Large-scale sources for negative hydrogen ions, capable of delivering an extracted ion current of several ten amperes, are a key component of the neutral beam injection system of the upcoming ITER fusion device. Since the created heat load of the inevitably co-extracted electrons after magnetic separation from the extracted beam limits their tolerable amount, special care must be taken for the reduction of coextracted electrons -in particular in deuterium operation, where the larger amount of co-extracted electrons often limits the source performance. By biasing the plasma grid (PG, first grid of the extraction system) positively with respect to the source body the plasma sheath in front of the PG can be changed from an electron repelling towards an electron attracting sheath. In this way, the flux of charged particles onto the PG can be varied, thus changing the bias current and inverse to it the amount of co-extracted electrons. The PG bias affects also the flux of surface-produced H − towards the plasma volume as well as the plasma symmetry in front of the plasma grid, strongly influenced by an ⃗ E × ⃗ B drift. The influence of varying PG sheath potential profile on the plasma drift, the negative hydrogen ion density and the source performance at the prototype H − source is presented, comparing hydrogen and deuterium operation. The transition in the PG sheath profile takes place in both isotopes, with a minimum of co-extracted electrons formed in case of the electron attracting PG sheath. The co-extracted electron density in deuterium operation is higher than in hydrogen operation, which is accompanied by an increased plasma density in deuterium.

Abstract. The cesium dynamics of the IPP prototype source for negative hydrogen ions is investigated using the Cs laser absorption spectroscopy spatially resolved at two lines of sight during vacuum and plasma phases. The spatial resolved measurement is of particular interest since the source has an intrinsic asymmetry due to the Cs oven position. Comparisons of the Cs homogeneity show an inhomogeneous distribution during vacuum phases and a more uniform distribution of neutral Cs during plasma phases. The Cs densities are compared to the extracted negative hydrogen ion and co-extracted electron current densities, and correlations are shown for different timescales. In addition, the cavity ring-down spectroscopy has been reinstalled for the measurement of the negative hydrogen ion density with a line of sight in the center of the ion source. A comparison of the extracted ion current density with the negative hydrogen ion density in the source shows a clearly linear correlation.

Maintenance-free RF sources for negative hydrogen ions with moderate extraction areas (100-200 cm 2 ) have been successfully developed in the last years at IPP Garching in the test facilities BATMAN and MANITU. A facility with larger extraction area (1000 cm 2 ), ELISE, is being designed with a "half-size" ITER-like extraction system, pulsed ion acceleration up to 60 kV for 10 s and plasma generation up to 1h. Due to the large size of the source, the magnetic filter field (FF) cannot be produced solely by permanent magnets. Therefore, an additional magnetic field produced by current flowing through the plasma grid (PG current) is required. The filter field homogeneity and the interaction with the electron suppression magnetic field have been studied in detail by finite element method (FEM) during the ELISE design phase. Significant improvements regarding the field homogeneity have been introduced compared to the ITER reference design. Also, for the same PG current a 50% higher field in front of the grid has been achieved by optimizing the plasma grid geometry. Hollow spaces have been introduced in the plasma grid for a more homogeneous PG current distribution. The introduction of hollow spaces also allows the insertion of permanent magnets in the plasma grid.

No abstract

Strict requirements are foreseen for the Neutral Beam Injection system (NBI) for ITER: a high extracted current density has to be achieved (33 mA/cm 2 for H − and 28.6 mA/cm 2 for D −) together with very small beam core divergence (< 7 mrad) and a beam uniformity of better than 90%, for a large beam extracted from 1280 apertures. The ion source filling pressure has been set to < 0.3 Pa, in order to keep the stripping losses in the accelerator to a tolerable level, and the ratio of co-extracted electrons to ions should be less than one. In the roadmap towards the development and design of the ITER NBI system, the ELISE test facility is an intermediate step, having half the size of the final ITER NBI source. As well as important scientific and engineering results, ELISE provides highly valuable experience in the operation and performance of a large RF-driven negative hydrogen ion source. At the ELISE test facility it is possible to have an insight into the physics of the large beam by means of several diagnostics. The Beam Emission Spectroscopy (BES) diagnostic provides information on the beam uniformity as well as the divergence, along a vertical and a horizontal profile. Analysis of infra-red (IR) imaging of the beam striking a calorimeter provides a 2D map of the beam power density. Three main topics will be here reported: 1) Studies of the vertical beam homogeneity often show a vertical (top/bottom) difference in terms of beam intensity and, as a consequence, in terms of divergence (i.e. different beam optics for different extracted beam currents). 2) The investigation of the possibility to measure the broad beam component (this being a small fraction of the beam with a significantly higher divergence than the majority) by means of the BES diagnostic, leads to different methods for a proper fit of the H  main Doppler peak. Different fit methods correspond to different hypotheses on the origins of the broad component itself. 3) The investigation of the stripping losses inside the extraction system aims to provide a robust method for the BES data analysis, in combination with modeling of the gas density profile along the beamline, in order to give a proper estimation of the stripping losses to be compared with predictions as extrapolated from calculation for ITER (< 10% up to the extraction grid).