Study On The Method Of Improving The Stability Of Laser Induced Breakdown Spectroscopy

The existing detection technology has the advantages of high accuracy and good sensitivity,but it needs digestion,is easy to secondary pollution,and is difficult to detect in a fast,large range,in situ and big data.Therefore,a new detection technology is urgently needed.Laser induced breakdown spectroscopy(LIBS) has been widely used in national environmental testing(soil,water quality,air,etc.),biological analysis,explosives analysis,archaeological analysis,food analysis,geological analysis,chemical analysis,etc Metallurgy,analysis and other fields.In recent years,laser-induced breakdown spectroscopy(LIBS) technology has developed rapidly at home and abroad,but the stability of LIBS technology is low,and the quantitative analysis still needs to be improved.In view of the poor stability of LIBS technology,this paper compares and analyzes the plasma image information and spectral intensity through image optimization method,so as to improve the stability of LIBS experiment.At the same time,aiming at the low stability of internal standard method and partial least squares(PLS) in quantitative analysis,the least squares support vector machine(LS-SVM) algorithm is proposed to improve the stability of quantitative analysis of LIBS technology.The specific research work and achievements are as follows:1.Image optimization method.The plasma image information of different concentrations was analyzed.Through the screening of spot area,spot near circle ratio and the color proportion of each circle,the plasma image with spot area of 3000m~2,near circle ratio of 95% and the color proportion of each circle of 8:0.5:1.5 is selected.The results show that the relative standard deviation(RSD)of different concentrations of samples are greatly improved.Without optimization,the RSD of each concentration is 5.39%,6.22%,7.56%,8.42% and 9.63% respectively;under optimization,the RSD of each concentration is 3.24%,4.47%,5.32%,6.13%and 7.21% respectively.The image optimization method effectively improves the stability of the spectrum.Finally,the internal standard method was used for the quantitative analysis of samples.Compared with the non optimization condition,the stability of the calibration model with optimization condition was greatly improved,R2 increased from 0.947 to 0.985.In LIBS detection,image optimization method can greatly improve the stability of LIBS technology.2.Study on the method of improving the stability of quantitative analysis.The background noise of characteristic spectral line is large and the baseline is too high.In this paper,the method of wavelet transform is used to correct the baseline of the spectrum,and the influence of the background noise and the base noise on the spectral information is eliminated.The calibration curve fitting coefficient R2 of 324.75nm Cu I in the sample was increased from 0.945 to 0.978,indicating that the baseline calibration has a great role in improving the stability of the experiment.In order to solve the problem of the stability of the internal standard method,the fitting coefficient obtained by the quantitative analysis of the internal standard method is still not very high,and the values of the root mean square error RMSE and the average relative error are are still very high.A calibration model of PLS is proposed to improve the stability of the experiment.Compared with the internal standard method,the fitting coefficient R2 increased from 0.978 to 0.9862,RMSEP decreased from3.5532wt.% to 0.1214wt%,RMSEP decreased from 1.2807wt.% to 0.2613wt%,are also decreased from 13.2978% to 7.5362%,which greatly improved the stability compared with the internal standard method.Compared with baseline corrected internal standard method and partial least squares(PLS),least squares support vector machine(LSSVM) solves the nonlinear interference in analysis.Through LSSVM algorithm,R2 is improved from 0.9862to 0.997,RMSEC and RMSEP are reduced from 0.1214wt% and 0.2613wt% to 0.0178wt% and 0.151wt%respectively.At the same time,are also reduced from 7.5362% to 2.2187%,which shows that LSSVM algorithm has good stability,can meet the experimental requirements,and also improves the stability of the calibration model.In order to improve the stability of LIBS technology,plasma image and quantitative analysis algorithm are used in this paper.The experimental results show that the image optimization method and the least squares support vector machine algorithm can improve the stability of LIBS technology,and provide effective power for the development of LIBS technology.