Abstract : Iron is an essential micro-nutrient for phytoplankton growth in the ocean. In broad areas of the ocean, iron limits primary production and therefore plays a role in the carbon cycle. However many questions remain about its marine cycle. Dusts and sediments are considered as the principal sources of iron to the surface open ocean. Because both sources display distinct iron isotopic compositions, iron isotopes in seawater were suggested as a promising new tracer of theses sources. In addition, iron undergoes numerous exchange processes between the various physical and chemical forms coexisting in the water column. Some of these processes have shown isotopic fractionations through in vitro experiments. Iron isotopes in the water column could also help to clarify these processes. At the beginning of my PhD, no dissolved iron isotopes measurements had been performed in the ocean. Because of the very low iron content and the concentrated salty matrix of a seawater sample, such a measurement represented a real analytical challenge. The recovery had to be high enough, with a blank of only a few ng, the matrix had to be efficiently removed and we needed a precise method to correct for the isotopic fractionation occurring during the procedure. We did develop such a method to measure iron isotopes in Fe-depleted seawater satisfying all of these requirements. The successful GEOTRACES intercalibration exercise contributed to validate our method. This method allowed acquiring the first data of dissolved iron isotopes in ocean. We also measured iron isotopes in suspended particles of the seawater, a measurement never performed either so far. The observed δ56Fe variations are significant and range from -0.71 to +0.58‰ with a precision of ±0.08‰ (2σ), the largest variations being in the dissolved phase. Through several oceanic regional studies, first interpretations of the iron isotope cycle in the ocean are highlighted. Below the surface layer of the water column, the Fe isotopic compositions (IC) seem consistent with i) the oceanic circulation, since similar δ56Fe are found in the same water masses sampled at locations separated by more than 4000 km,, and ii) with the biogeochemical properties of the water column. Isotopic fractionation associated with primary production, remineralization and sorption exchanges are suggested. All these processes would display moderated isotopic fractionations. The parallel study of particulate and dissolved Fe IC underlines the efficiency of the dissolved-particulate exchanges. Whereas the mechanism responsible for sedimentary Fe supplies had been widely proposed to be bacterial reductive dissolution so far (characterized by Δdiss-part≈ -3 to 1‰), our results suggest the significance of a different process, the non-reductive dissolution of sediments in seawater, displaying a mean fractionation of Δdiss part≈ +0.2‰, releasing heavy iron in the dissolved phase. New isotopic signatures of iron sources were also identified. δ56Fe of ~0.5‰ in our aerosol samples suggests that the atmospheric iron signature is more variable than expected, from ~0.1‰ to 0.5‰. Finally, the combination of DFe and PFe isotopic data shows strong interactions between these 2 fractions. Close to the continental margin, the seawater would get a new isotopic signature from the lithogenic particles whereas far from the sources, the internal processes induced fractionations are revealed. Even though the behavior of iron isotopes in the ocean is still not well understood, the Fe IC in the dissolved and particulate phases of seawater provide new insight into the iron cycle.