Diffusion–reaction in a thermal gradient: Implications for the genesis of anorthitic plagioclase, high alumina basalt and igneous mineral layering

Abstract

Piston-cylinder experiments investigating the interaction between basaltic andesite melt and partially molten gabbro in a thermal gradient provide insight into melt-rock reaction processes occurring during magma differentiation in the crust. In two experiments juxtaposing basaltic andesite and gabbro at 0.5 GPa pressure for durations of either 13 or 26 days, diffusive chemical exchange between the two materials results in mineral layering and notable mineral compositions such as anorthitic plagioclase. Specifically, the basaltic andesite gains Al 2O 3, MgO and CaO from the gabbro and loses Na 2O, K 2O, SiO 2 and FeO to it with a plagioclase-rich layer developing at the interface between the two materials in a process termed diffusion–reaction. The percent crystallinity of the basaltic andesite increases during the process and the plagioclase crystals within the interface region develop anorthitic cores (up to An 90) that abruptly shift in composition to thin rims that are in Na–Ca exchange equilibrium with the co-existing melt. Both the mineralogical layering and bulk compositional change occurring at the interface are reproduced in model simulations of diffusion–reaction. Isotopic tracers ( 45Ca, 6Li, 84Sr and 136Ba) initially deposited at the basaltic andesite–gabbro interface in the 13-day experiment were detected in the cores of the anorthitic plagioclase after the experiment, demonstrating that the melt chemically communicates with the plagioclase cores over the duration of the diffusion–reaction experiment. The formation of anorthitic plagioclase during diffusion–reaction may explain its widespread occurrence in terrestrial volcanic rocks without requiring the presence of ultra-calcic melts. Textures and mineralogical changes in the gabbro indicate that chemical transport occurs throughout the experiments despite temperatures at the cold end of the experimental capsule approaching 500 °C. For instance, apatite, FeNiS, olivine and almost pure albite occur at distinct, specific horizons in the gabbro within the 26-day experiment. Because the bulk element profiles indicating chemical transport reflect analyses of almost completely solid gabbro, equilibration between minerals and fluids/melts must be rapid. The overall effect of the diffusion–reaction process is to make an ascending magma more primitive in composition (and in this case, produce anorthitic plagioclase) while making surrounding crustal wall rocks more evolved. Several observations within igneous rocks support the occurrence of this process, suggesting that the genesis of porphyritic high alumina basalt, ubiquitously observed at convergent margins, could reflect a diffusion–reaction process in the crust.