Fig. 1
From: Sulfide resorption during crustal ascent and degassing of oceanic plateau basalts
Chalcophile element variations during crustal differentiation of OPB and MORB. a [Cu], b [Ag], c [S] and d [Se] versus [MgO]. The primitive Kroenke magmas have compositions which are similar to parental-MORB and are offset to higher [Cu] at a given [MgO] than samples from the ultra-slow spreading Gakkel Ridge, which are an example of low-degree partial melts50,51. Trajectories of increasing [Cu], [Ag], [S] and [Se] in a sulfide-under-saturated melt (gray-dashed triangle) are modeled assuming that these elements are perfectly incompatible and thus have a bulk-D of ~0. An additional trajectory for Cu in a sulfide-under-saturated melt is modeled using available partition coefficients for Cu in silicate phases (DCu values taken from ref. 39), giving the gray field on a. Many of the OPB samples plot below the “incompatible (D = 0) trajectories”, suggesting that the magmas are sulfide-saturated. Sulfide fractionation lines (red dashed lines, calculated using the slope for Cu and Ag in sulfide-saturated MORB magmas; ref. 1) suggest that the proportion of sulfides fractionated from OPB and MORB are comparable once saturation is achieved. OPB magmas saturate in sulfide at lower [MgO] than MORB. Unlike S, Se is not degassed from the Tamu Massif samples. Except Se (see Methods), data for MORB, parental-MORB and the Gakkel Ridge are taken from Jenner and O’Neill23, Jenner1 and Gale et al.50, respectively. Error bars for Se and Ag are given as the 1SD error on repeat analysis of the NWLSC Se and Ag standards (see Methods), with error bars on S* representing the propagated 1SD error of Se
