The Origin of Rhyolite and Plagiogranite in Oceanic Crust: An Experimental Study

Abstract

Hydrothermal and anhydrous experiments were conducted on a primitive oceanic tholeiite dredged from the Galapagos Spreading Center (DO8: mg = 0·65, K2O = 0·06) and on a more evolved oceanic tholeiite from Hawaii (Kilauea: mg = 0·53, K2O = 0·44) in order to characterize the minerals and residual melts present over the range 10–100 per cent melting. The hydrothermal experiments were done at fluid pressures in the range 1–3 kb, using the G-CH buffer (Wustite field) which means that fH2O ∼ 0·67 pfluid and fH2O ⋜ 0·33 Pfluid. The melts produced in the hydrothermal experiments, in contrast to similar amounts of melt produced under anhydrous conditions are characterized by strong SiO2 enrichment, a major enrichment in A12O3, and no FeO enrichment at ≦50 per cent residual melt. The liquid line of descent produced from the more evolved Kilauea basalt at Pfluid = 1 kb is similar to that followed by the Thingmuli, Icelandic magmas (Carmichael, 1964). Both the Thingmuli and the experimental (1 kb) Kilauea liquid lines of descent are characterized by an early iron enrichment due to the crystallization of olivine, pyroxene, and plagioclase, and a later SiO2 enrichment primarily as a result of Fe-Ti oxide crystallization. The Kilauea residual melt after 90 per cent crystallization (or 10 per cent partial melting) is a granite melt (SiO2 = 73·1, Na2O = 4·32, K2O = 2·39) which is similar in composition to the naturally occurring Icelandic rhyolites. Therefore, the conclusion is drawn that it is possible to produce a natural Icelandic granite from an evolved Kilauea-type basalt under low fO2⁠, Pfluid = 1 to 2 kb, and PH2O < Ptotal conditions, a major factor being the fractionation of Fe-Ti oxides. The residual melt produced from the DO8 tholeiite after 90 per cent crystallization at 1 kb is a low-K2O granitic melt which resembles the low-K2O granitic rocks (plagiogranites) found in M.O.R. dredge hauls and in ophiolite suites. Silicate liquid immiscibility, which occurred in 1 atm DO8 experiments, did not occur in the hydrothermal experiments. The data collected in this experimental study at low fO2⁠: together with that collected in previous higher fO2 experiments indicate that, regardless of fO2⁠;, amphibole coexists with melt only at pressures≧2 kb where PH2O is in the range 0·6 to 1 ·0 Pfluid- Unless amphibole is stabilized by some additional factor, i.e. F substitution for OH or lowering of aH2O⁠, it appears that amphibole fractionation from a basaltic melt to produce a high SiO2 melt is only possible at Pfluid ≥ 2 kb (>6 km). The results of this study also demonstrate that high SiO2 melts can be produced hydrothermally at low values of fO2 by fractionation of Fe-Ti oxides from a tholeiitic basalt at low pressures, i.e. at depths of 3–6 km in the oceanic crust.