with chemometrics methods was used to detect ternary system adulteration in camellia oil quantificationally. In order to get adulterated samples, rapeseed oil and peanut oil were added to pure camellia oil in different proportion. Visible/near infrared spectroscopy data of pure and adulterated camellia oil samples were acquired in the wavelength range of 350~1800nm, and samples were randomly divided into calibration set and prediction set. The adulteration models were optimized by comparing different wavelength ranges, pretreatment methods and calibration methods The results show that the optimal modeling wavelength ranges and pretreatment methods for the prediction models of rapeseed oil, peanut oil and total adulteration amount are 750~1 770, 900~1 770, 870~1 770 nm and Multiple scattering correction (MSC), Standard normal variate (SNV) and second order differentia, and the best modeling method is Least square support vector machine (LSSVM). The correlation coefficient (R(P)) in prediction set and the root mean square error predictions(RMSEPs) of optimal adulteration models for rapeseed oil, peanut oil and total adulteration are 0.963, 0.982, 0.993 and 2.1%, 1.5%, 1.8%, respectively. Thus it can be seen that visible /near infrared spectroscopy combined with chemometrics methods can be used for quantitative ternary system adulteration detection in camellia oil.

[Qualitative Detection of Procymidone in Edible Vegetable Oils by Near Infrared Spectroscopy and Variable Selection Methods].

In this research, near infrared (NIR) spectroscopy was used to detect procymidone in edible vegetable oils qualitatively. Edible vegetable oil samples with different procymidone contents were classified to two groups according to boundary line of maximum residue limit of procymidone in national standard. QualitySpec spectrometer was used to acquire spectra of two group samples, then uninformative variable elimination (UVE) and subwindow permutation analysis (SPA) were used to select informative wavelength variables. At last, several methods such as linear discriminant analysis (LDA), partial least squares-linear discriminant analysis (PLS-LDA) and discriminant partial least squares (DPLS) were used to develop classification models. The results indicate that NIR spectroscopy is feasible to classify the two group samples. UVE method can select informative wavelength variables effectively, and improve the performance of classification model. The best model is developed by UVE-DPLS method, the classification correct rate, sensitivity and specificity of prediction samples in this model are 98.7%, 95.0% and 100.0%, respectively.

Detection of Chromium Content in Soybean Oil by Laser Induced Breakdown Spectroscopy and UVE Method.

In order to monitor chromium (Cr) content in soybean oil, laser induced breakdown spectroscopy (LIBS) was used to detect Cr content in this research. Pine wood chips was used to enrich heavy metal of Cr, and the spectra of pine wood chips were acquired in the wavelength range of 206.28~481.77 nm by a two-channel high-precision spectrometer. Then, uninformative variable elimination (UVE) method was used to select sensitive wavelength variables for heavy metal of Cr, and calibration model of Cr in soybean oil was developed with partial least squares (PLS) regression, the performance of the calibration model was compared to univariate and full PLS calibration models. The results indicate that the performance of UVE-PLS calibration model is better than that of univariate and full PLS calibration models, the correlation coefficient, root mean square error of calibration (RMSEC), root mean square error of cross validation (RMSECV), root mean square error of prediction (RMSEP) are 0.990, 0.045 mg·g-1, 0.050 mg·g-1 and 0.054 mg·g-1, respectively. After UVE variable selection, the number of wavelength variables in UVE-PLS calibration model is about 2% of wavelength variables in full PLS calibration model. This means UVE is an effective variable selection method which can select correlative variables for heavy metal of Cr.