Assessing the atmospheric S burden of continental flood basalts through synchrotron light micro-XRF

Severe forcing of Large Igneous Provinces (LIPs) on the environment through massive outpours of volcanic gases such as SO2 and CO2 (Self et al., 2008) is suggested by their synchrony with major Phanerozoic mass extinctions, but gas contents of the basalts and gas emission rates are poorly constrained. Here we deal with three LIPs that are synchronous with, and may have triggered, major biotic crises, namely Siberian Traps (end-Permian, c. 250 Ma), CAMP (Central Atlantic magmatic province; end-Triassic, c. 201 Ma) and Deccan Traps (end-Cretaceous, c. 66 Ma) and two LIPs whose emplacement only had a minor impact on the biosphere, i.e. Early Jurassic Karoo-Ferrar (c. 181 Ma) and Early Cretaceous Paranà-Etendeka (PE, c. 134 Ma). Classically, the gas load of a continental flood basalt (CFB) is estimated through analyses of melt-inclusions from early crystallized olivines in degassed flows (Self et al., 2008), but these investigations may be hindered by the paucity of fresh olivine phenocrysts, as for CAMP or PE. Hence, here we illustrate an alternative approach in which the pyroxene/melt sulfur partition coefficient (KD) was appraised on experimentally crystallized clinopyroxenes (augites) from sulphide-saturated basalts (at 525 MPa and 1200° or 1175°C in a piston-cylinder apparatus). S (and Cl) contents were then measured by in-situ micro-XRF (Diamond synchrotron, UK) in augite phenocrysts from rocks chemically representative of the above mentioned LIPs. The magmatic S burden was thus calculated through the experimentally determined KD, starting from the analyzed S in clinopyroxenes. Transmission electron microscopy (TEM) analyses discarded any presence of sulphide or fluid inclusions in (CAMP) clinopyroxenes, thus justifying the use of the KD approach, which hinges on the concept of S being distributed at equilibrium between the magma and the crystal lattice. Data are presently available only for CAMP, PE and Deccan, but will shortly be completed with those from the remaining provinces. So far, S contents for Deccan basalts (0.04- 0.14 wt.%) are consistent with those obtained by Self et al. (2008) on melt inclusions. No paramount differences are highlighted between CFB provinces, but, surprisingly, PE high-Ti and CAMP low-Ti samples show the highest and lowest S values, respectively. Rationales for the decoupling between volcanic gas burden and severity of biotic crisis may either be due to PE magmas being emplaced during weaker magmatic pulses at lower eruption rates or S not being the primary cause of environmental perturbations (in favor of CO2 or other gases). Another hypothesis though hinges on the oxidation state of magmas influencing S solubility (Moretti and Baker, 2008). In this sense, strongly oxidized PE high-Ti basaltic magmas would have retained more S, with consequent reduced gas emissions and minor environmental impact.