Using Metrology in Chemistry for Validated Ultra-Trace Measurements at less than 10% Uncertainty: The Case of Fe in Seawater

Environmental studies require more and more the capacity to measure accurately the content of elements and chemical compounds in complex matrix samples at nanomolar levels or below. This is extremely challenging for various reasons including contamination issues, biases introduced by highly concentrated sample matrix components, lack of calibration strategies and certified reference material suitable for these ultra-low content levels etc. Also, convincing validation schemes for the measurement methods developed are essential but hard to establish. The determination of the iron content in seawater is such a case. Iron generally exists at sub-nanomolar concentrations in the open ocean, and may easily be a limiting micronutrient for primary production in the marine environment. Its distribution and transport has a profound effect on plankton communities and consequently, the global carbon cycle. This last point has been an important source of motivation for studying the complex marine biogeochemical cycle of iron. It is now well established that increasing amounts of CO2 released in the atmosphere are the cause for global climate changes. Thus, the possibility that fertilisation with iron of well chosen oceanic areas could lead to the definitive sequestration of significant amounts of atmospheric CO2 by the ocean has been investigated during the last two decades. Understanding the impact and mechanisms of greenhouse gas emissions on climate and carbon sinks is a research priority to the European Union (Council Decision 2002/834/EC). Despite important technological progresses in twenty years quantifying dissolved Fe in seawater remains a very difficult analytical challenge. Major obstacles, reported by analytical chemists, are differences in efficiency of the extraction of iron from the seawater matrix during pre-concentration methods, inaccuracies in system calibration, errors in the quantification of the analytical blank, and the incorrect estimation of the uncertainty of measurement results. This PhD work describes the application of isotope ratio and isotope dilution mass spectrometry (IDMS) measurements on elements at ultra trace levels in complex matrix samples, and particularly on iron in seawater down to 8 ng kg-1 Fe with combined uncertainty of 6 - 24 % (k=2).

PETROV Ivan; 

2010-06-16

JRC58867