Signal enhancement with double-pulse LIBS on biological samples and better discrimination of tissues through machine learning algorithms

Abstract

Laser-induced breakdown spectroscopy (LIBS) is a widely used technique in the field of spectroscopy, enabling the determination of the chemical composition of a variety of samples using typically single laser pulses. This paper presents the successful integration of femtosecond LIBS using double pulses that combined with machine learning algorithms enhanced the discrimination between two animal tissue types (liver and muscle) through the detection of atomic and molecular emissions. The double-pulse configuration was optimized on liver tissue, and the results demonstrated that at 1100 ps pulse delay, the signal was the highest overall for all identified lines, with a fivefold increase compared to single-pulse configuration at comparable energies. By employing femtosecond double-pulse LIBS, it is possible to achieve enhanced signal quality with a better signal-to-noise ratio. Both algorithms used here (Artificial Neural Network and Random Forest) consequently demonstrate superior performance in tissue type prediction when double pulses are employed, as compared to single pulses. The combination of femtosecond double-pulse LIBS with machine learning algorithms has the potential to be an effective technique for thin biological samples (for example biopsy sections), with minimal ablation.