Phosphorous-solubility in carbonatite melts: Apatite crystallization modeled via its solubility product

Abstract

We model apatite-saturation in carbonatite melts based on a compilation of experimental data ranging from 650 to 1430 °C and 1 to 60 kbar. The data show a very strong correlation of inverse temperature with the apatite solubility product, a relation expressed by the equation.

$\ln ({\left(CaO\right)}^5·{\left(P{O}_{2.5}\right)}^3)=-27450/T+5.79$

(T in Kelvin). Within the available dataset, F and Cl do not play a discernable role. Application of the solubility product to natural Ca-carbonatites indicates that a few rocks with >8 wt% P₂O₅ have cumulative apatite while most Ca-carbonatites (with typically <5 wt% P₂O₅) are apatite undersaturated at their liquidus temperatures, defined by calcite crystallization. To address true carbonatite liquids, we model calcite fractionation and melt evolution for natural rock compositions with 5, 10 and 20 mol% H₂O and/or (Na,K)₂CO₃ added, 5% representing the lower bound for any carbonatite formation model. Both H₂O or (Na,K)₂CO₃ cause very similar liquidus depressions of ∼10 °C/mol%. The model result is that saturation of apatite occurs in most natural carbonatite melts only after >45, 30–55, and 10–30 mol% calcite-fractionation for 5, 10, and 20 mol% fluxing components added, respectively. We further estimate the melt fractions necessary to dissolve all apatite in carbonatite melts generated from carbonated MORB and pelites, opening the discussion on an unlikely restitic nature of subducted apatites. In both the crystallization and forward melting cases, apatite crystallization or dissolution is mostly governed by temperature, surprisingly, carbonatite melt evolution through calcite-fractionation has a minor influence on the solubility product.