Laser-induced breakdown spectroscopy (LIBS) has been widely implemented in Mars explorations as a versatile technique for in-situ chemical analysis. Determination of elemental composition from Martian LIBS typically relies on uni/multi-variate models trained on specific laboratory datasets, which in turn requires similarities between this training dataset and Martian data. However, plasma excitation can affect this similarity. The inability to identify insufficiently excited data may ignore poor focusing for some observations or lead to unawareness of compositional uncertainty. This study establishes the LIBS Quality Index (LQI) for Martian LIBS data quality by evaluating the carbon ionic doublet emission around 658 nm for the Martian atmosphere, assessing its signal-to-noise ratio as well as its deviation from the line profile standard of the instrument-specific laboratory dataset. This method allows the LQI to be used as a confidence level for assessing LIBS excitation quality and measurement for similarity with instrument-specific laboratory datasets. The LQI evaluation routines were established for ChemCam, MarSCoDe, and SuperCam. Validation of LQI was carried out on the laboratory datasets and Mars data with known quality assessments of the three instruments. The index was verified to be sensitive to input irradiance, applicable with three current Mars LIBS instruments, robust against various material types (except carbonate), and moderate interference from H, Fe, and Ti. Applying the LQI, the study was able to identify the remote micro-imager (RMI) autofocus as an optimal mechanism, to discover the influence of environmental conditions on data quality, to discuss the impact of target material consolidation and morphology on data quality, and to investigate the contribution of data quality to quantification uncertainty. These application examples demonstrated the feasibility of LQI as a useful and sensitive tool to assess data quality and its potential to aid current and future Martian LIBS explorations.
Read full abstract