Abstract. The ITER IonCyclotronHeating and Current Drive system (IC H&CD) is designed to deliver 20MW to a broad range of plasma scenarios between 40 and 55MHz, during very long pulses. It consists of two broadband equatorial port plug antennas, their pre-matching and matching systems, transmission lines, Radio Frequency (RF) Sources and High Voltage Power Supplies. The overall project schedule has been revised and agreed by ITER Council; it re-integrates the second antenna and its power supplies in construction baseline and sets the dates for progressive installation with DT phase planned in 2035. Recent progress on ICRF subsystems is reported, covering design evolution, qualification of test articles and specific R&D results in domestic agencies, suppliers, associated laboratories and IO.
Fast-flowing plasmas in supersonic and super-Alfv´enic regime are generated in combined experiments of ioncyclotron resonance heating (ICRH) and acceleration in a magnetic nozzle. During radio-frequency (RF) wave excitation in a fast-flowing plasma produced by a magnet-plasma-dynamic arcjet (MPDA), strong ioncyclotronheating is clearly observed. Thermal energy in the heated plasma is converted into flow energy in a diverging magnetic nozzle, where the magnetic moment μ is nearly kept constant. Plasma flow energy can be controlled by changing the input RF power and/or modifying the magnetic nozzle configuration. In a strongly diverging magnetic nozzle, an Alfv´en Mach number as well as ion acoustic Mach number are more than unity, that is, supersonic and super-Alfv´enic plasma flow is realized.
Abstract. This paper describes three-dimensional self-consistent numerical simulations of wave propagation in magnetoplasmas of Electron cyclotron resonance ion sources (ECRIS). Numerical results can give useful information on the distribution of the absorbed RF power and/or efficiency of RFheating, especially in the case of alternative schemes such as mode-conversion based heating scenarios. Ray-tracing approximation is allowed only for small wavelength compared to the system scale lengths: as a consequence, full-wave solutions of Maxwell-Vlasov equation must be taken into account in compact and strongly inhomogeneous ECRIS plasmas. This contribution presents a multi-scale temporal domains approach for simultaneously including RF dynamics and plasma kinetics in a “cold-plasma”, and some perspectives for “hot-plasma” implementation. The presented results rely with the attempt to establish a modal-conversion scenario of OXB-type in double frequency heating inside an ECRIS testbench.
The electron cyclotron current drive (ECCD) is studied in Heliotron J and LHD plasmas using GNET code in order to study the ECCD physics in helical configurations. The magnetic configuration dependence of ECCD is investigated in the Heliotron J plasma. It is found that the current direction is reversed in high bumpiness configuration compared with the other configurations. The ECCD in LHD is also investigated by changing electron cyclotronheating points fixing the configuration. It is found that the direction of the current reverses when we change the heating point from the ripple top to the ripple bottom.
Release of large amounts of dusts was detected with the fast framing camera in 4.5-U port just before the plasma termination. It shows that the dusts were released from a closed divertor region in the inboard side of the tours near a lower port (4.5-L) as shown in Fig. 4. After the experimental campaign, traces of exfoliation of carbon- rich mixed material layers formed on brittle iron-rich lay- ers deposited on the surface of closed divertor components (a dome structure and an inner vacuum vessel) was found on the site . In the closed divertor configuration, some divertor plates (isotropic carbon) near the lower port are arranged for blocking the plasmas in the divertor legs in order to prevent neutral particles from escaping from the divertor region through the edge of the divertor compo- nents . It is probable that the divertor plates, which face to the inboard side of the tours, enhance the deposition lay- ers on the divertor components by physical and chemical sputtering on the plates.
3. Linear theory indicates that Σ h S h is signi ﬁ cantly raised, on average, to 0.10 ± 0.01, 0.15 ± 0.02, and 0.07 ± 0.02 during the three wave events, respectively. These are elevated thresholds for EMIC wave excitation with γ / Ω p = 0.001. A positive value of Σ h S h is not always accompanied by EMIC wave activity. Even a large positive value of Σ h S h is not a suf ﬁ cient condition for the excitation of an EMIC wave (e.g., after Wave #2 in Figure 4g): Limitations in linear theory, such as simplifying the ion distribution as Maxwellian [Khazanov et al., 2007] and not considering heavy ions (He + and O + ) and nonlinear effects, may be the reason for this. The wave-period enhancements of Σ h S h result from combined variations in A hp , n hp , n e , |B|, and β ||h . Negative n e gradients might also play an important role in the generation of the
first octopole O1 (functioning and details of this device will be explained in Sec. 2.2.3). Since O1 is surrounded by a cell, which is possible to be filled with an inert gas (e.g. He or Ar), this octopole is generally used as energy quencher chamber, where excited ions can be cooled to the ground state by collisions with the inert partner, reducing some of their internal energy (i.e. vibrational or elec- tronic). The presence of an unknown amount of ions in excited states represents a complication in the interpretation of the experimental results, because of the possibility to open endothermic chan- nel avoided at the nominal collision energy. The first quadrupole Q1, connected with O1 via a set of einzel lenses, is used to mass-select just the parent ion beam under investigation and the chosen cations are subsequently directed towards the second octopole O2 ion guide through a series of cylindrical lenses, the last of which operates as collimator of the beam, reducing its divergence before the reaction chamber. The second octopole O2 is used again as guide: the ions inside are trapped in the two transversal directions and forced to move towards the end of the set-up. O2 is surrounded by a scattering cell, where the incoming parent ions collide and react with a neu- tral target (either a gas or the vapor pressure of a liquid chemical). The pressure of the neutral compound inside the reaction chamber is monitored by a pressure gauge (spinning rotor vacuum gauge system SRG2 MKS instrument ). This device works by measuring, on a magnetically- levitated spinning metal sphere, the viscosity drag, which is directly correlated to the number of particles (atoms or molecules) and therefore to the pressure. The pressure inside the scattering cell is regulated to ensure the single-collision regime inside the cell (for multiple collision effects see Appendix A).
dominated) magnet, the higher energy machine must have superconducting coils and quite high (∼6T) peak fields. Injection is achieved, with the 60 MeV/amu energy with a series of electrostatic and magnetic elements that nudge the beam into the lowest-radius orbit. RF cavities again accelerate the beam to high energies, as quickly as possible to avoid beam losses due to stripping in the residual gas of the cyclotron. Extraction is done by a stripping foil; the H + 2 ion enters the foil but emerges as two protons. These spiral inwards, but because of the highly irregular magnetic field a stripper- foil location can be found that enables the protons to exit the machine cleanly, about 180 degrees from the foil position. This stripping technique avoids the need for clean turn separation, which as mentioned earlier becomes very difficult at high energies. The only consequence is that ions hitting the foil could come from one of several turns (hence have an energy spread ∼1%). It is easy to design the extraction channel to have momentum acceptance adequate to accommodate this.
limiters are minimized. One of the strategies was to use broad-limiter antennas. The release of W was reduced by up to ~40% in a single antenna operation in experiments operated with broad-limiter antennas . The other more advanced approach is the design and use of 3-strap antennas in AUG. The basic idea of this approach is to minimize the RF image currents on the antenna limiters by balancing the dipole contributions of the central strap with that of the outer straps . In 2015, a pair of W-coated 3-strap antennas has been installed in ASDEX Upgrade. The experimental results (Fig. 4(b)) show that the W-coated 3-strap antennas lead to almost the same values of edge and core W concentration as those for the B-coated 2-strap antennas. The heating effectiveness of the W-coated 3-strap antenna is not lower than that of B-coated 2-strap  while the ICRF heating power, plasma current, total input and radiated power are almost the same in the studied discharge. Together with the previous results (Fig4. (a)), we can conclude that the use of W-coated 3-strap antennas reduces the RF-induced Fig. 4. Central and edge W concentrations during ICRF heating with (a) B-coated and W-coated 2-strap antennas; (b) B-coated 2-strap and W-coated 3-strap antennas. Figure reproduced with permission from [7, 10].
The motivation for this work is provided by observations of both space and laboratory plasmas in which ﬂows have been reported, whose shear scale length can be of the order of the ion Larmor radius or smaller. In auroral phenomena; for example, electric ﬁelds called paired electrostatic shocks have been found to exist whose scale length of variation is of the order of the ion Larmor radius [15, 16]. These electric ﬁelds cause localized cross-ﬁeld ﬂows, which excite instability with frequency and growth rate in the vicinity of the ioncyclotron frequency [17, 18]. A distinctive feature of this instability is that it can exist even when the ﬁeld-aligned current is sub-critical. Another example of a strongly sheared system is given by the dynamical evolution of the plasma sheet just prior to the on set of a magnetic sub storm. During this time, the neutral plasma sheet region becomes thin and its width becomes smaller than the ion Larmor radius.
The steady-state excitation of ICE in ITER would depend also on the structure of the corresponding Alfv´en eigenmodes, and, if excited deep inside the plasma, it is unclear that ICE could propagate to a detector without being strongly absorbed en route, for example due to the ion–ion hybrid resonances that exist in DT plasmas . However, even if detectable levels of ICE are not normally produced in steady-state conditions in ITER, it is likely that it could be used as an additional method of studying fast ion redistribution and losses resulting from MHD activity. As noted previously, ICE eigenmodes have been predicted to be localized to the outer region of the plasma, and ICE could thus be excited by a transient flux of fast particles being ejected from the plasma core due to MHD events. In addition to the correlations between ICE and fishbones observed in DIII-D , clear evidence has also been found of links between this type of emission and both sawteeth and edge localized modes in JET  and with toroidal Alfv´en eigenmodes in LHD . The detection of ICE in ITER would provide information on energetic ion behaviour supplementing that obtained using other diagnostics, such as collective Thomson scattering and γ -ray detectors. It should be noted that the heat loads in the DT phase of ITER operation will make it impossible to use a conventional fast ion loss detector unless a reciprocating drive system is used . ICE detection could provide an alternative method of studying α-particle and beam ion losses.
In recently electrostatic Kelvin-Helmoholtz instability by parallel ﬂow velocity shear in presence of inhomogeneous d.c. electric ﬁeld and only density gradient has been studied by Pandey et al.  and velocity shear ion-cylotron instability with perpendicular a.c. electric ﬁled has been also studied by Pandet et al. .
The plasma in the end region can be regarded as collisionless, but γ was lower with ICRF heating than without the heating. A high T i|| component was identified in the heated case. As mentioned in a previous numerical study [B. Lin et al., Phys. Plasmas 23, 083508 (2016)], it is experimentally suggested that γ can be reduced by the high T i|| component resulting from the ICRF heating.
sociated with the parallel electric field exhibit spiky wave forms. These quasi-static, parallel electric field structures are thought to be responsible for particle acceleration in the au- roral acceleration region. The model presented in this study could explain the generation mechanism of these parallel propagation structures which show sawtooth or spiky wave forms. Observations by FAST (as shown in Fig. 1) can be explained naturally by assuming a mixture of near parallel and near perpendicular propagating wave modes. The free energy source for these waves could be either electron/ion beams or the field-aligned currents. The present model can generate various wave-forms ranging from ioncyclotron to ion acoustic modes, depending on the initial driver and the Mach number. We have not considered growth or damping of these waves. It would be interesting to study the stability of these structures. We have neglected the ion temperature ef- fects and treated ions as cold fluid. The inclusion of finite ion temperature will give rise to a dispersive effect, which may tend to broaden the structures. Therefore, the present model can be applied where T e T i , such as the auroral accel-
ion current was obtained is mapped with blue color. On the other hand, the region where electron current was mea- sured is mapped with red color. To evaluate the ion tem- perature, the inflow of bulk electrons should be suppressed over the range of the collector potential sweep. However, as shown in Fig. 5, the electron current is clearly measured when h is smaller than 1.0 mm. On the other hand, elec- tron current is almost zero when h is larger than 1.0 mm. Therefore, Fig. 5 indicates that in the case of V oﬀ = − 20 V, h > 1 . 0 mm is suitable for evaluating the ion temperature for these plasma parameters.
where all the protons are assumed to be hot and electrons to be cold. The calculations are performed for generalized distribution function reducible to an-isotropic Maxwellian plasma for j = 0 and loss cone distribution for j = 1 Figure 1 describes variation of the normalized growth rate and real frequency with normalized wave number. Fixed plasma parameters are defined in graph caption. The A.C. frequency affects the growth rate significantly. The maxima shifts towards lower values of k k as the frequency changes from 4 Hz to 12 Hz for a anisotropic maxwellian plasma at j = 0, for j = 1 the nature of the variations are similar but the growth rate increases by an order of magnitude and covers wider spectrum of k k . The resonance frequency is influenced by the frequency of the A.C. signal as long as it is less than the proton gyro frequency comparing well with AMPTE/CCE data of 1994 . However wave excitation takes place only when the proton perpendicular and parallel temperature ratio is more than or equal to 1.5.
High-ion temperature experiments in the Large Helical Device (LHD) are categorized in terms of the heating scenarios that are closely related to the development of neutral beam injection (NBI) systems. Although high- energy tangential negative-NBI heating has greatly contributed to extending the plasma parameter regime in LHD, the ion temperature does not increase because the electron heating is dominant with negative-NBIs. In the high-Z discharges, it was demonstrated that the ion temperature increased with an increasing ionheating power and achieved 13.5 keV with the negative-NBIs. Low-energy perpendicular positive-NBIs were installed for the ionheating, and the ion temperature was increased to more than 7 keV in hydrogen discharges. In the high-ion temperature plasmas, an ion internal transport barrier (ion ITB) was formed, and the impurity hole was observed in the core. Long-pulse ioncyclotron range of frequency heating (ICH)/electron cyclotron resonance heating (ECRH) helium discharges are eﬀective for wall conditioning, leading to a decrease in the neutral density and a peaked density profile. Consequently, the ionheating eﬃciency increases in the core, and the central T i is raised
A key issue in spherical tokamak (ST) research is to develop a method for initiating the plasma current and forming an ST configuration without using the central Ohmic solenoid. Various scenarios have been studied on ST devices. Electron cyclotronheating (ECH) start-up is one such method, and many experiments on ST devices have been carried out [1–13], but the current drive mech- anism is still not clearly identified. Moreover, the mech- anism of current jump, which is believed to occur when closed flux surfaces are formed, is not completely under- stood. A major diﬃculty in these studies is that the power deposition profile and the phase space region of wave- particle (electron) coupling are unknown. In such a situ- ation, experimental comparison of diﬀerent injection sce- narios (i.e., O-mode vs. X-mode) and various parameter dependences are crucial. There is no systematic experi- mental comparison between O-mode and X-mode polar- izations. The X-mode is expected to be more eﬀective than the O-mode, because X-mode absorption is eﬃcient even in low density plasmas and can be converted to the electron Bernstein wave (EBW), which can heat the plasma even at densities higher than the cuto ﬀ density.
Since the pioneer work of Ginzburg and Zheleznyakov (1958), Langmuir waves have been intensively studied in re- lation with Type III radio emission. Indeed, these electro- magnetic radio emissions at the plasma frequency, or twice of this frequency, are thought to result from wave coupling implying ion acoustic waves and Langmuir waves. From the observational point of view, numerous works have been de- voted to test the validity of this model (Lin et al., 1981, 1986; Gurnett et al., 1993; Thejappa et al., 1993; Cairns, 1995; Thejappa et al., 2003). New evidence of such wave cou- pling processes have been recently provided from the analy- sis of waveform data from the STEREO/WAVES experiment (Henri et al., 2009). Since the phase of the waves are avail- able from such measurements, it is possible to test the energy and momentum conservation laws induced by such coupling.
Although by no means the most commonly used method of additional heating, in many ways ECRH is the simplest. It is possible to heat electrons without the need for a large antenna, and without the EC wave having to cross an evanescent region, avoiding problems with the plasma edge. When waves are incident perpendicular to the background magnetic field they may travel in two modes viz the ordinary (O) and the extraordinary (X) mode. The O mode is a transverse wave whose electric field oscillates along the direction of the magnetic field while the X mode is a mixed transverse and longitudinal wave which has no electric field component in the direction of the magnetic field. The X mode's field can be composed into two rotating electric field components E+ and E_, the latter rotating in the same sense as the electrons (and the former in the same sense as the ions). As lE.I-IE+lv^^/c near the cyclotron frequency (Stix, 1962), V|-|j being the thermal velocity and c the velocity of light, the E_ component vanishes in the cold plasma approximation. Furthermore for electron cyclotronheating the E+ component is ignored as it rotates in the opposite sense to the electrons. It is therefore clear that the X mode cannot heat electrons at the electron cyclotron frequency in the cold plasma approximation. However in the hot theory E_ does not vanish and heating may take place. Charged particles in uniform magnetic fields move by spiralling around the magnetic field lines at the electron cyclotron frequency. When the cyclotron frequency matches that of electromagnetic waves the particle may interact strongly with the wave, either receiving or giving up energy from or to the wave and enhancing its perpendicular velocity. Resonance is also possible at harmonics of the cyclotron frequency, and, for the X mode, at the upper hybrid frequency cûuH^=œpe^+nce^. (This resonance does exist in a cold plasma).