Tungsten compound tops shield material comparison for mirror fusion magnets

Hammir-DT’s central cell and mirror end assemblies illustrate the whole-device geometry ParaTAN’s neutronics model was built to capture, in contrast to the simplified cylindrical shapes earlier 1D codes assumed

(Image courtesy of Realta Fusion)

UW-Madison researchers, working with spinout company Realta Fusion, have published a whole-device neutronics assessment of magnetic mirror fusion machines that directly addresses a gap left by the simplified 1D models used since the 1970s. Led by graduate student Hitarth Shah and Associate Professor Ben Lindley, the study uses a new parametric toolkit called ParaTAN to compare magnet shielding materials and tritium breeding blanket concepts across simple and tandem mirror designs.

From 1970s programs to a modern neutronics toolkit for mirror fusion

The classical tandem mirror concept envisioned a long solenoid central cell flanked by mirror end plugs, where intense heating would drive a hot, dense plasma confined by an electrostatic potential. It promised steady-state operation with confinement times near 10 seconds and system gains above unity. Achieving the required density ratios, mirror ratios, and plasma temperatures proved technologically infeasible at the time, and reduced funding led to the discontinuation of US mirror research.

Several enabling technologies have since shifted that landscape. Advances in high-energy negative-ion neutral beam injection, high-frequency gyrotrons, high-temperature superconducting magnets, and magnetohydrodynamic stabilization of axisymmetric mirrors now let simpler planar-coil designs reach substantially higher fields and mirror ratios than earlier generations achieved. These developments have reignited interest in tandem mirrors as candidates for both fusion power plants and intense neutron sources, pushing the technology into pre-conceptual design and creating a need for modern, whole-device neutronics tools.

Mirror fusion neutronics itself has sat largely untouched since major US mirror programs wound down in the late 1970s. The last generation of assessments, including the University of Wisconsin’s own WITAMIR-I study alongside TASKA and MARS, relied on 1D transport solvers focused mainly on the central cell, using the ANISN code with a P3-S8 approximation for cylindrical geometry. Because the focus stayed on the central cell, the simplified cylindrical modelling suited its time, but whole-device analysis needs a more accurate representation of geometry and physics.

Even so, those early models still produced useful baseline figures. Central cell blankets using lead-lithium eutectic achieved tritium breeding ratios above 1.05 alongside neutron energy multiplication factors near 1.36, and projected first-wall damage rates between 15 and 17 displacements per atom per full power year.

ParaTAN replaces the earlier approach with continuous-energy Monte Carlo modelling built on OpenMC, generating whole-device geometry rather than a simplified cylindrical approximation. The tool’s Python API supports parametric sweeps and automated post-processing, letting the team iterate through shielding and blanket configurations at a pace the earlier 1D codes could not match.

Shielding and breeding trade-offs the mirror fusion study reveals

The team ran two separate assessments. The first tested four candidate shield materials, boron carbide, tungsten carbide, tungsten boride, and tungsten pentaboride, for the high-field magnets of a simple mirror device. Tungsten pentaboride came out on top, and axial shielding proved substantially more effective than radial shielding at cutting radiation loads to the superconducting magnets.

The second assessment turned to tritium breeding across four blanket concepts for a tandem mirror central cell, covering liquid lithium, PbLi eutectic, FLiBe, and a lithium orthosilicate pebble bed. The helium-cooled pebble bed with a titanium beryllide (Be₁₂Ti) multiplier delivered the strongest breeding performance, approaching self-sufficiency at natural lithium enrichment for sufficiently large blanket dimensions.

The study found the pebble bed configuration proved less desirable in other respects, leaving PbLi as a competitive alternative given its liquid form suits combined breeding and cooling.

How mirror shielding trade-offs compare to tokamaks and ITER

The shielding-versus-breeding-margin tension ParaTAN quantifies for mirrors is familiar territory in tokamak design. The ARIES-ACT-2 project performed a detailed 3D neutronic assessment in which blanket thickness, shielding material, and magnet protection were sized against a tritium breeding ratio target near 1.05, finding that a thicker inboard shield extends magnet lifetime and reduces nuclear heating but adds size, cost, and reduces breeding margin.

A similar trade-off shaped the US ITER Magnet Shield study, which evaluated combinations of stainless steel, water, boron carbide, lead, and boron steel for the toroidal field magnets. Pure stainless steel and water shields failed to meet radiation damage and heating limits, and adding boron or lead proved necessary, with assembly gaps and limited inboard radial space pushing fast neutron fluence and insulator dose higher than shield thickness alone would suggest.

Mirror devices face the same fundamental constraint, balancing shield thickness against magnet protection and breeding margin. ParaTAN’s comparison of four shield materials for mirror magnets gives the sector its own dataset specific to mirror geometry, rather than an extrapolation from tokamak results.

For a sector still short on modern mirror-specific data, the practical value here is comparative. The study gives procurement and design teams evaluating mirror concepts a like-for-like reference across shield and blanket materials, the kind of comparative dataset mirror research has lacked since the field’s 1970s programs wound down.

The work sits inside a wider commercial push. Realta Fusion, co-founded by Lindley, has raised $36 million in Series A funding and a further $9.5 million SVB growth facility, backing its plan to commercialise compact magnetic mirror systems for industrial heat and power.

Next steps for tritium transport and a lead-lithium loop

The authors plan to extend ParaTAN toward neutral beam injection shielding, the effect of neutron multipliers on liquid breeder performance, and activation and shutdown dose rate calculations. UW-Madison and Realta Fusion will also continue their collaboration through a new lead-lithium MHD loop, led by Associate Professor Juliana Pacheco Duarte under the DOE-funded FIRE collaborative, testing how the winning blanket candidates behave under real magnetic field strength rather than in simulation alone.

Stay ahead in the fusion revolution explore more breakthroughs from leading innovators in clean energy technology.