Detection of K in soil using time-resolved laser-induced breakdown spectroscopy based on convolutional neural networks

One of the technical bottlenecks of traditional laser-induced breakdown spectroscopy (LIBS) is the difficulty in quantitative detection caused by the matrix effect. To troubleshoot this problem, this paper investigated a combination of time-resolved LIBS and convolutional neural networks (CNNs) to improve K determination in soil. The time-resolved LIBS contained the information of both wavelength and time dimension. The spectra of wavelength dimension showed the characteristic emission lines of elements, and those of time dimension presented the plasma decay trend. The one-dimensional data of LIBS intensity from the emission line at 766.49 nm were extracted and correlated with the K concentration, showing a poor correlation of R²_c=0.0967, which is caused by the matrix effect of heterogeneous soil. For the wavelength dimension, the two-dimensional data of traditional integrated LIBS were extracted and analyzed by an artificial neural network (ANN), showing R²_v=0.6318 and the root mean square error of validation (RMSEV)=0.6234. For the time dimension, the two-dimensional data of time-decay LIBS were extracted and analyzed by ANN, showing R²_v=0.7366 and RMSEV=0.7855. These higher determination coefficients reveal that both the non-K emission lines of wavelength dimension and the spectral decay of time dimension could assist in quantitative detection of K. However, due to limited calibration samples, the two-dimensional models presented over-fitting. The three-dimensional data of time-resolved LIBS were analyzed by CNNs, which extracted and integrated the information of both the wavelength and time dimension, showing the R²_v=0.9968 and RMSEV=0.0785. CNN analysis of time-resolved LIBS is capable of improving the determination of K in soil.     

Accuracy improvement of quantitative analysis of calorific value of coal by combining support vector machine and partial least square methods in laser-induced breakdown spectroscopy

Laser-induced breakdown spectroscopy(LIBS) is a potential technology for online coal property analysis,but successful quantitative measurement of calorific value using LIBS suffers from relatively low accuracy caused by the matrix effect.To solve this problem,the support vector machine(SVM) and the partial least square(PLS) were combined to increase the measurement accuracy of calorific value in this study.The combination model utilized SVM to classify coal samples into two groups according to their volatile matter contents to reduce the matrix effect,and then applied PLS to establish calibration models for each sample group respectively.The proposed model was applied to the measurement of calorific values of 53 coal samples,showing that the proposed model could greatly increase accuracy of the measurement of calorific values.Compared with the traditional PLS method,the coefficient of determination(R2) was improved from 0.93 to 0.97,the root-mean-square error of prediction was reduced from 1.68 MJ kg~(-1) to1.08 MJ kg~(-1),and the average relative error was decreased from 6.7% to 3.93%,showing an overall improvement.     

Quantitative analysis of the content of nitrogen and sulfur in coal based on laserinduced breakdown spectroscopy: effects of variable selection

Coal is a crucial fossil energy in today's society,and the detection of sulfir(S) and nitrogen(N)in coal is essential for the evaluation of coal quality.Therefore,an efficient method is needed to quantitatively analyze N and S content in coal,to achieve the purpose of clean utilization of coal.This study applied laser-induced breakdown spectroscopy(LIBS) to test coal quality,and combined two variable selection algorithms,competitive adaptive reweighted sampling(CARS) and the successive projections algorithm(SPA),to establish the corresponding partial least square(PLS) model.The results of the experiment were as follows.The PLS modeled with the full spectrum of 27,620 variables has poor accuracy,the coefficient of determination of the test set(R~2 P) and root mean square error of the test set(RMSEP) of nitrogen were 0.5172 and 0.2263,respectively,and those of sulfur were0.5784 and 0.5811,respectively.The CARS-PLS screened 37 and 25 variables respectively in the detection of N and S elements,but the prediction ability of the model did not improve significantly.SPA-PLS finally screened 14 and 11 variables respectively through successive projections,and obtained the best prediction effect among the three methods.The R~2 P and RMSEP of nitrogen were0.9873 and 0.0208,respectively,and those of sulfur were 0.9451 and 0.2082,respectively.In general,the predictive results of the two elements increased by about 90% for RMSEP and 60% for R2 P compared with PLS.The results show that LIBS combined with SPA-PLS has good potential for detecting N and S content in coal,and is a very promising technology for industrial application.     

Quantitative analysis of steel and iron by laser-induced breakdown spectroscopy using GA-KELM

According to the multiple researches in the last couple of years, laser-induced breakdown spectroscopy(LIBS) has shown a great potential for rapid analysis in steel industry.Nevertheless, the accuracy and precision may be limited by complex matrix effect and selfabsorption effect of LIBS seriously. A novel multivariate calibration method based on genetic algorithm-kernel extreme learning machine(GA-KELM) is proposed for quantitative analysis of multiple elements(Si, Mn, Cr, Ni, V, Ti, Cu, Mo) in forty-seven certified steel and iron samples.First, the standardized peak intensities of selected spectra lines are used as the input of model.Then, the genetic algorithm is adopted to optimize the model parameters due to its obvious capability in finding the global optimum solution. Based on these two steps above, the kernel method is introduced to create kernel matrix which is used to replace the hidden layer's output matrix. Finally, the least square is applied to calculate the model's output weight. In order to verify the predictive capability of the GA-KELM model, the R-square factor(R~2), Root-meansquare Errors of Calibration(RMSEC), Root-mean-square Errors of Prediction(RMSEP) of GAKELM model are compared with the traditional PLS algorithm, respectively. The results confirm that GA-KELM can reduce the interference from matrix effect and self-absorption effect and is suitable for multi-elements calibration of LIBS.     

Calibration curve and support vector regression methods applied for quantification of cement raw meal using laser-induced breakdown spectroscopy

Laser-induced breakdown spectroscopy(LIBS) is a qualitative and quantitative analytical technique with great potential in the cement industrial analysis. Calibration curve(CC) and support vector regression(SVR) methods coupled with LIBS technology were applied for the quantification of three types of cement raw meal samples to compare their analytical concentration range and the ability to reduce matrix effects, respectively. To reduce the effects of fluctuations of the pulse-to-pulse, the unstable ablation and improve the reproducibility, all of the analysis line intensities were normalized on a per-detector basis. The prediction results of the elements of interest in the three types of samples, Ca, Si, Fe, Al, Mg, Na, K and Ti, were compared with the results of the wet chemical analysis. The average relative error(ARE),relative standard deviation(RSD) and root mean squared error of prediction(RMSEP) were employed to investigate and evaluate the prediction accuracy and stability of the two prediction methods. The maximum average ARE of the CC and SVR methods is 34.62% instead of 6.13%,RSD is 40.89% instead of 7.60% and RMSEP is 1.34% instead of 0.43%. The results show that SVR method can accurately analyze samples within a wider concentration range and reduce the matrix effects, and LIBS coupled with it for a rapid, stable and accurate quantification of different types of cement raw meal samples is promising.     

Signal processing for real-time identification of similar metals by laser-induced breakdown spectroscopy

Laser-induced breakdown spectroscopy(LIBS) is regarded as a promising technique for realtime sorting of scrap metals due to its capability of fast multi-elemental and in-air analysis. This work reports a method for signal processing which ensures high accuracy and high speed during similar metal sorting by LIBS. Similar metals such as aluminum alloys or stainless steel are characterized by nearly the same constituent elements with slight variations in elemental concentration depending on metal type. In the proposed method, the original data matrix is substantially reduced for fast processing by selecting new input variables(spectral lines) using the information for the constituent elements of similar metals. Specifically, principal component analysis(PCA) of full-spectra LIBS data was performed and then, based on the loading plots, the input variables of greater significance were selected in the order of higher weights for each constituent element. The results for the classification test with aluminum alloy, copper alloy,stainless steel and cast steel showed that the classification accuracy of the proposed method was nearly the same as that of full-spectra PCA, but the computation time was reduced by a factor of 20 or more. The results demonstrated that incorporating the information for constituent elements can significantly accelerate classification speed without loss of accuracy.     

Quantitative analysis modeling for the ChemCam spectral data based on laser-induced breakdown spectroscopy using convolutional neural network

Laser-induced breakdown spectroscopy (LIBS) has been applied to many fields for the quantitative analysis of diverse materials. Improving the prediction accuracy of LIBS regression models is still of great significance for the Mars exploration in the near future. In this study, we explored the quantitative analysis of LIBS for the one-dimensional ChemCam (an instrument containing a LIBS spectrometer and a Remote Micro-Imager) spectral data whose spectra are produced by the ChemCam team using LIBS under the Mars-like atmospheric conditions. We constructed a convolutional neural network (CNN) regression model with unified parameters for all oxides, which is efficient and concise. CNN that has the excellent capability of feature extraction can effectively overcome the chemical matrix effects that impede the prediction accuracy of regression models. Firstly, we explored the effects of four activation functions on the performance of the CNN model. The results show that the CNN model with the hyperbolic tangent (tanh) function outperforms the CNN models with the other activation functions (the rectified linear unit function, the linear function and the Sigmoid function). Secondly, we compared the performance among the CNN models using different optimization methods. The CNN model with the stochastic gradient descent optimization and the initial learning rate=0.0005 achieves satisfactory performance compared to the other CNN models. Finally, we compared the performance of the CNN model, the model based on support vector regression (SVR) and the model based on partial least square regression (PLSR). The results exhibit the CNN model is superior to the SVR model and the PLSR model for all oxides. Based on the above analysis, we conclude the CNN regression model can effectively improve the prediction accuracy of LIBS.     

An efficient procedure in quantitative analysis using laser-induced breakdown spectroscopy

Laser-induced breakdown spectroscopy has become a general-purpose technique, and internal standard calibration is a common method for quantitative analysis. Calibration models should be reconstructed for different systems and application environments. This study presents an efficient procedure in the construction and selection of calibration models for LIBS analysis. The procedure concludes data preprocess, calibration model construction, and concentration calculation. These steps can be programmed without manual intervention. Results of the quantitative analysis of Ni-based alloys using the proposed procedure are presented in this study.Ten elements are calibrated, and most have an average relative standard error of less than 10%.The proposed procedure is an effective process for constructing and selecting calibration models.     

Geographical authenticity evaluation of Mentha haplocalyx by LIBS coupled with multivariate analyses

Mentha haplocalyx (mint) is a significant traditional Chinese medicine (TCM) listed in the Catalogue of ‘Medicinal and Food Homology’, therefore, its geographical origins (GOs) are critical to the medicinal and food value. Laser-induced breakdown spectroscopy (LIBS) is an advanced analytical technique for GOs certification, due to the fast multi-elemental analysis requiring minimal sample pretreatment. In this study, LIBS data of sampled mint from five GOs were investigated by LIBS coupled with multivariate statistical analyzes. The spectral data was analyzed by two chemometric algorithms, i.e. principal component analysis (PCA) and least squares support vector machines (LS-SVM). Specifically, the performance of LS-SVM with least kernel and radial basis function (RBF) kernel was explored in sensitivity and robustness tests. Both LS-SVM algorithms exhibited excellent performance of classification in sensitive test and good performance (a little inferior) in robustness test. Generally, LS-SVM with linear kernel equally outperformed LS-SVM based on RBF kernel. The result indicated the potential for future applications in herbs and food, especially for in situ GOs applications of TCM authenticity rapidly.     

A feature selection method combined with ridge regression and recursive feature elimination in quantitative analysis of laser induced breakdown spectroscopy

In the spectral analysis of laser-induced breakdown spectroscopy,abundant characteristic spectral lines and severe interference information exist simultaneously in the original spectral data.Here,a feature selection method called recursive feature elimination based on ridge regression(Ridge-RFE) for the original spectral data is recommended to make full use of the valid information of spectra.In the Ridge-RFE method,the absolute value of the ridge regression coefficient was used as a criterion to screen spectral characteristic,the feature with the absolute value of minimum weight in the input subset features was removed by recursive feature elimination(RFE),and the selected features were used as inputs of the partial least squares regression(PLS) model.The Ridge-RFE method based PLS model was used to measure the Fe,Si,Mg,Cu,Zn and Mn for 51 aluminum alloy samples,and the results showed that the root mean square error of prediction decreased greatly compared to the PLS model with full spectrum as input.The overall results demonstrate that the Ridge-RFE method is more efficient to extract the redundant features,make PLS model for better quantitative analysis results and improve model generalization ability.     

Nondestructive Determination of Cu Residue in Orange Peel by Laser Induced Breakdown Spectroscopy

Laser induced breakdown spectroscopy(LIBS) is an emerging tool with rapid,nondestructive,green characteristics in qualitative or quantitative analyses of composition in materials.But LIBS has its shortcomings in detect limit and sensitivity.In this work,heavy metal Cu in Gannan Navel Orange,which is one of famous fruits from Jiangxi of China,was analyzed.In view of LIBS's limit,it is difficult to determinate heavy metals in natural fruits.In this work,nine orange samples were pretreated in 50-500 μg/mL Cu solution,respectively.Another one orange sample was chosen as a control group without any pollution treatment.Previous researchers observed that the content of heavy metals is much higher in peel than in pulp.So,the content in pulp can be reflected by detecting peel.The real concentrations of Cu in peels were acquired by atomic absorption spectrophotometer(AAS).A calibration model of Cu I 324.7 and Cu Ⅰ 327.4was constructed between LIBS intensity and AAS concentration by six samples.The correlation coefficient of the two models is also 0.95.All of the samples were used to verify the accuracy of the model.The results show that the relative error(RE) between predicted and real concentration is less than 6.5%,and Cu Ⅰ 324.7 line has smaller RE than Cu Ⅰ 327.4.The analysis demonstrated that different characteristic lines decided different accuracy.The results prove the feasibility of detecting heavy metals in fruits by LIBS.But the results are limited in treated samples.The next work will focus on direct analysis of heavy metals in natural fruits without any pretreatment.This work is helpful to explore the distribution of heavy metals between pulp and peel.     

Rapid Detection of Oil Pollution in Soil by Using Laser-Induced Breakdown Spectroscopy

Detection of oil pollution in soil has been carried out using laser-induced breakdown spectroscopy(LIBS). A pulsed neodymium-doped yttrium aluminum garnet(Nd:YAG) laser(1,064 nm, 8 ns, 200 mJ) was focused onto pelletized soil samples. Emission spectra were obtained from oil-contaminated soil and clean soil. The contaminated soil had almost the same spectrum profile as the clean soil and contained the same major and minor elements. However, a C–H molecular band was clearly detected in the oil-contaminated soil, while no C–H band was detected in the clean soil. Linear calibration curve of the C–H molecular band was successfully made by using a soil sample containing various concentrations of oil. The limit of detection of the C–H band in the soil sample was 0.001 mL/g. Furthermore, the emission spectrum of the contaminated soil clearly displayed titanium(Ti) lines, which were not detected in the clean soil. The existence of the C–H band and Ti lines in oil-contaminated soil can be used to clearly distinguish contaminated soil from clean soil. For comparison, the emission spectra of contaminated and clean soil were also obtained using scanning electron microscope-energy dispersive X-ray(SEM/EDX) spectroscopy,showing that the spectra obtained using LIBS are much better than using SEM/EDX, as indicated by the signal to noise ratio(S/N ratio).     

Portable fiber-optic laser-induced breakdown spectroscopy system for the quantitative analysis of minor elements in steel

In this paper, we developed a portable laser-induced breakdown spectroscopy(LIBS) using an optical fiber to deliver laser energy and used it to quantitatively analyze minor elements in steel.The R~2 factors of calibration curves of elements Mn, Ti, V, and Cr in pig iron were 0.9965,0.9983, 0.9963, and 0.991, respectively, and their root mean square errors of cross-validation were 0.0501, 0.0054, 0.0205, and 0.0245 wt%, respectively. Six test samples were used for the validation of the performance of the calibration curves established by the portable LIBS. The average relative errors of elements Mn, Ti, V, and Cr were 2.5%, 11.7%, 13.0%, and 5.6%,respectively. These results were comparable with most results reported in traditional LIBS in steel or other matrices. However, the portable LIBS is flexible, compact, and robust, providing a promising prospect in industrial application.     

Ultrasonic Nebulizer Assisted LIBS: a Promising Metal Elements Detection Method for Aqueous Sample Analysis

A newly developed approach for trace metal elements detection for aqueous samples analysis is presented in this paper. The idea of this approach is to improve ablation efficiency by transforming the liquid sample into a dense cloud of droplets using an ultrasonic nebulizer. The resulting droplets are then subjected to analysis by laser induced breakdown spectroscopy (LIBS). A purpose-built ultrasonic nebulizer assisted LIBS (UN-LIBS) system has been applied to the analysis of aqueous samples at trace levels of concentration. Experimental investigations of solution samples were carried out with various dissolved trace metal elements (Mn, Zn, Cu, Pb, Fe, Mg and Na) using this approach. The characteristics of UN-LIBS signal of the elements were investigated regarding the lifetime and S/B ratio and the calibration curves for trace metal elements analyses. The obtained LODs are comparable or much better than the LODS of the reported signal enhancement approaches when the laser pulse energy was as low as 30 mJ. The good linearity of calibration curves and the low LODs shows the potential ability of this method for metal elements analysis application. The density of the electrons was calculated by measuring the Stark width of the line of Hα. The possible mechanism of the LIBS signal enhancement of this approach was briefly discussed.     

Identification of heavy metal-contaminated Tegillarca granosa using laser-induced breakdown spectroscopy and linear regression for classification

Tegillarca granosa (T. granosa) is susceptible to heavy metals, which may pose a threat to consumer health. Thus, healthy and polluted T. granosa should be distinguished quickly. This study aimed to rapidly identify heavy metal pollution by using laser-induced breakdown spectroscopy (LIBS) coupled with linear regression classification (LRC). Five types of T. granosa were studied, namely, Cd-, Zn-, Pb-contaminated, mixed contaminated, and control samples. Threshold method was applied to extract the significant variables from LIBS spectra. Then, LRC was used to classify the different types of T. granosa. Other classification models and feature selection methods were used for comparison. LRC was the best model, achieving an accuracy of 90.67%. Results indicated that LIBS combined with LRC is effective and feasible for T. granosa heavy metal detection.     

用于粒子加速器位置监测的线位置探测器静态特性的标定

本文介绍了一种线位置探测器(WPS)静态特性标定的方法。建立了一套由激光位移传感器、水平与垂直调节平台组成的静态特性标定系统。在垂直和水平两个方向对WPS的测量静态特性进行了研究。在WPS的测量范围内,以步长0.1 mm进行位移量输入,满量程观察了WPS的输出位移量。对输入输出数据进行了拟合,得到WPS灵敏度以及均方根误差。输入步长分别设为5、2和1 μm,考查输出量变化可得到仪器的分辨率。实验结果表明：垂直方向灵敏度K_V不低于-1.005 02,均方根误差为0.002 4 mm,分辨率达2 μm。水平方向灵敏度K_H不低于-1.001 55,均方根误差为0.002 1 mm,分辨率同样达2 μm。该WPS灵敏度、均方根误差和分辨率均达到了监测所需的精度。     

气体采样滤膜中铀含量的激光诱导击穿光谱分析方法研究

为对核设施尾气监测中气体采样滤膜中铀元素的含量进行测量分析,本文开展了激光诱导击穿光谱（LIBS）定量分析气体采样滤膜中铀含量的方法研究。通过优化LIBS测量延迟时间参数,最终选出了铀的定量分析有效特征谱线,并得到了浓度范围较低的工作曲线。结果表明,在选择的实验条件下,采用标准样品对筛选后的特征谱线进行实验验证,测量值的相对偏差绝对值小于20%,RSD小于10%。     

Quantitative Analysis of Mg in Pipeline Dirt Based on Laser-Induced Breakdown Spectroscopy

In order to maintain the pipeline better and remove the dirt more effectively,it was necessary to analyze the contents of elements in dirt.Mg in soil outside of the pipe and the dirt inside of the pipe was quantitatively analyzed and compared by using the laser-induced breakdown spectroscopy(LIBS).Firstly,Mg was quantitatively analyzed on the basis of Mg Ⅰ 285.213 nm by calibration curve for integrated intensity and peak intensity of the spectrum before and after subtracting noise,respectively.Then calibration curves on the basis of Mg Ⅱ 279.553 nm and MgⅡ 280.270 nm were analyzed.The results indicated that it is better to use integrated intensity after subtracting noise of the spectrum line with high relative intensity to make the calibration curve.     

A rapid classification method of aluminum alloy based on laser-induced breakdown spectroscopy and random forest algorithm

As an important non-ferrous metal structural material most used in industry and production,aluminum(Al) alloy shows its great value in the national economy and industrial manufacturing.How to classify Al alloy rapidly and accurately is a significant, popular and meaningful task.Classification methods based on laser-induced breakdown spectroscopy(LIBS) have been reported in recent years. Although LIBS is an advanced detection technology, it is necessary to combine it with some algorithm to reach the goal of rapid and accurate classification. As an important machine learning method, the random forest(RF) algorithm plays a great role in pattern recognition and material classification. This paper introduces a rapid classification method of Al alloy based on LIBS and the RF algorithm. The results show that the best accuracy that can be reached using this method to classify Al alloy samples is 98.59%, the average of which is 98.45%. It also reveals through the relationship laws that the accuracy varies with the number of trees in the RF and the size of the training sample set in the RF. According to the laws, researchers can find out the optimized parameters in the RF algorithm in order to achieve,as expected, a good result. These results prove that LIBS with the RF algorithm can exactly classify Al alloy effectively, precisely and rapidly with high accuracy, which obviously has significant practical value.     

Parameters Optimization of Laser-Induced Breakdown Spectroscopy Experimental Setup for the Case with Beam Expander

Improvement of measurement precision and repeatability is one of the issues cur?rently faced by the laser-induced breakdown spectroscopy (LIBS) technique, which is expected to be capable of precise and accurate quantitative analysis. It was found that there was great poten?tial to improve the signal quality and repeatability by reducing the laser beam divergence angle using a suitable beam expander (BE). In the present work, the influences of several experimental parameters for the case with BE are studied in order to optimize the analytical performances: the signal to noise ratio (SNR) and the relative standard deviation (RSD). We demonstrate that by selecting the optimal experimental parameters, the BE-included LIBS setup can give higher SNR and lower RSD values of the line intensity normalized by the whole spectrum area. For validation purposes, support vector machine (SVM) regression combined with principal component analysis (PCA) was used to establish a calibration model to realize the quantitative analysis of the ash content. Good agreement has been found between the laboratory measurement results from the LIBS method and those from the traditional method. The measurement accuracy presented here for ash content analysis is estimated to be 0.31%, while the average relative error is 2.36%.