1 OV/P-01

Overview of FTU Results

P. Buratti1, on behalf of FTU Team*, Z.O. Guimarães-Filho2 and E.Lazzaro3

1Associazione EURATOM-ENEA, CR ENEA-Frascati, C.P. 65, 00044 Frascati (Roma) Italy2Aix-Marseille Univ., IIFS-PIIM, UMR 7345, F-13397 Marseille, France3Associazione Euratom-ENEA sulla Fusione, IFP-CNR, Via R. Cozzi 53, 20125 Milano, Italy

E-mail contact of main author: paolo.buratti@enea.it

Abstract. Since the 2010 IAEA-FEC Conference, FTU has exploited improvements in cleaning procedures and in the density control system to complete a systematic exploration of access to high-density conditions in a wide range of plasma currents and magnetic fields. The line-averaged densities at the disruptive limit increased more than linearly with the toroidal field, while no dependence on plasma current was found, in fact the maximum density of 4.3×1020 m-3 was reached at B = 8 T even at the minimum current of 0.5 MA, corresponding to twice the Greenwald limit. The lack of plasma current dependence is due to the increase of density peaking with the safety factor. Experiments with the 140 GHz ECRH system were focused on the sawtooth period control and on the commissioning of the new launcher with real-time-steering capability that will act as the front-end actuator of a real time system for sawtooth period control and tearing modes stabilization. Various ECRH and ECCD modulation schemes have been used; with the fastest one, the sawtooth period synchronized with the 8 ms modulation period. The observed period variations were simulated using the JETTO code with a critical shear model for the crash trigger. The new launcher is of the plug-in type, allowing quick insertion and connection to the transmission line. Both beam characteristics and steering speed were in line with design expectation. Experimental results on the connection between improved coupling of lower hybrid waves in high-density plasmas and reduced wave spectral broadening have been interpreted by fully kinetic, non-linear model calculations. The effect of wall conditioning by lithium on MHD activities has been studied by comparing discharges with and without lithium conditioning at low-q and at the density limit. In both cases lithium conditioning has the same effect of reducing MHD modes associated with edge cooling by light impurities as a careful wall preparation. Experiments with the liquid lithium limiter inserted in the SOL, which have shown the formation of a radiative belt that acts as a virtual toroidal limiter, have been interpreted by the edge code TECXY as an effect of strong radiation from Li+ ions in non-coronal equilibrium.

1. Introduction

Since the 2010 IAEA-FEC Conference, FTU plasma operations have been largely ohmic, due to some faults in the auxiliary heating power supplies. A new density control system and an improved cleaning procedure ensuring very clean plasma conditions and rapid disruption recovery were successfully exploited for a systematic exploration of the density limit. The experimental sessions with 140 GHz EC system were devoted to experiments on sawtooth control and to commissioning of the new launcher with real-time steering capability.

2. Density Limit Dependence on the Toroidal Magnetic Field

The capability of predicting the maximum density that can be achieved in a tokamak plasma is of crucial importance to establish the ultimate performance of a fusion power plant. The Greenwald limit [1], commonly used as an empirical scaling law, predicts that the maximum line-averaged density (expressed in units of 1020 particles m-3) is given by nG = Ip /π a2 for an elliptical plasma cross section with minor radius a (in m), where Ip (in MA) is the plasma current. The Greenwald limit has been exceeded is several experiments that produced peaked density profiles (e.g. due to pellet injection); in particular, spontaneous density peaking leading to densities above the Greenwald limit has been recently observed in FTU discharges when using the Liquid Lithium Limiter at high values of the edge safety factor, thus suggesting a possible dependence of the density limit on the toroidal magnetic field [2, 3]. Following this indication, the high density domain has been systematically explored in a wide

* See Appendix

2 OV/P-01

range of plasma current (Ip = 0.5, 0.7, 0.9 MA) and toroidal field (BT = 4.0, 5.2, 6.0, 7.2, 8.0 T) values [4]. For each (Ip, BT) configuration, the gas flow waveform has been carefully designed to meet the disruption for density limit with slowly ramping density during the current flat-top. The stainless steel vacuum wall and the molybdenum toroidal limiter were conditioned with boron coating at the beginning of the experimental campaign, resulting in clean machine conditions with Zeff = 1.5-1.0. The electron density was measured by a high-resolution CO2/CO fast-scanning interferometer along more than 30 vertical chords; in particular, the axial and peripheral line-averaged densities reported in the following refer to a chord that crosses the plasma axis and to one with 0.8a impact parameter respectively. Time traces for these chords are shown in FIG.1, together with central (r/a = 0) and peripheral (r/a = 0.8) electron temperature and Dα emission. At t = 0.7 s (vertical dashed line), a dramatic increase of Dα emission and a decrease of peripheral temperature mark the onset of a MARFE (Multifaceted Asymmetric Radiation From the Edge); the density peaking increases afterwards, as can be seen from the increasing separation between axial and peripheral density time traces.

2.1. Comparison with the Greenwald Density Limit

The Hugill plot of disruption (solid symbols) and MARFE onset (open symbols) conditions for the complete dataset is shown in FIG.2a. For the circular FTU cross section, the inverse cylindrical edge safety factor 1/qcyl = RIp /(5a2BT) is proportional to the average plasma current density, and the Greenwald limit corresponds to a straight line with π /5 slope (as shown by the solid line in FIG.2a). Disruption data do not follow the Greenwald scaling, which is exceeded at high qcyl (points below the solid line), while it is not reached at low qcyl (points above the solid line). By contrast, MARFE onset points (open symbols) are well ordered around a straight line at 40% of the Greenwald limit (dashed line in FIG.2a).

FIG. 1. Time traces of some relevant quantities for a specific discharge with Ip = 0.7 MA and BT = 7.2 T. (a) Line-averaged density at r/a = 0 (green) and 0.8 (blue). (b) Electron temperature at r/a = 0 (green) and 0.8 (blue). (c) Dα emission. The dashed vertical line marks the MARFE onset.

FIG. 2. (a) Hugill plot of disruptions (solid symbols) and MARFE onsets (open symbols). BT / Ip are coded by symbol shape/colour. The solid line corresponds to n = nG, the dashed one to n = 0.4·nG. (b) Modified Hugill plot with the peripheral line-averaged density instead of the axial one: A Greenwald-like scaling is recovered at n4/5 = 0.36·nG (solid line).

3 OV/P-01

2.2. The Role of Density Peaking

A closer inspection of density profiles, which have been reconstructed with a particular care to avoid interference from MARFE density oscillations [2, 4, 5], reveals that high qcyl discharges that disrupt at densities above the Greenwald limit attain large (2–2.7) values of the density peaking, which is defined as central to volume-averaged density ratio (FIG.3). This indicates that the peripheral line-averaged density, rather than the axial one, may be the limiting factor. As a matter of fact, both MARFE onset and disruption data get ordered along straight lines in a modified Hugill plot considering the peripheral density instead of the axial one, as shown in FIG.2b.

The observation that density profile peaking starts just after MARFE formation suggests that the peaking mechanism should be searched in the MARFE properties, such as increased neutral particle penetration to the plasma core [4-6].

2.3. Magnetic Field Dependence

The substantial insensitivity to plasma current (as confirmed by FIG.4a) and the strong dependence on qcyl translate into a toroidal field dependence of the density limit. FIG4b shows that the achievable density scales more than linearly with BT, the best fit exponent being 1.5±0.1 [4]. The strong BT dependence allowed attaining line-average density as high as 4.3×1020 m-3 at BT = 8 T and Ip = 0.5 MA, which corresponds to twice the Greenwald limit.

2.4. Specific MARFE Studies

The toroidally symmetric belt of high density, strongly radiating plasma that constitutes the MARFE has been studied in FTU by means of a fast-framing camera. The widely accepted cause of this phenomenon is a thermal instability of the plasma edge, which gives rise to a poloidally narrow belt of cold plasma that tends to be localized at the high field side due to poloidal asymmetry of the heat flux from the plasma core. Cyclic vertical oscillations of the belt between the edges of the toroidal limiter have been observed, depending on density and plasma current conditions, which produce large oscillations on the interferometer signals. A preliminary attempt to explain these limit-cycle-like oscillations in terms of power balance at the upper and lower edges of the radiating plasma belt has been presented in [7]

FIG. 3. Density peaking at MARFE onset (open symbols) and at disruption (solid symbols) versus qcyl. BT and Ip are coded like in FIG.2 (see inserts). The two lines are only guides for the eye.

FIG. 4. Axial line-averaged density at the disruption versus (a) the average plasma current density and (b) the toroidal field. BT

and Ip are coded like in FIG.2 (see inserts).

4 OV/P-01

3. Sawtooth Control by ECRH and ECCD

The control of sawtooth (ST) activity is a critical issue, since big crashes after long periods of magnetic energy accumulation produce magnetic islands that can exceed the critical value above which metastable tearing modes are triggered. On the other hand, the ST activity is useful to prevent impurity accumulation in the plasma core. The goal of these studies is then the attainment of sufficiently fine control of the ST period in order to keep seed islands below the critical amplitude, while preserving impurity ejection. A promising method for ST period control, which has the further advantage of reducing the required energy budget, is the use of modulated Electron Cyclotron Heating (ECH) and Current Drive (ECCD). This method has been investigated in FTU using 500 ms of repetitive pulses from two 140 GHz gyrotrons with up to 0.8 MW total input power [8]. Three modulation schemes have been used, i) 40 ms on / 10 ms off, i.e. 20 Hz, 80% duty cycle (d.c.), ii) 10 ms on / 40 ms off (20 Hz, 20% d.c.), iii) 4 ms on / 4 ms off (125 Hz, 50% d.c.). In all cases the EC power was injected during repetitive discharges with 0.5 MA plasma current and 0.6×1020 m-3 line averaged density, with ohmic sawtooth period τST_OH = 6.4 ms. The magnetic field was ramped from 5.1 T to 5.9 T during EC power injection in order to scan the EC deposition location across the ST inversion radius.

The first modulation scheme, being the EC-on phase (40 ms) much longer than τST_OH , essentially gives the response to continuous power injection. In this case the ST period (τST) settles at 4 ms (38% below the ohmic value) as long as the EC resonance is inside the inversion radius (rinv), independent of the precise EC deposition radius (rdep), while for rdep > rinv (i.e. BT > 5.5 T), τST increases to 10 ms (FIG.5a). The 20 Hz / 10 ms on scheme gives, as expected, irregular variations of τST, depending on the lag between the pulse front and the previous ohmic crash. The 125 Hz, 50% d.c. modulation gives the unforeseen result that for BT ≈ 5.3 T (deposition inside the inversion radius), τST increases and locks to the 8 ms modulation period (FIG.5b). The observed locking has been reproduced by transport simulations supplemented by a critical shear model for the sawtooth trigger and by the Kadomtsev reconnection model [9].

FIG. 5. Sawtooth period during 5.1-5.9 T BT scans (symbols) and EC power waveform (gray lines). Triangles (red dots) refer to ST periods at start (end) of EC pulses. BT = 5.5 T (rdep = rinv) at t = 0.8 s. (a) 20 Hz, 80% d.c. modulation. (b) 125 Hz, 50% d.c. modulation.

5 OV/P-01

4. Development of Tools for Feedback MHD Control

EC power deposition at or near magnetic reconnection sites has proven to be an effective tool to modify the stability characteristics of different kinds of MHD activities. In particular, tearing modes that form magnetic island chains by reconnecting field lines around rational surfaces such as q = 2, 3/2, 4/3, can be mitigated or suppressed provided that the EC deposition volume is sufficiently close to the island location. The tasks of a real time control (RTC) system is then to detect the mode and its location, switch on the EC power and aim it at the mode location. The same system can be used for sawtooth period control, by imposing that modulated EC power is deposited at a prescribed distance from the q=1 radius. The main objectives of the RTC system developed for FTU are mitigation of tearing modes with 3/2 and 2/1 poloidal/toroidal mode numbers, control of the sawtooth period, and control of heating in overdense plasmas. The implementations related to the first objective will be reported in this section.

4.1. The Real-Time Control System

The MHD control system detects the presence of modes, determines mode numbers and island locations, and provides reference toroidal and poloidal launching angles for the EC launcher, together with trigger signals for the gyrotrons that feed the launcher. The key plant inputs are temperature oscillations from a 12-channels ECE ploychromator, magnetic oscillations from a set of Mirnov coils, two ECH power monitors and plasma equilibrium signals [10]. Raw data are elaborated using algorithms developed in order to detect the presence of the MHD activity, to track its radial position, to determine its poloidal and toroidal mode numbers and to timely switch on/off the EC power. This set of signals undergoes a two-stage elaboration through the control chain. The open-source MARTe (Multi-threaded Application Real-Time executor) framework is used [11].

In the first stage ECE, Mirnov coil and ECH signals are acquired at 20 kHz sampling rate, pre-processed and cross-correlated. ECH signals are processed to obtain the “plasma thermal image” (PTI), which accounts for the electron temperature response to modulated heating. Each ECE channel is calibrated and multiplexed by means of a low-pass filter and two band-pass filters, one optimized for the evaluation of the deposition radius by ECE-PTI correlation during ad-hoc modulation of the ECH power, and the other for mode detection and location by ECE-ECE and ECE-Mirnov coil correlation. Data from a set of Mirnov coils at different poloidal and toroidal locations are also processed by a singular value decomposition (SVD) code in order to obtain early mode detection and a real-time estimate of the toroidal (n) and poloidal (m) mode numbers [12]. All these actions are implemented in a fast thread composed by several GAMs (Generic Application Modules). The execution time for the codes running on the fast thread is in the range 1-50 µs.

In the second stage, the results from the first stage are processed along with data from the FTU main control system by a MARTe real time thread at the lower rate of 1 KHz. Island locations (rMHD) are identified between pairs of channels in which ECE-Mirnov coil cross-correlation changes sign. In principle, the same information can be extracted from ECE-ECE cross-correlations, but the latter are disturbed by the occurrence of sawtooth crashes; on the other hand, they can be conveniently employed for detecting the q = 1 radius in sawtooth control experiments. A parabolic fit to the maximum of the ECE-PTI cross-correlation profile is used to estimate of the actual EC deposition radius (rDEP).

Independent “a priori” estimates of rMHD and rDEP are obtained by means of a fast equilibrium reconstruction (FASTEQ) that employs a best fit procedure of plasma boundary and density

6 OV/P-01

signals (from the FTU main control system) and real-time temperature profiles (from low-pass filtered ECE data) to a large database of FTU equilibria [13]. The resulting q-profile, together with the estimated mode numbers, gives the island location, while a fast EC beam ray tracing gives the deposition radius. The a priori estimates are used to optimize the overall time response of the system and to supply a reliable status for the actuator even if the diagnostic information is not continuously available [14]. The switching from a-priori-based control to measurements-based control and tracking of EC launching angles is handled by the final launcher control GAM [15]. Comprehensive system tests are reported in [16, 17].

4.2. Launcher Tests

The launcher can inject into the plasma two independent high-power beams (each with 0.4 MW power, 140 GHz frequency, 0.5 s maximum duration). A front steering option has been adopted that gives a wide poloidal range, covering more than 80% of the plasma cross-section, and large toroidal angles as well (±40°), which are desirable for ECCD and for overdense plasma heating. The steering mirrors have small inertia, in order to guarantee sweeping in the poloidal direction by ±25°, with 1°/10 ms angular speed. The spot size at the plasma center can be varied from 19 to 26 mm by moving a shaping mirror.

The launcher plugged into FTU vessel has been characterized using a low power source inserted into ECRH transmission line and a thermopile as detector. The dynamic performance under plasma operation has subsequently been tested by performing pre-programmed angular scans during plasma discharges. The maximum deviation between reference and measured angles was 0.3° (poloidal) and 0.06° (toroidal). The power deposition location during angular scans was estimated by the real-time ray tracing (FIG.6) and ECE-PTI cross-correlation (FIG.7) GAMs of the control system; the results agreed with off-line beam ray tracing calculations to within 2 cm [18].

FIG. 6. Equilibrium contours from FASTEQ and traced rays for two different poloidal launch angles. The dashed line shows EC resonance location. Symbols represent equatorial projections of the deposition volumes.

FIG. 7. Profiles of cross-correlations between temperature oscillations and plasma thermal image for two different poloidal launch angles. Symbols represent equatorial projections of the deposition volumes from ray tracing and FASTEQ.

7 OV/P-01

5. LHCD Studies

No new LHCD experiments have been performed since the 2010 IAEA-FEC Conference, because of faults in the electrical supplies; however the 8 GHz plant has been recently recommissioned. New analyses of previously collected data are presented in this section.

5.1. Progress in Understanding LHCD at High Density

Effective electron acceleration by LH waves that propagate to the core of high-density plasmas has been achieved in FTU [19] following the theoretical indication that increasing electron temperature at the plasma edge should lead to reduced spectral broadening of the injected waves, and then to improved coupling to the plasma core [20]. Three different regimes can be distinguished on the basis of LH effects, namely i) the low-density regime, in which current-drive effects are regularly observed; ii) the “standard” high-density regime (line-averaged ne_av > 1020 m-3 typical), in which the hard X-ray (HXR) emission by LH-accelerated electrons disappears and the coupled LH power appears to be dissipated at the plasma edge; iii) the “new” high-density regime, in which the HXR signature of core electron acceleration (as measured by a camera with spatial and energy resolution) is recovered. Observations on the RF spectral broadening as measured by a loop antenna in the three regimes and results of modelling based on parametric instabilities of the launched waves are summarized in the following. Comparison with results from other tokamaks and consideration of alternative dissipation mechanisms based on non resonant collisional absorption or on linear scattering by edge density fluctuations can be found in [21-23].

Parametric instabilities (PI) driven by coupling of quasi-electrostatic waves in the domain of the 8 GHz pump wave with ion-sound plasma quasi modes produce a frequency broadening of the pump wave that can be observed in the spectra of signals collected by RF loop antennas, and a corresponding broadening of the parallel wave refractive index (n//) spectrum that is essential for determining propagation and damping properties of the coupled LH power. The spectral broadening has been calculated by solving the full non linear mode coupling equations with a fully kinetic plasma model [20].

In the low-density regime, the spectral broadening is moderate, a few MHz at -35 dB below the pump power peak value at 8 GHz. Moreover, the spectral broadening is symmetric. PI calculations result in sidebands carrying about 10% of the injected power, with n// distributed from the nominal antenna value (n//0 ≈ 1.8) to a cut-off value n//cut-off ≈ 3-5 [20]. This broadening of the n// spectrum is sufficient to enhance the quasi-linear interaction of the whole coupled LH power with plasma electrons. Due to their small shift from the injected power frequency, both the upper and the lower sidebands are LH propagating modes, in agreement with the observed symmetry of the measured frequency spectrum.

LHCD effects disappear and HXR are below the noise level in the “standard” high-density regime, despite the full accessibility to the plasma centre of the nominal antenna n//. Measured frequency spectra become broader (∆f ≈ 15 MHz) and markedly asymmetric (see FIG.8). According to PI modelling, the key factor that determines such an increased spectral broadening is the presence of a sufficiently wide region of low (< 30 eV) electron temperature [20]. A strong PI regime is entered in this condition and the frequency range of lower sidebands broadens, meanwhile upper sideband waves with large frequency shift become evanescent. The cut-off parallel index increases to n//cut-off ≈ 30 and the computed LH power deposition is localized at the very plasma edge, consistent with the absence of LHCD effects. The estimated energy of accelerated electrons is < 10 keV, below the HXR range of detection.

8 OV/P-01

FIG. 8. Frequency spectra of the RF signal collected by a loop antenna in the standard high-density regime (red) and in the “new” one with higher temperature at the plasma periphery (green). Zero frequency corresponds to the (8 GHz) injected frequency.

FIG. 9. Measured and calculated reflection coefficients for a row of a conventional FTU grill launcher with 75° phasing.

The “new” high-density regime has been obtained in FTU by increasing the electron temperature at the plasma periphery (Te ≈ 0.2 keV at r/a = 0.8) with the plasma edge conditioned by the insertion of a liquid lithium limiter in the scrape-off layer. The spectral broadening as measured by the loop antenna (FIG.8) is strongly reduced with respect to the “standard” regime and HXR from the plasma core (r/a ≈ 0.4) are detected. A strongly reduced PI-produced spectral broadening is expected in these experimental conditions, owing to the fact that the low temperature (Te < 30 eV) region is limited to a narrow layer at the antenna-plasma interface and then the strong PI regime is prevented by convective losses. Assuming that the peak of the launched spectrum should be up-shifted to n// = 5 and broadened by PI up to n//1 = 7.5, numerical simulation also shows that LH waves accelerate plasma electrons, mainly around r/a ≈ 0.4, at energies in the 40 keV÷80 keV range, consistently with data from the HXR camera

5.2. Modeling of LHCD Coupling

An exhaustive benchmark of TOPLHA and GRILL3D-U LH coupling codes has been performed with reference to a row (12 phased waveguides) of the FTU conventional grill launcher, which is fully monitored in terns of direct and reflected power, and equipped with Langmuir probes for local density measurements. Very good agreement between predicted and measured data was found for various launched spectra in a wide range of edge densities [24]. Reflection coefficients as measured and as calculated by TOPLHA and GRILL3D-U codes for 75° grill phasing are shown in FIG.9 together with results from the Brambilla code.

6. New Diagnostics

Refractometry: A dual-frequency pulsed refractometer has been developed and tested on FTU [25]. The chord-averaged plasma refraction index has been measured in a double-pass scheme with a reflection at the inner wall by launching from the low field side waves at 51.5 GHz and 60.5 with extraordinary polarization. The time-of-flight measurement in the pulsed operating mode allows to avoid phase jumps problems and to get rid of parasitic reflections. The launched wave frequencies lie in the transparency window between the lower

9 OV/P-01

cut-off for the extraordinary wave and the electron cyclotron resonance. In this window the time delay is a functional of the density profile that reduces to a simple line-average at frequencies well above the cut-off; in principle, simultaneous measurements at different distances from the cut-off can be used to retrieve details of the density profile.

The dual-frequency system has been developed as a first step towards a multi-frequency diagnostic. In order to test the diagnostic capability, a simple analytical profile n(r) = n0(0.2 + 0.8(1-r2/a2)α) has been assumed and the n0 and α parameters have been determined by a fitting procedure to the two measured delays [25]. The line-averaged density and the profile peaking factor (i.e. the peak to volume-average ratio) have then been calculated from the fitted profile and compared to the ones measured by the high-resolution CO2/CO interferometer, see FIG.10. The two diagnostics are in excellent agreement for densities above 5×1019 m-3. At lower densities the profile peaking estimate fails, due to the fact that the cut-off frequency is too far below the probing frequencies.

Electro-Optical Probe for Dust Detection: A new electro-optical (EO) probe diagnostic for the detection of hypervelocity dust impacts in tokamaks has been developed in the frame of a collaboration with IFP-CNR Milan, KTH-Stockholm and the Universities of Naples and Molise. The EO probe has been successfully tested simulating the dust impact by means of a laser beam [26]. The EO probe has been recently installed in FTU.

7. Liquid Lithium Limiter Experiments

Edge Cooling: Progressive insertion of the liquid lithium limiter (LLL) into the shadow of the Mo toroidal limiter resulted in the formation of a radiative belt that acts as a virtual toroidal limiter [27]. The observed decrease of SOL electron temperature has been modeled by the edge code TECXY as an effect of strong radiation from Li+ ions in non-coronal equilibrium. Runaway plasma cooling has been observed for LLL front-end temperature above 550°C.

MHD Effects of Lithium Conditioning: Comparative studies on the MHD behavior with and without wall conditioning by lithium have been done in low-qa and in density-limit discharges. In both cases lithium conditioning resulted as effective as a careful wall preparation by baking, glow discharge and boronization as regards the reduction of MHD activity associated with edge cooling by hydrocarbon impurities [28].

8. Theory and Modeling of MHD Observations

Analysis of Fishbone-Like Modes: A comparative study of the MHD modes identified as electron fishbone-like modes in lower hybrid current drive discharges in FTU and Tore Supra has been carried on. A transition between continuous and bursting evolution of modes with n = 1 modes has been found with increasing LH power in FTU. The frequency evolution during bursts is consistent with a non adiabatic downward chirping produced by mode-particle pumping [29]. Inverse cascades from n = 4 to n = 1 slowly evolving lines have been found in Tore Supra. The different characteristics of frequency evolution in the two machines are

FIG. 10. (a) Comparison between line-averaged density time traces from the laser interferometer (solid line) and from the new time-of-flight refractometer (dots). (b) Comparison between density peaking factors from the two diagnostics.

10 OV/P-01

thought to be due to different LH power densities (about ten times larger in FTU) and different q-profiles. [30].

Electron Fisbones Simulation: The XHMGC hybrid MHD-Gyrokinetic code (see [31] and references therein) has been used to simulate, for the first time, energetic electron driven fishbone instabilities (e-fishbones) [31]. The XHMGC code includes both thermal ion collisionless response to low-frequency Alfvénic modes and finite parallel electric field due to parallel thermal electron pressure gradient. Meanwhile, XHMGC is capable of treating up to three independent particle populations kinetically, assuming different equilibrium distribution functions for each species. A FTU-like reference equilibrium has been considered, with a slightly inverted q profile, and strongly anisotropic Maxwellian distribution function for the energetic electrons (TEe = 50 KeV); thermal ions (Ti0 = 2 keV) are instead described by isotropic Maxwellian distribution function: in this way, both resonant excitation by supra-thermal electrons as well as thermal ion Landau damping and finite compressibility are accounted for. An unstable mode driven by the energetic electrons has been observed in simulations with toroidal mode number n = 1, provided the energetic electron density is above a critical threshold; this mode disappears if the trapped energetic electrons response is artificially suppressed. As expected, the mode rotates in the direction of the diamagnetic velocity of supra-thermal electrons. The power exchange with energetic electrons and bulk ions is shown in FIG.11 for the radial shell inside the minimum q radius, where the power exchange between particles and mode is peaked: trapped energetic electrons drive the mode, whereas co- and counter-passing ions provide collisionless damping.

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

0 2 4 6

-1

0

1

all particles

u^

µ^-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

0 2 4 6-4

-2

0

2

4

all particles

u^

µ^

FIG. 11. Wave-particle power exchange in the normalized parallel velocity and normalized magnetic momentum space, for energetic electrons (left) and bulk ions (right). Green to red areas correspond to driving contributions, cyan to violet to damping ones. Solid lines show the trapped/circulating boundary.

Acknowledgements: This work was supported by EURATOM and carried out within the framework of the European Fusion Development Agreement. The views and opinions expressed herein do not necessarily reflect those of the European Commission.

References

[1] GREENWALD, M., “Density limits in toroidal plasmas”, Plasma Phys. Contr. Fusion 44 (2002) R27.

11 OV/P-01

[2] TUDISCO, O., et al., “Peaked density profiles and MHD activity on FTU in lithium dominated discharges”, Fusion Engineering and Design 85 (2010) 902.

[3] MAZZITELLI, G., et al., “FTU results with a liquid lithium limiter”, Nucl. Fusion 51 (2011) 073006.

[4] PUCELLA, G., et al., “Dependence of the density limit on the toroidal magnetic field on FTU”, submitted to Nucl. Fusion (2012).

[5] PUCELLA, G., et al., “Density Limit Experiments on FTU”, EX/P8-03, this Conf.

[6] TUDISCO, O., et al., “Effect of magnetic field on density limit in FTU” 39th EPS Conference (Stockholm 2012), ECA 36F (ISBN 2-914771-79-7) paper P4.091.

[7] ZENG, L., et al., “Density Fluctuations Induced by MARFE in FTU”, 20th International Conference on Plasma Surface Interactions (Aachen 2012), paper P2-84.

[8] NOWAK, S., et al., “Sawteeth (de)stabilization by ECH and ECCD in FTU” ”, Proc. 38th EPS Conference on Plasma Physics (Strasbourg 2011), ECA 35G, P5.085 (2011).

[9] NOWAK, S., et al., “Control of Sawteeth Periods by Pulsed ECH and ECCD in FTU Tokamak”, EX-P4-20, this Conference.

[10] ALESSI, E., et al, “Fast elaboration of diagnostic data for real time control in FTU tokamak”, EPJ Web of Conferences 32 (2012) 02015.

[11] BONCAGNI, L., “Progress in the migration towards the real-time framework MARTe at the FTU tokamak”, Fus. Eng. Design 86 (2011) 1061.

[12] MARCHETTO, C., et al., “Application of SVD algorithm to a set of Real Time Mirnov coil signals in FTU tokamak”, 39th EPS Conference on Plasma Physics (Stockholm 2012), ECA 36F, P2.064 (2012).

[13] NOWAK, S., et al., “Fast equilibrium reconstruction (FASTEQ) for the control in real time of MHD instabilities in FTU tokamak”, Proc. of the 38th EPS Conference on Plasma Physics (Strasbourg 2011), ECA 35G, P4.085 (2011).

[14] D’ANTONA, G., et al., “The MHD Control System for the FTU Tokamak”, IEEE Trans. on Nuclear Science, 58 (2011) 1503.

[15] GALPERTI, C,. et al., “Specifications and implementation of the RT MHD control system for the EC launcher of FTU”, EPJ Web of Conferences 32 (2012) 04015.

[16] SOZZI, C., et al., “A Real-Time System for Data Acquisition, Elaboration and Actuator's Control for MHD Instabilities in the FTU Tokamak”, Submitted to IEEE.

[17] SOZZI, C., et al., “First operations of the real-time ECRH/ECCD system for control of magnetohydrodynamics instabilities in the FTU tokamak”, EX/P6-04, this Conference.

[18] MORO, A., et al., “In vessel characterization and first power tests on plasma of the Real-Time controllable EC launcher on FTU Tokamak”, EPJ Web of Conferences 32 (2012) 02018.

[19] CESARIO, R., et al., “Current drive at plasma densities required for thermonuclear reactors”, Nature Commun. 1 (2010) 55.

[20] CESARIO, R., et al., “Spectral Broadening and Lower Hybrid Current Drive in High Density Tokamak Plasmas”, EX/P6-29, this Conference.

[21] TUCCILLO, A.A., et al., “On the Use of Lower Hybrid Waves at ITER Relevant Density”, ITR/P1-09, this Conference.

[22] BARBATO, E., et al., “The role of non-resonant collision dissipation of Lower Hybrid current driven plasmas”, Nucl. Fusion 51(2011) 103032.

12 OV/P-01

[23] PERICOLI-RIDOLFINI, V., et al., “LHCD efficiency in tokamaks and wave scattering by density fluctuation at the plasma edge”, Nucl. Fusion 51 (2011) 113023.

[24] CECCUZZI, S., et al., “Validation of Lower Hybrid coupling codes (Brambilla, GRILL3D-U,. TOPLHA) with the FTU conventional grill”, Proc. of the 38th EPS Conference on Plasma Physics (Strasbourg 2011), ECA 35G, P1.097 (2011).

[25] PETROV, V.G., et al., “First Results from Plasma Density Measurements in the FTU Tokamak by means of a Two-Frequency Pulsed Time-of-Flight Refractometer”, Plasma Physics Reports, 38 (2012) 343.

[26] DE ANGELI, M., et al., “Simultaneous electrical and optical detection of expanding dense partially ionized vapour clouds”, Rev. Sci. Instrum. 82 (2011) 106101.

[27] APICELLA, M.L., et al., “Lithization of the FTU tokamak with a critical amount of lithium injection” Plasma Phys. Control. Fusion 54 (2012) 035001.

[28] BOTRUGNO, A., et al., “” Proc. 38th EPS Conf. on Plasma Physics, (Strasbourg 2011), ECA 35G P4.110 (2011).

[29] GUIMARÃES-FILHO, Z.O., et al, “Electron fishbones in FTU and Tore Supra Tokamaks”, Nucl. Fusion 52 (2012) 094009.

[30] GUIMARÃES-FILHO, Z.O., et al, “Electron fishbones in LHCD plasmas on FTU and Tore Supra”, EX/P6-07, this Conference.

[31] VLAD, G., et al., “Electron Fishbone Simulations in FTU-like Equilibria Using XHMGC”, TH/P6-03, this Conference.

Appendix: Members of the FTU Team

E. Alessi1, L. Amicucci4, B. Angelini2, M. L. Apicella2, G. Apruzzese2, G. Artaserse2, E. Barbato2, F. Belli2, A. Bertocchi2, W. Bin1, L. Boncagni2, A. Botrugno2, S. Briguglio2, A. Bruschi1, P. Buratti2, G. Calabrò2, A. Cardinali2, C. Castaldo2, S. Ceccuzzi2, C. Centioli2, R. Cesario2, C. Cianfarani2, S. Cirant1, F. Crisanti2, O. D'Arcangelo1, M. De Angeli1, R. De Angelis2, L. Di Matteo2, C. Di Troia2, B. Esposito2, D. Farina1, L. Figini1, G. Fogaccia2, D. Frigione2, V. Fusco2, L. Gabellieri2, C. Galperti1, S. Garavaglia1, E. Giovannozzi2, G. Granucci1, G. Grossetti1, G. Grosso1, F. Iannone2, A. Krivska4, H. Kroegler2, M. Lontano1, G. Maddaluno2, C. Marchetto1, M. Marinucci2, D. Marocco2, G. Mazzitelli2, C. Mazzotta2, A. Milovanov2, D. Minelli1, F. C. Mirizzi2, G. A. Moro1, F. Napoli4, S. Nowak1, F. P. Orsitto2, D. Pacella2, L. Panaccione3, M. Panella, V. Pericoli-Ridolfini2, S. Podda2, A. Pizzuto2, G. Pucella2, G. Ramogida2, G. Ravera2, A. Romano2, C. Sozzi1, A. A. Tuccillo2, O. Tudisco2, B. Viola2, E. Vitale2, G. Vlad2, V. Zanza2, M. Zerbini2, F. Zonca2, M. Aquilini2, P. Cefali2, E. Di Ferdinando2, S. Di Giovenale2, G. Giacomi2, F. Gravanti2, A. Grosso2, V. Mellera1, M. Mezzacappa2, A. Pensa2, P. Petrolini2, V. Piergotti2, B. Raspante2, G. Rocchi2, A. Sibio2, B. Tilia2, C. Torelli2, R. Tulli2, M. Vellucci2, D. Zannetti2.1Associazione Euratom-ENEA sulla Fusione, IFP-CNR, Via R. Cozzi 53, 20125 Milano, Italy.2Associazione Euratom-ENEA sulla Fusione, C. R. ENEA Frascati, Via E. Fermi 45, 00044Frascati, Roma, Italy.3Associazione Euratom-ENEA sulla Fusione, CREATE, Via E. Fermi 45, 00044 Frascati, Roma, Italy.4ENEA Guest.