Search

During 2016, a joint experimental campaign was carried out by QST and Consorzio RFX on the Negative Ion Test Stand (NITS) at the QST Naka Fusion Institute, Japan, with the purpose of validating some design solutions adopted in MITICA, which is the full-scale prototype of the ITER NBI, presently under construction at Consorzio RFX, Padova, Italy. The main purpose of the campaign was to test a novel technique, for suppressing the beamlet criss-cross magnetic deflection. This new technique, involving a set of permanent magnets embedded in the Extraction Grid, named Asymmetric Deflection Compensation Magnets (ADCM), is potentially more performing and robust than the traditional electrostatic compensation methods. The results of this first campaign confirmed the effectiveness of the new magnetic configuration in reducing the criss-cross magnetic deflection. Nonetheless, contrary to expectations, a complete deflection correction was not achieved. By analyzing in detail the results, we found indications that a physical process, taking place just upstream of the plasma grid, was giving an important contribution to the final deflection of the negative ion beam. This process appears to be related to the drift of negative ions inside the plasma source, in the presence of a magnetic field transverse to the extraction direction, and results in a non-uniform ion current density extracted at the meniscus. Therefore, the numerical models adopted in the design were improved by including this previously disregarded effect, so as to obtain a much better matching with the experimental results. Based on the results of the first campaign, new permanent magnets were designed and installed on the Extraction Grid of NITS. A second QST-Consorzio RFX joint experimental campaign was then carried out in 2017, demonstrating the complete correction of the criss-cross deflection and confirming the validity of the novel magnetic configuration and of the hypothesis behind the new models. This contribution presents the results of the second joint experimental campaign on NITS along with the overall data analysis of both campaigns, and the description of the improved models. A general picture is given of the relation among magnetic field, beam energy, meniscus non-uniformity and beamlet deflection, constituting a useful database for the design of future machines. [ABSTRACT FROM AUTHOR]

A thermal beam emittance measurement method is developed specifically to measure high power density beams for the ITER neutral beam injection system (NBI)’s. A one-dimensional carbon-fiber composite (CFC) target is utilized to withstand the high current densities of the negative hydrogen beam making the beam emittance measurement possible. Schemes for the emittance measurement were discussed since beam instabilities are present in various stages of the beam phase. The allowable time for the measurement will be less than 2 seconds and it’s necessary to monitor the CFC temperature to not exceed 2000°C. The proposed measurement scheme to minimize beam instabilities is through the constant operation of the arc discharge and pulsed beam extraction. Mechanical structures are also being designed for rapid measurements to select different phases of the beam and for implementing the measurement system to long pulse beam operation. [ABSTRACT FROM AUTHOR]

In order to evaluate deposition of cesium (Cs) in the plasma grid (PG) and investigate the origin of the gradual degradation of the negative ion current in a 100 s long pulsed operation in the JT-60SA negative ion source, Cs behavior in the source is model. To do this, Cs adsorption/desorption coefficient with a wide range of temperature is experimentally measured by a surface ionization detector and a quartz crystal microbalance. As a result, both of the adsorption and desorption are drastically varied at temperature range from room temperature to 100 °C. Based on these data, the model calculates the amounts of the Cs mass in the chamber wall and PG. The calculation result shows that, in the operation, an increase of temperature in the chamber wall (Twall) causes the evaporation of the huge amount of Cs in it. This leads to over pilling-up of Cs in the PG and the slight degradation of the negative ion production. Therefore, it is found that the constant temperature operation is needed to sustain the thickness of the Cs monolayer in the PG, and the Twall is desired to be less than 60 °C to suppress the evaporation of Cs in the chamber wall. According to this model, it is showed that there are many combinations of the temperatures to achieve it with low Cs consumption. [ABSTRACT FROM AUTHOR]

A high power density and a long pulse negative ion beam accelerator has been developed in QST to realize the ITER neutral beam accelerator. A main target was the H- ion beam acceleration up to 1 MeV with 200 A/m² for 60 s. After the achievement of 1 MeV at a short pulse, the pulse length has been gradually extended by modifications of grid configurations to suppress grid heat loads due to negative ions and electrons. However, the beam energy was limited up to 0.7 MeV at 60 s operation. One issue was an unstable negative ion production for a long time due to arcing and excess temperature rise of a plasma grid for optimum negative ion production. To suppress arcing, a cathode filament shape was modified. The plasma grid was modified to have a large heat capacitance in order to suppress temperature rise for 60 s operation at high discharge power required for 1 MeV operation. Another issue is to optimize beam optics more precisely. Fine tuning of the 1st acceleration stage was carried to improve beam optics for high current density beam. With these modifications, high power density beam acceleration of 0.97 MeV, 190 A/m² has been achieved for 60 s. This is the first long pulse acceleration of ITER class high power density beam. [ABSTRACT FROM AUTHOR]

In order to understand the physics mechanism of a negative ion extraction in negative ion sources, an emission surface of the negative ions around an aperture at a plasma grid, so-called a meniscus, has been analyzed by an inverse calculation of the negative ion trajectory in a two dimensional beam analysis code. In this method, the meniscus is defined as the final position of the negative ion trajectories which are inversely calculated from the measured beam profile to the plasma grid. In a case of the volume-produced negative ions, the calculated meniscus by the inverse calculation was similar to that obtained in conventional beam simulation codes for positive ion extractions such as BEAMORBT and SLACCAD. The negative ion current density was uniform along the meniscus. This indicates that the negative ions produced in the plasma are transported to the plasma grid uniformly as considered in the transportation of the positive ions. However, in a surface production case of negative ions, where the negative ions are generated near the plasma grid with lower work function by seeding cesium, the current density in the peripheral region of the meniscus close to the plasma grid surface was estimated to be 2 times larger than the center region, which suggested that the extraction process of the surfaceproduced negative ions was much different with that for the positive ions. Because this non-uniform profile of the current density made the meniscus shape strongly concave, the beam extracted from the peripheral region could have a large divergence angle, which might be one of origins of so-called beam halo. This is the first results of the determination of the meniscus based on the experiment, which is useful to improve the prediction of the meniscus shape and heat loads based on the beam trajectories including beam halo. [ABSTRACT FROM AUTHOR]

An innovative magnetic configuration called ADCM (Asymmetric Deflection Compensation Magnets) has been recently proposed for the cancellation of the "criss-cross" deflection of ion trajectories in multi-beamlet negative ion accelerators. Such undesired deflection occurs due to the transverse magnetic field, which is necessary for the suppression of co-extracted electrons. In view of the application of this concept to the ITER Heating Neutral Beam Injector, the new configuration has been experimentally tested for the first time in 2016, in the framework of a research collaboration between Consorzio RFX and QST. [ABSTRACT FROM AUTHOR]

The energy relaxation processes of the surface produced H- ions due to the Coulomb collision and the charge exchange collision is verified by using 3D3V-PIC (three dimensions in real space and three dimensions in velocity space particle in cell) simulation. Under the present simulation condition (ne = 1.0 ×1018 m-3, TH+ = 0.25 eV, TH- = 1.5 eV (the initial temperature of the surface produced H- ion), nH/nH2 =0.5, nH = 9.4 ×1018 m-3, TH- = 0.1 eV), it is indicated the Coulomb collision is dominant for the energy relaxation processes. Moreover, it is indicated from the emittance diagrams that the H- ion beam optics is improved by the energy relaxation processes. [ABSTRACT FROM AUTHOR]

The optics design of the multi-beamlet negative ion accelerator for MITICA has been verified by means of specific experiments in the NITS test facility at the QST lab, employing a scaled-down configuration having the same features as the accelerator of MITICA and ITER HNB. The experiments confirmed the validity of the accelerator design. However it was found that, in the presence of transverse magnetic field in the ion extraction stage, the ion beamlet deflection in numerical simulations was somehow underestimated with respect to the beamlet deflection inferred from thermal images of a CFC target exposed to the negative ion beam. A detailed benchmark, using also another (completely independent) numerical model of the accelerator, has allowed the identification of possible origins of the differences. [ABSTRACT FROM AUTHOR]

The negative ion beam optics is studied quantitatively from the viewpoints of the heat loads due to the beam halo in the accelerator and the emittance of the negative ion beam by using the 3D3V PIC model combined with a Monte Carlo calculation, in which negative ion beam from plasma meniscus formation to the beam acceleration is modeled. As for the heat loads of the A2G and the GRG, the simulation result almost agrees with the experimental result. The simulation result indicates that the secondary electrons from the extractor also contribute to the heat load of the A1G. Moreover, the normalized rms emittance of the negative ion beam after the exit of the GRG are estimated to be around 0.25π mm mrad, which are compatible with the typical values of the negative ion sources. [ABSTRACT FROM AUTHOR]

The temperature control system of the large-size plasma grid has been developed to realize the long pulse production of high-current negative ions for JT-60SA. By using this prototype system for the JT-60SA ion source, 15 A negative ions has been sustained for 100 s for the first time, which is three times longer than that obtained in JT-60U. In this system, a high-temperature fluorinated fluid with a high boiling point of 270 degree Celsius is circulated in the cooling channels of the plasma grids (PG) where a cesium (Cs) coverage is formed to enhance the negative ion production. Because the PG temperature control had been applied to only 10% of the extraction area previously, the prototype PG with the full extraction area (110 cm x 45 cm) was developed to increase the negative ion current in this time. In the preliminary results of long pulse productions of high-current negative ions at a Cs conditioning phase, the negative ion production was gradually degraded in the last half of 100 s pulse where the temperature of an arc chamber wall was not saturated. From the spectroscopic measurements, it was found that the Cs flux released from the wall might affect to the negative ion production, which implied the wall temperature should be kept low to control the Cs flux to the PG for the long-pulse high-current production. The obtained results of long-pulse production and the PG temperature control method contributes the design of the ITER ion source. [ABSTRACT FROM AUTHOR]

R&D of high power density negative ion beam acceleration has been carried out at MeV test facility in JAEA to realize ITER neutral beam accelerator. The main target is H- ion beam acceleration up to 1 MeV with 200 A/m² for 60 s whose pulse length is the present facility limit. For long pulse acceleration at high power density, new extraction grid (EXG) has been developed with high cooling capability, which electron suppression magnet is placed under cooling channel similar to ITER. In addition, aperture size of electron suppression grid (ESG) is enlarged from 14 mm to 16 mm to reduce direct interception on the ESG and emission of secondary electron which leads to high heat load on the upstream acceleration grid. By enlarging ESG aperture, beam current increased 10% at high current beam and total acceleration grid heat load reduced from 13% to 10% of input power at long pulse beam. In addition, heat load by back stream positive ion into the EXG is measured for the first time and is estimated as 0.3% of beam power, while heat load by back stream ion into the source chamber is estimated as 3.5 ~ 4.0% of beam power. Beam acceleration up to 60 s which is the facility limit, has achieved at 683 keV, 100 A/m² of negative ion beam, whose energy density increases two orders of magnitude since 2011. [ABSTRACT FROM AUTHOR]

A new type of traveling wave antenna which excites the lower hybrid wave directly was developed. This antenna is similar to the inductively-coupled combline antenna in that only the first element of the antenna array is excited externally, and subsequent elements are excited passively by mutual coupling between adjacent elements. The main difference is that whereas the inductively-coupled combline antenna makes use of mutual inductance, the presently proposed antenna makes use of mutual capacitance. The radiating elements are located at the voltage maximum, and the electric field induced in the plasma is in the toroidal direction rather than the poloidal direction, matching the polarization of the lower hybrid wave. Optimization studies were carried out to obtain a band-pass characteristic centered around 200 MHz, and a unidirectional wavenumber spectrum with the parallel index of refraction corresponding to approximately 5. Plasma current ramp-up to 2 kA has been achieved on the TST-2 spherical tokamak with 12 kW of RF power at 200 MHz during the initial experimental period using this antenna. Further optimization studies are being performed. [ABSTRACT FROM AUTHOR]

Noninductive plasma current (Ip) start-up experiments using RF power in the lower hybrid frequency range are being conducted on the TST-2 spherical tokamak. The lower hybrid wave (LHW) has demonstrated efficient current drive in conventional tokamaks. However, in spherical tokamak (ST) plasmas with very high dielectric constants (;[variant_greek_epsilon]∼ωpe2/Ωe2>1), accessibility of the LHW to the plasma core is severely limited. Our approach is to keep the plasma density low (such that [variant_greek_epsilon]∼1) during Ip ramp-up. Once Ip reaches a level sufficiently high for neutral beam current drive, plasma can be densified and transformed into an advanced tokamak plasma dominated by the self-driven bootstrap current. Initial plasma start-up experiments were performed on TST-2 using a combline antenna which excites a traveling fast wave. After formation of toroidal flux surfaces, RF power and vertical field were ramped up to increase Ip. Up to 12 kA of Ip has been obtained by this method. Soft X-ray measurements indicate that the electron temperature increases as Ip increases, and hard X-ray spectroscopy indicates that energetic electrons build up as Ip is ramped up. [ABSTRACT FROM AUTHOR]

Plasma current start-up experiments using the lower hybrid wave (LHW) excited by a new dielectric loaded waveguide array antenna are planned on TST-2. The wave excitation efficiency of this antenna is investigated using a versatile FEM solver package, COMSOL. Propagation and absorption of the LHW are calculated by iterating between the TORLH full wave code and the CQL3D Fokker-Planck code. Wave absorption is very weak for Maxwellian plasma, but becomes stronger as the quasilinear deformation of the distribution function develops. [ABSTRACT FROM AUTHOR]

Plasma current start-up and formation of the ST configuration without the use of the central solenoid is a critical issue in ST research. In the TST-2 spherical tokamak (R = 0.38 m, a = 0.25 m), sustainment of an ECRF (2.45 GHz) produced ST plasma by low frequency (21 MHz) RF power alone was demonstrated. Since direct RF current drive can be ruled out, this result implies that the ST configuration is sustained entirely by pressure-driven currents. The Grad-Shafranov equilibrium was generalized to take into account the open field line region with finite plasma current and pressure (truncated equilibrium). In addition to the precessional current of trapped particles, Pfirsch-Schlüter current flowing along the open field line (and partially returning through the vacuum vessel) contributes to the toroidal plasma current. Three phases of plasma start-up are analyzed: (i) the current formation phase, (ii) the current jump phase, and (iii) the current sustainment phase. In the current formation phase, the plasma current is formed and increases slowly, roughly proportional to the stored energy. Closed flux surfaces do not exist and the current density profile is peaked on the outboard side. Once the plasma current increases beyond a critical value (approximately equal to the level necessary to satisfy the major radial force balance), the plasma current increases rapidly (current jump). Closed flux surfaces appear not at the beginning, but at the end of the current jump. In the sustainment phase, plasma has a high poloidal beta, βp = O(1). The equilibrium is characterized by the hollowness of the current density profile, which also determines the fraction of the plasma current inside the last closed flux surface. The plasma equilibrium does not differ greatly between EC-sustained and RF-sustained plasmas. However, RF-sustained plasmas are more turbulent, with larger fluctuation levels over a wide frequency band (0–20 kHz), and in many cases MHD activity in the 5–10 kHz frequency band increases rapidly and terminates the discharge. [ABSTRACT FROM AUTHOR]