Comparison of sedimentary iron speciation obtained by sequential extraction and X-ray absorption spectroscopy

Iron (Fe) speciation in marine sediments and suspended particulate matter (SPM) are widely used to investigate the role of Fe minerals in marine biogeochemical cycling and to reconstruct paleo-redox conditions. A traditional method for the determination of Fe speciation are sequential extractions, in which operationally defined Fe pools are dissolved by specific extraction solutions. More recently, many studies make use of synchrotron radiation -based X-ray absorption near edge spectroscopy (XANES), especially if little sample material is available or a high spatial resolution is required. However, few studies have systematically compared Fe speciation obtained by the two methods. Here, we report Fe speciation for marine sediments (both wet and freeze-dried) and SPM from contrasting marine environments (western Baltic Sea and Amazon shelf) obtained by sequential extraction and XANES. The two methods yield comparable results regarding the approximate proportion of reactive Fe ((oxy-hydr)oxide, carbonate and sulfide minerals) in the total Fe pool and the extent to which reactive Fe minerals have been converted to pyrite. However, methodological issues associated with both approaches complicate a direct comparison of specific mineral groups. Pyrite concentrations obtained by XANES are lower than those obtained by extraction, which is attributed to an underestimation by XANES due to self-absorption effects and an over-estimation by the chromium reduction method due to dissolution of reduced sulfur phases that are not associated with Fe. As substantial amounts of pyrite are oxidized during freeze-drying and sample storage, we recommend to analyze wet sediments in modern marine environments. The XANES spectra of Fe (oxyhydr)oxide and Fe(III)-containing (ferric) clay minerals are similar, which is why these components cannot be differentiated by XANES. Similarly, sequential extraction schemes are known to dissolve a variable proportion of ferric clay minerals along with Fe carbonates and Fe (oxyhydr)oxides. Further investigating the role of ferric clay minerals in the low -temperature marine Fe cycle may therefore require the use of additional methods. Our findings may help to assess the transferability of XANES-based Fe speciation for small sediment and SPM samples to those reported in more traditional studies on Fe biogeochemical cycling and paleo-redox conditions.