Permeation behaviour of hydrogen isotopes in molten FLiBe (2LiF-BeF2): Identifying sources of uncertainty and associated measurement challenges

Abstract: This paper presents results from systematic investigations conducted in the HYPERION facility to quantify permeabilities of hydrogen isotopes in FLiBe over a temperature range of 773K - 973K. To address the knowledge gap resulting from widely scattered transport parameters reported in the earlier studies, HYPERION experiments incorporate specific provisions to probe unsubstantiated assumptions, including a one-dimensional permeation, membrane surface coverage by the salt, and ideal wettability of the salt-metal interface. Suppression of the isotopic transport behaviour for metal-side charging highlights the permeation barrier characteristics of a bubble-laden Ni-FLiBe interface, impacting permeability by up to 77%. This work attempts to resolve this issue via salt-side charging and informs constraints on the permeant charging methodology for future studies. The initial insights from HYPERION experiments call into question the design choices and estimated transport properties in earlier works. This study provides a plausible explanation for the observed scatter in H/D/T transport in FLiBe.

• The paper demonstrates that interface bubble nucleation suppresses hydrogen isotope permeation in molten FLiBe, reconciling discrepancies in prior literature.
• Experiments using a Ni-200 cell with controlled salt- and metal-side charging yield permeation data with uncertainties within ±13% and validate Arrhenius behavior.
• Findings inform fusion blanket design by revealing how salt coverage and interfacial phenomena impact tritium transport and flux measurements.

Permeation Behavior of Hydrogen Isotopes in Molten FLiBe: Systematic Characterization and Interface Effects

Introduction

The quantification of hydrogen isotope permeation in molten FLiBe (2LiF-BeF$_2$) is critical for the modeling and operation of fusion breeding blankets, where accurate tritium management underpins fuel cycle closure and self-sufficiency. Prior studies report permeation data with significant scatter and unexplained variability, as well as non-trivial uncertainties arising from methodological and system-specific factors. This work presents controlled experiments in the HYPERION facility, designed to probe and explicitly characterize the underlying mechanisms and uncertainties that influence hydrogen/deuterium transport in FLiBe over $773$–$973$ K.

Experimental Framework and Methodological Rigor

Experiments utilized a custom-fabricated Ni-200 permeation cell with a thermal-sprayed Al$_2$O$_3$ coating, operated in a rigorously purged inert (Ar) glovebox environment to suppress extrinsic impurity and background effects. The dual-volume system enables both metal-side and salt-side isotope charging, allowing decoupling of transport contributions from Ni, interfacial phenomena, and bulk FLiBe. Real-time gas chromatography with sub-ppm detection and precise flow/pressure control, together with periodic visual inspection of the salt-membrane interface, provide a high degree of quantification for both transport flux and system state.

Comprehensive impurity analysis of the FLiBe (ICP-MS) reveals cationic and transition metal content meeting or exceeding MSRE and contemporary operational thresholds, providing fidelity in referencing results to “clean” and “impure” salt regimes. Explicit calculation accounts for radial and axial fluxes, using configurations with bounding permeation reduction factors (PRFs) for the Al$_2$O$_3$ barrier, ensuring all inferred transport coefficients are systematically bracketed.

Key Findings: Permeability, Isotope Effects, and Interface Limitations

Validation and Baseline Permeabilities

Dry runs using only the Ni-200 membrane demonstrate excellent agreement ($<$3% deviation) with the literature for H$_2$/D$_2$ permeabilities. Arrhenius behavior, uncertainty budgets (within $773$0), and isotopic separation ($773$1factor 1.4–1.6) calibrate expectations for the system and validate the analytic approach.

Impact of Salt Coverage and Interface Bubbles

The transition from partial to complete FLiBe coverage on the Ni membrane induces a two-order-of-magnitude drop in the permeation flux, consistent with effective separation of metal and salt transport characteristics. Notably, during metal-side Q$773$2 charging, transient oscillations are observed in the permeation signal; these are directly correlated with spontaneous or forced bubble release events at the Ni-FLiBe interface. The accumulation of H$773$3 at the interface, precipitated by drastically lower bulk FLiBe permeability, drives supersaturation and interfacial bubble nucleation. These stochastic bubble events introduce apparent “jumps” in measured flux unrelated to steady-state diffusive transport, elucidating the origin of overestimated permeabilities in prior literature lacking real-time flux monitoring and interface state visualization.

Isotope Masking and Interface-Limited Transport

Direct comparison of H$773$4 versus D$773$5 permeation under interface-limited, bubble-forming conditions reveals substantial isotopic masking: fluxes converge regardless of isotope, and steady-state flux is governed primarily by bubble formation/detachment kinetics rather than by intrinsic solubility or diffusivity. In contrast, under interface-optimized (bubble-suppressed) conditions achievable with salt-side charging, the expected isotopic separation is restored, and the fluxes scale as predicted by classical theory ($773$6factor 2.3–2.4, activation energy difference $773$77 kJ/mol).

Salt-Side Charging and Wetting Interventions

Reversing the charging configuration to the low permeance (FLiBe) side suppresses interfacial bubble nucleation, as evidenced by “restored” isotopic and Arrhenius behavior in permeability measurements. Shaking-induced enhancements in interfacial wetting further increase observed fluxes by 11–36%, reinforcing the inference that interface imperfections act as rate-limiting traps.

Across $773$8–$773$9 K, interface bubble effects decrease observable permeability by up to 77% in the metal-side configuration, with diminished suppression at elevated temperatures, suggesting temperature-activated interfacial mechanisms. At $973$0 K, bubble dynamics (growth, detachment, and perhaps natural convection) complicate flux interpretation, decreasing confidence in permeability estimations and requiring careful experimental design to isolate transport phenomena of interest.

Comparative Analysis, System Design Choices, and Literature Discrepancies

The HYPERION results clarify several previously unexplained features in the literature:

• Large Discrepancies in Reported Transport Coefficients: At a given T, permeabilities reported by prior studies differ by factors up to 10, a gap directly attributable to error sources such as partial salt coverage, inadequate interface monitoring, impurity-induced wettability variations, and unaccounted-for radial losses.
• Radial versus Axial Configurations: Radial permeation experiments, especially with multi-layer tube arrangements (e.g., Monel-400 in Nakamura et al.), are systematically shown to dampen or mask flux oscillations through secondary interfaces and unpurged “dead” volumes, which, coupled with ambiguous accounting for permeation area and salt coverage, lead to overestimated permeabilities.
• Effect of Isotopic Mixture and Downstream Q$973$1 Partial Pressure: In scenarios with simultaneous multi-isotope presence, interfacial isotopic exchange accelerates downstream transport, further complicating attributions of measured flux to intrinsic properties.

Arrhenius Quantification and Intrinsic Transport Parameters

FLiBe permeabilities for H$973$2 (salt-side charging, $973$3–$973$4 K): $973$5 (PRF=1): $973$6 (H$973$7/m·s·Pa)

$973$8 (PRF=$973$9): $_2$0

For D$_2$1 (up to $_2$2 K): $_2$3 (PRF=1): $_2$4 (D$_2$5/m·s·Pa)

The H/D permeability ratio is consistently $_2$62.3–2.4 in this window, and activation energy differences match theoretical expectations, validating that intrinsic transport mechanisms are preserved only when bubble and interface effects are explicitly suppressed or accounted for.

Implications for Fusion Blanket Operations and Modeling

Quantitative understanding of hydrogen isotope transport in FLiBe is essential for tritium inventory management, fuel processing, and safety analyses in fusion blankets. Interface-induced permeation suppression linked to bubble nucleation alters both steady-state and dynamic transport, challenging previously held assumptions of one-dimensional, series-coupled diffusion. These findings necessitate reconsideration of “intrinsic” transport coefficients used in fusion system models, inform practical engineering of blanket and tritium extraction systems, and provide experimentally grounded constraints on permissible impurity levels and surface condition requirements to manage interface phenomena.

Outlook and Future Directions in Hydrogen Transport Studies

The authors propose several paths for further investigation:

• Direct visualization and in situ monitoring of interface bubble nucleation and growth,
• Systematic parametric studies of surface roughness, salt chemistry, redox control, and operational configuration to decouple and isolate interface versus bulk effects,
• Expansion to multi-dimensional modeling (2D/3D) using validated platforms such as FESTIM,
• Rigorous extension to tritium transport and quantification of isotope dependence for fuel cycle database enrichment.

Conclusion

The HYPERION study systematically deconvolutes the contributions of interfacial phenomena—particularly bubble nucleation and growth—to hydrogen isotope permeation in molten FLiBe, providing direct evidence for interface-controlled suppression and isotopic masking under non-ideal conditions. The results reconcile decades of scattered literature data, identify explicit design and procedural factors responsible for experimental variability, and yield robust, configuration-dependent Arrhenius parameters for hydrogen and deuterium transport. This work establishes a new technical standard for future permeation studies in molten salts relevant to fusion energy systems, with direct implications for breeder blanket design, model validation, and safety assurance (2603.28991).