# Apatite saturation revisited: new model formulations and applications to igneous rocks

## TL;DR

This paper introduces new models to predict when apatite forms in igneous rocks, improving accuracy and understanding of its behavior under various conditions.

## Contribution

The paper presents two new, accurate models for apatite saturation in igneous melts, outperforming existing methods.

## Key findings

- The first model predicts apatite saturation temperature with ~32°C accuracy, surpassing existing models.
- The second model estimates melt P2O5 content at apatite saturation with better than a factor of two accuracy.
- Apatite stability is not significantly affected by volatile species or pressure in the tested ranges.

## Abstract

Apatite is the primary phosphate mineral in the Earth’s crust and is present in a range of intrusive and extrusive igneous rocks, typically at minor to trace quantities. Similarly, apatite is a stable phase in many equilibrium crystallization experiments conducted with phosphorus-bearing starting compositions. We leverage this experimental stability to produce a large compilation of apatite-saturated liquid compositions, supplemented by additional apatite trace element and volatile partitioning and explicit apatite solubility experiments, as well as analyses of natural rhyolitic glasses. Using this compilation, we calibrate two new, independent models: for apatite saturation temperature as a function of melt P2O5 and SiO2 contents and aluminum saturation index (ASI; molar Al/(2Ca + Na + K)); and for melt P2O5 contents at apatite saturation as a function of temperature, melt SiO2 contents and ASI. The first model reproduces apatite saturation temperatures with an accuracy of ~ 32 °C, significantly outperforming existing apatite saturation models as well as recent zircon saturation thermometers. The second model reproduces melt P2O5 contents at apatite saturation by better than a factor of two across four orders of magnitude. Our new calibrations show that, within uncertainty, apatite stability does not differ regardless of the specific volatile species present (H2O, F or Cl) or on the quantity of H2O dissolved in the melt. Further, we find that apatite stability is not sensitive to pressures at least in the range of 1 atm to 2 GPa. This model accurately describes the saturation of apatite in a wide range of liquids, including metaluminous liquids, moderately alkaline liquids, and most peraluminous liquids. Experimental and natural peraluminous liquids with unusually high P2O5 contents that are not well described by our model contain CaO:P2O5 ratios below apatite stoichiometry (‘perphosphorous melts’), indicating that apatite saturation in these liquids is at minimum jointly controlled by CaO and P2O5 contents.

The online version contains supplementary material available at 10.1007/s00410-026-02300-5.