RESEARCH PLAN FOR DIVERTOR SIMULATION STUDIES AND ITS RECENT RESULTSUSING THE GAMMMA 10 TANNDEM MIRROR

Y. Nakashima1, H. Takeda1, R. Yonenaga1, K. Hosoi1, H. Ozawa1, T. Ishii1, N. Nishino2, M. Ichimura1, T. Kariya1,I. Katanuma1, R. Minami1, Y. Miyata1, Y. Yamaguchi1, M. Yoshikawa1 and T. Imai1

1Plasma Research Center, University of Tsukuba, Tsukuba, Ibaraki 305-8577, Japan2Graduate School of Engineering, Hiroshima University, Higashi-Hiroshima, 739-8527, Japan

As the new research plan of Plasma Research Centerof the University of Tsukuba, we are planning to start astudy of divertor simulation under the closely resemble toactual fusion plasmas environment making an advantageof the GAMMA 10 tandem mirror and to contribute thesolution for realizing the divertor in future toroidalsystems. In the research plan, the concepts of two divertordevices are introduced. One has an axi-symmetricdivertor configuration with separatrix (A-Div.) and theother is a high heat flux divertor simulator by using anend-mirror exit of the large tandem mirror device (E-Div.). Preparative experiments have been successfullystarted at the end-mirror region of GAMMA 10 anddetailed behavior of end-loss particles has beeninvestigated by using newly developed diagnosticinstruments. In standard hot-ion mode plasmas (ne0 ~2×1018 m-3, Ti0 ~ 5 keV), the heat flux density of 0.8MW/m2 and the particle flux density of 4×1022 /s•m2 wereobserved at 30 cm downstream of the end-mirror exit onthe machine axis. It is confirmed that the heat flux densityincreases in proportion to the applied RF power.Superimposing the ECH pulse induces a remarkableenhancement of heat flux and a peak value in the net heatflux density of 8 MW/m2 was attained during the ECHinjection, which almost comes up to the heat load level ofthe divertor plate of ITER. Two-dimensional visible imagemeasurement of newly installed target plates using high-speed camera revealed a significant difference in thebehavior of visible emission from plasma-materialinteraction. The above results give a clear prospect ofgenerating the required performance and providing usefulinformation for divertor studies in GAMMA 10.

I. INTRODUCTION

It is one of the important issues in fusion devices tooptimally design the divertor that is an essentialcomponent for exhaust of He ash and impurity particles[1]. In ITER and fusion reactor plasmas, the heat load tothe divertor materials is expected to be more than 10MW/m2 during several hundreds sec [2]. The steady-state

sustainment of tokamak discharge with detached-plasmacondition is an important key as an operation scenario offusion reactors to reduce such high heat-load to thedivertor plates [3]. In order to solve such problems, alarge number of so-called divertor simulators have beendeveloped [4-8]. In the Plasma Research Center (PRC) ofUniversity of Tsukuba, the following two projects areinitiated to upgrade the our conventional research plan forthe physics study on the control of plasma potential andradial electric field using wave heating techniques, suchas ECH.(1) Research on improvement of plasma confinement by

potential / electric field will be extended from the coreregion to boundary region.

(2) We will introduce new divertor systems (A-Div. andE-Div.) to existing tandem mirror device (GAMMA 10)and develop the study of divertor simulation.

For this purpose, we started a study of divertorsimulation under circumstances with close resemblance toactual fusion plasma in order to directly contribute therealization of the divertor in ITER and fusion reactors,making use of the advantage of a large-scale tandemmirror device.

In GAMMA 10 [9,10], many plasma production andheating devices with the same scale of present-day fusiondevices, such as radio-frequency (RF) wave, microwaveand neutral beam injection systems have been equippedand high-temperature plasmas have been produced. In thispaper, a brief description of the concept of divertorsimulation studies and the detailed results of investigationon the plasma flow from the end-mirror exit of GAMMA10 performed to examine its performance relevant to thedivertor simulation studies are given.

In the next section, outline of new research plan isintroduced and the concepts of two divertors areexplained. In section 3, the recent results of thepreparative experiments carried out in the end-mirrorregion are presented and the research plan for divertorstudies in near future is reported in section 4. Finallysummary of this paper is presented in section 5.

II. OUTLINE OF THE NEW DIVERTORRESEARCH

II.A. GAMMA 10 Device and Objectives of TwoDivertor Projects

Figure 1 shows the schematic view of GAMMA 10tandem mirror device that is the main experimental devicefor the divertor project. GAMMA 10 is an effectively axi-symmetrized tandem mirror with minimum-B anchor andmainly separated into four sections, from the central to theend [9]. In the central-cell, main plasma is produced andheated by ion cyclotron range of frequency (ICRF) waves.In each minimum-B anchor-cell which is connected to theboth ends of the central-cell, MHD stabilization of theplasma is preserved. Plasma confining potential isproduced by ECH using high power gyrotrons atplug/barrier region in the end-mirror cell and the end-lossplasma leaking to the end-cell is plugged with thisconfining potential. Plasma particles flowing out of theplug/barrier-cell hit the end plates and neutralized gas ispumped out with large-scale cryo-pump system installedon the end cell [11].

In this research plan, we develop the new divertorcontaining a separatrix configuration based on a novelidea (A-Divertor) and a high-heat flux divertor simulatorby use of open-ended device (E-Divertor). By using theabove devices, we approach the following researchsubjects of urgency for the ITER project.(a) Plasma-PFC interaction and impurity transport in the

divertor region (surface physics under ELM-like highheat flux environment).

(b) Steady-state sustainment of detached plasmas and itsphysical mechanism (atomic-molecular processes in thedivertor plasma, etc.).

(c) Heat and particle control in pedestal and divertorregion.

(d) Divertor pumping and particle exhaust experimentsusing the large-capacity cryopumping system.

II.B. Axisymmetric Divertor Configuration withSeparatrix Structure (A-Div.)

Figure 2 shows the schematic view of the newdivertor configuration and the example of introducing anaxi-symmetric divertor into the GAMMA 10 device. Thisaxi-symmetric divertor has a similar magnetic structurewith the toroidal divertor of torus systems. As shown inFig. 2(a), this configuration has a separatrix, whichenables to define the separation of the core plasma and theperiphery one in the mirror device for the first time. Inthis design example shown in Fig. 2(b), a set ofminimum-B coils in the west anchor cell is replaced withanother set of divertor coils. In the experiment with newdivertor configuration, the plasma stability is controlledby conventional minimum-B of one side together with thenew divertor. Designing of the magnetic configuration isin progress under the optimized condition of MHDstability [12] and compatibility with existing electricpower systems.

II.C. High Heat Flux Divertor Simulator at the End-mirror (E-Div.)

Fig.1. Schematic view of the GAMMA 10 tandemmirror device.

(a)

(b)

Fig.2. Schematic view of A-Divertor. (a)Conceptual view of axi-symmetric divertor. (b)Design example of A-divertor in GAMMA 10.

Figure 3 shows the conceptual view of open magneticconfiguration using a simple mirror and an example ofschematic of the divertor simulator using an end-mirrorexit of GAMMA 10. Open magnetic field configuration isan optimal structure as a divertor simulator. Some kind ofmodification will be planned to achieve this concept. Alarge-scale cryo-pumping system enables to achievepumped divertor configuration [13, 14] which existingsmall divertor simulators cannot realize. By using high-power plasma heating systems installed in GAMMA 10(ICRF, ECH and NBI), a large diameter and high heatflux plasma flow is expected. Pulse operation using strongelectron heating using high-power gyrotrons also cancontribute to divertor plasma control experiments, such asELM simulation, under a similar environment of divertorsin actual fusion plasmas (high Ti, Te and dischargewithout arc-discharge electrodes)

Based on the experimental results obtained inGAMMA 10, axial distributions of heat and particle fluxdensities have been estimated. In Fig. 4(a), the calculatedheat flux density and the cross-sectional area of themagnetic flux tube S are plotted against the distance fromthe center of the end-mirror coil (zEXIT). The magneticflux tube is expanded from the exit of end-mirror coil andthen each flux density decreases with the distance fromthe exit. In this calculation, the plasma density at themirror exit is assumed to be 1×1019 m-3 and the results areshown as a parameter of parallel ion energy Ei// in therange from 10 eV to 200 eV. The heat flux density, asshown in Fig. 4(a), the heat flux density of >10 MW/m2

will be achieved in the case that the plasma density of1019 m-3 with the parallel ion energy of 100 eV at theperiphery of the mirror exit. In the calculation of particleflux density, as shown in Fig. 4(b), the plasma density1018-1019 m-3 and the Ei// = 200 eV at the mirror exit canprovide the particle flux of 1022-1024 particles/s•m2 withinthe range of 1 m from the mirror exit. From the aboveresults, it is expected that hot and high density plasma

production (>1019 m-3 and several hundreds eV) by usingthe existing plasma heating systems in the end-mirrorregion generate the particle flux and heat flux relevant todivertor simulation studies (1023-1024/s•m2 and 1 - severaltens MW/m2) at 0.5 m downstream from the end-mirrorexit.

III. PREPARATIVE EXPERIMENTS AT THE END-MIRROR OF GAMMA 10

III.A. Experimental Apparatus of High Heat-FluxExperiment at the End-mirror Region

In the beginning of FY 2009, measurements of heatand particle fluxes at the end-mirror exit have been startedas a preparative experiment of new project in the westend-mirror region of GAMMA 10. Figure 5 shows theschematic view of the vacuum vessel and the shape of theplasma in the west end-mirror region, together with thelocation of the diagnostic equipment installed for thisexperiment. In order to perform a simultaneousmeasurement of heat and particle fluxes from the end-mirror exit, a set of calorimeter and directional probe wasmanufactured.

Figure 6(a) shows the assembly photograph of thediagnostic instrument. This diagnostic instrument islocated at 30 cm downstream from the end-mirror coil(zEXIT = 30 cm) and can be inserted from the bottom of thevacuum vessel up to the center axis of GAMMA 10. This

Fig.3. Conceptual view of high heat flux divertorsimulator using GAMMA 10 west end-cell.

(a)

0.01

0.1

1

10

100

0

0.2

0.4

0.6

0.8

0 0.2 0.4 0.6 0.8 1

Ei// = 200 eV Ei// = 100 eV Ei// = 50 eV Ei// = 20 eV Ei// = 10 eV

Mag. flux tube

Heat

Flu

x De

nsity

[ M

W/m

2 ]

S [m2]

ZEXIT [m]

ni0 = 1x1019 [m-3]

(b)

1021

1022

1023

1024

1025

0

0.2

0.4

0.6

0.8

0 0.2 0.4 0.6 0.8 1

ni0 = 2x1019 m-3

ni0 = 1x1019 m-3

ni0 = 5x1018 m-3

ni0 = 2x1018 m-3

ni0 = 1x1018 m-3

Mag. flux tube

Γi

[ par

ticle

s/s•

m2

]

S [m2]

ZEXIT [m]

Ei // = 200 eV

Fig.4. Calculation results of the heat flux (a) andparticle flux (b).

instrument is also capable of rotation around the shaft,which enables to measure the direction of the plasma flowto the magnetic field line. A set of movable target plateswas also made to obtain visible spectroscopic data fromthe interactions between the plasma and the targetmaterials. As shown in Fig.6 (b), the target plates consistof a 10 cmφ carbon disk, a 10 cmφ tungsten disk, a

calorimeter array mounted on the stainless steel disk and adirectional probe. The set of target plates is installed atzEXIT = 70 cm through the horizontal viewing port. Thetarget system has a capability of changing the targetmaterial by rotating the shaft without the break of air tothe vacuum vessel and has a function measuring heat fluxand particle flux. Behavior of the visible emission fromthe interaction between plasma and the target materials isobserved with high-speed cameras.

III.B. Heat and Particle Flux Measurements

III.B.1. Characteristics of ICRF Produced Plasmas

In typical hot-ion-mode plasmas (ne0 ~2×1018 m-3, Ti0

~5 keV), measurement of heat and particle fluxes fromthe end-mirror exit has been carried out. In the experiment,ICRF wave heating of 150 kW, 190 ms for plasmaproduction/heating are applied to the initial plasmainjected from the east side plasma gun.

Figure 7 shows the angle dependence of ion fluxdensity measured with the directional probe at zEXIT = 30cm together with that of heat flux measured with thecalorimeter by rotating the diagnostic instrument in thecase of only RF plasmas. Each result shows the similardependence, which indicate that the heat source isdominated by ions flowing out of the end-mirror exit.

In Fig. 8, the dependence of ICRF power for the ionheating in the central-cell (RF2) on the heat and particleflux density is presented. As seen in the figure, the heatflux density of 0.8 MW/m2 and the particle flux density of4×1022 /s•m2 were observed at zEXIT = 30 cm It is alsofound that the heat flux density increases almost inproportion to the RF power. However the particle fluxdensity near the axis (x = 2 cm) hardly changes withincreasing power. This indicates that increase ofperpendicular ion energy (Ei⊥) according to the increase ofRF power induces the enhancement of the energy in end-

Fig.5. Schematic view of west end-mirror vacuumvessel and the location of diagnostic equipment.

Fig.6. Photograph of the diagnostics. (a) Calorimeterand directional probe. (b) Movable target plates.

0

0.2

0.4

0.6

0.8

0

1.0

2.0

3.0

4.0

0 90 180 270

PHeat

(r = 0 cm)

Prb1:Γi

Prb2:Γi

Prb4:Γi

PH

eat (

MW

/m2

)

Γi ( × 10

22 particles/s•m2 )

Angle (deg.)

tsample

120-160 ms

Vbias

= -250 V

(Down) (North) (Up) (South)

rPrb

= 2 cm

RF only 190 ms ZEXIT = 30 cm

Fig.7. Angular dependence of ion flux and that of heatflux by rotating the diagnostic instrument.

0

0.2

0.4

0.6

0.8

1

1.2

0

0.2

0.4

0.6

0.8

1.0

1.2

60 80 100 120 140 160 180 200

PHeat

(x = 0 cm)

PHeat

(x = -5 cm)

Γi (x = 2 cm)

Γi (x = -3 cm)

P Heat (

MW

/ m

2 )

Γi (× 10

23 particles/s•m2 )

RF2 Power (kW)

RF only 190 ms ZExit

= 30 cm

Fig.8. ICRF power dependence of ion flux and that ofheat flux.

loss ions. In the periphery region, on the other hand, bothheat and particle fluxes increase with the RF power. Thisimplies that the radial distribution of the plasma flowingout of the end-mirror exit spreads out with the RF power.

Figure 9 shows the radial profiles of heat andparticle flux densities measure at different positions (zEXIT

= 30 and 70 cm). Each position is normalized at thecentral-cell midplane position according to the shape ofthe magnetic flux tube. In both heat and particle fluxprofiles, it is found that a good agreement is obtained.From the above results, it is confirmed that the axialdistribution of the heat and particle fluxes from the end-mirror exit is governed by the magnetic fieldconfiguration in the end-mirror region.

III.B.2. Effect of the Additional Plasma Heating

It is well understood that the ECH for potentialformation (p-ECH) produce electron flow with highenergy along the magnetic field line. Here, increment ofheat flux during additional heating of ECH into the RFproduced plasma is evaluated in order to investigate theeffect of superimposing the ECH pulse.

The radial profile of the heat flux density measuredwith the calorimeter is shown in Fig. 10. Here, the ECHpulse with gyrotrons of 150~300 kW, 20~25ms areapplied to the ICRF produced plasma of 190 ms. Asshown in the figure, superimposing the ECH pulse of 300kW, 25 ms attained the maximum heat flux up to 2MW/m2 on axis. According to the increasing ECH power,the heat flux continues to increase and radial spread of theheat flux density is also enhanced. The peak value of thenet heat flux density defined as follows,

PHeatnet ECH ≡ (QHeat

with ECH-QHeatonly RF) / tECH).

Here, QHeatwith ECH is the total calorific value in the plasma

discharge with ECH and tECH is the duration of the ECHpulse. During 300 kW ECH injection PHeat

net ECH wasestimated to be 8 MW/m2. This value almost comes up tothe heat load of the divertor plate of ITER, which gives aclear prospect of generating the required heat density fordivertor studies by building up heating systems to theend-mirror cell.

III.C. Two-Dimensional Visible Image Measurement

Initial experiments of plasma irradiation to the targetinstalled at 70 cm from the end-mirror exit have beencarried out. Figure 11 shows the two-dimensional imagesof visible emission due to plasma-material interactionsobtained from the high-speed camera in cases of astainless steed plate mounted with calorimeters, atungsten plate and a carbon plate. In spite of irradiatingsame flux onto the targets, observed light intensity haslarge difference in each material. From the time evolutionof visible image captured with the high-speed camera, it isobserved that a rapid reduction of light emission during p-ECH in all cases of material. This indicates that theplasma confining potential produced by p-ECHsuppresses the end-loss ions and reduces the lightemission during the p-ECH injection.

Radial distribution of brightness in visible emission isshown in Fig. 12 (a). There is a significant difference inthe peak intensity between materials. However, the shape

0

0.5

1

1.5

2

2.5

0

0.2

0.4

0.6

0.8

1.0

15 10 5 0 5 10

PHeat

zEXIT

= 30 cm

PHeat

zEXIT

= 70 cm (x 10)

Γi z

EXIT = 30 cm

Γi z

EXIT = 70 cm (x 5)

PH

eat (

MW

/m2 )

Γi ( × 10

23 particles/s•m2)

rcc

(cm)

RF only 190 ms ZExit

= 30 cm / 70 cm

Fig.9. Radial profiles of ion flux and the heat flux.Each position is normalized at the central-cellmidplane position

0

0.5

1

1.5

2

2.5

-8.0 -6.0 -4.0 -2.0 0

+ ECH-p 300kW, 25ms

+ ECH-p 230kW, 20ms

+ ECH-p 150kW, 20ms

only RF 150 kW, 190ms

PH

eat (

MW

/m2 )

X (cm)

HWHM4.5 (cm)

2.8 (cm)3.0 (cm)

zExit = 30 cm

Fig.10. Radial profile of heat flux in the case withsuperimposing the ECH pulse into RF plasma.

Stainless steel (216287) Tungsten (216288) Carbon (216289)

Y

X

Plasma

(a) (b) (c)

Fig.11. Comparison of 2-D visible image of emissionfrom different target materials. (a) Stainless steel,(b) Tungsten, (c) Carbon.

of each profile has no remarkable change. The intensityprofile along the z-axis, on the other hand, reveals asignificant difference as shown in Fig. 12(b). The FWHMvalue in the case of the tungsten plate is measured to belarger than that in the case of carbon by more than threetimes. This result is explained from the reason that energyreflection coefficient of tungsten is much larger than thatof carbon [15], which causes the enhancement ofpenetration of reflected hydrogen atoms from the plate.

IV. FUTURE PLAN FOR THE DIVERTORSTUDIES

In the Plasma Research Center, the followingresearch plan has been considered in near future. As aresearch plan for E-Div., based on the above shownresults, the following experimental issues are planned:(1) Installation of the divertor experimental-cell (D-cell)

in the end vacuum chamber and generation of strongplasma-surface interactions and realization of plasmadetachment using gas injection,

(2) Installation of test material pieces into D-cell and highthe heat flux plasma irradiation and surface analyses,

(3) Divertor pumping experiment using the large capacityhelium cryopumping system.

As to A-Div. project, detailed design of the coilsystem for the axisymmetric divertor configuration and ofplasma heating systems has been carried out based on theFokker-Planck simulation [16].

V. SUMMARY

Divertor simulation study has been started as the newresearch plan, by the use of a large tandem mirror device.In the research plan, two divertor devices are in progress.One has an axi-symmetric magnetic field configurationwith separatrix structure and the other is a high heat fluxdivertor simulator by using an end-mirror exit. Theexperiment of the generation of high heat and particle fluxis successfully performed at an end-mirror exit ofGAMMA 10. In standard hot-ion mode plasmas, the heatflux density of 0.8 MW/m2 and the particle flux density of4×1022 /s•m2 were observed at 30 cm downstream of the

end-mirror exit. It is confirmed that the heat flux densityincreases in proportion to the applied RF power.Superimposing the ECH pulse induces a remarkableenhancement of heat flux and a peak value in the net heatflux density of 8 MW/m2 was attained during the ECHinjection, which almost comes up to the heat load level ofthe divertor plate of ITER. Angular and radial profilemeasurement of heat and particle fluxes showed that theheat source is dominated by ions flowing out of the end-mirror exit and that the end-loss flux is governed by themagnetic field configuration in the end-mirror region. Aninitial stage of plasma irradiation experiment has beenstarted and a significant difference between materials isconfirmed in the behavior of visible emission from thetarget plates. The above results give a clear prospect ofgenerating the required performance and providing usefulinformation for divertor studies by building up the plasmaheating and diagnostic systems to the end-mirror cell.

ACKNOWLEDGMENTS

The authors would like to acknowledge the membersof the GAMMA 10 group, University of Tsukuba for theircollaboration in the experiments. This work wasperformed with the support by the bidirectionalcollaboration research program (NIFS09KUGM037).

REFERENCES

1. J. Roth et al., J. Nucl. Mater. 390-391, 1 (2009).2. W. Fundamenski et al., ibid. 10 (2009).3. G. F. Matthews J. Nucl. Mater. 220-222, 104 (1995).4. D. M.Goebel et al., Nucl. Fusion 28, 1041 (1988).5. Y. Hirooka et al., J. Vac. Sci. Technol. A8, 1790

(1990).6. S. Masuzaki et al., Jpn. J. Appl. Phys. 29, 2853

(1990).7. N. Ohno et al., Nucl. Fusion 41, 1055 (2001).8. B. Koch et al., J. Nucl. Mater 290-293, 653 (2001).9. M. Inutake et al., Phys. Rev. Lett. 55, 939 (1985).10. T. Tamano, Phys. Plasmas 2, 2321 (1995).11. Y. Nakashima et al.,Vacuum 41 No.4-6, 1561 (1990).12. I. Katanuma, Y. Tatematsu, K. Ishii, T. Tamano, K.

Yatsu, J. Phys. Soc. Jpn. 69, No.10, 3244 (2000).13. Y. Nakashima et al., J. Nucl. Mater. 196-198, 493

(1992).14. Y. Nakashima et al., J. Nucl. Mater. 220-222, 580

(1995).15. R. Behrisch and W. Eckstein, Physics of Plasma-

Wall Interactions in Controlled Fusion, pp. 413-438,D. Post and R. Behrisch Ed., Plenum Press, NewYork (1986).

16. I. Katanuma, Y. Kiwamoto, K. Sawada, S. Miyoshi,Phys. Fluids, 30, No.4, 1142 (1987).

(a) (b)

0

0.2

0.4

0.6

0.8

1

1.2

120 140 160 180 200 220 240

Stainless steel

Tungsten

Carbon

X-pixel (Z-axis)

Brig

htne

ss (R

el. U

nit)

21.4 pixel 5.8 pixel7.0 pixelFWHM value

0

50

100

150

200

250

300

50 75 100 125 150

Stainless steel Tungsten Carbon

Brig

htne

ss (R

el. U

nit)

Y-pixel (Y-axis)

Fig.12. Spatial distribution of 2-D image of thevisible emission from the target. (a) Radialprofile. (b) axial profile.