Apatite dissolution into peraluminous haplogranitic melts: An experimental study of solubilities and mechanisms

Apatite (Ap) dissolution and solubility in the peraluminous haplogranite system were investigated in three sets of experiments at 750°C and 200 MPa (PH2o), as functions of Ap grain size, abundance (melt: Ap ratio), and melt ASI [alumina saturation index: mol. Al2O3/(Na2O + K2O + CaO)]. Between an ASI of 1–1.1 (metaluminous to mildly peraluminous), the solubility of Ap, as measured by P2O5 in the melt, is low (0.1 wt% P2O5), which agrees with previous work. In equilibrium runs with more strongly peraluminous compositions, however, Ap solubility increases linearly with ASI, to P2O5 ~ 0.63 wt% in melt at ASI = 1.3, and can be denned by a simple equation: P2O5 = −3.4 + 3.1 × ASI (R = 0.833), which agrees within error with previous calculations reported in the literature. In experiments where Ap grains are sparce and widely separated (i.e., simulating Ap distribution in an Ap-bearing peraluminous protolith undergoing anatexis) the dissolution of Ap involves transient disequilibrium that promotes higher solubility of P2O5 (up to nearly six times the equilibrium values) adjacent to Ap at a given bulk ASI of melt, and this increases the Al:Si of melt in the Ap dissolution aureole. The approach to equilibrium between Ap and melt is hastened by increasing the Ap:melt ratio within the local melt pool (e.g., the capsule) or by decreasing Ap grain size (for a similar melt: Ap ratio). Diffusivity calculations, using inverse error function methods for the normal diffusion of P and Ca, yield Dp ~ 10−11 cm2/s and DCa ~ 10−10 cm2/s, confirming prior results that diffusion of P away from Ap surfaces is the rate-limiting factor in Ap dissolution. These results, combined with previous work at higher temperatures but on similar hydrous melts, yield an activation energy (E) for P diffusion of 16.5 kcal/mol and a frequency factor (D0) of 1.4 × 10−3 cm2/s. Phosphorus diffusion through the melt is accompanied by uphill diffusion of Al, resulting in an increase of up to 2 wt% Al2O3 in melt near the Ap. Within the Ap dissolution aureole, melt contains one additional Al cation for every P cation (i.e., ΔAl:ΔP = 1:1) which is suggestive of an associated AlPO4 complex. The exchange reaction in melt is represented by AlPSi−1, in which 2 TO4 clusters are created in the place of 1; this exchange should be accompanied by discernable changes in melt properties (e.g., density, molar volume, viscosity). The results have applications to equilibrium and local disequilibrium effects during anatexis of Apbearing metapelites, and to the saturation limits of Ap during crystallization of peraluminous magmas. The dissolution of flux-bearing minerals such as Ap (e.g., those containing P, B, or F) may influence the stability and solubility of other trace element-bearing minerals during anatexis of metapelites. Texture (e.g., grain size and distribution, and spacial relationship to other minerals) plays an important role in reaction kinetics and, at least temporarily, in mineral solubilities. The disequilibrium behavior of accessory minerals has critical consequences for REE modelling of granites.