High δ56Fe values in Samoan basalts

Abstract:

Fe isotope fractionation spans ~0-0.4 permil in igneous systems, which cannot all be attributed to variable source compositions since peridotites barely overlap these compositions. Other processes may fractionate Fe isotopes such as variations in the degree of partial melting, magmatic differentiation, fluid addition related to the final stages of melt evolution, and kinetic fractionation related to diffusion. An important observation in igneous systems is the trend of increasing Fe isotope values against an index of magmatic fractionation (e.g. SiO₂; [1]). The data strongly curve from δ⁵⁶Fe >0.3 permil for SiO₂ >70 wt% down to values around 0.09 permil from ~65 wt% down to 40 wt% SiO₂ of basalts. However, ocean island basalts (OIBs) have a slightly larger δ⁵⁶Fe variability than mid ocean ridge basalts (MORBs; [e.g. 2]). We present Fe isotope data on samples from the Samoan Islands (OIB) that have unusually high δ⁵⁶Fe values for their SiO₂ content. We rule out alteration by using fresh samples, and further test for the effects of magmatic processes on the δ⁵⁶Fe values. In order to model the largest possible fractionation, unusually small degrees of melting with extreme fractionation factors are modeled with fractional crystallization of olivine alone, but such processing fails to fractionate the Fe isotopes to the observed values. Moreover, Samoan lavas likely also fractionated clinopyroxene, and its lower fractionation factor would limit the final δ⁵⁶Fe value of the melt. We therefore suggest the mantle source of Samoan lavas must have had unusually high δ⁵⁶Fe. However, there is no clear correlation with the highly radiogenic isotope signatures that reflect the unique source compositions of Samoa. Instead, increasing melt extraction correlates with lower δ⁵⁶Fe values in peridotites assumed to be driven by the preference for the melt phase by heavy Fe³⁺, while high values may be related to metasomatism [3]. The latter would be in line with metasomatized xenoliths from Samoa [4].

[1] Heimann et al., 2008, doi:10.1016/j.gca.2008.06.009

[2] Teng et al., 2013, doi:10.1016/j.gca.2012.12.027

[3] Williams et al., 2004, doi: 10.1126/science.1095679

[4] Hauri et al., 1993, doi: 10.1038/365221a0