Formation of hybrid arc andesites beneath thick continental crust Susanne M. Straub a,b, ⁎, Arturo Gomez-Tuena c , Finlay M. Stuart d , Georg F. Zellmer b , Ramon Espinasa-Perena e , Yue Cai a,f , Yoshiyuki Iizuka b a Lamont Doherty Earth Observatory at the Columbia University, 61 Route 9W, Palisades NY 10964, USA b Institute of Earth Sciences, Academia Sinica, 128 Academia Road, Sec. 2, Nankang, Taipei 11529, Taiwan, ROC c Centro de Geociencias, Universidad Nacional Autónoma de México, Querétaro 76230, Mexico d Isotope Geosciences Unit, Scottish Universities Research and Reactor Centre, East Kilbride G75 0QF, UK e Instituto de Geofisica, Universidad Nacional Autónoma de México, Ciudad Universitaria, Mexico, D.F. 04510, Mexico f Department of Earth and Environmental Sciences, Columbia University, 61 Route 9W, Palisades NY 10964, USA abstract article info Article history: Received 7 November 2010 Received in revised form 12 January 2011 Accepted 13 January 2011 Editor: R.W. Carlson Keywords: helium isotopes high-Ni olivine andesite formation Mexican Volcanic Belt Andesite magmatism at convergent margins is essential for the differentiation of silicate Earth, but no consensus exists as to andesite petrogenesis. Models proposing origin of primary andesite melts from mantle and/or slab materials remain in deadlock with the seemingly irrefutable petrographic and chemical evidence for andesite formation through mixing of basaltic mantle melts with silicic components from the overlying crust. Here we use 3 He/ 4 He ratios of high-Ni olivines to demonstrate the mantle origin of basaltic to andesitic arc magmas in the central Mexican Volcanic Belt (MVB) that is constructed on ~ 50 km thick continental crust. We propose that the central MVB arc magmas are hybrids of high-Mg# N 70 basaltic and dacitic initial mantle melts which were produced by melting of a peridotite subarc mantle interspersed with silica-deficient and silica-excess pyroxenite veins. These veins formed by infiltration of reactive silicic components from the subducting slab. Partial melts from pyroxenites, and minor component melts from peridotite, mix in variable proportions to produce high-Mg# basaltic, andesitic and dacitic magmas. Moderate fractional crystallization and recharge melt mixing in the overlying crust produces then the lower-Mg# magmas erupted. Our model accounts for the contrast between the arc-typical SiO 2 variability at a given Mg# and the strong correlation between major element oxides SiO 2 , MgO and FeO which is not reproduced by mantle–crust mixing models. Our data further indicate that viscous high-silica mantle magmas may preferentially be emplaced as intrusive silicic plutonic rocks in the crust rather than erupt. Ultimately, our results imply a stronger turnover of slab and mantle materials in subduction zones with a negligible, or lesser dilution, by materials from the overlying crust. © 2011 Elsevier B.V. All rights reserved. 1. Introduction Andesite magmas at convergent margins are enriched in silica compared to magmas erupting at mid-ocean ridges and intra-plate volcanoes. Determining the cause(s) of silica enrichment is funda- mental for models of continental crust formation, arc growth rates and across-arc mass balances (Plank and Langmuir, 1993; Rudnick, 1995; White et al., 2006). Andesite petrogenesis, however, has long been controversial, with no consensus even whether andesitic magmas form in the subarc mantle or in the overlying crust (Rudnick, 1995; Rudnick and Gao, 2002). Many models assume a basaltic flux from mantle to arc crust, with silicic magmas evolving subsequently in the upper plate crust through fractional crystallization and crustal assimiliation of up half of the mass of the erupted melt (e.g. Eichelberger, 1978; Leeman, 1983; Hildreth and Moorbath, 1988; Plank and Langmuir, 1988; Streck et al., 2007; Tatsumi et al., 2008; Reubi and Blundy, 2009). Other models propose primary andesite formation beneath the Moho, which may occur by various mecha- nisms such as hydrous melting of peridotite (Hirose, 1997; Moore and Carmichael, 1998; Blatter and Carmichael, 1998b; Carmichael, 2002), slab melting (Defant and Drummond, 1990) or hybridization of slab and mantle materials by melt rock-reaction processes (Kay, 1978; Yogodzinksi et al., 1994; Kelemen, 1995; Yogodzinski et al., 1995; Rapp et al., 1999; Kelemen et al., 2003, 2004; Gomez-Tuena et al., 2007). These two approaches differ substantially with respect to the turnover of slab and mantle materials in subduction zones, the rate of crustal growth, and the overall connectivity between arc magmatism and the other geochemical cycles of Earth. A recent study suggests that ‘high-Ni’ olivines, that have been by now reported from several arcs, e.g. Mexico, Cascades, Setouchi, Kamchatka and the Aleutians (e.g., GeoROC, 2009) may provide Earth and Planetary Science Letters 303 (2011) 337–347 ⁎ Corresponding author at: Lamont Doherty Earth Observatory at the Columbia University, 61 Route 9W, Palisades NY 10964, USA. Tel.: + 1 845 365 8464; fax: + 1 845 365 8155. E-mail address: smstraub@ldeo.columbia.edu (S.M. Straub). 0012-821X/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.epsl.2011.01.013 Contents lists available at ScienceDirect Earth and Planetary Science Letters journal homepage: www.elsevier.com/locate/epsl