Abstract

Adakites are an enigmatic class of Na-rich volcanic rocks of andesitic-dacitic composition, characterised by strong depletions in Y and heavy rare earth elements. Their origin is still a matter of debate, although similar depletions in Archean TTDs (tonalitestrondhjemites-granodiorites) are often used with other evidence in support of an origin through slab melting. However, recent work suggests that many of the chemical characteristics of adakites can be produced from partial melting of newly underplated mafic lower crust. Irrespective of tectonic setting, the only compositionally appropriate protoliths for adakite magmas are metabasalts, with either fluid-absent or wet partial melting. Temperatures 650C are required for fluid-present slab melting at appropriate depths, and 850-900C where no fluid phase in initially present. How does one tell apart magmas generated by partial melting of similar source materials (metabasalts) yet formed by markedly different tectonic processes? We believe the key factors are the initial water contents of the magmas (trapped in early melt inclusions), the isotopic composition of the melts (recorded in the inclusions and their host crystals), and the eruption temperature. Recent experimental work suggests that while >15 wt.% H2O may be required to generate adakite magmas with appropriate phenocryst concentrations, the upper parts of sea mounts (for example the Nazca Ridge offshore Peru), could have been a substantial reservoir of H2O due to extensive sea floor alteration. Partial melting of such material could begin at lower temperatures and generate initial melts rich in H2O. To test the idea of fluid present melting origin for adakites, we have sampled 5Ma old rocks from the Yungay volcanic group, Peru. These rocks are rich in phenocrysts of zoned plagioclase and quartz, in which melt and fluid inclusions abound. Using Raman spectroscopic analysis of trapped melt and fluid inclusions from phenocrysts, it is possible to detect the presence or absence of H2Orich magmas considered diagnostic of low temperature, fluid-present slab melting. A range of synthetic glasses with adakite compositions and varying water contents have been prepared in order to calibrate the Raman instrument for the detection and quantification of H2O in melt inclusions. Together with microthermometry of both fluid and melt inclusions, this will allow interpretation of the melt trapping temperatures, and H2O contents of the magmas.