ANN-LIBS analysis of mixture plasmas: detection of xenon

We developed an artificial neural network method for characterising crucial physical plasma parameters (i.e., temperature, electron density, and abundance ratios of ionisation states) in a fast and precise manner that mitigates common issues arising in evaluation of laser-induced breakdown spectra. The neural network was trained on a set of laser-induced breakdown spectra of xenon, a particularly physically and geochemically intriguing noble gas. The artificial neural network results were subsequently compared to a standard local thermodynamic equilibrium model. Speciation analysis of Xe was performed in a model atmosphere, mimicking gaseous systems relevant for tracing noble gases in geochemistry. The results demonstrate a comprehensive method for geochemical analyses, particularly a new concept of Xe detection in geochemical systems with an order-of-magnitude speed enhancement and requiring minimal input information. The method can be used for determination of Xe plasma physical parameters in industrial as well as scientific applications.

Automated summary

We developed an artificial neural network method to accurately characterize key physical parameters of plasma and mitigate common issues in evaluating laser-induced breakdown spectra. We trained the neural network on xenon spectra and compared the results with a standard model. This method offers a comprehensive approach for geochemical analysis, particularly in detecting xenon in geochemical systems, with a significant speed improvement and minimal input information required.