| Laser-induced breakdown spectroscopy (LIBS) is a newly developed technique forelement analysis. It was based on analyzing the atomic lines or ionic lines in the transientplasma emission produced by the ablation of investigated substance with a high power laserpulse. The excellent advantages of LIBS include simultaneous multi-element detection,in-situ and stand-off analysis capability without any sample pretreatment. In recentlyyears, the potential application of LIBS in the field of ocean resources investigation andocean chemistry environmental monitoring has promoted the research progress ofunderwater LIBS. While in practice, how to realize analysis quantitatively with areasonable accuracy is a big challenge. In this thesis, taking metal elements as theinvestigating target and quantitative analysis as the technical focus, several experimentalinvestigations were carried out, with the special attention paid to the suitable spectrumdetection scheme, comparison of quantitative analysis method and the impact factoraroused in the procedure of LIBS.The thesis begins with an overview of LIBS and its potential significance in oceanapplications. A detailed review on the development of quantitative analysis using LIBS isgiven in Chapter2to show the basis of LIBS, various methods used in quantitative analysisand the impact factors aroused in the procedure of LIBS. The Chapter3is the descriptionof the relevant instrumentation and samples used in the experimental investigations. Thebulk of author’s contribution, within the general group effort, was to perform theexperiments which are described in Chapters4,5and6.Chapter4deals with the suitable detection scheme for quantitative analysis using LIBS.Element Mo in steel alloy samples was taken as the detection target. Using CCD detect,laser-induced Mo plasma emission were spatially resolved. Based on the obtainedspatially resolved spectra, the optimum detection range was determined and applied to thequantitative analysis of Mo element in alloy. The time-resolved laser-induced plasmaemission spectra were taken by a PMT in conjunction with a Boxcar system. Theoptimum detection time delay was determined and used in the quantitative analysis of Mo.It was found that the achieved results based on spatially resolved spectra was better thanthat based on time-resolved spectra, with the relative errors of2.13%for CCD detectionscheme and5.57%for PMT+Boxcar detection scheme. The obtained results suggestedthat CCD detection scheme could be taken as a suitable scheme for quantitative analysisusing LIBS. Based on CCD detection scheme, various methods were applied to the quantitativeanalysis of Mn element in steel alloy samples and detailed performance is given in Chapter5. The analysis was carried out with the methods based on direct intensity calibration,internal calibration, multiple linear regression, non-linear multivariate regression and partialleast squares respectively. It turned out that the direct intensity calibration method hardlyserve the quantitative analysis purpose for multi-components steel alloy samples. Withthe internal calibration of Fe I404.581nm line, the reasonable analysis results wereobtained with the relative error of3.54%and9.70%fro two target samples. Using IMn,IFeas variables, the results achieved by non-linear multivariate regression were found notgood with higher relative errors of9.69%and9.38%. With the introduction of IMn, IFeand CCras variables in non-linear multivariate regression analysis, the quantitative analysisresults were improved with the relative error down to3.29%or even less, withconsideration of dimension reduction in binary quadric terms. Regarding the analysisspeed, the method based on partial least squares(PLS) turned to be the best among all themethods applied, although the analysis accuracy was not stable and expected to beimproved further. Based on PLS, the concentrations of Mn in two target samples weredetermined with the relative errors of6.62%and1.49%respectively.There are many impact factors aroused in the procedure of LIBS, the relevantinvestigation is described in Chapter6with brief discussion. The focal lens to sampledistance(LTSD) is a key factor in the ablation process. It was found that thesignal-to-background ratio of LIBS signal hence the resultant analysis accuracy would bechanged with LTSD varied. It was suggested that LTSD must be kept the same during allthe spectra taken to minimize variation. The self-absorption of laser-induced plasma isanother impact factor in the analysis using LIBS. The obtained results showed that theimpact of self-absorption would be reduced effectively by the means of investigated lineselection, ablation using larger laser pulse energy and detection with longer detection delay.In the case of LIBS under water, the light attenuation should be take into considerations.A relationship between laser pulse energy before getting into water Eiopt(r) and thedetection range reached has been established. The simulation results suggested that toachieve effective LIBS detection,1064nm laser beam was better choice within thedetection range of5cm. On increasing the detection distance beyond10cm,532nm laserbeam may have better performance..Finally, a summery of the work so far the author has done and some suggestions forpossible future development are given in Chapter7. |
PDF Full Text Request