EXL-50U Spherical Tokamak: Advanced Fusion Device
- EXL-50U is a spherical tokamak featuring a low aspect ratio, compact design, and innovative plasma control methods for non-inductive start-up.
- It utilizes multi-harmonic electron cyclotron heating with gyrotrons and real-time equilibrium reconstruction to achieve 1 MA-class plasma currents.
- Enhanced diagnostic systems, advanced wall conditioning, and impurity control strategies support scalable performance for future hydrogen–boron fusion reactors.
The EXL-50U spherical tokamak is a compact, low-aspect-ratio fusion device designed for advanced plasma confinement, high-efficiency non-inductive current drive, and reactor-relevant hydrogen–boron fuel experiments. Upgraded from the original EXL-50, EXL-50U implements major engineering, diagnostic, and operational enhancements, targeting 1 MA-class plasma currents, hot-ion modes, and scalable empirical demonstrations supportive of next-generation spherical tokamak (ST) reactors. Research on EXL-50U provides a definitive platform for exploring multi-harmonic electron cyclotron wave (ECW) current drive, real-time equilibrium control, and advanced wall conditioning for low-Z, impurity-managed operation.
1. Device Geometry, Magnet System, and Core Parameters
EXL-50U features a major radius –$0.8$ m, minor radius –$0.45$ m, and aspect ratio –$1.85$ (Shi et al., 7 Feb 2025, Liu et al., 2024, Jiang et al., 22 Dec 2025). An all-metallic vacuum vessel, reinforced for mechanical stiffness, houses the full-metal center column with a 0.5–0.6 Vs-capable central solenoid (CS), water-cooled copper toroidal field (TF) coils (providing T at ), and a flexible set of four to twelve poloidal field (PF) coils for shaping and stability. The device operates with both limiter and single-null divertor configurations, and incorporates actively cooled, tungsten first-wall components. Upgrades include expanded port access for Thomson scattering, bolometry, interferometry, and charge-exchange recombination spectroscopy (CXRS), as well as fast magnetic diagnostics.
Nominal plasma performance targets and achievements include:
| Parameter | Value | Comments |
|---|---|---|
| $0.6$–$0.8$0 m | Extended from EXL-50 for improved shaping | |
| $0.8$1 | $0.8$2–$0.8$3 m | Implies $0.8$4 |
| $0.8$5 | $0.8$6–$0.8$7 T | On-axis, tunable by TF current |
| $0.8$8 | $0.8$9 MA | Limiter/divertor, non-inductive ramp-up achieved |
| 0 | 1–2 | Elongation at LCFS |
| 3 | 4–5 m6 | Line-averaged, Thomson/interferometer |
| 7 | 8–9 keV | Core, with 800 kW ECRH |
| $0.45$0 | Up to $0.45$1 keV target | Goal for p–B campaign |
| $0.45$2 | $0.45$3 | Edge safety factor at $0.45$4 MA, $0.45$5 T |
Confinement and equilibrium analyses employ the Grad–Shafranov formalism, multi-fluid models, and real-time plasma reconstruction (Zheng et al., 18 Jan 2026).
2. Heating, Current Drive, and Non-Inductive Operation
EXL-50U's core innovation is non-inductive current start-up and ramp-up relying on multi-harmonic ECW in synergy with Ohmic assist from the CS. Six gyrotrons (three at 28 GHz, two at 50 GHz, up to 200 kW each, 2 s pulse duration) enable O/X-mode launching from the low-field-side (Jiang et al., 22 Dec 2025, Shi, 12 May 2025). By controlling $0.45$6, the radial location and number of accessible EC resonance layers (n=1–5) are adjusted such that multiple harmonics lie within the plasma—a prerequisite for high-efficiency, non-linear multi-harmonic EC absorption and current drive (Banerjee et al., 2021).
Non-inductive start-up efficiency increases monotonically with the number of in-plasma harmonics: for 28 GHz at $0.45$7 m, raising from 1 to 3 layers increased non-inductive $0.45$8 from 25 to 78 kA; similar scaling is shown for 50 GHz (Table below).
| EC Harmonics Accessible | Peak $0.45$9 (kA, 28 GHz) | Peak 0 (kA, 50 GHz) |
|---|---|---|
| One layer | 25 | — |
| Two layers | 45 | — |
| Three layers | 78 | — |
| 2–4 layers (by limiter/BT scan) | — | ≈96 |
With multi-pass absorption (wall reflection and O1X conversion), ECW-driven current-drive efficiency (2) reaches 0.06 MA MW3m4 for three or more harmonics, and net non-inductive fraction (5) can reach 70% at 6 MA (with ECW 7 CS) (Jiang et al., 22 Dec 2025).
Underlying mechanisms are:
- Multi-harmonic heating: Doppler-broadened cyclotron absorption at successive harmonic layers accelerates electrons to high energies.
- Multiple reflections: O-mode not absorbed reflects from the vessel wall, converts with probability 8 to X-mode, enabling further resonance.
- Multi-pass absorption and mode conversion: Post-reflection, X-mode or EBW is absorbed at the upper hybrid resonance (UHR), heating electrons to 9100 keV (Banerjee et al., 2021, Jiang et al., 22 Dec 2025).
Modeling with GENRAY+CQL3D and X-ray diagnostics validate that $1.85$060% of current is carried by this energetic electron population.
3. Equilibrium Reconstruction and Real-Time Control
Real-time MHD equilibrium is achieved on EXL-50U using PTEFIT—a modular, GPU-accelerated solver mapping the Grad–Shafranov equation and a least-squares system to a PyTorch/TensorRT computational graph (Zheng et al., 18 Jan 2026). The algorithm reconstructs plasma flux and profiles at 129$1.85$1129 resolution, attaining $1.85$2 ms time per slice, $1.85$3 faster than conventional offline EFIT, and sub-centimeter LCFS accuracy. The workflow incorporates Green’s function evaluation, Picard iteration, quadratic O/X-point search, and flux-surface averaging for safety factor and geometric indices.
Closed-loop feedback (PID for $1.85$4, multi-point “isoflux” for strike/X-point positioning) is realized at $1.85$51 ms cycle times, stabilizing the plasma to within 2 cm of target locations even in the presence of disturbances. The framework is immediately extensible to transport, stability, and kinetic solvers, and supports device portability through PyTorch–ONNX–TensorRT graph export.
4. Fueling, Impurity Control, and Wall Conditioning
EXL-50U implements integrated boronization for impurity control, wall conditioning, and real-time fueling (Shi, 12 May 2025). Mixtures with 30% B$1.85$6H$1.85$7 and 70% H$1.85$8 yield core boron fractions $1.85$9, stabilized by simultaneous boron powder injection during discharge. This reduces 0, mitigates radiative loss, and improves ramp-up rates—evidenced by a 78% increase in 1 for H–B shots versus hydrogen alone.
Wall boronization (gas, powder, pellet) is regularly applied to maintain low-Z plasma-facing surfaces, and dynamic gas-puffing with divertor cryopumping enables control over edge recycling and impurity influx. Impurity transport is characterized using spectroscopic and bolometric techniques, with modeling based on diffusive–pinch frameworks.
5. Diagnostics Suite and Turbulence Measurement
Comprehensive diagnostics installed on EXL-50U include multi-chord CO2 and 140–330 GHz interferometry, Thomson scattering (profiled, 35% accuracy), CXRS for ion temperature, 32-channel bolometry, as well as high-frequency magnetic probes, Mirnov arrays, and hard X-ray spectrometry for energetic-electron studies (Shi, 12 May 2025, Zheng et al., 18 Jan 2026).
For microturbulence, the Doppler backscattering (DBS) diagnostic concept (Liang et al., 23 Sep 2025) uses a U-band (40–60 GHz) beamline with toroidal and poloidal steering, optimized to match the high (%%%%64$0.45$565%%%%) magnetic pitch angle at the outboard midplane. SCOTTY beam-tracing predicts accessible 6 coverage of 7–8 cm9, spanning 0–1, with frequency tuning and angle matching to minimize signal attenuation due to mismatch.
6. Physics Results: Hydrogen–Boron Operation and Advanced Regimes
EXL-50U demonstrated the first 1 MA-class hydrogen–boron plasma for a ST at 2 T, 3 m, with non-inductive start-up and rapid ECRH+CS ramp-up (Shi, 12 May 2025). Real-time boron fueling and wall conditioning yielded low impurity (4), improved 5, and core 6 up to 3–3.5 keV. The operation regime achieved 7, 8, and 9 s (Ohmic), supporting 0 non-inductive current fractions (Jiang et al., 22 Dec 2025).
Physics insights:
- Low loop-voltage start-up with ECRH–CS synergy reduces stress on the CS system.
- Real-time boronization is an effective strategy for radiative and impurity control in metal-walled devices.
- Two-stage ramp-up profiles (fast initial, then slow to avoid plasma-wetting instability and disruptive expansion) are critical for device stability.
- Demonstrated 1 A W2 for multi-harmonic X-mode ECW, up to 3 A W4 in multi-pass scenarios (Banerjee et al., 2021, Shi et al., 7 Feb 2025).
EXL-50U thus validates the empirical scaling required for future reactor-class STs, especially EHL-2, targeting 3 MA, 5 m, 6 keV, and advanced 7 ratios (Liu et al., 2024).
7. Scientific Roadmap and Research Program
The EXL-50U program, operated by ENN Science and Technology and associated research teams, serves as the critical test bed for ST-based proton–boron fusion (Liu et al., 2024, Shi et al., 7 Feb 2025). Key objectives and milestones include:
- Demonstrate 8 MA non-inductively at high 9, $0.6$0>3 keV, $0.6$1 s, and toroidal $0.6$2>30%.
- Execute systematic boron fueling and wall-conditioning campaigns for impurity control and edge physics studies.
- Integrate and operate multi-MW heating and current-drive systems for D–T and p–$0.6$3B–relevant scenarios (including NBI, LHCD, ICRH upgrades).
- Deliver empirical guidance and validated models for the scaling, engineering, and physics integration of EHL-2—a 3 MA, $0.6$4 T, hot-ion, advanced-fuel ST for the late 2020s (Liu et al., 2024).
EXL-50U is scheduled to undergo further upgrades, targeting $0.6$5 MA, $0.6$6 T, and higher elongation, with extended high-$0.6$7 and pulse-length operation, advanced equilibrium feedback, and continuous impurity management.
References: (Banerjee et al., 2021, Shi et al., 2021, Liu et al., 2024, Shi et al., 7 Feb 2025, Shi, 12 May 2025, Jiang et al., 22 Dec 2025, Liang et al., 23 Sep 2025, Zheng et al., 18 Jan 2026)