EPS2005, Session "P-5" Abstracts

Session Author PosterTitle Numerical Simulation of the Collisionless Diffusion of Particles Across a Magnetic Field at P-5.001 M.Shoucri a Plasma Edge Carbon source from the toroidal pumped limiter during long discharge operation in Tore P-5.002 E.Dufour Supra Surface layer analysis on the Toroidal Pumped Limiter of Tore Supra using IR data during P-5.003 Y.Corre disruptions Recombining low temperature plasma in the divertor simulator Mistral-B radiative emission P-5.004 A.Escarguel study with the SOPHIA code P-5.005 M.Sakamoto Impact of a movable limiter on the global wall recycling in TRIAM-1M P-5.006 N.Asakura ELM propagation and fluctuations in SOL and divertor on JT-60U Tokamak Emission Rates of CH/CD and C2 spectral Bands for Hydrocarbon-Loss-Events Measured P-5.007 T.Nakano in JT-60U Divertor Plasmas High-Density Deuterium Plasma Discharge in NAGDIS-II with Double Anode P-5.008 D.Nishijima Configuration P-5.009 H.Masuda Optimum Generation of High Heat Flux Toroidal Plasma by RF Ohmic Discharge Effect of plasma Biasing on Suppression of Electrostatic Fluctuation in the Edge Region of P-5.010 K.Yambe STP-3 M Reversed Field Pinch An analysis of the magnetic presheath entrance in a multi-component plasma with ExB and P-5.011 N.Jelic diamagnetic drifts P-5.012 KarinSchombourg Synchronisation of ELMs within magnetic perturbation bursts in TCV P-5.013 M.Wischmeier Impurities from main chamber walls as trigger for detachment in TCV deuterium discharges Comparing magnetic triggering of ELMs in ASDEX Upgrade and TCV with the DINA-CH P-5.014 S.H.Kim tokamak simulator P-5.015 E.Delchambre Issues related to power loading studies in MAST with an infrared camera P-5.016 Zheng,Y.Z Studies of Internal Magnetic Fluctuation by runaway transport in the HL-1M P-5.017 Y.Huang Statistical Estimate of Long-Range Time Dependency in HL-1M edge Plasma Turbulence P-5.018 J.Brotánková Fluctuation measurements with 2D matrix of Langmuir probes on the CASTOR tokamak P-5.019 J.Zajac Multifractal analysis of tokamak plasma turbulence in biasing experiments P-5.020 R.Panek Anomalous impurity diffusion in an experimentally measured turbulent potential P-5.021 G.Pokol Analysis of transient MHD modes of Wendelstein 7-AS by coherence techniques Detection of radially localized and poloidally symmetric structures in the poloidal flow of P-5.022 M.Berta tokamak plasmas P-5.023 S.Zoletnik Anomalous transport events in the core plasma of the Wendelstein 7-AS stellarator P-5.024 D.López-Bruna Effects of magnetic shear on confinement in TJ-II ECRH discharges Direct evidence of coupling between density tails and turbulent transport in the scrape-off P-5.025 E.Calderón layer region in the TJ-II stellarator P-5.026 F.Medina Spatial distribution of lost ripple-trapped suprathermal electrons P-5.027 J.A.Alonso High-speed turbulence imaging in TJ-II edge plasmas. P-5.028 L.Krupnik Electron internal transport barriers, rationals and fluctuations in the TJ-II P-5.029 M.A.Ochando Effect of suprathermal electrons on impurity ionization state P-5.030 M.A.Pedrosa Experimental investigation of ExB sheared flow development in the TJ-II stellarator Influence of ExB sheared velocities in the statistical properties of fluctuations in the plasma P-5.031 R.O.Orozco boundary of the TJ-II stellarator P-5.032 V.I.Vargas Local transport in density and rotational transform scans in TJ-II ECRH discharges Dynamical coupling between parallel flows and radial turbulent transport in a linear plasma P-5.033 E.Anabitarte machine Interpretation of perturbative transport experiments based on modulation techniques in P-5.034 J.A.Mier DTEM numerical turbulence P-5.035 J.M.Delgado Effect of poloidal flow on fluctuations P-5.036 J.Dies Theoretical transport analysis in TJ-II scenarios with enhanced heat confinement P-5.037 JerónimoGarcía Internal transport barrier simulation in the LHD P-5.038 T.C.Luce Search for Threshold Behavior in DIII-D Electron Transport P-5.039 A.N.Simakov Ion and electron viscosity for arbitrary-mean-free-path plasma P-5.040 B.Coppi Magnetic Reconnection and Associated Transport of Plasma Thermal Energy P-5.041 M.H.Redi Testing Gyrokinetics on C-Mod and NSTX

P-5.042 S.M.Kaye Confinement and Transport Scaling in NSTX P-5.043 W.M.Solomon Neoclassical Poloidal Rotation Studies in High Temperature Plasmas P-5.044 P.J.Catto Neoclassical Radial Electric Field and Flows in a Collisional Tokamak P-5.045 G.R.Tynan Observation of Turbulent Driven Zonal Flow in a Cylindrical Plasma Device Studies of Density Fluctuation Dynamics in L-H Transitions and Pedestal Formation in P-5.046 C.Holland DIII-D Using Beam Emission Spectroscopy P-5.047 J.A.Boedo Modification of Intermittency by External Perturbations Spatially resolved ion heating measurements during a sawtooth crash on the Madison P-5.048 SanjayGangadhara Symmetric Torus P-5.049 F.Villone Analysis of RWM with a 3D Model of Conducting Structures P-5.050 D.Bonfiglio MHD dynamo and charge separation in Reversed Field Pinch plasmas First results on the analysis of magnetic fluctuations in toroidal geometry and comparison P-5.051 PaoloZanca with numerical simulations for the RFX reversed field pinch Imaging of magnetic chaos reduction and coherent structures in the MST reversed field P-5.052 P.Piovesan pinch P-5.053 L.Marrelli Transport properties and magnetic fluctuations in RFX-mod First Reversed Field Pinch plasmas with new magnetic boundary in the RFX-mod P-5.054 S.Martini experiment P-5.055 P.Buratti Inter-Machine Scaling of Alfvén-like Modes Excited by Magnetic Islands in Tokamaks P-5.056 H.Reimerdes Resistive Wall Mode Stability in High Beta Plasmas in DIII-D and JET Measurement of the Instability Threshold for Toroidal Alfvén Eigenmodes in JET Tokamak P-5.057 DuccioTesta Plasmas P-5.058 M.K.Zedda Novelty Detection for on-line disruption prediction systems

P-5.059 C.J.Boswell Experimental Observations of n 0 Mode Driven by Energetic Particles on JET Interplay of Error Field and Neoclassical Tearing Mode drives and rotation for the 2/1 P-5.060 R.J.Buttery mode on JET and DIII-D P-5.061 A.V.Burdakov Stable operation regimes in the multimirror trap GOL-3 P-5.062 V.V.Postupaev Features of MHD activity in beam-heated plasma in multimirror trap GOL-3 P-5.063 S.Yu.Medvedev New Adaptive Grid Plasma Evolution Code SPIDER P-5.064 S.Yu.Medvedev Magnetic ELM Triggering and Edge Stability of Tokamak Plasma P-5.065 A.A.Martynov Reversed Current Density in Tokamaks - Equilibrium and Stability Issues Investigation of internal transport barrier in sawtooth oscillation OH and sawtooth P-5.066 I.S.Bel'bas stabilized by off-axis ECRH heating modes. P-5.067 Yu.V.Gott Experimental Studies of the Ion Plasma Component Behavior during Irregularity of the Magnetic Island Rotation under External Helical Magnetic Perturbation P-5.068 N.V.Ivanov in T-10 Tokamak P-5.069 V.D.Pustovitov Analysis of RWM feedback systems with internal coils P-5.070 V.Ilgisonis Sufficient stability condition for axisymmetric equilibrium of flowing magnetized plasma P-5.071 L.G.Askinazi Plasma potential structure nearby the magnetic island in the TUMAN-3M tokamak P-5.072 G.Granucci Mitigation of disruption-generated runaways by means of ECRH P-5.073 E.Joffrin Steady state and stationary long pulse operation in JET P-5.074 L.Laborde A multiple-time-scale approach to the control of ITBs on JET P-5.075 YAndrew H-mode Threshold Experiments on JET Comparison of High Density Discharges Heated Ohmically and with NBI in the Globus-M P-5.076 V.K.Gusev Spherical Tokamak P-5.077 A.V.Anikeev The SHIP experiment at GDT First experimental results P-5.078 Yu.V.Mitrishkin Simulation of burning plasma in multivariable kinetic feedback control system Analysis of the of the supra-thermal electrons during disruption instability in the T-10 P-5.079 P.V.Savrukhin tokamak P-5.080 V.N.Dokuka Linearization of ITER plasma equilibrium model on DINA code P-5.081 C.Ionita Direct Measurements of the Electron Temperature by a Ball-pen/Langmuir probe New Results on the Laser Heated Emissive Probe for Direct Measurements of the Plasma P-5.082 R.Schrittwieser Potential P-5.083 A.M.M.Fonseca Rational surfaces determination from ECE radiometry in the TCABR tokamak P-5.084 L.Giannone A prototype resistive bolometer for ITER P-5.085 M.Garcia-Muñoz Fast Ion Loss Diagnostic in ASDEX Upgrade Use of a High Resolution Overview Spectrometer for the Visible Range in Fusion P-5.086 S.Brezinsek Boundary Plasmas P-5.087 JSvensson Bayesian Modelling of Spectrometer Systems and Plasma Profile Inversions from Spectra P-5.088 D.Hildebrandt Thermal Properties of the CFC-material BN31 at surface heat loading High spatial and temporal resolution FM-CW reflectometry to study pellet triggered ELMs P-5.089 L.Fattorini at ASDEX Upgrade Spatial asymmetries in ELM induced pedestal density collapse in ASDEX Upgrade H- P-5.090 I.Nunes modes A tool for the parallel calculation and ready visualization of the Choi-Williams distribution P-5.091 A.C.A.Figueiredo application to fusion plasma signals P-5.092 T.I.Madeira First results obtained with the real-time PHA diagnostic in the tokamak TCV Study of core and edge plasma structures in TEXTOR with the burst-mode 10 kHz P-5.093 S.K.Varshney Thomson scattering system P-5.094 G.Kamelander Experimental validation of improved pellet ablation models P-5.095 T.Hellsten On Parasitic Absorption in FWCD Experiments in ITB Plasmas P-5.096 R.Farengo Current drive with oscillating magnetic fields in RFPs and FRCs P-5.097 M.Taguchi Radial diffusion equation for rf-driven current density in tokamaks P-5.098 K.Hamamatsu Monte-Calro simulation of electron cyclotron cuurent drive in NTM magnetic islands P-5.099 H.Ohwaki Modeling of Plasma Current Decay during the Disruption Axisymmetric MHD Simulation of Disruption Dynamics in High-beta Reversed Shear P-5.100 H.Tsutsui Plasmas P-5.101 D.Applegate Micro-Tearing Modes in MAST P-5.102 G.P.Maddison Analytic approximations of divertor behaviour and application to MAST P-5.103 A.Surkov Parametric decay instability accompanying electron Bernstein wave heating in MAST Spesific feature of ICR heating on the spherical tokamak Globus-M at high concentration of P-5.104 V light ion component P-5.105 A.Bendib Effect of electron collisions on nonlocal transport coefficients in laser heated plasmas P-5.106 A.V.Brantov Relaxation of the localized temperature perturbation in the collisional plasma P-5.107 J.Ramírez Upgraded version of the ICF simulation code MULTI P-5.108 Dubroca Magnetic Field Generation in Pasmas due to Anisotropic Laser Heating P-5.109 R.A.Fonseca OSIRIS 2.0 an integrated framework for parallel PIC simulations P-5.110 F.Hamadi Numerical modelling of metallic plasma produced by UV laser beam ablation

P-5.111 BongJuLee Hyperthermal neutral beam for the material processing P-5.112 M.Fiore Magnetic field generation in the Weibel instability P-5.113 A.V.Kochetov Phase Evolution Of Signals Scattered By Modified Ionosphere Comparison of various mechanisms of electron heating at the irradiation of dense targets by P-5.114 V.P.Krainov a super-intense femtosecond laser pulse P-5.115 M.V.Goldman Electron holes and Buneman instabilities in nonuniform current-carrying plasma equilibria P-5.116 S.Hamaguchi Plasma discharge in supercritical fluids P-5.117 M.A.Erukhimova Stimulated emission without inversion in ensembles of classical electrons P-5.118 N.Ezumi Electron Temperature Control in a Hot Cathode Arc Discharge Plasma P-5.119 MuhammadShafiq Wake potential of a test charge using the stationary phase method Supersonic Rotation Exceeding the Alfven Ionization Limit in the Maryland Centrifugal P-5.120 R.Ellis Experiment P-5.121 G.Sorasio Kinetic Theory of Electrostatic Collisionless Shocks in Plasmas Analysis of the influence of excited configurations and plasma interaction in atomic and P-5.122 R.Florido plasma magnitudes of interest in NLTE plasmas using an analytical potential Analytical potential for determining atomic properties of ions in plasmas for a wide range P-5.123 R.Florido of plasma coupling parameter. An application to calculate total photoionization cross section. Using sparse matrices techniques and iterative solvers in the calculation of level P-5.124 R.Florido populations for NLTE plasmas P-5.125 J.G.Kwak mm-wave collective scattering study of density fluctuations in helicon plasmas P-5.126 G.Sorasio Orbital motion theory of the shielding around an electron emitting body Calculation of the radiative opacity of laser-produced plasmas using a relativistic-screened P-5.127 R.Florido hydrogenic model for ions including plasma effects P-5.128 K.N.Sato Development of a new non-diaphragm type shock tube for high density plasmas P-5.129 S.Tzortzakis X and XUV time-resolved emission of laser-created hot Xe and Kr plasmas P-5.001, Friday July 1, 2005

Numerical Simulation of the Collisionless Diffusion of Particles Across a Magnetic Field at a Plasma Edge M. Shoucri1, E. Pohn2, G. Kamelander2 1 Institut de recherche d’Hydro-Québec, Varennes Québec Canada J3X 1S1 2 Forschungszentrum Seibersdorf, Austria

There are many situations at the plasma edge of a tokamak where both ions and electrons are accelerated along the magnetic field lines. This acceleration of the particles along the magnetic field lines is generally coupled to a diffusion of the particles across the magnetic field lines. We study this problem using a two- dimensional (2D) gyro-kinetic Vlasov code that include the finite Larmor radius correction and the polarization drift, in a slab geometry. The ExB drift is charge and mass independent. The ions finite Larmor radius correction allows the existence of a charge separation between ions and electrons, and the polarization drift, which has opposite signs for ions and electrons, has a tendency to accentuate the charge separation in a time varying electric field and in the presence of strong gradients. The Kelvin-Helmholtz instability which results from this charge separation can produce a low frequency turbulent spectrum. We study the phase-space dynamics of the particles in this turbulent spectrum. In the present 2D slab layer, z represents the toroidal direction, assumed homogeneous, x represents the poloidal periodic direction and y represents the radial direction. The magnetic field is located in the x-z plane, very close to the toroidal direction. It is found in the present geometry that the collisionless

diffusion coefficients Dy in the radial direction and Dv in velocity space along the magnetic field are related by a relation which links the electron dynamics in the x-y real space and in the y-v phase space. This relation is verified using test particles diagnostic techniques.The Vlasov code associated with test particles diagnostic techniques provides a powerful tool to study particle diffusion in space and in phase- space, especially in the low-density regions of the distribution function, because of the very low noise level associated with the Vlasov code.

P-5.002, Friday July 1, 2005

('XIRXU&%URVVHW&/RZU\&'HVJUDQJHV55HLFKOH50LWWHDX-%XFDORVVL -*XQQ30RQLHU*DUEHW%3pJRXULp(7VLWURQH35RXELQ&0DUWLQ &$UQDV3&KDSSXLV<&RUUH5*XLUOHW-+RJDQ7/RDUHU37KRPDV

!∀#∃%&∀∋()#%&∀∗∋+, (−−,..∋,∃/∃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ĮVSHFWUDOOLQHEULJKWQHVVHV LQVHULHVRIH[SHULPHQWVZKHUHSODVPDHGJHFRQGLWLRQVDUHFKDQJHGE\XVLQJGLIIHUHQWOHYHOV RI /+ RU ,&5 KHDWLQJ DQG GLIIHUHQW FRUH IXHOOLQJ PHWKRGV JDV SXII DQG SHOOHW LQMHFWLRQ HGJH ([SHULPHQWDOFDUERQVSXWWHULQJ\LHOGVRIa DUHIRXQGIRUQH a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riday July 1, 2005

!∀#∃ !!∀

%&∋∋(!(∋ &)&&&(∋∋(& ! (∋ ∀∋(& &&( % ∗ ( ∋(∋(&∋(( (&()!!∋!)%∀∋( ( (( (&! &! ∋ ( ∋ ( + & & &(, − +(−! & ∋(,∋(∗(%∋∋&)(! ∗&&!&∋

∀ & ( ∀ #∋ ∋./0& (∀&∋+∀,∀1(2/ 34(1&&∋& !+(1 &5 µ .2µ & 67/µ, # ∋)& 1 ∋((∋&∀∋& +)8/,9!((&() &&−∋&∋)∋ +!(!,&(&∋&%∗∀#∋ !&!&∋(&!!! &∋&∀∋&∋)&&∋ &!(∋!(&&&&! ( ∀ ) ! ∗ ! + ∋& ∋ )&(!!#&∃%&∋,∀4& ( ( &∋& ∀ (( & (∋& )!+∋,)&(%∋(∋!∀ &&∋4&∋&&)!!∗&&! &∋)!&&!%∋(∋! 9&&∋( P-5.004, Friday July 1, 2005

Recombining low temperature plasma in the divertor simulator Mistral-B : radiative emission study with the SOPHIA code

A. Escarguel, F Rosmej, C. Brault, R. Stamm, Th. Pierre, D. Guyomarc’h, K. Quotb PIIM-UMR 6633 CNRS/Université de Provence, centre de St-Jérôme 13397 Marseille, France

Low temperature recombining plasmas are observed in edge divertor tokamak plasmas. They are of crucial importance, because of their role in reducing the plasma heat flux to plasma facing components. Similar plasmas can be obtained in the laboratory Mistral-B experiment at PIIM, Marseille [1]. The Mistral-B device consists in a linear high density (~2×1018 m-3 typically on the axis) magnetized plasma column. A diagnostic method for the determination of non-equilibrium conditions is applied to the cold recombining helium plasma (0.6 eV ≤ Te ≤ 2 eV) obtained by radial ejection in a limiter shadow [2]. The method is based on the perturbation of ionic and atomic fractions due to particle flow and diffusion which drive anomalous population flow of three-body and collisional excitation processes. In neutral helium these effects can be observed by bound-bound transitions from the singlet and triplet systems because their response is different as the triplet system cannot decay radiatively to the ground state. Simulations are carried out with the collisional radiative code SOPHIA for the line series 2-n in neutral helium taking fully into account the singlet and triplet structure until high quantum numbers [3]. The simulations are compared with the emission spectra from MISTRAL showing transitions up to principal quantum number n = 19. Different plasma regimes will be discussed along with simulations of the radiative emission properties.

[1] C. Brault et al., 12th International Conference on Plasma Physics, Nice, France (2004) [2] A. Escarguel et al., 17th International Conference on Spectral Line Shape, Paris, France (2004) [3] F. Rosmej et al., J. Nucl. Mat., 337-339 (2005) P-5.005, Friday July 1, 2005

Impact of a movable limiter on the global wall recycling in TRIAM-1M M. Sakamoto1, M. Ogawa2, K. Takaki2, H. Zushi1, K. Nakashima2, N. Maezono2, T. Sugata2, Y. Nakashima3, Y. Higashizono3, Y. Kubota3, A. Higashijima1, H. Nakashima1, S. Kawasaki1, A. Iyomasa1, M. Hasegawa1, H. Idei1, K. Hanada1, K. Nakamura1, K.N. Sato1 1 Advanced Fusion Research Center, Research Institute for Applied Mechanics, Kyushu University, Kasuga, Fukuoka, Japan 2Interdisciplinary Graduate School of Engineering Sciences, Kyushu University, Kasuga, Fukuoka, Japan 3 Plasma Research Center, Tsukuba University, Tsukuba, Ibaragi, Japan

Steady state operation (SSO) is one of the critical issues for the future fusion reactor. Understanding of the global wall recycling is important from the viewpoint of the density control for the SSO, since the wall plays a role of the particle sink and source. In TRIAM-1M, the ultra-long discharge with the duration of 5 h 16 min was achieved using a local movable limiter with good cooling capability. Moreover, effects of the movable limiter on the global wall recycling have been studied. Due to insertion of the movable limiter to the plasma edge, SOL parameters, the toroidal profile of the neutral particle density and the heat load on the wall are changed to influence the global wall recycling. In 5 h 16 min discharge, the temperature increase in the plasma facing components could be successfully suppressed by the insertion of the movable limiter. Long time global wall saturation (i.e. global balance between particle absorption and release of the wall) was not observed. The averaged wall pumping rate is evaluated to be ~8.6 x 1016 atoms m-2 s-1, which is about 3.6 times higher than that evaluated in the previous ultra-long discharge without the movable limiter. It seems to be attributed to suppression of the hydrogen release from the wall due to less temperature increase. The wall inventory continued to increase during the discharge and finally reached up to 8 x 1021 atoms at the end of the discharge. The whole toroidal profile of H-alpha emission was changed by the insertion of the movable limiter. The H-alpha emission at the movable limiter and its neighborhood increased but that of other parts, i.e. the main chamber and fixed poloidal limiters decreased. The contribution ratio of the main chamber to the whole wall recycling decreases by about half due to the movable limiter. That of the section of the movable limiter reaches about 45%. P-5.006, Friday July 1, 2005

ELM propagation and fluctuations in SOL and divertor on JT-60U Tokamak N. Asakura1, M. Takechi1, G. Matshunaga1, N. Oyama1, T. Nakano1, H. Miyoshi2, N. Ohno3, S. Takamura2, Y. Uesugi4 1Japan Atomic Energy Research Institute, Naka, Ibaraki 311-0193, Japan 2Department of Energy Engineering and Science, Graduate School of Engineering, Nagoya Univ, Nagoya 3EcoTopia Science Research Institute, Nagoya Univ., Nagoya 464-8603, Japan 4Department of Electrical and Electronic Engineering, Kanazawa Univ. Kanazawa 920-8667, Japan Transient heat and particle loading caused by Edge Localized Mode (ELM) is crucial for determining the lifetime of ITER divertor materials. At the same time, understanding of the transient radial transport is important to determine the heat and particle loading to the first wall. The radial and parallel transport of the ELM plasma has been investigated in JT-60U, using a midplane reciprocating Mach probe, target Langmuir probes and magnetic pick-up coils at the first wall. The sampling rate was recently improved from 200 to 500kHz. Multi

mid peaks in the waveform of ion saturation current of the Mach probe, js , were measured clearly during an ELM event, and time and amplitude of the burst were determined. Enhancement and propagation of the peak plasma flux was observed even at far SOL flux surfaces, i.e. distance of the reciprocating probe location to separatrix of !rmid~16 cm. Radial mid ∀ mid propagation velocity at the outer midplane, Vpara , was evaluated from the time lag, perp , mid mid ! mid ∀ mi mid between the first peaks in js and magnetic coil signal: Vpara = r / perp . Vpara near the separatrix (!rmid = 2-4cm) and at far SOL (!rmid = 8-10cm) were ~1 and ~2 km/s, respectively.

Floating potential, Vf,, was also enhanced to |Vf| ~400-600V near the separatrix and 30-80V at far SOL. As a result, the radial propagation of the ELM plasma is accelerated at far SOL

mid while the amplitude of flux density is reduced. At the same time, cross-correlation of js turbulences at the ion and electron sides of the Mach probe (downstream and upstream sides, respectively) became large over all !rmid, suggesting radial transport of the ELM plasma. At ELM event, radial plasma transport is enhanced to the far SOL, and ELM transport model is discussed. Influence of the parallel and radial transport on transient plasma loading profile to

div div the divertor is also discussed compared to measured js and Vf profiles. Characteristics of the fluctuations were investigated between ELMs. Fluctuation level, # mid mid js /, was increased from 40% (near separatrix) to 100% (at far SOL), which included mid large power components of high frequencies (f>20kHz). Small and fast positive bursts in js was found most frequently at far SOL (!rmid~7cm), suggesting intermittent transport. However, the bursts may be different from so call "blob", which is mostly generated near separatrix. The characteristics (amplitude, frequency and localization) are discussed. P-5.007, Friday July 1, 2005

Emission Rates of CH/CD and C2 Spectral Bands for Hydrocarbon-Loss-Events Measured in JT-60U Divertor Plasmas

T. Nakano, K. Tsuzuki, S. Higashijima, H. Kubo, N. Asakura

Japan Atomic Energy Research Institute, Naka, Ibaraki, 311-0193, Japan

The CH4 sputtering yield has been measured in almost all fusion experimental devices with carbon divertor plates. But the sputtering yields are substantially different from one another. This discrepancy is considered to be due to variety of emission rates of CH spectral bands. The CH flux is expressed as Γ = Iobs − B · Iobs ·LEPCH if contribution of CH emission 4 CH4 ³ CH C2 ´ CH4 originating from heavier hydrocarbon such as C2H6 is taken into account. Here B is a branching ratio of CH to C emission, originating from heavier hydrocarbon, and LEPCH stands for ’CH 2 CH4 4 Loss-Events / CH Photon’, which is defined as a reciprocal of the number of CH photons emitted until one CH4 is lost in plasmas by collision processes such as dissociation. By using known B and LEPCH , Γ can be estimated from measured intensities of CH and C spectral bands CH4 CH4 2 ( Iobs and Iobs, respectively). However, B and LEPCH depend on the transport during the long CH C2 CH4 break-up chain from CH4 to CH, which is influenced by the divertor geometry. Hence, B and LEPCH are not identical amongst devices but it is necessary to determine them in each device. CH4 In this work, to determine LEPCH and B in JT-60U, CH and C H were injected at a known CH4 4 2 6 flow rate into spectroscopic observation volume around the outer strike point. LEPCH : as shown in Fig. 1, LEPCH and LEPCD measured in JT-60U and LEPCH in CH4 CH4 CD4 CH4 PISCES-A are in good agreement at 30 eV. At Te < 30 eV, this agreement is considered to be hold by extrapolation of measured LEPCH along the model calculation. CH4 B : contribution of CH spectral band emission origi- 250 nating from C2H6 was found to be more significant than CD reported in [1]. Higher B by ∼ 50% was obtained. 200 4 In the previous report on the CH sputtering yield [2], 4 150 CH LEPCH and B taken in PISCES-A were used. By replac- 4 CH4 100 ing them by ones determined here, the CH sputtering 4 Model calculation CH4 ∼ 50 PISCES-A yield was revised: lower by 50 %, due to the higher Loss-Events / CH Photon 4

B. Revised C2Hy sputtering yield will be also reported. CH 0 20 40 60 80 T (eV) References e Figure 1: CH4 Loss-Events / CH Photon as a function of T . Data taken in PISCES [1] A. Pospieszczyk et al., UCLA-PPG-1251 (1989) e -A [1] and data from model calculation by the authors are also shown. [2] T. Nakano, et al., Nucl. Fusion 42, 689 (2002) P-5.008, Friday July 1, 2005

High-Density Deuterium Plasma Discharge in NAGDIS-II with Double Anode Configuration

D. Nishijima, I. Yagyu, D. Tanaka, M. Takagi, N. Ohno and S. Takamura

Nagoya University, Nagoya, Japan

Linear divertor plasma simulators such as NAGDIS-II have been taking a significant role in

basic understanding of divertor plasma of large tokamaks such as International Thermonuclear

Experimental Reactor (ITER). It has contributed especially to the physics of plasma detachment

and plasma-wall interactions.

The plasma density of divertor plasma in ITER is expected to be an order of 1020 m-3 or more.

However, the plasma density in NAGDIS-II is not so high compared with that of the divertor

plasma, especially for hydrogen isotopes plasma. Maximum electron density (ne) of deuterium

plasma in NAGDIS-II so far is early 1018 m-3, which is limited an instability of discharge. The

voltage of direct-current (d.c.) discharge increases and begins to fluctuate as the discharge current

increases, and finally the discharge disappears. In order to stabilize the d.c. discharge and to obtain

high-density deuterium plasma, the second auxiliary anode is put in NAGDIS-II at a distance of 0.2

m from the first anode. The structure of the second anode is hollow type, the same as the first one.

The hollow hole has a diameter of 27 mm with a length of 70 mm. Such an anode structure

increases the effective surface of anode and decreases the conductance between the plasma source

region and divertor test region. We expected that a reduction of neutral gas pressure in the test

region could increase plasma density because it decreases the rate of charge exchange process

between deuterium ion and neutral atom/molecule. Contrary to our expectation, ne was drastically

reduced in the test region with low gas pressure condition while it increased as the gas pressure was

increased by deuterium gas puffing in the test region. In addition, the gas puffing decreased the

discharge voltage compared to the old configuration. The limit of discharge current is expanded

from around 100 A to 160 A due to the double anode configuration. P-5.009, Friday July 1, 2005

Optimum Generation of High Heat Flux Toroidal Plasma by RF Ohmic Discharge H. Masuda1, H. Kitahara1, M. Nagase1, N. Ohno2, S. Takamura1 and M. Takagi1 1 Department of Energy Engineering & Science, Nagoya University, Nagoya , Japan 2 EcoTopia Science Institute, Nagoya University, Nagoya, Japan

Linear divertor plasma simulator (L-DPS) has contributed to understanding the various physics in boundary plasmas of fusion devices, since it has a good accessibility for comprehensive diagnostics and flexible control of plasma parameters. However, L-DPS, which has a simple magnetic configuration, can not simulate some important phenomena appeared in tokamak SOL plasmas related to the magnetic configuration, such as curvature and grad B effects. To overcome these problems, we have developed a new toroidal divertor plasma simulator, NAGDIS-T (NAGoya DIvertor plasma Simulator with Troidal magnetic configuration) to investigate the SOL plasma physics. In the NAGDIS-T, the plasma is generated by DC and/or RF Joule discharge. In order to generate a high density plasma relevant to boundary plasmas in fusion devices, we need to optimize experimental parameters for RF Joule discharge such as gas pressure, driving frequency, transformer truns-ratio, plasma volume, etc. We have investigated such an experimental condition for RF Joule discharge to get 18 -3 high electron density ne of H2 plasma above 10 m .When the maximum toroidal magnetic field is 1 kG and the maximum vertical field is 10 Gauss, the connection length of magnetic line of force is 28 m. The driving frequency was varied from 100 kHz to 300 kHz by a SIT inverter source (9kVA: 300V-30A). 17 -3 In the present system, we got the electron density of ne=4×10 m and the electron

temperature Te=10 eV in the full volume of vacuum chamber. A characteristic

horizontal profile of ne has a hollow profile which may be explained by RF skin effect. The hollow profile is changed by changing the driving frequency, the gas pressure and the input RF power. The plasma density increases up to 1×1018 m-3 by using a plasma limiter in order to reduce the discharge volume, especially the low-field side. The high power RF source of 30kW will work in near future so that we will have much higher plasma density relevant to the boundary plasma in fusion devices. In the conference, we will report on a preliminary result in high power RF Joule discharge. P-5.010, Friday July 1, 2005

Effect of plasma Biasing on Suppression of Electrostatic Fluctuation in the Edge Region of STP-3(M) Reversed Field Pinch K. Yambe1, S. Masamune2, H. Arimoto1, K. I. Sato1, Y. Yagi3 1 Department of Energy Engineering and Science, Nagoya University, Furocho, Chikusa-ku, Nagoya, 464-8603, Japan 2 Department of Electronics and Information Science, Kyoto Institute of Technology, Matsugasaki, Sakyo-ku, Kyoto, 606-8585, Japan 3 National Institute of Advanced Industrial Science and Technology, 1-1-1 Umezono, Tskuba, Ibaraki, 305-8568, Japan

The reversed field pinch (RFP) is one of the magnetic confinement systems for nuclear fusion research. The core confinement of standard RFP is dominated by magnetic fluctuations due to tearing modes, while electrostatic fluctuations affect both the particle and energy transport in the edge region. In this work, we will describe the detailed study on electrostatic transport with plasma biasing experiment in the edge region of an RFP. The plasma biasing experiment using an insertable electrode has been carried out in the

STP-3(M) RFP device. The profiles of electron density ne , electron temperature Te and

plasma potential Vs were measured with an electrostatic probe array in the edge region of the plasma. The effect of biasing is observed as an increase of edge electron density and

suppression of fluctuation amplitudes in ne , Te and Vs . As a result, both the particle and energy fluxes due to the electrostatic fluctuation are reduced and confinement improvement is observed together with a change of the E· B velocity shear. The reduction in both fluctuation amplitudes and coherence between fluctuating quantities contributes to this improvement. Moreover, the velocity shear frequency was estimated as .1( 57 ‒ .0 50) ·108 rad s , and both the electrostatic particle and energy fluxes broadly decreased, and the velocity shear frequency was high enough to decorrelate turbulence. It was suggested that the velocity shear strength affected reduction of particle and energy fluxes due to the electrostatic fluctuation. P-5.011, Friday July 1, 2005

An analysis of the magnetic presheath entrance in a multi-component plasma with E×B and diamagnetic drifts M. Stanojevic1, J. Duhovnik1, N. Jelic2, S. Kuhn2 1 LECAD Laboratory, Faculty of Mechanical Engineering, University of Ljubljana, Aškerceva Street 6, SI-1000 Ljubljana, Slovenia 2 Plasma and Energy Physics Group, Association EURATOM-ÖAW, Department of Theoretical Physics, University of Innsbruck, Technikerstrasse 25, A-6020 Innsbruck, Austria

In recent years, the magnetic presheath, or Chodura sheath, has become a subject of intense investigations. The main reason is that the boundary conditions for the fluid codes used for simulating fusion edge plasmas are imposed at the magnetic presheath entrance (MPSE), where the fluid equations on which these codes are based cease to be valid. For the ion fluid velocity at the MPSE, an “intuitive” boundary condition is usually applied. It ρ is based on the assumption that the total ion fluid velocity component parallel to B at the ρρ MPSE is the vector sum of the parallel velocity, the EB× drift velocity, and the diamagnetic drift velocity, and that it equals the ion-acoustic velocity. However, besides the fact that it was not derived from the basic fluid equations, it can yield some unphysical solutions when incorporated into the fluid codes, such as highly supersonic ion fluid velocity at the MPSE or ion flow in the wrong direction, i.e., from the magnetic presheath or the wall towards the plasma presheath. In this paper we present an analysis ρρ of the MPSE in a multi-component plasma with EB× and diamagnetic drifts, from which the boundary condition for the ion component fluid velocities at the MPSE is derived from the basic fluid equations. This condition takes into account gradients of the drift velocities and can be incorporated into multi-fluid codes used for modelling tokamak boundary plasmas and other magnetically confined plasmas.

P-5.012, Friday July 1, 2005

Synchronisation of ELMs within magnetic perturbation bursts in TCV K. Schombourg, J.B. Lister, Y.R. Martin CRPP-EPFL, CH-1015 Lausanne, Switzerland

Edge Localised Modes frequently appear in H-mode, which, due to its good confinement properties, is the regime foreseen for ITER operations. ELMs facilitate a stationary H- mode and are useful to remove impurities, but for ITER the amount of energy expelled onto the divertor would be too large. Since the expelled energy usually scales inversely

with the frequency of the ELMs, a control of felm would be suitable to avoid damage. The frequency of the ELMs has been successfully controlled on TCV by applying perturbing currents on the vertical position control coils. The fast induced deviations of about 2 centimetres in the plasma vertical position modify the current at the plasma edge by induction, as demonstrated by simulations done with the DINA code. Our experiments show that ELMs are synchronised with the rapidly increasing edge current. When the polarity of the perturbation is inversed and the edge current rapidly decreases, ELMs are momentarily inhibited. We are investigating whether a single perturbation, i.e. a prompt perturbation, directly triggers an ELM or if a repeated perturbation, i.e. cyclic stimulation, is necessary. In this latter case, the ELM limit cycle would be modified in the Current:Pressure-gradient diagram and thus the ELM frequency would be changed. Our results indicate that both prompt and cyclic perturbations have an effect on the ELMs. In this paper their relative importance will be discussed. In the figure below, ELMs are triggered directly by the 1st perturbation, but the synchronisation becomes better after more than one perturbation. In addition, we observed that the delay between the perturbation and the ELM does not usually depend on the delay between two successive perturbations.

Delay between 4th Fig.1: Delays between 1st and 4th perturbations 20 perturbation and the following ELM and the ELMs are shown. The analysis is done with bursts of perturbations with constant 0 Delay between 1st frequency, with the driving frequency differing 20 perturbation and the following ELM from one burst to another.

0 0 2 4 6 ∆ tpELM [ms] P-5.013, Friday July 1, 2005

Impurities from main chamber walls as trigger for detachment in TCV deuterium discharges

M. Wischmeier1, R.A. Pitts1, J. Horacek1, D. Coster2, D. Reiter3

1 Centre de Recherches en Physique des Plasmas, Association EURATOM-Conféderation Suisse, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland 2 Max-Planck-Institut für Plasmaphysik, EURATOM-Association, D-85748 Garching, Germany 3 FZJ-Forschungszentrum Jülich, Association EURATOM-FZJ, D-52425 Jülich, Germany

Tokamak divertor detachment, expected to play a major role in allowing high power fluxes expected in reactor grade devices to be tolerated, is still not completely understood. Although the various physics components of the detachment are thought to be well known, their relative importance and the degree to which each may affect the others in determining the final detached state, can only be understood using complex codes, such as the two dimensional, coupled fluid-Monte Carlo neutral code package, SOLPS. Predictive modeling of detachment and how the divertor should be designed to optimise it is currently being performed for ITER using SOLPS. In TCV deuterium discharges, divertor detachment in the simplest of situations is observed experimentally to be anomalous [1] with respect to the behaviour seen in larger tokamaks, or those with more closed divertor geometries. In helium plasmas, the character of detachment is dramatically modified, as might be expected given the all-carbon first wall in TCV and the absence of chemical sputtering in the case of an inert fuel species. Using the SOLPS5.0 code, a careful comparison has been made between simulation and experiment in both deuterium and helium. This contribution will present a selection of results from this study which provides conclusive evidence for the TCV detachment anomaly being directly linked to an enhanced interaction between the plasma and the graphite main chamber walls at high density, almost certainly due to increased levels of anomalous perpendicular transport in the far SOL. Since ion interaction at the main walls is not correctly accounted for in SOLPS, these results highlight an important case in which the complexity of the real situation is inadequately represented by the numerical model.

References [1] R. A. Pitts et al. Detachment in variable divertor geometry on TCV. Proc. 18th IAEA Fusion Energy Conference (Sorrento, Italy), IAEA-CN-77-EXP/P4-23, 2000. P-5.014, Friday July 1, 2005

Comparing magnetic triggering of ELMs in ASDEX Upgrade and TCV with the DINA-CH tokamak simulator

S.H. Kim, M.Cavinato4, V. Dokuka3, R.R. Khayrutdinov3, P.T. Lang5, J.B. Lister, V.E. Lukash2, Y.R. Martin, L. Villard

Centre de Recherches en Physique des Plasmas, Association EURATOM-Confédération Suisse, EPFL, 1015 Lausanne, Switzerland 2RRC Kurchatov, Moscow, Russia 3TRINITI, Moscow Region, Russia 4Consorzio RFX, Associazione Euratom-ENEA per la fusione, Padova, Italy 5Max-Planck-Institut für Plasmaphysik, EURATOM Association, Garching, Germany

Magnetic triggering of edge localized modes (ELMs) in ASDEX Upgrade was reported showing frequent ELM events when the plasma is rapidly moved vertically. However, the ELMs are triggered when the plasma is moving down, contrary to the previous observation on TCV in which the ELMs occurred when the plasma moved upward. To investigate this opposite behavior of triggered ELMs, the DINA-CH tokamak simulator has been used. Two passive stablilisers located inside the vacuum vessel of ASDEX Upgrade play a significant role during the triggered ELM events. They produce similar external linking flux changes with those observed in the G-coils of TCV for opposite vertical plasma movements. However, these passive structures create different shapes and magnitudes of the plasma flux surface deformation near them, compared with the simulation results of TCV in which the movement preserves the plasma shape. The time variation of toroidal plasma current in the edge region depends on the vertical plasma movement and shows a reversed time response inside the edge region in the two tokamaks. For a more precise understanding of the effect of plasma flux surface deformation by these vertical plasma movements and the passive stabilisers, stability analysis has been started with the finite element stability code, KINX.

P-5.015, Friday July 1, 2005

Issues related to power loading studies in MAST with an infrared camera E. Delchambre, G. Counsell, J. Dowling, A. Kirk, F. Lott, S. Shibaev EURATOM/UKAEA Fusion association, Culham Science Centre, Abingdon, Oxfordshire, OX14 3DB, UK

The target heat fluxes in MAST are deduced from infra-red (IR) measurements of the surface temperature using THEODOR, a 2D inverse heat transfer code [1] and more recently, TACO, a 3D code which allows calculation of the surface heat flux distribution. The 3D code allows for better interpretation of toroidally asymmetric power loads (e.g. during certain transient phenomena and as a result of plasma facing component imbrication). Negative heat fluxes are sometimes derived during surface cooling following a transient power pulse [2] and may be caused by an over-estimation of the temperature during the heating phase. IR measurements on MAST with high spatial resolution relative to gradients in the surface heat flux are combined with the 3D heat flux calculations to examine the role of dust, hot spots and surface roughness (all with dimensions smaller than the spatial resolution) in this over-estimation. Band-pass filters are used to determine the impact of IR wavelength on calculated heat flux and to clarify the role of hot-spots, target surface conditions and plasma radiation in the IR. The effect of the finite time resolution of the IR camera, as well as the spatial resolution, during transient events will also be discussed. Using improvements to the target power measurements in MAST resulting from this study, an overview of power loading on the new divertor and first wall in MAST, both during quiescent phases and transients such as edge localised modes and disruptions, will be presented. Detailed comparisons of target heat loads calculated from the analysis of the extensive array of Langmuir probes in MAST are compared with IR data for a wide range of plasma conditions. With the assumption of equal ion and electron temperatures, good agreement is found during L-Mode discharges but significant discrepancies are observed during H-Mode, especially during ELMs. Wide angle views of 70 % of the vessel during disruptions have shown first wall plasma interactions and toroidal asymmetries in power loadings during later phases of vertical displacement events compared to the evenly distributed flux during early phases. This work is jointly funded by Euratom and the United Kingdom Engineering and Physical Sciences Research Council. The authors would like to thank IPP-Garching for providing the THEODOR code. Dr Delchambre would like to acknowledge a Euratom Training Fellowship grant. [1] A. Herrmann, et al., Plasma Phys. Control. Fusion 37 (1995) 17 [2] S.Clement et al., J. Nucl. Mat. Vol. 266-269 p.285-290 (1999) P-5.016, Friday July 1, 2005

Studies of Internal Magnetic Fluctuation by runaway transport in the

HL-1M ZHENG Yongzhen, DING Xuatong, LEE Wenzhong. Southwestern Institute of Physics, P.O.Box 432, Chengdu, 610041, China Email: [email protected] Abstract The transport of runaway electrons in hot plasma can be comparatively easily measured by perturbation or steady state experiments, which provide runaway

electron diffusivity Dr. The runaway diffusion coefficient has been obtained using three different perturbation techniques and steady state approach, each sensitive to a

different region of the plasma. (1)Dr is deduced from Plasma shift experiment using a vertical magnetic for the plasma displacement; (2)Dr is deduced from its sawteeth oscillations behaviors of HXR flux, and of SXR intensity, recorded simultaneously;

(3)Dr is deduced from microwave radiation intensity sawtoothing behaviors; (4) Dr is deduced from hard-X-ray bremsstrahlung spectra. The mentioned methods give local 2 -1 values for Dr in the range of (0.4-10) m s with a decreasing dependence on toroidal –1.5 magnetic field as BT . The diffusivity can be interpreted in terms of a magnetic ~ fluctuation level. The internal magnetic fluctuations level (br /BT) is estimated to be about (2-4) ×10-4 in the HL-1M plasma. The results presented here demonstrate the effectiveness of using runaway

transport techniques for determining internal magnetic fluctuations. A profile of the magnetic fluctuation level as a function of plasma radius in the HL-1M can be

estimated from the runaway diffusion coefficient Dr and shown as Figure 1. We can confirm that magnetic fluctuations values increase from the edge to the inner minor radius. It appears that magnetic fluctuation will not be a major contribution to the transport of thermal species in the HL-1M.

P-5.017, Friday July 1, 2005

Statistical Estimate of Long-Range Time Dependency

in HL-1M edge Plasma Turbulence

Yuan Huang, Xiaoming Qiu Southwestern Institute of Physics, P O Box 432, Chengdu 610041

Abstract Many different plasma turbulence models have predicted the existence of long-range time dependencies, which could significantly affect the cross-field turbulent transport in magnetically confined plasmas by means of radial “avalanches” or “heat pulses”. Experimentally, a lot of data analyses are utilized to look for the evidence of long-range dependencies. There are roughly two types of tools used in assessing the presence of such dependencies in experimental data: (1) spectral scaling exponent c estimate analyses in power spectral density (PSD), such as spectral exponent estimate methods via Fourier transform, or spectral wavelet analysis(SWA), and so on. (2) Hurst exponent H assessment methods, such as rescaled range (R/S) technique, structure functions (SF’s) technique, detrended fluctuation analysis (DFA), and so on. In this work, for the same experimental data of turbulent particle flux measured at the HL-1M tokamak plasma edge by Langmuir probe diagnostic, Fourier transform and SWA methods were used to get spectral exponent, while R/S technique, DFA and SF’s methods were utilized to get Hurst exponent. The HL-1M plasma is circular in cross-section with major radius R = 1.02 m and minor radius a = 0.26 m. The data utilized by this work are taken from measurements in the ohmic-heating regime of hydrogen discharges during the period of the plasma current plateau. The results are, within self-similarity range(100~ v ~ 2000 os ), c ~ .1 05 by FFT and c ~ .0 48 by SWA,

H (q ? )2 ~ .0 74 by SF’s, H ~ .0 76 by R/S and H ~ .0 77 by DFA. The relationship between the spectral exponent c and Hurst exponent H is known to be c ? 2H -1 for fractional Brownian motion(FBM) and c ? 2H /1for fractional Gaussian noise(FGN). By checking if spectral exponent c and Hurst exponent H satisfy such relations, results show that SWA is a proper method to obtain the spectral exponent and SF’s method shows advantages to get Hurst exponent. Considering c ? 2H /1 relation, SWA and SF’s (or R/S, DFA) methods are not independent to each other. P-5.018, Friday July 1, 2005

Fluctuation measurements with 2D matrix of Langmuir probes on the CASTOR tokamak

J Brotánková1, J Stöckel1, E Martines2, G Van Oost3, V Svoboda4, T Görler5, T Hansen3

1Institute of Plasma Physics, Association Euratom/IPP.CR, AS CR, Prague, Czech Republic 2Consorzio RFX, Associazione Euratom/ENEA sulla Fusione, Padova, Italy 3Department of Applied Physics, Gent University, Gent, Belgium 4Faculty of Nuclear Science and Physical Engineering, Czech Technical University, Prague 5University of Ulm, Germany

Edge plasma parameters and their fluctuations are studied on the CASTOR tokamak (Bt= 1.3 T, Ip = 10 kA, ne = 10^19 m-3) during biasing discharges. Two probe arrays are used to characterize the edge plasma with a sufficiently high temporal and spatial resolution. A two- dimensional matrix of Langmuir probes (8x8) measures simultaneously in the poloidal and radial directions while a rake of radially distributed Langmuir probes is located at a different toroidal position. Both arrays measure either the floating potential or the ion saturation current. Correlation analysis is used for data processing. In these experiments, a graphite electrode is inserted into the scrape-off layer and biased either by DC or AC voltage with respect to the vacuum vessel. Consequently, a biased flux tube is formed. The projections of the biased flux tube, which emanates from the electrode upstream and downstream along the magnetic field lines are seen as bright patterns in the figure. This contribution is devoted to detailed spatial characterization of these flux tubes at different discharge conditions and their impact on particle transport in the SOL. P-5.019, Friday July 1, 2005

Multifractal analysis of tokamak plasma turbulence in biasing experiments J. Zajac1, V. Weinzettl1, E. Dufkova1, V.P. Budaev2 1 Institute of Plasma Physics, Prague, Czech Republic 2 Nuclear Fusion Institute, RRC Kurchatov Institute, Moscow, Russia

Various data analysis can be performed to study turbulent structures in magnetically confined plasmas. Multifractal analysis complements “traditional” methods (correlation functions, spectral analysis, probability distribution functions etc.) and illustrates namely burst properties of the processes in the plasma. Experiments as well as some stochastic models have shown that behaviour of the transport in the edge plasma in fusion devices should be considered as a multi-scale phenomenon. Large-scale intermittent bursts are followed by formation of coherent structures being a result of self-organization. These structures which can transfer much of energy are signalized by strong correlations and non-Gaussian statistics. Small-scale diffusion produces high-frequency part of fluctuations. The multifractal formalism based on the wavelet decomposition of analyzed data allows one to study scaling properties of the turbulence. The multifractal statistics describes processes as a multiplicative cascade with non-trivial self- similarity. Reducing of fluctuations and turbulent transport by means of externally driven electric field (biasing) has been investigating on the Castor tokamak systematically. The multifractal analysis enables to demonstrate clearly the turbulent properties of plasma during the biasing. Calculated multifractal exponent relates directly to the relative level of large- scale events. Influence of various biasing parameters is shown including creation of low multifractality region in the edge plasma. Positive biasing voltage has strong effect on turbulence suppression in the area between the electrode and the tokamak wall. P-5.020, Friday July 1, 2005

Anomalous impurity diffusion in an experimentally measured turbulent potential R. Pánek1, L.Krlín1, J. Stöckel1, D. Tskhakaya jr.2*, S. Kuhn2, P. Pavlo1, V. Svoboda3, M. Tendler4 1Association EURATOM-IPP.CR, Institute of Plasma Physics, Academy of Sciences of the Czech Republic, CZ-182 21 Prague 8, Czech Republic 2Association EURATOM-ÖAW, Department of Theoretical Physics, University of Innsbruck, A-6020 Innsbruck, Austria 3 Czech Technical University, CZ-115 19 Prague 1, Czech Republic 4Association EURATOM-VR, Alfvén Laboratory, Fusion Plasma Physics, Royal Institute of Technology, S-1044 Stockholm, Sweden *Permanent address: Institute of Physics, Georgian Academy of Sciences, 380077 Tbilisi, Georgia

Using a simple model of tokamak edge-plasma turbulence in the form of a spatially periodic and time-independent electrostatic potential, we have recently found a new type of anomalous impurity diffusion. In the present contribution, we estimate this diffusion for carbon ions (C+) in the test-particle approach for a real turbulent potential obtained experimentally by means of a 2D array of Langmuir probes in the CASTOR tokamak. A significant difference in the impurity dynamics between the Hamiltonian and drift approaches is observed. The drift approximation yields a radial diffusion coefficient about four times lower than the Hamiltonian approach. As an interesting consequence of these dynamics, the possibility of radial-electric-field generation in the turbulence regimes is discussed.

P-5.021, Friday July 1, 2005

Analysis of transient MHD modes of Wendelstein 7-AS by coherence techniques G. Papp1, G. Pokol1, G. Por1, S. Zoletnik2, W7-AS team3 1 BME Department of Nuclear Techniques, Association EURATOM, Budapest, Hungary 2 KFKI-RMKI, Association EURATOM, Budapest, Hungary 3 IPP, Association EURATOM, Garching bei München, Germany

Bursts in the Mirnov-coil signals are frequently observed on the Wendelstein 7-AS stellarator in otherwise MHD inactive plasmas. The duration of these bursts - often referred to as transient MHD modes - is very short, typically on the 100µs timescale, yet they possess a complex spatiotemporal structure. Transient MHD modes can also be observed in the Lithium Beam Emission Spectroscopy (LiBES) signals. In this paper the spatial and temporal structure of these transients is analysed by statistical techniques based on estimation of coherence, phase and cross-correlation functions and on continuous time-frequency analysis methods. These MHD transients are probably connected to anomalous transport events. Only main arguments for this statement are mentioned in this paper. A more comprehensive discussion on this topic can be found in [1]. Given this connection between transient MHD bursts and transport phenomena,

primarily data from discharges exploring the confinement transition around ιa/2π=1/3 edge rotational transform were processed. Several differences in the statistic properties of the transient MHD modes between good and bad confinement states in different experimental series are discussed in detail.

References: [1] S. Zoletnik, et al., Anomalous transport events in the core plasma of the Wendelstein 7-AS stellarator, 32nd EPS Plasma Physics Conference (2005) P-5.022, Friday July 1, 2005

Detection of radially localized and poloidally symmetric structures in the poloidal ﬂow of tokamak plasmas

A. Bencze1, M. Berta2;3, S. Zoletnik1, J. Stockel4, J. Adámek4, M. Hron4

1KFKI-RMKI, Association EURATOM, P.O.Box 49, H-1525 Budapest, Hungary 2Széchenyi Univ. of Applied Sciences,Association EURATOM, Gyor.˝ Hungary 3BME-NTI, Association EURATOM, Budapest, Hungary 4Institute of Plasma Physics, Association EURATOM/IIP.CR, Prague, Czech Republic

A new method to detect plasma ﬂow velocity modulation has been recently developed and theoretically investigated [1]. The method is based on single point correlation measurements,

concentrating on the case when E ¢ B ﬂow velocity dominates the local temporal correlation of

= τ τ turbulence – i.e. vp wp c, where vp, wp and c are the fluctuating poloidal flow velocity, the poloidal correlation length and the intrinsic decorrelation time, respectively. In this case the time variation of the poloidal flow velocity can be directly characterized by modulations in the properly defined half width of the autocorrelation function of the density fluctuations. The above mentioned method was applied for the first time for fluctuation data measured at the Castor tokamak using an appropriate system of Langmuir probes. Poloidally extended and radially localized structures in the modulation of the poloidal flow hav been identified in the edge plasa of ohmically heated low temperature and low density discharges. Their characteristic time scale has been found to be about 1 ms (1 kHz). These observed features closely resemble the zonal flows seen in numerical simulations [2]. Correlated changes in the radial density profile were also detected in good agreement with the expected impact of zonal flows on radial transport. Our results were compared to measurements recently published in [3]. These measurements dealt with poloidal correlations of radial electric field fluctuations and reasonable agreement with our results has been shown.

[1] A. Bencze and S. Zoletnik, submitted to Physics of Plasmas, (2004). [2] Z. Lin, T. S. Hahm, W. W. Lee, W. M. Tang and R.B. White, Science 281, 1835 (1998). [3] G. S. Xu, B. N. Wan, M. Song and J. Li, Phys. Rev. Lett. 91, 125001,(2003). P-5.023, Friday July 1, 2005

Anomalous transport events in the core plasma of the Wendelstein 7-AS stellarator

S. Zoletnik1, N.P. Basse2, A. Bencze1, D. Dunai1, M. Hirsch3, G. Pokol4, G. Por4 1KFKI-RMKI, Association EURATOM, Budapest, Hungary 2Plasma Science and Fusion Center, MIT, Cambridge, USA 3IPP, Teilinstitut Greifswald, Association EURATOM, Greifswald, Germany 4BME Institute of Nuclear Techniques, Association EURATOM, Budapest, Hungary

Over the past years data from several diagnostics have revealed the existence of transients both in the profiles and turbulence on the Wendelstein 7-AS stellarator. The radial extent and statistical properties of these phenomena show high sensitivity to the rotational transform close to low order rational values. As the energy confinement time is also known to change considerably around low order rationals[1] it appears to be important to analyze the relation between transients and confinement, and details of the transients themselves. Experiments revealed that these transients are associated with the following events:

• Flattening of the temperature profile along a whole flux surface on a sub ms timescale followed by a relaxation.[2] • A poloidally symmetric, about 100µs long burst in the small-scale density turbulence.[3] • An ≈ 100µs long burst in the MHD activity of low-m modes[4] seen by both magnetic pick-up coils and edge density fluctuation measurement[5].

Although it is appealing to consider the burst of MHD modes as driving radial transport, this is most probably not the case as examples are found where MHD modes are absent[5]. Instead we try to reconcile the above observations by the following hypothesis. Some yet unknown transport event (e.g an explosive instability[6]) opens up an effective transport channel between two flux surfaces. It is highly unlikely that this happens in a poloidally symmetric way. The resulting poloidal pressure asymmetry will smooth out on one flux surface on an ion transit timescale which is several 10 µs. During this time the plasma is not in MHD force equilibrium, therefore an MHD wave appears as the harmonic response of the plasma. As soon as the perturbation smoothes out along flux surfaces steeper density gradients appear at the edges of the flattened radial region which results in an increase of the small-scale density fluctuation amplitude in a poloidally symmetric way. The paper presents an overview of the above transients, and discusses their rele- vance to transport.

References. [1] R. Brakel et al. Nucl. Fusion 42,903 (2002) [2] M. Hirsch, et al. Plasma Phys. Control. Fusion 42, A231 (2000) [3] N.P. Basse, et al. Phys. Plasmas 12 012507(2005) [4] G. Pokol, et el al. IAEA Conf. 2004 [5] S. Zoletnik, et al., Phys. Plasmas 6, 4239 (1999) [6] H.R. Wilson, S.C. Cowley, Phys. Rev. Lett. 92 175006 (2004) P-5.024, Friday July 1, 2005

Effects of magnetic shear on confinement in TJ-II ECRH discharges

D. López-Bruna, F. Castejón, T. Estrada, J. A. Romero, F. Medina, M. Ochando, A. López- Fraguas, E. Ascasíbar, J. Herranz, E. Sánchez, E. de la Luna, I. Pastor

Laboratorio Nacional de Fusión –Asociación Euratom/Ciemat, 28040-Madrid, Spain

Traditionally, the evolution of the rotational transform ! in magnetic confinement devices has been related either to tokamak physics or to bootstrap current in stellarators. In the latter case, the main purpose has been controlling !/2∀ so as to avoid low order rationals inside the plasma, which are especially deleterious in low shear machines (see, e. g. [1]). On the other hand, it is precisely in these confinement devices that the evolution of the rotational transform and its shear can be best estimated as long as bootstrap currents can be ruled out. In the present work, we show this evolution with corresponding transport analysis for TJ-II discharges with moderate (negligible modification of the magnetic field strength) ohmic induction where, nonetheless, the bootstrap current is presumed to be of second importance in the perturbed rotational transform. The current profile has been estimated by assuming that the resistivity obeys a Spitzer-like dependence on electron density and temperature, corrected by the ripple of the device. The appropriateness of Spitzer resistivity is found after checking that our measured loop voltage (along the central conductor) can be considered as a diagnostic of plasma loop voltage: the evolution of plasma current density imposing the experimental net plasma current as boundary condition yields loop voltages in fair agreement with the diagnostic when the resistivity is as mentioned above. The TJ-II heliac, which is characterised by an almost flat vacuum ! profile [2], offers valuable information on the influence of the magnetic shear in electron cyclotron resonance heated discharges because it remains rather uniform across the bulk plasma during the ohmic induction process. We find that heat and particle transport are affected differently by magnetic shear: positive shear increases particle transport while heat transport seems to be less affected or increases with shear. On the other hand, negative shear systematically improves particle transport while heat transport remains roughly the same or, if any, diminishes giving rise to an overall enhancement of confinement.

[1] R. Brakel et al., Plasma Phys. Control. Fusion 39, B273–B86 (1997). [2] C. Alejaldre et al., Fusion Technol. 17, 131 (1990). P-5.025, Friday July 1, 2005

Direct evidence of coupling between density tails and turbulent transport in the scrape-off layer region in the TJ-II stellarator

E. Calderón, M.A. Pedrosa, C. Hidalgo, M. Liniers, A. Fernández, C. Fuentes, A. Cappa EURATOM-CIEMAT, 28040-Madrid, Spain P. Balan, R. Schrittwieser Institute for Ion Physics, University of Innsbruck, Innsbruck, Austria The importance of intermittent plasma turbulence to explain non-exponential decays of the plasma parameters profiles in the scrape-off layer (SOL) has been investigated in tokamak plasmas [1]. This paper shows a direct link between non-exponential SOL decays and the statistical properties of turbulent particle flux in the plasma boundary region of the TJ-II stellarator. First comparative studies of the structure of turbulence in ECRH (200 - 400 kW) and NBI (400 kW port-through) plasmas in the TJ-II stellarator have shown a drastic decrease in the level of turbulence in the transition from ECRH to NBI plasmas. Frequency spectra show that the reduction of fluctuations is due to the decreasing in a wide range (10 – 200 kHz) of frequency components. Perpendicular phase velocity of fluctuations strongly depends on plasma density in ECRH plasmas and increases up to 2 km/s in NBI regimes. The radial velocity of transport decreases by about a factor of ten in the transition from ECRH to NBI heated plasmas and SOL non-exponential tails usually observed in ECRH plasmas disappear in the NBI regime. These findings provide the first direct coupling between radial velocity of turbulent transport events and the structure of SOL profiles. The SOL plasma response is different at densities below and above the threshold value to trigger the spontaneous development of ExB sheared flows [2, 3]. At low densities (i.e. without edge velocity shear layer) ion saturation current measured in the SOL increases with density. Above the threshold value (i.e. once the edge sheared layer is developed) the ion saturation current remains almost constant in the SOL. As plasma density approaches the highest NBI density (4 x 1019 m-3), an increase in the level of fluctuations and ExB transport has been systematically observed in the SOL region. Considering that these densities fall within the stellarator density limit scaling law, this result can be interpreted in terms of the increasing of edge transport near the density limit of TJ-II. [1] J.A. Boedo et al., Phys. of Plasmas 10 (2003) 1670 [2] C. Hidalgo et al., PR-E 70 (2004) 067402 [3] M.A. Pedrosa et al., PPCF (2005) in press P-5.026, Friday July 1, 2005

Spatial distribution of lost ripple-trapped suprathermal electrons. F. Medina, M. A. Ochando.

Laboratorio Nacional de Fusión. Asociación EURATOM/CIEMAT, 28040 Madrid, Spain The ECRH driven convective flux of ripple-trapped suprathermal electrons in low density stellarator plasmas has been mentioned as being related to various features and phenomena: the suprathermal feature and the burst-like phenomena in the ECE emissions [1], the electron-root feature [2] and the phenomenon of electric pulsation [3]. In a previous investigation on TJ-II plasmas [4], fast transitions in confinement regimes were found to be associated to changes in the electron distribution function and to changes in the direct convective losses at the heating region. There, low order rational surfaces located at the gradient region seemed to be affecting the trapping/de-trapping rates. It was also reported [5] that the plasma potential at the core region increased with the characteristic energy of the electron suprathermal electron tail. Here we present the spatial distribution of electron losses to the vacuum vessel detected through their bremsstrahlung emission in the range of soft x rays. Electron losses are monitored with the SXR tomography diagnostic [6] using thick Be filters to reduce the contribution of plasma emission. The energy range of the detected photons was the expected for the electron energy distribution function at the plasma core in the studied TJ-II plasmas. Three distinct losses regions were observed at the inboard side of the vessel wall. The emissions coming from each of them behave differently with electron density, heating method (ECR on-axis vs. off-axis) and magnetic configuration. By monitoring the relative intensity between the three bands of losses it is intended to extend to the plasma boundary the assessment of the effect of low order rational surfaces on the ripple particles losses. The experiments were carried out, in steady state plasmas as well as in plasmas showing transitions between different confinement regimes for several magnetic configurations.

[1] Romé M et al 1997 Plasma Phys. Control. Fusion 39 117 [2] Kick M et al 1999 Plasma Phys. Control. Fusion 41 A549 [3] Fujisawa A et al.1999 Plasma Phys. Control. Fusion 41 A561 [4] Ochando M A and Medina F et al Plasma Phys. Control. Fusion 45 221 (2003) [5] Medina F. et al Proc. 31st EPS, London 2004) (in CD-Rom) [6] Medina F. et al, Rew.Sci.Instrum. 70, 642 (1999) P-5.027, Friday July 1, 2005

High-speed turbulence imaging in TJ-II edge plasmas.

J.A. Alonso, S.J. Zweben1, H. Thomsen2, C. Hidalgo, T. Klinger2, B.Ph. van Milligen, J.L. de Pablos, M.A. Pedrosa Laboratorio Nacional de Fusión, Asociación EURATOM-CIEMAT, 28040 Madrid, Spain 1 Princeton Plasma Physics Laboratory, Princeton, NJ, USA 2 Max-Planck-Institut für Plasma Physik, EURATOM Ass., 17491 Greifswald, Germany

Transport in fusion devices is a phenomenon with high degree of complexity. Two-dimensional images of edge plasma turbulence have been obtained by high-speed imaging in the visible range in the edge of tokamak devices [1], [2]. Recently, a 2-D visualization of transport has been investigated in the plasma edge of the TJ-II stellarator. A Princeton Scientific Instruments intensified camera with CCD sensor (PSI-5) was used with Hα filter, achieving recording ratios up to 250.000 frames per second. The storage capacity is 300 frames with 64 by 64 pixels resolution, thus giving 1.2ms total recording time at maximum speed with an image every 4µs. The view plane is in a near-poloidal cross-section with optimized B-field perpendicularity. Neutral recycling at the poloidal limiter is used to light up the outer plasma region (ρ ∼ 0.8- 1). Bright, long-living structures are frequently seen with a spatial extent of few centimeters (see Fig.1). Those structures, previously referred to as “blobs”, show predominant poloidal movements with typical speeds of 103 - 104 m s−1 in agreement with the expected E × B drift rotation direction. Figure 1: Blob moving poloidally over four frames. 0µs 4µs 8µs 12µs

Limiter

out 11cm

Image analysis techniques are being implemented for characterising blob geometry (aspect ratio, orientation...) in TJ-II different velocity shear regimes. Automatic target detection using 2-D directional wavelets seems to be the most suitable analyzing tool.

References [1] Zweben S.J.et al, Nucl. Fusion 44, (2004) 134.

[2] Terry J.L.et al, Phys. Plasmas 10, (2003) 1739. P-5.028, Friday July 1, 2005

Electron internal transport barriers, rationals and fluctuations in the TJ-II L. Krupnik1, J.A. Alonso, A. Cappa, A.A. Chmyga1, N.B Dreval1, L. Eliseev2, T. Estrada, A. Fernández, C. Fuentes, C. Hidalgo, S.M. Khrebtov1, A.D. Komarov1, A.S. Kozachok1, A.V.Melnikov2, I. S. Nedzelskiy3, M. Liniers, J. L. de Pablos, M.A. Pedrosa, Y. Tashchev3, V. Tereshin1, E. Sánchez Laboratorio Nacional de Fusión. EURATOM-CIEMAT, 28040 Madrid, Spain 1Institute of Plasma Physics, NSC KIPT, 310108 Kharkov, Ukrain 2Institute of Nuclear Fusion, RNC Kurchatov Institute, Moscow, Russia 3 Associação EURATOM/IST, Centro de Fusão Nuclear, 1049-001 Lisboa, Portugal

The influence of the magnetic topology on e-ITB formation has recently been studied in the TJ-II stellarator. A configuration scan (changing iota from 1.55 to 1.61) has shown that the plasma current value at which this transition takes place depends on the magnetic configuration. This result points to the presence of a low order rational surface (m=3 / n=2) close to the plasma core as being a necessary condition for triggering e-ITB formation [1]. During the e-ITB formation, the plasma potential, as measured by HIBP, increases in the plasma core region, increasing the radial electric field in a factor of three. Simultaneous measurements of the total beam intensity (proportional to plasma density) indicate that, at the transition, plasma density profiles are more hollow.

A recent improvement in the signal to noise ratio of the HIBP system has allowed to characterize the radial structure of plasma modes from the edge to the plasma core region in the TJ-II stellarator. Quasi-coherent modes (10–20 kHz) have been recently identified in the HIBP profiles during the development of e-ITB. These modes are clearly observed in the beam current (20 kHz) at the plasma region where core ExB sheared flows are developed. These modes are also observable in the ECE signals localized around ! ! 0.3, close to the foot of the e-ITB. Present results show a decreasing in the mode amplitude as e-ITBs are fully developed. This finding points out the role of MHD modes to trigger core plasma transitions.

A magnetic configuration scan shows evidence of quasi-coherent modes (! 20 kHz) in the plasma edge gradient region. HIBP shows a strong correlation with Langmuir probes and Mirnov coil measurements [2]. Configuration and density scans have shown that these modes appear in a narrow configuration window and their amplitude increases with plasma density. These results suggest the combined role of threshold gradients and low order rationals.

[1] T. Estrada et al., Plasma Physics and Cont. Fusion 46 (2004) 277. P-5.029, Friday July 1, 2005

Effect of suprathermal electrons on impurity ionization state

M. A. Ochando, F. Medina, B. Zurro, K. J. McCarthy, J. A. Jiménez, A. Baciero, D. Rapisarda, J. M. Carmona and D. Jiménez Laboratorio Nacional de Fusión. Asociación EURATOM/CIEMAT, 28040 Madrid, Spain

A large part of the experimental investigations in TJ-II are dedicated to obtaining more evidence concerning the effect of rational surfaces, or the associated chains of islands, on particle transport and confinement. For instance, in previous impurity transport studies, robust structures in chord-integrated radiation profiles of the 227.1 nm line emission of CV were detected at radial positions where the rotational transform has a rational value [1]. It was also found that the relaxation of the global emission perturbation caused by non-intrinsic impurity ions could not be fitted to a pure exponential function [2]. Moreover, the adjusting parameters needed to describe the intensity decay showed a non-monotonic dependence on plasma minor radius. On the other hand, near the periphery (from 0.6 to 0.9 in effective radius), it was found that rational surfaces are good confinement regions for high-energy electrons, that is, the radial transport of such electrons cannot be expressed as smooth functions of minor radius [3]. In recent laser blow-off (LBO) impurity injection experiments performed in ECRH plasmas, radiation signals showed some unexpected behavior: e.g., high-energy (10-100 keV) photon fluxes increased and global emissivity had an atypical ‘temporal (and toroidal) structure’. Furthermore, highly ionised states of iron (Fe XX and Fe XXI) were detected. The experimental data presented are interpreted as evidence of the over-ionisation of iron in the plasma periphery resulting from the interaction of ingoing impurities with suprathermal electrons. As the ‘suprathermal effect’ seems to be a rather universal phenomenon, this circumstance should be considered when impurity transport coefficients are evaluated in the frame of diffusive/convective models (i.e., the inward convective velocity is proportional to the charge state of the impurity [4]), and when the time dependent diffusion coefficient is investigated in turbulent plasmas (i.e., the Larmor radius has a strong effect on impurity ion transport [5]).

[1] A. Baciero et al., Plasma Phys Control. Fusion 43, 1039 (2001). [2] B. Zurro et al., 29th EPS Conf. Plasma Phys. Control. Fusion, Vol. 26B P-5.025 Montreux 2002. [3] F. Medina et al., Rev. Sci. Instrum. 72, 471 (2001) [4] H. Nozato et al., Phys. Plasmas 11,1920 (2004) [5] M. Vlad et al., Plasma Phys Control. Fusion 47, 281 (2005) P-5.030, Friday July 1, 2005

Experimental investigation of ExB sheared flow development in the TJ-II stellarator M.A. Pedrosa, C. Hidalgo, J.A. Alonso, A. Cappa, E. Calderón, A.A. Chmyga(1), N.B. Dreval(1), L. Eliseev(2), T. Estrada, A. Fernández, L. Krupnik(1), A.V.Melnikov(2), O. Orozco, J.L. de Pablos, S.J. Zweben(3) Laboratorio Nacional de Fusion, Euratom-Ciemat, 28040 Madrid, Spain (1) Institute of Plasma Physics, NSC KIPT, 310108 Kharkov, Ukraine (2) Institute of Nuclear Fusion, RNC Kurchatov Institute, Moscow, Russia (3) Princeton Plasma Physics Laboratory, Princeton, NJ, USA Recent experiments in the TJ-II stellarator have shown that the generation of spontaneous perpendicular sheared flow (i.e. naturally occurring shear layer) requires a minimum plasma density. Near this threshold density, the level of edge turbulent transport and the turbulent kinetic energy significantly increase in the plasma edge. Above this threshold value, the level of turbulence decreases with a concomitant development of ExB sheared flow that can be observed as a sharp change of the floating potential [1, 2]. The physics of ExB sheared flow is being investigated by means of multiple diagnostics in the TJ-II stellarator, including Langmuir probes, Ultra Fast Speed cameras [3], Heavy Ion Beam (HIBP) diagnostic [4] and reflectometry [5]. The reversal in the ExB rotation first reported by means of Langmuir probes measurements, has been recently 2-D visualized by means of Ultra Fast Speed cameras. Both diagnostics show consistent results: perpendicular phase velocity of fluctuations and estimated blobs velocity, which rotation is predominantly in the perpendicular direction, are in the range of 103 – 104 m/s [3]. HIBP measurements in ECR heating plasmas show that the plasma potential increases up to 1 kV near the magnetic axis (i.e. radial electric fields are radially outwards) [4]. In addition, the plasma potential shows a strong dependence on the plasma density. At plasma densities near the threshold value evidence of radially inwards edge radial electric fields has been observed, whereas in the plasma core radial electric field (as measured by the HIBP system) remains positive. This result implies the simultaneous development of two sheared flows at the threshold density: one located in the proximity of the LCFS (r/a ! 1) and the other one near r/a ! 0.7. Reflectometry measurements have permitted the characterization of the velocity shear layer that develops spontaneously in the edge of TJ-II plasmas above a certain critical density [5]; experimental findings shows that the inner shear layer moves inwards as density increases above the threshold. The link between the development of sheared flows and plasma density in TJ-II has been observed in different plasma magnetic configurations and plasma regimes. The systematic investigation of the effect of the iota value and the magnetic ripple on the threshold value of the density is under study. Those results have a direct impact in the understanding of the mechanisms underlying the generation of sheared flows in fusion plasmas.

[1] C. Hidalgo et al., Phys. Rev. E 70, 067402 (2004) [2] M.A. Pedrosa et al., Plasma Phys. Control. Fusion, in press [3] A. Alonso et al., This Conference [4] L. Krupnik et al. Proc. 31th EPS Conf. (2004). [5] T. Estrada et al., This Conference P-5.031, Friday July 1, 2005

Influence of ExB sheared velocities in the statistical properties of fluctuations in the plasma boundary of the TJ-II stellarator R.O. Orozco1, E. Faleiro2, C. Hidalgo1, M.A. Pedrosa1, E. Sánchez1, J.M. Gómez3, E. Calderón1 1 EURATOM-CIEMAT. Madrid, Spain 2 Dpto. de Física Aplicada. EUIT Industrial. UPM. Madrid, Spain 3 Dpto. de Física Atómica Molecular y Nuclear. UCM. Madrid, Spain The statistical properties of plasma fluctuations have been investigated in the plasma boundary region of fusion and non-fusion plasma devices [1]. Those experiments have shown that fluctuations have a non-Gausssian character, mainly, in the scrape off layer (SOL). Furthermore, fluctuations show sporadic pulses whose time asymmetry changes in the velocity shear layer regions, and such changes may be generated via coupling between fluctuations and the shear flow. This work reports the investigation of the influence of ExB sheared velocities and magnetic topology in the statistical properties of turbulence in the plasma boundary region of the TJ-II stellarator. Experiments were done in plasma regimes with and without edge velocity shear layer [2, 3]. In agreement with previous findings, fluctuations in ion saturation current show clearly a non Gaussian behaviour with an additional feature; moving down into the Last Closed Flux Surface (LCFS), its non gaussianity increases in the SOL region and decreases in the proximity of the LCFS. Fluctuations in floating potential show sporadic pulses that are asymmetric in time. The degree of asymmetry is minimum near the LCFS and increases in the SOL region. However, time asymmetry increases in plasma regimes without edge sheared flows. Fluctuations in density scan and ECRH power modulation were also investigated. Results show that the deviation from a Gaussian distribution is dynamically coupled with the shift of velocities in the proximity of LCFS region, pointing to sheared velocities as a mechanism involved in the modification of the statistical properties of turbulence. Present findings show the importance of both magnetic configuration and sheared velocities to determine the statistical properties of fluctuations in the plasma boundary region, calling into question the leading role of competition between fluctuation driving and decorrelation mechanisms at the shear layer location to explain the time asymmetry of turbulent bursts reported in fusion plasmas.

[1] E. Sánchez et al., Phys. Plasmas 7, 1408 (2000) [2] C. Hidalgo et al., Phys. Rev. E 70, 067402 (2004) [3] M.A. Pedrosa et al., Plasma Phys. Control. Fusion, in press P-5.032, Friday July 1, 2005

Local transport in density and rotational transform scans in TJ-II ECRH discharges

V. I. Vargas, D. López-Bruna, T. Estrada, E. Ascasíbar, E. de la Luna, M. Ochando, F. Medina, J. Herranz, A. Fernández, E. Sánchez

Laboratorio Nacional de Fusión –Asociación Euratom/Ciemat, 28040-Madrid, Spain

Aside from a number of analysis devoted to the study of particular discharges, systematic local (flux surface averaged) transport in the TJ-II heliac remains to be done in connection with confinement time scalings. In this work we present the results of performing interpretative transport on two series of electron cyclotron resonance heated discharges, one belonging to a rotational transform scan, the other one to a density scan. After boronisation, TJ-II matches reasonably well [1] ISS95 scaling [2] with rotational transform but the density dependence found in TJ-II (exponential factor 1.06) is much stronger than the one in ISS95 (exponential factor 0.51). The dependence on rotational transform is weaker (0.34 instead 0.4) although still marginally congruent with ISS95. This support is relevant because the !/2∀- range in TJ-II data is considerably wider than the ISS95 one. The ISS95 dependences for TJ- II thermal confinement are commented in light of the estimated thermal diffusivities for the two scans. We also present estimates of particle transport in the bulk plasma even though its interpretation is strongly affected by uncertainties on the particle source profiles. Finally, we present some results using a simplified transport model [3] that seems to capture the radial dependences of the experimental heat transport coefficient profiles. According to this model, transport increases strongly with trapped particle fraction and linearly with collisionality.

[1] E. Ascasíbar, T. Estrada, F. Castejón et al. “Magnetic configuration and plasma parameter dependence of the energy confinement time in ECR heated TJ-II plasmas”. Nuclear Fusion In press. [2] U. Stroth et al., Nucl. Fusion 36, 1063 (1996). [3] B. B. Kadomtsev, O. P. Pogutse, Nucl. Fusion 11, 67 (1971). P-5.033, Friday July 1, 2005

Dynamical coupling between parallel flows and radial turbulent transport in a linear plasma machine O.F. Castellanos, E. Anabitarte, J. M. Sentíes Departamento de Física Aplicada. Universidad de Cantabria, 39005 Santander, Spain

The mechanisms underlying the generation of plasma flows play a crucial role to understand transport in magnetically confined plasmas ([1,2] and references therein). The best performance of existing fusion plasma devices has been obtained in plasma conditions where E×B shear stabilization mechanisms are likely to play a key role: both edge and core transport barriers are related to a large increase in the E×B sheared flows.

The possible link between turbulence and flow generation have been discussed both from the theoretical and experimental point of view. Sheared flow can be generated when turbulence driven flows play a dominant role in the momentum balance. In this case, sheared perpendicular flow is expected to increase when the turbulent energy is large enough to overcome the flow damping. Recent experiments in the TJ-II stellarator have reported that development of the naturally occurring perpendicular velocity shear layer requires a minimum plasma pressure gradients providing an experimental evidence of coupling between sheared flows development and increasing in the level of edge turbulence.

In this context, experimental evidence of parallel flows dynamically coupled to radial turbulent transport are reported in a linear plasma machine. Probability Density Functions (PDFs) for parallel flows and transport are different. PDFs for transport show clear non- gaussian features with large and sporadic burst and PDFs of parallel flows are rather gaussian. Plasma profiles of parallel Mach number (in the range 0.1 – 0.3) show gradients

dMparallel / dr comparable to 1/Ln. Experimental results show that the dynamical coupling between transport and flows shows differences at different plasma radii. In particular, the behavior of this dynamical coupling with the microwave incident power has been investigated. Microwave incident power is an effective tool in order to modify the profile of the expected Mach numbers. The link between radial turbulent transport and parallel dynamics suggest a role of turbulence in the parallel dynamics. [1] Burrell K 1999 Phys. Plasmas 6 4418 [2] Terry P W 2000 Rev. Mod. Phys. 72 109

P-5.034, Friday July 1, 2005

¡£¢¥¤§¦©¨ ¦£¤¥¢ ¢¥ ¡¨ ¤§¦£¢§©¦£ ¢§ £¤¢§¦ £¡ ¨ ¦£¢

¤ £¨ ¤§¦£ ¤§¡£¢ ¤§ ¡ ©£! £¢¥ ¡¢¥¤¥"#¡£ $ £¤ ¡

£¢¥¤¥%¡©©¤§¦£ " !&¢§©¦£§£!¥¤§¡£"©¤

' ( )§( *+ , - .0/1(32 4 - 5 + 464 708 9 (;:;4 705 <0, =

>@?0A B C D E A F0G0F¥H@G0D I J;EK K K@F0CLG0F0D A F3M

B F0G0O;F0CPI G©>@?0A B C D E A F0G0F¥Q;R S TC UVG0?0W E S LG0F0D A F0S0X Y@G3A ?

Z[, - \ ]3- ^04 \ + _0,`\ a0b;4 c§4 bd\ - 4 7;e f;a0- \, g3f;, - + c¥, 73\ e£^04 e , 8da07f;a0h , -£c§a;83];i 4 \ + a07\ , 5 <070+ j3];,0e

+ , i 8+ 7c¥4 7 5 4 e , e@- , e ];i \ e1a;l 80+ l l + 5 ];i \ + 73\ , - f0- , \ 4 \ + a0783];, \ a¥\ <0, 5 i aVe , 7;, e e1a;l\ <0, f;i 4 e c4

k k

f0- a0l + i ,3e@\ a4 i a;5 4 i[5 - + \ + 5 4 im\ <3- , e <0a;i 8£n o . p;q (0r 7¥\ <0+ e1h a0- b0.Vh , ]Ve ,§4 \ ]0- ^3];i , 705 ,§5 a;80, ^;4 e , 8

a07d\ <0,s70a070i + 7;, 4 - , g05 + \ 4 \ + a07ta;l©83- + l \h 4 _;, e + 7t\ <0,s80+ e e + f;4 \ + _0, u \ - 4 f0f;, 80u , i , 5 \- a070u c§a;80,

v w xy x

* z£- , {0+ c¥, 7t+ 7&5 i + 7;83- + 5 4 i¥{V, a0c¥, \- n |0q©\ a}, g3f;i a0- ,`\ <0+ e + e e ];,3( <0,sf0- a0l + i ,s+ e

k k

4 i i a0h , 8©\ a£, _0a0i _0,e a©\ <04 \180, f;, 7080, 7;5 + , e a0l1{;i a;^04 i \ - 4 7;e f;a0- \@h + \ <©8 7;4 c¥+ 5 4 ifV4 - 4 c¥, \, - e

- , i 4 \, 8£\ a + \ .@4 e / ~ .@5 4 7 ^0,£e \ ];80+ , 8V( <0,©5 - + \ + 5 4 i@\ <3- , e <0a;i 8 + e - , i 4 \, 80. + 7 \ <0+ e 5 43e , . \ a4

5 , - \ 4 + 7ti a;5 4 i_04 i ];,sa0l\ <0,`f;4 - 4 c¥, \, - /~ / .§4 ^0a0_0,`h <0+ 5

5 <04 707;, im+ e14 5 \ + _04 \, 8§70, 4 - \ <0, - , e a0704 73\[- 4 \ + a0704 ie ]0- l 4 5 , e (0`, <;4 _0,6f;, - l a0- c§, 8c§a;83];i 4 \ + a;7

, g3f;, - + c¥, 73\ e6h <0, 7©\ <0,©e e \, c+ e 83- + _0, 7£_;, - l 4 - 4 708£_;, - 5 i ame ,¥\ a£+ \ e 5 - + \ + 5 4 i f;a;+ 73\1^

k k k k

5 <04 73{;+ 70{\ <0,£^04 5 b0{0- a0]070880+ l l ]Ve + _;+ \ ( <0+ e¥4 i i a0h¥e¥l a0-¥4 7, 4 e + , -¥+ 73\ , - f0- , \ 4 \ + a07a;l ^0a0\ <

4 c f;i + \ ];80,64 708 f0<043e ,Pf0- a0f;4 {;4 \ + a07§80+ 4 {0- 4 ce (

a0c f;4 - + e a07Ve}a0l\ <0, e ,- , e ];i \ eth + \ <f;a0h , - u ^04 i 4 705 ,\ - 4 7;e f;a0- \s- , e ]0i \ eth + i is4 i e a ^;,

80+ e 5 ]Ve e , 8V(

x v

C ;G0I IC G0D E A J0?¥3

n o q[( 9 \, - .;2 ( 4 - 83+ 70+ . . 60 .[o |00 p;00|3z

C ;G0I IC G0D E A J0?¥3

n p;q[§( /1( a0];.02§(326+ - ]0= = + . . 60 .o o0o p;00|3z

w x

E O0@I A F;E

n |0q ( )§( 4 - - , - 4 e .V§(V:;+ 80+ b;c¥4 70.;Z ( ¥( + 4 c§a;7080.;Z ( ( , - - 4 708¥/1(02 4 - 5 + 4 .3@

.0|[o0o o 03p0z

P-5.035, Friday July 1, 2005

¢¡£¡£¤¦¥¨§ © ¡ ©© ¡©© ¡ ¥¨§£ §£©

.0/213 546+8798:;+<.0/<$= %1?>¨@¦&8ACBED37GFH¨+JI ! "# %$'&!(*)¦+!,¨-

.KMLNPO!QJRSTNVU¨WÜ=XMW¨RTYPZS\[[[]^¨_¨àäcb?QTdW¨LeJS]!fgWÜ¨RTNVU¨]5hi WNVL

j¦k k'v

STZ¨l!NVWl!NPm

wyxzP{ z;{ xz z z x { x z { z

@¦(P&¢-¨H¦&!B, &!H -¨H¦+!(|9J}( H¦,57 9!+!(u~E7+8H @¦-¨7G~ -¦,¨&8(\-¦&!,5)¦&'~E7+8H @¦-¨7G~ +877 &87 -¨7 +~ -¨H

x { x z x xz z {¢z x

+ ,¨&J&!(;-¨@¦&!, 8 \¨& -¦,¨&!( + +8@¨@¦( &!,~G-~E¨& ~Fq,<}-¨c~E¨&=1~G-~E7+8H ~ -¨H,<}

zx x z z xz z { x|z x z z

\ -¨7 + +)¦&8H¨&87+!( AC+~ -¨H-¨(;-¦9!+!(<~E7+8H ~ -¦H -¦,¨&!( H<q& ~ )¦+~G&!,&!+87( &87 P¦

xz { z z x z x xz z z

\¨&+ 9 -¦,¦&!(|9!-¨H2~G+ H¨&!,&!2Fq+~ -¨H -¨73~E¨&¢(P&J&!(|-¦(Fq9~Fq+~ -¨H /\,¨&8H ~}@¨7-¦ (P&¢+!H¨, -¦H

{ z x z z zx

~G& @¦&87+~F7&@¨7-¨ (P& ?#&¨+J&+!,¨,¨&!,+H¨& &!2F+~ -¨H-¨7t~E¨&+&87+8)¦&!,@¦-¦(;- ,¨+!((;- |\

z z x { z { { z { z { x z xz

&!2Fq+~ -¨H9!-¨H2~G+ H +(;- ,¨+ @ H¨)~G&87 ,2Fq&~G- +8)¦H¨&~ 9@

{ x x xx z xxz x x xx {£x zPx z x

~G&87 /+8H¨,=~E¨&t ¡&J}

z z { { { x z

9!-¨H¨, ~ -¨H-¨~E-2~G+8(¦+!H5)

z x z z z x

¥¦-¨7, &87&!H2~+!(Fq& -¨'~E¨& H5@

z z x z z x x zPx z zx

~E @¨7-¨ (P& +!H¨,@¦-¨(;- ,¨+!((;- FH2~ (+ ~G&!+!,<} ~G+~G& 7&!+!98¨&!, |\¨&9!+!(P9Fq(P+~ -¨H 7&8@¦&!+~G&!,

z x z z x6x xx { zx x

, &87&!H2~ +!(Fq& -¦M~E¨&¢9!-¦&! 9 &!H2~3-¨M~E¨& ¡&J}

xx xx z z x x

~E¨&6&!&!9~?-¨u ¡&J}

(;-

z z { z

¥¦-¨7¨+© ¢q&!,ª(;- +!(Fq&-¦3~E¨& H¨@

t« «

z xzx x{ xx xx z z zPx w¨x

(Fq9~Fq+~ -¨H &87¤} +!(P( ¨&!H~E¨&# ¡&J}

z z z x x z x z {¢z x x x x { x

9!-¦&8 9 &!H2~ H¨987&!+ & / ~37&!+!9!¨& +¢@¦- H2~ ¦&87&~E¨&,<}

x z zP{ z x x zP{z z z z

- 9 (;(P+~G-¨7G}7&8) &!/ 9!9!-¨77& @¦-¨H¨, ~E-~E¨&¨( ~c9J}<9!(P&6@&87 -¦, 9t&!¦+J -¨7-¨|+6@¨7&!,¨+~G-57B@¨7&}

{ zx z;{ z z x x x

-¦,¨&!( 5§Ht~E 7&8) &!/<~E¨&7+!, +!(+J&87+8)¦&6-¨~E¨& (F9~Fq+T~ -¨H(P&J&!(,¦&!987&!+ & /¦+!H¨,t~E¨& ¦&!+87|(;-

z x x w x x x xx z z {¢z x x

H¦987&!+ & -<&+t~E57& ¨-¦(;,+!(F&-¨?~E¨&' ¡&J}

x z x x z x x z zx z x

Fq,¨,¨&!H¨(V} H¦987&!+ & +!H¨,~E¨&©(Fq9~Fq+~ -¨H +87&,57+ ~ 9!+!(P(V}7&!,2Fq9!&!, \\ +8)¦+ H \\¨& ¨&!+87¨(;-

z z;{ z z x¨z z x xz z z zx xzx

~E&J¢<@&87 &!H2~G+!( H¦, H5) H~E¨& ~E+ &!9!-¨H¨,¨B-¨7,¨&87¡~07+8H ~ -¦H+!H¨, &!¨+J -¨7 9!-¨H ~G&!H2~

\ £B§§ ~G&8(;(P+87+~G-57 ®¨

¯¨°8±°C²!°!³ ´!°!µ

x xxz

8 ! %"# J$'&8(*)+!,¨-¦/¦&T~¦+8( /¦®¨¶!· ¸¹|ºM» ¨»2~¤ %ºu&T~G&87 5F7)¦/2 F + ª¦¶¨¶¨®

z { x

P¦ºM %$ + -¦H¨,¨/¦&~¦+8( /<º¦<} % ¡&J %1M&~P~¼ C½5¾¦/5¿5À¨¿¨ª88Á¨Á5Â

z x

®¨Ã % ,¨+!(;)¦-¦/¦&T~¦+8( /<º¦<} % ¡&J %¹Ä½¨Å5/¦¶¨À¨Æ5Â¦¶5ª¨¶¨¶5Â « P-5.036, Friday July 1, 2005

Theoretical transport analysis in TJ-II scenarios with enhanced heat confinement

J. Dies1, F. Castejón2, J. Garcia1, J. Izquierdo1 1. Fusion Energy Engineering Laboratory (FEEL), Departament de Física i Enginyeria Nuclear, ETSEIB, Universitat Politècnica de Catalunya (UPC), Barcelona, SPAIN

2. Laboratorio Nacional de Fusión, Asociación EURATOM-CIEMAT, Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas, Madrid, SPAIN

Experimental data show that high central electron temperature with high temperature gradients are obtained in TJ-II stellataratori for high absorbed power density. These scenarios are the so called enhanced heat confinement (ECH) regimesii. These plasmas are heated with ECRH and have low central electron density and flat or hollow density profile. In order to study these shots, the code PRETOR-Stellaratoriii has been used. It was developed in the “Departament de Física i Enginyeria Nuclear” of “Universitat Politècnica de Catalunya” DFEN with a new module capable to take into account, to some extent, the stellarator geometry and, hence allowing to simulate TJ-II shots. TJ-IIi is a medium size four periods stellarator, of flexible heliac type, with magnetic

field Bt0ø1.2T, major radius R0=1.5 m, average minor radius ø 0.2 m. TJ-II has an almost flat rotational transform profile that can achieve a wide range of rotational transform values at the center: 0.9ø 畏緯尉/2 ø2.2. A new transport model that accounts for the anomalous transport suppression due to sheared electric field has been added to the mentioned transport code. The structure of this electric field is studied with theoretical models. First results show that even there is only one solution to the ambipolar equation and the transition between electron root and ion root should be soft, a reduction of anomalous transport to neoclassical values in the

range 0< ７<0.15 for the enhanced heat confinement shots where radial electric field Er

is large and dEr/dr > 0. Nevertheless more work is necessary to elucidate if the key role for transport reduction is played by the field or by shear flow.

i C. Alejaldre et. al. Fusion Technology 17, 131 (1990). ii F. Castejón et al. Nuclear Fusion 42, 271-280 (2002) iii J. Dies et al., ECA (EPS-CCFPP), vol. 26B, pp 5.027.1-5.027.4, June 2002. P-5.037, Friday July 1, 2005

Internal transport barrier simulation in the LHD

J. García (1), K.Yamazaki (2), J. Dies (1), J.Izquierdo (1)

1. Fusion Energy Engineering Laboratory (FEEL),Departament de Física i Enginyeria Nuclear, ETSEIB, Universitat Politècnica de Catalunya (UPC), Barcelona (Spain). e-mail: [email protected] 2. National Institute for Fusion Science, Toki, Gifu, 509-5292, Japan. e-mail: [email protected]

High electron temperature plasmas with peaked profiles have been obtained in the Large Helical Device (LHD) [1] as well as in others stellarators devices, as the Compact Helical System (CHS) [2] and the TJ-II [3]. These shots share the common characteristic of having a high positive electric field at the plasma core with a large shear. Both, electric field and its shear are supposed to suppress neoclassical transport and anomalous one respectively.

In order to study the electron heat transport channel and to clarify the electron thermal diffusivity dependence with some plasma parameters in LHD shots with internal transport barrier (ITB), some transport models have been added to TOTAL [4] code. These models can be divided into two categories: Bohm and GyroBohm-like models and drift wave models. A sketch of mixed short and long wavelength models has been derived for this study as a good candidate for the ITB explaining.

The effect of anomalous transport reduction by the electric field shear has been introduced by i /1 means of the factor*+1(- vyExB ) . This factor has been previously checked as a good candidate to drive anomalous transport in tokamak plasmas [5], as well as, also derived from theoretical models [6].

Preliminary results show that a combination of short wavelength and long wavelength together with the electric field shear can explain the transition between non-ITB and ITB shots. The central temperature dependence with density is also well simulated. In the case of GyroBohm models, they fit also temperature profiles, although central temperature dependence with density is higher.

[1] K.Ida et al., Phys. Rev. Lett. 91, 085003 (2003) [2] A.Fujisawa, et al., Phys. Rev. Lett. 82, 2669(1999) [3] F.Castejon et al., Nucl. Fusion 42, 271-280 (2002) [4] K.Yamazaki and T.Amano, Nucl. Fusion 32, 633 (1992) [5] S.Jachmich and R.R. Weynants, Plasma Phys. Controlled Fusion 42, A147-A152 (2000) [6] C.F. Figarella et al., Phys. Rev. Lett. 90, 015002 (2003)

This work has been done in the framework of the project ENE2004-05647 P-5.038, Friday July 1, 2005

Search for Threshold Behavior in DIII–D Electron Transport*

T.C. Luce1, J.C. DeBoo1, N. Howard2, C. Liu1, and C.C. Petty1

1General Atomics, P.O. Box 85608, San Diego, California 92186-5608, USA 2University of Illinois, Urbana-Champaign, Illinois, USA

Theoretical work points to a critical gradient or critical scale length description of tokamak transport. Power balance analysis of experiments cannot affirm this type of model, because of intrinsic limitations. Perturbation techniques are more sensitive and may yield a clearer conclusion about the threshold hypothesis. Analysis of modulated ECH experiments in the DIII-D tokamak shows no direct evidence of a threshold for the conditions tested. Two approaches to analysis of DIII-D perturbation data are taken. First, a general form of the heat flux is applied that includes the possibility of diffusive and convective terms in the equilibrium heat flux. To determine what terms are required, reliable uncertainty estimates for the Fourier spectrum amplitudes and phases must be supplied. A fitting procedure finds best estimates for the “effective” transport coefficients that appear in the linearized energy transport equation. The analysis clearly shows that multiple frequency analysis is necessary. A purely diffusive model applied to the fundamental frequency would indicate a high incremental diffusivity, but the multiple frequency analysis indicates that the effective convection and damping terms are necessary to describe the heat pulse evolution. An intrinsic convective term in the equilibrium heat flux is needed to fully describe the observed transport. The second approach is to impose a model on the data. A critical scale length model has been tested by methods like those applied on ASDEX-Upgrade [1] and FT-U [2]. One method varies the local heat flux near the half radius, while keeping the total flux to the boundary constant. The inferred heat pulse diffusivity (assuming a purely diffusive model) is typically a factor of 2 larger than the power balance value and is an increasing function of the inverse scale length. The data point to a threshold value that is quite small, such that even the low auxiliary power L-mode plasmas tested are above the threshold. In addition, the ratio of the incremental diffusivity to the power balance value (one definition of stiffness) is modest (~2-3). A second method varies the gradient locally while maintaining fixed temperature, by modulating two ECH sources out of phase at two slightly displaced locations. No threshold behavior is observed. This sets strong limits on the threshold value in these plasmas. [1] F. Ryter, et al., Nucl. Fusion 43, 1396 (2003). [2] S. Cirant, et al., Plasma Phys. Control. Fusion Research (Proc. 19th IAEA Conf., Lyon, 2002) IAEA-CN- 94/EX/C4-2Rb.

*Work supported by the U.S. Department of Energy under DE-FC02-04ER54698. P-5.039, Friday July 1, 2005

Ion and electron viscosity for arbitrary-mean-free-path plasma* Andrei N. Simakov1 and Peter J. Catto2 1 Los Alamos National Laboratory, Los Alamos, NM 87544, USA 2 MIT Plasma Science and Fusion Center, Cambridge, MA 02139, USA

The fluid equations of Braginskii [1] are known to contain incomplete expressions for species’ viscosities. More specifically, these expressions miss contributions from gradients of species’ heat fluxes. Their absence is a consequence of ordering the plasma flow velocity comparable to the sound speed. Earlier attempts to improve the equations (Mikhailovskii and Tsypin [2]) did not cure the problem entirely, since the employed Sonine polynomial expansion of the distribution function was truncated too soon, the non-linear contributions from the ion-ion collision operator were neglected, and electrons were not considered. We recently revisited the problem, addressed these shortcomings, and derived a complete closure scheme for short-mean-free-path ions and electrons [3]. When a subsequent attempt to recover these improved equations from the widely-used drift kinetic equation of Hazeltine [4] failed, we realized that this kinetic equation is only exact through the first order in the small Larmor radius expansion. However, the second order accuracy is needed to account for the effects of the gyro-viscous stress tensor. We corrected the drift kinetic equation and used the formalism developed to derive expressions for the gyro-viscosity [5] and (ion) perpendicular viscosity for plasmas of arbitrary collisionality in terms of a few velocity integrals of the zeroth- and first-order (in the small Larmor radius expansion) pieces of the gyrophase- averaged distribution function.

1 S. I. Braginskii, Zh. Eksp. Teor. Fiz. 33, 459 (1957) [Sov. Phys. JETP 6, 358 (1958)]. 2 A. B. Mikhailovskii and V. S. Tsypin, Beitr. Plasmaphys. 24, 335 (1984) and references therein. 3 P. J. Catto and A. N. Simakov, Phys. Plasmas 11, 90 (2004). 4 R. D. Hazeltine, Plasma Phys. 15, 77 (1973). 5 A. N. Simakov and P. J. Catto, Phys. Plasmas 12, 012105 (2005).

*Work supported by U.S. D.o.E. P-5.040, Friday July 1, 2005

Magnetic Reconnection and Associated Transport of Plasma Thermal Energy*

V. Roytershteyn1, B. Coppi1, C. Yarim2 1MIT, Cambridge, MA, USA 2I.T.U. Istanbul, Turkey

The excitation of drift-tearing modes in the high-temperature regimes that are achieved in present day experiments is an important theoretical issue that remains to be resolved. The original theory of these modes [1] was carried out for collisional regimes where the effects of finite thermal conductivity are not important. A modern linearized theory [2] that includes the effects of a small but finite thermal conductivity shows that the structure of the mode is drastically changed relative to that of the regime with “adiabatic” electrons. When these effects become relevant, the threshold for the mode excitation, that is related to the gradients of the plasma current density and pressure, becomes difficult to overcome. This is consistent with a previous analysis that found the mode to become stable in the presence of a significant electron temperature gradient in collissionless regimes [3]. On the other hand, we have pointed out earlier that in problems involving singular perturbations, non-linear effects become important for relatively small amplitudes. Thus the limitations on the linear mode amplitude necessary for the validity of the linearized approximation are identified for the present case. Then, we assume that if the threshold is overcome initially, the stabilizing finite thermal conductivity terms can be reduced locally, for instance by the formation of small islands. Three possibilities have been explored: a local th depression of the longitudinal electron thermal diffusion coefficient D e|| , a local flattening of the electron temperature profile, and a combination of both. In all these cases the ∆ ≈ “threshold” from which the growth rate is evaluated can be reduced to the level ( 'crit 0 ) that is found in the highly collisional limit while the frequency of oscillation is of the order of the electron diamagnetic frequency and is higher than the relevant resistive growth rate.

*Sponsored in part by the U.S. department of energy

[1] B. Coppi, Phys. Fluids, 8, 2273 (1965). [2] B. Coppi, V. Roytershteyn and C. Yarim, Paper IAEA-CN-116/TH/P2-29, I.A.E.A. Fusion Conference 2004 (Publ. I.A.E.A., Vienna, 2004) [3] B. Coppi, J.W.-K. Mark and L. Sugiyama, Phys. Rev. Lett., 42, 1058 (1979)

P-5.041, Friday July 1, 2005

Testing Gyrokinetics on C-Mod and NSTX* M. H. Redi1, W. Dorland2, C. L. Fiore3, D. Stutman4, S. M. Kaye1, J. Menard1, G. Rewoldt1 1Princeton Plasma Physics Laboratory, Princeton University, Princeton, NJ 08540 USA 2University of Maryland, College Park, MD 20742 USA 3Plasma Science Fusion Center, MIT, Cambridge, MA 02139 USA 4Johns Hopkins University, Baltimore, MD 21218 USA Quantitative benchmarks of computational physics codes against experiment are essential for the credible application of such codes. Fluctuation measurements can provide necessary critical tests of nonlinear gyrokinetic simulations, but these simulations require extraordinary computational resources. In this paper we show how linear calculations with GS2 [1] have been used to test the gyrokinetic model for ITG microturbulence on C-Mod [2]. The criterion is the linear mixing length model, which is applied at the onset time for internal transport barrier formation in H-mode experiments [3]. Mixing length estimates of transport exceed experimental estimates by less than a factor of two and are in agreement with experiment in the plasma core. Definitive testing of gyrokinetics is much more difficult for NSTX [4] than for C- Mod, since the long wavelength modes are not ITG/TEM only, as in the case described above. Instead calculations reveal kinetic ballooning modes and microtearing modes as well as ITG/TEM modes. Microtearing modes are unstable in every NSTX plasma examined, unless the plasma has reversed shear or is assumed collisionless. L-mode plasmas exhibit improved electron transport [5] and less strongly destabilized microtearing than do H-modes. ETG and microtearing modes occur in different regions of NSTX plasmas and so appear to compete for the same free energy source. We show the effects on all these modes of a scan of both magnetic shear and temperature scale length variations. A more complex mixing length model is needed for NSTX to properly include all long wavelength modes. Testing nonlinear gyrokinetics will be challenging because of the number and variety of unstable modes. *Supported by USDoE contract No. DE-AC02-76CH03073 [1] M. Kotschenreuther, et al., Comput. Phys. Commun. 88, 128 (1995). [2] I. H. Hutchinson, et al., Phys. Plas. 1, 1511 (1994). [3] C. L. Fiore, et al. Plasma Physics and Controlled Fusion 46, B281-291 (2004). [4] M. Ono, et al. Proc. 18th Int. Conf. on Fusion Energy (Sorrento) IAEA, Vienna, 2000. [5] D. Stutman, et al., “Studies of Improved Electron Confinement on NSTX”, submitted to Nucl. Fus. (2004). P-5.042, Friday July 1, 2005

Confinement and Transport Scaling in NSTX S.M. Kaye1, M. G. Bell1, R. E. Bell1, E. D. Fredrickson1, B. P.LeBlanc1, K.C. Lee2, S. A. Sabbagh3, D. Stutman4 and the NSTX Team 1 Princeton Plasma Physics Laboratory, Princeton University, Princeton, N.J. 08543 2 University of California at Davis, Davis, Cal. 3 Dept. of Applied Physics, Columbia Univ., NYC, N.Y. 4 Johns Hopkins University, Baltimore, Md.

Data from neutral beam heating experiments in the low aspect ratio National Spherical Torus Experiment (NSTX) greatly expands the range of data available for establishing both confinement and transport trends in toroidal confinement devices. For instance, the NSTX data extend the range of toroidal beta from the conventional aspect ratio tokamak confinement database by a factor of 3, to values over 35%. The aspect ratio of NSTX is 1.3, a factor of two lower than those of conventional aspect ratio devices. Consequently, NSTX data provide a good test of conventional aspect ratio confinement scalings and transport trends, as well as an examination of the effects of extreme toroidicity, kinetic effects and high beta. Both thermal and global confinement times in NSTX L and H-modes can exceed values given by conventional aspect ratio scalings, and these data indicate a strong toroidal field. At fixed toroidal field, the current scaling is approximately linear and the power degradation is less severe than that at higher aspect ratio. Both density and magnetic fluctuation levels can influence the confinement, as can severity of ELMs. Both dimensional and dimensionless confinement scalings will be presented. Local transport in NSTX is dominated by losses through the electron channel, with the ions being at or above neoclassical transport levels in the plasma core (r/a~0.4). The transport is highly sensitive to the q-profile, with transport barriers developing during periods of reversed magnetic shear.

This work is supported by U.S. DOE contract # DE-AC02-76CH03073 P-5.043, Friday July 1, 2005

Neoclassical poloidal rotation studies in high temperature plasmas

W.M. Solomon1, K.H. Burrell2, R. Andre1, L.R. Baylor3, R.E. Bell1, R. Budny1, E.J. Doyle4, R.J. Fonck5, P. Gohil2, D.J. Gupta5, R.J. Groebner2, G.J. Kramer1, G.R. McKee5, R. Nazikian1, G. Rewoldt1, T.L. Rhodes4, E.J. Synakowski1, W. Wang1, and M.C. Zarnstorff1 1 Princeton Plasma Physics Laboratory, Princeton, New Jersey, USA 2 General Atomics, San Diego, California, USA 3 Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA 4 University of California, Los Angeles, California, USA 5 University of Wisconsin, Madison, Wisconsin, USA

Rotation plays a critical role in high performance fusion plasmas, yet at present, predictive knowledge of rotation and momentum confinement is lacking. Neoclassical theory provides predictions for the poloidal rotation, and hence experimental verification of these predictions is highly desirable. In this work, we focus on investigating measurements of poloidal rotation from charge exchange recombination spectroscopy. The determination of poloidal rotation from charge exchange measurements is greatly complicated by various atomic physics effects. The energy-dependence of the charge exchange cross-section, in combination with the finite lifetime of the excited state and the gyro-motion of the ion, can generate apparent poloidal velocities that are many times greater than the neoclassical predictions. Hence, any attempt to compare experimentally determined poloidal rotation profiles with neoclassical predictions must carefully handle these corrections. Our analysis using a simplified model for the charge exchange cross-section indicated that the neoclassically predicted poloidal rotation, as determined from the code NCLASS, was too low by an order of magnitude. Also, the rotation was predicted to be in the opposite direction than what was inferred experimentally. Further refinements have been made to use more sophisticated and accurate charge exchange cross- sections to determine the corrections. In addition, the effect of halo neutrals is investigated and accounted for. Given the complexity of the measurement, the radial electric field is compared for three different impurity species, C, He, and Ne (calculated through radial force balance). Each species has very different cross-sections and lifetimes and provides a further verification of the measurement technique.

¡ Work supported by US Department of Energy under DE-AC02-76CH03073, DE-FC02-04ER54698, DE- AC05-00OR22725, DE-FG03-01ER54615, and DE-FG03-96ER54373. P-5.044, Friday July 1, 2005

Neoclassical Radial Electric Field and Flows in a Collisional Tokamak

Peter J. Catto1 and Andrei N. Simakov2

1 Plasma Science and FusionCenter, NW16-250 Massachusetts Institute of Technology 167 Albany Street, Cambridge, MA 02139

2 Los Alamos National Laboratory MS K717, T-15 Los Alamos, NM 87545

Abstract

The neoclassical electric field in a tokamak is determined by the conservation of toroidal angular momentum. In the steady state in the absence of momentum sources and sinks, it is explicitly evaluated by the condition that radial flux of toroidal angular momentum vanish. Of course, for an isothermal plasma the viscosity must simplify to an expression that vanishes for a radial Maxwell-Boltzmann ion response. In the presence of ion temperature variation obtaining the electric field in terms of density and temperature gradients is far more complicated. For a collisional or Pfirsch-Schlüter short mean free path ordering in which the plasma flow is sub-sonic [1] we find that there are two limiting cases of interest [2]. The first is the simpler case of an extremely up-down asymmetric tokamak (for example, just inside the separatrix of a single null configuration) for which the lowest order gyroviscosity does not vanish and must be balanced by the leading order collisional viscosity in order to determine the radial electric field. The second case is the more complicated case of an up- down symmetric tokamak for which the gyroviscosity must be evaluated to higher order and balanced by the lowest order collisional viscosity to determine the radial electric field. For equal ion and electron temperatures our result agrees with the more recent large aspect ratio, concentric, circular flux surface expression found by Claassen and Gerhauser [1]. In general both contributions must be retained, and standard drift kinetics can not be used to obtain the results [4]. The magnetic topology of a tokamak impacts the radial electric field and ion flow velocity by reversing the direction of (i) the plasma current, (ii) the toroidal magnetic field, or (iii) both; or (iv) by changing from lower single null to upper single null operation.

[1] P. J. Catto and A. N. Simakov, Phys. Plasmas 11, 90 (2004). [2] P. J. Catto and A. N. Simakov, Phys. Plasmas 12, 012501 (2005). [3] H. A. Claassen and H. Gerhauser, Czech. J. Phys. 49, 69 (1999). [4] A. N. Simakov and P. J. Catto, Phys. Plasmas 12, 012105 (2005). P-5.045, Friday July 1, 2005

Observation of Turbulent Driven Zonal Flow in a Cylindrical Plasma Device G.R. Tynan1, C. Holland1, M.J. Burin2, A. James1, D. Nishijima1, M. Shimada1

1Department of Mechanical and Aerospace Engineering University of Californaia San Diego La Jolla CA 92093

2 Department of Astrophysical Sciences Princeton University Princeton NJ

An azimuthally symmetric sheared flow is observed in a turbulent cylindrical magnetized plasma that has no apparent source of angular momentum. Measurements show that the shear layer is maintained against ion-ion viscosity and ion-neutral flow damping by the turbulent electrostatic Reynolds stress which is driven by collisional drift turbulence, providing clear evidence for the formation of a sheared zonal flow in a simple plasma system. P-5.046, Friday July 1, 2005

Studies of Density Fluctuation Dynamics in L-H Transitions and Pedestal Formation in DIII-D Using Beam Emission Spectroscopy*

C. Holland1, G.R. Tynan1, R.J. Fonck2, G.R. McKee2, D.K. Gupta2, D.J. Schlossberg2, M.W. Shafer 2, and R.J. Groebner3

1Dept. of Mechanical and Aerospace Engineering, University of California, San Diego, California, USA 2University of Wisconsin-Madison, Madison, Wisconsin, USA 3General Atomics, San Diego, California, USA

Edge plasma transport barriers play a critical role in global plasma performance and are expected to determine core plasma conditions in ITER. It is therefore particularly important to understand the mechanism(s) that govern the width and height of the edge pedestal. One such mechanism is believed to be sheared ExB decorrelation of turbulence, which is observed during the L-H transition, and may also continue during subsequent development of edge barriers and lead to a reduction in the cross-field transport rates during pedestal formation. In this presentation, we report on studies of density fluctuation dynamics across the edge and pedestal region in DIII-D, during and beyond the L-H transition, measured with a newly upgraded beam emission spectroscopy diagnostic [1]. It is found that in the pedestal region, the turbulence is rapidly suppressed at the transition, but then begins to grow in amplitude as the pedestal height increases and the Er well deepens, such that it has returned to (or even exceeds) L-mode intensity levels when ELMs begin. The dynamics of power spectra, coherences, cross phases, and correlation lengths are also presented, along with initial measurements of poloidal velocities obtained via time-delay estimation [2]. Comparisons to models of edge barrier formation are also discussed.

[1] D.K. Gupta, et al., Rev. Sci. Instru. 75, 3493 (2004). [2] M. Jakubowski, R.J. Fonck, and G.R. McKee, Phys. Rev. Lett. 89, 265003-1/4 (2002).

*Work supported in part by the U.S. Department of Energy under DE-FG03-01ER54615, DE-FG03- 96ER54373, and DE-FC02-04ER54698. P-5.047, Friday July 1, 2005

Modification of Intermittency by External Perturbations*

J.A. Boedo1, D.L. Rudakov1, T.E. Evans2, M.J. Schaffer2, R.A. Moyer1, J.G. Watkins3, S.L. Allen4, M.E. Fenstermacher4, E.M. Hollmann1, C.J. Lasnier4, A.W. Leonard2, M.A. Mahdavi2, A.G. McLean5, and W.P. West2

1University of California, San Diego, California, USA 2 General Atomics, P.O. Box 85608, San Diego, California 92186-5608, USA 3Sandia National Laboratories, Albuquerque, New Mexico, USA 4Lawrence Livermore National Laboratory, Livermore, California, USA 5University of Toronto Institute for Aerospace Studies, Toronto, M3H 5T6 Canada

Extensive evidence exists indicating the importance of intermittent transport in the region of the separatrix and scrape-off layer (SOL) in tokamaks. Methods that allow a degree of control of the intermittent transport may be useful in the control of divertor and particle fluxes and may also be applicable to the control of edge localized modes (ELMs), that can be interpreted as large blobs. The ability to control the extreme heat and particle fluxes expected in ITER could be very advantageous. We present results from a study on effects of applied perturbations on radial turbulent transport, both broadband and intermittent. Two such perturbations have been applied in DIII-D stochastic magnetic fields via an internal resistive wall mode coil and electric fields via SOL biasing.

The application of an stochastic field [1] results on broadening of the edge/SOL Te and Ne profiles and concomitant changes in the intermittent transport. The intermittent objects become hotter and denser at a given radius when the stochastic fields are applied. Measured turbulent radial particle and heat fluxes increase and are consistent with the broadening of the profile. The application of electric fields, using the DIII-D outer-divertor ring electrode [2], results in quite marked changes in the edge/SOL profiles and also in the private region, which becomes almost depleted of plasma. The SOL and the separatrix were biased by positioning the plasma on the ring electrode. The bias voltage was changed from –100 V to +100 V in 50 V steps. It was found that both the particle and heat flux to the divertor can be modified. Some of these changes can be understood in terms of EB× drifts, but changes to the anomalous radial transport including intermittency must be invoked as well.

[1] T.E. Evans, R.A. Moyer, P.R. Thomas, et al., Phys. Rev. Lett. 92, 1 (2004). [2] M.J. Schaffer, R. Maingi, A.W. Hyatt, et al., Nucl. Fusion 36, 495 (1996).

*Work supported by the U.S. Department of Energy under DE-FG02-04ER54758, DE-FC02-04ER54698, DE- AC04-94AL85000, and W-7405-ENG-48. P-5.048, Friday July 1, 2005

Spatially resolved ion heating measurements during a sawtooth crash on the Madison Symmetric Torus (MST)

S. Gangadhara, D. Craig, D.A. Ennis, D.J. Den Hartog, G. Fiksel, D.J. Holly, S.C. Prager, V.V. Mirnov, V.A. Svidzinki University of Wisconsin-Madison Center for Magnetic Self-Organization in Laboratory and Astrophysical Plasmas

Self-heating of ions during magnetic reconnection is a process important to a number of laboratory and astrophysical plasmas. To study this phenomenon, measurements have

recently been made – for the first time – of the impurity ion temperature (Ti) profile evolution during a fast (~ 100 µs) reconnection event (sawtooth crash) on the MST reversed field pinch. Results have been obtained at a number of different radial locations, and have been compared to bulk ion temperature measurements at a single point. Measurements suggest that during the crash the bulk ion distribution is non-Maxwellian; conversely, a Maxwellian distribution appears to fit the impurity ion data quite well. These results also indicate that the impurities are more strongly heated than the bulk during a sawtooth crash. The profile of the impurity ion temperature rise suggests that the heat source is broad and active throughout the plasma volume. The time scale for ion heating (~ 100 µs) is similar at all measured locations,

while the time scale over which Ti returns to its pre-crash value is significantly longer at most radii. Impurity measurements were obtained using a new custom-built spectrometer, recently

installed on MST to provide localized measurements of impurity Ti with fast time resolution using charge exchange recombination spectroscopy (CHERS). A sophisticated model of the emission data has also been developed to fit the CHERS results with high accuracy. A number of theories are being examined to explain the full set of results. These include viscous damping of magnetic fluctuations, and electric field acceleration of ions. Preliminary modeling results will also be presented.

Work supported by U.S.D.O.E. and N.S.F. P-5.049, Friday July 1, 2005

Analysis of RWM with a 3D Model of Conducting Structures F. Villone1, G. Rubinacci2, Y.Q. Liu3, Y. Gribov4 1 Ass. EURATOM/ENEA/CREATE, DAEIMI, Univ. di Cassino, Cassino (FR), ITALY 2 Ass. EURATOM/ENEA/CREATE, DIEL, Univ. Federico II di Napoli, ITALY 3 Ass. EURATOM/VR, Chalmers Univ. Technology, Gothenburg, SWEDEN 4 Physics Unit, ITER Naka Joint Work Site, Naka, Ibaraki, JAPAN

Ideal external kink instabilities of low n”0 toroidal mode numbers limit the achievable normalized dN. The presence of a stabilizing ideal conducting wall can mitigate this limit, thanks to eddy currents induced by plasma perturbations. The non-vanishing resistivity of any real wall causes a decay of eddy currents and hence a loss of stability. The resulting resistive wall modes (RWM) grow at the time scale of the wall time, thus allowing a possible active stabilization of such modes with magnetic coils. A critical issue is a correct three-dimensional model of the conducting structures (walls) surrounding the plasma, and of the feedback coils. Aim of this paper is to describe the coupling between the toroidal stability code MARS-F [1] and the three-dimensional magneto-quasi-static code CARIDDI [2]. MARS-F has been used for numerical simulation of RWM in ITER [3]. It represents active feedback coils and passive conducting structures with simplified 2D models (e.g. neglecting the presence of ports in the vessels and the physical shape of coils). CARIDDI is a 3D eddy currents code based on an integral formulation, hence requiring a discretization of the conducting structures only, and allowing an easy coupling with external circuitry. The introduction of a two-component electric vector potential and the use of edge elements give rise to a very accurate and effective code, which has been extensively used for fusion applications, also coupled with a 2D linearized plasma response model [4]. We envisage a “weak” coupling procedure: starting from an initial 2D guess of plasma perturbation provided by MARS-F, CARIDDI computes the currents in the 3D conducting structures. These are in turn used by MARS-F to compute an upgraded estimate of plasma perturbation, which is fed back to CARIDDI until convergence is reached. Initial results on the RWM study using this scheme will be reported.

[1] Y.Q. Liu, et al., Phys. Plasmas 7, 3681 (2000) [2] R. Albanese, G. Rubinacci, Adv. Im. El. Phys.,102, 1-86, Academic Press (1998) [3] Y.Q. Liu, et al., Nucl. Fusion 44, 232 (2004) [4] R. Albanese, R. Fresa, G. Rubinacci, F. Villone, IEEE Trans. Mag., 36, 1804 (2000) P-5.050, Friday July 1, 2005

!∀ ! !∀ ! #!∀ ∃% !∀ !∀&∋(∀ )∗+,−∗+ ∋( ! ∃ ! ∀ !∀ ∃ , , ∋( ∀ ∀ ∋( ∀(∀,./0∃( ∀ ∃1!!∀ !∀ ∀ ( ∀, ∀,2∃ ! ∃ 2,∀ # !∀ ∃∃ ∀1∀∀ ∀∃∀∀∀ ∀∋(∀ 3!∋(% ∋(∀∃ !∀∀ !!! ( ∃ ∋( ∀ # !∀ 2,3∃∀ ∀ ./0∀∃4)5,/)6∗∗∗7)7) P-5.051, Friday July 1, 2005

32nd EPS Conference on Controlled Fusion and Plasma Physics

First results on the analysis of magnetic fluctuations in toroidal geometry and comparison with numerical simulations for the RFX-mod reversed field pinch

D. Terranova, P. Zanca, R. Paccagnella, T. Bolzonella and S. Martini

Consorzio RFX, Associazione Euratom-ENEA sulla fusione Corso Stati Uniti 4, 35127 Padova, Italy.

The modified RFX experiment (RFX-mod, a=0.45 m, R=2 m, I = 1 MA) is equipped with magnetic probes which allow a detailed analysis of magnetic fluctuations. The new layout consists of four toroidal arrays of 48 pick-up coils for measuring both the toroidal and radial component of the magnetic field outside the vacuum vessel, along with two toroidal arrays inside the vacuum vessel for measuring the toroidal component. In comparison to the previous layout, this configuration allows a better spectral resolution of the poloidal harmonics. This is important, since the unstable modes produce poloidal sidebands due to the toroidal geometry. RFX-mod has a new magnetic boundary (a shell thinner and closer to the plasma), together with a new first wall and a better control of the equilibrium. Therefore we present a comparison between RFX and RFX-mod in terms of amplitudes, phase relations and radial profiles of the magnetic perturbations. The analysis is performed using the method described in [1], which adopts flux co-ordinates in toroidal geometry and provides a radial profile of the perturbation eigenfunctions. In order to quantify the importance of non linear effects the experimental data are also compared with 3D simulations in presence of a resistive wall after deconvolving the toroidal contribution, evaluated as outlined above. We also address the phenomenon of the phase-locking of the modes, which is crucial for RFPs [2].

[1] P. Zanca, D. Terranova Plasma Phys. Control. Fusion 46 (2004) 1115 [2] R. Fitzpatrick, P. Zanca Physics of Plasmas 9 (2002) 2707 P-5.052, Friday July 1, 2005

!∀∀ !∀#∃#∀% #∃%&∀∋&∃&∃ %&(∃)&∗&∀ ∀∗∃&)&+∀,− .,/ 0 ∃ ∀ ∗ &1∃ ∀ ∃& &∃∀ ∀& &∗&∃&∃&∀2& ∗ ∃% ( ∀& ∃ ∃& ∀ & & ∗& &∃ ∃% % ∀ &% ( &∃∀&∃&&∀% ∀∀∃ ∀

∃ & ∗ , && 3 ∃ ∀ &&%∋∀1∃&%&&%%∃& &&∀∗&∀∗∀1∃&% (&∃&&∀∃∀%&&∀∀ 1∃&)∗&&∃&∃ ∀ &4∃ & & .5/ ( ∃&∃ ∃ ∗ 6) %∃44∗&&∃&( ∃& % && ∀ 1∃&∃∃ ∀ ∗& 3%∗∃∃∃∀&%∀∗ &∀∃&∃(7∋&&∀∀&%8 ∋&.8/(&∃∀∃&∀& ∗ &% ∃&( 3 % &∃&%&∃&)∗∃ ∃&%&&)&∀∀&1∃∃ ∗∗∃∀&∗%&% .,/%),59::,+5::;− .5/&%),<<5,<<9+5:::− .8/%)589::,+5::;− P-5.053, Friday July 1, 2005

!!∀#∃ %& ∋(&)∃∗)!∗+ !∀ # ∃ ∃%!∃& ∋()∗&+ %!∋,) −%! + + + #.∃& ∃ /+∋0) & 12.3 ∃ ∃ +&+ & + % ∃−%!∃ + ∋().∃ ,445(5644( ∋,) ,445,0644( ∋0)7∃∀ ,445(((6 P-5.054, Friday July 1, 2005

First results on Reversed Field Pinch plasmas with new magnetic boundary in the RFX-mod experiment S.Martini and the RFX team Consorzio RFX, Associazione Euratom-ENEA sulla Fusione, Corso Stati Uniti 4, I-35127, Padova, Italy

Reversed Field Pinch experiments resumed in December 2004 in the modified RFX device (RFX-mod). The magnetic boundary is now made of a thin shell (50 ms Bv penetration time) plus a feedback controlled vertical equilibrium circuit. New power supplies provide a more flexible control of the toroidal field coils. A new full graphite tile coverage of the vacuum vessel has also been installed, which should prevent highly localized power deposition on the tile edges. A set of 48 x 4 saddle coils outside the thin shell allows to feedback control both dynamo modes and RWMs. The greatly enhanced set of internal and external magnetic pick up coils is available in RFX-mod, which, along with the detailed measurements of temperature density and impurity emission profiles, permits to do a complete characterization of the pulses, both in terms of MHD mode dynamics and of plasma confinement. The paper presents the first set of experiments at relatively low currents of 300-400 kA (see figure), which have been done to assess the performance of the new machine front end. After optimizing RFP setting up, density control and equilibrium control, the plasma performance has been benchmarked versus selected reference pulses of the previous database. Up to now we have been able to achieve well sustained pulses with duration in excess of 1.5 the shell time constant with field reversal and pinch parameter (F, Θ), resistive loop voltage, density and temperature comparable to those of the reference pulses. Mode dynamic is also relatively benign, with amplitudes and spectra similar to the Ip [A] past ones, and no evidence of RWM growth is seen on the present

pulse time scale. Bφ & [T] #15696 Experiments at higher current are ongoing and will also be presented at the conference. Time [s] P-5.055, Friday July 1, 2005

Inter-Machine Scaling of Alfvén-like Modes Excited by Magnetic Islands in FTU and TEXTOR P. Buratti1, O. Zimmermann2, M. De Benedetti1, Y. Liang2, H.R. Koslowski2, G. Regnoli1, P. Smeulders1, F. Zonca1 1EURATOM-ENEA Association, C.R. Frascati, CP 65, 00044 Frascati, Italy. 2Institut für Plasmaphysik, Forschungszentrum Jülich GmbH, EURATOM Association, Trilateral Euregio Cluster, 52425 Jülich, Germany.

The excitation of high-frequency (HF) waves in the presence of large magnetic islands has been recently discovered in tokamak plasmas [1, 2]. Wave frequencies are well above the island rotation frequency, one order of magnitude below the first toroidal gap in the Alfvén continuum and of the same order of the low frequency gap introduced by finite beta effects. HF waves appear in pairs of branches that are separated by twice the island rotation frequency, and have typical average frequencies of 50 kHz in FTU and 20 kHz in TEXTOR. If the island frequency decreases to zero due to locking, the two branches of each pair coalesce into a single line that remains at high frequency; this excludes the possibility of explaining HF waves as island deformations due for example to toroidal ripple. There is however a clear causal relationship between the presence of large magnetic islands and the excitation of HF waves, in fact the latter only appear above a threshold island amplitude; in addition, HF waves can be found in the absence of fast ions that could drive Alfvén modes. These observations indicate that HF waves should tap their energy from the magnetic island. The excitation of HF modes by magnetic islands could have important implications on island dynamics that are outside the scope of this work. In this paper we will focus on identifying the excitation mechanism. There are at least two different approaches to the interpretation of HF modes excitation by magnetic islands. The first one is based on nonlinear interaction by three-wave coupling between the island and beta-induced Alfvén eigenmodes (BAE). In the second interpretation the magnetic island is considered as a new (helical) equilibrium in which shear-Alfvén waves become marginally stable. The two approaches produce different scalings for the wave frequency, which will be compared with FTU and TEXTOR data. This inter-machine comparison is particularly useful in that it allows comparing very different magnetic field values, from 2 T to 7 T. [1] P. Buratti et al. Fusion Science and Technology 45 350 (2004). [2] O. Zimmermann et al. this Conference. P-5.056, Friday July 1, 2005

Resistive wall mode stability in high beta plasmas in DIII-D and JET

H. Reimerdes1, M. Bigi2, A.M. Garofalo1, M.P. Gryaznevich2, T.C. Hender2, D.F. Howell2, G.L. Jackson3, R.J. La Haye3, G.A. Navratil1, M. Okabayashi4, S.D. Pinches5, J.T. Scoville3 and E.J. Strait3 1 Columbia University, New York, USA 2EURATOM/UKAEA Fusion Association, Culham, UK 3General Atomics, San Diego, USA 4Princeton Plasma Physics Laboratory, Princeton, USA 5Max-Planck-Institut für Plasmaphysik, Garching, Germany

A quantitative comparison of resistive wall mode (RWM) stability in rotating high-β plasmas has been carried out in both the DIII-D and the JET tokamaks. The JET conﬁguration with similar safety factor and pressure proﬁles has been developed in DIII-D, providing plasmas with the same ideal MHD stability properties but different wall properties, notably a different plasma- wall distance and a different characteristic decay time for wall eddy currents [H. Reimerdes, et al., Bull. Am. Phys. Soc. 49, 142 (2004)]. The stability is studied passively by measuring

the critical rotation required for RWM stability, Ωcrit, and actively by probing the plasma with

externally applied resonant, n = 1 magnetic fields. In both devices the value of Ωcrit at the q = 2 surface is found to be of the order of a percent of the local inverse Alfvén time. Likewise, the resonant field amplification (RFA) of the externally applied field in both devices increases significantly once β is close to or above the no-wall kink limit. Evaluating the RFA at the plasma boundary in plasmas with the same normalized 0.4 plasma rotation and at the same normalized distance DIII-D

between the no-wall limit, βno−wall, and ideal-wall limit, JET °

° 0.3 β − , results in quantitative agreement (see Figure). = 0 ideal wall ϕ = 90 ϕ The apparent independence of the RWM stability from the wall properties shows that the stabilization of the 0.2 no-wall limit RWM is provided by the fast bulk plasma rotation relative ideal-wall limit

Plasma response at 0.1 to the quasi-static magnetic perturbation of the mode, Externally applied field whereas the position of the wall sets the ultimate, ideal- 0.0 wall β-limit and the conductivity of the wall only prevents 0.0 0.2 0.4 0.6 0.8 1.0 (β - βno-wall) / (βideal-wall - βno-wall) the mode from rotating with the bulk plasma. This work was supported in part by the US Department of Energy under DE-FG02- 89ER53297, DE-FC02-04ER54698, and DE-AC02-76CH03073. P-5.057, Friday July 1, 2005

Measurement of the Instability Threshold for Toroidal Alfvén Eigenmodes in JET Tokamak Plasmas ∗ D.Testa1, D.Borba2, C.Boswell3, A.Fasoli1,3, S.Sharapov4, JET-EFDA contributors [1] CRPP, Association EURATOM – Confédération Suisse, EPFL, Lausanne, Switzerland [2] Associacão EURATOM/IST, Av. Rovisco Pais, 1049-001 Lisboa, Portugal [3] Plasma Science and Fusion Center, Massachusetts Institute of Technology, Boston, USA [4] EURATOM-UKAEA Fusion Association, Culham Science Centre, OX14 3DB, Abingdon Controlling the interaction between α’s and modes in the Alfvén frequency range is a crucial issue for ITER operation, as these modes can be driven unstable by the slowing-down α’s up to amplitudes at which they could cause rapid radial transport of the α’s themselves. To assess the regimes under which such losses may occur, this experimental study focuses on the dependence of the stability threshold of Alfvén Eigenmodes (AEs) on different Ion Cyclotron Resonance Frequency (ICRF) heating schemes and background plasma parameters. First, the role of the direction of the ion ∇B-drift is considered. As previously reported [1,2], the damping rate of n=1 Toroidal AEs (TAEs) is a factor of three larger when the ion ∇B-drift is directed away from the divertor. A possible explanation for this observation is related to the effect of slightly different density and temperature profiles (due to the different flows) on the continuum and Landau damping at the plasma edge. This idea can be tested by determining the instability threshold for TAEs as function of their toroidal mode number, since high-n core localised TAEs will be less affected by the edge damping mechanisms than lower n’s TAEs, which are typically more global modes. Second, we consider the TAE instability threshold for 3He and H minority ICRF heating in both deuterium and 4He plasmas for similar ICRF power levels and background plasma parameters. For the most usual H(D) minority heating scheme, we typically observe many TAEs being excited with n’s in the range n=4÷10. On the other hand, with the 3He(4He) heating scenario lower mode numbers for the unstable TAEs were observed, usually n=1÷3. The differences in the spectrum of excited n’s can be attributed to a combination of a different fast ion drive (as determined from the measurement of the fast ion distribution function) and Landau damping from the bulk plasma (as determined with the measurement of the n=1 TAE damping rate in ohmic D and 4He plasmas). These experimental observations are also supported with calculations performed with the CASTOR-K [3] code. Third, we consider the decay time for fast-ion driven TAEs at the ICRF switch-off as function of their toroidal mode numbers. This decay-time depends on the n’s and it is observed to be a factor 2÷3 shorter than the fast ion slowing-down time. The relative roles of MHD modes in reducing the fast ion energy content on time scales shorter than the slowing-down time compared to the ICRF-induced diffusion can then be discussed based on these observations, as previously predicted [4].

This work has been conducted under the European Fusion Development Agreement. D.Testa and A.Fasoli were partly supported by the Fond National Suisse pour la Recherche Scientifique, Grant 620-062924. C.Boswell was partly supported by the DoE contract No. DE-FG02-99ER54563.

[1] D.Testa et al., Plasma Phys. Control. Fusion 46 (2004), 1. [2] D.Testa et al., Paper EX/P4-45, 20th IAEA Fusion Energy Conference, 2004. [3] D.Borba, W.Kerner, Journ. Comp. Phys. 153 (1999), 101; D.Borba et al., Nucl. Fus. 43 (2002), 1029. [4] T.Hellsten et al., 8th IAEA-TCM on Energetic Particles, San Diego (USA), 2003.

∗ Annex 1 of J.Pamela et al., “Overview of JET Results”, OV-1/1.2, Proceedings 20th IAEA Fusion Energy Conference, Vilamoura, Portugal, 1-6 November 2004; to appear in Nuclear Fusion, 2005. P-5.058, Friday July 1, 2005

Novelty Detection for on-line disruption prediction systems Barbara Cannas1, Alessandra Fanni1, Piergiorgio Sonato2, Maria Katiuscia Zedda1 and contributors to the EFDA-JET workprogramme1 1Electrical and Electronic Engineering Dept. - University of Cagliari, Cagliari - Italy. 2 Consorzio RFX, Associazione EURATOM-ENEA sulla Fusione, Padova - Italy.

Disruptions are one of the major issues in current nuclear fusion tokamak research. Recently, the disruption predicting capability of Neural Networks was stressed in [1-3]. It was also shown that disruption related indicators or diagnostic signals could be used to provide a composite impending neural network disruption warning indicator. The drawback of this approach is that the trained network could deteriorate its performance once it is on- line. In fact, a network that is trained to discriminate between inputs coming from a set of distributions will be completely confused when input data comes from an entirely new distribution. This could be the case of JET machine, where new plasma configurations can lead to unknown discharges. A possible improvement can be made using Novelty Detection techniques [4]. Recently, Novelty Detection methods have been proposed in the literature to determine the degree of novelty of a given input dataset, based both on statistical and neural network approaches. In this paper, different Novelty Detector (ND) architectures are tested, providing information about the reliability of the outputs of the Neural Networks disruption predictor nowadays installed in the JET. In the test phase, the system triggered 10 over 86 false alarms and 23 over 86 missed alarms [2]. In the on-line application, the ND should be used to assess the reliability of the network output, i.e., samples having a low confidence have to be discarded or at least flagged as suspicious and used to update the disruption predictor. Two approaches to Novelty Detection will be compared, that use Self Organising Maps and Support Vector Machines. Results obtained using a simple ND based on a Self Organising Map, showed that 8 over the 10 false alarms and 13 over the 23 missed alarms correspond to low confidence predictor outputs. On the other hand, the ND discards 23 correctly predicted disruptions. These preliminary results show that the ND application results in an increase of missed alarms; thus a further tuning of ND parameters is needed.

[1] G. Pautasso, C.Tichmann, et al. “ On-line prediction and mitigation of disruption in ASDEX Upgrade”, Nucl. Fusion 37,pp. 100 2002. [2] B. Cannas, A. Fanni, G. Sias, P. Sonato, M.K. Zedda, and JET EFDA contributors “Neural approaches to disruption prediction at JET”, 31th EPS Conf. on Plasma Phys., Vol.28G, p-1.167, London, UK, 2004. [3] R. Yoshino “Neural-net disruption predictor in JT-60U” Nucl. Fusion 43, pag 1771-1786, (Dec 2003). [4] C. M. Bishop, “Novelty detection and neural network validation”, Proceedings of the IEE Conf. on Vision and Image Processing, pp. 217-222, 1994.

1 See the Appendix of J.Pamela et al., Proc. 20th IAEA Fusion Energy Conference 2004 (Vilamoura, 2004) P-5.059, Friday July 1, 2005

Experimental Observations of n =0 Mode Driven by Energetic Particles on JET C. Boswell1, H.L. Berk2, T. Johnson3, M.F.F. Nave4, S.D. Pinches5, S.E. Sharapov6, L. Zakharov7 and contributors to the EFDA-JET workprogramme∗ 1 Plasma Science and Fusion Center, MIT, Cambridge, MA 02139, USA 2 Institute for Fusion Studies, University of Texas, Austin, TX, USA 3 Royal Institute of Technology, Stockholm, Sweden 4 Associac¸ao˜ EURATOM/IST, Centre de Fusao˜ Nuclear, 1049-001 Lisbon, Portugal 5 Max-Planck Institute for Plasma Physics, EURATOM Association, Boltzmannstrasse 2, D-85748 Garching, Germany 6 Euratom/UKAEA, Culham Science Centre, Culham, UK 7 Princeton Plasma Physics Laboratory, Princeton, NJ, USA ∗ See the Appendix of J. Pamela et al., Fusion Energy 2004 (Proc. 20th Int. Conf. Vilamoura, 2004) IAEA, Vienna (2004) A mode with toroidal mode number of zero

log (‘ B (T)) (n = 0) and excited by energetic particles has JPN 54895, KC1F Probe: H304 10 been observed on JET. This mode is observed (5 38 to be ∼ a/R0 below the frequency of Toroidal 36 (5.5 Alfven´ Eigenmodes and was ﬁrst described in 34 reference [1] where it was noted because of the (6 chirping behavior it displayed (see ﬁgure 1). This 32

6.5 30 ( mode cannot be excited by a spatial gradient in Frequency (kHz)

the energetic particle population, since to tap this 28 (7

energy would require a non-zero toroidal mode 26 number. A possible resonant phase space drive, (7.5 2.83 2.84 2.85 2.86 2.87 2.88 2.89 2.9 2.91 at frequencies much less than the ion cyclotron Time (s) frequency, is from energy inversion at a constant Figure 1: n =0 chirping mode excited by magnetic moment of the energetic particle distri- energetic particles during a reversed shear bution function. This mode has been observed in discharge. several discharges on JET all with high-field-side ICRF heating. The observed chirping can be described as a weak resonance instability where the frequency at each instant is equal to an orbit resonance frequency. The orbit resonance that is in the appropriate range of frequencies is the bounce frequency of the trapped particles. The changing mode frequency can be interpreted as a synchronism between the mode frequency and the resonance frequency as explained in the model developed in reference [2]. The basic plasma mode that is excited by the energetic particles has not been unambiguously identified, but the mode frequency has recently been found to scale with the electron temperature gradient near the mid-radius of the plasma. No detectable degradation of the plasma confinement was observed in the presence of this mode. [1] Heeter R F 1999 Alfven´ eigenmode and ion Bernstein wave studies for controlling fusion alpha particles. Unpublished Ph.D. dissertation, Princeton University [2] Berk H L, Breizman B N, Petviashvili N V 1997 Phys. Lett. A 234 213 P-5.060, Friday July 1, 2005

Interplay of Error Field and Neoclassical Tearing Mode drives and rotation for the 2/1 mode on JET and DIII-D R. J. Buttery1, R. J. La Haye2, J. T. Scoville2, EFDA-JET contributors* and the DIII-D team2. 1EURATOM/UKAEA Fusion Association, Culham Science Centre, OX14 3DB, UK. 2General Atomics, San Diego, USA. *see appendix: J. Pamela, Proc. 20th IAEA Vilamoura (2004). 2/1 tearing modes pose a concern to performance of baseline and hybrid scenarios in next step fusion devices such as ITER, potentially causing dramatic reductions in confinement or plasma terminations. Previous studies have focussed on two aspects that lead to such modes. With low momentum injection (such as during scenario formation prior to main heating) natural axisymmetric ‘error’ fields in a device can resonate with the plasma, stopping its rotation and driving island growth. Conversely, at high performance, helical holes in the pressure gradient driven ‘bootstrap’ current, can drive small ‘seed’ islands to large size ‘neoclassical tearing modes’ (NTMs). Because of their different origins, it might be assumed that these aspects can be treated separately. However, new evidence suggests that in some circumstances the two drives can combine to lower thresholds for mode triggering. This interplay is interesting as a mechanism to explore the basic underlying physics, and important 3 to understand in terms of potential limits to fusion power. error field ramps

(G) at constant Although earlier work had observed a lowering of error heating power field thresholds close to ideal stability limits1, the first clear 2 evidence of an interplay with NTMs was identified when it was found that 2/1 NTMs could be triggered at lower than 1 (corrected for linear expected dP on JET when error field ramps were applied. A density dependence)

Error threshold Error 0 detailed scan (Fig 1) found modes at dP up to one third lower 234N with 2-3 Gauss (at q=2 surface) of applied error field. With Fig 1: 2/1 error field threshold error fields present the modes always formed locked, as a function of dP on JET. indicating an enhanced error field sensitivity close to the NTM d limit, possibly related to error field amplification effects2. Similarity studies on DIII-D yielded a similar lowering of thresholds (Fig 2), but in these cases (with purer 2/1 fields and d ramps), medium levels of error field led to a rotating mode onset. This suggests a different mechanism with the error field changing the underlying NTM stability, rather than directly braking rotation to induce tearing. Preliminary data confirms this, with estimates of island rotation in the E·B frame indicating it falling through zero in the medium error field cases, potentially reversing ion polarisation current effects to drive island growth. Further DIII-D studies varied bulk plasma rotation by changing neutral beam injection angle, but even with a near halving of rotation, 4 ramps at NTM dP thresholds were only slightly lower. This constant suggests a less significant role for mode coupling in 2/1 error field NTM triggering, which may therefore be more related to N (corrected for density, toroidal field underlying changes in stability, such as via ion and q95 dependence) polarisation or F' effects. Detailed observations and Rotating onset interpretations of this will be explained in the paper. Locked onset 0 This work was partly funded by EURATOM, the UK 012/1 Error Field (Gauss) 2EPSRC and US DOE under DE-FC02-04ER54698. 1 Fig 2: 2/1 NTM dP thresholds fall with R. J. La Haye et al., Nuc. Fus. 32 (1992) 2119. 2 larger error field in DIII-D dP ramps. A.H. Boozer, et al., Phys. Rev. Lett. 86 (2001) 5059. P-5.061, Friday July 1, 2005

Stable operation regimes in the multimirror trap GOL-3

A.V. Burdakov, A.V. Arzhannikov, V.T. Astrelin, A.M. Averkov, A.D. Beklemishev, I.A. Ivanov, V.S. Koidan, K.I. Mekler, S.V. Polosatkin, V.V. Postupaev, A.F. Rovenskikh, S.L. Sinitsky, Yu.S. Sulyaev E.R. Zubairov Budker Institute of Nuclear Physics, 630090 Novosibirsk, Russia

New experimental results on high performance regimes with stable plasma confinement in the GOL-3 multimirror trap are presented. In the last few years, the device was step by step transformed into the multimirror trap with weak (4.8T/3.2T) corrugation of the magnetic field in 12-meters long solenoid [1, 2]. Collective plasma heating by ~120 kJ 21 -3 (~8 os) relativistic electron beam results in Te~2 keV at ~10 m density. High Te exists for

~10 os. To this time Ti reaches ~2 keV. Ion temperature keeps at the high level during ~1 ms. The final aim of experiments carried out at GOL-3 facility is a proof of a perspective of the multimirror trap for the fusion. Conditions for plasma macroscopic stability in the axisymetric configuration of corrugated solenoid was found. A special experimental campaign was devoted to studies of mechanisms for the stability. In the GOL-3 conditions the magnetic shear was shown to be the important factor for good plasma confinement. Sheared structure of the magnetic field is formed by axial guiding magnetic field of the solenoid and by azimuthal magnetic field, which is generated by axial currents in the plasma. Radial profile of local current density is created by three main sources of current, two of which are external (current of relativistic electron beam and current of the preliminary linear discharge) and the third internal source is the return current to the beam generator which also runs through the plasma. Direct measurements of the safety factor q within the core plasma show the existence of strongly sheared magnetic field, which has q(0)<1 and different sign of q(a). Experiments shown that in macroscopically stable regimes the energy confinement 18 -3 time sufficiently increases and a value of nvE =(1.5„3)©10 m s is obtained. Prospects of plasma stabilization by sheared magnetic field on reactor related multimirror trap are discussed. [1] V.S. Koidan, et al., Progress in multimirror trap GOL-3. // Transactions of fusion science and technology, Vol.47, No.1T, 2005, p.35-42. [2] A.V. Burdakov, et al., Fast Heating of Ions in GOL-3 Multiple Mirror Trap, 31th EPS Plasma Physics Conf. (London, 2004). P-5.062, Friday July 1, 2005

ΦεατυρεσοφΜΗαχτιϖιτψινβεα−ηεατεδπλασαινυλτιιρρορτραπΓΟΛ−3 ς.ς.Ποστυπαεϖ,Α.ς.Αρζηαννικοϖ,ς.Τ.Αστρελιν,Α..Βεκλεισηεϖ,Α.ς.Βυρδακοϖ, Ι.Α.Ιϖανοϖ,Μ.ς.Ιϖαντσιϖσκψ,ς.Σ.Κοιδαν,Κ.Ι.Μεκλερ,Α.Φ.Ροϖενσκικη,Σ.ς.Πολοσατκιν, Ι.ς.Σχηωαβ,Σ.Λ.Σινιτσκψ,Ψυ.Σ.Συλψαεϖ,ς.Π.Ζηυκοϖ,Ε.Ρ.Ζυβαιροϖ ΒυδκερΙνστιτυτεοφΝυχλεαρΠηψσιχσ,Νοϖοσιβιρσκ630090,Ρυσσια

Μυλτιιρροραππροαχητοπλασαχονφινεεν τισστυδιεδαττηεΓΟΛ−3φαχιλιτψιν Νοϖοσιβιρσκ.Τηεπλασαοφ10 201022−3δενσιτψισχονφινεδινα12−ετερ−λονγσολενοιδ,

ωηιχηχονσιστσοφ55χορρυγατιονχελλσωιτη Βαξ/Βιν=4.8/3.2Τ.Τηεπλασαισηεατεδυπτο 1−2κεςιοντεπερατυρε(ατ∼1021−3δενσιτψ)βψαηιγηποωερρελατιϖιστιχελεχτρονβεα(∼1 Μες,∼30κΑ,∼8∝σ,∼120κϑ).Ενεργψχονφινεεν ττιεισ∼1στηατ ισιναγρεεεντωιτη τηεορψπρεδιχτιονσ. Τηεπλασαηεατινγανδχονφινεεντιντηε ΓΟΛ−3φαχιλιτψαρεδε τερινεδβψσεϖεραλ λινκεδχολλεχτιϖεπροχεσσεσ,ωη ιχηλεαδτοαηιγη−λεϖελτυρβ υλενχεανδλαργεοδυλατιονοφ πλασαπαραετερσ.Ιτισυνυσυαλτηατιν τηεαξισψετριχαγνετιχφιελδοφτηεΓΟΛ−3 φαχιλιτψτηεεντιονεδτυρβυλ ενχειπροϖεσγλοβαλενεργψχονφινεενττιεδυετοδεχρεασε οφαξιαλελεχτρονηεατλοσσεσ. Φαστπλασαηεατινγανδσυππρε σσεδαξιαληεαττρανσπορτδυρινγ τηεηεατινγπηασερεσυλτινσοεοβσερϖεδφεατ υρεσοφτηεπλασαβεηαϖιουρ,ωηιχηινχλυδεσ οτιονοφτηεπλασαανδΜΗαχτιϖιτψ. Αξιαλπροφιλεοφεισσιονοφ−νευτρονσ ισπεακεδιντηερεγιονοφπεακπλασα πρεσσυρε.Ατσοετιεαφτερτηεενδοφτη εβεαινϕεχτιοντηελοχαλνευτρονεισσιον βεχοεσηιγηλψοδυλατεδ.Νευτρονδετεχτορσ πλαχεδινονεχορρυγα τιονχελλσηοωπεριοδ− ανδπηασε−χορρελατεδσιγναλσ,περιοδσιναδϕαχεν τχελλσδιφφερ.Μοδυλατιονοφνευτρονσιγναλσ εξιστσφορσεϖεραλτενσοφοσχιλλατιονπεριοδσ .Συχηοδυλατιονχανβεαττριβυτεδτοα ωηιστλε−λικεεχηανισ,ωηιχηχανεξιστιν απλασαωιτηλοχαλλψτραππεδιονσδυετο πλασαφλοωαλονγτηεαγνετιχφιελδανδλοχαλ (ωιτηιντηεφιελδχορρ υγατιονχελλσ)πρεσσυρε γραδιεντσ. Ανοτηερδισχυσσεδφεατυρειστηεδψναιχσοφτηεσηεαρεδστρυχτυρεοφτηεαγνετιχφιελδ, ωηιχηισινιτιαλλψφορεδδυρινγτηεβεαινϕεχτιονπηασεωιτη θ(0)∼0.3.Αφτερτηεενδοφτηε βεαινϕεχτιοντηεσηεαρεδστρυχτυρεοφτηε αγνετιχφιελδεξιστσφορσοετιεανδτηεν δεχαψσ.Εϖολυτιονοφπλασαχυρρεντσωασαλ σοοδελλεδωιτη2−ΜΗχοδε,ωηιχησηοωσ χλοσεσιιλαριτψοφχορεπλασαδψναιχστο τηεωελλ−κνοωνΚαδοτσεϖοδελφορσαωτεετη ιντοκαακ. P-5.063, Friday July 1, 2005

New Adaptive Grid Plasma Evolution Code SPIDER A.A.Ivanov1, R.R.Khayrutdinov2, S.Yu.Medvedev1, Yu.Yu.Poshekhonov1 1Keldysh Institute of Applied Mathematics, Moscow, Russia 2TRINITI, Moscow Region, Russia Non-linear time-evolution codes are essential tools for modeling of existing and future tokamak experiments. There exist several well known free-boundary equilibrium evolution codes, such as DINA [1], PET [2] and TSC [3]. In all existing evolution codes 2D equilibrium problem is solved on rectangular grid and special mapping technique is required to solve 1D transport and magnetic field diffusion equations self-consistently with 2D free- boundary equilibrium. To achieve high accuracy of simulations a lot of iterations between 2D equilibrium and 1D set of equations are required, that leads to a substantial increase of the simulation time. Another problem arises in case of plasma equilibrium with high edge or non-monotonic plasma current density. The size of the rectangular mesh needed to resolve such equilibrium features should be increased several times leading to an order of magnitude increase in the time of simulation. A NEW nonlinear time-evolution free-boundary equilibrium code SPIDER has been developed, in which 2D equilibrium problem is solved using adaptive grid method. As a result the coordinates of magnetic surfaces are obtained and no additional mapping is required to solve the 1D set of transport equations. Additional acceleration of the computations and high accuracy are achieved by simultaneous solution of 2D equilibrium, 1D transport and magnetic field diffusion equations, and circuit equations for PF coils and vacuum vessel eddy currents. Plasma equilibrium with high edge or non-monotonic plasma current density does not require any additional increase of mesh size and does not increase the time of computation. Examples of test simulations for ITER plasma will be presented and evolution of equilibria with high edge or non-monotonic plasma current density will be studied. References [1] R. R. Khayrutdinov, V. E. Lukash. Studies of Plasma Equilibrium and Transport in a Tokamak Fusion Device with the Inverse-Variable Technique, J. Comput. Physics, v. 109, 193 (1993) . [2] S.A. Galkin, V.V. Drozdov, A.A. Ivanov, S.Yu. Medvedev, Yu.Yu. Poshekhonov, GA Report, GA-A23045, May 1999 [3] S.C. Jardin, N. Pomphrey, and J. DeLucia. Dynamic Modelling of Transport and Positional Control in a Tokamak. Journal of Computational Physics, v. 66, 418 (1986) P-5.064, Friday July 1, 2005

Magnetic ELM Triggering and Edge Stability of Tokamak Plasma

S.Yu.Medvedev1, A.A.Ivanov1, A.A.Martynov1, Yu.Yu.Poshekhonov1, S.H.Kim2, J.B.Lister2, Y.R.Martin2, O.Sauter2, L.Villard2, R.R.Khayrutdinov3 1Keldysh Institute, Russian Academy of Sciences, Moscow, Russia 2CRPP, Association Euratom-Confédération Suisse, EPFL, Lausanne, Switzerland 3TRINITI, Moscow Region, Russia

Ideal MHD external kink modes driven by large current density and pressure gradient values in the pedestal region of the tokamak plasma are believed to trigger the edge localized modes (ELM). Since the ELM triggering mechanism depends on the edge current and pressure profiles, a modification of these parameters can lead to a variation of the ELM cycle controlling their frequency and amplitude. In the experiments on TCV the vertical oscillation of the plasma generates edge current which, in turn, triggers or inhibits ELMs depending on the sign of the induced edge current [1]. Similar experiments have been performed on ASDEX Upgrade (AUG) to test this technique on an additionally heated H-mode plasma with type I ELMs [2]. Detail quasi equilibrium modelling of the edge current induction in the ELM triggering experiments is performed with the PET code integrated into the DINA-CH Simulink environment taking into account the magnetic field diffusion and with realistic plasma profiles in the pedestal region. The dependence of the edge current perturbations on the plasma temperature is studied. Ideal MHD stability of the equilibrium sequences in the ELM controlling schemes are investigated and compared to the edge stability scalings [3] derived from the results of the calculations with the KINX code that includes plasma up to the separatrix. For better understanding of the underlying physics and differences in the general behavior at AUG compared to TCV the stability boundaries against edge kink/ballooning modes are investigated under variations of the pedestal profiles and plasma boundary shape including lower and upper triangularity and also higher moments of the plasma geometry.

References [1] A.W.Degeling et al. Plasma Phys. Control. Fusion 45 (2003) 1637 [2] P.T.Lang et al. 20th IAEA Fusion Energy Conference, Vilamoura, Portugal, 1-6 November 2004, IAEA-CN-116/EX/2-6 [3] S.Yu.Medvedev et al. 31th EPS Conference on Contr. Fusion and Plasma Phys., London, 28 June - 2 July 2004, ECA Vol. 28G, P-2.147

Acknowledgements The CRPP authors are supported in part by the Swiss National Science Foundation. P-5.065, Friday July 1, 2005

Reversed Current Density in Tokamaks: Equilibrium and Stability Issues

A.A.Martynov1, S.A.Galkin1, S.Yu.Medvedev1, L.Villard2 1Keldysh Institute, Russian Academy of Sciences, Moscow, Russia 2CRPP, Association Euratom-Conf´ed´eration Suisse, EPFL, Lausanne, Switzerland

The ”current hole” tokamak regimes with nearly zero toroidal current in the central region feature a robust formation of the internal transport barrier, good confinement properties and may be naturally attained in larger toroidal plasmas. Systematic studies of the MHD equilibrium and stability of such configurations require a development of the numerical tools capable to deal with the new types of magnetic surface configuration. The use of unstructured adaptive grids presents an attractive possibility of an efficient and versatile approach to the MHD computations with variable magnetic surface topology. To solve the Grad-Shafranov equation on triangular grids standard finite element code was developed. The grid adaptation technology is based on a combination of isotropic refinement of ”macrocells” and edge refinement which allows to generate anisotropic grid (triangular cells stretched in the direction of slow change of the solution) of acceptable quality. For the first experiments two families of force-free configurations were chosen: reversed current density equilibria with piece-wise constant ff ′ [1] and configurations with linear profiles f(ψ) = λψ reproducing analytical examples from [2]. For stability study of the configurations without nested magnetic surfaces the most general and promising approach would be to refuse from the use of special coordinates and magnetic surface projections. From this point of view the ideas of hybrid finite elements and role of the grid edge alinement to the magnetic surfaces for good approximation of MHD spectrum are discussed. For the cylindrical plasma with piece-wise constant current density profile it is possible to get analytically exact criteria of external kink mode stability. ”Current hole” profiles with zero or negative core current density are considered. Specifics of the configurations are a presence of zero poloidal field surface and internal resonant surface for all the mode numbers m, n. The 1D model shows how the location of the surfaces influences external kink mode stability: ”current hole” profiles make the stability properties worse compared to monotone profiles – higher values of global shear are needed for stability; negative core current has destabilizing effect due to outward shift of the zero poloidal field surface.

[1] A.A.Martynov, S.Yu.Medvedev, L.Villard, Phys. Rev. Lett. 91(2003)85004 [2] A.A.Martynov, S.Yu.Medvedev, Plasma Phys. Reports, 28, 259-267 (2002)

Acknowledgements The CRPP author is supported in part by the Swiss National Science Foundation. P-5.066, Friday July 1, 2005

Investigation of internal transport barrier in sawtooth oscillation (OH) and sawtooth stabilized by off-axis ECRH heating modes.

I.S.Bel’bas1, V.F.Andreev1, A.V.Gorshkov1, V.I.Pozdnayk1 , K.A.Razumova1, V.V.Sannikov1, ,V.A.Zhuravlev1, H.J. van der Meiden2, T. Oyevaar2 1 RRC ”Kurchatov institute”, Moscow, Russia 2 FOM-Institute for Plasma Physics, Neiuwegein, Netherlands

ITBs are studied on the T-10 tokamak (R=1.5m, a = 0.3m, B!2.4 T, I ! 150 ∀#, ne!1.5*10-19 m-3) in ohmic heating mode and in discharges with stabilized sawtooth oscillations. Electron temperature profiles (Te) are measured with high spatial resolution (0.007 m) TV thomson scattering and a 21-channels radiometer using the second harmonic of the electron-cyclotron emission. Microwave reflectometer register appearance of a barrier on electron density in OH mode. Existence of ITB in OH mode is confirmed that switching central ECRH force

electron temperature growth only in the area restricted within approximately r=rq=1, during first few millisecond. Afterwards, being reached a critical gradient of temperature, rise of temperature spreads over out plasma region. The same phenomena are observed under the condition of switching on ECRH in sawtooth stabilized discharges, but barrier occurs on radius corresponding somewhat more q ! 1.5. In paper are considered evolution and possible physical reasons of ITB weakness in above mentioned modes. This work is done in accordance with grant NWO 047.016.015. P-5.067, Friday July 1, 2005

Experimental Studies of the Ion Plasma Component Behavior during Disruptive Instability in Tokamaks N. N. Brevnov, Yu. V. Gott, and V. A. Shurygin Russian Research Centre Kurchatov Institute, Kurchatov sq., Moscow, 123182 Russia

Results are presented from experimental studies of the plasma ion component behavior during disruptive instability in the TVD and DAVAMAND tokamaks. It is shown that in major and minor disruption ions behave in different ways. During a major disruption the ion temperature increases by a factor of 1.5--2. The ions are accelerated predominantly across the magnetic field near the rational magnetic surfaces. Results on the ion acceleration along the magnetic field indicate that disruptions are accompanied by the generation of longitudinal electric fields that are aligned in opposite directions at the plasma periphery and near the plasma axis. During a minor disruption the ion temperature practically do not change and some amount of ions is accelerated predominantly along magnetic field. It was proposed that the ion acceleration by longitudinal electric fields which are generated during disruption take place both in major and minor disruption. During the major disruption velocity of accelerated by this electric field electrons are greater then a threshold of some ion-cyclotron instability excitation and we have a resonance ion acceleration across the magnetic field.

P-5.068, Friday July 1, 2005

Irregularity of the Magnetic Island Rotation under External Helical Magnetic Perturbation in T-10 Tokamak

N.V. Ivanov1, A.M. Kakurin1, I.I. Orlovskiy1 1 Nuclear Fusion Institute, RRC «Kurchatov Institute», Moscow, Russia E-mail: [email protected]

The T-10 tokamak experimental observations of the m = 2 mode rotation irregularity is analysed to develop a method of the error field identification. The structure of the MHD perturbation is measured with poloidal magnetic field sensors located at the inner surface of the vacuum vessel wall. The MHD signal processing procedure includes the space Fourier transform. It is supposed that the observed variations of the tearing mode amplitude and rotation velocity, which take place along with the magnetic island rotation, can be attributed to the effect of the error field. The experimental results are simulated using a rotation model for the non-linear Rutherford tearing mode in the presence of an external helical magnetic field perturbation of the same helicity. The rotation velocity of the mode is represented by a superposition of the velocity of the mode with respect to the plasma that arises as a characteristic of an externally driven non-linear tearing mode and the plasma rotation velocity affected by the plasma current interaction with the external magnetic field. In the figure one can see a comparison between the T-10 experimental data and the result of the simulation for the conditions of the experiment. In the calculations, the error field m = 2, n = 1 component equal to 4©10-4 T was chosen to produce the experimental level of the

irregularity of the instantaneous frequency, Y *t+ ? d]arctg*BS BC +_ dt .

(a) (b) 12 12

)

t 6 T] 6 ) t T] -4 ( C -4 0 ( 0 Experimental (a) and C B [10

-6 B

[10 -6 -12 -12 simulated (b) waveforms 12 12 of the cosine, Bc(t), and ) ) T]

t 6 T]

t 6 -4 ( -4 ( S 0 S 0 sine, Bs(t), components, B B [10 -6 [10 -6 amplitude, AmplBし(t), -12 -12 )

) 12 12 t t and instantaneous ( ( s s T] 8 T]

B 8 B -4 -4 frequency, っ(t), of the 4 4 [10 [10 tearing mode m = 2, Ampl 0 Ampl 0 12 12 n = 1 polidal magnetic ) t

) 8 8

t field perturbation ( ( Y

4 [rad/s] 4 Y

[rad/s] 0 0 470 472 474 476 478 480 72 74 76 78 80 t [ms] t [ms]

P-5.069, Friday July 1, 2005

Analysis of RWM feedback systems with internal coils V.D. Pustovitov, M.S. Mayorova Russian Research Centre “Kurchatov Institute”, Moscow, 123182, Russia

In recent experiments in the DIII-D tokamak, the feedback using new Internal-Coils was found to be more effective when compared to previously operated system with the External-Coils located outside the vacuum vessel [1]. Here we analyse the Resistive Wall Mode (RWM) feedback system with internal coils (inside the vacuum vessel) using analytical approach based on cylindrical approximation, as described in [2]. The model, which was tested against the numerical results and was found reliable [3], is generalized so that it allows now to consider the RWM feedback control with either external or internal correction coils, or with both. We compare the efficiency of the RWM suppression for system with different positioning of the correction coils and with different choice of the magnetic sensors in the feedback chain: with radial field sensors, poloidal field sensors inside the vessel, and poloidal sensors outside. The comparison is performed using the mode rigidity assumption, which means independence of the mode structure in the plasma on the applied feedback field. The feedback schemes are compared via the gain factor K necessary for suppressing the mode with a given growth rate. For quantitative estimates, a proportional control is assumed. The analysis shows that, when the radial magnetic field is used as the input signal, the system with internal coils is always better than a similar one with external coils: smaller K is needed to suppress the mode. The same is true for a system with external poloidal probes. When internal poloidal probes are used, the result depends on the growth rate of the mode to be stabilized. Several tests are proposed for experimental verification.

[1] Okabayashi M. et al, Control of the Resistive Wall Mode with Internal Coils in the DIII-D Tokamak, Paper IAEA-CN-116/EX/3-1Ra, in Fusion Energy 2004 (Proc. 20th Int. Conf. Vilamoura, 2004). [2] Pustovitov V.D., Plasma Physics and Controlled Fusion, 44 (2002) 259. [3] Bondeson A. et al, Nucl. Fusion 42 (2002) 768.

P-5.070, Friday July 1, 2005

Sufficient stability condition for axisymmetric equilibrium of flowing magnetized plasma V. Ilgisonis1, I. Khalzov1 1 RRC Kurchatov Institute, Moscow, Russia

New variational principle for linear stability of three-dimensional, inhomogenious,

compressible, moving magnetized plasma is suggested. The sufficient stability

condition is obtained, which appears to be softer (easier to be satisfied) than all

previously known variational stability conditions. The key point of the analysis is a

conservation in variations of new integrals inherent in the linearized equation of the

motion that was not earlier discussed in the literature. The principle is concretized for

axisymmetric plasma equilibrium with stationary toroidal flows. P-5.071 , Friday July 1, 2005

Plasma potential structure nearby the magnetic island in the TUMAN-3M tokamak Askinazi L.G., Lebedev S.V., Kornev V.A., Krikunov S.V., Tukachinsky A.S., Vildjunas M.I., Zhubr N.A. Ioffe Institute, 194021, St.Petersburg, Russia

Radial electric field may play a role in transport barrier formation near rational q surface. The paper presents the results of experimental study of peripheral plasma potential and electron density in TUMAN-3M tokamak in presence of a rotating tearing mode island. It is believed that plasma potential and density should be constant on the magnetic surface. Thus, measuring spatial distribution of these parameters in a region affected by the rotating island, one can deduce the structure of magnetic surfaces. On the other hand, if surfaces of

φ=const and ne=const do not correspond one to another, it may indicate that plasma potential is not constant on the magnetic surface. In this work we measure plasma floating potential and electron density using electrostatic probes immersed into the core plasma periphery down to the 4cm inside LCFS. The probe design and materials allowed it to withstand strong thermal flux from the core plasma. The potential, as well as electron density, was measured simultaneously in several

radial locations. In a process of φ=const and ne=const surfaces structure reconstruction, information on poloidal number m obtained from Mirnov coil array was used. Experimentally

measured φ=const and ne=const surfaces structure is compared with one calculated using Mirnov coil data. The method works better for the MHD oscillations with sufficiently high m =3…5, when the island is located close to the plasma periphery.

P-5.072, Friday July 1, 2005

Mitigation of disruption-generated runaways by means of ECRH

B. Esposito1, G. Granucci2, J.R. Martin-Solis3, M. Leigheb1, L. Gabellieri1, F. Gandini2, D. Marocco1, C. Mazzotta1, P. Smeulders1

1 Associazione Euratom-ENEA sulla Fusione, CR Frascati, C.P.65, 00044 Frascati, Roma, Italy 2 Associazione Euratom-CNR sulla Fusione, IFP-CNR, Via R, Cozzi 53, 20125 Milano, Italy 3 Universidad Carlos III de Madrid, Avenida de la Universidad 30, 28911 Madrid, Spain

When a population of runaway electrons is present in the plasma current plateau phase of FTU discharges, it is found experimentally that, during electron cyclotron resonance heating (ECRH), the runaway energy decreases and the whole runaway population is often quenched. An analysis of the runaway dynamics based on a test particle model, including acceleration in the electric field, collision with the plasma particles and syncrotron radiation losses, has shown that the cause of such runaway suppression is a drop of the plasma resistivity (and in turn of the toroidal electric field) due to ECRH heating [1]. In order to evaluate the suitability of this scheme for the mitigation of disruption- generated runaways, experiments have been carried out in FTU in which controlled

disruptions (Ip=500 kA, Bt=5.3 T) have been triggered by injection of impurities (typically Mo) through laser blow-off (LBO). On-axis ECRH pulses of 20-100 ms duration have been started within a few ms before the time of the disruption. With respect to ohmic disruptions, the following differences have been found: the current decay is much lengthened; small or no g-ray signals (as measured by a NE213 scintillator working in current mode) due to disruption-generated runaways are observed at an ECRH power level of 0.65 MW; photoneutrons (from runaways) are also found to be reduced. ECRH power as low as 35% of the ohmic power produces a softening of the plasma current decay (i.e. longer decay) and, in several cases, even prevents the disruption when ECRH application is early enough with respect to the disruption time. Additional experiments are being prepared in which the loop voltage signal exceeding a given threshold is used to trigger the ECRH pulse at the disruption.

[1] J.R. Martin-Solis et al., Nucl. Fusion 44 (2004) 974 P-5.073, Friday July 1, 2005

 !∀#!∃∀ #%&∋()

#∗#+,∗∗∀!−. ,/0###∗∀+∗#∃(12!3/0 !#/#14#∗56 !∀ #∃%∀ & & ∋(#∃) ∗+ ∀ ∋,∃) !−∋ ,)∀# . /∀ #∃ 0 ∀ 1,23∀ 4∋!5,)/36,∃ 7896/69: ∋8) ; −70<∃= > 9?7?0∋ ! #! ,) !

≅0∀>,∃∃Α∋,∃9+,∃∃Α)−−+∋,∃∃Α) P-5.074, Friday July 1, 2005

A multiple-time-scale approach to the control of ITBs on JET

L. Laborde1, D. Mazon1, D. Moreau1, 2, T. Tala3, M. Ariola4, and contributors to the EFDA- JET workprogramme*

1 Association Euratom -CEA, DSM/DRFC CEA-Cadarache, 13108, St Paul lez Durance, France 2 EFDA-JET Close Support Unit, Culham Science Centre, Abingdon, OX14 3DB, UK 3 Assoc. Euratom-ENEA-CREATE, Univ. Napoli Federico II, Via Claudio 21, I-80125 Napoli, Italy

The simultaneous real-time control of the current and pressure profiles could lead to the steady state sustainment of an internal transport barrier (ITB) and so to a stationary optimized plasma regime. Recent experiments in JET have demonstrated significant progress in achieving such a control: different current and temperature gradient target profiles have been reached and sustained for several seconds using a controller based on a static linear model identification. Nevertheless, these experiments have shown that the controller was sensitive to rapid plasma events such as transient ITBs during the safety factor profile evolution or MHD instabilities which modify the pressure profiles on the confinement time scale. The control technique has been improved by using a multiple-time-scale approximation in order to better respond to these rapid plasma events. The paper describes the theoretical analysis and closed- loop simulations using the controller that will be tested experimentally in JET during the forthcoming campaign. First, the identification of a fully dynamic linearized model (which differs from the modular approach [1]) has been addressed, taking into account the physical structure and couplings of the transport equations. The model is designed to best reproduce the response of the current and temperature gradient profiles to power modulations. The identification technique is set-up using semi-empirical transport models (implemented in JETTO and ASTRA codes), and is then applied to real experimental data. The full dynamic response is then used to construct a two-time-scale model and design a controller which can respond faster to rapid plasma events, while converging slowly towards the requested high performance plasma state (on the resistive time scale). Simulation tests of this two-time-scale controller and comparisons with the technique used in previous experimental campaigns have already shown a faster and more accurate control. In parallel, a detailed correlation analysis of the experimental data is in progress in order to identify parameters which perturb the measurements or the plasma evolution in a systematic way and could be taken into account in the controller design. For example, a change in the density profile has been found to affect the current density profile measurements. The paper will also describe how a feedforward correction of such perturbation effects can be included in the controller scheme. [1] G. De Tommasi, “Identification of a dynamic model of plasma current density profile”, submitted to 32nd EPS, Tarragona, Spain.

*See the appendix of J.Pamela et al, Fusion Energy 2004 (Proc. 20th Int. Conf. Vilamoura, 2004) IAEA, Vienna. P-5.075, Friday July 1, 2005

H-mode Threshold Experiments on JET Y. Andrew1, R. Sartori2, R. Barnsley1, E. de la Luna3, S. Hacquin4, NC Hawkes1, LD Horton5, A Huber6, A Korotkov1 and contributors to EFDA-JET work programme 1. Euratom/UKAEA Fusion Association, Culham Science Centre, Abingdon, OX14 3DB, UK, 2. EFDA CSU Bolzmannstrasse 2, 85748 Garching, Germany, 3. Asociasion EURATOM-CIEMAT para Fusion, CIEMAT, Madrid, Spain, 4. CFN-Associacao EURATOM/IST, Av. Rovisco Pais, 1049-001 Lisboa, Portugal, 5. Max-Planck-Institut fur Plasmaphysik, EURATOM Association, D-85748, Garching, Germany, 6. Institut fur Plasmaphysik, Forschungszentrum Julich GmbH, EURATOM Association, Trilateral Euregio Cluster, D-52425 Julich, Germany Previous studies of the L-H transition on JET have shown the divertor geometry to have a strong influence on H-mode access with the X-point is sufficiently close to the divertor floor [1,2]. These experiments were conducted under conditions of relatively high plasma density and constant plasma current. A systematic study of the influence of the ion ∇B direction with the MKIIGB divertor with the septum removed, Septum Replacement Plate (SRP) divertor, showed field direction to have little effect on H-mode access on JET [3]. This earlier work is extended here by comparison of H-mode access on JET under conditions of very low plasma density and with slow current ramps. Operation under a

ped 19 -3 very low density regime ne = 0.8-1.8 x10 m with the divertor septum was found to

have a significant effect on the dependence of the L-H threshold power, Pth, on global

plasma parameters. A very sharp ‘turning point’ below which the Pth increases with

decreasing ne is found for Ip/Bt values of 2.5 MA/2.7 T, 2.5 MA/2.4 T and 2.2 MA/2.4 T.

The value of ne for the turning point appears to decrease with increasing q95 from these

three data sets. The low density septum shots also demonstrate a dependence of Pth on q95, which disappears as the density is increased above the ‘turning point’. The lack of

dependence of Pth on q95 at higher densities is further confirmed by data from slow Ip ramp experiments. The reference septum shots were repeated with the SRP at 2.5 MA/2.7 T and

this ne scan showed no evidence of a Pth ‘turning point’. This result indicates that removal of the septum lowered the ‘turning point’ for H-mode access or changed L-H transition behaviour in the low density regime.

In the final part of this study, results demonstrating the similarity of Pth for forward and reversed field operation with the SRP divertor [3] are confirmed by comparison with similar experiments conducted with the septum in place. [1] Horton LD, PPCF 42 (2000) A37-A49., [2] Y Andrew et al., PPCF 46 (2004) A87- A93., [3] Y Andrew et al., PPCF 46(2004) 337-347. Acknowledgement This work has been performed under European Fusion Development Agreement (EFDA) and was partly funded by UK Engineering and Physical Sciences Research Council and by Euratom. P-5.076, Friday July 1, 2005

Comparison of High Density Discharges Heated Ohmically and with NBI in the Globus-M Spherical Tokamak V.K. Gusev1, A.G. Barsukov2, F.V. Chernyshev1, I.N. Chugunov1, V.E. Golant1, V.G. Kapralov3, S.V. Krikunov1, G.S. Kurskiev1, V.M. Leonov2, R.G. Levin1, V.B. Minaev1, A.B. Mineev4, I.V. Miroshnikov3, E.E. Mukhin1, M.I. Patrov1, Yu.V. Petrov1, K.A. Podushnikova1, V.V. Rozhdestvenskii1, N.V. Sakharov1, A.E. Shevelev1, A.S. Smirnov2, A.V. Sushkov2, G.N. Tilinin2, S.Yu. Tolstyakov1, V.I. Varfolomeev1, M.I. Vildjunas1 1 A.F. Ioffe Physico-Technical Institute, St. Petersburg, Russia 2.NFI RRC “Kurchatov Institute”, Moscow, Russia 3 Saint-Petersburg State Politechnical University, St. Petersburg, Russia 4 D.V. Efremov Institute of Electrophysical Apparatus, St. Petersburg, Russia One of the most attractive fusion relevant scenarios is a high plasma density regime which was investigated in the course of Globus-M experiments (the machine is described in [1]). The experiments were performed at conditions: toroidal field – 0.4 T, major and minor radii ~ 0.35 m and ~0.23 m respectively, plasma current – 0.2-0.28 MA, safety factor at 95% flux surface 3.5-6, plasma elongations ~1.5-1.7. Magnetic configuration was mainly limiter one with few shots when plasma detached from the central column and exhibited separatrix configuration. In the shots with NB injection the neutral beam power was ranged from 0.4 to 0.6 MW, which was added to OH power ranged from 0.3 to 0.5 MW. Comparison of density

limits in pure OH discharges with increased power deposition (POH > 0.5MW), high current and high current density was made to auxiliary NBI heated discharges. In the regimes with high power deposition plasma average density increase up to 8 1019m-3 was measured, which is currently the record value for Globus-M. Such regimes are characterized by low value of

q95 ~ 3.5 and demonstrated low value of MHD activity recorded by Mirnov probes. Experimental data analysis, including data from magnetic diagnostics, Thomson scattering and NPA helped to compare relative electron and ion heating efficiencies in the regimes with high plasma density. Obtained density limits are described and compared with achieved previously. Reasons currently limited maximum densities in the Globus-M spherical tokamak are discussed, among which is disruptive MHD instability. The work is supported by RF Ministry of Education and Science, Rosatom RF and RFBR grant 03-02-17659.

[1] V.K. Gusev, V.E. Golant, E.Z. Gusakov, et al., Tech. Physics, v.44, (1999) No.9, pp. 1054-1057 P-5.077, Friday July 1, 2005

!!∀#∃∀% ∀&∀∋#∀∀∋(∀∃)&∗+∀,) !∀ #∃%&∀#!∀ ∋&()!∗+,−)∀

−,!∀−&+∀∀./,! !,,∀&,∀−&!,!.&&− ∀−&!&∀∀∀&01&∀ ∀&∀∀−/,,&∀∀∀− ∀∀&,∀.&−∀∀∀,∀∀&.!& 2∀&/,,∀−/, .∀.&.−∀,∀ ,∀,∀−/&∋34(∀−,56 /∀!,∀/&.!∀− & ∀∀&!0(1∀&.,∀−∀ ,.∋778 ,!9−,! .&−∀&!∀−/ ∀& −∀&∀,∀&∀&∀− ,−&∀ ∀−∋:& ∀∀ ∀;∋,∀−&/.,∀&−∀,, ,∀∀/&!∀−)9:73 9) /&&,∀ .∀77/∀.<,∀&,!<,∀, ∀:7∀∀−!&∀∀/!∀−= ∀∀ ∀− . 9, ∀ ,∀∀ / &!

3 9) −7 &!.∀7 ,∀.∀ − ∀.& ∀− ,∀∀/∀−,,∀.!∀−(∀&& &,∀.∀−∀−∀&

56 !! ∀ ∗ +∀, # .∃ / )∀/∀∋0123#4## (∀.∀ 4(∀ 9873=0∋77∋1 P-5.078, Friday July 1, 2005

Simulation of burning plasma in multivariable kinetic feedback control system V.M. Leonov1, Yu.V. Mitrishkin2 1Nuclear Fusion Institute, 1 Kurchatov Sq., Moscow, Russia, [email protected] 2Inst. of Cont. Sci., 65 Profsoyuznaya St., Moscow, Russia, [email protected]

The paper is devoted to developing methods and systems of identification and kinetic feedback control of burning plasma. Numerical simulations are carried out for ITER parameters. In the approach given plasma column is considered as a nonlinear multi-linked dynamical system. It has space-distributed parameters, input and output constraints (saturations) and is described by a set of coupled differential equations of diffusion type. Plasma-physics code ASTRA has been used for the qualitative investigation of plant internal couplings by means of determining plasma responses of Pfus, ne, Uloop, J(r), q(r), Te, Ti, くp,

Pdivertor on test-step actions PNBI, PRF, fuelling, Arseeding. In doing so, one observes two inherent times of plasma parameters change. One of these times is appropriate to the plasma transport time, and the other one is related with the reconstruction of current profile and essentially exceeds the first one (about 2 orders). After transient responses on the test-step actions the output signals return to nearby vicinity of their initial values. This phenomenon shows the existence of a self-compensation of input actions by internal couplings of burning plasma.

The 4-loops control system containing separate feedback channels is used to stabilize Pfus, ne,

Uloop, Pdivertor simultaneously with corresponding plant input actions PNBI, fuelling, PRF,

Arseeding. Simulations on ASTRA code show that proportional regulations in each control channel stabilize output signals around their specified values. The reasonable amount of stabilization accuracy is achieved at the action of impulse disturbances of various natures. The results are compared with the simulations when the feedbacks are disconnected. Therewith the considerable plasma parameters perturbations are observed which sometimes lead to the whole cooling of plasma column. The stabilization without decoupling says about weak internal plant couplings, which permits to apply multivariable diagonal controller without additional external cross-couplings. Sometimes the hierarchical principle of control priorities change is used at the stabilization of the set of output variables. When the control power PNBI saturates the other actuator is automatically connected up to stabilize Pfus and the stabilization loop of that actuator is broken at the same time. Further stages of the work are discussed, which assume applying separate and combined control of radial profiles of ITER plasma basic parameters taking into account technical constraints adopted in the ITER project. P-5.079, Friday July 1, 2005

Analysis of the of the supra-thermal electrons during disruption instability in the T-10 tokamak

P.V.Savrukhin, A.V.Sushkov, E.V.Alexandrov Russian Research Centre "Kurchatov Institute", 123182, Moscow, Russia

Generation of electron beams with nonthermal energies (Eγ ~100keV) is one of the common feature of the disruption instability in tokamaks. Resent experiments in T-10 tokamak have indicated that the beams are often characterised by narrow localisation around magnetic surfaces with ) T10-31556

A 5 )

rational values of the safety factor M 0.2 major ( V ( ,

p , disruption l (q=1,2). Such localisation of the I 0.15 U 1 beams can be connected with ) 2 energy 2-30 keV .u. a

( quench gas detector

strong electric fields induced due a r x 0 xwd35 I to reconnection of the magnetic 2-150 keV 2 CdTe field lines during growth of the ) tangential .u.

a view (

large-scale MHD perturbations. y a r

x -1 Analysis indicated, that while the I 0.2 ms 0 density of the nonthermal 800 T i m e , ( m s ) 820 840

electrons induced during the )

.u. 2 -1 a

reconnection is two orders of ( 3-5 ms

y a r x magnitude smaller than the I 1 equilibrium plasma density, they 829 T i m e , ( m s ) 831 832 833 can substitute considerable 1.6 -1

) 30-40 ms

fraction of the plasma current .u. 1.4 a (

y a

around the rational q surfaces and r

x 1.2 I can lead to growth of the runaway 831.8 831.9 832.0 T i m e , ( m s ) 832.3 avalanches during major Fig.1. Intensive x-ray bursts are observed during disruption. Present paper major disruption in T-10 tokamak using CdTe represents new results on studies of spatial localisation and temporal evolution of the nonthermal electron beams during density limit disruption and evaluate role of the primary electrons in intensive hard x-ray spikes in post-disruptive plasma in the T-10 tokamak (see Fig.1). The work is supported by Russian Fund for Basic Research (Grant 05-02-17294). P-5.080, Friday July 1, 2005

/LQHDUL]DWLRQRI,7(5SODVPDHTXLOLEULXPPRGHORQ',1$FRGH 91'RNXND

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

ES DQGOL SDUDPHWHUVDUHWDNHQLQWRDFFRXQWLQOLQHDUPRGHO7KHSURFHGXUHRIQXPHULFDO OLQHDUL]DWLRQRIQRQOLQHDUPRGHORI,7(5SODVPDHTXLOLEULXPKDVEHHQFDUULHGRXWIRUJULG VL]H QUuQ] ZLWKWKHQXPEHURIQRGHVuDQGuLQWKHVROXWLRQGRPDLQRI*UDG 6KDIUDQRYHTXDWLRQDQGIRUSDVVLYHFRQGXFWLQJVWUXFWXUHVDSSUR[LPDWHGDVDVHWRIDQG D[LDOV\PPHWULFFRQGXFWRUV,QWKHIUDPHRIOLQHDUPRGHOZHKDYHDQDO\VHGWKHLQIOXHQFH RIWKHJULGVL]HVDQGWKH QXPEHURIXVHGILODPHQWVRQWRSODVPDHYROXWLRQYHUWLFDOJURZWKWLPH DQGWKHSDVVLYHVWDELOL]DWLRQTXDOLW\RIYDULDQWVSUHVHQWHG7KHFRPSDULVRQVRIWUDQVLHQW

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riday July 1, 2005

Direct Measurements of the Electron Temperature by a Ball-pen/Langmuir probe

J. Adámek, C. Ionita,1 R. Schrittwieser,1 J. Stöckel, M. Tichy,2 G. Van Oost3 Institute of Plasma Physics, Association EURATOM/IPP.CR, Prague, Czech Republic; 1University of Innsbruck, Austria; 2Charles University, Prague; 3Ghent University, Belgium.

We have used a combination of a ball-pen probe [1] with a cold probe to determine the

plasma potential Hpl and the floating potential Vfl at the same position in the edge region of the CASTOR tokamak simultaneously (see figure). The measurement of the plasma potential by means of the ball-pen probe utilises the difference between the electron and ion gyroradii: by shifting the collector inside the boron nitride screening tube, the electron current that reaches the collector can be adjusted to the same magnitude as that of the ion current. In this case, in a Maxwellian plasma the floating potential of the collector is equal to the plasma potential. In the combination shown in the figure, the normal cold probe floating potential

Vfl is measured by a ring-shaped Langmuir probe of 3 mm diameter located around the tip of the BN screening tube. The probe is made of a 0.2 mm diameter tungsten wire. The electron temperature can be derived from the relation between the two potentials:

e i Vfl = Hpl / cTe, where c = ln*I sat I sat + represents the ratio of the electron and ion saturation currents to the Langmuir probe-ring. The theoretical value of c in hydrogen plasma is about 2.5. In this contribution we present first measurements of the electron temperature in the

CASTOR edge region (R = 40 cm, a = 8.5 cm, BT = 1.3 T, IP B 10 kA) by using this technique. Furthermore the measured values are compared to the electron temperatures obtained from the I-V characteristics of the probe-ring and the exposed collector, where it acts as normal cold probe.

References: [1] J. Adámek et al., 31st EPS Conf. Plasma Phys. (London, Great Britain, 2004), Europhys. Conf. Abst. 28G (2004), p. 5.120. P-5.082, Friday July 1, 2005

New Results on the Laser Heated Emissive Probe for Direct Measurements of the Plasma Potential R. Schrittwieser,a A. Sarma,a,b G. Amarandei,a,c C. Ionita,a T. Klingerd, O. Grulke,d A. Vogelsang,d T. Windischd aInstitute for Ion Physics, University of Innsbruck, Austria, bBirla Institute of Technology, Mesra, Ranchi, India, cFaculty of Physics, "Al. I. Cuza" University, Iasi, Romania, dMax Planck Institute for Plasma Physics, Greifswald Branch, EURATOM Association, Germany

The direct measurement of the plasma potential poses a major challenge to plasma diagnostics, and the sole tools for localized measurements of this essential plasma parameter are certain forms of electric probes. In principle the plasma potential can be derived from the current-voltage characteristic of a cold probe, or it can be determined from its floating potential, provided the electron temperature is known. But these methods work only indirectly and without sufficient temporal resolution. In addition, there are systematic errors when there is an electron drift or beam. However, the floating potential of a probe, which emits a strong electron current into the plasma, thereby compensating the plasma electron current, is in principle able to yield a good direct approximation of the plasma potential. We present theoretical considerations and experimental results on the heating of dif-

ferent materials (in particular LaB6, and graphite) by an infrared laser to temperatures where they emit a sufficiently strong electron current. We have taken into account the absorption coefficients, reflection coefficients and work functions of the respective materials. We have

constructed two types of emissive probes consisting of cylindrical pieces of LaB6 or graphite, respectively (2 mm diameter, 1 mm height), held by molybdenum wires and heated by fo- cussing the laser light on them through a quartz window. An infrared diode laser with a maximum power of 50 W and a wavelength of 808 nm was used. Initial tests were performed in a vacuum chamber of 20 cm length and 14 cm diameter to measure the probe temperatures and to verify their electron emission. Then the probes were inserted into the edge region of the linear helicon discharge plasma machine VINETA at the MPI Greifswald. We found the electron emission from both materials to be sufficiently strong even for dense laboratory and experimental fusion plasmas. For comparison at the same time either a conventional emissive tungsten wire probe or an emissive carbon fibre probe or a normal cold Langmuir probe were inserted into VINETA. P-5.083, Friday July 1, 2005

Rational surfaces determination from ECE radiometry in the TCABR tokamak A.M. M. Fonseca1, R.P. da Silva1, Yu. K. Kuznetsov1, A. Vannucci1, C.J.A. Pires1, L.F. Ruchko1, V.S. Tsypin1 1 Instituto de Física, Universidade de São Paulo, Rua do Matão, travessa R, 187

São Paulo, 05508-900, SP, Brazil.

E-mail: [email protected]

Measurements of the electron temperature profiles have been performed in the Tokamak Chauffage Alfvén Brésilien (TCABR) with a high radial resolution, using an Electron Cyclotron Emission (ECE) radiometer. Both the absolute temperatures and

the low frequency Te perturbations along the radial direction of the plasma column show distinct peculiarities, which could be associated to the presence of magnetic

islands in the TCABR tokamak circular plasmas. Measurements of Te(r) fluctuations, originated by the heat propagation after sawtooth collapses, is proposed here as an effective and direct method to evaluate the radial positions and widths of the magnetic islands. P-5.084, Friday July 1, 2005

A prototype resistive bolometer for ITER

L.Giannone, D.Queen1, F.Hellman1, J.C.Fuchs and the ASDEX Upgrade Team

Max-Planck-Institut für Plasmaphysik, EURATOM-IPP Association, D-85748 Garching, Germany 1University of California, Berkeley, CA 94270-7300, USA

A prototype of a radiation hard resistive bolometer, with a platinum absorber and meander on a low stress silicon nitride substrate, has been produced and installed in ASDEX Upgrade. The aim is to test its long term stability in a tokamak environment. The original versions of the bolometer with a gold absorber and resistive meander on either a Kapton or mica isolating substrate [1] have been shown to perform unreliably in an environment with the high neutron fluences expected in ITER [2]. The neutron thermal cross section of platinum is a factor 9 smaller than that of gold and suggests that the suspected transmutation problem experienced with the gold meander could be slowed down by substitution with platinum. Silicon nitride is currently being investigated as one of a number of ceramics to be used in the ITER environment. The prototype bolometer could then be a suitable radiation hard bolometer for operation in ITER. Aspects of resistive metal bolometer foil design and possible alternatives for ITER have also been previously considered [3,4]. The prototype bolometer has been successfully operated for two months, indicating that the 1.5 micron thick ceramic substrate is robust enough for use in a tokamak. Absorber thicknesses up to 1.5 microns have been produced but this thickness needs to be increased to 10 microns for ITER. The calibration factors of the prototype bolometer and an original Kapton isolator and gold bolometer have been measured as a function of temperature. The temperature coefficient of the 300 nm thick Pt meander is only 5% larger than that of the 200 nm thick Au meander. The prototype bolometer has a cooling time constant a factor of 3 smaller and a normalised heat capacity a factor of 2 smaller than the original bolometer. It is important for long pulse operation that the meander resistance, normalised heat capacity and cooling time constant of the measurement and reference foil are closely matched, so that a temperature rise of the bolometer head during a discharge leads only to a minimal offset drift. A comparison of the normalised difference of these values shows that the prototype bolometer needs to be improved.

[1] K.F. Mast, J.C. Vallet, C. Andelfinger, P. Betzler, H. Krause, and G. Schramm, Rev. Sci. Instrum., 62, 744, 1991. [2] R. Reichle, T. Nishitani, E.R. Hodgson et al., Europhysics Conference Abstracts Vol. 25A (EPS, 2001), pp. 1293-1296 [3] R. Reichle, J.C. Fuchs, R.M. Giannella et al., Diagnostics for experimental thermonuclear fusion reactors, (Plenum Press, New York, 1996), pp. 559-570. [4] R. Reichle, M. Di Maio and L.C. Ingesson et al. in Diagnostics for experimental thermonuclear fusion reactors 2, (Plenum Press, New York, 1998), pp. 389-398.

P-5.085, Friday July 1, 2005

¢¡¤£¦¥¨§¦© © ££¡ © £¥ "!$#&%"'(*)¡¤+,

-/.10325476892;:<- =?> .HGI.HJ*25KB4ML¤256NK .QPR8TSMS7UT8 4 .QX =BK¤25=S 4 .HZ[8 6NK S . =]M]^4

@BADC E*F E O @BVDW E¤O W W E Y @ W E*\ Y AD_ E

.1` Kacbed G 4725j 25a

F A YfRgih _BV W3k(W

lnmoDprqtsum5vxwNy{z|v}¦~~~¢ Rqtsum;}¦m¦¤}¦Tyt 7m5~;p}N}w¦m5~r;vz^qqm5 ^wNv;

3p<B^* ¢¡¦ ¦m5v¤

d¤£2SM]¥8 j892 SM]78T6£ 4td G 4^25j K¤25S¦L j U j .[©ª]¦8TS¦U 6§2] j«2]t]^K

AD@ VD@A A YfRgih _V W WW§@ W¦¨DW AD_ W A W W

a8Tj U92 j a¨2 8 =BUT2] 4§. K 6 a ] ££25SM] 254]^8T6U St8®S £¢6§4M=B6§8T25U*8Ta 4]N2 6

_ @BW W VDW @ _ A k W AD@¬¤@W W§@ A _ W A _A @ W

£ 4 £¯=°]^=B4 UT254 £¯=S78 j 8T6 S§. a²258 S =B4M6 S ££25SM](8 U S7S S?62 L jB8®SM]^8 =B8®S7K j .

A W VDW A5@ W¦¨ W k¥±tA @ A W A AD@ A W @ W @BV W

J8®47S]^U®³ 8 SM=´6§8 ]µ6 a ] 4 4M]78 S £¦]^K a²2 ]^8®6 U®j/] U ³¶a¨2³·6§25=BS

E @ W§@ AD@¬¤@W W§@ _ AD_ W W A W VD@BW ¬¤W AD_A ADV W

U S7S S £ 8 S 4^2;] j¸8 ]^K j 4 8 KB8T6NK]7K U 2 ]^K UT25S7a²2 8u]^K¹]^K K U

A W A A5@ VDW§@W W @ WW VDW WV AD@xE5± W§@ W ¨5W W_ ± W± A W

4 ³»º 4 a ]U SMS SN¼¦. 6 jBUu³ a²2 ] K½³jB4 j³ 2a²8®6 4M]^=47L¤2]78 Sa²2³ K¤2 6

W@BW V _ A _ A W Y°W AD@ E VD@BW A A @ _W AD@ W§@ @ W

]7K ]74^2 S 4M] £ £25SM]µ8 SI£¯4 a UT25S7a²26 4 ] ]7K j .¾J 2SM]{8 Sµ254 ]^K 6] j

W @ _A A AD@ A _ A W A W«W VDW AD@ W W@¿W<ÀMW W

] ]7K 2UTUÁ;UT8Ta8®] 4L½³n6§UT25S7SM8T62Ui478 U :L¤2 2 2 4 j 4M]7=B47L¤2;]^8 Sº8. . - S^¼¦.©

A W± W _B_ W @ @ A W VDW¸_W A5@ W gÂ @

25jjB8®]78 ] ]7K j ] 478 4^2]78 £*K 2;]^8 ´6§8 6¦³{]7K 4 S7=U®]^8 4iU 2jµa²2³IL 6 a

AD@ A W W W A AD@{A W @V$W W§@ W W @BV$_ A±¦W A W A W

2 4 LBU a 8®]^K ]^= SM] U®8Ta8®] 47S2 j¾]^K 4 £ 4 8®]$K¤25S] L 2 8Tj j 4$a8 8®a²8 j8

_ A W ± @BV W@ W @ W W A W A W ¨5A W A @ C§W @

]¢S] 4M8Ta ]^S.

@BW|Ã W_¶W¦Ã°_ W W§@

K £25S]$8 jB8T2 S]^8®6{a 25SM=B4 S]^K 4 ³¾2 j 8u]^6NK·2 U jB8TS]^4M8TLB=]78 £ 4:

k W AD@ VD@BA W W W¹W§@BW V @ _ @BV W AD@ A W§@W

]^8®6¨U SM]{8 SI£¯4 aÄ]7K UT25S7a²2 6 4 L½³»a 2 S £ ]^K U®8 K½] a8®]M] jÅL½³Å2¶SM6§8 ]78TU®U92] 4

VDW A AD@ A W«_ A W W @ A W V W W @ A

=BSM8 ] j8®Æ 4 ]¥SM³S] aS§Ç2¹È È 625a 472 8®]7K«2µKB8 K¨S 2]78925U*4 S U®=]^8 2 j"2{S ] £

@BV ±¦A W W@ W f W ± V _ W A AD@ @ W A

Z- S 8®]7K«2IKB8 K¸]^8®a 4 S UT=]78 .iZ4 U®8Ta8 254³µ4 SM=BU®]7St]N2Ë 8 ]7K U®8Ta8®] 4iSMK¤25j

É5Ê

k ± V W W A AD@ W @ W W§@ @ W W A±

8u]^K¸jB8®Æ 4 ]8 S =B476 StS 8®]76NK j 6 S 6=]^8 Uu³¸£¯4 aÌ]^K 247L½³ =°]^4^2U¤8 6] 4

± W W§@ AD@ A W ± W AD@ AD@ W ¨DW A W@BW @BW @;ÀMW A

SMK j«]^K 6] j"6 474 U92;] j 8®]76NK«2 U 2 j 4 ³jB8®SM]7478TL=]^8 .iX 4 a ]tU SMS S

A±¦W WW¦Ã°_ W W A W W _ @V W @ W§@BW V AD@ A{_ A _ A W

£¯4 aÍ8 6¦] 4tÎ 4 S L 625=S £*]^K U ] 4 8Tj¤2U°jB8®SM]^2 6 ] ]^K jB892 SM]78T6.i-

A @;ÀMW A ±¦W W WW§@ W W A W AD@BV A A @ W A W V5@BA F¢f

8 jB=B6 j¾U S7S S 4 8Tj ]^8 j¶j=B478 ]^K 4 S 6 £º Î¼ 6§UT25S7SM8T62U] 254M8 a j

@ W A W ±tW W W@ ¬BW @BV Wµ_ W W@ WA E @BW§A W @BV A W

256¦]^8 8u]³ S7K 8 2 4M³¨KB8 K 4 jB8®6§=BUT254 U 68®]³6 a ]2 j«] ]^25U 4 8 S¥L :

¨ E A± @BV ¨DW V _W _ W§@ ¨5W A A _ AD@BW§@ @ A W§@BW V W W

] Î 2 j Ë .- 2S7=B4 a ]^S ££25S]8 U SMS SµjB=B478 ]7K 4 S 6 £ ]7K 4

Ê É5Ê

±tW§W@ @ WÏ W W W§@ A AD@ A W @BV W"_ W W@ WA A W

a j S UT8®Ë d SMKL S$2 j - S 8TU®UxL 4 S ] j .

A W WRk g EB¬ AD@BW @ gÂ ± WI_ W W§@ W

@ÐA W A V @ W W @ W @ @BV¶A W W¨W W @ A W|Ã°_ A W @ A

a²2]^8 - 4]^=B4ML¤2]78 6 ]N258 j/8 ]7K £25SM]38 6]^8 2 8 ] 4725Uia ]7K j

AD@¾AD@ F¢fÑ_ W AD@ AD@ @BW @ W AD@·W<ÀMW AD@ E @ @ W§V W A

K¤2S$L j U jn6 SM8TS]^8 £]^K £25S]R8 j ] 6] 4$] ]7K 4 8u]^Kn2«6 j 26NË2 .

WW§@ W¨5W AD_W AD@ @V"A W A5@ W W A A5VDW W ± A W_ VDW

K 6 j 256NË;2 S78®a¹=BU92;] S¹]7K £2SM]8 ]^472 6¦] 4M8 S£¯4 aÒ8®]7S 478 8 ºL 2a 2]7] 4

k W A W_ VDW W W AD@ ÀMW A W A A V @ W _ W @

6§25UT6=BU92] jnL½³/J°dJiX 6 j ¼$] ]7K 2UTU 4Rj ] 6] 4{8 6§8®j 6 º¯£25S]{8 ]^472 6¦] 4M8 S

W gÓ A W A W± A W W A @ W§@ W AD@ ÀMW A W

A _ W @BVnF Y @ Óf A W A _ AD@ W ±tW§W@ A

UTUT8 2 j]7K a 25S7=4 j£25SM]8 U SMS S 8®UTUx25UTS L 4 S ] jx.

W @BV @ W W W AD@ A W ± A WR_ W W@ W P-5.086, Friday July 1, 2005

Use of a High Resolution Overview Spectrometer for the Visible Range in Fusion Boundary Plasmas S.Brezinsek1, Ph.Mertens1, A.Pospieszczyk1, G.Sergienko1, M.F.Stamp2

1Institut fürPlasmaphysik, Forschungszentrum Jülich GmbH, EURATOM-Association, Trilateral Euregio Cluster, D-52425 Jülich, Germany 2EURATOM/UKAEA Fusion Association, Culham Science Centre, Oxon OX14 3DB, UK

Present-day fusion devices use graphite as a plasma-facing material at the locations of highest particle inﬂux. Owing to erosion and deposition processes, a complex hydrocarbon chemistry takes place at the surface, including the formation of hydrocarbon layers. To monitor the diﬀerent carbon-containing species which are released from the surface, only a limited number of in situ diagnostics are currently available. Passive spectroscopy in the visible range is a standard technique for the investigation

of atomic (D, C, He, B, Ne etc.) and molecular (CD,D2,C2, BD etc.) particle fluxes in the edge of fusion plasmas. To observe and analyse the emission of atoms and molecules simultaneously from a given observation volume, a spectrometer system for the plasma edge has to fulfil high demands with respect to the spectral resolution, wavelength range, time resolution, sensitivity and dynamic range of the detection system. An échelle spectrometer, based on the principle of cross-dispersion, was used as compro- mise to fulfil most of the requirements. The spectrometer (aperture f = 7.0, focal length

fL = 190mm, entrance slit d = 25µm × 75µm) consists of a cross dispersion order-sorter (prism) and a highly dispersive ´echelle grating (ruling 31.6g/mm, size 10.0cm × 3.0cm, blaze angle 63.5◦). The spectrum is reconstructed from the 85th to the 150th spectral order and covers the wavelength range between 372nm and 680nm without gap, providing a constant resolving power λ/∆λ of more than 21000. A cooled 16 bit camera (model: Andor DV434, back-illuminated CCD, 1024 × 1024pixel) with high quantum eﬃciency (Q > 0.9 at 500nm) and 1MHz read-out time is used as a detector.

We present measurements performed in TEXTOR during the injection of C3H4 and of

CH4 in the high-temperature plasma boundary layer. Amongst the atomic species, the + molecular break-up products CH, CH and C2 could be detected simultaneously. Spec- tra of diﬀerent transitions, namely the A − X and B − X band of CH and Swan-band of

C2, could be ro-vibrationally analysed. A similar analysis was also performed on spectra deduced from intrinsically released hydrocarbons in the JET divertor. Signiﬁcant diﬀer- ences in the population of the carbon dimer were observed, especially in the cold inner divertor plasma. Both series of measurements demonstrate the potential and the ability of the system for the use in fusion edge plasmas. P-5.087, Friday July 1, 2005

Bayesian Modelling of Spectrometer Systems and Plasma Profile Inversions from Spectra J. Svensson, R. W. T. König 1 Max Planck Institute for Plasma Physics, Greifswald, Germany

The analysis of a spectrometer measurement is dependent on auxiliary measurements of dispersion curves, calibration curves and instrument function, in turn dependent on spectrometer characteristics, such as focal length, grating constant, CCD chip specifics etc.. We show in this paper how a Bayesian model of the system treated as a whole, can take into account the uncertainties and inter-dependencies of these quantities and their influences on the deconvolution of spectral features. This model is then used to assess the relative influence of these quantities on plasma parameters determined from spectral analysis. In a next step, we show that with such probabilistic models specified for each spectrometer channel, it is possible to join the models for individual channels with differing lines of sight to create, by full radial inversions, the local physics quantities (such as temperatures, rotation velocities etc) which determine the local spectral characteristics. In effect, the inversion process reconstructs radial profiles of the frequency distribution of the local photon emission, parameterised by physics quantities of interest. Also, because of the usage of full Bayesian models for each channel, the inversion process can take into account uncertainties in specific channels and uncertainties in viewing geometry, thus providing ‘error bars’, here Bayesian marginal posterior distributions, on inferred local plasma quantities.

P-5.088, Friday July 1, 2005

Thermal Properties of the CFC-material BN31 at surface heat loading

D. Hildebrandt

EURATOM Association Max-Planck-Institut für Plasmaphysik, Teilinstitut Greifswald, Wendelsteinstr.1, D-17491 Greifswald, Germany

The CFC-material BN31 has been chosen as the plasma facing material for the target elements of the Wendelstein 7-X stellarator presently under construction. The water-cooled targets are designed to receive stationary heat fluxes up to 10 MW / m2. Surface temperature measurements by IR- thermography are supposed for supervision of the heat load arriving at these plates. Specimens of the target material BN 31 and for comparison of the CFC-material N11 have been examined with respect to its thermal response at the surface during heat loading. This paper presents results of experiments involving well defined heat pulses by a diode laser. The laser pulses arrived at the surface with normal incidence. Heat fluxes with densities up to 10 MW / m2 and with durations up to 1s were applied . The surface temperature excursion during and after the laser heat pulse have been measured by different IR-cameras operating in the wavelength region between 3-5 mm (InSb-sensor) or 8-15 mm (VOx-microbolometer). The measurements were made with a spatial resolution of 30 µm (3-5 µm) and 100 µm (8-15 µm), respectively. In order to validate the accuracy of the measured surface temperature evolutions they are compared with corresponding values of an analytic solution of the three-dimensional heat diffusion equation. The temperature pattern of the N11 material shows smaller morphological effects than earlier observed in similar experiments with glancing incident heat fluxes. However, the samples of the material BN31 showed an extremely non-uniform surface temperature distribution even at normal incidence indicating that many microscopic surface parts have poor thermal contact to the bulk material. This causes, in particular for the shorter wavelength region, strong deviations of the measured results from the predicted behaviour for this material. Values of the thermal conductivity at the surface derived from these measurements are compared to those given for the bulk material.

http://eps2005.ciemat.es/ P-5.089, Friday July 1, 2005

High spatial and temporal resolution FM-CW reflectometry to study pellet triggered ELMs at ASDEX Upgrade L. Fattorini1, P. T. Lang2, M. E. Manso1, S. Kalvin3, G. Kocsis3 and the ASDEX Upgrade Team2 1 Centro de Fusão Nuclear, Associação EURATOM / IST, Instituto Superior Técnico, Av. Rovisco Pais, P-1049-001 Lisboa, Portugal 2 Max-Planck-Institut für Plasmaphysik, EURATOM Association, Boltzmannstr. 2, D-85748 Garching, Germany 3 KFKI-RMKI, EURATOM Association, P.O. Box 49, H-1525 Budapest-114, Hungary

Triggering ELMs by cryogenic pellets has been shown a promising technique to investigate ELM physics. Reflectometry with its high temporal and spatial resolution is a diagnostic specially suited for this type of experiments. The ASDEX Upgrade FM-CW reflectometry has 5 channels at low field side (LFS) and 4 channels at high field side (HFS) probing the plasma simultaneously. By operating in swept frequency mode we obtain the plasma response in a large probed plasma region, with a minimum repetition rate of 35 たs. The extracted group delay due to wave propagation to the density cutoff layers is the relevant data for density profile reconstruction. A new method has been developed that gives not only the direct evolution of density profile but also the localized fluctuations caused by the ELMs. It is based on the time evolution of the Fast Fourier Transform of the swept frequency signal at selected plasma layers. Complementary operation in fixed frequency mode resolves the turbulence spectrum at selected layers with higher temporal resolution (1 たs sampling). This diagnostic technique gives information on the characteristics of natural and pellet triggered ELMs. Preliminary results confirm earlier findings[1] that the spatial and temporal dynamics of triggered and intrinsic ELMs are entirely identical. Especially it is shown that triggered ELMs start – like their intrinsic counterpart - to evolve at the LFS despite the fact that pellets impose their massive initial perturbation close to their entrance point into the plasma at the HFS resulting in the peaking of the density profile at the HFS. Combining observations from the reflectometry with data from the fast Mirnov coil system and a fast framing camera system allows investigating with high precision the correlation between the imposed perturbation and its impact on the ELM release. [1] P. T. Lang et al., Nuclear Fusion 44 (2004) 665-677. P-5.090, Friday July 1, 2005

Spatial asymmetries in ELM induced pedestal density collapse in ASDEX Upgrade H-modes

I Nunes1,MManso1, F Serra1, L D Horton2,GDConway2 and the CFN1 Reflectometry and ASDEX Upgrade2 Teams 1 Centro de Fusão Nuclear, Associa¸cão EURATOM-IST, Lisboa, Portugal 2 MPI für Plasmaphysik, EURATOM-Association IPP, Garching D-85748, Germany

Abstract During an ELM event, the time evolution of the density profiles in ASDEX Upgrade type I ELMy H-modes, shows a collapse of the pedestal density and a broadening of the scrape-off layer at both high-field (HFS) and low-field (LFS) side midplane. The collapse is seen first at the LFS leading to the important question of the origin of the density collapse at the HFS. By analyzing the density profiles at both sides in single and double null configurations, it is possible to build a physics picture to explain the density collapse at the HFS. Previous results from ASDEX Upgrade (single null configuration) have shown that the delay between the HFS and LFS density collapse is approximately equal with the ion

parallel transport time (τ//) from the LFS to the HFS midplane. The radial extent of the density collapse in space coordinates is similar but when the HFS density profiles are mapped to the LFS, the radial extent is found to be smaller at the HFS. Density and magnetic fluctuations measurements also show an increase of fluctuations in both LFS and HFS suggesting a perturbation at the HFS. Another way to analyze the density crash is to look at the change in the number of particles at both HFS and LFS at the pedestal and scrape-off layer regions. By determining the number of particles in the region of closed field lines (pedestal) and in the region of open field lines (scrape-off layer) is possible to determine if there is any correlation between the number of particles lost at the LFS after the density crash and those that appear at the scrape-off layer at the HFS. From the set of data analyzed so far, no correlation has been found. Currently there are two possible explanations for the observation of the collapse of the density profile at the HFS. One explanation is that the ELM starts at the LFS where, due to the ELM, there is an outflux of particles across the separatrix into the scrape-off layer (SOL), leaving a density hole (i.e. collapse of the density in the pedestal) which then fills up with particles flowing from the HFS, which in consequence pulls down the density pedestal there. This interpretation is proposed by Oyama et al. based on measurements on JT60-U. Another explanation is that the ELM MHD perturbation begins at the LFS and propagates around to the HFS. The perturbation induces particle transport across the separatrix locally at both the HFS and LFS. In order to further clarify this question we will present results obtained in double null configuration where the flow of particles along the open field lines is prevented. P-5.091, Friday July 1, 2005

A tool for the parallel calculation and ready visualization of the Choi–Williams distribution: application to fusion plasma signals A.C.A. Figueiredo1, J.S. Ferreira1, B.B. Carvalho1, and M.F.F. Nave1 1 Centro de Fusão Nuclear, Associação Euratom–IST, Instituto Superior Técnico, Lisboa, Portugal

The Choi-Williams distribution has been successfully used to analyse some nonstationary fusion plasma signals, for which the short-time Fourier transform spectrogram, wavelets, and the Wigner distribution do not constitute the best approach [1, 2]. Nevertheless, from the practical point of view, a possible drawback of using the Choi-Williams distribution, particularly when compared with the spectrogram, is its larger computation time. In fact, while the spectrogram requires the calculation of a single summation, which is usually done using the fast Fourier transform (FFT) algorithm, in the case of the Choi-Williams distribution, because of its bilinearity, a double summation must be computed. Although this calculation can in part be accelerated by using the FFT to compute one of the summations, further optimization can be achieved by performing independent parts of the calculation in parallel, using several processors. Here, a tool is presented that explores this possibility. The tool relies on a parallel Fortran code that uses the MPI library, and on an IDL graphical interface that allows the user to interactively launch, control and display the results of the calculation in the time–frequency plane. For convenience, the spectrogram can also be calculated and displayed, either separately or simultaneously with the Choi-Williams distribution. The calculations are done in a cluster with 16 nodes running Linux. Magnetic signals from a set of 1 MHz pick-up coils installed in a poloidal cross section of the ISTTOK tokamak [3] are analysed and a comparison of calculation times using 1 to 16 processors is made, thereby showing the advantages of the parallel implementation of the Choi-Williams distribution.

[1] Figueiredo A.C.A., Nave M.F.F., and EFDA-JET Contributors 2004 Nucl. Fusion 44 L17 [2] Figueiredo A.C.A., Nave M.F.F., and EFDA-JET Contributors 2004 Rev. Sci. Instrum. 75 4268 [3] Carvalho B.B., Fernandes H., and Varandas C.A.F. 2003 Rev. Sci. Instrum. 74 1799 P-5.092, Friday July 1, 2005

First results obtained with the real-time PHA diagnostic in the tokamak TCV T.I.Madeira, P.Amorim1, A.P.Rodrigues, B.P.Duval2, C.A.F.Varandas Associação EURATOM-IST, Centro de Fusão Nuclear, Av. Rovisco Pais, P-1049-001Lisboa, Portugal 1Universidade de Lisboa, Faculdade de Ciências, Departamento de Física, Centro de Fusão Nuclear, Portugal 2Association EURATOM-CRPP, Centre de Recherches en Physique des Plasmas, Lausanne, Suisse

Soft X-ray Pulse High Analysis (PHA) is a diagnostic that measures the X-ray spectral plasma emission from high temperature plasmas, from which we can derive Te, Zeff and detect the presence of heavy impurities and/or non-maxwellian electron velocity distributions, using direct numerical analysis. The vertical PHA at TCV is a re-vamped version for the traditional PHA systems. It combines a new generation compact Silicon Drift Detector (SDD) with Digital Signal Processing (DSP) capabilities. The small diode features an integrated thermally stabilized thermoelectric cooler, which can maintain the diode at –10 oC without additional hardware or the need of maintenance, also allowing flexible geometrical arrangements and positioning around the plasma device. Pulse conditioning and processing are performed by a special devoted DSP based VME module. Data is collected into spectra of up to 8k channels with up to 1024 time slices available in the module’s local memory, which can be passed to the host without interrupting the collection process. The optical étendue is real-time regulated by the DSP and provides a robust signal for plasma temperature and impurity content measurement.This combination exhibits an energy resolution (FWHM ~180 eV at 5.9 keV) with short shaping time-constants (0.25-0.5os), high throughput flux (~ 700 kHz) associated to fast signal processing performance (a few os) as it is mandatory in tokamaks of short pulse duration and high-energy fluxes. It addresses the demands of fusion devices for speed, feedback and real-time information. Increasing the frequency response of the system for fast acquisition capabilities and increasing the throughput for better statistics; allows surveys of more spectral time slices, and more detailed study of plasma behavior during the plasma discharge. The performance of the electronic analysis, which permits spectra to be obtained in a period close to the TCV ion confinement time (a few ms), will be explained. The first results showing the effects of ECRH, ECCD,

density modulation, and the generation of eITBs will be presented. P-5.093, Friday July 1, 2005

Study of core and edge plasma structures in TEXTOR with the burst- mode 10 kHz Thomson scattering system

S.K. Varshney1, H. J. van der Meiden1, C.J. Barth1, T. Oyevaar1, M.Yu. Kantor2, D.V. Kouprienko2, R. Jaspers1, A.J.H. Donné1, E. Westerhof1, B. Unterberg3, E. Uzgel3, W. Biel3, A. Pospieszczyk3 and the TEXTOR team

1. FOM-Institute for Plasma Physics Rijnhuizen*, Association EURATOM-FOM, P.O. Box 1207, 3430 BE Nieuwegein, The Netherlands, www.rijnh.nl 2. Ioffe Institute, RAS, Saint Petersburg, 194021, Russia 3. Institut für Plasmaphysik*, Forschungszentrum Jülich GmbH, Association EURATOM- FZJ, D-52425 Jülich, Germany * Partners in the Trilateral Euregio Cluster

A novel TV Thomson scattering diagnostic capable of measuring plasma electron

temperature (Te) and density (ne) profiles at 10 kHz in burst mode has become operational at TEXTOR [1]. The new system features a multi-burst 10 kHz ruby laser and an ultra-fast detection system equipped with two CMOS cameras and a cascaded intensifier stage. The detection branch enables to diagnose either the full plasma diameter of 900 mm with 120 spatial elements of 7.5 mm each, or a 160 mm long top edge chord with 98 spatial elements of 1.7 mm each. The recent operation of the system has been performed in a single burst mode consisting of 12 laser pulses with average energy of about 15 J giving a sequence of 19 -3 core profiles with statistical errors as low as 7% in Te and 3% in ne (at 2.0×10 m ). Similar error bars are expected in the edge profiles which are being presently analysed. After installation of a low Cr-doped ruby rod, the performance of the system is expected to be enhanced towards the design goal of 3-4 bursts of 40 laser pulses each, with pulse energies of about 15 J. The first core observations made on large m=2 island structures induced by the Dynamic Ergodic Divertor (DED) have shown the potential of the new diagnostic. The DED [2] is a tool to produce a moving or stationary ergodization of the magnetic field at the plasma boundary by breaking the closed magnetic flux surfaces in this region. The magnetic island structures are formed in the transition region between the unperturbed and ergodic zones. Detailed dynamic measurements of electron temperature and density distributions inside the island structures will be presented. The effect of DED structures on the plasma edge layer has been observed with the core system. It is found that the temperature profile is lowered in the region of ergodization when the DED is switched on. The ‘magnified’ observations of the DED structures and their effects in the plasma boundary are presently addressed with the dedicated edge Thomson scattering system.

[1] H. J. van der Meiden et al., Rev. Sci. Instrum. 75(10) 2004, pp 3849 [2] K.H. Finken et al., Phys. Rev. Lett. 94, 2005, pp 015003 P-5.094, Friday July 1, 2005

Experimental validation of improved pellet ablation models G. Kamelander, M.König, M.Pachole, G. Sdouz, A Tischner, G.Weimann Austrian Research Center, A-2444 Seibersdorf

Injection of pellets is currently the most developed method to refuel ITER-like plasmas. It has been demonstrated by many experiments that the fuelling effectivity can be considerably f f enhanced if the pellets are injected from the low field side1. This effect based on the E · B drift originating from the polarization of the ablatant due to the curved magnetic field is included into the model of the pellet code PELDEP described in Ref.[1]. This code is appropriate to be combined with transport codes for numerical simulation of tokamak scenarios as it has been done for the transport code system ASTRA2 . It has been shown that the results are realistic on condition that some input data, especially the homogenization time, are realistic and experimentally validated. The homogenization time v h is well known for Tore Supra but for JET and above all for ITER-FEAT we rely just on estimations. Self-consistent methods are needed to calculate that sensitive parameter. The model should be so simple as possible in order to use the pellet code as a module for ASTRA. Our investigations are starting with an improved theory of Rozhansky3 for the shift of ablated pellet material.

The homogenization time has been obtained developing a relation between mass shift and v h . In order to calculate mass shift and homogenization time the time derivative of the pellet radius r$p depends on the ablation model. We have compared the NGS model with more advanced ablation models and validated these models by the experiment. The density profiles predicted by the improved theory has been compared with profiles from JET experiments e.g. #53212 and other shots before and after inboard injections. An improved agreement with the experiment has been achieved. We present furthermore JET-shot simulations obtained by our improved transport system ASTRA-PELDEP using Bohm-gyroBohm and MMM transport models. Acknowledgement: The work has been performed within EURATOM-ÖAW-Association. The content of the publication is in the responsibility of the authors and does not necessarily represent the views of EC. Thanks also to L.Garzotti (JET, RFX), B.Pégourié and G. Pereverzev (IPP) for their precious advices

1 Pégourié B., Garzotti L. 24th EPS, Berchtesgaden, Contr.Fus.Pl.Phys, 21A,213 2 Pereverzev G., Yushmanov P., IPP 5/98 3 Rozhansky V. 30th EPS, St. Petersburg, Vol 27A, P-3.155 P-5.095, Friday July 1, 2005

On Parasitic absorption in FWCD Experiments in ITB Plasmas

T. Hellsten1, M. Laxåback1, T. Bergkvist1, T. Johnson1, F. Meo2, M. Mantsinen3, J.-M.

Noterdaeme4,5, C. C. Petty6, T. Tala7, P. Andrew8, V. Bobkov4, J. Brzozowski1, L.-G.

Eriksson9, E. Joffrin9, J. Mailloux8, M.-L. Mayoral8, A.A. Tuccillo10, and JET-EFDA

contributors.

1Alfvén Laboratory, Association VR-Euratom, Sweden, 2Association Euratom-Risø, National

Laboratory, Denmark, 3HUT, Association Euratom-Tekes, Finland, 4Max-Planck-Institut für

Plasmaphysik, EURATOM Association, Garching, Germany, 5EESA Department, University

Gent, Belgium, 6General Atomics, San Diego, USA, 7VTT Processes, Association Euratom-

Tekes, Finland, 8UKAEA Fusion Association, Culham Science Centre, U.K., 9Association

Euratom-CEA, CEA-Cadarache, France, 10 Associazione Euratom-ENEA, Frascati, Italy.

Fast wave current drive (FWCD) experiments have been performed in lower hybrid current drive preheated plasmas with internal transport barriers (ITBs). Although the current drive efficiency of the power absorbed on the electrons is fairly high it is difficult to affect the central plasma current. The main reasons are the parasitic absorption, the strongly inductive nature of the plasma current due to the high electric conductivity at the high electron temperature and the interplay between the fast wave driven current and the bootstrap current.

The latter is caused by the amount of the large bootstrap current in the barrier changes thereby counteracting the FWCD. The heating in the current drive experiments with +/-90° phasing at low single pass damping is found to be strongly degraded compared to that of dipole. A large fraction of the power is found to be lost as it is neither delivered to the divertor nor radiated into the main chamber. Pulses with strong degradation of heating and current drive also show strongly enhanced line radiation from BeII and CIV along sight lines passing through the divertor. P-5.096, Friday July 1, 2005

Current drive with oscillating magnetic fields in RFPs and FRCs

R. Farengo1, H. E. Ferrari1 and R. A. Clemente2 1 Centro Atómico Bariloche e Instituto Balseiro, Bariloche, Argentina 2 Instituto de Física Gleb Wataghin, Universidade Estadual de Campinas, Campinas, Brasil

The use of traveling (helical) and rotating (transverse) magnetic fields to drive continuous currents in RFPs and FRCs is studied. We present a complete numerical study on the use of oscillating helical fields (double helix) to drive current in RFPs. An infinitely long plasma column of radius a subject to external traveling magnetic fields varying like exp{i(+kz-t)} and a uniform steady axial magnetic

field B0 is considered. Helical coordinates x1=r/a, x2=+kz and x3:ignorable are employed. The ions are considered fixed, the electrons are described with a non linear Ohm’s law and density and temperature gradients are neglected. Fig. 1 shows the time evolution of the

normalized toroidal and poloidal currents (II/a0B) and external toroidal field (BB/B 2 2 for =45 (=B/en B: amplitude of helical field), =16 ( =a 0/2) and B0=1. Fig. 2 shows the averaged (over the azimuthal angle) magnetic field profiles of the steady-state for the same values of the parameters. Different ratios of toroidal to poloidal current can be obtained by changing h (h=ka).

4 1.0

2 B 0.5 ext 0 Fig.2. Radial Fig. 1. Time 0.0 -2 B profiles of the evolution of -0.5 -4 averaged I -6 the external -1.0

B (over ) -8 field and -1.5 magnetic field -10 currents B I -2.0 z -12 z -14 -2.5

-16 -3.0 0 100 200 300 400 500 600 0.0 0.2 0.4r/a 0.6 0.8 1.0 time

The use of two counter-rotating magnetic fields to drive current in FRCs was studied using a two fluids model that includes electron inertia. Analytical calculations including higher order harmonics (only the first harmonic was included in previous studies) showed that the torque applied on electrons and ions has an oscillatory component that is larger than the average value. Numerical calculations including the full time dependence and viscosity showed that although on the average (over one period) it is possible to deposit opposite torques on electrons and ions the large torque oscillations destroy the configuration. P-5.097, Friday July 1, 2005

Radial diffusion equation for rf-driven current density in tokamaks

M. Taguchi

College of Industrial Technology, Nihon University 2-11-1 Shin-ei, Narashino, Chiba 275-0005, Japan

Effects of fluctuations on the radial profile of a rf-driven current density is theoretically studied in tokamaks. Applying a renormalized perturbation theory to a drift kinetic equation in a presence of fluctuations, we can obtain a closed set of equations for an ensemble-averaged distribution function and a response function to an infinitesimal external perturbation [1]. In this paper, from this rather complicated closed set of equations, we derive a following more tractable radial diffusion equation for the rf-driven current density J:

1 ∂ ∂ rD hBJi−hBJi = −hBJ i, (1) r ∂r ∂r 0 h·i where J0 is the rf-driven current density in the absence of fluctuations and denotes a flux- surface average. The anomalous diffusion coefficient D due to fluctuations is expressed as

1 h dvS χ i D = − R rf 2 , (2) ν h dvS χ i R rf 1 ν χ where Srf represents the velocity-diffusion due to rf waves; is the collision frequency; 1 is χ the Spitzer function; and 2 is determined using the response function. Assuming specific forms for the correlation function of electric and magnetic fluctuations, we explicitly calculate the anomalous diffusion coefficient as a function of the amplitude and correlation lengths of fluctuations, and the collision frequency. Solving the diffusion equation with our obtained diffusion coefficient, we study the impact of the anomalous diffusion on the current profile.

References [1] M. Taguchi , Phys. Plasmas 10, 37 (2003). P-5.098, Friday July 1, 2005

Monte-Calro simulation of electron cyclotron cuurent drive in NTM magnetic islands

K.Hamamatsu, T.Takizuka, N.Hayashi, T.Ozeki

Naka Fusion Research Establishment, Japan Atomic Energy Research Institute, Naka-city, Ibaraki-ken, 311-0193, Japan

The control of neoclassical tearing modes (NTMs) is considered as one of the most important issues of burning plasmas such as the ITER. The electron cyclotron current drive (ECCD) attracts much attention as an effective method for stabilization of NTMs. Recently, by taking advantage of highly localized driven current due to ECCD, an experiment of a current drive inside a magnetic island in phase with the rotation of island and its effect is also discussed within the MHD theory. The current driven by EC wave is evaluated by a spacial profile of EC power deposition and deformation of velocity distribution. As a conventional numerical method, the EC power deposition is calculated by the ray tracing method and the velocity distribution is obtained by the bounce-averaged Fokker Planck (FP) equation. In the analysis of bounce-averaged FP equation, the magnetic field is assumed to be axisymmetric and the magnetic surfaces are also assumed to be nested. However, when the magnetic islands are formed by NTM, these assumptions become incomplete and the validity of ECCD analysis based on the bounce-averaged FP equation becomes questionable. In this paper, the electron guiding center motions are followed in full three dimensional magnetic field configuration with the Coulomb collision and quasilinear diffusion due to EC wave, which are simulated by the Monte-Carlo methods. As the first step of study, the momentum conservation in the Coulomb collision and the toroidal rotation of magnetic island are not yet included in our numerical model. We obtain the following result: When the EC wave is absorbed arround the O-point at a localized toroidal position, the driven current is diffused across helical magnetic surfaces but then never reach the separatrix with the result that the driven current is confined in the magnetic island, i.e., SNAKE-like driven current is formed. In the conference, a feature of ECCD in a magnetic island will be analyzed from the view point of electron motion. Furthermore, an averaged FP equation on flux surfaces in island will be formulated. P-5.099, Friday July 1, 2005

ΜοδελινγοφΠλασαΧυρρεντεχαψΠροχεσσδυρινγτηεισρυπτιον Η.Οηωακι1,2,Μ.Συγιηαρα3,Ψ.Καωανο1,ς.Ε.Λυκαση4,Ρ.Ρ.Κηαψρυτδινοϖ5, ς.Ζηογολεϖ4,Τ.Οζεκι1,Α.Ηαταψαα2 1ϑαπανΑτοιχΕνεργψΡεσεαρχηΙνστιτυτε,Νακα−σηι,Ιβαρακι−κεν,ϑαπαν311−0193 2ΚειοΥνιϖερσιτψ,Ηιψοσηι,Ψοκοηαα,ϑαπαν223−8522 3ΙΤΕΡΙντερνατιοναλΤεα,ΝακαϑΣ,Νακα−σηι,Ιβαρακι−κεν,ϑαπαν311−0193 4ΡΡΧΚυρχηατοϖΙνστιτυτεφορΑτοιχΕνεργψ,Μοσχοω,ΡυσσιανΦεδερατιον 5ΤΡΙΝΙΤΙ,Τροιτσκ,ΡυσσιανΦεδερατιον

Ελεχτροαγνετιχ(ΕΜ)λοαδβψτηεδισρυπτιονοντηειν−ϖεσσελχοπονεντσανδϖαχυυ ϖεσσελχανβεσιγνιφιχαντλψλαργεινΙΤΕΡ.ΑναχχυρατεεστιατιονοφτηεΕΜλοαδισοφ πριαρψιπορτανχεινΙΤΕΡδεσιγν.ΤηεΕΜλοαδισινδυχεδβψτηεραπιδπλασαχυρρεντ ρεδυχτιον(χυρρεντθυενχη)δυρινγτηεδισρυπτιον.Τηερεφορε,ιτισνεχεσσαρψτοδεϖελοπτηε νυεριχαλοδελωηιχηεστιατεστηεπλασαχυρρεντδεχαψτιε. ηεντηετηεραλθυενχηοχχυρσ,σπυττερεδορεϖαπορατεδιπυριτιεσφροτηεπλασα φαχινγχοπονεντσπενετρατειντοτηεπλασα.υρινγτηεχυρρεντθυενχηπηασε,τηειπυριτψ χοντεντισυσυαλλψϖερψηιγη,σοτηαττηειπυριτψραδιατιονισυχηλαργερτηαντηεορδιναρψ

τρανσπορτλοσσανδαυξιλιαρψηεατινγ.Τηισραδιατιονλοσσ, Πραδ,ιστηενβαλανχεδωιτητηεϕουλε

ινπυτποωερ, Πϕουλε,χονϖερτεδφροτηεαγνετιχενεργψοφτηεπλασαχυρρενταταχερταιν τεπερατυρε,

Πϕουλε ()Τε = Πραδ ()Τε ,(1)

ωηερε Τειστηεελεχτροντεπερατυρε.Τηεσετωοποωερσστρονγλψδεπενδονιπυριτψ χονχεντρατιονσ.Ιντηισοδελ,τηεφολλοωινγρατεεθυατιονινχλυδινγανιπυριτψτρανσπορτ εφφεχτισυσεδτοσολϖετηετιεβεηαϖιοροφιπυριτψδενσιτιεσπρεχισελψ,

δν ιον ρεχ ιον ρεχ ν κ =−σσϖνν() ϖ+σσ ϖ νν+− ϖνν κ ,(2) τ κ1− εκ1− κ κ εκ κ1+ εκ1+ τ δ ιπ

ιον ρεχ ωηερενκ,νε,<σϖ>κ ανδ<σϖ>κ αρεδενσιτιεσοφ κ−τηχηαργεστατειονσανδελεχτρονσ,ρατε

χοεφφιχιεντσοφιονιζατιονανδρεχοβινατιονοφτηε κ−τηχηαργεστατε,ρεσπεχτιϖελψ,ανδ τιπισ τηελιφετιεοφιπυριτιεσιντηεπλασα.Ιναδδιτιον,τηεδεταιλεδδισρυπτιονσιυλατιονχοδε ΙΝΑ[1]ωηιχηχαλχυλατεσανεϖολυτιονοφτηεπλασαΜΗεθυιλιβριυανδχυρρεντδεχαψ χονσιστεντλψ,ισσιυλτανεουσλψσολϖεδωιτηΕθσ.(1)ανδ(2).

Ιντηισνυεριχαλχοδε, νζ,τηειπυριτψχοντεντιντηεπλασα,ανδ τιπαρεδοιναντ αδϕυσταβλεπαραετερσ,ανδτηεψωιλλβεεξαινεδινδεταιλτορεπροδυχετηεεξπεριενταλ χυρρεντδεχαψτιεσανδωαϖεφορσινϑΤ−60Υ. [1]Ρ.Ρ.Κηαψρυτδινοϖ,ς.Ε.Λυκαση,ϑ.Χοπυτ.Πηψσ.,109,193(1993). P-5.100, Friday July 1, 2005

Axisymmetric MHD Simulation of Disruption Dynamics in High-β Reversed Shear Plasmas

H. Tsutsui1, N. Takei1, Y. Nakamura2, S. Tsuji-Iio1, R. Shimada1, S. C. Jardin3

1 Tokyo Institute of Technology, Tokyo, Japan 2 Japan Atomic Energy Research Institute, Ibaraki, Japan 3 Princeton Plasma Physics Laboratory, Princeton, New Jersey, USA

The advanced tokamak operation requires high-β, high-performance fusion plasmas with the internal transport barrier (ITB). In JT-60U reversed shear (RS) plasmas with a small bootstrap (BS) current fraction, a fast current quench of major disruptions was recently observed [1]. Hence, an understanding of disruption characteristics of high-β, RS plasmas is particularly important for designing high performance tokamak reactors, where the location of the BS current generated by ITB is closely linked with the magnetic shear profile. A dramatic change of the current profile from hollow to parabolic are accordingly caused by high-β RS plasma disruptions, and leads to a fast current quench. In the previous work [2, 3], disruptions with the ITB crash were investigated by an equilibrium model, in which its pressure/temperature profile was self consistently determined by its magnetic shear profile while its energy transport was not solved. Although this model repro- duced the fast current quench, it was clarified that this phenomenon strongly depends on the current diffusion process, because the first current quench phase is essentially determined by the change of a temperature/current distribution. This paper presents a detailed process of the ITB crash and following disruption dynamics of RS plasmas through an axisymmetric MHD simulation using the Tokamak Simulation Code [4] with transport and runaway-electron models. Detailed discussions about the governing mechanism for disruption behavior of strong ITB, RS plasmas will be given at the forthcoming EPS conference.

References [1] Y. Kawano et al., 30th EPS, P-2.129, St Petersburg, 2003. [2] N. Takei et al., J. Plasma Fusion Res. SERIES 6, 554–557 (2004) [3] N. Takei et al., 31th EPS, P2-141, London, 2004. [4] S.C. Jardin, N. Pomphery and J. Delucia, J. Comput. Phys. 66, 481 (1986) P-5.101, Friday July 1, 2005

Micro-Tearing Modes in MAST

D. Applegate1, S. C. Cowley1, W. Dorland3, N. Joiner1, C. M. Roach2

1 Imperial College, London, England

2 Culham Science Centre, Abingdon, England

3 University of Maryland, College Park, USA

Recent linear gyrokinetic calculations carried out for a wide range of spherical tokamaks suggest that these devices are susceptible to magnetic reconnection at ion gyro-radius scale lengths. These micro-tearing instabilities have been seen in studies of equilibria from MAST at Culham, NSTX at Princeton, and also the conceptual power plant. It has been reported that heat transport occurs predominantly through the electron channel in NSTX, and it is important to assess whether micro-tearing instabilities might provide a possible explanation for this observation. Here we present GS2 results based on a MAST, ELMy H-mode equilibrium, which have revealed basic characteristics of the tearing instability. This work shows the instability is mainly due to electron physics, despite occurring in the waveband usually associated with ion- driven instabilities. Indeed the ion response to these modes can reasonably be assumed to be Boltzmann and the growth rate is shown to be especially sensitive to electron collisions, electron temperature gradient, and finite β, although other parameters including the density gradient and magnetic shear also play an important role. Poincaré plots have been derived from the perturbed vector potential associated with the instability. These demonstrate the tearing nature of the mode, showing clear island structures at moderate amplitudes and stochastic fields at higher amplitudes. Non-linear studies are underway to investigate saturation mechanisms and to assess the importance of these modes for transport. Initial mixing length estimates suggest these microtearing modes could provide substantial heat flux. However, the radial wave-number of the mode increases along the field line, and since the mode can extend far along the field line it is sensitive to dissipation effects which could in turn lead to early saturation of the mode. Non- linear GS2 calculations are in progress and will hopefully shed light on this issue.

This work is funded jointly by the UK Engineering and Physical Sciences Research Council and EURATOM. P-5.102, Friday July 1, 2005

Analytic approximations of divertor behaviour and application to MAST

G P Maddison, A Turner

EURATOM/UKAEA Fusion Association, Culham, Abingdon, Oxon. OX14 3DB, UK.

Fuelling of tokamak plasmas always consists partly of impinging neutral-particle fluxes from their surrounding vessel space, fed by recycling from boundary surfaces and perhaps directly from gas puffing. In its simplest form, particle balance can therefore be described by global rate equations for the numbers of ions in the plasma, atoms and molecules in the surrounding volume, and equivalent atoms adsorbed by walls. Recycling coefficients, active pumps and fuel inputs determine net sinks and sources. This representation can then be solved exactly for steady states, showing the relative effects of fuelling efficiency eΦ , as well as wall or active

pumps, on vessel gas density nD2 . For eΦ = 0 , equivalent to puffing gas just into the vessel

space itself, nD2 / ni (where ni is plasma ion density) becomes independent of any sinks.

Stronger cryopumping from tokamaks is promoted by raising gas compression in front of the duct, usually by enclosing recycling regions in a baffled divertor. This general situation can in turn be described by separate sets of global rate equations for the divertor and main-vessel (“tank”) zones, connected by parameterized conductances (ie inverses of closure) for atoms and molecules. In a slightly reduced form including only ions, molecules and active divertor pumping, exact steady-state solutions can again be derived. If fuelling can achieve perfect t efficiency at the plasma edge eΦ = 1 , normalized gas density in the tank nD2 / ni is the same wherever such a source is added. For more realistic eΦ < 1 , however, behaviour differs significantly when input is within the divertor or the tank. In the latter case, there is a critical t t value of divertor conductance dependent upon eΦ below which nD2 / ni rises asymptotically to the previous open-vessel result in sequential steady states with progressively higher d t divertor cryopumping. In contrast, for puffing in the divertor with similar eΦ = 0 , nD2 / ni again becomes independent of cryopumping altogether. Hence divertor compression ratio d t nD2 / nD2 is in fact entirely independent of cryopumping for divertor puffing with arbitrary d t efficiency eΦ , but falls as pumping increases for tank fuelling unless eΦ = 1 .

In the MAST spherical tokamak, like its predecessor START, plasmas are formed inside a large vacuum vessel with a ratio of respective volumes typically ≈ 1 : 10 , so that fuelling from the gas reservoir and pumping by a large wall area are particularly accentuated. Considering more efficient fuelling at the plasma edge eΦ > 0 , the walls, when boronized, are already very t t effective in reducing nD2 / ni . Similarly, an in-vessel divertor could lower nD2 / ni significantly through closure, even before adding a cryopump. Numerical solutions of the full global model with atoms, molecules, plus both wall and active pumping, show in addition that rapid decrease of plasma density after switching off gas input, necessary for feedback control, would be possible with a moderately closed, cryopumped divertor. Unlike wall sinks, this arrangement also would not degrade through saturation in long pulses.

Work funded jointly by the United Kingdom Engineering & Physical Sciences Research Council and by EURATOM. P-5.103, Friday July 1, 2005

Parametric decay instability accompanying electron Bernstein wave heating in MAST

A. Surkov1, G. Cunningham2, A. Gurchenko1, E. Gusakov1, V. Shevchenko2, F. Volpe2 1 Ioffe Institute, St.-Petersburg, Russia 2 EURATOM/UKAEA Fusion Association, Culham Science Center, Abingdon, OX14 3DB, UK

Electron cyclotron heating was shown to be an effective tool for plasma heating in stellarator and tokamak plasmas. Unfortunately its application to heating in spherical tokamaks (STs) is substantially limited by the high plasma density and relatively low magnetic field typical for these devices. This ST feature has a strong effect on the electromagnetic wave propagation. In the microwave frequency region, characteristic surfaces, like the upper hybrid resonance (UHR) and cut-offs are very close to the plasma edge. As a result, the electromagnetic (EM) waves are unable to penetrate into the plasma interior. The only way to overcome this difficulty is to use the linear conversion of the incident EM wave into the electron Bernstein wave (EBW) at the UHR. This has no density limit and can, in principle, carry the radio frequency power deep into the plasma. The feasibility of this plasma heating scheme is under investigation now on the MAST tokamak at Culham, UK, where up to 1 MW microwave power at 60 GHz is launched into the H-mode plasma in 7 beams and the coupling to EBW is optimized by a system of steerable focusing mirrors. Nonlinear effects can influence the wave propagation close to the UHR, where the wave electric field increases, in particular by parametric decay instabilities, which cause anomalous reflection and absorption of the incident power. In the present paper the first observations of lower hybrid (LH) waves generated by parametric decay in EBW heating experiments on the MAST tokamak are reported. At a certain orientation of the microwave antenna the steep growth of 100 MHz RF radiation during the heating pulse was registered by the specially designed RF antenna situated in the edge plasma at 5 cm outside the separatrix. Due to the fact that the observed RF signals have frequencies close to the LH resonance frequency in the UHR region we suppose the radiation is generated by the parametric decay instability occurring in the UHR. The theoretical study of UH wave decay into another UH wave and LH wave has been performed. The threshold power is estimated for typical parameters in MAST. It is shown that the backscattering parametric decay instability can arise only if the pump power exceeds 60 kW in any beam, so there is an indication of substantial coupling of the microwave power to the UHR in the MAST experiment. The work was partly funded by the UK Engineering and Physical Sciences Research Council and by the EURATOM. P-5.104, Friday July 1, 2005

Specific features of ICR heating on the spherical tokamak Globus-M at high concentration of light ion component V.V.Dyachenko1, B.B.Ayushin1, F.V.Chernyshev1, V.K.Gusev1, Yu.V.Petrov1, N.V.Sakharov1, O.N.Shcherbinin1, S.Yu.Tolstyakov1, V.M.Leonov2 1 A.F.Ioffe Physico-Technical Institute, RAS, 194021, St.-Petersburg, Russia 2 Nuclear Fusion Institute, RRC “Kurchatov Institute”, Moscow, Russia

In spherical tokamaks the conventional ICR plasma heating has a number of specific features and therefore requires additional investigations. The first and the most important point consists in simultaneous existence of several IC harmonics in the plasma cross-section, in which RF power absorption is possible. The other feature is intrinsic property of the Globus- M tokamak: the amount of hydrogen in a deuterium plasma can vary in the range 10-50% in different experimental conditions. For this reason, the efficiency of wave energy absorption in the vicinity of some cyclotron harmonics can change also. The ICRH experiments are performed on the low aspect ratio tokamak Globus-M (R=0.36 m, a=0.24 m, B0=0.3-0.4 T,

Ip=0.15-0.25 MA, vertical elongation 1.2-2) at RF power input level up to 200 kW at frequencies 7.5 and 9.1 MHz /1/. The ion temperature behaviour was observed by a 12- channel neutral particle analyzer ACORD-12, which measured simultaneously hydrogen and deuterium fluxes and relative concentration of ion components. The 1-D modeling /2/ of wave propagation foretells effective RF power absorption both by ions and electrons in broad range of relative hydrogen concentration. All possible mechanisms of RF absorption (cyclotron, TTMP, Landau) were taken into account. In the experiment the ion temperature increases by 2 times, but the ion heating efficiency depends from the location of the cyclotron harmonics and on the concentration of the light ion impurity (hydrogen). The estimates show that the RF power absorbed by ions is comparable to the power of ion ohmic heating. The predicted electron heating was not detected in the experiment, probably because the RF power absorbed by electrons (about 100 kW) is much less than the ohmic power. The transport code ASTRA modeling carried out in the assumption of the neoclassical transport coefficients confirms the ion heating efficiency observed in the experiment. 1. V.K.Gusev, V.V.Dyachenko, F.V.Chernyshev et al, Tech.Phys.Lett.,30(8),2004,690-692. 2. M.A.Irzak, E.N.Tregubova, O.N.Shcherbinin, Plasma Physics Reports, 25(8), (1999), 659- 667.

P-5.105, Friday July 1, 2005

Effect of electron collisions on nonlocal transport coefficients in laser heated plasmas

A. Bendib, A. Tahraoui and K. Bendib-Kalache Laboratoire d’Electronique Quantique, Faculté de Physique, USTHB, El Alia BP 32, Bab Ezzouar 16111, Algiers, Algeria

A complete set of nonlocal transport coefficients in laser heated plasmas is derived solving numerically the electron Fokker-Planck equation, using a perturbation method,

parametrized as a function of the electron mean-free-path nei compared to the spatial scales L. The electron-electron Landau collision terms are kept in the analysis up to the tenth anisotropy, and therefore the results are valid for arbitrary atomic number Z. Practical numerical fits of the transport coefficients are presented as functions of Z and

the collisionality parameter nei / L . An application to parametric instabilities in laser- created plasmas is also proposed.

P-5.106, Friday July 1, 2005

Relaxation of the localized temperature perturbation in the collisional plasma A. V. Brantov1,2, V. Yu. Bychenkov2, W. Rozmus1 1 Physics Department University of Alberta, Edmonton, Canada 2 P.N. Lebedev Physics Institute, Moscow, Russia

Localized temperature perturbations in a speckled laser beam are fundamental features of laser heated plasmas. Evolution of these hot spots depends on plasma collisionality, spatial scale length of the perturbation, its amplitude and the magnetic field effects. Typical transverse dimensions of the heated region are on the order of the electron mean free path resulting in the nonlocal thermal conductivity. Also weak plasma collisionality means that the initial rapid relaxation of the hot spot is essentially collisionless before taking the form of the nonlocal hydrodynamical evolution. Such a change in the heat conduction leads to the time nonlocality of electron thermal conductivity [1]. We report on results of theoretical and numerical modelling of the hot spot evolution. Nonlocal and nonstationary closure relation for thermal flux has been derived by solving linearized electron kinetic equation with exact collision operators in the Landau form. Extension of the linear transport relations into nonlinear regime is proposed and tested by comparing them with Fokker-Planck simulations. An initial value problem and the laser absorption and plasma heating case have been considered. We have examined effects of external magnetic field [2]. In particular we have considered the heat wave propagation across the magnetic field as in the proposed magnetized plasma extension of the gas jet laser plasma heating experiment [3]. This work was partly supported by the Natural Sciences and Engineering Research Council of Canada and the Russian Foundation for Basic Research (Grant No. 03-02-16428).

[1] A. V. Brantov, et al., Phys. Rev. Lett., 93, 125002 (2004). [2] A. V. Brantov, et al., Phys. Plasmas 10, 4633 (2003). [3] G. Gregori, et al, Phys. Rev. Lett., 92, 205006 (2004).

P-5.107, Friday July 1, 2005

Upgraded version of the ICF simulation code MULTI

J. Ramírez and J. Ramis

E. T. S. I. Aeronáuticos, Universidad Politécnica de Madrid

The radiation-hydrodynamics code MULTI [1, 2, 3] has been widely used by the laser plasma community since the eighties (numerical results appear in around 100 scientific papers). Now, a new upgraded version has just been developed. The 1D and 2D versions have been integrated (together with equation of state and opacities generation codes MPQEOS [4] and SNOP [5]) in a unique package. Output file format, graphic interface, table protocol access, etc, have been unified. The code has been structured in modules and layers. The top level code is written in a weakly- typed language (called r94), and a preprocessor is used to translate it into standard C. This allows a maximum simplicity in the interfaces between modules. Current capacities include: 1D planar, cylindrical or spherical geometries, 2D axially symmetric configuration with unstructured grid; Lagrangian, Eulerian or mixed hydrodynamics; radiation transport with angular resolution (in 2D) and frequency resolution (in 1D); DT ignition physics including α-particle diffusion (in 1D) and transport (in 2D), and laser and ion-beam ray-tracing packages. The code has been partially adapted to parallel MPI-computing.

References [1] http://server.faia.upm.es/multi [2] RAMIS, R., MEYER-TER-VEHN, J., and SCHMALTZ, R., Comp. Phys. Comm. 49 (1988) 475. [3] RAMIS, R. and RAMÍREZ, J., Nucl. Fusion 44 (2004) 720-730 [4] KEMP, A. J., Das Zustandsgleichung-Modell QEOS für heisse, dichte Materie Report MPQ229, Max-Planck-Institut für Quantenoptik, Garching, Germany (1998) [5] EIDMANN, K., Laser Particle Beams 12 (1994) 223. P-5.108, Friday July 1, 2005

Magnetic Field Generation in Pasmas due to Anisotropic Laser Heating

A. Sangam 1,2 , B. Dubroca 1, P. Charrier 1, V. T. Tikhonchuk 1, J.- P. Morreeuw 3

1 Centre Lasers Intenses et Applications, UMR 5107 CEA, CNRS, Université Bordeaux 1, 351 Cours de la Libération, 33405 Talence Cedex, France 2 Laboratoire MAB, Université Bordeaux 1, 351, Cours de la Libération, 33405 Talence Cedex, France 3 CEA- CESTA, B.P. 2, 33114 Le Barp, France

Generation of magnetic fields around the laser speckles in underdense plasma due to an anisotropic laser heating is shown to be an important effect under the conditions previewed for the inertial confinement fusion. Due to the preferential laser heating of electrons in the direction of the laser electric field, the electron distribution acquires an anisotropy, which can be characterized by the second angular harmonic of the distribution function and/or by the pressure tensor. The source of magnetic field generation is proportional to gradients of the laser intensity and the pressure tensor. The growing magnetic field induces the Weibel- like instabilities, which cannot be stabilized in the frame of conventional hydrodynamics. We developed a new equation for the off- diagonal components of the pressure tensor. It accounts for the spatial diffusion of the tensorial part of the heat flux and allows one to stabilize the short- wavelength Weibel instability. This model is compared to kinetic simulations of the semi- collisional Weibel instability. This work is based on both theoretical analysis and numerical computations. The equations of flow are solved by using a fully nonlinear implicit in time hydrodynamics coupled with the equation for the magnetic field. P-5.109, Friday July 1, 2005

OSIRIS 2.0: an integrated framework for parallel PIC simulations

R. A. Fonseca1,2, M. Marti1, S. F. Martins1, L. O. Silva1, F. Tsung3, W. B. Mori3

1 GoLP/CFP, Instituto Superior Técnico, Portugal 2 DCTI, Instituto Superior Ciências do Trabalho e da Empresa, Portugal 3 University of California Los Angeles, CA 90095, U.S.A.

We describe OSIRIS 2.0 framework, an integrated framework for particle-in-cell (PIC) simulations. This framework is based on a three-dimensional, fully relativistic, massively parallel, object oriented particle-in-cell code, that has successfully been applied to a number of problems, ranging from laser-plasma interaction and inertial fusion to plasma shell collisions in astrophysical scenarios. The OSIRIS 2.0 framework is the new version of the OSIRIS code [1]. Developed in Fortran 95, the code runs on multiple platforms and can be easily ported to new ones. Details on the capabilities of the framework are given, focusing on the new capabilities introduced, such as bessel beams, binary collisions, tunnel (ADK) and impact ionization, and new diagnostics, and also static load balancing and parallel I/O. This framework also includes a visualization and data-analysis infrastructure [2], tightly integrated into the framework, developed to post-process the scalar and vector results from our simulations.

References [1] R. A. Fonseca et al., LNCS 2331, 342-351, (Springer, Heidelberg, 2002). [2] R. A. Fonseca, Proceedings of ISSS-7, (to be published, 2005). P-5.110, Friday July 1, 2005

Numerical modelling of metallic plasma produced by UV laser beam ablation

F.Hamadi, E.H.Amara, D.Doumaz

Advanced Technologies Development Centre (CDTA) Laser Material Processing Group Po. Box 17, Baba-Hassen Algiers, Algeria

Abstract

In the present paper, we study a numerical modelling of the plasma or vapour expansion into a target chamber, produced by a nanosecond UV laser beam on a metallic sample surface. The vaporised material (plasma) ejected is composed of neutrals, metallic ion species and electrons, which are modelled as an ideal gases. The velocity, temperature, and density distribution of the plasma are calculated using the computational fluids dynamics software “fluent”, based on a finite volume discretization. The dynamical state of the plasma species are described by the compressible Navier-Stokes conservation equation for mass, momentum and energy.

P-5.111, Friday July 1, 2005

Hyperthermal neutral beam for the material processing Bong Ju Lee a, DC Kim a, T. Lho a, SJ Yoo a, M. Joungb a Korea Basic Science Institute, Daejeon, Korea 305-333 b Dept. of Physics, POSTECH, Pohang, Korea 790-784

Hyperthermal neutrals with energy level between 2eV and 100eV from the plasma can be utilized for a diagnostic tool for the boundary region of fusion plasmas, for the space environment on earth orbiting spacecraft, and for a surface modification and material’s processing. Hyperthermal neutrals in our device is generated through the Auger neutralization (surface neutralization or reflection) of the accelerated ions, which processes the capture of electron within a few atomic layer from the surface of conducting material, namely the reflector, which is biased negatively. For high flux and directionality of Hyperthermal neutral beam (HNB), the operating parameters of Inductively Coupled Plasma (ICP) source, such as geometries of a chamber, reflector and collimator and operating pressure have been optimized. The limiter to assure the plasma-free and radiation-free conditions at the processing chamber is required also. The extraction of HNB from the plasma was demonstrated by the a-Carbon etching experiment using oxygen plasma and implanting hyperthermal neutrals into the Si wafer covered with SiO2 thin film with Nitrogen plasma. The estimated HNB flux via the Photoresistor (PR) etching experiment was above 1016 cm-2 s-1, which is an enough quantity for the real applications, the ion current density and electron current density at the target were none and below – 1x10-4 mA/cm2, respectively and no radiation at the processing chamber exists because the line of sight is prevented from the plasma. Another feature of our device is to process an unlimited target size and currently 8 inch wafer can be processed. Energy of HNB was measured by Neutral Particle Analyzer (NPA) and is proportional to the ion energy, which can be given by the difference between the plasma potential and reflector bias voltage. NPA measurement showed that the energy of HNB was about 60% of the incident ion energy below 150eV and could be controlled by the negative bias voltage applied to the reflector.

P-5.112, Friday July 1, 2005

Magnetic field generation in the Weibel instability

1 1 1 2 3 M. Fiore , R. A. Fonseca , L. O. Silva , W. B. Mori , M. Medvedev 1 GoLP/CFP, Instituto Superior Técnico, Lisbon, Portugal 2 University of California Los Angeles, CA 90095, USA 3 University of Kansas, KS 66045, USA

In recent years, the Weibel instability [1] has received an increasing interest due to its potentially important role in Astrophysics [2] and in other fields of Plasma Physics, such as Fast Ignition [3]. The magnetic field generation in gamma-ray bursts represents one of the unsolved problems of the astrophysics of extreme events. Synchrotron radiation measurements in GRBs indicate the presence of near-equipartition magnetic fields, whose generation can be explained by the onset of the Weibel instability. In order to understand the saturated level of the magnetic field, it is crucial to determine

the growth rate (!) of the Weibel instability. In fact, the saturation level Bsat of the

2 magnetic field is related to ! according to Bsat ∀ ! [4]. Using relativistic kinetic theory [3], we present a theoretical model to study the linear stage of the Weibel instability including hot plasma, space charge effects and different mixtures of plasma components. Due to its versatility, the model allows for calculation of the growth rate of the Weibel instability for a wide range of scenarios, thus leading to the maximum value of B. The long-time evolution of the magnetic field after saturation is examined with two- dimensional and three-dimensional fully relativistic particle-in-cell simulations, carried out with OSIRIS 2.0 [5], and compared with a theoretical two-dimensional (toy) model [6]. We find a very good agreement with a correlation length for the B field evolving as

∃ #B ∀ t with ∃ ~ 1. [1] E. S. Weibel, Phys. Rev. Lett. 2, 83 (1959) [2] M. Medvedev & A. Loeb, ApJ 526, 697 (1999); L. O. Silva et al., ApJLett. 596, L121 (2003) [3] L. O. Silva et al., Phys. Plasmas 9, 2458 (2002) [4] T-Y. B. Yang, J. Arons & A. Langdon, Phys. Plasmas 1, 3059 (1994) [5] R. A. Fonseca et al., LNCS 2331, 342 (2002) [6] M. Medvedev, M. Fiore, R.A. Fonseca et al., ApJLett. 618, L75 (2005) P-5.113, Friday July 1, 2005

Evolution of Signals Scattered by Modified Ionosphere A.V. Kochetov1, G.I. Terina2 1Institute of Applied Physics of RAS, Nizhny Novgorod, Russian Federation 2Radiophysical Research Institute, Nizhny Novgorod, Russian Federation

Presented work is the continuation of experimental and theoretical study of nonlinear dynamic processes arising in ionospheric plasma under the action of powerful radio waves The amplitude and phase characteristics of the “caviton” signal (CS) and the main signal (MS) of probing transmitter were investigated. The temporal evolution of CS and MS phase was revealed. The experimentsl were carried out in Nizhegorodsky region of Russia at the heating facility “Zimenki” by means of artificially disturbed ionosphere sounding by probing radio pulses (G.I. Terina, J. Atm. Terr. Phys., 1995, v.57, pp. 273-278). The heating transmitter radiated radio waves of ordinary polarization and was switched on periodically for 20s and switched off for the same duration. The probing transmitter radiated pulses of 50µs duration with the ordinary polarization too. When the heating frequency was greater than probing one at the initial heater stage the fast decrease of CS phase occurred for hundreds milliseconds. It was replaced then by more slow increase of phase during the heater time. The similar phase increase of main signal of probing transmitter was observed after the heater turning on, which was replaced by its decrease after heater turning off. The observed evolution of time phase dependencies of CS and MS showed the dynamics of plasma density structures and the displacement of reflection region of probing radio wave during the heating time. The theoretical simulations were carried out in the frameworks of driven nonlinear Schrödinger equation modified for inhomogeneous plasma (A. V. Kochetov, et. al., Advances in Space Res., 2002, vol. 29, pp. 1369-1373). The phase characteristics of the backscattered heating and probing radio waves were calculated for different frequency deviation between heating and probing waves. The spatial and temporal plasma density structures dynamics was considered. Qualitative conformity of theoretical and experimental results showed the dynamics of strong turbulence structures excitation and relaxation in the reflection region of powerful radio waves and the chances of its diagnostics. This work was supported in part by the Russian Foundation for Basic Research (grants 03-02-16309, 05-02-17370). P-5.114, Friday July 1, 2005

COMPARISON OF VARIOUS MECHANISMS OF ELECTRON HEATING AT THE IRRADIATION OF DENSE TARGETS BY A SUPER-INTENSE FEMTOSECOND LASER PULSE

Krainov V.P. Moscow Institute of Physics and Technology, 141700 Dolgoprudny, Moscow Region, Russia Various mechanisms of electron heating are considered at the irradiation of solid targets and dense atomic clusters by a super-intense femtosecond laser pulse. Free electron cannot absorb photons of the monochromatic laser field. A strong breaking of the electron motion with respect to the laser field is required in order to absorb laser photons. Such a breaking can be produced by various reasons. The dominating mechanism depends significantly on the parameters of the laser pulse (laser intensity and frequency, pulse duration, contrast) and on the properties of the solid target. We discuss the next mechanisms of the electron heating by a femtosecond pulse: 1. Induced inverse bremsstrahlung at the elastic scattering of electrons on atomic ions of the target after the field ionization. This mechanism is important in the underdense region of plasma cloud at the relatively low laser intensities, of the order of 1015 W/cm2 and lower, when the rate of electron – ion collisions is sufficiently high. 2. Vacuum (Brunel) heating, when an electron is ejected from the dense plasma by the oblique p-polarized laser field and returns then back approximately in a half of the laser period acquiring the energy of the order of its oscillation energy in the laser field. This regime is realized for lasers with high contrast and without any pre-pulse. 3. Electron heating at their elastic reflections from the surface of the ionized cluster. The double oscillation energy is absorbed by an electron at each collision. 4. Electron heating by the field of the longitudinal plasma (wake) wave. 5. Stochastic (diffusion) heating of electrons ejected from overdense plasma in the summary field of the incident and reflected laser waves because of dynamic chaos. 6. Electron heating at their scattering on the plasma density, or potential gradient. 7. Ponderomotive heating along the polarization of the laser field after ionization. 8. Relativistic ponderomotive heating along the propagation of the laser pulse, produced by the magnetic part of the Lorentz force and by charge separation. This work was supported by ISTC (project N 2917) and RFBR (project N 04-02- 16499). P-5.115, Friday July 1, 2005

Electron holes and Buneman instabilities in nonuniform current-carrying plasma equilibria* M. V. Goldman1, D.L. Newman1, A. Mangeney2 and F. Califano3 1 Physics Dept. and CIPS, University of Colorado, Boulder, CO, USA 2 DESPA, Observatoire de Paris à Meudon, France 3 University of Pisa, Pisa, Italy

Analytical theory, numerical analysis and Vlasov simulations are employed to study linear eigenmodes of Buneman instabilities and their nonlinear evolution into slow electron holes in nonuniform current-carrying plasma equilbria. Two classes of equilibria are considered. The first is 2-D, with strongly magnetized electrons and shear in the parallel electron current (sinusoidal variation of parallel velocity in the direction perpendicular to B). Magnetic reconnection simulationsi have revealed electric field structures explained in terms of unsheared Buneman instabilities. The unstable waves trap electrons and form electron phase space holes which have been identified as a source for the small-scale dissipation needed during reconnection.i The present sheared studies should apply to the edges of current sheets formed during reconnection and also to current-reversal regions of the auroral ionosphere. It is found that even a small amount of shear can lead to very different nonlinear hole evolution even though the growth rates of the linear eigenmodes are only slightly changed Shear in electron velocity and compensating shear in the density to keep the electron flux and current constant is also studied. The second class of nonuniformities studied are 1-D stationary nonlinear cold fluid equilibria with constant fluxes of counterstreaming electrons and ions. The stability of such equilibria is relevant to Buneman instabilities embedded in double layers.ii The growth of large electron holes (spinners) localized in the potential ramp of a double layer has been observed to accompany the breakup of laminar double layers in 1-D simulations. i Drake, J. F., M. Swisdak, C. Cattell, M. A. Shay, B. N. Rogers, and A. Zeiler, Formation of Electron Holes and Particle Energization During Magnetic Reconnection, Science, 299, (2003) ii Goldman, M.V., D.L. Newman and R.E. Ergun, Phase-space holes due to electron and ion beams accelerated by a current-driven potential ramp, Nonlinear Processes in Geophysics, 10, 37-44, (2003) *Work supported by DOE, NSF and NASA P-5.116, Friday July 1, 2005

Plasma discharge in supercritical fluids Satoshi Hamaguchi1, and Suresh C. Sharma1,* 1 STAMIC, Graduate School of Engineering, Osaka University, Osaka, Japan

Parallel plate discharges in CO2 fluids near supercritical conditions are known to exhibit a peculiar Paschen's law [1,2]: the breakdown voltage becomes significantly reduced at the critical pressure from the corresponding value for the ideal gas. In the present work a theory to account for this phenomenon is proposed, based on kinetic theory of discharge breakdown in supercritical fluids[3]. According to this theory, the reduction of

breakdown voltage at the critical point should not be limited to supercritical CO2 but it is a generic property for all supercritical fluids. In Fig. 1, the Paschen's law for a supercritical fluid obtained from our theoretical model is given by the magenta curve. The central issue regarding discharge of a fluid at its critical point is the formation of micro clusters of fluid molecules. The formation of such clusters creates micro cavities (i.e. electron channels), where electrons gain higher kinetic Fig. 1: Normalized breakdown voltage V as a energies from the applied function of the normalized pressure-distance product. The blue and magenta curves represent the electric field than those in a cases for the ideal gas and the fluid near supercritical conditions. At the critical pressure liquid under similar conditions. ( Pd 0 45 in this case), the breakdown voltage is This effectively increases the shown to be lowered. ionization rate and therefore reduces the breakdown voltage. *Permanent address: Physics Department, GPMCE (G.G.S. Indraprastha University, Delhi), India. References [1] Ito and Terashima, Appl. Phys. Lett. 80, 2854 (2002). [2] Ito, Fujiwara, and Terashima, J. Appl. Phys. 94, 5411 (2003). [3] S. Hamaguchi and S.C. Sharma, in preparation. P-5.117, Friday July 1, 2005

Stimulated emission without inversion in ensembles of classical electrons

M. A. Erukhimova, M. D. Tokman

Institute of Applied Physics RAS, Nizhny Novgorod, Russia

The stimulated emission (lasing) without inversion (LWI) as one of the basic effects in the physics of radiation processes has been intensively studied for the last several years. This effect was predicted initially for quantum systems [1]. But like the usual stimulated emission, the LWI has been found as a general physical effect. The classical LWI models include both monochromatic systems with spatial division of electron beam into two fractions [2] and systems where spatially homogeneous ensemble of electrons interacts with the bichromatic radiation [3]. In the systems of the second type, discussed here the parametric coupling of radiation components by means of medium oscillations underlies the LWI effect. From this viewpoint such effect is related to another coherent parametric effect: Electromagnetically In- duced Transparency, which is widely discussed both in quantum electronics and plasmas ([4]). Here we present stationary nonlinear regimes of three-waves interaction resulting in "in- versionless" cyclotron amplification of two HF waves due to coupling by means of permanent LF pumping. In the absence of LF pumping the linear regime of interaction between HF waves with the medium characterized either by the field absorption (due to resonant cyclotron interaction with the "inverted" distribution function) or by no energy exchange between the medium and the field (due to the absence of resonant particles). As result of these investigations further analogies between classical and quantum electronics, where large number of AWI schemes with permanent LF pumping is known, are revealed. Besides, a thorough research of use environment of the Manley-Rowe relations, which define multi-photon process behaviour, is presented.

References [1] O. Kocharovskaya, Phys.Rep, 219, 175 (1992) [2] Y. Rostovtsev, S. Trendaﬁlov et al., Phys.Rev.Lett, 90, 214802 (2003) [3] A. V. Gaponov-Grekhov, M. D. Tokman, JETP, 85, 640 (1997), M. A. Erukhimova, M. D. Tokman, Radiophysics and quantum electronics, 45, 249 (2003), M. A. Erukhimova, M. D. Tokman, Proc. of 31th EPS Conference on Plasma Phys., London, 28June-2 July 2004 ECA, 28G, P-4.211 (2004). [4] A. G. Litvak, M. D. Tokman, Phys.Rev.Lett, 88, 095003 (2002), G. Shvets, J. S. Wurtele, Phys.Rev.Lett, 89, 115003 (2002) P-5.118, Friday July 1, 2005

Electron Temperature Control in a Hot Cathode Arc Discharge Plasma N. Ezumi Dept. of Electronics and Control Engineering, Nagano National College of Technology, 716 Tokuma, Nagano 381-8550, Japan

Control of electron temperature Te has lots of interest in advanced plasma applications as well as basic plasma researches. So far, some methods for the temperature control were applied such as pulse modulation, biased grid, RF heating and so on. In this

study, we have attempted that the Te control of a hot cathode arc discharge plasma by means of varying a quantity of thermal electron emission and a distance of the discharge gap, because the thermal electron emission and the gap distance are essential to sustain arc discharge plasmas. Experiments were performed by using the Compact Arc Plasma Generator with a

LaB6 hot cathode and a variable gap driving system. The device has a glass vacuum vessel with 0.8 m in length and 0.14 m in diameter, and is equipped with two solenoidal magnetic coils, which produce a flux density of 10 mT in steady state. Argon plasma was produced in steady state by sophisticated dc discharge system, which is designed to improve the ionization efficiency using a divergent magnetic configuration and the direct heating hot

17 cathode. Typical the electron density ne and Te in the plasma test region are about 5 x10 m-3 and less than 10 eV, respectively. The plasma diameter is about 0.01 m. Neutral

pressure measured in the plasma test region Pn was less than 2 mTorr in this study. So far, decreasing the heating current of the cathode, probe measurements showed

decrease of Te of high-energy tail component for Argon plasmas. Simultaneously, a change of cathode heating mode was observed. Moreover, strong dependence of discharge voltage

on gap distance was also observed. These results indicate a possibility of the control of Te (or energy distribution function) by using these methods. P-5.119, Friday July 1, 2005

Wake potential of a test charge using the stationary phase method

Michael A. Raadu and Muhammad Shafiq

Alfvén Laboratory, Royal Institute of Technology, Division of Plasma Physics, SE-10044, Stockholm, Sweden

The linear response of a dusty (complex) plasma to a moving test charge can be determined using an appropriate plasma dielectric function and a three dimensional Fourier analysis. Many analytical results have been found for a slowly moving test charge. For intermediate and large velocities numerical methods of integration are commonly used. However general asymptotic results valid at large distances can be found, using a combination of the residue calculus applied to the zeroes of the dielectric function and the method of stationary phase for integration over a wave vector component. The method can be expressed in terms of conditions on the group and phase velocity of waves in the reference frame of the moving test charge. In particular for a given radial direction the asymptotic response is determined by wave vectors for which the group velocity is directed radially outwards. The analysis is close to that due to Kelvin for ship waves in deep water. An essential difference is that in the present case three spatial dimensions are involved instead of only two, and also that the dispersion relation for plasma waves is more complicated. P-5.120, Friday July 1, 2005

Supersonic Rotation Exceeding the Alfven Ionization Limit in the Maryland Centrifugal Experiment R. Ellis,S. Messer, A. Case, R. Elton, J. Ghosh, H. Griem, A. Hassam, R. Lunsford, C. Teodorescu University of Maryland, College Park, MD, USA

The Maryland Centrifugal Experiment(MCX) has achieved supersonic rotation produced by the application of a steady state electric field perpendicular to a linear confining magnetic field and most recently MCX has demonstrated an enhanced mode of operation with high rotation velocities which exceed the Alfven ionization limit ( ion directed kinetic energy equal to neutral ionization potential) , which is approximately 50 km/s for hydrogen. MCX has a magnetic mirror geometry of 2.6m length, variable mirror ratio (2-20), maximum mirror field of 1.9T, maximum midplane field of 0.5T. The MCX plasma is terminated axially by insulating discs located near the mirror maxima. Biasing of an inner coaxial core relative to the outer wall produces a radial electric field which drives the azimuthal rotation. A 10 KV capacitor bank is the power source for the discharge. Azimuthal velocity is measured spectroscopically employing doppler shift of impurity lines; doppler broadening of these lines determines the ion temperature. Plasma density is measured employing a laser interferometer. MCX has achieved high density (n>1020 m-3) fully ionized plasmas rotating supersonically and has two modes of operation. The first mode( O mode ) has azimuthal velocities in the range of 100 km/sec at midplane for discharge times exceeding 8 ms under a wide range of conditions with ion temperatures typically 30 eV and sonic mach numbers

(vphi/vti) in the range 1-2 and Alfven mach numbers (vphi/vA) of somewhat less than one. The O mode rotation velocity is close the limit predicted by the Alfven ionization condition, applied at the end insulators. A second higher rotation mode ( HR mode ) has been discovered which has rotation velocities greater than 200 km/s, ion temperatures similar to O mode and sonic mach numbers greater than 3, with somewhat lower Alfven mach numbers. The HR mode velocity substantially exceeds the Alfven limit at the insulators and there is evidence the HR mode plasma is detached from the end insulators. In general, MCX O mode plasmas remain stationary for many milliseconds but the HR mode is transient, lasting only a few milliseconds. Both lifetimes are much longer than MHD instability timescales but magnetic probe data show significant MHD activity, which is stronger in HR mode. P-5.121, Friday July 1, 2005

Kinetic Theory of Electrostatic Collisionless Shocks in Plasmas G. Sorasio, M. Marti, L. O. Silva GoLP/Centro de Física de Plasmas, Instituto de Superior Técnico, Lisbon, Portugal

The collision of plasma shells with different properties is quite an ubiquitous scenario: it can be found both in the universe, as in the collision of supernova's expanding plasma shells, and in the laboratory, as during the propagation of a laser-induced-shocks in a solid. In a general scheme, two different plasmas flow one towards the other. Previous theories [1-2] analyzed situations where the plasma was either colliding towards a perfecting reflecting mirror, or against another plasma with exactly the same properties, but so far none or little effort has been done to consider more complex scenarios. In the present work, we developed a new theoretical framework which describes, in a kinetic fashion, the properties of the electrostatic structures arising from the collision of slabs of plasma with different properties (temperatures, densities, etc.). At the present stage, our theoretical framework considers purely electrostatic shocks arising from the collision of two slabs of plasma with different electron temperatures and particle densities. We have found that the energy dispersion mechanism that allows the shock formation is due both electron trapping [1] and to ion reflection [2]. The first effect is found to be crucial when considering collisions of plasmas with different temperatures. We also found that the condition over the maximum allowed shock propagation velocity is modified by the presence of electrons with different temperatures. The consistency of the theoretical assumptions and the credibility of the theoretical predictions have been compared with one-dimensional and two-dimensional fully relativistic particle-in-cell simulations, carried out with OSIRIS 2.0 framework [3].

[1] D. W. Forslung and J. P. Freidberg, Physical Rev. Lett. 27, 1189 (1971). [2] R.Z. Sagdeev , Rev. Plasma Phys., 4, 23 (1966). [3] R. A. Fonseca et al., LNCS 2331, 342 (2002).

P-5.122, Friday July 1, 2005

Analysis of the influence of excited configurations and plasma interaction in atomic and plasma magnitudes of interest in NLTE plasmas using an analytical potential R. Rodríguez1, J.M. Gil1, R. Florido1, J. G. Rubiano1, P. Martel1 and E. Mínguez2 1Departamento de Física (Universidad de Las Palmas de Gran Canaria), Las Palmas de Gran Canaria, Spain 2Instituto de Fusión Nuclear (Universidad Politécnica de Madrid), Madrid, Spain

In the study of hot plasmas the availability of atomic properties results essential because the first stage in the radiation modeling is usually the calculation or the compilation of the necessary atomic data base. These atomic properties have to be obtained for all the ions immersed into the plasma, which are able to be in ground configuration and in excited configurations as well. This fact means that the amount of atomic data to handle with is huge and sometimes it is necessary, in order to reduce the complexity of the problem and the computing time, to resort to some more simple models that provide these data in a quite accurate manner, such as the effective analytical potentials. In this work we propose an analytical potential which includes excited configurations and plasma interaction simultaneously. With the help of this potential we have analyzed in this work the influence of the doubly excited configurations and the plasma effect in the oscillator strengths and level energies, because they are relevant atomic magnitudes in the resolution of rates equations and it is also studied this influence in level populations and ionic abundance for plasmas under NLTE conditions. These magnitudes are computed solving the rate equations in the Collisional Radiative Steady State (CRSS) model using ATOM3R code [1].

References: [1] Florido R, Gil JM, Rodríguez, R, Rubiano JG, Martel P, Mínguez E. ATOM3R a code for calculation of NLTE plasma populations. Low and high density limits behavior. Proceedings of the XXVIII European Conference on Laser Interaction with Matter. (In Press)

P-5.123, Friday July 1, 2005

Analytical potential for determining atomic properties of ions in plasmas for a wide range of plasma coupling parameter. An application to calculate total photoionization cross section. J.M. Gil1, R. Rodríguez1, R. Florido1, J.G. Rubiano1, P. Martel1 and E. Mínguez2 1Departamento de Física (Universidad de Las Palmas de Gran Canaria), Las Palmas de Gran Canaria, Spain 2Instituto de Fusión Nuclear (Universidad Politécnica de Madrid), Madrid, Spain In the calculation of optical properties of plasmas, it is needed to obtain atomic data for all the ions immersed into the plasma, which are able to be in ground configuration and in excited configurations as well. For a plasma in the regime of medium and high electron density, the atomic properties must be corrected taking into account the effect of the surrounding plasma into the atomic potential. This fact means that the huge amount of necessary atomic data has to be calculated for each set of plasma conditions. As a consequence, sophisticated self-consistent models become absolutely unpractical in this context and more simple models, such as the effective analytical potentials, are a very useful tool to provide these data with reasonable accuracy and to reduce the complexity of the problem as well as the computing time. In this work we propose a procedure for obtaining an atomic potential which includes interaction of the bound electrons of the ion with the surrounding plasma in a wide range of plasma coupling parameters. The potential has been developed considering free electrons charge distribution in the neighbourhood of the ion, linearized Boltzmann charge distribution for free electron and ions at medium and large distances of the nucleus and the reaction plasma-charge density to the optical electron in a relativistic quantum state (nlj). Also, a correction for degenerated free electrons has been taken into account. In addition, an analytical expression for this non isolated potential is proposed. Finally, this model is used to calculate total photoionization cross sections [1] in a wide range of the plasma coupling parameter, as in LTE as in NLTE regimes. References: [1] R. Rodríguez, J.M. Gil, J.G. Rubiano, R. Florido, P. Martel, E. Mínguez. J. Quant. Spectrosc. Radiat. Transfer 91 (2005) 393-413. P-5.124, Friday July 1, 2005

Using sparse matrices techniques and iterative solvers in the calculation of level populations for NLTE plasmas R. Florido1, J.M. Gil1, R. Rodríguez1, J.G. Rubiano1, P. Martel1, E. Mínguez2 1Departamento de Física (Universidad de Las Palmas de Gran Canaria), Las Palmas de Gran Canaria, Spain 2Instituto de Fusión Nuclear (Universidad Politécnica de Madrid), Madrid, Spain Calculations of populations of ion species existing in plasmas comprise an important research area for ICF and astrophysical plasmas. Ionic charge distribution is required for the determination of energy balance, emission, absorption and other properties of plasmas. If we take into account the different ionic stages in the plasma, which are able to be in ground or excited levels, we find two main difficulties in the determination of populations. On one hand, a huge amount of atomic data is needed to get the rate coefficients of the different collisional and radiative atomic processes that occurs in the plasma and, on the other hand, we have to solve a large number of coupled rate equations to obtain the level populations. We have developed ATOM3R code [1] to carry out the calculation described above for low and medium-Z plasmas. In our model, the determination of the atomic data becomes easy and fast by using analytical expressions for the effective atomic potential as well as for the rate coefficients. Then, we obtain with reasonable accuracy ionic abundances and averages properties of the plasma [2] solving the rate equations by traditional algebra methods. However, if we deal with high-Z plasmas or we are interesting in the plasma spectroscopy a very large number of configurations has to be considered and direct methods become unpractical to solve the rate equations. So, in this work the sparse matrix technique is employed to store the matrix elements of the rate equations and we show the utility of iterative methods to solve them. Finally, a brief discussion about how the number of configurations considered affects to average ionization and ionic charge distribution for different plasma conditions is presented. References [1] Florido R, Gil JM, Rodríguez, R, Rubiano JG, Martel P, Mínguez E. ATOM3R a code for calculation of NLTE plasma populations. Low and high density limits behavior. Proceedings of the XXVIII European Conference on Laser Interaction with Matter (In Press). [2] Ralchenko Y, Lee RW, Bowen C. Review of the third non-LTE code comparison workshop. Atomic Processes in Plasmas (AIP Conference Proceedings) 2004; 730:151-160. P-5.125, Friday July 1, 2005

mm-wave collective scattering study of density fluctuations in helicon plasmas

J. G. Kwak1, S. J. Wang1, S. K. Kim1 , S. Cho2 1 Korea Atomic Energy Research Institute, Daejeon, Korea 2 Kyunggi University, Suwon, Korea

Recently studies on low-frequency fluctuations in the helicon plasmas raised the important issue of linear and nonlinear microprocess which could alter the general characteristics of helicon plasmas.[1] Most of the measurement results, however, have been made with probes in the downstream region of the helicon discharges. In this paper, collective scattering measurement results obtained from the antenna region of our helicon source are discussed as well as comparison with probe results. The scattering system operates at 140GHz with an IF of 70 MHz, and a feedforward tracking circuit is used to handle frequency drifts and to reduce noise. The scattering angle ranges 18-24 degree with the corresponding wavenumber 9.3-12.73 cm-1.

[1] C.S. Corr et al, Phys. Plasmas, 11, 4596(2004)

P-5.126, Friday July 1, 2005

Orbital motion theory of the shielding around an electron emitting body 1G.L. Delzanno, 2A. Bruno, 3G. Sorasio, 1G. Lapenta 1Plasma Theory Group (T-15), Los Alamos National Laboratory, USA 2 Burning Plasma Research Group, Politecnico di Torino, Italy 3GoLP/Centro de Física de Plasmas, Instituto de Superior Técnico, Lisbon, Portugal

The problem of the shielding potential around a body in a collisionless plasma is investigated for the case of a body emitting electrons due to thermionic emission, photoemission, and secondary emission. The system is investigated by means of an analytic, self-consistent kinetic theory as well as Particle-In-Cell (PIC) simulations. The theory takes into account charged particles collection by the body as well as electron emission, and assumes spherical symmetry so that conservation of energy and angular momentum can be used to calculate the plasma distribution functions at any given point in phase space. Far away from the body the plasma is assumed unperturbed, and acts as a source of particles, balancing particle loss by grain collection. The theory is formulated for positively charged bodies and for radial profiles of the shielding potential presenting an attractive well, as found with PIC simulations by Delzanno et al. [1] in the case of thermionic emission. Once the plasma currents and densities are calculated by integrating the distribution functions in phase space, Poisson’s equation is solved numerically. The theory is shown to be in good agreement with PIC simulations. Several cases are presented for the three different emission mechanisms, showing that shielding potentials having an attractive well are also possible for photoemission and secondary emission [2]. The influence of physical parameters such as ion to electron mass ratio, temperature of the emitted electrons, body work function, etc are also shown. Finally, an analytical criterion for the presence of the potential well is derived for a simplified version of the kinetic theory presented here. Despite its simplifications, the model shows reasonable agreement with the full orbital model, at least for the cases considered.

[1] G. L. Delzanno, G. Lapenta, M. Rosenberg, Phys. Rev. Lett. 92, 035002 (2004). [2] G. L. Delzanno, A. Bruno, G. Sorasio, G. Lapenta, submitted Phys. Plasmas (2005).

P-5.127, Friday July 1, 2005

Calculation of the radiative opacity of laser-produced plasmas using a relativistic-screened hydrogenic model for ions including plasma effects J. G. Rubiano1, R. Florido1, R. Rodríguez1, J.M. Gil1, P. Martel1 and E. Mínguez2 1Departamento de Física de la Universidad de Las Palmas de Gran Canaria. 35017 Las Palmas de Gran Canaria (Spain) 2Instituto de Fusión Nuclear, Universidad Politécnica de Madrid. 28006 Madrid (Spain)

In the study of laser – plasma interaction is necessary to deal with an enormous amount of atomic data. In particular, the study of the radiative properties of hot dense matter requires the knowledge of the atomic structure of different ions in the plasmas, energy levels for ground and excited states, line transitions and plasmas interactions. A detailed numerical calculation of these quantities is a very time–consuming computer task. To deal with this problem some computer codes use analytical or semi–analytical models to computed the atomic data. One model widely used in plasma physics is the screened hydrogenic model. In this work, we use a relativistic–screened hydrogenic model to compute the radiative opacity of laser-produced plasmas. The model is based a set of screening charges which allow to calculate easily atomic properties ions immersed in plasma. These screened charges are derived from an analytical potential for non isolated ions using a simple analytical iterative procedure. In the opacity model used, ionic populations are obtained by solving the Saha equation including degeneracy corrections. Bound–bound transitions are determined using a Voigt profile for lineshape which includes natural, collisional, Doppler and UTA widths. Bound– free and free–free opacities are evaluated using the Kramer cross sections with appropriated corrections. Scattering processes are computed through the use of the Thomson formula with corrections. The results are compared with other screened hydrogenic models and more sophisticated self-consistent codes.

References: [1] Rubiano J.G., Florido R., Rodríguez R., Gil J.M., Martel P., Mínguez, E. Calculation of the radiative opacity of laser-produced plasmas using a relativistic-screened hydrogenic model. JQSRT 83 (2004) 159-182. [2] Gil JM, Martel P, Minguez E, Rubiano JG, Rodriguez R, Ruano FH. An effective analytical potential including plasma effects. JQSRT 75 (2002) 539-557. P-5.128, Friday July 1, 2005

Development of a New Non-Diaphragm Type Shock Tube for High Density Plasmas

K.N.Sato1, S.Kugimiya2, S.Nourgostar2, T.Aoki2, S.Kawasaki1, TRIAM Exp.Group1 琨Research Institute for Applied Mechanics, Kyushu University, Japan 2Interdisciplinary Graduate School of Engineering Sciences, Kyushu University, Japan

The shock tube of this type has been designed to be operated without a diaphragm which the conventional shock tube has in order to separate the high and low pressure regions. By this new scheme, several advantageous properties are expected to be obtained. The vacuum vessel of the conventional shock tube should be opened to atmosphere every time after each shot in order to replace a diaphragm by a new one. This process always introduces impurities into the vacuum vessel. The reproducibility of experiments sometimes depends on the reproducibility of diaphragm-breaking-phenomena. In the non-diaphragm type shock tube, there is no necessity for replacing a diaphragm after each experiment. Thus, it will be possible to reduce influx of impurities in the experiment. In addition, the reproducibility and the efficiency of experiments may be expected to be largely improved. High pressure driven pistons in the non-diaphragm shock tube are shown in Fig.1. The cylinder sections C and D are filled with a higher pressure gas than that of the driver room. Since a small pinhole (l=1mm) is drilled at the center of the piston B, the pressures of cylinder sections C and D are equal at the initial stage. When the gas is released from the C part through exhaust valve, the pressure of section C quickly decreases. Because of a low conductance through the pinhole between C and D, a pressure imbalance quickly develops and this will lead the fast movement of the piston B toward left hand side. When the gas in the section D flows out through lateral holes, the pressure imbalance between the section D and the driver room causes a fast movement of the piston A to the left side, and this makes the high pressure gas in the driver room flows into the driven room. Consequently, A non-diaphragm type shock tube is being developed for the gas-dynamic laser research. a shock wave will occur. Several improvements have been carried out in order to have quick movement of pistons, and finally the shock wave propagation has been successfully obtained. The dynamic behaviors of pressure changes caused by the arrival of shock waves have been observed, and the shock wave velocity has been measured by using piezo electric gauges .

Lateral holes InletInflow pipe pipe

Exhaust valve

Fig.1 High pressure driven pistons in the non-diaphragm type shock tube P-5.129, Friday July 1, 2005

X and XUV time-resolved emission of laser-created hot Xe and Kr plasmas

S. Tzortzakis1, J.-R. Marques1, N. Kontogiannopoulos1, S. Bastiani-Ceccotti1, L. Lecherbourg1, F. Thais2, I. Matsushima3, O. Peyrusse4, and C. Chenais-Popovics1

1Laboratoire pour l’Utilisation des Lasers Intenses, CNRS UMR7605, Ecole Polytechnique, 91128 Palaiseau, France 2CEA/DSM/SPAM, Centre d’Etudes de Saclay, 91191 GIF-sur-Yvette Cedex, France 3 National Institute of Advanced Industrial Science and Technology, Tsukuba, 3058568, Japan 4CELIA, Université Bordeaux 1, 33405 Talence, France

In last years it has been recognized, by the international scientific community, the need of well adapted experiments for the benchmarking of non-local-thermodynamic-equilibrium collisional-radiative codes for high-Z elements in hot plasmas. The current experiment presents an important contribution to this objective.

A frequency-doubled (0.53-nm) laser beam of the Nd:glass kilojoule nanosecond LULI2000 facility was focused on a Xe (Kr) gas jet. Time integrated x-ray spectra (in the regions: 12- 16 Å for Xe and 6-8 Å for Kr) and time-resolved XUV spectra (in the region: 20-200 Å) emitted from the plasma were simultaneously recorded. Time-resolved XUV spectra from gas jet plasmas are reported here for the first time, to our knowledge, showing interesting characteristics on the temporal evolution of the plasma emission.

Concurrently with the spectra, the plasma electron and ion densities, as well as electron temperatures were monitored using time-resolved Thomson scattering diagnostics. The information provided by the Thomson diagnostics was used as input for the AVERROES superconfiguration code [1]. This allowed direct comparisons between the measured and calculated X and XUV emission spectra and the corresponding charge-state distributions.

[1] O. Peyrusse, J. Phys. B 33, 4303 (2000); ibid JQSRT 71, 571 (2001).