Origin of Giant Miocene and Pliocene Cu-Mo Deposits in Central Chile: Role of Ridge Subduction, Decreased Subduction Angle, Subduction Erosion, Crustal Thickening, and Long-Lived, Batholith-Size, Open-System Magma Chambers

Three of the world's largest Cu-Mo deposits, Los Pelambres, Rio Blanco-Los Bronces and El Teniente, formed in close temporal association with southward migration of the locus of subduction of the Juan Fernandez Ridge and the resultant decrease in subduction angle below central Chile during the Miocene and Pliocene. All three contain large Cu-mineralised magmatic-hydrothermal biotite itourmaline ±anhydrite breccia pipes generated by exsolution of saline, high-temperature fluids from crystallising magmas. Sr, Nd, Pb, S, Os, 0 and H isotopic data indicate that the metals these breccias contain, and aqueous fluids responsible for their emplacement, were derived from the same magmas that produced igneous rocks associated with each deposit. Isotopic data are consistent with derivation of these magmas from subduction-modified subarc mantle, and suggest that formation of these deposits did not involve either dehydration or melting of continental crust. Each deposit formed by multiple mineralising events occurring over a >2 m.y. period during which there is no evidence for coeval volcanic activity. Assuming an average Andean magma with 100 ppm Cu, the original lOOxlO6 tonnes of Cu in each deposit prior to erosion requires a parent body of magma with a batholith-size dimension of approximately >600 km3. We suggest that the multiple Cu-mineralised breccia pipes in each deposit were generated by exsolution of magmatic fluids from the roofs of large, long-lived, open-system magma chambers, crystallising at depths of >4 km below the palaeosurface as indicated by

geologic constraints. Input of mantle-derived mafic magmas into the base of these chambers provided heat for their progressive growth and persistence, as well as Cu, S, Fe, Ca and Cl-rich aqueous fluids which migrated to the tops of the chambers due to thermal gradients. As mafic magma supply from the mantle decreased, due to decreasing subduction angle caused by ridge subduction in the latest Miocene and Pliocene, the progressive growth stage of these magma chambers ended and they solidified. Crustal thickening, uplift and erosion speeded this crystallisation and de-fluidisation process, and caused telescoping of Cu-mineralised biotite followed by tourmaline breccias both as the roofs of these chambers became progressively closer to the surface and as multiple brecciation events caused confining pressures to change from lithostatic to hydrostatic conditions.  Each deposit contains small, late, barren or weakly mineralised felsic porphyry intrusions with high "adakite-like" Sr/Y and La/Yb ratios. These felsic porphyries are both too small to have been the source of the enormous amount of Cu in these giant deposits, and they post-date the emplacement of the main mineralised breccias.  Furthermore, they are not true adakites.  Their isotopic compositions, which are similar to all the other igneous rocks associated with these deposits, indicate that they are not products of either slab-melting, nor melting of continental crust. Their unique chemical characteristics may result from crystal-liquid fractionation within, and extensive fluid transfer into and out of the tops of the crystallising batholith-size magma chambers that generated the mineralised breccias in each deposit.