Contamination of basalt through silicic melts: The first chaotic dynamics experiments with Parana-Etendeka starting materials.

This work is the first attempt to simulate experimentally the impact of underplating high-Ti basaltic melt from Paraná-Etendeka Magmatic Province (PEMP) on a pre-existing continental crust. Our final aim is to unravel the origin of high volumes of high-Ti Chapecó dacitic melts. A campaign of three new chaotic mixing experiments was performed starting from a natural basaltic melt composition (i.e., high-Ti Pitanga-type) while solely varying the chemical composition of rhyolitic end-member (i.e., the contaminant). The choice of end-members was based on the major oxide composition and isotopic signatures. After experiments, chemical analyses were performed by electron microprobe microscopy along contact zones between the interconnected end-members on the experimental charges produced.

The results confirm and quantify elemental homogenization in the melt, which occurs by diffusion. The degree of homogenization varies according to the element, and the differential mobility in the mixing system culminates in non-linear correlation in inter-elemental plots. The mobility of major/minor elements along contact areas was quantified calculating the normalized variance, which consistently indicates the connection between viscosity and mobility. Orbicular structures containing dendritic crystals and remnant portions of glass in the basaltic area were described in detailed for the first time. The observed dendritic areas of the basaltic regions point towards an early crystallization process during the initial quench phase of the experiment, and that the crystallization process does not happen homogeneously.

Chaotic mixing experiments confirmed that mixing between melts with vastly different viscosities is physically possible. Chaotic dynamics governs this process initially producing vortex structures, and stretched and folded filaments within regions that appear to correspond to remnants of the unmixed end-members. The created filaments become systematically thinner as continuously stretched and folded process develops. It exponentially increases the contact surfaces and, therefore, the mixing efficiency. In respect to the formation of the PEMP high-Ti Chapecó dacites, our results lead us to infer that basalts commonly melt crustal rocks, partially or totally, on their way to the surface. Initial calculations using the linear mixing model are in disagreement with our experimental results regarding the best contaminant candidate, which reflects the very complex simulated scenario. Nevertheless, the reproduced features such as the variability of basaltic compositions and chemical similarities with PEMP intermediary rare outcrops point towards short interaction times, low convective forces, and a predominance of density driven separation of contrasting melts (i.e., simulated conditions) as possible mechanisms involved in the genesis of Chapecó dacites.