The anomalous nondipolar and nonaxisymmetric magnetic fields of Uranus and Neptune have long challenged conventional views of planetary dynamos. A thin-shell dynamo conjecture captures the observed phenomena but leaves unexplained the fundamental material basis and underlying mechanism. Here we report extensive quantum-mechanical calculations of polymorphism in the hydrogen-oxygen system at the pressures and temperatures of the deep interiors of these ice giant planets (to >600 GPa and 7,000 K). The results reveal the surprising stability of solid and fluid trihydrogen oxide (H₃O) at these extreme conditions. Fluid H₃O is metallic and calculated to be stable near the cores of Uranus and Neptune. As a convecting fluid, the material could give rise to the magnetic field consistent with the thin-shell dynamo model proposed for these planets. H₃O could also be a major component in both solid and superionic forms in other (e.g., nonconvecting) layers. The results thus provide a materials basis for understanding the enigmatic magnetic-field anomalies and other aspects of the interiors of Uranus and Neptune. These findings have direct implications for the internal structure, composition, and dynamos of related exoplanets.

Abstract Various metastable ice phases and their complicated transition pathways have been found by pressurization at low temperatures at which slow kinetics and high metastability are easily achieved. By contrast, such diversity is less expected at room or elevated temperatures. Here, using a combination of a dynamic diamond anvil cell and X-ray free electron laser techniques, we demonstrate that supercompressed water transforms into ice VI through multiple freezing–melting pathways at room temperature, hidden within the pressure region of ice VI. These multiple transition pathways occur via a metastable ice (more specifically, ice XXI with body-centred tetragonal structure ( $$I\bar{4}2d$$ I 4 ¯ 2 d )) discovered in this study and a metastable ice VII that exists within the pressure range of ice VI. We find that supercompressed water structurally evolves from high-density water to very-high-density water, causing multiple transition pathways. These findings provide an insight to find more metastable ice phases and their transition pathways at elevated temperatures. Here, using a combination of a dynamic diamond anvil cell and X-ray free electron laser techniques, we demonstrate that supercompressed water transforms into ice VI through multiple freezing–melting pathways at room temperature, hidden within the pressure region of ice VI. These multiple transition pathways occur via a metastable ice (more specifically, ice XXI with body-centred tetragonal structure ( I 4 ¯ 2 d )) discovered in this study and a metastable ice VII that exists within the pressure range of ice VI. We find that supercompressed water structurally evolves from high-density water to very-high-density water, causing multiple transition pathways. Main Water (H 2 O), composed of only two elements, forms numerous polymorphic phases from ice I h to ice XX (refs. 1 – 6 ) and four amorphous phases 7 , 8 . Understanding the formation and transition pathways of the diverse H 2 O phases has been of interest in high-pressure physics and the search for life in space and on icy moons 9 – 11 for a century. Generally, the abundant H 2 O phases result from the configurational changes of hydrogen-bond networks (HBNs), which are tuned by the interplay of increasing packing density and lowering bonding energy over a wide range of temperatures and pressures 12 . However, the instantaneous cooperative motion of H 2 O molecules in HBN can still be hampered at elevated temperatures, when highly compressed water crystallizes rapidly into dense and complicated hydrogen-bonded ice phases. For example, it has been reported that metastable ice VII (refs. 17 , 18 ) and high-density amorphous ice (HDA) 19 , 20 are formed during the fast crystallization of highly overpressurized water at room temperature within the stable ice VI or ice VII pressure regime. Theoretical studies have shown that plastic ice VII with defective hydrogen bonds can form from supercompressed water (SW) at elevated temperatures 21 – 23 , due to the decoupling of rotational and translational ordering of H 2 O molecules on rapid crystallization 21 . In addition, the formation of metastable phases in SW reflects the similar local structure of SW and the metastable phases, which can lower the nucleation barrier for the metastable phases rather than for stable phases 17 , 21 , 24 – 30 . This implies that SW can structurally evolve further with pressure, influencing the phase selection and freezing–melting pathways. The pressurization cycles are obtained within 10 s. a , Crystal growth is observed after the pressure drop and during the long melting plateau in type 1. b , First and second crystallizations are completed within 0.3 ms and 1 ms, respectively, in type 2. Inset: high-speed imaging at a rate of 50,000 fps, revealing that the first phase crystallizes within 40 μs (Supplementary Fig. 3 ). c , Mixture phase of ms-ice VII and water forms and melts away at the end of the pressure plateau at 1.6 GPa. d , e , Images of the second crystallization on decompression in type 4 ( d ) and type 5 ( e ), which occurs at 1.6 GPa (see the arrows in the P – t curves). This behavior is consistent with the type 2 crystallization sequence shown in Fig. 2b . Remarkably, the ice phase transforms into not only ice VI (Fig. 3e ) but also ms-ice VII (Fig. 3f and Supplementary Fig. 7b ). It should be noted that the reverse transition (from ms-ice VII to ice XXI) has not been observed in this study, even though both metastable phases can form directly from SW. The red lines in a – c denote the ice XXI phase, which are presumed, based on experimental observations (see details in the main text). d – f , Simulation results of SW are obtained using the SPCfw potential. d , Potential energy and density versus pressure. The black solid lines are guides for the eye. Data are presented as mean values, with the error bars indicating the s.d. calculated from 500 data points sampled every 10 ps over trajectories of 15–20 ns. e , PDFs of O–O atoms. Specifically, the former allows deviations of the H–O–H angle and O–H length of the water molecule from their equilibrium values caused by the formation of distorted hydrogen bonds, like in the structure of ice VI. Both models yield a similar trend of water properties with pressure, which qualitatively coincides with the experiment. The simulation system is a rectangular parallelepiped having 2,880 water molecules. Periodic boundary conditions are imposed in the x , y and z directions.

The elements hydrogen, carbon, nitrogen and oxygen are assumed to comprise the bulk of the interiors of the ice giant planets Uranus, Neptune, and sub-Neptune exoplanets. The details of their interior structures have remained largely unknown because it is not understood how the compounds H₂O, NH₃ and CH₄ behave and react once they have been accreted and exposed to high pressures and temperatures. Here we study thirteen H-C-N-O compounds with ab initio computer simulations and demonstrate that they assume a superionic state at elevated temperatures, in which the hydrogen ions diffuse through a stable sublattice that is provided by the larger nuclei. At yet higher temperatures, four of the thirteen compounds undergo a second transition to a novel doubly superionic state, in which the smallest of the heavy nuclei diffuse simultaneously with hydrogen ions through the remaining sublattice. Since this transition and the melting transition at yet higher temperatures are both of first order, this may introduce additional layers in the mantle of ice giant planets and alter their convective patterns.

We present a machine learning interatomic potential for water designed to capture its complex multiphase behavior, including both molecular and superionic ice phases. The potential is based on the atomic cluster expansion (ACE) formulation and has been parameterized to enable high-fidelity molecular dynamics simulations of water under extreme conditions, for pressures up to 100 GPa and for temperatures between 500 and 6000 K. A diverse range of configurations was generated through ab initio molecular dynamics (AI-MD) simulations, covering insulating and superionic ice phases, liquid water, and dissociated plasma phase. We demonstrate that the H 2 O ACE potential accurately reproduces experimental and DFT predicted isotherms and Hugoniots. Crucially, the potential is able to capture the intricate phase behavior of water, including the transition from molecular fluid to the appropriate solid ice phases, and the superionic ice phases. This work provides a robust interatomic potential that can be used for large-scale, accurate simulations of water under extreme thermodynamic conditions.

sulfur. There are two ice giants in the Solar System: Uranus and Neptune. In astrophysics and planetary science, the term "ice" refers to volatile chemical An ice giant is a giant planet composed mainly of elements heavier than hydrogen and helium, such as oxygen, carbon, nitrogen, and sulfur. There are two ice giants in the Solar System: Uranus and Neptune. In astrophysics and planetary science, the term "ice" refers to volatile chemical compounds with freezing points above about 100 K, such as water, ammonia, or methane, with freezing points of 273 An ice giant is a giant planet composed mainly of elements heavier than hydrogen and helium, such as oxygen, carbon, nitrogen, and sulfur. There are two ice giants in the Solar System: Uranus and Neptune. In astrophysics and planetary science, the term "ice" refers to volatile chemical compounds with freezing points above about 100 K, such as water, ammonia, or methane, with freezing points of 273 K (0 °C), 195 K (−78 °C), and 91 K (−182 °C), respectively. In the 1990s, it was determined (primarily by Voyager 2) that Uranus and Neptune were a distinct class of giant planet, separate from the other giant planets, Jupiter and Saturn, which are gas giants predominantly composed of hydrogen and helium. Neptune and Uranus are now referred to as ice giants. Lacking well-defined solid surfaces, they are primarily composed of gases and liquids. Their constituent compounds were solids when they were primarily incorporated into the planets during their formation, either directly in the form of ice or trapped in water ice. Today, very little of the water in Uranus and Neptune remains in the form of ice. Instead, water primarily exists as supercritical fluid at the temperatures and pressures within these planets. Uranus and Neptune consist of only about 20% hydrogen and helium by mass, compared to the Solar System's gas giants, Jupiter and Saturn, which are more than 90% hydrogen and helium by mass.