No abstract

[1] We analyzed d 29 Si of dissolved silicate for eight water column profiles across the Southern Ocean (south of Australia in spring 2001) from the Seasonal Ice Zone (SIZ) north to the Subantarctic Zone (SAZ), including the first isotopic compositions measured for Si-depleted seawaters. All profiles display mixed layer enrichments in heavy Si isotopes relative to deep water in accordance with preferential uptake of the light isotope by diatoms. As silicate levels decrease from the SIZ northward across the Polar Front Zone (PFZ) to the SAZ, surface and mesopelagic d 29 Si signatures generally become progressively heavier, but the most Si-depleted SAZ waters do not exhibit d 29 Si values heavier than in the PFZ. This intricacy appears to derive from variations in the vertical and horizontal supply of silicate to surface waters, and by applying a steady state open system model, we estimate a fractionation factor, 29 e, between diatoms and seawater of À0.45 ± 0.17%, independently of zones and phytoplankton community. Though encouraging, these results are related to latitudinal changes in mesopelagic d 29 Si values, complexity in surface silicateÀd 29 Si correlations, and differences from previous studies, which underline the need for caution in the use of silicon isotopes in paleoceanographic studies until systematic efforts have been undertaken to better understand modern variations.

The cation-exchange purification technique used here does not remove anions (in our case, mostly Cl-, SO 4 2-and to a lesser extent NO 3-) from solutions. In this case, the addition of a known artificial matrix in excess in both the sample and standard solution can be used to dilute the natural concentration of the contaminant and to homogenize sample and standard matrices (doping method, Georg et al., 2006; Hughes et al., 2011). Indeed, dissimilar matrices will affect differently the plasma and ionization efficiency and will induce artificial bias in the delta measurements, invalidating the use of the standard-sample bracketing technique. In our samples, Cl-originating from seawater can be neglected compared to Cladded as HCl (Merck Suprapur) to dissolve the brucite; and as solutions were analyzed in a HCl matrix largely in excess (up to 0.5 mol L-1) compared to natural Cl-concentration. Similarly, the occurrence of NO 3-in seawater was resolved by the use of HNO 3 (Merck Suprapur, 0.5 mol L-1) as a solvent in both the samples and standards. For Depth Depth (m) (m)

No abstract

No abstract