- 1Korzhinskii Institute of Experimental Mineralogy RAS, Laboratory of Magmatism, Metamorphism and Geodynamics of Lithosphere, Chernogolovka, Russian Federation (oleg@iem.ac.ru)
- 2Department of Petrology and Volcanology, Geological Faculty, Moscow State University, Moscow, Russian Federation
- 3Department of Geology, University of Johannesburg, Johannesburg, South Africa
IR spectra of garnets from mantle xenoliths, diamond inclusions and UHP metamorphic rocks indicate that this mineral can be a principal participant in a water balance of the mantle and subduction zones. This conclusion is consistent with experiments showing that H2O content in garnet increases with pressure and could reach up to >2000 ppm H2O at the transition zone conditions (Liu et al., 2024; Chen et al., 2025). Although being controversial (dependent on starting materials, i.e. crystalline natural garnets vs. synthesized ones), available experimental data allow determination regularities of the H2O solubility in garnet with respect to P, T, fO2 and garnet composition. Present study shows an attempt to parametrize these regularities.
The parametrization is not possible for the experiments with starting crystalline garnets (Lu, Keppler, 1997; Zhang et al., 2022; Zhang, Yang, 2025). These data are not consistent between each other, and the reason for the inconsistency is not clear. The data on garnets synthesized from oxide mixtures are better self-consistent. 54 data points from (Geiger et al., 1991; Khomenko et al., 1994; Withers et al., 1998; Mookherjee, Karato, 2010; Fan et al., 2017; Bolfan-Casanova et al., 2000; Katayama et al., 2003; Thomas et al., 2015; Panero et al., 2020; Liu et al., 2021, 2024) represent intervals 2 – 25 GPa and 900 - 2000°C for a wide range of garnet composition including majoritic ones. The H2O content (the Bell et al., 1995 calibration) in this set varies from 130 to 1620 ppm (the above mentioned data on the H2O content >2000 ppm in garnet were excluded). The data were approximated with an equation DH - TDS + (p-1)DV – nRTlnfH2O + RTlnCH2O + WAl*XAl2 + WSi*XSi2 + WCa*XCa3 = 0, where CH2O is the H2O content in garnet, DH = 0 kJ/mol, DS = -92.96(±3.94) J/mol/K, DV = 0.475(±0.022) J/mol/bar are thermodynamic effects of the reaction Grt + nH2O = Grt*nH2O, fH2O is a H2O fugacity (Pitzer, Sterner, 1995), n = 0.5, WAl = -30554.3(±4515) J/mol, WSi = -81777.4(±29415) J/mol, WCa = -498078.1(±131842) J/mol, XAl = [Al]/2 and XSi = ([Si] - 3])/2 – Al and Si mole fractions in the VI site and XCa = [Ca]/3 – Ca mole fraction in the VIII site ([Al], [Si], [Ca] – a.p.f.u. per 12 О). The equation reproduces the H2O content in garnet from 54 data points with a mean accuracy ±280 ppm.
Showing an increase of the H2O solubility in garnet with pressure and a decrease with temperature, the equation predicts a solubility maximum, which is dependent on temperature (for pyrope, it is 2400 ppm at 18.5 GPa for 1000°C and 1270 ppm at 22 GPa for 1500°C). The H2O solubility decreases with an increase of the majorite component in garnet. Following to these effects, the H2O content in garnet in the upper mantle is expected to be about 600-800 ppm along the sub-cratonic geotherm.
The study is fulfilled under support of the RSCF project 23-17-00066.
How to cite: Safonov, O.: Water content in garnet: review of available experimental data and parameterization with respect to temperature, pressure and composition , EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-173, https://doi.org/10.5194/egusphere-egu26-173, 2026.