Visualising Fe speciation diversity in ocean particulate samples by micro X-ray absorption near-edge spectroscopy
Marcus, Matthew A.
Lam, Phoebe J.
MetadataShow full item record
It is a well known truism that natural materials are inhomogeneous, so analysing them on a point-by-point basis can generate a large volume of data, from which it becomes challenging to extract understanding. In this paper, we show an example in which particles taken from the ocean in two different regions (the Western Subarctic Pacific and the Australian sector of the Southern Ocean, south of Tasmania) are studied by Fe K-edge micro X-ray absorption near-edge spectroscopy (μXANES). The resulting set of data consists of 209 spectra from the Western Subarctic Pacific and 126 from the Southern Ocean. We show the use of principal components analysis with an interactive projection visualisation tool to reduce the complexity of the data to something manageable. The Western Subarctic Pacific particles were grouped into four main populations, each of which was characterised by spectra consistent with mixtures of 1–3 minerals: (1) Fe3+ oxyhydroxides + Fe3+ clays + Fe2+ phyllosilicates, (2) Fe3+ clays, (3) mixed-valence phyllosilicates and (4) magnetite + Fe3+ clays + Fe2+ silicates, listed in order of abundance. The Southern Ocean particles break into three clusters: (1) Fe3+-bearing clays + Fe3+ oxyhydroxides, (2) Fe2+ silicates + Fe3+ oxyhydroxides and (3) Fe3+ oxides + Fe3+-bearing clays + Fe2+ silicates, in abundance order. Although there was some overlap between the two regions, this analysis shows that the particulate Fe mineral assemblage is distinct between the Western Subarctic Pacific and the Southern Ocean, with potential implications for the bioavailability of particulate Fe in these two iron-limited regions. We then discuss possible advances in the methods, including automatic methods for characterising the structure of the data.
This paper is not subject to U.S. copyright. The definitive version was published in Environmental Chemistry 11 (2014): 10-17, doi:10.1071/EN13075.