Lithium isotope ratio analysis of geological samples using atomic absorption spectrometry with improved spectral resolution

• Published in: Spectrochimica acta. Part B: Atomic spectroscopy Vol. 220; p. 107013
• Main Authors: Morcillo, Dalia, Winckelmann, Alexander, Frick, Daniel A., Jacobsen, Lars, Seger, Tino, Florek, Stefan, Richter, Silke, Vogl, Jochen, Recknagel, Sebastian, Panne, Ulrich, Abad, Carlos
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.10.2024
• Subjects: Echelle spectrometer; High-resolution continuum source atomic absorption spectroscopy; Isotope; lithium; Machine learning; High-resolution continuum source atomic absorption spectroscopy; lithium; Echelle spectrometer; Isotope; Machine learning
• ISSN: 0584-8547
• DOI: 10.1016/j.sab.2024.107013

This study introduces an improved spectrometric method with enhanced precision to determine isotope ratios in geological samples without chromatographic separation. Firstly, the improvement is achieved by increasing the spectral resolution of the spectrometer applied in well-known high-resolution continuum source atomic absorption spectrometry (HR-CS-AAS). The resulting resolving power and linear dispersion of the upgraded setup, which is denoted in the following as HR+CS-AAS, is well adapted to the line widths of the Li isotope components we investigated. Secondly, our proposed method combines optical absorption spectrometry with machine learning data analysis using an extreme gradient boosting algorithm (XGBoost). This method was applied to analyze certified geological reference materials with δLSVEC(7Li/6Li) (hereafter δ7Li) values ranging from −0.5 ‰ to 4.5 ‰. With a pixel related optical resolving power of λ/∆λ ≈ 780 000, we obtain precisions in δ7Li measurements from 1.0 ‰ to 2.5 ‰. The method is validated by comparing the results with multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS), confirming its metrological compatibility. This work presents a fast, robust, and reliable method for δ7Li measurement in geological samples. [Display omitted] •HR+CS-AAS enhances spectral resolution for precise lithium isotope ratio determination in geological samples.•Optical absorption spectrometry combined with machine learning provides fast, reliable analysis and precise δ7Li measurements.•Validated vs. MC-ICP-MS, the method is promising in determining lithium deposit origins and distinguishing sample matrices.