Experimental determination of hydrogen isotopic equilibrium in the system H2O(l)-H2(g) from 3 to 90 °C

Molecular hydrogen (H₂) is found in a variety of settings on and in the Earth from low-temperature sediments to hydrothermal vents, and is actively being considered as an energy resource for the transition to a green energy future. The hydrogen isotopic composition of H₂, given as D/H ratios or δD, varies in nature by hundreds of per mil from ∼−800 ‰ in hydrothermal and sedimentary systems to ∼+450 ‰ in the stratosphere. This range reflects a variety of processes, including kinetic isotope effects associated with formation and destruction and equilibration with water, the latter proceeding at fast (order year) timescales at low temperatures (<100 °C). At isotopic equilibrium, the D/H fractionation factor between liquid water and hydrogen (^Dα_H2O(l)-H2(g)) is a function of temperature and can thus be used as a geothermometer for H₂ formation or re-equilibration temperatures. Multiple studies have produced theoretical calculations for hydrogen isotopic equilibrium between H₂ and water vapor. However, only three published experimental calibrations used in geochemistry exist for the H₂O-H₂ system: two between 51 and 742 °C for H₂O_(g)-H_2(g) (Suess, 1949, Cerrai et al., 1954), and one in the H₂O_(l)-H_2(g) system for temperatures <100 °C (Rolston et al., 1976). Despite these calibrations existing, there is uncertainty on their accuracy at low temperatures (<100 °C; e.g., Horibe and Craig, 1995).

Here we present a new experimental calibration of the equilibrium hydrogen isotopic fractionation factor for liquid water and molecular hydrogen from 3 to 90 °C. Equilibration was achieved using platinum catalysts and verified via experimental bracketing by approaching final values of ^Dα_H2O(l)-H2(g) at a given temperature from both higher (top-bracket) and lower (bottom-bracket) initial ^Dα values. Our calibration yields the following equation:

$1000{\times \ln }^D{\alpha }_{H2O\left(l\right)-H2\left(g\right)}=-526.48+\frac{559,316}{T}$

Where T is in Kelvin. We find that our calibrations differ from prior experimental calibrations by, on average, up to 20 ‰ and prior theoretical results by up to, on average, 25 ‰. Good agreement with theoretical results (<11 ‰ differences) is found for calculations that consider both anharmonic effects and the Diagonal Born-Oppenheimer correction.