The effect of diffusive re-equilibration time on trace element partitioning between alkali feldspar and trachytic melts

We present new experimental data on major and trace element partition coefficients between alkali feldspar and trachytic melt. Experiments were conducted at 500 MPa, 870 890 {\deg}C to investigate through short disequilibrium and long near equilibrium experiments the influence of diffusive re-equilibration on trace element partitioning during crystallization. Our data show that Ba and Sr behave compatibly, and their partition coefficients are influenced by re-equilibration time, orthoclase (Or) content, growth rate and cation order-disorder. High field strength elements (HFSE) and rare earth elements (except Eu) are strongly incompatible, but alkali feldspar efficiently fractionates light (LREE) from heavy rare earth elements (HREE). Our crystallization experiments reveal a strong influence of disequilibrium crystal growth on the partitioning of Ba and Sr. In particular, short-duration experiments show that rapid alkali feldspar crystal growth after nucleation, promotes disordered growth and less selectivity in the partitioning of compatible trace elements that easily enter the crystal lattice (e.g., Ba and Sr)….

💡 Research Summary

This study investigates how diffusive re‑equilibration time and crystal growth kinetics control trace‑element partitioning between alkali feldspar and trachytic melt. Experiments were performed at 500 MPa and 870–890 °C under water‑undersaturated conditions (~2 wt % H₂O). Two contrasting experimental regimes were employed: short‑duration runs (6–39 h) that produce rapid, disequilibrium crystal growth, and long‑duration runs (up to 192 h) that allow near‑equilibrium growth. The starting material was a natural trachytic obsidian (ZAC) from the Campi Flegrei ignimbrite, homogenised, spiked with a trace‑element mix (TX7), and then melted to generate a glass of known composition.

Textural analysis using back‑scattered electron imaging showed that short experiments yielded small (50–120 µm) feldspar crystals with high growth rates (~10⁻⁶ m s⁻¹) and irregular morphologies, whereas long experiments produced larger (200–350 µm) crystals growing at ~10⁻⁸ m s⁻¹ with smooth, well‑developed faces. Orthoclase (Or) content increased from ~0.55 in short runs to ~0.75 in long runs, indicating progressive Na–K exchange and lattice ordering during slower growth.

Major‑element compositions were measured by EMPA, and trace‑element concentrations (including Ba, Sr, HFSE, and REE) were obtained by LA‑ICP‑MS. The partition coefficients (D = C_crystal/C_melt) for Ba and Sr displayed a strong dependence on experiment duration. In short, fast‑growth experiments D_Ba ranged from 1.6 to 2.0 and D_Sr from 1.4 to 1.8, reflecting a compatible behaviour. In contrast, long, near‑equilibrium experiments yielded D_Ba ≈ 0.8–1.0 and D_Sr ≈ 0.7–0.9, indicating reduced compatibility. High‑field‑strength elements (Zr, Nb, Ti) and most rare‑earth elements (except Eu) remained strongly incompatible (D < 0.05) regardless of growth rate. Light REE showed modest partitioning (D ≈ 0.2–0.4) while heavy REE were even less compatible (D ≈ 0.05), demonstrating that alkali feldspar efficiently fractionates LREE from HREE.

The authors interpret the Ba–Sr behaviour in terms of lattice‑strain and cation ordering. Rapid growth creates a disordered crystal lattice with many defects, allowing large divalent cations to be incorporated more readily, thus raising D. As growth slows, the lattice becomes more ordered, and the size‑ and charge‑selective nature of the 12‑fold coordinated sites becomes dominant, lowering D. A clear non‑linear relationship emerges between growth rate and D: once the growth rate exceeds ~10⁻⁷ m s⁻¹, D for Ba and Sr increases sharply, suggesting an “order‑disorder” threshold.

Applying these experimental calibrations to natural alkali feldspar from the Campi Flegrei volcanic complex, the authors estimate that the observed Ba and Sr partition coefficients correspond to crystallization times of up to 6 days under disequilibrium conditions and a minimum of 9 days when growth approaches equilibrium. This provides a quantitative “clock” for pre‑eruptive magma storage and ascent.

The study highlights that trace‑element partitioning is not solely a function of temperature, pressure, and melt composition; kinetic factors such as diffusive re‑equilibration time and crystal growth rate can dominate, especially for compatible large cations. The findings extend traditional equilibrium partitioning models by incorporating growth‑rate‑dependent ordering effects, offering a more realistic framework for interpreting magmatic processes in evolved, alkali‑rich systems. Future work should explore the influence of water content, higher pressures, and different melt compositions, as well as direct imaging of the diffusive boundary layer using high‑resolution TEM or atom‑probe tomography.