Masaki Osakabe

This paper summarizes the contributions presented in the ‘magnetic fusion experiments including validation (EX)’ category at the 30th IAEA fusion energy conference 2025, held in Chengdu, People’s Republic of China, from 13–18 October 2025. Conference materials are available at the FEC 2025 conference website ( www.iaea.org/events/fec2025 ), under the category ‘magnetic fusion experiments including validation (EX)’.

We investigated the phase-space dynamics of helically trapped fast ions during energetic-particle-driven magnetohydrodynamic (MHD) bursts in the Large Helical Device (LHD) using a fast-sampling fast-ion D α (F-FIDA) diagnostic based on charge-exchange recombination spectroscopy. Conditional averaging synchronized with the timing of radio-frequency (RF) bursts accompanying the MHD bursts enabled the extraction of F-FIDA signals associated with the bursts. The phase-space dynamics inferred from the F-FIDA spectra showed two distinct behaviors: a rapid increase in the F-FIDA signal near the plasma edge immediately after the burst, and a delayed decrease inside the plasma occurring a few milliseconds later. The former is likely related to localized deformation and magnetic reconnection during the early phase of the MHD event, while the latter appears to be caused by rotating, downward-chirping MHD modes, leading to phase-space dynamics of fast ions. These results contribute to understanding the fast-ion transport processes during energetic-particle-induced MHD activity in fusion plasmas.

A two-dimensional fluid model of the RF ion source in NIFS test stand was established using COMSOL Multiphysics to examine the influence of the cusp magnets on plasma transport and RF power coupling. Particular attention was paid to the magnetic field intensity of the cusp magnets on driver backplane and its impact on plasma behaviour in both the driver and the extraction region. Simulation results reveal that strong magnetic confinement from the upper cusp magnets suppresses cross-field plasma transport, leading to reduced electron density near the plasma grid and less efficient RF power deposition. To explore potential optimization, a modified configuration with two smaller magnets of half the original size was investigated. This adjustment significantly enhances plasma density, increasing the peak density in the driver by nearly a factor of two. However, the enhanced plasma density is accompanied by increased spatial gradients in the extraction region, indicating that plasma uniformity is not improved under the modified magnetic configuration. These results reveal a trade-off between plasma density enhancement and spatial uniformity, arising from the interplay between magnetic confinement and plasma transport. This study provides insight into the role of cusp magnetic fields in regulating plasma generation and spatial gradients, and highlights the importance of balancing plasma density and uniformity in the optimization of RF negative ion sources.

We have developed a simulation code for the plasma current driven by the neutral beam in the Large Helical Device (LHD). To compare the results between the experiment and the simulation, the current component, not only the beam-driven current, but also the other components, must be estimated. In this paper, we have integrated the codes that solve electromagnetic induction, neoclassical transport (DKES/PENTA), and the fast-ion distribution function (TASK3D-a and TASK/FP) to achieve NBCD simulation. The code was validated through model simulations, which assume steady-state beam injection. Then the code was applied to actual LHD discharges. Although the simulated value became similar to the measured one, the time constant of the current evolution disagreed. By considering the mutual induction between the plasma and the external coils, the simulated time constant of the plasma current evolution became closer to the experimental one.

In the Large Helical Device (LHD) experiments using fast protons and fast deuterons, hybrid simulations were performed with the MEGA code to investigate the synergistic effects of multiple fast-ion species on instabilities and fast-ion transport, where multiple fast ions interact with the magnetohydrodynamic (MHD) fluid. In the MEGA simulations, the burst amplitude of Alfvén eigenmodes (AEs) after the fast-ion stored energy reaches a steady state is larger than that before the steady state, leading to a significant increase in the fast-ion loss rate. The dominant AE frequencies obtained in the simulations are in good agreement with those observed in the experiments. In the case of multiple fast-ion species, additional AEs appear that are absent when only a single fast-ion species drives the AEs, and these modes are caused by the redistribution of multiple fast ions. Consequently, these additional AEs lead to increased loss rates of both fast protons and fast deuterons. The synergistic effect of multiple fast-ion species is thus shown to enhance fast-ion transport and losses in the LHD.

Proton-boron-11 (p- ¹¹ B) fusion is an attractive concept for advanced fusion energy because of its nearly aneutronic characteristics. However, efficient use of beam-injected fast ions may be limited by their slowing down through Coulomb collisions, which lead to lower fusion reactivity. In this study we investigate whether ion cyclotron range of frequencies (ICRF) heating can ameliorate this effect by resonantly accelerating fast ions in the Large Helical Device (LHD) of a stellarator. Using the tangential neutral beam (NB) injection to supply energetic protons at ∼160-170 keV, together with boron seeding, we conducted a series of experiments with superimposed hydrogen minority ICRF heating for beam proton acceleration. Three complementary experimental methods, ICRF power modulation, NB termination, and NB energy scan, consistently showed that ICRF waves accelerate beam ions and enhance the p- ¹¹ B fusion yield. This led to measurable enhancements in the p- ¹¹ B fusion alpha-particle yield, as observed with a dedicated fusion-product detector. These findings provide the first direct experimental evidence that ICRF heating can improve the efficiency of beam-driven p- ¹¹ B fusion reactions.

The impact of the radial electric field (Er) on the confinement of energetic particles in the Large Helical Device is studied through gyrokinetic simulation. Particle orbits are characterized as passing, helically trapped, and barely trapped. Er has little effect on passing particles. A positive Er is found to improve the confinement properties of trapped particles by increasing their precession rate, thereby reducing radial drift over a bounce motion. A negative Er is found to do the opposite. Er can induce particles to change orbit topology. The effects of an experimental Er profile are discussed. Overall, positive Er is found to improve while negative Er is found to degrade confinement of energetic particles. Furthermore, the lost energetic particles originate from localized regions of phase space. Knowledge of which can be utilized to induce radial current to control Er by injecting neutral beams into these regions.

A study of aneutronic fusion, which refers to fusion reactions that do not produce neutrons, has been conducted on the Large Helical Device (LHD). For the D–³He study, feasibility investigations of detecting charged fusion products have been performed. High-energy neutral beams, lost proton detectors, and gamma-ray detectors were planned for installation to validate the D–³He fusion reactions. Numerical calculations show that the expected D–³He fusion rate is 2.7 × 10 ¹⁶ s ^-1 with most of the protons being lost immediately from the plasma. For the p–¹¹B study, the total alpha particle emission rate was estimated to be 10 ¹⁴ s ^-1 , and the loss points of alpha particle distribution were calculated. Based on these numerical simulations, we installed an alpha particle detector on a manipulator and positioned it at the bottom of the LHD. The time evolution of the alpha particle detection rate, measured by the detector, was found to be consistent with the predictions from the numerical simulations, demonstrating the first successful observation of the p–¹¹B reaction in a magnetic confinement system.

Tritium yields in all six Large Helical Device (LHD) deuterium experiment campaigns, or 41,064 discharges, have been numerically estimated. Usually, tritium yields are estimated from neutron yields by assuming that they are the same. In present fusion devices, this assumption is inaccurate because a fusion reaction between thermal-deuterons and fast-deuterons is dominant. In this paper, the energy distribution of fast-deuterons injected by neutral beam injectors is taken into account for the estimation of the ratio of the tritium yields to the neutron yields, Yt/Yn. By integrated simulation, the Yt / Yn is approximately 0.94 in each campaign. Because assumptions applied in the simulation aim to avoid under-estimation of the tritium yields compared to the actual value, the Yt/Yn should be 0.85 < Yt/Yn < 0.94.

A boron charge exchange spectroscopy system has been applied to measure the radial profiles of absolute boron density. The in situ calibration of charge exchange spectroscopy was performed using the boron pellet in a Large Helical Device. The calibration factor was derived by comparing the increment of boron density measured with a charge exchange spectroscopy system by the increment of electron density measured with a YAG Thomson scattering system. The radial profiles of absolute boron density are derived for the plasma using boron pellets, tracer-encapsulated solid pellets (TESPEL), and boron impurity power drops (IPD). By using boron pellets, TESPEL, and IPD, the boron concentration (nB/ne) was controlled in the wide range from 4% to 17%. The radial profiles of boron are peaked in the core for the boron pellet and TESPEL discharge, while they are slightly hollow for the boron IPD plasma.