Ce(III) and Ce(IV) (re)distribution and fractionation in a laterite profile from Madagascar: Insights from in situ XANES spectroscopy at the Ce LIII-edge

The distribution of trivalent and tetravalent cerium, Ce(III) and Ce(IV) respectively, in a lateritic profile from Madagascar, has been characterized by X-ray-absorption near-edge structure (XANES) spectroscopy at the Ce LIII-edge on the LUCIA beamline (SOLEIL synchrotron, France). XANES spectra were acquired on bulk-rock samples as well as on specific lateritic minerals or polymineral zones (in-situ measurements) of the tonalite bedrock and the three overlying weathered horizons (C-, B- and A-horizons). Geochemically, the bedrock, and the A- and C-horizons show similar rare earth element content (REE = 363–405 mg/kg). They also display the same positive Ce-anomaly (CeCN/Ce∗ = 1.12–1.45), which is therefore likely to be inherited from the bedrock. In the B-horizon, the higher REE content (REE = 2194 mg/kg) and the larger Ce-anomaly (CeCN/Ce∗ = 4.26) are consistent with an accumulation zone caused by the evaporation of groundwater during the dry season. There is a good agreement between the Ce(III)/Cetotal ratio (XCe(III)) deduced from the positive Ce-anomaly (bulk-rock geochemical data) and that derived from XANES spectroscopy on the same bulk-rock samples (BR-XCe(III)-XANES) in the bedrock, and the C- and B-horizons. In the A-horizon, XANES measurements on bulk rock and minerals revealed a higher BR-XCe(III)-XANES (up to 100%) compared to the XCe(III) deduced from geochemical data (XCe(III) = 79%). The preservation of a positive Ce-anomaly in the A-horizon suggests that the Ce mobilization and redistribution during weathering occurred with no significant Ce fractionation from other trivalent REE. Remarkably, the only investigated sample where cerianite is observed belongs to the B-horizon. Within this horizon, Ce oxidation state varies depending on the microstructural position (porosity, cracks, clay-rich groundmass). The highest Ce(IV) concentrations are measured in cerianite (and aluminophosphates) localized in pores at the vicinity of Mn-rich domains (XCe(III)-XANES = 30–51%). Therefore, Ce fractionation from other REE is attributed to a Ce oxidation and precipitation potentially assisted by oxyhydroxide scavenging. In the C-horizon, Ce(III) and Ce(IV) are mainly distributed in REE-minerals of the rhabdophane group found in pores and cracks. The similarity between the Ce(III) proportion of rhabdophane grains (XCe(III)-XANES = 74–89%) with that of the bedrock (BR-XCe(III)-XANES = 79%) suggests no significant fractionation of Ce(III) and Ce(IV) between solution and mineral during the successive stages of primary REE-mineral alteration, transport in solution and secondary precipitation in the incipient stages of weathering. Overall, our novel spectroscopic approach shows that Ce is not necessarily oxidized nor fractionated from other REE during weathering in lateritic conditions. This implies that like Ce(III), Ce(IV) can be mobilized in aqueous fluids during weathering, possibly thanks to complexation with organic molecules, and can precipitate together with Ce(III) in secondary REE-bearing minerals. The corollary is that (paleo)redox reconstructions in soils and/or sediments based on Ce-anomaly in weathered rocks or minerals must be interpreted with caution.

Late Quaternary history of the Vakinankaratra volcanic field (central Madagascar): insights from luminescence dating of phreatomagmatic eruption deposits

The Quaternary Vakinankaratra volcanic field in the central Madagascar highlands consists of scoria cones, lava flows, tuff rings, and maars. These volcanic landforms are the result of processes triggered by intracontinental rifting and overlie Precambrian basement or Neogene volcanic rocks. Infrared-stimulated luminescence (IRSL) dating was applied to 13 samples taken from phreatomagmatic eruption deposits in the Antsirabe–Betafo region with the aim of constraining the chronology of the volcanic activity. Establishing such a chronology is important for evaluating volcanic hazards in this densely populated area. Stratigraphic correlations of eruption deposits and IRSL ages suggest at least five phreatomagmatic eruption events in Late Pleistocene times. In the Lake Andraikiba region, two such eruption layers can be clearly distinguished. The older one yields ages between 109 ± 15 and 90 ± 11 ka and is possibly related to an eruption at the Amboniloha volcanic complex to the north. The younger one gives ages between 58 ± 4 and 47 ± 7 ka and is clearly related to the phreatomagmatic eruption that formed Lake Andraikiba. IRSL ages of a similar eruption deposit directly overlying basement laterite in the vicinity of the Fizinana and Ampasamihaiky volcanic complexes yield coherent ages of 68 ± 7 and 65 ± 8 ka. These ages provide the upper age limit for the subsequently developed Iavoko, Antsifotra, and Fizinana scoria cones and their associated lava flows. Two phreatomagmatic deposits, identified near Lake Tritrivakely, yield the youngest IRSL ages in the region, with respective ages of 32 ± 3 and 19 ± 2 ka. The reported K-feldspar IRSL ages are the first recorded numerical ages of phreatomagmatic eruption deposits in Madagascar, and our results confirm the huge potential of this dating approach for reconstructing the volcanic activity of Late Pleistocene to Holocene volcanic provinces.