Study Of The Calibration-Free Laser-induced Break Down Spectroscopy

Laser-Induced-Breakdown Spectroscopy (LIBS) is a recognized laser detection technique for sensing the chemical composition of a wide range of materials including environmental monitoring, industrial processing, Biomedical studies, military and safety needs, space exploration, and art works analysis. However the concentration of major components is difficult to measure by this method because of the so called "matrix effect". Fortunately, Calibration-Free laser-induced break down spectroscopy (CF-LIBS) which can correct the matrix effect was put forward by A.CIUCCI in 1999. It is a promising approach for quantitative analysis without using certified samples and calibration curves. So from then on, many researchers committed their efforts to this field and applied it to several other fields; however, much work is needed in the future study.CF-LIBS has advantages of eliminating the use of calibration curve and avoiding the matrix effect. It is a potential technique for analysis in many fields; however, it must find a suitbale observation winddow and a complex procedure is needed. So in this article, the main goal is to establish a spectrum data processing procedure which can intelligently fulfill the peek-seeking, overlapped spectral-correcting and qualitative analysis. Then study on the LTE condition and its effect on the accuracy of CF-LIBS.In chapter one, the basic concept, characteristic, diagnostic techniques, and the application of the laser-induced plasma are introduced briefly. Broadening of the spectral lines and two important characters of the emission spectral and the method to measure the electron temperature are presented.In the second chapter, a brief introduction of the CF-LIBS is presented. Then the theory of the CF-LIBS is introduced in detail. Finally it lists some factor that will affect the accuracy of CF-LIBS.In chapter three, the experimental setup and method are introduced in detail with the description of the ablation laser and spectrum detection system. An intelligent data preprocessing models was founded in this work. The automatic peak-seeking algorithm was based on the covariance and second derivative. For the resolution of overlapped spectral lines, a model based on the line-fitting and genetic algorithm was established and tested. The qualitative analysis rule was founded for automatic spectral line identification.In chapter four, the model for calculating the electron density and electron temperature of the plasma was founded. The electron density is gotten from the Stark-broadening of line at 422.67 nm, the electron density is gotten from the Saha-boltzmann plot. The temporal history of the plasma is obtained by recording the emission features at predetermined delays and at a fixed gate width (500ns). For each spectrum both electron density and excitation temperature are calculated for each delay time, the LTE condition is studied in shown that the LTE condition is validated when the delay time ranged from 1000ns to 1500ns. Then based on the principle of CF-LIBS, a procedure for elemental analysis is proposed. With help of the procedure, we studied the factors that affect the accuracy of the analysis results.Finally, the results of the whole research work are summarized and the directions for the further studies suggested.