Routine measurement of high-precision potassium stable isotope compositions using a continuous-flow Neoma MC-ICPMS/MS†

Natural processes, from cosmochemistry to human homeostasis, can be traced by means of the mass-dependent fractionation of K isotopes because the ³⁹K/⁴¹K ratio is characterized by a wide range of variations (ca. 3‰). The measurement of the ³⁹K/⁴¹K ratio is traditionally achieved by multi-collector inductively coupled plasma mass spectrometers (MC-ICPMS), but is significantly impeded by large isobaric argide interferences on K isotopes. A new generation of MC-ICPMS equipped with a collision/reaction cell allows the quantitative elimination of argide interferences using H₂ as a reaction gas. We report on a set of high-precision K isotopic data obtained with the recently released ThermoScientific Neoma MC-ICPMS/MS equipped with a prefiltering system consisting of a double-Wien filter and a collision/reaction cell. In the low-resolution mode, the mass resolving power is ca. 2200, resulting in a K sensitivity of ca. 1000 V ppm⁻¹ for ³⁹K in dry mode with an Apex Omega desolvator. This large mass revolving power allows the observation of yet undetected interferences on the high-mass shoulders of ³⁹K and ⁴¹K. The interference is ca. 25 mV on ³⁹K and 1 mV on ⁴¹K in the low-resolution mode, similar in K and blank HNO₃ (0.05 M) solutions, increases when N₂ is added in the desolvator and decreases when He is added as a collision gas. The presence of these interferences, which contribute modestly < 0.02% of the K signal, is probably the result of the formation of complex organic compounds in the collision/reaction cell. However, blank subtraction is a critical step to achieve steady and accurate analysis of the ³⁹K/⁴¹K ratio. The overall stability of the analysis of the ³⁹K/⁴¹K ratio is greatly improved by using a continuous-flow microFAST Isotope autosampler. A survey on the potential effects of sample-standard mismatches reveals significant offsets for matrix elements (Ca, Mg and Na), no offset for acid molarity. Regarding the effect of sample-standard concentration mismatch, we show that the amplitude of the offset is session-dependent, such that no general correction could be applied. We use the autosampler adjustable injection flow rate to correct for a concentration mismatch up to ± 30% to recover expected K isotope composition within the ± 0.05‰ uncertainty. In these conditions, short-term external precision and long-term reproducibility are 0.07‰ (2SD, n = 500) and 0.08‰ (2SD, n = 66), respectively. For validation of the overall method, we finally purified K using a single step chemical separation by ion exchange chromatography, and measured the K isotope composition of geological and biological reference materials, for which we found values similar to the literature. Our study shows that the Neoma MC-ICPMS/MS with continuous flow injection is a robust instrumentation that will contribute to expediting high-precision K isotopic measurements for various applications.