Validation of a portable flow injection - chemiluminescence (FI-CL) method for the determination of dissolved iron in Atlantic open ocean and shelf waters by comparison with isotope dilution - inductively coupled plasma mass spectrometry (ID-ICPMS)

Abstract

Environmental context. The importance of iron as a limiting micronutrient for primary production in the marine environment and its complex marine biogeochemical cycle necessitate accurate methods for the determination of iron in seawater. Current analytical challenges include the detection of trace concentrations (sub-nanomolar) and the high potential for contamination and matrix interferences. To improve confidence in dissolved iron data, intercomparison exercises of commonly used analytical methods are required that demonstrate their applicability to different water masses. Abstract. A blind intercomparison exercise was carried out to validate a well documented, portable flow injection–chemiluminescence (FI-CL) method for the determination of iron in seawater. This was done by the analysis of a variety of filtered Atlantic Ocean samples using FI-CL and a potential primary method of measurement, isotope dilution–inductively coupled plasma mass spectrometry (ID-ICPMS). To investigate the effect of the seawater matrix at various concentrations of iron, samples were collected at various depths (0–200 m) from different water masses (European Continental Shelf, the South Atlantic Ocean) and filtered through both 0.02- and 0.2-μm pore size filters. The exercise was conducted under controlled conditions using the same bottles transported between laboratories to avoid between-bottle inhomogeneity. The results generally showed good agreement between the two methods for dissolved iron over the concentration range 0.15 to 2.1 nM. However, some samples were not in agreement according to estimated uncertainties and this was attributed to random errors arising from contamination during sample handling and matrix effects (i.e. variable interferences) rather than systematic errors.