Search

In multi-beamlet negative ion accelerators for neutral beam injectors, the transverse magnetic field necessary for suppressing the co-extracted electrons induces a deflection of the negative ion beamlets that must be corrected. For the design, particle-tracing simulation codes are used to compute ion trajectories and optical properties of the beamlets in the acceleration stage. In these codes, uniform boundary conditions are normally assumed for the ion current density distribution at the surface (called meniscus), where negative ions are extracted from the plasma and form beamlets, which are accelerated across the apertures of the accelerator grids. Recently, experimental campaigns dedicated to the accurate measurement of the beamlet deflection in the acceleration phase revealed higher deflection than foreseen by simulations. In this work, we demonstrate that an agreement with the experimental data can be obtained by incorporating in the numerical simulations a non-radially symmetric distribution of the ion current density extracted across the meniscus surface. In the first part, the asymmetry of the ion current density at the meniscus is studied in an empirical way, by analysing and fitting the experimental results obtained with different operating parameters. In the second part, we show that the ion current asymmetry estimated by this procedure is well consistent with the flow pattern of H− ions calculated in the meniscus zone using a detailed particle in cell (PIC) model of the ion source in the presence of transverse magnetic field. [ABSTRACT FROM AUTHOR]

Long pulse acceleration of hydrogen negative ion beams with the power density over 70 MW/m2 and the pulse length over 100 s has been demonstrated for the first time by using a multi-aperture 3-stage accelerator. Such long pulse acceleration was achieved by integrating the design of beam optics and voltage holding capability to meet the requirements of JT-60SA. By using the newly designed accelerator for JT-60SA, voltage holding at 500 kV with beam acceleration was stably sustained even after 5 g of cesium was seeded, and heat load on each acceleration grid was reduced below the allowable level for long pulse, less than 5% of total acceleration power. As a result, 500 keV, 154 A/m2 for 118 s beam acceleration was achieved, which satisfies the requirement of the negative ion source for JT-60SA. This pulse length of such high-power density beams is longest in the world. In addition, the result contributes to the long pulse acceleration of multi-stage electrostatic accelerators, such as 1 MeV negative ion accelerator for ITER. [ABSTRACT FROM AUTHOR]

The physical mechanism of the formation of the negative ion beam halo and the heat loads of the multi-stage acceleration grids are investigated with the 3D PIC (particle in cell) simulation. The following physical mechanism of the beam halo formation is verified: The beam core and the halo consist of the negative ions extracted from the center and the periphery of the meniscus, respectively. This difference of negative ion extraction location results in a geometrical aberration. Furthermore, it is shown that the heat loads on the first acceleration grid and the second acceleration grid are quantitatively improved compared with those for the 2D PIC simulation result. [ABSTRACT FROM AUTHOR]

• Joint experimental campaigns carried out in QST lab in Naka, Ibaraki, Japan. • Full compensation of magnetic criss-cross deflection of beamlets obtained. • Non-uniform extraction assumed for interpreting the experimental results. • Numerical simulations can reproduce the experimental results with good accuracy. In the framework of the collaboration between Consorzio RFX (Padova, Italy) and QST (Naka, Japan) two experimental campaigns have been organized on the Negative Ion Test Stand (NITS) in Naka, employing an ITER-like multi-beamlet configuration. These campaigns were aimed at compensating for the undesired magnetic deflection of the ion beam. The experiments have clarified the influence of operating parameters on the optical properties of the ion beam. The physical phenomena acting on the optics have been investigated, in particular the effect of an asymmetry of the ion current density at meniscus. Comparisons with simulations have been made and the modeling has been improved consequently. During the second experimental campaign, a perfect compensation of the ion deflection has been achieved. [ABSTRACT FROM AUTHOR]

• Experimental campaign on MTF carried out in QST lab in Naka, Ibaraki, Japan. • Opera, SLACCAD, COMSOL, EAMCC3D correctly predict beam divergence. • Semi-empirical Opera model with current density asymmetry correctly predicts the beam deflection. • COMSOL and EAMCC3D slightly underestimate heat loads on the grids. It is a widely recognized phenomenon that, in negative ion sources for NBI, the magnetic fields in the source cause an asymmetry of the current density at the meniscus, which affects the beam optics. In order to achieve a more realistic estimation of the negative ion beam optics, the ion current density profile at the beamlet meniscus is sometimes assumed to be non-symmetric with respect to the beamlet axis. The new hypothesis tested in this paper is an empirical law obtained by recent investigations on a similar Kamaboko ion source, which relates this asymmetry to the magnetic field strength at the meniscus and to the power injected in the ion source. For the first time this hypothesis is used to directly and correctly predict the value of the beamlet deflection. The MeV ion source Test Facility (MTF) is a high voltage negative ion beam accelerator dedicated to fusion experiments in QST, Naka, Ibaraki, Japan. MTF can be operated as a three- or five-stage multi-aperture accelerator with a configuration and beam energy comparable to the future ITER NBI. In 2019, experiments were conducted on MTF aiming at accelerating a H- beam with high energy and good optics for a long duration of the beam (around 100 s). These experiments allowed to measure the beamlet divergence and deflection as well as the heat loads due to particle impinging on the accelerator grids, over a wide range of operating parameters. The predictions of several beam simulation codes (Opera, COMSOL, SLACCAD, EAMCC3D) have been compared to the experimental measurements, which allows to obtain a good agreement in terms of beam optics and a fair agreement in terms of heat loads. [ABSTRACT FROM AUTHOR]