The role of plants in ironstone evolution: iron and aluminium cycling in the rhizosphere.

The Carajás plateaus in Brazil host endemic epilithic vegetation ("campo rupestre") on top of ironstone duricrusts, known as canga. This capping rock is primarily composed of iron(III) oxide minerals and forms a physically resistant horizon. Field observations reveal an intimate interaction between canga's surface and two native sedges (Rhynchospora barbata and Bulbostylis cangae). These observations suggest that certain plants contribute to the biogeochemical cycling of iron. Iron dissolution features at the root-rock interface were characterised using synchrotron-based techniques, Raman spectroscopy and scanning electron microscopy. These microscale characterisations indicate that iron is preferentially leached in the rhizosphere, enriching the comparatively insoluble aluminium around root channels. Oxalic acid and other exudates were detected in active root channels, signifying ligand-controlled iron oxide dissolution, likely driven by the plants' requirements for goethite-associated nutrients such as phosphorus. The excess iron not uptaken by the plant can reprecipitate in and around roots, line root channels and cement detrital fragments in the soil crust at the base of the plants. The reprecipitation of iron is significant as it provides a continuously forming cement, which makes canga horizons a 'self-healing' cover and contributes to them being the world's most stable continuously exposed land surfaces. Aluminium hydroxide precipitates ("gibbsite cutans") were also detected, coating some of the root cavities, often in alternating layers with goethite. This alternating pattern may correspond with oscillating oxygen concentrations in the rhizosphere. Microbial lineages known to contain iron-reducing bacteria were identified in the sedge rhizospheric microbiome and likely contribute to the reductive dissolution of iron(III) oxides within canga. Drying or percolation of oxygenated water to these anaerobic niches have led to iron mineralisation of biofilms, detected in many root channels. This study sheds light on plants' direct and indirect involvement in canga evolution, with possible implications for revegetation and surface restoration of iron mine sites.