Temperature-, pressure-, and composition-dependence of 18O/16O fractionation involving H2O-NaCl fluids

Abstract

Oxygen stable isotope fractionation involving water is a function of temperature, water density (or pressure), and dissolved components, namely salts. While temperature effects are well understood, knowledge of pressure and salinity effects is still patchy. A review of available experimental ¹⁸O/¹⁶O quartz-water and calcite-water fractionation data provides consistent evidence for a measurable pressure effect, i.e., both mineral–water fractionation factors decrease similarly with increasing pressure. While the decrease is small to negligible at temperatures above ca. 500 °C it amounts to ca. 1.5 ‰ when going from saturated water vapor pressure to 1500 MPa at 250 °C (equivalent to a water density increase from 800 to 1200 kg m⁻³) and explains most of the differences between different experimental calibrations. Fits to the experimental data and a theoretical analysis based on published concepts are used to construct a novel graphic representation of the ¹⁸O/¹⁶O reduced isotope partition function ratio (RIPFR) of water, demonstrating in a single diagram the effects of pressure and liquid–vapor fractionation. In a plot of the RIPFR vs. inverse squared temperature, isochores show much simpler trends than isobars and the change of the RIPFR can be represented as a single, empirical function of temperature and water density as

$1000\ln {\beta }^{\ast }=\left(3.695\times {10}^{-5}\frac{{10}^6}{T^2}-3.508\times {10}^{-5}\frac{{10}^3}{T}\right){\rho }^{1.5}$

with $1000\ln {\beta }^{\ast }$ being the normalized RIPFR obtained by subtracting the ideal gas contribution, $T$ temperature in Kelvin, and $\rho$ water density [kg m⁻³]. It is then demonstrated that at high temperatures the formula can also be used for aqueous NaCl solutions, using the solution’s density rather than the pure water density. Differences to low temperature data can be eliminated with a simple deviation term such that eventually a single formula allows reproducing almost all mineral–water/-solution and liquid–vapor (water or NaCl solution) experimental results within their uncertainties. Applying this formula to published experimental mineral–water fractionation data allows deriving mineral–water fractionation curves at constant water density. These corrected mineral–water fractionation lines can then be combined to derive consistent mineral–mineral fractionation factors. Where high-quality hydrothermal experiments are available, the latter then agree within error with those determined directly in anhydrous mineral–mineral exchange experiments, i.e., the previously purported, enigmatic discrepancies between the two experimental techniques appear to be inexistent in these cases. Namely, for quartz-calcite and magnetite-calcite excellent agreement is obtained and there appears to be no need to invoke cryptic salt effects assigned to dissolved minerals as a means to explain inferred discrepancies. For several other mineral pairs, the previously inferred discrepancies often simply result from inconclusive, scattered experimental data and/or too simple fits of experimental data that missed non-linear behavior at high temperatures.