New insights into H2O wt.% content in a residual melt during hornblendite crystallization in the Fuegian Batholith (Patagonia) 

<p>Hornblendites in the Patagonian Batholith are less known than in other well-studied sites (e.g., Adamello Pluton, Bonanza arc). However, they offer a remarkable opportunity to constrain the structural depth of differentiation processes within the arc and the structure of magmatic plumbing systems in the southernmost Andes. It is widely accepted that arc magmas can undergo significant modification at the base of the arc crust without leaving evidence in their mineral assemblages due to cumulate-melt reaction, as seen in deep crustal sections worldwide. Arc cumulates have the potential to record textural and chemical information from hidden and deep-seated crustal physicochemical processes in the first stages of magma differentiation. The presence of volatiles in subduction zones plays a fundamental role in many processes involved in magma genesis and differentiation, and in this case, water is the most abundant. It has a significant effect on the chemistry of growing crystals and the lifetime of the evolution of a magmatic system.</p><p>The Fuegian Batholith (53-56&#186;S) represents the roots of a Mesozoic to Cenozoic magmatic arc build-up by episodically assembled calc-alkaline plutons. The early mafic and ultramafic plutons of the Fuegian Batholith were emplaced contemporaneously with steep subduction of the oceanic slab and the development of the Late Jurassic-Early Cretaceous Rocas Verdes back-arc basin with oceanic-type crust.</p><p>New hygrometric calculations based on Ridolfi and Renzulli (2012) indicate a residual melt segregated from amphibole-rich mushes with H₂O contents up to ~9-11wt%.</p><p>We interpret plagioclase-hornblendites as the result of progressive fractional crystallization of a primitive initial batch of a hydrous (~3 wt.% of water) basaltic melt, constituting a high-pressure (7-8 Kbar at 980-1010 &#176;C) hydrous crystallization trend. The early stabilization of amphibole is reflected by the high Mg# (0.71) and Al^Vl, which indicates crystallization at lower crustal depths. Geothermobarometric constraints are consistent with experimental studies and the occurrence of Mg-rich amphibole with high Al^VI content (2.1 a.p.f.u.) in similar amphibole&#8211;bearing ultramafic cumulates reported in other lower crustal sections (e.g. Larocque et al., 2009).&#160;</p><p>Due to the lack of other mafic phases, it is assumed that amphibole is either formed by a peritectic reaction consuming clinopyroxene, leaving no trace or that it crystallized directly from the melt. In either case, early and cryptic crystallization of clinopyroxene and olivine in earlier magmatic reservoirs cannot be discarded. The early-stage fractionation of these minerals would increase the H₂O content in residual melts. In addition, by numerical simulations, Petrelli et al. (2018) suggest an initial batch of magma containing 3.9&#177;0.4 wt.% H₂O (constant global average of arc volcano mafic magmas), stored at mid-to deep-crustal levels (~20-35 km), can evolve to a water-rich (H₂O ~6-9 wt.%) residual melt in timescales ranging from a few to several thousand years.</p><p>These results are in agreement with our observations. Moreover, amphibole-bearing lithologies have been shown to enhance melt productivity within the crust. Therefore, it could be a reservoir to recycle water and incompatible elements into the mantle. This study was supported by the Fondecyt grants 1161818, 1211906 and the Anillo Project ACT-105.</p>