Laser-induced breakdown spectroscopy of biological tissues: Plasma diagnostics and a comparison of quantification approaches

Abstract

While calibration-free laser-induced breakdown spectroscopy (CF-LIBS) is a well-known technique, its applications to biosamples, particularly soft tissues, remain limited. Our research focuses on the plasma diagnostics and CF-LIBS analysis of fatty and non-fatty samples (zooplankton, fish, and plants). The selection of zooplankton as an object was motivated by previously reported anomalous accumulation of certain chemical elements. It was observed that temperature and electron number density of laser plasmas induced on these samples, as determined with the use of Mg I-II and Hα emission lines, do not appear to be systematically affected by minor element composition. Plasma conditions across varying samples are most similar at early stages of plasma evolution.Our investigation yielded four sets of analytical outcomes: calibration-based (CB); calibration-free calculations without (CF1) and with (CF2) spatial resolution, as well as data based on the emission from the plasma central zone (CF1a). The same set of analytical lines of B, Ca, K, Li, Mg, Na, P, and Sr was used across all the four cases. To reduce error for CF-LIBS calculations, each element was measured independently from the others using element-to‑carbon (E/C) ratios. Conversely, in non-fatty samples, E/Mg ratios were shown to fit the Local Thermodynamic Equilibrium for all considered transitions. For a more nuanced evaluation of our results, we have proposed a new variant of the distance metric termed the Accuracy Factor (AF). This metric offers a lucid representation of the method accuracy (for instance, AF = 1.5 indicates that the results are accurate within an average factor of 1.5, i.e., they range between 0.67 and 1.5 times the respective certified values).The performance of the CF technique for our biological samples aligns closely with calibration-based LIBS. Using CF-LIBS, we have successfully quantified Li, an element for which constructing calibration curves proved untenable. Our findings underscore that spatially resolved measurements yield more accurate results compared to spatially-integrated measurements, especially for fatty samples.