How partial melts of mafic lower crust affect ascending magmas at oceanic ridges

Abstract

We present experiments showing that the lower oceanic crust should melt efficiently and quickly when heated by hot ascending magmas. Average plagioclase–olivine and plagioclase–augite pairs from the lower crust at the Southwest Indian Ridge have melt–mineral saturation boundaries at 1,190 and 1,154°C, respectively, and melt rapidly (>0.01 mm/h) at 50°C or more above these temperatures. Melting experiments performed on olivine–plagioclase and augite–plagioclase mineral pairs from actual oceanic lower crustal rock samples and under conditions applicable to a MOR setting (1,220–1,330°C, 1 atm, quartz–fayalite–magnetite oxygen buffer, 0.25–24 h) indicate that the resulting disequilibrium melts are linear mixes of the mineral compositions. The rates of melting are slower than the rate of heat-diffusion into a sample and are approximated as:

$$ \begin{aligned}{} & \xi = 2.43 \times 10^{{( - 26)}} \times {\text{e}}^{{0.004109 \times T({\deg \text{C}})}} {\text{ }}({\text{m}}/{\text{s}}^{{1/2}} ){\text{ for}}\,{\text{augite}}\,{\text{melted}}\,{\text{with}}\,{\text{plagioclase}}; \\ & \xi = 3.47 \times 10^{{( - 15)}} \times {\text{e}}^{{0.002041 \times T({\deg \text{C}})}} \,{\text{for}}\,{\text{plagioclase}}\,{\text{melted}}\,{\text{with}}\,{\text{augite}}; \\ & {\text{and }}\xi = 1.79 \times 10^{{( - 21)}} \times {\text{e}}^{{0.0032 \times T({\deg \text{C}})}} \,{\text{for}}\,{\text{plagioclase}}\,{\text{melted}}\,{\text{with}}\,{\text{olivine}}{\text{.}} \\ \end{aligned} $$

Our results indicate that great care must be taken in backward models using basalt chemistry alone to explore mantle-melting processes, assuming only crystallization and fractionation during ascent, as partial melts may mix with intruded hot magma.