Fig. 5
From: Sulfide resorption during crustal ascent and degassing of oceanic plateau basalts
Relative bulk-partitioning of trace elements during crustal differentiation of OPB and MORB. Relative bulk-partitioning values (slopes; see Methods for details) for each element are calculated following the method of O’Neill and Jenner29. Two slopes are modeled for OPB (including and excluding two outliers which are enriched in highly incompatible elements (e.g., Rb, Cs and Th) at a given [MgO] compared to the other OPB samples) to provide a best-estimate of the range in compatibility of each element during crustal differentiation. Elements are ordered by increasing compatibility (i.e., more positive bulk-partitioning values) during MORB differentiation (MORB slopes taken from Jenner1, except for Mo, Se, Sn, which were calculated using the new MORB data presented in Supplementary Table 4 and are listed in Supplementary Table 5; calculations performed without plume-influenced MORB). Elements with a slope < 0 are incompatible (i.e., preferentially remain in the melt phase during fractional crystallization; bulk-D < 1) whereas elements with a slope > 0 are compatible (i.e., preferentially incorporated into silicates, sulfide or oxides; bulk-D > 1). Together, CSE in OPB with a Dsulf/sil > 300 (yellow symbols) are bulk-incompatible (bulk-D < 1), but have more positive slopes than Rb, which has an estimated bulk-D of ~0 and is therefore taken as a proxy for the behavior of CSE in a sulfide-under-saturated melt. These offsets (black arrows) suggest that the melts reached sulfide-saturation during differentiation. Importantly, Cu is offset to a higher slope compared to a modeled bulk-D which assumes the behavior of Cu in the melt during differentiation is controlled solely by partitioning into silicate phases. Error bars are the standard error of the slope (see Methods)
