Quantifying Multidimensional Transport Effects on Permeability Inference in FLiBe Systems Using a Validation-Informed Modeling Framework

This study employs a validation-informed, multi-dimensional modeling framework to demonstrate that relying on simplified one-dimensional interpretations of permeation experiments can lead to inaccurate inference of hydrogen isotope permeability in FLiBe systems due to significant multidomain transport effects and boundary condition sensitivities.

The Gist

Imagine you are trying to measure how fast water leaks through a specific type of sponge. You set up a simple experiment: you pour water on one side of the sponge and measure how much comes out the other side. In a perfect world, you could just do the math, and you'd know exactly how "leaky" that sponge is.

But in the real world, things are messier. What if the water also sneaks out through the sides of the bucket holding the sponge? Or what if the bucket itself is made of a material that soaks up some water and leaks it elsewhere? If you ignore those side paths and just look at the water coming out the bottom, your calculation of the sponge's leakiness will be wrong.

This paper is about doing exactly that kind of "messy" math for molten salt used in future fusion power plants. Specifically, they are studying FLiBe, a special hot liquid salt, and how hydrogen isotopes (like tritium, a fuel for fusion) move through it.

Here is the breakdown of their findings using simple analogies:

The Problem: The "One-Dimensional" Trap

Scientists often try to figure out how fast hydrogen moves through FLiBe by using a 1D model. Think of this like measuring traffic flow on a straight, single-lane road. You assume cars only go forward.

However, in the real experiment (called HYPERION at MIT), the setup is more like a busy city intersection. The hydrogen doesn't just go straight through the salt and a metal wall; it also:

1. Sneaks around the sides: It travels through the metal walls of the container.
2. Leaks out the back: It escapes into the surrounding room (the glovebox) if the container isn't perfectly sealed.

If you use the "straight road" (1D) math to analyze data from this "city intersection," your answer for how "leaky" the salt is will be completely off.

The Experiment: The "Leaky Bucket"

The researchers built a test rig with:

• Hot FLiBe salt on one side.
• A Nickel metal wall in the middle.
• A gas collection area on the other side.

They wanted to see how fast hydrogen moved from the salt, through the nickel, to the gas collector. But they realized the nickel container itself was acting like a second, hidden highway for the hydrogen.

The Solution: A "3D Detective" Approach

Instead of using the simple "straight road" math, they used a powerful computer simulation (called FESTIM) that acts like a 3D detective. It tracks every single hydrogen atom, whether it's going straight through the salt, sneaking through the side walls, or leaking out into the room.

They tested two extreme scenarios for the outside of the container:

1. The "Perfect Seal" (Ideal Coating): Imagine the outside of the bucket is wrapped in a magical, impermeable tape. Nothing can escape the sides.
2. The "Open Bucket" (Uncoated): Imagine the bucket is bare metal, and hydrogen can easily leak out into the room.

The Big Discoveries

1. The "Sidewall Highway" is Real and Huge
The computer model showed that the side walls of the container are not just passive containers; they are active highways.

• In the "Perfect Seal" scenario: The side walls actually helped the hydrogen get to the detector faster by providing a bypass route around the salt. It was like a shortcut.
• In the "Open Bucket" scenario: The side walls acted like a drain, sucking the hydrogen away before it could reach the detector. It was like a leaky pipe.

2. The "Leakiness" Number Changes Drastically
Because the side walls change the flow so much, the number they calculated for how "leaky" the FLiBe salt is changed by more than 10 times (an order of magnitude) depending on which scenario they assumed!

• If they assumed the bucket was perfectly sealed, the salt looked less leaky.
• If they assumed the bucket was open, the salt looked more leaky.

3. The Old Math Was Wrong
When they compared their new 3D detective method to the old 1D "straight road" math:

• The old math underestimated the flow when the bucket was sealed (because it missed the side shortcuts).
• The old math overestimated the flow when the bucket was open (because it missed the side leaks).

The Takeaway

The main point of this paper is: You cannot accurately measure how a material behaves if you ignore the shape of the container and the environment around it.

If you want to know the true "leakiness" of FLiBe salt for fusion power plants, you can't just use a simple formula. You have to build a complex, 3D model that accounts for every possible path the hydrogen can take, including the sneaky side routes and the leaks to the outside world.

The authors aren't saying the salt is definitely more or less leaky than we thought; they are saying that previous studies might have been measuring the "leakiness of the whole experiment" rather than just the "leakiness of the salt." To get the real answer, we need to stop using simple 1D maps and start using detailed 3D GPS tracking for the hydrogen atoms.

Technical Summary

Technical Summary: Quantifying Multidimensional Transport Effects on Permeability Inference in FLiBe Systems

Problem Statement
Accurate prediction of hydrogen isotope transport in molten salt fusion blankets, particularly those using FLiBe, relies on reliable permeability data. However, reported permeability values for FLiBe exhibit significant variability (spanning one to two orders of magnitude), often attributed to material factors like redox state and impurities. This work identifies a critical, often overlooked source of discrepancy: the interpretation of permeation experiments themselves. Conventional analyses typically employ simplified one-dimensional (1D) transport models that assume transport occurs along a single effective pathway under idealized boundary conditions. In realistic experimental systems, such as the HYPERION facility at MIT, hydrogen isotopes undergo coupled transport across gas, liquid, and solid domains. These systems feature parallel pathways, including radial diffusion through structural sidewalls and exchange with the external environment, which 1D models cannot represent. Consequently, permeability values inferred from 1D models may reflect system-dependent parameters (geometry and boundary conditions) rather than intrinsic material properties.

Methodology
To address these limitations, the authors developed a multidimensional, multi-material transport modeling framework using the FESTIM code, benchmarked against experimental data from the HYPERION facility. The experimental configuration involves a molten FLiBe region separated from a downstream gas measurement region by a nickel (Ni-200) membrane, all contained within a nickel vessel.

The methodology involves several key components:

1. Multidimensional Modeling: The model explicitly resolves transport across the molten salt, nickel membrane, and surrounding nickel containment structure using an axisymmetric (r-z) formulation. It accounts for dissolution, diffusion, and thermodynamic partitioning at interfaces (Henry's law in FLiBe, Sieverts' law in nickel).
2. Validation-Informed Inverse Framework: Instead of assuming a permeability value, the study employs an inverse procedure. The FLiBe permeability is iteratively adjusted until the simulated steady-state flux matches the experimentally measured flux. This ensures the extracted value reflects the coupled system response.
3. Boundary Condition Envelope: Recognizing the uncertainty in external vessel surface performance (specifically protective coatings), the study defines a physically motivated envelope using two limiting boundary conditions:
   • Ideal Coating: A zero-flux condition suppressing hydrogen transport across the outer vessel wall.
   • Uncoated Vessel: A condition where the outer surface is in local thermodynamic equilibrium with the glovebox atmosphere, allowing for hydrogen exchange.
4. Uncertainty Quantification: The framework incorporates measurement uncertainties (Type A and Type B) and propagates them to the inferred permeability. It also accounts for temperature-dependent geometry changes (salt thickness) to isolate material properties from geometric variations.

Key Results
The study analyzed hydrogen and deuterium permeation data across a temperature range of 773 K to 973 K. The results highlight the profound impact of multidimensional transport:

• Sidewall Transport Significance: The nickel sidewalls act as significant transport pathways. Under ideal coating conditions, sidewall bypass pathways augment the downstream flux by 10–17% for hydrogen and 25–27% for deuterium. Under uncoated conditions, sidewalls act as a dominant loss mechanism to the environment, reducing the downstream flux. At 973 K, sidewall losses exceed the flux reaching the detector by factors of ~2.3 (hydrogen) and ~2.7 (deuterium).
• Permeability Inference Range: The inferred FLiBe permeability spans more than an order of magnitude depending on the assumed boundary condition. The "uncoated" assumption yields permeability values systematically higher (by >10x) than the "ideal coating" assumption. Both inferred ranges exhibit Arrhenius behavior but differ significantly in pre-exponential factors and activation energies.
• Bias in 1D Models: Comparisons between 1D and 2D models reveal that 1D formulations produce qualitatively opposite errors depending on the boundary condition. Under ideal coating, 1D models underpredict flux (leading to an artificial overestimation of permeability if used for inference). Under uncoated conditions, 1D models overpredict flux (leading to an underestimation of permeability).
• Literature Comparison: The inferred permeability values generally lie in the lower portion of reported literature correlations. The study suggests that the scatter in existing literature (often 1–2 orders of magnitude) may be partially attributable to differences in experimental geometry and boundary conditions rather than solely intrinsic material variability.

Significance and Claims
The paper claims to provide a "physically grounded framework" for interpreting permeation measurements in coupled liquid-metal systems. Its primary contribution is demonstrating that using one-dimensional formulations to describe permeation experiments in systems with structural transport pathways is inadequate for extracting accurate, transferable material properties.

The authors assert that:

1. System-Level Interpretation is Necessary: Measured permeation fluxes represent a coupled system response. Ignoring multidomain transport and external boundary assumptions leads to permeability values that are specific to the experimental geometry rather than intrinsic to the material.
2. Multidimensional Modeling is Essential: To reliably infer intrinsic transport properties for use in reactor-scale blanket simulations, multidomain transport modeling must be used to separate geometric and boundary-condition effects from the material response.
3. Re-evaluation of Literature Data: The observed variability in FLiBe permeability data may stem significantly from how different experimental configurations were analyzed (i.e., the treatment of sidewall losses and boundary conditions) rather than just material inconsistencies.

The study concludes that future permeation experiments and their analysis should be co-designed with multidimensional modeling capabilities to reduce systematic uncertainty and enable the reliable extraction of intrinsic transport properties for fusion-relevant molten-salt environments.