First measurements of p¹¹B fusion in a magnetically confined plasma

The study addresses whether p¹¹B fusion reactions can be produced and diagnosed in a magnetically confined, thermonuclear plasma. While p¹¹B fusion offers advantages such as abundant, non-toxic, non-radioactive fuel and aneutronic primary reactions, it presents physics challenges due to the much higher required temperatures compared to deuterium-tritium and associated radiative losses. Prior to this work, p¹¹B fusion had not been demonstrated in a magnetically confined plasma environment, limiting understanding of collective plasma effects on the fusion rate. The purpose of the study is to create conditions for p¹¹B fusion in the Large Helical Device (LHD) using high-energy neutral beam-injected protons and boron powder injection, and to detect resulting MeV alpha particles. Demonstrating such reactions and correlating them with simulations is important for the development of aneutronic fusion concepts and to enable further studies of fast-ion-driven and non-thermal effects relevant to p¹¹B reactors.

Proton–boron fusion has been explored primarily in theoretical studies, nuclear physics experiments, and laser-produced plasmas, including beam-target experiments where interactions are limited to binary collisions. Updated cross-section evaluations and kinetic treatments suggest that thermal p¹¹B plasmas can achieve high Q and potentially ignition under appropriate conditions, especially with low internal magnetic fields and ion–electron temperature separation to reduce radiation losses. Concepts such as beam-driven field-reversed configurations (FRCs) envision maintaining non-thermal ion populations to enhance fusion. Prior work also indicates potential collective effects (e.g., beam-driven waves) that could boost fusion output, and resonance structures in the p¹¹B cross section may influence rates in the presence of high-energy tails and non-Maxwellian distributions. However, magnetically confined plasma experiments directly measuring p¹¹B fusion had not been reported before this study.

Platform and operating scenario: Experiments were conducted on the Large Helical Device (LHD), a heliotron stellarator with a few-tesla magnetic field providing good confinement of fast ions. High-energy protons were produced using tangential negative-ion neutral beams (NBs) operated at 160 kV; LHD has three such beams (and two lower-energy positive-ion beams). For the reported discharges, two tangential beams (NB1 and NB3) were fired at 160 kV for 4 s each in sequence to provide continuous high-energy proton injection between t ≈ 3.3–7.3 s. Typical plasma parameters included toroidal field Bt ≈ 2.75 T, Rax = 3.60 m, central Te ≈ 2 keV, and line-averaged ne ≈ 2 × 10¹⁹ m⁻³. Boron fueling via real-time boronization: Boron was introduced using the Impurity Powder Dropper (IPD), delivering sub-millimeter grains of boron or boron nitride into the plasma edge. Besides wall conditioning and confinement improvement, the method led to significant boron accumulation at mid-radius as measured by charge exchange recombination spectroscopy, producing substantial boron densities during injection. EUV spectroscopy provided relative boron density evolution used in simulations. Alpha detection diagnostic: A custom alpha particle detector was built around a 2000 mm² Passivated Implanted Planar Silicon (PIPS) diode (Mirion PD 2000-40-300 AM), operated at 72 V reverse bias (partially depleted). The detector was housed in a tungsten shield with a graphite collimator (28° full-angle acceptance) and installed on a movable manipulator entering from the 10.5-L diagnostic port into the divertor region. To minimize contamination by X-rays and low-energy particles, the detector had no direct line-of-sight to the core; instead, the ~3 T magnetic field guided alpha orbits to it. A 2 µm platinum foil in front of the detector blocked photons below ~4 keV with minimal attenuation of MeV alphas (range ≈ 5.9 µm in Pt at 4 MeV). Detector orientation was optimized using Lorentz orbit tracing calculations to maximize alpha collection while avoiding direct plasma sightlines. Electronics and calibration: The detector current pulses were read by a transimpedance amplifier (Femto HCA-2M-1M-C, 2 MHz corner) outside the vacuum vessel over a ~9 m single-conductor cable. Bench tests with a ²¹⁰Po 5.41 MeV alpha source characterized signal attenuation and cable effects, with capacitance and inductance management implemented. Additional calibration used a ²⁴¹Am source to verify pulse shapes and amplitudes. A graphite heat shield and DS-4 graphite collimator were thermally designed via ANSYS to withstand divertor heat fluxes (~1 MW/m²) during 10 s pulses, ensuring detector temperatures remained below 100 °C (verified by thermocouples). Data acquisition and analysis: Time traces of NB voltages/powers, EUV B⁺ line intensity (proxy for boron content), and detector pulses were recorded. Pulse detection counted negative-going spikes consistent with calibration pulse shapes. Discharges with and without boron injection were compared under otherwise identical NB conditions to isolate p¹¹B fusion signals. Numerical modeling using the FBURN code computed the global p¹¹B fusion rate based on measured NB parameters, bulk plasma profiles, and boron density profiles (from charge exchange and EUV). Fast ion dynamics, energy spectra, and spatial profiles were modeled to generate a time-resolved fusion rate for comparison with measured pulse count rates.

• Clear alpha particle detection correlated with high-energy NB injection and the presence of boron. Negative-going pulses (~150 mV) appeared coincident with NB turn-on, with pulse shapes matching calibration with a ²⁴¹Am source.
• Discharges with boron powder injection showed dramatically higher pulse count rates than those without boron under otherwise identical NB conditions. Peak count rate with boron was about 150 thousand counts per second (kcps), while the case without boron was <1 kcps, likely due to residual wall-deposited boron.
• Time dynamics: The leading-edge rise of measured count rate tracked the simulated global p¹¹B fusion rate from FBURN, reflecting boron accumulation timescales. The trailing-edge drop at NB turn-off was fast, indicating dominance of beam-ion dynamics once beams ceased, even with boron still present.
• Simulation inputs and uncertainties: Two tangential NBs operated at 160 kV from t ≈ 3.3–7.3 s. The simulated global fusion rate uncertainty reflected ±1 kV beam energy measurement error. Measurement error bars were based on Poisson counting statistics in 10 ms bins.
• Rate magnitude expectations: Calculations predict that experimentally achieved boron densities together with simultaneous operation of all three high-energy beams would yield global p¹¹B fusion rates on the order of 10¹⁴ s⁻¹. Large alpha Larmor radii (~10 cm) imply most alphas are lost to walls/divertor within a few orbits, enabling detection outside the core plasma.
• Boron accumulation: Real-time boron powder injection resulted in significant boron density build-up in the mid-radius/core region during injection, enhancing p¹¹B reaction opportunities with well-confined fast protons.

The experiments demonstrate, for the first time, generation and detection of p¹¹B fusion-born alphas in a magnetically confined plasma, validating a key step toward aneutronic fusion energy concepts. The strong correlation between alpha count rates and both boron content and high-energy proton injection confirms the reaction’s occurrence in the LHD environment. The measured temporal evolution of the count rate agrees well in a relative sense with numerical predictions that incorporate fast-ion physics and boron profiles, elucidating the roles of boron accumulation and beam-ion dynamics. These findings establish experimental capability to study collective and non-thermal effects unique to p¹¹B plasmas in magnetic confinement, such as resonance-enhanced reaction rates with high-energy ion tails and beam-driven wave interactions. The results are significant because they shift p¹¹B research from solely theory, accelerators, and laser plasmas into the practical setting of magnetically confined plasmas where reactor-relevant physics, including fast-ion transport and profile control, can be optimized for improved performance.

This work reports the first realization and measurement of p¹¹B fusion in a magnetically confined plasma, achieved by combining boron powder injection to accumulate boron in the core with high-energy (160 kV) neutral beam-injected protons in the LHD. A custom PIPS-based alpha detector, engineered to reject photon contamination and survive divertor heat fluxes, detected MeV alpha particles whose count rates correlated strongly with NB operation and boron presence, and matched the simulated global p¹¹B fusion rate dynamics. This establishes a platform for systematic experimental studies of p¹¹B fusion physics in magnetic confinement. Future research will pursue increased fusion gain via alpha channeling, profile tuning, phase-space engineering, fusion-product current drive, and exploitation of collective beam-induced heating, as well as improved absolute diagnostics of alpha spectra and fast-ion distributions to reduce survivorship biases and enable quantitative absolute rate comparisons.

• Absolute rate comparison is limited by survivorship bias: the detector measures only alphas that reach it, not the full source spectrum in the plasma. Accurate absolute benchmarking would require detailed knowledge of the in-plasma alpha energy/angular spectrum and comprehensive orbit-loss modeling.
• Most alphas have large Larmor radii and are lost to walls/divertor within a few orbits, reducing detected fraction and complicating absolute efficiency estimates.
• Background in boron-free shots likely arises from residual boron deposited on walls, limiting dynamic range of the “no-boron” baseline.
• Measurement uncertainties include Poisson counting statistics and ±1 kV uncertainty in beam energy affecting simulation-derived rates.
• Detector orientation and shielding, while reducing photon contamination, constrain acceptance and may filter parts of the alpha phase space.