Direct analysis of lanthanides by ICPMS in calcium-rich water samples using a modular high-efficiency sample introduction system–membrane desolvator
Aliaga-Campuzano, MP and Bernal, JP and Briceno-Prieto, SB and Perez-Arvizu, O and Lounejeva, E, Direct analysis of lanthanides by ICPMS in calcium-rich water samples using a modular high-efficiency sample introduction system-membrane desolvator, Journal of Analytical Atomic Spectrometry, 28, (7) pp. 1102-1109. ISSN 0267-9477 (2013) [Refereed Article]
Lanthanoids (also known as rare-earth elements – REE) are well-known tracers of water–rock interaction processes due to their behavior being coherent with their atomic radii, and the capability of Ce and Eu to reflect redox conditions. However, analysis of REE in natural waters is hampered by their relatively low concentration (sub-pg g−1), and the concentration of dissolved solids ranging in the hundreds of μg g−1, imposing important matrix effects to the analysis. Because of this, different analytical techniques have been developed to analyze them by Q-ICPMS and SF-ICPMS, which usually involve matrix removal and analyte pre-concentration. Analysis of REE in karstic waters, those naturally saturated in CaCO3, presents an additional challenge due to the presence of high Ba concentrations (several ng g−1) which can bias the analysis due to the formation of BaO+ and BaOH+ ions during ionization, interfering with the mass range of 148–155 m/z+, including both Eu isotopes, and significantly hampering the direct analysis of such samples. Here, using a modular high-efficiency sample introduction system and desolvator we eliminate the formation of BaO+, and significantly reduce BaOH+, allowing us to analyze small samples (1–2 ml) with Ba/Eu ratios as high as 8 × 104 (mol/mol), and up to 600 μg g−1 of Ca, without any sample pretreatment. Our methodology was validated by analyzing VIDAC18 and SERMIN1 reference materials, and permits the quantification of all REE with detection limits ranging between 500 and 30 fg g−1 and with RSD ~ 10% for 1 pg g−1, controlled by counting statistics.