Thermal modeling of runaway electron induced damage in the SPARC tokamak

This paper presents the first systematic thermal analysis of tungsten-based plasma-facing components in the SPARC tokamak, evaluating damage characteristics like melt depth and vaporization losses caused by runaway electron beams during vertical displacement events.

The Gist

Imagine a fusion reactor, like the SPARC tokamak described in this paper, as a giant, super-hot star trapped inside a magnetic cage. The goal is to squeeze this star to create clean energy. But sometimes, the cage wobbles, and the star gets angry.

When things go wrong during a "disruption" (a sudden crash of the plasma), a tiny fraction of the electrons inside the star get kicked out of the magnetic cage. These aren't just normal electrons; they become Runaway Electrons (REs). Think of them as a swarm of microscopic, super-speedy bullets, each carrying the energy of a speeding car, all heading straight for the reactor's walls.

This paper is a "damage report" written by scientists trying to figure out what happens when these electron bullets hit the reactor's armor.

The Armor: The Tiles

The inside of the reactor is lined with special tiles made mostly of Tungsten. You can think of Tungsten as the "titanium" of the nuclear world—it's incredibly hard and has a very high melting point. However, even the toughest armor can be beaten if hit hard enough.

The scientists modeled a specific part of the armor: the limiters on the side of the reactor. These are like the bumpers on a car, designed to take the hit so the rest of the engine doesn't get destroyed.

The Simulation: A Three-Step Detective Story

To predict the damage, the team didn't just guess; they built a digital laboratory using a three-step process:

1. The "What If" Scenarios (Dream Code): First, they used a super-computer program called Dream to simulate a reactor crash. They asked: "If the reactor crashes, how many bullets are there? How fast are they going? Are they flying straight or at an angle?" They created two types of "bullet storms": one with a steady stream of bullets and another with a chaotic mix of speeds and angles.
2. The "Bullet Impact" (Geant4): Next, they used a program called Geant4 (which is like a virtual billiards table for subatomic particles) to see exactly how these bullets hit the Tungsten armor. They discovered that the armor isn't flat; it's curved. Just like rain hitting a curved roof, the bullets don't hit evenly. Some parts get soaked, while others stay dry. Also, when a bullet hits, it doesn't just stop; it bounces off or creates a shower of smaller particles. The simulation had to account for these bounces to know exactly where the heat lands.
3. The "Heat Check" (MEMENTO): Finally, they used a program called MEMENTO to calculate the heat. They asked: "If this much energy hits this spot for 1 millisecond (a blink of an eye) or 10 milliseconds, how hot does it get? Does it melt? Does it boil away?"

The Big Findings: It's Not Just About Speed

The scientists found some surprising things that you can't guess just by looking at the speed of the bullets:

• The "Curved Roof" Effect: Because the tiles are curved, the bullets don't hit the "front" of the tile the way you'd expect. The shape of the tile and the magnetic field lines cause the energy to spread out in weird ways. A bullet hitting at a shallow angle might actually deposit more heat deep inside the tile than a bullet hitting straight on.
• The "Explosion" Risk: When the heat is intense enough, the surface of the Tungsten doesn't just melt; it can violently explode. Imagine a pot of water boiling so fast that the steam blows the lid off. In the reactor, this "explosion" throws hot debris everywhere, which is bad news for the machine. The paper found that for very fast bullets (50 MeV), the heat gets trapped under the surface, creating a pressure cooker effect that leads to these explosions.
• Time Matters: If the bullets hit in a split second (1 ms), the heat is so concentrated it causes massive melting. If they hit over a slightly longer time (10 ms), the heat has time to spread out, but the surface might boil away (vaporize) instead of melting. It's the difference between a laser cutting through metal (fast, deep cut) and a blowtorch (slower, wider burn).
• The "Tail" of the Storm: The most dangerous part of the "bullet storm" isn't the average speed, but the few super-fast bullets at the very end of the distribution. These high-speed outliers can penetrate deep into the armor and change where the damage happens, shifting the "hot spot" to a different part of the tile.

The Bottom Line

This paper is a warning label for the future of fusion energy. It tells us that if a runaway electron event happens in the SPARC reactor, the damage won't be uniform. It will be a complex, chaotic mix of melting, boiling, and exploding, heavily influenced by the shape of the tiles and the speed of the electrons.

The scientists are essentially saying: "We know the armor is strong, but if the 'bullets' are fast enough and hit the right angle, they can punch right through or blow the armor apart."

This research is crucial because it helps engineers design better armor and safety systems for future fusion power plants, ensuring that when we finally harness the power of the stars, the machine doesn't blow itself up in the process.

Technical Summary

Here is a detailed technical summary of the paper "Thermal modeling of runaway electron induced damage in the SPARC tokamak."

1. Problem Statement

Runaway electrons (REs) represent a critical threat to the integrity of plasma-facing components (PFCs) in high-performance tokamaks like SPARC and ITER. During unmitigated disruptions, a significant portion of the plasma current can convert into REs, which strike the first wall with high kinetic energies. This impact causes:

• Deep volumetric melting and potential thermal-stress-driven explosions (debris release).
• Elevated temperatures at the bond interface between the PFC and the cooling substrate, risking coolant leaks.
• Material erosion via vaporization.

While previous studies focused on graphite or simplified tungsten (W) models, this paper addresses the specific challenge of modeling RE-induced damage on realistic, tungsten-heavy alloy (WHA) tiles in the SPARC tokamak (a compact, high-field device with $I_P = 8.7$ MA). The study focuses on the outboard off-midplane limiters, which are the primary contact point during a Vertical Displacement Event (VDE).

2. Methodology

The authors employed a three-stage, one-way coupled workflow to simulate the thermal response:

A. RE Loading and Distribution (Input)

Two distinct approaches were used to define the RE beam characteristics:

1. Parametric Scans: Simplified inputs assuming mono-energetic beams (0.5, 1, 10, and 50 MeV) with pitch angles of 0°, 5°, and 10°. Total energies ranged from 20 kJ to 100 kJ over loading times of 1 ms and 10 ms.
2. DREAM Simulations: Realistic distribution functions generated using the Dream fluid-kinetic code. Two avalanche models were tested:
   • Hesslow fluid source term: Yields an average energy of ~12 MeV.
   • Rosenbluth-Putvinski kinetic source term: Yields an average energy of ~17.5 MeV.
   • Both scenarios assumed a total RE energy of 100 kJ distributed over ~13–14 tiles.

B. Energy Deposition (Geant4)

Monte Carlo simulations using Geant4 calculated the volumetric energy deposition:

• Geometry: A realistic, curved panel design (notional but high-fidelity) was used to capture magnetic field inclination ($\alpha$) and shadowing effects.
• Material: Tungsten Heavy Alloy (97% W, 2% Ni, 1% Fe). Simulations assumed pure W properties for interaction cross-sections, as the 3% alloying elements had negligible impact on energy deposition profiles.
• Two-step approach:
  1. Full panel simulation: Moderate statistics to capture global geometry and backscattering effects.
  2. Simplified slab simulation: High statistics (up to $10^7$particles) with fine mesh resolution (20$\mu$m near surface) to generate accurate depth-dependent energy maps.
• Loss Channels: The simulations tracked energy losses via electron backscattering (EBS), positron backscattering (PBS), photon transmission (bremsstrahlung), and neutron production.

C. Thermal Response (MEMENTO)

The MEMENTO code performed heat transfer simulations:

• Input: Volumetric heat sources derived from Geant4 maps.
• Material Properties: Temperature-dependent specific heat ($c_p$) and thermal conductivity ($k$) were constructed for WHA (solid) and pure W (liquid/vapor).
• Physics: Included non-linear effects such as vaporization cooling (moving boundary condition), thermal radiation, and latent heats.
• Output: In-depth temperature profiles, melt depth, and vaporization losses.

3. Key Contributions

• First Systematic Analysis for SPARC: This is the first study to model RE-induced thermal damage specifically for the realistic PFC geometry and magnetic topology of the SPARC tokamak.
• Realistic Geometry Integration: Unlike previous 1D or flat-surface models, this work incorporates the curved panel geometry and local magnetic field inclination, revealing that 3D effects significantly alter energy partitioning.
• Avalanche Model Comparison: The study contrasts two different RE generation models (Hesslow vs. Rosenbluth-Putvinski) to understand how the high-energy tail of the distribution affects damage localization.
• Comprehensive Loss Accounting: Detailed quantification of energy losses due to backscattering and photon/neutron transmission, showing that high-energy REs lose significant energy (>20% at 50 MeV) via bremsstrahlung.

4. Key Results

Energy Deposition and Partitioning

• Deviation from Geometric Scaling: Energy partitioning does not follow a simple $\sin(\alpha)$ scaling. Due to the "smearing out" effect of backscattered electrons and large Larmor radii at high energies, the energy distribution becomes non-local.
• Sector Damage:
  • For low energies ($\le$ 10 MeV), the sector with the steepest magnetic field inclination (Sector 4) receives the most energy and suffers the deepest melting.
  • For high energies (50 MeV), the high-energy tail deposits energy more deeply and broadly, shifting the maximum energy density to sectors with shallower angles (Sector 1 or 2).
• Pitch Angle Effect: Non-zero pitch angles redistribute energy, generally reducing peak temperatures in steep-angle sectors but increasing them in shallow-angle sectors for high-energy beams.

Thermal Response and Damage

• Melting Depths:
  • 100 kJ / 1 ms loading: 10 MeV REs cause melt depths of ~1 mm. 50 MeV REs cause ~0.5 mm.
  • Low Energy (0.5–1 MeV): These cause the most severe surface erosion due to intense vaporization. For 0.5 MeV, ~350 $\mu$m is eroded in 1 ms, with >55% of energy dissipated by vaporization.
• Temperature Profiles:
  • Non-monotonic Profiles: High-energy impacts (10–50 MeV) often result in temperature peaks beneath the surface (sub-surface melting). This is a precursor to PFC explosions and debris release.
  • Surface Loading: Low-energy impacts (0.5 MeV) result in monotonic temperature profiles with maximum heat at the surface.
• Time Scale Dependence:
  • 1 ms loading: Leads to extreme surface temperatures and vaporization-dominated erosion for low energies.
  • 10 ms loading: Allows more heat diffusion, reducing peak surface temperatures but allowing deeper melt penetration for moderate energies.
• Energy Losses:
  • Total energy loss increases with RE energy: ~5% at 0.5 MeV to >20% at 50 MeV.
  • Bremsstrahlung is the dominant loss mechanism for energies >10.5 MeV, converting electron energy into photons that escape the panel.
  • Backscattering yields decrease with increasing energy.

DREAM vs. Mono-energetic Comparison

• Simulations using realistic DREAM distributions showed melt depths 10–20% deeper in the most damaged sectors compared to mono-energetic 10 MeV cases.
• The high-energy tail in DREAM distributions significantly alters the energy partition across the curved panel, proving that single-point energy/pitch approximations are insufficient for accurate damage prediction.

5. Significance and Outlook

• Design Implications: The study highlights that SPARC limiters are susceptible to deep melting and explosive debris release, particularly under 10 MeV+ RE loads. The non-monotonic temperature profiles suggest a high risk of mechanical failure.
• Mitigation Strategy: The results indicate that mitigation strategies must account for the high-energy tail of the RE distribution, as these particles deposit energy non-locally and can damage areas not predicted by simple geometric models.
• Future Work:
  • Full thermomechanical modeling (including viscoplasticity and hydrodynamics) is required to predict the exact mechanics of PFC explosions.
  • Assessment of photon transmission (bremsstrahlung) impacts on surrounding structural materials and superconducting magnets.
  • Experimental validation using MeV-range electron beams on pure W samples.
  • Application of these findings to the ARC tokamak (SPARC's successor).

In conclusion, this paper establishes a robust, high-fidelity framework for predicting RE damage in next-generation fusion devices, demonstrating that realistic geometry and full distribution functions are essential for accurate safety assessments.