Decoupled Ca and Fe + Mg content of S-type granites: An investigation on the factors that control the Ca budget of S-type granites

The compositional variation of S-type granites is thought to derive from source-controlled processes whereby magmas acquire their composition as a result of entrainment of variable proportions of a peritectic-rich residual mineral assemblage into the melt. A corollary of this model is that the entrained phases keep the stoichiometric proportions they have in the partially melted source. These proportions are controlled by the composition of the subsolidus minerals. Consequently, it is expected that the ratio between maficity (FeOt + MgO) and other compatible major elements in granites would be the same as those of the source. This is true for Ti but not for Ca which shows systematically higher Ca:maficity ratios. In order to explore the factors that cause this discrepancy, thermodynamic calculations were used to study the partial melting of a representative metapelite and a metagreywacke under a range of crustal pressure and temperature conditions. The modelling included segregation of melt or magma in pulses, disequilibrium melting of plagioclase (in which only a fraction of this phase participates in the melting system and produces two populations of plagioclase, the unreacted cores with the original, Na-rich composition and a second one with a Ca-rich composition) and the entrainment of up to 40 wt% minerals into the melt. This entrained assemblage consists of two fractions, one containing the minerals that are most likely to be surrounded by melt in the source and representing 80 wt% of the total assemblage, and another one consisting of the phases that are most likely to make up the solid framework, with the minerals within each fraction keeping stoichiometric proportions. The modelled magmas failed to reproduce the observed high Ca:maficity ratios. This indicates that there are other processes that increase the Ca content of granites. Several processes were evaluated, including the presence of apatite, enhanced availability of subsolidus plagioclase in the melting system, disequilibrium Ca-rich melts and enhanced entrainment of Ca-rich plagioclase. The last three processes were found to increase the Ca content but only enhanced entrainment of Ca-rich plagioclase increases the Ca:maficity ratio to values similar to those of the granites. A possible explanation for the preferential entrainment of Ca-rich plagioclase is that the density of this phase is different to that of garnet and ilmenite and similar to the density of the melt. Consequently, a significant fraction of the plagioclase in granites may actually be of residual origin. Other relevant findings of this study are that metagreywackes seem to be the main source of S-type granites, as previous investigations suggested, and that cordierite-rich granitoids represent mineral-rich magmas that are formed only from metapelites at pressures below 0.75 GPa. This investigation emphasizes the importance of source-controlled processes in the composition of S-type granites and indicates that entrainment of a mineral assemblage into the melt is not a simple and straightforward process as previously thought.