Distribution and isotopic signature of Thorium and REE-bearing phases in marine particles and sediments

Particle-reactive tracers like Thorium (Th) and Neodymium (Nd) are widely used to evaluate marine particle fluxes but theirs host phases are controversial. We study the Th and Nd-rich phases in marine sediments from the Mediterranean Sea (DYFAMED site) in order to: (1) constrain the nature of the phases carrying Th isotopes (232Th which is a “lithogenic” tracer but also, potentially, in situ produced 230Thex) and (2) determine the nature of the phases involved in the Boundary Exchange which is a key feature of the Nd cycle in the ocean. Using systematic survey and quantitative analysis by scanning electron microscopy (SEM), we have identified on the Th-REE rich phases of clearly lithogenic origin such as zircon ZrSiO4, xenotime YPO4, monazite (REE, Th)PO4, allanite (REE,Y)2(Al,Fe3)3(SiO4)3(OH)6 but also florencite (REE)Al3(PO4)3(OH)6, a mineral of possible authigenic origin. The two main 232Th-carrying phases are monazite and zircon. The average U/Th ratio in zircon (0.6 g/g) is higher than the crustal average (0,3) but this ratio is very heterogeneous among the zircon crystal population. The U/Th ratio of monazite grains is (~ 0.20 g/g) lower than the crustal average. The inventory of these grains indicates that a large fraction of the 232Th and of the Nd contained in the sediment could reside in monazite (for Th), zircon and florencite (for Th and Nd). Mineral separates micron size-phases were obtained by density and chemical separation for isotopic analyses by Thermal Ionization Mass Spectrometer (TIMS). Their purity was checked by SEM. The low 230Th/232Th ratio (~ 2.7×10-6 mol/mol) of the monazite fraction is significantly below the crustal average (4-5×10-6), in agreement with the low U/Th ratio of these minerals (assuming secular equilibrium between 238U and 230Th). The 230Th/232Th ratio of florencite shows no 230Thxs enrichment compared to secular equilibrium, indicating that it is of lithogenic origin and leads us to the conclusion that 230Th et 232Th may not be carried by the same particles in the ocean (at least in these Mediterranean sediments). It’s consistent with the Nd analysis demonstrating identical isotopic compositions in monazite phase, florencite phase and bulk samples. These results help to understand the large variability of the U/Th ratio in marine particles collected in sediment traps moored in the Mediterranean Sea and to better evaluate the 230Thxs in these samples. Concerning boundary exchange, it suggests that most Nd and Th in marine sediments is locked in refractory minerals and therefore may not participate to this isotopic exchange. Analyses of Th and Nd isotopes in small filtered particles and large trapped particles collected in the Mediterranean Sea following the same protocol are underway in order to improve these first observations. Besides, the study of samples from the Austral Ocean, where the lithogenic inputs are much lower than in the Mediterranean sea, are also in progress to ameliorate our previous results about the 230Thxs bearing phases.