Yun Tang

Accuracy improvement of boron by molecular emission with a genetic algorithm and partial least squares regression model in laser-induced breakdown spectroscopy

Laser-induced breakdown spectroscopy (LIBS) is an atomic emission spectrometry technique for material component analysis. However, the spectral signal distortion and the low analytical accuracy remain as challenges due to the self-absorption effect of atomic lines in LIBS. Here, to overcome this flaw, we demonstrated a method to build calibration with molecular emission, which was measured from a mixture of H3BO3 and C6H12O6·H2O in powder form. We compared the calibration established by typical atomic emission and molecular emission of boron monoxide. The results showed that the self-absorption effect and R2 values were improved by using molecular spectra. Furthermore, to improve the accuracy of molecular emission content determination, a genetic algorithm and partial least squares regression (GA-PLSR) combination model was adopted. The achieved root mean square error of prediction (RMSEP) and the mean prediction error (MPE) for the GA-PLSR model were 0.8667 wt% and 10.9685%, respectively. The results demonstrated that it is a potential method to overcome the self-absorption effect with molecular emission and the accuracy of boron content determination of molecular emission can be improved with the GA-PLSR model.

Evaluation of the self-absorption reduction of minor elements in laser-induced breakdown spectroscopy assisted with laser-stimulated absorption

Reducing self-absorption is one of the most important steps in achieving accurate analyses in laser-induced breakdown spectroscopy (LIBS). In this work, LIBS assisted with laser-stimulated absorption (LSA-LIBS) was used to reduce self-absorption effects in LIBS. A wavelength-tunable laser was used to resonantly excite cold atoms and reduce their amount in laser-induced plasmas, and thus fewer cold atoms at the plasma periphery re-absorb light from the plasma center. Accordingly, the self-absorption effect was reduced. Copper (Cu) and chromium (Cr) elements in steels were taken as examples to evaluate and compare the self-absorption effects in LIBS and LSA-LIBS. The results of calculated self-absorption coefficients (SAs) supported the effectiveness in reducing self-absorption and improving analytical linearity using LSA-LIBS.

Investigation on self-absorption reduction in laser-induced breakdown spectroscopy assisted with spatially selective laser-stimulated absorption

The method of spatially selective laser-stimulated absorption (SS-LSA) was proposed to reduce self-absorption in laser-induced breakdown spectroscopy (LIBS). Through scanning intensity maps of analytes, it was discovered that elemental distribution was inhomogeneous in the formed plasma. Taking copper (Cu) and chromium (Cr) elements in steel as examples, the distribution of Cu and Cr has its own characteristics. By focusing the laser beam on the optimal locations of the plasma with an optical parametric oscillator (OPO) wavelength-tunable laser, the linear determination coefficient (R2 factor) for Cu and Cr elements can reach 0.993 and 0.999, respectively. For Cu determination, compared with LIBS and LSA-LIBS, the root-mean-square error of cross-validation (RMSECV) of SS-LSA-LIBS decreased by 81 and 52%, respectively. The analytical accuracy of the quantitative analysis was much higher than those of conventional LIBS and LSA-LIBS, which indicates that the proposed method can effectively reduce self-absorption and improve LIBS analytical performance.

Accuracy and stability improvement in detecting Wuchang rice adulteration by piece-wise multiplicative scatter correction in the hyperspectral imaging system

The adulteration of rice in the food industry is a very serious problem nowadays. To realize the rapid and stable identification of adulterated Wuchang rice, a hyperspectral imaging system (380–1000 nm) has been introduced in this study. Piece-wise multiplicative scatter correction (PMSC) was first used to correct the non-linear additive and multiplicative scatter effects. Then, the adulterated rice samples were identified via support vector machines (SVM). The PMSC-SVM model was attained over the whole spectral range, with the correct classification rate (CCR) increased from 95.47% to 99.20%, the kappa coefficient increased from 0.95 to 0.99, and the prediction CV (coefficient of variation) decreased from 3.04% to 1.56%. Furthermore, a simplified PMSC-SVM model was established, where 13 principal components were selected using 5-fold cross-validation. The CCR was increased from 95.40% to 99.08%, the kappa coefficient was increased from 0.94 to 0.99, and the prediction CV was decreased from 3.02% to 1.72%. The results demonstrated that the accuracy and stability for identifying adulterated rice has been improved by PMSC in the hyperspectral imaging system.

Accuracy improvement of iron ore analysis using laser-induced breakdown spectroscopy with a hybrid sparse partial least squares and least-squares support vector machine model

The quantitative analysis of iron ore by laser-induced breakdown spectroscopy (LIBS) is usually complicated due to nonlinear self-absorption and matrix effects. To overcome this challenge, a hybrid sparse partial least squares (SPLS) and least-squares support vector machine (LS-SVM) model was proposed to analyze the content of total iron (TFe) and oxides SiO2, Al2O3, CaO, and MgO in iron ore. In this study, 24 samples were used for calibration and 12 for prediction. Sparse partial least squares was used for variable selection and establishing the multilinear regression model between spectral data and concentrations; LS-SVM was used to fit the residual errors of the SPLS regression model to compensate for the nonlinear effects. The model parameters were determined by using the tenfold cross-validation (CV) method. With the hybrid model, the root-mean-square-error of prediction (RMSEP) values of TFe, SiO2, Al2O3, CaO, and MgO were 0.6242, 0.3569, 0.0456, 0.0962, and 0.2157 wt%, respectively. The results showed that the hybrid model yielded better performance than only the conventional SPLS or LS-SVM model. This study demonstrated that the hybrid model is a competitive data processing method for iron ore analysis using LIBS.

Discrimination of nasopharyngeal carcinoma serum using laser-induced breakdown spectroscopy combined with an extreme learning machine and random forest method

The early diagnosis of malignant solid tumours remains a challenge. Here, we propose an efficient way to discriminate between nasopharyngeal carcinoma (NPC) serum and healthy control serum by using laser-induced breakdown spectroscopy (LIBS). Serum was dripped onto a boric acid substrate for LIBS spectrum acquisition. The focus elements (Na, K, Zn, Mg, etc.) were selected for diagnosing NPC using LIBS. With the random forest (RF), characteristic spectral lines were selected based on the variable importance. The spectral lines with variable importance greater than the average were selected. The selected spectral lines are the input of the extreme learning machine (ELM) classifier. Using the RF combined with the ELM classifier, the accuracy rate, sensitivity, and specificity of NPC serum and healthy controls reached 98.330%, 99.0222% and 97.751%, respectively. This demonstrates that LIBS combined with a RF-ELM model can be used to identify NPC with a high rate of accuracy.

Identification accuracy improvement for steel species using a least squares support vector machine and laser-induced breakdown spectroscopy

The classification of discarded alloy steel for recycling purposes and the identification of the alloy steels used in spare parts are very important in modern society. In this work, two typical classification methods, partial least squares discriminant analysis (PLS-DA) and a support vector machine (SVM), were used to study the classification of steels with similar constituents. Forty (40) steel species were detected using laser-induced breakdown spectroscopy (LIBS), and the identification by the PLS-DA and SVM models was 96.25% and 95% accurate, respectively. Based on these two classification algorithms, the least squares support vector machine (LSSVM) algorithm was used to further improve the identification accuracy. The kernel function parameter and error penalty factor of the LSSVM model were 0.4353 and 6.9644, respectively. The results showed that the identification accuracy reached 100%; therefore, combining LIBS with the LSSVM algorithm proved to be an effective approach for accurately identifying steel species with similar constituents.