One of the persistent concerns in plasma physics is whether results observed on a single machine reflect universal physics or are artifacts of that machine’s particular geometry, heating systems, or wall conditions. This concern is especially acute for negative triangularity (NT), where most of the world’s data comes from two devices: DIII-D in San Diego and TCV in Lausanne, Switzerland. Both are conventional aspect-ratio tokamaks with broadly similar geometries. To build confidence in NT as a pilot plant scenario, we need to know whether the same physics appears on fundamentally different machines.
MAST-U, the Mega Amp Spherical Tokamak Upgrade at the UK Atomic Energy Authority, is about as different from DIII-D as a tokamak can be. Spherical tokamaks have a very tight central column, giving them a low aspect ratio — roughly the shape of a cored apple rather than a donut. They operate in a different regime of magnetic curvature and plasma shaping, and they have historically been difficult to shape strongly in either positive or negative triangularity. From 2023 to 2025, I led a project to implement negative triangularity configurations on MAST-U for the first time, extending the NT database to this entirely new machine class.
The core scientific question was straightforward: does the mechanism that suppresses ELMs in NT plasmas on conventional tokamaks also operate in the spherical tokamak geometry? The NT ELM suppression effect on DIII-D is linked to a change in the magnetic shear at the plasma edge that is induced by the negative triangularity shape. If the same shear-induced stabilization mechanism is present on MAST-U, it would provide strong evidence that the physics is geometry-robust — and therefore likely to hold in future pilot plants with NT shapes that will not look exactly like any current experiment.
The DIII-D tokamak has been central to establishing the physics of negative triangularity, but validating those results across multiple machines is essential for pilot plant confidence.
This MAST-U campaign is one piece of a broader international collaboration on negative triangularity physics that I help lead as working group co-leader of the EU/US NT collaboration. This initiative brings together experimental teams from DIII-D, TCV, MAST-U, and AUG (the ASDEX Upgrade tokamak in Germany) to systematically compare NT results across machines, identify which features are universal, and build the cross-machine database needed to project NT performance to pilot plant scales.
International collaboration in fusion has a long and productive history — ITER is its grandest expression — but the kind of tightly-coordinated multi-machine campaigns we are running for NT are still relatively unusual. They require building shared analysis frameworks, agreeing on common experimental protocols, and developing enough mutual trust to share data before publication. When they work well, they produce physics understanding that no single machine could provide alone. A result that holds on DIII-D, TCV, MAST-U, and AUG simultaneously is not a coincidence — it is evidence that we have found something real.
Selected publications on this subject:
First Access to ELM-free Negative Triangularity at Low Aspect Ratio
Nelson, A. O., Vincent, C., Anand, H., Lovell, J., Parisi, J. F., Wilson, H. S., Imada, K., Wehner, W. P., Kochan, M., Blackmore, G., McArdle, G., Guizzo, S., Rondini, L., Freiberger, S. & Paz-Soldan, C., Nuclear Fusion 64, 124004 (2024).
ELM suppression and confinement in negative triangularity with stronger shaping in ASDEX Upgrade
Vanovac, B., Hobirk, J., Nelson, A. O., Sauter, O., Dunne, M., Faitsch, M., Puetterich, F., Rainer, D., Stieglitz, E., Strumberger, L., Xianzi, O., Grover, A., Kappatou, P., David, G., Tardini, P., Mantica, P. & White, A., Nuclear Fusion 66 (2026), in review.
Initial Simulations of Negative Triangularity Plasmas on MAST-U
Nelson, A. O., Freiberger, H. S., Wilson, L., Rondini, L. & Paz-Soldan, C., 21st International Spherical Torus Workshop (2022).