Overview of TJ-II flexible heliac results

Fusion Engineering and Design 56–57 (2001) 145–154

Overview of TJ-II flexible heliac results

E. Ascasıbar *, C. Alejaldre, J. Alonso, L. Almoguera, A. Baciero, R. Balbın,M. Blaumoser, J. Botija, B. Branas, E. de la Cal, A. Cappa, J. Castellano,R. Carrasco, F. Castejon, J.R. Cepero, C. Cremy, J. Doncel, S. Eguilior,

T. Estrada, A. Fernandez, C. Fuentes, A. Garcıa, I. Garcıa-Cortes, J. Guasp,J. Herranz, C. Hidalgo, J.A. Jimenez, I. Kirpitchev, V. Krivenski, I. Labrador,

F. Lapayese, K. Likin, M. Liniers, A. Lopez-Fraguas, A. Lopez-Sanchez,E. de la Luna, R. Martın, L. Martınez-Laso, M. Medrano, P. Mendez,K.J. McCarthy, F. Medina, B. van Milligen, M. Ochando, L. Pacios,

I. Pastor, M.A. Pedrosa, A. de la Pena, A. Portas, J. Qin,L. Rodrıguez-Rodrigo, J. Romero, A. Salas, E. Sanchez, J. Sanchez,

F. Tabares, D. Tafalla, V. Tribaldos, J. Vega, B. ZurroLaboratorio Nacional de Fusion, Asociacion EURATOM-CIEMAT. A�. Complutense 22, 28040 Madrid, Spain

Abstract

The TJ-II is a four period, low magnetic shear stellarator, with high degree of configuration flexibility (rotationaltransform from 0.9 to 2.5) which has been operating in Madrid since 1998 (R=1.5 m, a�0.22 m, B0=1 T,PECRH�600 kW, PNBI�3 MW under installation). This paper reviews the main technical aspects of the TJ-II heliacas well as the principal physics results obtained in the most recent TJ-II experimental campaign carried out in 2000.© 2001 Elsevier Science B.V. All rights reserved.

Keywords: TJ-II; Flexible; Magnetic

www.elsevier.com/locate/fusengdes

1. Introduction

TJ-II is a four period, medium size, low mag-netic shear stellarator of the heliac type (R=1.5m, a�0.22 m, B0�1 T, PECRH�600 kW,

PNBI�3 MW under installation) which was con-structed in Madrid between 1991 and 1997 andwhich has been operating since 1998 [1,2].

TJ-II was designed to have a high degree ofmagnetic configuration flexibility, as is inherent toheliac devices. The rotational transform can bevaried between 0.9 and 2.5, the magnetic well maybe changed from −1 to 6%, while theory predictsmaximum beta values as high as 6%.

* Corresponding author. Tel.: +34-91-3466-369; fax: +34-91-3466-124.

E-mail address: [email protected] (E. Ascasıbar).

0920-3796/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.

PII: S0920 -3796 (01 )00237 -X

E. Ascasıbar et al. / Fusion Engineering and Design 56–57 (2001) 145–154146

In the most recent experimental campaign, itwas possible to couple the full ECH power (600kW) to the plasma using two ECRH transmissionlines with different power densities (1 vs. 15 W/cm3) and steering launching capabilities (fixed ver-sus poloidal and toroidal variation).

2. Technological issues of TJ-II

2.1. Technical concept

The design of the TJ-II is characterised by thefollowing design features:� Mechanically de-coupled systems. This means

that each of the main components is individu-ally supported by the general support structureand the components are not directly coupled toeach other.

� No coils inside the vacuum vessel. Althoughthe Hard Core (see below) needs to be veryclose to the plasma, it was decided to locate itoutside of the vacuum vessel in order to avoidmanufacturing, assembly and operation prob-lems [3,4].

� Modular design of the vacuum vessel. Thevacuum vessel consists of octants, whichformed the elementary units for the on-siteassembly [4].

� Human access to the vacuum vessel. Althoughits dimensions are rather small, it is equippedwith eight manholes to allow a person to enterthe vacuum vessel and to work inside it.

� Divisible circular toroidal field coils. Thesewere necessary for assembly reasons [5].

� All field coils are water cooled and made ofhollow copper conductors [3].

� The use of ferromagnetic materials wasavoided in the stellarator and its vicinity. Themaximum relative magnetic permeability al-lowed was 1.01.

� An air core transformer.The very compact design of the device, together

with the high magnetic field precision required,resulted in very narrow tolerances and high me-chanical and thermal loads on the components.

2.2. Main components of TJ-II stellarator

2.2.1. Hard coreThe Hard Core is the central part of the TJ-II

stellarator. It consists of a circular (CC) coil anda helical (HX) one. The CC coil is enclosed in astainless steel casing, while the HX one is woundaround this casing in phase with the helical pathof the Toroidal Field (TF) coils and the vacuumvessel [3].

The CC coil has 24 turns connected in seriesand two hydraulic cooling circuits. The maximumcoil current is 280 kA turn. The coil major diame-ter is 3 m and the coil cross-section is 65×68mm. A design was chosen so that no brazing ofthe conductor is located inside the coil. The coil isvacuum impregnated and has a diameter tolerancebetter than �1 mm.

The manufacture of the HX coil was very com-plicated as it spirals around the CC one. A specialsemi-automatic winding device was needed forthis purpose. The coil is composed of two sub-coils individually supplied in order to allow shearvariation. Each sub-coil has 12 windings. Thecross section of the two coils is 92×46 mm andthe total maximum coil current is 260 kA turn,which results in a maximum current density of120 MA/m2. This current density increases thecoil temperature by approximately 40 °C during apulse with a 200 ms flat-top.

2.2.2. Vacuum �esselThe VV is made of stainless steel 304 LN and is

composed of 32 sectors and 32 rings [4]. Thesectors support the 96 portholes, eight of whichallow human access to the VV. Sectors and ringswere factory-assembled to form octants, with eachoctant containing four sectors and three rings.The octants were connected to each other on-sitewith closing rings, the closing rings being manu-ally welded from within the VV in order toachieve vacuum tightness. Furthermore, in orderto achieve the required mechanical strength, addi-tional welding was done on the outer side of theclosing rings. Tolerances for the assembled VVwere very demanding due to the proximity to theCC/HX coils and the TF coils, the latter beinglocated directly above the rings. In general, the

E. Ascasıbar et al. / Fusion Engineering and Design 56–57 (2001) 145–154 147

wall thickness of the VV is 10 mm, but thisreduces to 7 mm in those areas where sufficientspace is not available.

2.2.3. Toroidal field coilsThe TJ-II toroidal field system consists of 32

circular coils distributed in vertical planes withfour-fold symmetry [3]. The coils provide a maxi-mum toroidal magnetic field of 1 T on axis. Thetoroidal position of the coils is modulated inorder to minimise the toroidal field ripple. Also,four of the coils are slightly enlarged in order togive better access for neutral beam injection. TheTF coils were built in two halves in order topermit their assembly on the VV and overlappingjoints connect the two halves of each TF coilwhich is enclosed in a non-magnetic stainless steelcasing (304 LN). Finally, the coil and casing set issupported by an external frame, which is also splitinto two halves for assembly reasons and which isconnected to the general support structure of themachine by four legs [5].

2.3. Heating systems

2.3.1. ECRH systemThe ECRH system of the TJ-II consists of two

gyrotron oscillators operated at 53.2 GHz. Eachgyrotron can deliver up to 300 kW of microwavepower at the output window for a pulse length upto 1 s. Both are triode type tubes and have abuilt-up converter forming the Gaussian outputwave beam with a purity of 96%. The operatingbeam voltage is 65 kV and the beam current is14.8 A [6]. The gyrotrons are driven by two anodemodulators and it is possible to modulate theanode voltage up to 50 kHz.

Two mirror lines transmit the microwave powerfrom the gyrotrons to the plasma. The first trans-mission uses eight mirrors and has a fixed injec-tion geometry. Line losses for this line areapproximately 12%, the beam width at the plasmaborder is approximately 50 mm, and the powerdensity is about 1 W/cm3. The second transmis-sion line has ten mirrors and includes a movablemirror located inside the vacuum vessel, permit-ting experiments on current drive, central andoff-axis heating to be carried out. The beam is

focused at the plasma centre and its waist is about9.5 mm. The total losses along this line are about14%, and the power density is 15 W/cm3.

A kinetic model that takes account of the fullionisation chain has been developed. It is used tosimulate plasma start-up by microwaves at thesecond harmonic of the electron cyclotron fre-quency. Good agreement between model predic-tions and the experimental data has been obtained[7].

2.3.2. NBI systemNB injectors and ion sources have been re-

ceived from Oak Ridge National Laboratory andthese are presently being installed on TJ-II. Theequipment, which is on loan, is undergoing refur-bishment, while auxiliary systems such as theHigh Voltage Power Supplies, the Cooling Sys-tem, the Primary Vacuum System, and the Con-trol System, are being supplied as new byCIEMAT [8]. It is expected that the first injectorwill be ready for plasma heating by the end of theyear 2001.

When the NBI is operating, the plasma will beheated by two H0 beams, each with approximately1 MW of power, mounted in a co-counterconfiguration. The beam energy is 40 keV in orderto achieve optimum absorption, while the maxi-mum pulse length is 300 ms. The estimated beamtransmission through the TJ-II apertures isgreater than 60%. This is because about 20% ofthe neutral beam power (350 kW) is purposelylost to a cooled diaphragm in the beam duct. Thisis done to protect sensitive structures such as theinner edge of the first toroidal field coil. Thermalloads are acceptable due to the moderate pulselength. However, a large effort is being devoted todesigning and constructing convenient thermalshields for the groove protection tiles and thebeam exit [9,10].

Physics studies show that for central plasmadensities in the range of 1020 per m3, the totalabsorbed beam power will be in excess of 60%.Semi-empirical central density limits for 1.2 MWof absorbed power are approximately 1.6×1020

per m3; at this limit the central � reaches 3.8%[11].


2.4. Technical status and operation

The current technical status of the TJ-IIis satisfactory. All systems are in a goodoperational condition and no major technicalproblems were encountered during the lastyears. A comprehensive maintenance program forthe device and its periphery, together with a so-phisticated monitoring and control system,guarantee safe and reliable operation. Normally,the device is operated 3 days a week, while Mon-days and Fridays are reserved for maintenance,system installations, data evaluation and meet-ings.

The data acquisition system provides resourcesfor remote control of instrumentation mainframes(VXI, VME and CAMAC), remote operation ofcontrol systems of diagnostics, on-line and off-linedata integration as well as a host-centralised data-base. The VXI modules were designed and devel-oped at CIEMAT and some of them supply realtime processing capabilities by means of digitalsignal processors [12]. A special application soft-ware was developed to provide access to TJ-IIinstrumentation modules and data visualisation[13]. Data are stored compressed withoutdistortion by a special technique, which showscompression rates over 80% with negligible com-putation time [14]. The database is organised ac-cording to a multi-layer model [15]. Data accessand addition of new data to the database isachieved by means of a distributed environmentbased on ONC RPC, environment which presentlyincludes Alpha AXP/UNIX, Sparc/Solaris, Intel/LINUX, Cray/UNICOS, Windows 9x and WindowsNT platforms [16].

As the shine through of the neutral beamscauses heat loads of �5 MW/m2 on the grooveof the vacuum vessel, it is essential to providesome heat resistant protection for the stainlesssteel groove protecting tiles. In order to avoid areduction in the plasma cross section by protec-tive elements, these steel plates will be covered bya 200 mm thick layer of boron-carbide (B4C). Atest program for B4C-covered stainless steel platesis now under way and it has produced promisingresults to date [10].

3. Recent experimental results in TJ-II

3.1. Particle control and wall conditioning

3.1.1. Baking of TJ-II �acuum chamberThe stainless steel vacuum vessel of TJ-II has a

volume of 6 m3 and a total inner surface area ofabout 75 m2. This vessel is pumped by a systemconsisting of four turbomolecular pumps with atotal effective pumping speed of about 4000 l/s[17]. Before starting the 1999 experimental cam-paign, the vacuum vessel was baked at 150 °C ina thermal cycle that lasted 30 h. The partialremoval of residual water by baking, which wasconfirmed by RGA data, led to a decrease in thetotal base pressure. Subsequently, base pressuresof the order of 5×10−8 mbar were achieved.These conditions were preserved even after vent-ing the vessel (using back-fillings with dry N2) fordiagnostic maintenance. In addition, the residualwater can be further reduced using glow dis-charges in helium.

3.1.2. Helium and argon glow-discharge cleaningThe wall conditioning procedure followed dur-

ing the 1998 and 1999 campaigns was overnightHe glow discharge cleaning (GDC) at room tem-perature [18]. The principal effect of this proce-dure was to remove the hydrogen implanted onthe walls during the day’s plasma discharges andto produce an activated surface with typical wallpumping behaviour. This latter effect was de-duced from an observed decrease of about 10–20% in residual pressure after a He GDC. Also,the residual gas analyser (RGA) placed directly inthe TJ-II vacuum chamber showed a strong de-crease of masses 18 and 32. The improvement inthe base pressure achieved by decreasing the resid-ual water (baking of the chamber followed by HeGDC) has had an important effect on the repro-ducibility of TJ-II discharges. This is primarilydue to the suppression of sources of electrons, asthese electrons can be accelerated during the riseof the currents in the magnetic coils. However,during the 1999 campaign, density control forhigh ECRH injected powers (�300 kW) was stillextremely difficult. The reason for this was thatthe He implanted in the walls during the


overnight GDC conditioning was released duringplasma discharges thereby producing an uncon-trolled increase in the plasma density [19]. As anintensive He GDC seemed to be essential forTJ-II in order to achieve reproducible discharges,an additional procedure was needed to remove theimplanted He from the walls before plasma opera-tion. With this objective, prior to beginning oper-ation, an Ar GDC was performed for �30 minafter the overnight He GDC had been stopped.The release of He atoms from the walls by Arbombardment has been measured, and the inte-grated flux released during the 30 min typicallyyields a total of 1021 He-atoms. This correspondsto �1 monolayer of He from the TJ-II vacuumchamber surface area (�75 m2). This amount ofHe depletion is sufficient to permit operation un-der high injected power using density control byexternal gas puffing [19]. The development of agas injection and monitoring system for very lowgas flows (�0.05 Pa m3 per s) has also helped inthis respect.

Low Z effective values (1.5–2.5 in H plasmasand 3–4.5 in He plasmas) and low radiated pow-ers (�20% of injected power) were typicallyachieved during the year 2000 experimental cam-paign for all heating schemes applied. These factsare consistent with spectroscopic observations(small amounts of metallic species, i.e. Cr, Fe, Ni,Cu) [20] and with the very low values of desorbedAr deduced from mass spectrometry [19].

3.2. Magnetic configuration and confinement

Magnetic configuration (iota �1.28–2.24) andplasma volume (0.6–1.1 m3) scans have beenmade and stored plasma energies (W) reaching1.5 kJ have been measured for electron densitiesand temperatures up to 1.2×1019 per m3 and 2.0keV, respectively, with PECRH�600 kW. The de-pendence of diamagnetic energy normalised to thecalculated plasma volume as a function of rota-tional transform for helium and hydrogen plas-mas has been studied [21]. Although a positiveiota dependence (in the low iota range) was re-ported in [21] for helium plasmas, very recentresults, which have still to be fully analysed, showno clear dependence of normalised energy content

for hydrogen plasmas in the same iota region.These last discharges have been obtained with theimproved vacuum vessel conditions (additional ArGDC) described in Section 3.1. Thus, the differentbehaviour found in helium and hydrogen plasmasmay just indicate that the older helium resultscorrespond to discharges in which optimum confi-nement was yet not reached. Obviously, furtherexperiments are needed to elucidate this point.

Maximum chord-averaged CV temperature andpoloidal rotation data are plotted in Fig. 1 as afunction of the central iota value. It is observedthat the lowest impurity temperature value corre-sponds to the lowest rotational transform andthat in this range, the shell where CV emissionpeaks, rotates in the negative direction [22].

High spatial resolution, Thomson scatteringmeasurements have revealed the presence of finestructures in both the density and temperatureprofiles of all magnetic configurations [23]. Aninvestigation of impurity radiation profiles usingan automated pattern recognition procedure hasalso indicated the presence of such topologicalstructures [24]. A possible link to the iota profile

Fig. 1. Maximum CV temperature (top) and poloidal rotationvelocity (bottom) versus central iota value. Both graphs corre-spond to chord-averaged values in He plasmas.


Fig. 2. Upper box: MXR flux from two discharges with thesame magnetic configuration (iota (a)=1.61) as a function ofthe normalised effective radius. In shot c2615 (thick line), theECRH pulse lasted 2 ms and in shot c2625 (thin line) noECRH heating was applied. Lower box: effect of the thermalplasma on the apparent position of the confinement region intwo discharges with the same magnetic configuration as inupper box.

is close to the position of the n/m=8/5 resonantsurface. Fig. 2 (lower box) shows MXR flux fromshots c3476 and c3477, which have the samemagnetic configuration as those in the upper box.In c3477 (40 ms long ECRH pulse), as inc2615 (upper box), the probe was inserted about50 ms after the plasma was extinguished, while inc3476 (200 ms long ECRH pulse) plasma wasstill present when the probe was inserted. Theresult found was that there was no difference inthe position at which fast electrons are confinedwhen no thermal electrons remain in the vacuumvessel (c2615 and c3477). In fact, only thenumber of electrons increases with pulse length(note the factor 10 increase in flux intensity).However, when the probe was inserted during thegyrotron pulse (c3476) a slight outward radialdisplacement was observed in the position of theconfinement region. This result can be understoodas the effect of the bootstrap current on therotational transform radial profile.

3.3. Transport and turbulence studies

Electron heat diffusivity has been investigatedduring ECRH power modulation experiments andusing power balance analysis. The heat conductiv-ity is about 4 m2/s in the plasma core region andincreases towards the plasma boundary region[28].

In addition, electron temperature profile mea-surements made using electron cyclotron emission(ECE) and Thomson scattering diagnostics haveshown evidence of internal heat transport barri-ers. This is illustrated in Fig. 3. Transport analysisshows a reduction in transport coefficients totransport rates that are consistent with neo-classi-cal predictions based on Monte Carlo simulations[29]. The cause of this observed enhancement inconfinement could be electrons trapped by rippleand pumped out by ECRH to create a shearedelectric field while ExB de-correlation effectscould diminish turbulent transport. Indeed, non-Maxwellian features have been observed in elec-tron distribution functions in the 1–5 keV energyrange, a result that could be related to an ECRHinduced deformation of the distribution function.

(i.e. rationals) and the influence of plasmaparameters (collisionality, instabilities) is beinginvestigated.

Enhancements in bremsstrahlung emissionscaused by high energy electrons colliding with afast reciprocating probe tip have been found [25].They can be associated to the presence of relevantlow order resonances [26]. Fig. 2 (upper box)shows the intermediate energy (20–200 keV) X-ray (MXR) flux measured in two shots with thesame magnetic configuration. In shot c2625, noECRH pulse was applied, therefore the emissionscome from runaway electrons. In shot c2615, arunaway electron suppresser was operated beforethe 300 kW pulse (2 ms long) was triggered [27]and the probe was inserted 80 ms after plasmaextinction. As seen, flux enhancement begins atthe same radial location as that of the probe tip(expressed in terms of normalised effective ra-dius), which, according to theoretical calculations,


The energy spectra of Charge Exchange (CX)neutral particles have also been measured. As aresult, ion temperatures, Ti, in the range 50–70 eVwere obtained from the slope of the energy spec-trum in the 100–200 eV region, whereas Ti valuesof 90–120 eV were obtained using the same anal-ysis from fast atoms with energies above 250 eV.The analyser, which is placed in a cross sectionwith the highest magnetic ripple (3% at �eff=0and about 35% at �eff=1), scans a poloidal crosssection along a vertical line. In this position, thecharge exchange analyser mainly measures the ionenergy distribution of trapped particles. In addi-tion, part of the Ti(r) radial profile has beenmeasured for the 100–44–64 configuration (cen-tral iota 1.551). This is shown in Fig. 4 and it canbe fitted by the expression: Ti(�)=Ti(0)× [1−��]�, where ��4 and ��3 for ��0.6.

ELM-like transport events have been measuredin plasmas with stored energies of at least 1 kJ.

Fig. 4. Ion temperature radial profile for the 100–44–64configuration (central rotational transform 1.551) with 300 kWheating power.

The plasma develops bursts of magnetic activity(observed by Mirnov coils) followed by a largespike in the Ha signal, see Fig. 5 [30]. Also, theelectron temperature measured by the ECE sys-tem shows a pivot point at the plasma radius�eff�0.6 (at this point the temperature is 100–200 eV). As a consequence, the electron tempera-ture profile is flattened at this position. Thisflattening is due to a factor of 2 increase in theelectron thermal diffusivity, caused by the trigger-ing of the ELM instability. These events are lo-calised in the pressure gradient region therebysuggesting a possible role of resistive ballooninginstabilities.

Edge parameters (electron density and tempera-ture) have been investigated in a fixed magneticconfiguration (named the 100–40–63, with rota-tional transform at axis 1.509) for both H and Heplasmas. For this, He and Li atomic beams, aswell as Langmuir probes, were employed [31].Density and power scans have also been con-ducted, but no significant differences are observedin electron temperatures close to the last closedflux surface (LCFS) when the power is changed.Diffusion coefficients of the order of DBohm andglobal particle confinement times (�p) rangingfrom 14 to 3 ms have been obtained. A cleardegradation of particle confinement with injectedpower has also been found, together with indica-tions of confinement enhancement with density.

TJ-II has a large edge magnetic ripple value(�35% for the standard configuration). As aconsequence relatively large fractions of trapped

Fig. 3. Thomson scattering (electron temperature and density)profiles of helium plasmas with 300 kW injected power, rota-tional transform iota (0)=1.51 and different electron density.An ECE temperature profile (open squares in the upper box)is displayed for the low density case in which the ITB appears.


particles together with enhanced energy and parti-cle losses should be expected in the plasmaboundary region (�eff�0.7) [32]. This may ex-plain the particularly low densities and tempera-tures observed by edge diagnostics in this region[31]. Along this line, experiments in which theplasma volume has been decreased by insertingmobile poloidal limiters have been made and theresults show improved confinement when com-pared with plasmas bounded by the usual belttoroidal/helical limiter (TL) [21,33]. This improve-ment is interpreted in terms of a modification of

the SOL due to partial suppression of the tenuousTL plasma boundary, which is very transparent toneutrals [33].

Experimental evidence of ExB sheared flowslinked to rational surfaces has been obtained inthe plasma edge region for different magneticconfigurations [34,35]. The presence of the 4/2rational surface predicted by vacuum magneticfield calculations has been detected by two clearfootprints, namely, the flattening of edge profiles,and the modification in the root mean square(rms) of floating potential fluctuations and in thepoloidal phase velocity of fluctuations. The mea-sured correlation time of fluctuations (10 �s) turnsout to be comparable to the inverse of the ExBde-correlation rate, thereby suggesting a possiblerole of rational surfaces in accessing high confi-nement regimes. The resulting ExB sheared flowsassociated with rational surfaces could thereforedepend on the competition between mechanismsthat drive flow processes (i.e. Reynolds stress,ion-electron flux differences) and that dampenthem (i.e. magnetic viscosity).

The measurement of wave number spectra isdifficult in fusion devices due to the fact thatspatially resolved information is required. A com-parative study of wave number spectra obtainedusing both the high-resolution Thomson scatter-ing system (core plasma region) and the Langmuirprobes (edge region) is under way. The (radial)wave number spectra show a remarkable similar-ity in their shapes as these do not appear to besignificantly dependent on the measuring tech-nique employed or on the plasma region or coreconditions. In particular, the wave number spec-tra obtained by the Thomson scattering diagnos-tic are similar to those obtained by othertechniques or devices. This may show that thedetailed structure observed in the Thomson den-sity and temperature profiles indicate a footprintfor ‘common’ turbulence as well as the influenceof magnetic topology [23,36].

In order to investigate transport characteristicsclose to instability thresholds, radial profiles ofion saturation current and of floating potential, aswell as their fluctuations, have been measured formagnetic configurations with the same rotationaltransform (iota (a)�1.8) but with magnetic well

Fig. 5. Time traces of: (a) H� signal, (b) Mirnov coil signaland (c) ECE signals at four radial positions showing ELM-likeactivity in TJ-II shot c2156.


ranges from 0.2 to 2%. The level of fluctuations atthe plasma boundary increases in plasma configu-rations with a magnetic hill at the plasma edge. Theincrease in the fluctuation level is caused by fluctu-ations in the frequency range (1–30 kHz). Interest-ingly, the breaking point in the frequency spectra(i.e. 1/f region) can be directly related to the levelof fluctuations. These experimental results show theimportant role of the magnetic well in stabilisingpressure gradient instabilities in the TJ-II stellara-tor and open the possibility of investigating theproperties of turbulent transport in the proximityof instability thresholds [37].

4. Conclusions

During the year 2000 campaign, suitable glowdischarge cleaning of the TJ-II vessel combinedwith a very accurate gas puffing system has alloweddensity control of both helium and hydrogen plas-mas. Internal heat transport barriers appear in lowdensity and high ECRH power density plasmas.

The flexibility of the TJ-II device is proving tobe a very powerful tool for plasma physics studies,as demonstrated by the following results:� Fine structures in temperature and density

profiles appear in all TJ-II discharges andconfigurations. Its possible relationship withmagnetic topology is under study.

� Rotational transform scans have been per-formed but, so far, no clear dependence ofconfinement on iota has been found.

� Experimental evidence of fast electronsconfined near low order rational surfaces lo-cated in the plasma edge has been found.

� ELM-like transport events increase heat con-ductivity. Such events seem to be triggered by alow order rational surface interacting with aresistive ballooning instability.

� Experimental evidence of ExB sheared flowslinked to rational surfaces has been obtained inthe plasma edge region for different magneticconfigurations.

� Magnetic well scans for a fixed iota-value yieldresults that point to the possible role played bythe magnetic well in stabilising pressure gradi-ent driven instabilities in TJ-II.

References

[1] C. Alejaldre, et al., Confinement studies in the TJ-IIstellarator, Plasma Phys. Control Fusion 41 (1999) B109–B117.

[2] E. Ascasıbar, et al., Magnetic surfaces mapping experi-ments in TJ-II Heliac, J. Plasma Fusion Res. 1 (1998)183–186.

[3] J. Alonso et al., TJ-II magnetic field coils: present status,Proc. 17th Symp. on Fusion Tech., Rome 1992, in FusionTechnology 1 (1992) 763-767.

[4] J. Botija et al., Fabrication of the vacuum vessel of theSpanish stellarator TJ-II, Proc. 18th Symp. on FusionTech., Karlsruhe 1994, in Fusion Technology 1 (1994)771-774.

[5] J. Botija et al., Final assembly and present status of theSpanish stellarator TJ-II, Proc. 19th Symp. on FusionTech., Lisbon 1996, in Fusion Technology 1 (1996) 985-988.

[6] R. Martın et al., Complex for ECRH and ECCD experi-ments on TJ-II, Proc. 20th Symposium on Fusion Tech-nology, Marseille 1998, in Fusion Technology 1 (1998)403-406.

[7] A. Cappa, et al., Plasma breakdown by ECRH at 2ndharmonic in TJ-II, Nuclear Fusion 41 (4) (2001) 363.

[8] M. Liniers et al., Neutral Beam Injection System forTJ-II, Proc. 20th Symp. on Fusion Tech., Marseille 1998,Fusion Technology 1 (1998) 307-310.

[9] J. Guasp, et al., Thermal load calculations at TJ-II vac-uum vessel under NBI, Fusion Technol. 35 (1999) 32–41.

[10] C. Fuentes et al., Development and tests of B4C coveredheat shields for TJ-II, presented in this 21st SOFT Con-ference, Madrid 2000.

[11] J. Guasp et al., Transport Studies in TJ-II, Proc. 10th Int.Conf. on Stellarators, Madrid. Report EUR-CIEMAT 30(1995) 322-325.

[12] J.L. de Pablos, et al., Improvements in the treatment ofsignals used for plasma diagnostics, IEEE Trans. NuclearSci. 43 (1996) 229.

[13] J. Vega, et al., The TJ-II data acquisition system: anoverview, Fusion Eng. Design 43 (1999) 309.

[14] J. Vega, et al., Encoding technique for a high datacompaction in data bases of fusion devices, Rev. Sci.Instrum. 67 (1996) 4154.

[15] J. Vega, et al., Data management in the TJ-II multi-layerdatabase, Fusion Eng. Design 48 (2000) 69–79.

[16] E. Sanchez, et al., Accesing TJ-II data with RPC, Rev.Sci. Instrum. 72 (1) (2001) 525–529.

[17] F.L. Tabares, et al., The vacuum system of the TJ-IIstellarator, Vacuum 45 (1994) 1059.

[18] F.L. Tabares, et al., Plasma-wall interaction in theSpanish stellarator TJ-II. Diagnostics and first results, J.Nucl. Mater. 266–269 (1999) 1273–1278.

[19] D. Tafalla et al., Wall conditioning and density control inTJ-II stellarator, Proc. 14th Int. Conf. on Plasma SurfaceInteractions, Rosenheim, Germany 2000, J. Nucl. Mater.290-293 (2001) 1195–1198.


[20] K.J. McCarthy et al., Impurity behaviour studies in theTJ-II Stellarator, Proc. 27th EPS Conf. on ControlledFusion and Plasma Physics, Budapest 2000, ECA Vol.24B (2000) 1244–1247.

[21] E. Ascasıbar et al., Magnetic configuration dependence ofthe plasma energy content in TJ-II stellarator, Proc. 27thEPS Conf. on Controlled Fusion and Plasma Physics,Budapest 2000, ECA 24B (2000) 209–212.

[22] B. Zurro et al., Impurity ion rotation and temperaturebehaviour in the TJ-II stellarator as a function of injectedpower and magnetic configuration, Proc. 27th EPS Conf.on Controlled Fusion and Plasma Physics, Budapest2000, ECA Vol. 24B (2000) 189–192.

[23] J. Herranz et al., Profile structures of TJ-II Stellaratorplasmas, Phys. Rev. Lett., 85 (22) (2000) 4715–4718.

[24] A. Baciero et al., The search for topological structures inTJ-II radiation profiles using an automated pattern recog-nition procedure, Proc. 27th EPS Conf. on ControlledFusion and Plasma Physics, Budapest 2000, ECA 24B(2000) 1248–1251.

[25] M.A. Ochando et al., Confinement of fast electrons atrational surfaces in the TJ-II Stellarator, Proc. 12th IAEAInt. Stellarator Workshop, Madison Wisconsin USA(1999) (in CD-ROM).

[26] F. Medina, et al., Filamentary current detection in stel-larator plasmas, Rev. Sci. Intrum. 72 (2001) 471.

[27] L. Rodrıguez-Rodrigo et al., Generation of fast electronsin TJ-II, Proc. 26th EPS Conf. on Controlled Fusion andPlasma Physics, Maastricht, The Netherlands, ECA 23J(1999) 353–356.

[28] F. Castejon et al., Transport evaluation of TJ-II plasmas,

Proc. 12th IAEA International Stellarator Workshop,Madison Wisconsin USA (1999) (in CD-ROM).

[29] F. Castejon et al., Internal heat transport barriers in TJ-IIflexible heliac, submitted to Nuclear Fusion, 2001.

[30] I. Garcıa-Cortes, et al., ELM-like transport events inTJ-II stellarator, Nuclear Fusion 40 (11) (2000) 1867.

[31] F.L. Tabares et al., Edge transport characterisation ofhydrogen and helium plasmas in the TJ-II stellarator,Proc. 27th EPS Conf. on Controlled Fusion and PlasmaPhysics, Budapest 2000, ECA 24B (2000) 193–196.

[32] V. Tribaldos et al., Monte Carlo computations of neoclas-sical transport in TJ-II stellarator, Proc. 26th EPS Conf.on Controlled Fusion and Plasma Physics, MaastrichtThe Netherlands, ECA 23J (1999), 349–352.

[33] E. de la Cal et al., Plasma boundary and S.O.L. studies ofECH plasmas in TJ-II stellarator with diagnosed mobilepoloidal limiters, Proc. 14th Int. Conf. on Plasma SurfaceInteractions, Rosenheim, Germany 2000, J. Nucl. Mater.290–293 (2001) 579–583.

[34] C. Hidalgo, et al., Generation of sheared poloidal flowsvia Reynold stress and transport barrier physics, PlasmaPhys. Contr. Fusion 42 (2000) A153.

[35] M.A. Pedrosa et al., Edge fluctuations and transport inthe TJ-II stellarator, Proc. 27th EPS Conf. on ControlledFusion and Plasma Physics, Budapest 2000, ECA 24B(2000) 9–12.

[36] B. Van Milligen, et al., k spectra of Thomson scatteringtemperature profiles at the TJ-II stellarator, Nuclear Fu-sion 41 (4) (2001) 447–453.

[37] C. Alejaldre et al., Review of confinement and transportstudies in the TJ-II flexible heliac, Nuclear Fusion 2001,in press.

Documents

Overview of TJ-II flexible heliac results