Quantitative Analysis of Carbon with Laser-Induced Breakdown Spectroscopy (LIBS) Using Genetic Algorithm and Back...

Area of Science:

• Analytical Chemistry
• Materials Science
• Spectroscopy

Background:

• Accurate carbon content detection is crucial for metal smelting and classification.
• Laser-induced breakdown spectroscopy (LIBS) is a common method for elemental analysis.
• Interference from iron lines (Fe) with carbon lines (C) complicates LIBS analysis of steel.

Purpose of the Study:

• To develop an optimized method for quantitative carbon analysis in steel using LIBS.
• To address spectral interferences between C(I) 247.86 nm and Fe(II) 247.86 nm lines.
• To improve the accuracy and reliability of carbon detection in steels and iron-based alloys.

Main Methods:

• Utilized laser-induced breakdown spectroscopy (LIBS) for elemental analysis.
• Employed a back propagation (BP) neural network for spectral data modeling.
• Optimized the BP neural network using a genetic algorithm (GA-BP) for enhanced parameter modeling and prediction.

Main Results:

• The GA-BP model achieved a root mean square error of 0.0114 for carbon content analysis.
• Significant improvement in the linear correlation coefficient was observed after applying the GA-BP correction.
• The proposed method demonstrated effectiveness in quantitative analysis of the C(I) 247.86 nm line.

Conclusions:

• The GA-BP method provides an effective solution for resolving spectral interferences in carbon detection via LIBS.
• The developed approach is concise, easy to implement, and suitable for industrial applications in steel and iron-based alloy analysis.
• This technique enhances the precision of carbon content determination in metallurgical processes.