Hydrogen Superionic Diffusion

Updated 2 December 2025

Hydrogen superionic diffusion is a state where hydrogen ions flow like a liquid within a fixed crystalline lattice, marked by dramatic increases in ionic conductivity.
It depends on the host lattice’s bonding topology, stoichiometry, and structural flexibility to create interconnected pathways for rapid proton migration.
This phenomenon impacts energy storage, planetary magnetic field generation, and advanced material design, studied using ab initio MD and machine learning techniques.

Hydrogen superionic diffusion describes the regime in which hydrogen ions exhibit liquid-like mobility within a solid or highly structured matrix, while the host lattice (typically formed by heavier elements such as O, N, metals, or other framework constituents) remains crystalline on relevant timescales. This phenomenon underlies a range of behaviors in planetary ices, hydrous minerals, hydrides, and engineered materials and is closely linked to the functional properties of planetary interiors, energy storage, geochemistry, and materials design. The transition to the superionic state, the character of hydrogen migration, and the consequences for macroscopic properties are each highly sensitive to the local bonding topology, stoichiometry, external conditions (pressure, temperature), and quantum effects.

1. Phenomenology and Definition

Superionic hydrogen diffusion is characterized by the highly mobile, effectively delocalized motion of hydrogen ions (usually H⁺ or protons) within a robust, typically immobile framework. In this state, the mean-squared displacement (MSD) of hydrogen grows linearly with time,

$\text{MSD}(t) \sim 2dDt,$

where $D$ is the self-diffusion coefficient and $d$ the spatial dimension. This is in sharp contrast to the finite, plateaued MSD of a bound solid or the joint diffusive motion of all atoms in a melt. The heavy atom sublattice (O, Fe, metals, etc.) remains essentially static over timescales relevant to hydrogen diffusion.

Superionicity is frequently identified operationally via a marked increase in $D$ (6–10 orders of magnitude above the solid state) and a corresponding jump in ionic conductivity, typically

$D > 10^{-6}~\text{cm}^2/\text{s},\qquad \sigma > 1~\text{S}/\text{cm}$

at high $P$ and $T$ , though the precise electronic, vibrational, and thermodynamic boundaries are compound-specific. The transition is generally described by either a sharp or continuous temperature onset $T_{\text{SI}}$ which is in some materials distinct from melting. The majority of prototypes (ices, hydrides, polyhydrides, complex oxides) show Arrhenius or super-Arrhenius temperature dependence of $D$ ; quantum zero-point and tunneling corrections are frequently critical.

2. Structural and Bonding Prerequisites

The onset and nature of hydrogen superionic diffusion is dictated by the host lattice topology and the character of the hydrogen sublattice:

Hydrogen-rich lattices and open frameworks: High proton fractions and the presence of quasi-molecular units or large, interconnected interstitial sites support percolating diffusion channels. In compounds such as $LaH_{10-x}$ , hydrogen densely occupies a three-dimensional clathrate-like framework formed by 32f cage sites, yielding a high-connectivity H network (Zhou et al., 24 Aug 2024).
Layered and one-dimensional motifs: In $Pmn2_1$ - $H_4O$ , the H₃⁺ clusters and –H–O–H–O– chains yield pronounced anisotropy. Only when weakly bound H₃ cluster rotations become mobile (onset at ~900 K) does interlayer hopping become activated, leading to a sharp, one-dimensional superionic state (960–1000 K) prior to the destruction of in-plane H–O covalency and transition to three-dimensional conduction above 1000 K (Liang et al., 17 Mar 2025).
Rigidity and flexibility of heavy atom sublattices: Close-packed phases (e.g., fcc, hcp H₂O ices) support isotropic diffusion via uniform octahedral and tetrahedral pathways, whereas lability in the backbone (as in Pa $\overline{3}$ H₂O₂) can further enhance $D$ due to larger voids and reduced migration barriers (Militzer et al., 2018).
Hydrogen bonding and network flexibility: In hydrous silica and nanoconfined superionic water, short O–O distances ( $\lesssim2.7~\text{\AA}$ ) and less than fourfold hydrogen bonding per molecule (yielding "dangling" bonds) lead to significant reduction in the free energy barrier for Grotthuss-like proton transfer (Coles et al., 20 May 2025).
Stoichiometric and compositional control: The proton fraction ( $\varphi$ ) and elemental makeup (e.g., ammonium polyhydrides $NH_x$ at $x=7-24$ ) modulate both the onset temperature for superionicity and the character of the transition. Above $\varphi\approx0.97$ in such systems, direct melting predominates over a well-resolved superionic window (Villa et al., 26 Nov 2025).

3. Dynamical Regimes, Diffusion Laws, and Activation Barriers

Hydrogen motion in the superionic regime is quantified via the self-diffusion coefficient $D$ , generically determined from the long-time slope of the MSD. The temperature and pressure dependence of $D$ is typically Arrhenius,

$D(T,P) = D_0(P)\,\exp\left(-\frac{E_a(P)}{k_BT}\right),$

with $D_0$ and $E_a$ extracted from MD or NMR fits. Reported values (see Table 1) span $D_0 \sim 10^{-7}$ to $10^{-5}~\text{m}^2/\text{s}$ and $E_a \sim 0.1$ – $1.2~\text{eV}$ across different high-pressure compounds (Militzer et al., 2018, Li et al., 24 Feb 2024, Zhou et al., 24 Aug 2024, He et al., 2018). For superionic water at $100~\text{GPa}$ , $E_a \sim 0.3~\text{eV}$ ; in $LaH_{10-x}$ the corresponding value is $0.12~\text{eV}$ (Zhou et al., 24 Aug 2024).

Compound/Phase	$D$ (m²/s)	$T_{SI}$ (K)	$E_a$ (eV)	Reference
$Pmn2_1$ - $H_4O$	$0.9$– $1.1×10^{-10}$	960–1000	≈0.1	(Liang et al., 17 Mar 2025)
fcc–H₂O	$1.2×10^{-7}$	4000	0.32	(Militzer et al., 2018)
$LaH_{10-x}$	$1×10^{-6}$	300	0.12	(Zhou et al., 24 Aug 2024)
LiH₂	$5×10^{-7}$ (at 1400 K)	900–1430	1.41	(Li et al., 24 Feb 2024)
NH₉–P2₁/m	$10^{-6}$ – $10^{-5}$	400–600	0.35–0.60	(Villa et al., 26 Nov 2025)

Non-Arrhenius ("super-Arrhenius") behavior is reported in chemically heterogeneous hosts such as multi-principal-element alloys (MPEAs) (Shuang et al., 22 Sep 2024). In these, H diffusion is best fit by the Vogel–Fulcher–Tammann law,

$D(T) = D_0\,\exp\left[-\frac{Q_{\mathrm{VFT}}}{k_B(T-T_0)}\right],$

where $T_0$ is a composition-dependent Vogel temperature, with physical origin in the presence of deep trapping sites and a configurationally rugged energy landscape. This regime is ubiquitous in glassy superionic transport.

4. Microscopic Diffusion Pathways

Multiple atomic-scale mechanisms for proton mobility are realized in superionic materials:

Hopping between interstitial sites: In high-symmetry clathrates and open frameworks, H hops between equivalent cage centers or channel sites (e.g., 32f → 32f in $LaH_{10-x}$ , or along the chains in $Pmn2_1$ - $H_4O$ and LiH₂) with migration barriers dictated by geometry, local bonds, and quantum zero-point correction.
Grotthuss-like transfer and paddle-wheel exchange: In hydrogen-bonded and molecular materials, correlated proton transfer occurs via sequential breaking and reforming of hydrogen bonds, as in nanoconfined molecular superionic water and ammonium polyhydrides. These processes are enabled by flexible networks and reduced rotational barriers (Coles et al., 20 May 2025, Villa et al., 26 Nov 2025).
Cluster rotations and emergent magnetism: In $Pmn2_1$ - $H_4O$ , rotation of triangular H₃⁺ ions (above 900 K) produces instantaneous, albeit randomly directed, ionic magnetic moments of order $10^{-26}~\text{A}\,\text{m}^2$ ( $0.001\,\mu_B$ ), a unique feature of cluster-based diffusion (Liang et al., 17 Mar 2025).
Direct hopping and ion exchange: In defective or vacancy-rich systems (e.g., pyrite-FeO₂Hₓ), migration proceeds by vacancy-mediated hopping among O-coordinated sites, traversing a three-dimensional network with barriers of 0.7–1.2 eV (He et al., 2018).

5. Materials Classes and Experimental Evidence

Hydrogen superionic diffusion has been observed or predicted in numerous systems:

High-pressure ices: $H_2O$ , $H_2O_2$ , and $H_9O_4$ exhibit superionic phases at multi-megabar pressures and $T>2000$ K, with $D$ strongly dependent on lattice type and hydrogen sublattice topology (Militzer et al., 2018, Cheng et al., 2021).
Hydrous minerals: Earth’s lower mantle hosts phases (py-FeO₂H, δ-AlOOH) in which pressure-induced O–H–O symmetrization triggers superionic proton mobility at geophysically relevant $P$ - $T$ (He et al., 2018).
Metal hydrides and superhydrides: Room temperature superionic hydrogen diffusion is experimentally confirmed in $LaH_{10–x}$ at >160 GPa, supported by NMR and ab initio MD, with $D\sim 10^{-6}$ cm²/s (Zhou et al., 24 Aug 2024).
Ammonium polyhydrides: DFT-MD shows a superionic window (350–700 K at 100–300 GPa) in $NH_x$ compounds, with transition temperatures and activation energies systematically decreasing as the hydrogen fraction increases (Villa et al., 26 Nov 2025).
Nanoconfined superionics: Machine-learning MD and DFT reveal molecular superionic water in $\sim$ 5 Å graphene slits at 400–600 K, 1–12 GPa—well below bulk superionic transitions—where rapid Grotthuss proton chains yield $D\sim10^{-9}$ m²/s (Coles et al., 20 May 2025).
Artificial architectures: Acid/WO₃/ITO “job-share” sandwich structures exhibit room-temperature hydrogen diffusion up to $1.6\times10^{-1}~\text{cm}^2/\text{s}$ (~ $10^6$ times bulk WO₃) by spatially decoupling ion and electron transport, with an effective activation barrier $E_a\sim0.1$ –$0.2$ eV (Li et al., 2022).
Multi-principal element alloys (MPEAs): ML-driven kinetic Monte Carlo studies on BCC Mo–Nb–Ta–W reveal a glassy, super-Arrhenius H-diffusion regime. Analytical expressions relate $D(T)$ to composition, local barriers, solution energy spectra, and SRO (Shuang et al., 22 Sep 2024).

6. Macroscopic Consequences and Applications

The presence of superionic hydrogen imparts dramatic macroscopic properties:

Ionic conductivities: Ionic conductivities can reach $10^3$ – $10^4$ S/m in superionic water at 100 GPa and 3000 K, with lower (but still high) values in hydrides and complex oxides (Cheng et al., 2021).
Elastic and transport softening: The superionic transition is associated with pronounced softening of elastic moduli (bulk, shear, Young’s), and thus longitudinal and transverse sound velocities, as observed in LiH₂ and related premelting softening phenomena (Li et al., 24 Feb 2024).
Electromagnetic effects: Rotating hydrogen clusters and highly anisotropic conduction in layered hydrous ices (notably $Pmn2_1$ - $H_4O$ ) are theoretically linked to the generation of planetary magnetic fields, providing an explanation for the non-dipolar, nonaxisymmetric fields of Uranus and Neptune (Liang et al., 17 Mar 2025).
Phase metastability: Rapid hydrogen diffusion and one-dimensional hopping drive transient phases and time-dependent decomposition, as observed in metal superhydrides ( $LaH_{10-x}$ ), where hydrogen loss over weeks quenches high- $T_c$ superconductivity (Zhou et al., 24 Aug 2024).
Energy technology: Interface-engineered oxide devices utilize ultrafast H transport for solid-state electrochemistry, batteries, fuel cells, and neuromorphic electronics (Li et al., 2022). Nanoconfined molecular superionics are proposed as proton-conducting membranes at moderate $T$ (Coles et al., 20 May 2025).

7. Methodological Approaches and Theoretical Models

First-principles DFT-MD, born–Oppenheimer MD, and path-integral MD dominate the ab initio paper of hydrogen superionic diffusion. The Einstein relation for self-diffusion ( $D$ from MSD), Arrhenius/VFT/fractional models for temperature dependence, and analysis of local hopping pathways via bond valence sums, NEB calculations, and machine learned potentials provide comprehensive characterization (Militzer et al., 2018, Zhou et al., 24 Aug 2024, Shuang et al., 22 Sep 2024, Coles et al., 20 May 2025).

Key equations and concepts include:

Einstein relation: $D = \lim_{t\to\infty}\frac{1}{6t}\langle|r(t)-r(0)|^2\rangle$ .
Arrhenius law: $D(T) = D_0\exp(-E_a/k_BT)$ .
Vogel–Fulcher–Tammann law: $D(T) = D_0\exp[-Q_{\rm VFT}/(k_B(T-T_0))]$ .
Nernst–Einstein relation (linking $D$ to ionic conductivity): $\sigma = n q^2 D/(k_BT)$ .

Advanced methodologies, such as hybrid machine learning potentials (NNP, MLFF), integration with neural network KMC, and symbolic regression, permit compositional dependency mapping and identification of super-Arrhenius transport regimes (Shuang et al., 22 Sep 2024, Coles et al., 20 May 2025).

Hydrogen superionic diffusion is now established as a universal, structurally programmable phenomenon, with implications for planetary field generation, energy transport, elastic softening, and materials design for ionic conductors. Its realization depends critically on lattice topology, composition, and quantum coherence; models capturing these aspects have matured into predictive frameworks applicable to both geophysical phenomena and engineered functionality. The interplay of structure, dimensionality, disorder, and thermodynamics continues to drive the search for new superionic materials and innovative functional devices.