Viscous flow behavior of tholeiitic and alkaline Fe-rich martian basalts

Abstract

The chemical compositions of martian basalts are enriched in iron with respect to terrestrial basalts. Their rheology is poorly known and liquids of this chemical composition have not been experimentally investigated. Here, we determine the viscosity of five synthetic silicate liquids having compositions representative of the diversity of martian volcanic rocks including primary martian mantle melts and alkali basalts. The concentric cylinder method has been employed between 1500°C and the respective liquidus temperatures of these liquids. The viscosity near the glass transition has been derived from calorimetric measurements of the glass transition. Although some glass heterogeneity limits the accuracy of the data near the glass transition, it was nevertheless possible to determine the parameters of the non-Arrhenian temperature-dependence of viscosity over a wide temperature range (1500°C to the glass transition temperature). At superliquidus conditions, the martian basalt viscosities are as low as those of the Fe–Ti-rich lunar basalts, similar to the lowest viscosities recorded for terrestrial ferrobasalts, and 0.5 to 1 order of magnitude lower than terrestrial tholeiitic basalts. Comparison with empirical models reveals that Giordano et al. (2008) offers the best approximation, whereas the model proposed by Hui and Zhang (2007) is inappropriate for the compositions considered.The slightly lower viscosities exhibited by the melts produced by low degree of mantle partial melting versus melts produced at high degree of mantle partial melting (likely corresponding to the early history of Mars), is not deemed sufficient to lead to viscosity variations large enough to produce an overall shift of martian lava flow morphologies over time. Rather, the details of the crystallization sequence (and in particular the ability of some of these magmas to form spinifex texture) is proposed to be a dominant effect on the viscosity during martian lava flow emplacement and may explain the lower range of viscosities (102–104Pas) inferred from lava flow morphology. Further, the differences between the rheological behaviors of tholeiitic vs. trachy-basalts are significant enough to affect their emplacement as intrusive bodies or as effusive lava flows. The upper range of viscosities (106–108Pas) suggested from lava flow morphology is found consistent with the occurrence of alkali basalt documented from in situ analyses and does not necessarily imply the occurrence of basaltic-andesite or andesitic rocks.