Scientists have synthesized two previously unknown iron oxyhydroxides that remain stable at the immense pressures and temperatures of Earth’s deep mantle. The phases are likely the first experimentally demonstrated water-bearing solids synthesized and stable under conditions overlapping a substantial portion of the mantle geotherm—providing a long-sought mineralogical host for water near the base of the mantle.

The study identifies the hexagonal compounds Fe₅O₁₂Hₓ (x ≥ 9) and Fe₇O₁₂Hₓ (x ≥ 3). They formed in laser-heated diamond anvil cells between 78 and 198 gigapascals and at temperatures of 2,400–2,800 kelvin. These conditions extend across much of the deep lower-mantle pressure range and overlap the temperatures expected along the mantle geotherm.

Water profoundly affects mantle melting, deformation and chemical transport, yet the major minerals of the lower mantle can store only very small amounts of hydrogen. Previously proposed hydrous phases either require unusually cold subducting slabs, restricted chemical compositions or break down at high mantle temperatures. The new iron oxyhydroxides therefore fill an important gap: they are dense, highly hydrated solids that can coexist with the principal minerals of the lower mantle.

Recreating the deep mantle in the laboratory

The international team included Center for High Pressure Science and Technology Advanced Research (HPSTAR), the Bayerisches Geoinstitut at the University of Bayreuth, the ESRF, ETH Zürich, the Shanghai Institute of Applied Physics, DESY and Goethe University Frankfurt.

They came to the ESRF’s beamline ID11, ID14 and ID15B where they compressed iron-bearing starting materials between diamond anvils and heated microscopic samples with lasers, using the dedicated laser-heating setup of ID14, while collecting synchrotron X-ray diffraction data. “Iron oxyhydroxides phases form as tiny crystals several microns in size. Thus, the usage of the X-ray beam of micron size or below, provided by ID11 and ID15B beamlines is indispensable to identify their structures”, explains Ilya Kupenko, co-author of the publication and ESRF scientist. 

In situ single-crystal diffraction revealed two hexagonal structures built from unusually highly coordinated iron–oxygen polyhedra. The compounds formed not only from deliberately water-rich mixtures, but also from nominally anhydrous starting materials containing only trace adsorbed moisture. This shows that small amounts of available hydrogen may be sufficient to stabilise highly hydrated iron-rich domains.

A reservoir for primordial and recycled water

Geochemical signatures in volcanic rocks from places such as Baffin Island and Iceland have been interpreted as evidence that some primordial water survived deep inside Earth. One possible reservoir is the dense residue of an ancient basal magma ocean that crystallized above the core–mantle boundary. As that magma ocean cooled, its remaining melt would have become enriched in both iron and water—the ingredients needed to form the newly identified phases.

The same compounds could also capture recycled water delivered downward by subducting tectonic plates. Because the oxyhydroxides are dense and compatible with bridgmanite, post-perovskite, ferropericlase and davemaoite, they could accumulate near the core–mantle boundary rather than being readily swept upward by convection. In this way, they may chemically anchor both primordial and recycled water in the lowermost mantle.

The Fe₅O₁₂Hₓ phase is especially hydrogen-rich: the conservative composition x = 9 corresponds to about 1.7 wt% hydrogen. The authors estimate that a global abundance of only about 1% of this phase could, in principle, accommodate the estimated hydrogen inventory of the bulk silicate Earth. In a water-poor ambient mantle, however, the phases would probably occur in small, concentrated domains rather than as a major global mineral component.

Possible fingerprints at the base of the mantle

Formation of the oxyhydroxides extracts iron from surrounding silicate minerals. In the experiments, coexisting bridgmanite became strongly depleted in iron, a change that may sharpen the transition to post-perovskite and help explain regional variations in the seismic Dʺ discontinuity. Dense accumulations of iron–oxygen–hydrogen material could also contribute to ultralow-velocity zones—thin, seismically anomalous patches immediately above the core–mantle boundary.

The researchers further propose that these water-rich piles could act as a volatile “fuse” for mantle plumes. If the oxyhydroxides melt more readily than the surrounding silicates during heating from the core, they could release stored water and other volatiles into rising melts, helping connect deep reservoirs with volcanic activity at Earth’s surface.

The experiments demonstrate stability up to approximately 2,800 K. The authors caution that survival at the very hottest temperatures estimated directly along the core–mantle boundary—roughly 3,000–3,500 K—has not yet been tested. Direct measurements of hydrogen positions and concentrations, for example by future spectroscopic or neutron-diffraction studies, will also be needed to refine the compositions.

Reference:

H. Yuan, et al., Nature Geoscience (2026). DOI: 10.1038/s41561-026-02088-w.