Level of Confidence in Elemental Analysis by LIBS

• NCJ Number: 252772
• Author(s): Matthieu Baudelet
• Date Published: March 2019
• Length: 29 pages
• Grant Number(s): 2012-DN-BX-K027
• Agencies: NIJ-Sponsored
• Publication Type: Research (Applied/Empirical); Report (Study/Research); Report (Grant Sponsored); Program/Project Description

Annotation

Findings and methodology are presented for a project with the goal of establishing a parameter to quantify the level of confidence that can be given to an elemental analysis by Laser -Inducted Breakdown Spectroscopy (LIBS).

Abstract

The project succeeded in developing a novel approach to the evaluation of spectral interferences in the line assignment in LIBS. By combining fundamental knowledge of the plasma emission, eventual prior knowledge of the sample and the information a full spectrum can provide, this algorithm provides a quantitative measurement of how much a line is interfered for its assignment. The approach was demonstrated on spectra from a pure silicon sample and a more complex NIST SRM610 glass. The algorithm is used in this project for LIBS data; however, it could be applied to other types of elemental emission techniques, such as ICP-OES. The need for only a single spectrum and a database for analysis makes this approach a strong candidate for the analysis of unknown samples when plasma modeling is not possible. The parameter was built considering a LIBS spectrum as a distributed sequence of weighing distributions (the emission peaks with their own profile) over a discrete list of wavelength-ordered emitters. The statistical weight for each emitter line in the list is not equivalent and depends on the emission probability of the transition, the prior knowledge of the sample elemental composition, and the parameters of the LIBS plasma. The quality of the spectra and the amount of spectral lines available for decision is also crucial in the confidence for assignment of lines. The parameter was exposed to different laser conditions, as well as detection and analysis conditions. 8 figures, 2 tables, and 30 references