Main content area

Isotope Dilution Mass Spectrometry for Highly Precise Determination of Dissolved Inorganic Carbon in Seawater Aiming at Climate Change Studies

Freije-Carrelo, Laura, Alonso Sobrado, Laura, Moldovan, Mariella, Encinar, Jorge Ruiz, García Alonso, J. Ignacio
Analytical chemistry 2018 v.90 no.7 pp. 4677-4685
acidification, adsorption, bicarbonates, carbon dioxide, climate change, dissolved inorganic carbon, isotope dilution technique, isotopes, mass spectrometry, seawater, spectrometers, uncertainty
Dissolved inorganic carbon (DIC) is one of the most important parameters to be measured in seawaters for climate change studies. Its quantitative assessment requires analytical methodologies with overall uncertainties around 0.05% RSD for clear evaluation of temporal trends. Herein, two alternative isotope dilution mass spectrometry (IDMS) methodologies (online and species-specific) using an isotope ratio mass spectrometer (IRMS) and two calculation procedures for each methodology have been compared. As a result, a new method for the determination of DIC in seawaters, based on species-specific IDMS with isotope pattern deconvolution calculation, was developed and validated. A ¹³C-enriched bicarbonate tracer was added to the sample and, after equilibration and acidification, the isotope abundances at CO₂ masses 44, 45, and 46 were measured on an IRMS instrument. Notably, early spiking allows correcting for evaporations and/or adsorptions during sample preparation and storage and could be carried out immediately after sampling. Full uncertainty budgets were calculated taking into account all the factors involved in the determination (initial weights, concentration and isotope abundances of standards, and final IRMS measurements). The average DIC value obtained for CRM seawater agreed very well with the certified value. Propagated precision obtained ranged from 0.035 to 0.050% RSD for individual sample triplicates. Reproducibility, assessed by three independent experiments carried out in different working days, was excellent as well (−0.01% and 0.057%, error and full combined uncertainty, respectively). Additionally, the approach proposed improves on established methods by simplicity, higher throughput (15 min per sample), and lower volume requirements (10 mL).