Negative Hydrogen Ion Sources for Fusion: From Plasma Generation to Beam Properties

Abstract

The neutral beam injection systems for the international fusion experiment ITER used for heating, current drive, and diagnostic purposes are based on RF-driven negative hydrogen ion sources with a source area of roughly 0.9 m × 1.9 m. The sources operate at 0.3 Pa in hydrogen and in deuterium using a total available RF generator power of 800 kW per source at a frequency of 1 MHz. In order to fulfill the challenging requirements for ITER and beyond (like a DEMOnstration power plant, DEMO), worldwide developments are underway addressing the topics of plasma generation, ion extraction together with the issue of reducing and stabilizing the co-extracted electron current, and the beam properties. At the example of the activities at the ITER prototype source and the size scaling experiment ELISE, the present status and its challenges are summarized. The RF power transfer efficiency of these sources is only about 65% in maximum, giving significant room for improvements to relax the demands on the RF generator and ensure reliable operation. The plasma uniformity in front of the large extraction system is the result of plasma drifts. They have a huge impact on the nonuniformity of the co-extracted electrons and influence the ions and thus the beam properties as well. Understanding the optics of such large beams composed of hundreds of beamlets is a crucial task and is under continuous improvement. The main challenge, however, is still the fulfillment of the ITER requirements for deuterium, in particular, for long pulses. The management of caesium, which is evaporated into the source to generate sufficient negative ions by the surface conversion process, is one of the keys for stable and reliable operation.

Tutorial: Modeling of the extraction and acceleration of negative ions from plasma sources using particle-based methods

“…The computational time to reach convergence is about 150 h using 4096 cores. The negative ion current density extracted from the model 144 j H − , ext is on the order of 250 A.m −2 similar in principle to experimental observations but with (1) a negligible amount of particles coming from the flat side of the PG facing the plasma while it represents ∼ 25% of the total surface area, (2) a higher ratio of ions originating from the plasma volume (∼ 35% instead of ∼ 10% 34 ), and (3) the remaining (∼ 65%) coming from the chamfered surface with a high divergence angle, all being in contradiction with measured data in the experiments 30 . The latter also predicts an increase of the extracted negative ion current somewhat proportional to the aperture surface area in a cesiated ion source 41 .…”

“…The evaporation of tungsten from the filaments and its re-deposition on the grid surfaces has a deleterious impact on the production of negative ions 33 . The power injected in the BATMAN source is on the order of 50-100 kW and the pressure is around 0.3 Pa. One of the critical issues is to increase the power coupling efficiency that is actually limited to 60% 30 . The RF driver for the SNS machine is very similar (but with a smaller volume), the pressure is around 3.2 Pa (24 mTorr), the frequency is 2 MHz, and the injected power is in the range of 55-65 kW 22 .…”

“…The volume of the discharge is 0.4 l in the Linac4 source, the injected RF power is ∼ 40 kW (1 MHz frequency), and the gas pressure is about 3 Pa 18 . Note that the plasma in the driver region is mainly an electropositive plasma, and the electron temperature is too high for generating negative ions through attachment processes 30 . The H 2 gas is used in these ion sources.…”

“…The electrodes must be polarized positively with respect to the plasma potential for extracting negative charges. In BATMAN upgrade, the plasma grid (PG) is at -45 kV, the extraction grid (EG) is between 5 and 10 kV above the PG voltage, and finally the grounded grid (GG) at ground potential 30 . Extraction and grounded grids are made of copper while the PG is in molybdenum 35 .…”

“…For comparison, the values determined from Langmuir probe measurements also at a distance of 2.6 cm from the PG are included (open symbols). Note that the filter field strengths are given for an axial position of 2 cm above the PG surface and both vertically and horizontally centered 30 . Reproduced with permission from Fantz et al, Front.…”