1 EX/P1-38

Fully Non-inductive Current DriveExperiments using 28 GHz and 8.2 GHzElectron Cyclotron Waves in QUEST

H. Idei1, T. Kariya2, T. Imai2, K. Mishra3, O. Watanabe1, H. Zushi1, K. Hanada1,T. Onchi1, A. Ejiri4, K. Nakamura1, A. Fujisawa1, Y. Nagashima1, M. Hasegawa1,K. Matsuoka1, A. Fukuyama5, S. Kubo6, T. Shimozuma6, M. Yoshikawa2, M. Sakamoto2,S. Kawasaki1, H. Nakashima1, A. Higashijima1, S. Ide7, T. Maekawa8, Y. Takase4

and K. Toi6

1Research Institute for Fusion Science, Kasuga, 816-8580, Japan2Plasma Research Center, University of Tsukuba, 305-8577, Ibaraki, Japan3Interdisciplinary Grad. School of Eng. Sci., Kyushu Univ., Kasuga, 816-8580, Japan4Department of Complexity Sci. and Eng., Univ. of Tokyo, 277-8561, Kashiwa, Japan5Department of Nuclear Engineering, Kyoto Univ., Kyoto, 606-8501, Japan6National Institute for Fusion Science, Toki, 509-5292, Japan7Japan Atomic Energy Agency, Naka, 311-0193, Japan8Graduate School of Energy Science, Kyoto Univ., Kyoto, 606-8501, Japan

Corresponding Author: idei@triam.kyushu-u.ac.jp

Abstract:

28 GHz Electron Cyclotron Current Drive (ECCD) effect was clearly observed in Ohmicallyheated plasmas with feedback regulation of Center Solenoid (CS) coil current in 2nd har-monic inboard off-axis heating scenario. In non-inductive current drive experiments onlyby the 28 GHz injection, the 54 kA plasma current Ip was sustained for 0.9 sec. HigherIp of 66 kA was non-inductively obtained by slow ramp-up of vertical field Bv using the28 GHz ECH/ECCD. Non-inductive high-density/current plasma start-up, which is a keyissue for fusion reactor design has been demonstrated using 2nd harmonic ECH/ECCD.Density jump across 8.2 GHz cutoff density was observed in superposed 28 GHz / 8.2 GHzinjections. The 50 kA plasmas were sustained by the 8.2 GHz injection into the 28 GHztarget plasma if the stable plasma shaping was obtained.

1 Introduction

The Q-shu University Experiments with Steady-State Spherical Tokamak (QUEST) wasproposed at Kyushu University, and the QUEST device was constructed [1]. Sphericaltokamaks (STs) can attain a higher β than conventional tokamaks. The ultimate goalof the QUEST project is steady-state operation with relatively high beta (<10%) in

EX/P1-38 2

the ST configuration under controlled plasma wall interactions. For steady-state tokamakoperation, a plasma current drive method should be established. Electron Bernstein WaveHeating and Current Drive (EBWH/EBWCD) is an attractive candidate for sustaininga steady-state ST plasma. In ST experiments, the plasma frequency may become largerthan the electron cyclotron frequency in the operating density range because of the lowmagnetic field, and electron cyclotron waves cannot propagate into the plasma beyondthe cutoff. EBWH/CD experiments require mode conversion processes from the electroncyclotron (electromagnetic) waves to the Electron Bernstein (electrostatic) Wave (EBW).In O-X-B mode conversion experiments, the launching angle and elliptical polarizationstate should be controlled to achieve high-efficiency mode conversion into the EBW. Aphased-array antenna (PAA) system that enabled us to control the launching angle andelliptical polarization state was proposed for the EBWH/EBWCD experiments in theQUEST [2] . The CW PAA′s performance was evaluated in the QUEST vacuum vessel ata high power level [3]. RF startup and sustainment experiments using the CWPAA systemhave progressed to expand the obtained plasma current region, but their densities werestill lower than the cutoff density. The high density experiments has been conducted toconfirm the EBWH/EBWCD effect for last 3 years in the QUEST, and a new operationalwindow for sustained plasma current was observed in the higher RF incident power [3].The EBWH/EBWCD effect was observed in superposed 8.2 GHz injection into over denseOhmically sustained plasmas of 30 kA [3].

In order to conduct high density and plasma current experiments on the EBWH/EBWCD subject, Kyushu University, University of Tsukuba and the National Institutefor Fusion Science (NIFS) have begun bi-directional collaborative research on ElectronCyclotron Heating and Current Drive (ECH/ECCD) experiments using a high-frequencypower tube, gyrotron. If the high-frequency power source is available, accessible density inthe experiments increases as square of the operation frequency. University of Tsukuba hasdeveloped a 28 GHz gyrotron with 1MW output [4]. The high power facilities to operatethe gyrotron were prepared in Kyushu University, and the 28 GHz experiments has begunrecently in Kyushu University under the bi-directional collaboration framework.

This paper is organized as follows. Section 2 introduces 28 GHz ECH/ECCD systemsetup and experimental scenarios. Sections 3 describes 28 GHz ECCD experiments inOhmically sustained plasmas. Section 4 describes non-inductive current sustainment onlyby the 28 GHz injection. In Section 5, the 8.2 GHz injection experiments into the 28 GHznon-inductive plasma are explained, and the summary is finally given in Section 6.

2 28 GHz ECH/ECCD System and Experimental Sce-

narios

2.1 28 GHz ECH/ECCD System

A 28 GHz 1 MW 1 s gyrotron with TE8,3 cavity and triode magnetron injection gun(MIG) has been developed at University of Tsukuba [5]. Maximum output power of 1MW level was obtained at the beam voltage Vb = −80 kV and the beam current Ib = 40 A.

3 EX/P1-38

In addition, a high efficiency of 40% without collector potential depression was obtainedat 0.8 MW. In Kyushu University, a high voltage power supply system was prepared forECH/ECCD experiments by a 170 GHz diode MIG gyrotron on the TRIAM-1M tokamak[7]. Maximum beam voltage and current have been limited to −75 kV and 25 A at thegyrotron power supply. The operating parameter region of the 28 GHz gyrotron wasconsidered and surveyed in order to operate it with the lower beam voltage and current.A 400 kW power level capability was confirmed within the operating limits of the powersupply. A high-speed, high-voltage CW switch system that integrated multiple stages ofMOSFETs connected in series was prepared to turn the anode voltage on/off and to cutthe voltage upon overcurrent detection as well as a conventional high-voltage power supplyof 50 kV. An overcurrent detection system for the anode current and a gyrotron-tank withthe anode and body electrode insulators were prepared in addition to the gyrotron systemat Kyushu University.

Figure 1(a) shows a schematic of 28 GHz ECH/ECCD system in QUEST. Circularcorrugated waveguides of a diameter D = 0.0635 m were set with a miter-bend. Thewaveguide line length was about 11 m. The waveguide was inserted into a mid-plane portof the QUEST vessel as a launcher with electrical isolation at a major radius R ∼ 1.5m. There are no mirror systems to focus and to steer the incident beam. Since thegyrotron output field is in horizontal direction, the incident beam becomes O-mode wavein the ECH/ECCD experiments. The polarizer system was not prepared. Since a mainobject of the 28 GHz injection was high-density plasma start-up and sustainment on theEBWH/EBWCD subject, simplest transmission and launcher systems were prepared asa first step.

2.2 Experimental Scenario

Figure 1(b) shows fundamental (1st) and second (2nd) harmonic Electron Cyclotron Res-onance (ECR) frequencies, fce and 2fce , and O-mode cutoff frequencies fpe for variousdensity profiles as functions of R. Here toroidal magnetic field Bt was 0.25 T at R = 0.64m, and parabolic electron density profiles were assumed with specified central densitiesne0 in the figure. The 2fce at 28 GHz was located at the high field side, while fce positionat 8.2 GHz was near the plasma center. The O-mode cutoff density is 1×1019 m−3 at 28GHz. The incident 28 GHz wave can propagate into the plasma beyond the cutoff densityfor 8.2 GHz (∼ 8×1017 m−3).

Figure 1(c) shows intensity and N∥ profiles of the beam radiated from the simplewaveguide launcher after 1.3 m propagation corresponding to the distance between thewaveguide launcher and the 2fce position. Here N∥ was refractive index in parallel to themagnetic field. The toroidal coordinate in the profiles is in perpendicular to the propa-gating −R direction. The radiation field from the waveguide launcher was evaluated by aKirchhoff integral based on Huygen principle [6]. Although the launcher waveguide diam-eter D was oversized well for the 28 GHz wavelength, the radiated beam was significantlyexpanding due to the long propagation. The parallel refractive index N∥ became large inthe expanding beam under the condition of RNψ = const.. The toroidal refractive indexNψ was a dominant term in N∥ here. The parallel refractive index N∥ was important

EX/P1-38 4

FIG. 1: (a): Schematic of 28 GHz ECH/ECCD system in QUEST. (b): Fundamentaland second harmonic Electron Cyclotron Resonance frequencies, fce and 2fce, as functionsof major radius R. Toroidal magnetic field Bt is 0.25 T, and parabolic electron densityprofiles are assumed with specified central densities ne0 in the figure.(c): Intensity andN∥ profiles of radiated beam from a simple waveguide launcher after 1.3 m propagationcorresponding to distance between the waveguide launcher and 2fce position in (b). HereN∥ is refractive index in parallel to the magnetic field. The toroidal coordinate in theprofiles is in perpendicular to the propagating −R direction.

when the resonant condition is considered.

Since the 28 GHz resonance is 2nd harmonic, the incident O-mode wave is not effectivefor one-path absorption. However, relativistic Doppler up-shifted 1st ECR can be consid-ered with N∥ evaluated in the higher-field side than the 2nd cold ECR if there were highenergetic electrons by the 28 and 8.2 GHz injections. The resonant electron energy andvelocity can be explained with resonant ellipses in parallel- and perpendicular- velocity(v∥ and v⊥) space [7]. In the resonant ellipse, the resonant energy range is characterizedby minimum and maximum parallel energies E min

∥ , E∥max at v⊥ = 0, and maximum per-

pendicular energy E max⊥ at v∥ with (E min

∥ + E max∥ )/2. Characteristic 1st ECR energies

or velocity, E min∥ , E max

∥ , E max⊥ and pitch factor [v∥/v⊥] at E

max⊥ are plotted as functions

of R in Fig.2(a). All E min∥ and E max

∥ were appeared in positive v∥ region of the resonant

ellipse. Non-resonant region with R larger than maximum ECR position at E min∥ = E max

∥and no ray crossing region are shown in the figure. High energetic electrons more than40 keV were resonant or the incident 28 GHz O-mode due to the Dopper shifted effect inthe one-path absorption.

In main absorption at the 2nd harmonic resonance, scrambled ECR absorption de-pending on the resonance area distribution was expected by multiple wall reflections inthe incident O-mode condition. Various parallel refractive indexes of |N∥| 51 should betaken at the resonance due to ray scrambling by the multiple wall reflections. Figure 2(b)shows E min

∥ (R) and E max∥ (R) for the 2nd harmonic ECR condition with various parallel

refractive indexes of |N∥| 5 1. Since E min∥ and E max

∥ were appeared in negative and pos-itive v∥ regions of the resonant ellipse, the parameters with signs of v∥ are plotted in thefigure. The Doppler up- and down-shifted resonances were in lower and higher field sidesfor the cold resonance depending on signs of N∥ and v∥, respectively. There was an asym-metric structure stemming from the relativistic mass effect on fce between the Doppler

5 EX/P1-38

FIG. 2: (a): Characteristic 1st ECR energies or velocity, E min∥ (R), E max

∥ (R), E max⊥ (R)

and pitch factor [v∥/v⊥](R) at Emax

⊥ . (b): Contor plot of E min∥ (R) and E max

∥ (R) for 2nd

harmonic ECR condition with various parallel refractive indexes of |N∥| 5 1.

up- and down-shifted resonances. If MeV-level energetic electrons were considered, therewas the 2nd ECR in a whole of the confined plasma region.

3 2nd harmonic Off-axis ECCD effect in Ohmically

Sustained Plasmas

First 28 GHz off-axis ECH/ECCD effect was confirmed in Ohmically sustained plasmawith feedback regulation of CS coil current. Figure 3 shows time evolution of Bt, Bv, CScoil current and plasma current Ip with and without the superposed 28 GHz injection.Since the Ohmic discharges have been conducted with rather low field, the plasma currentwas started-up to the 30 kA level at the low field, and later the field was ramped up to themaximum of 0.25 T. Although the plasma current has been controlled with the feedback,it began to increase by the 28 GHz injection against retarding electric field by ramp-upof the CS coil current. Recharging phenomena in the CS coil power supply were clearlyobserved. The plasma current was fully driven only by the 2nd harmonic off-axis 28 GHzinjection.

4 28 GHz Plasma Sustainment in Low Aspect Con-

figuration

Fully non-inductive current drive experiments have been conducted with the 28 GHzECH/ECCD system in the QUEST to demonstrate high-density/current plasma start-upand sustainment with 2nd harmonic ECH/ECCD. Figure 4 shows time evolution of Ip,Bv, loop voltage, and the plasma shaping parameters of aspect ratio A, elongation κ andtriangularity δ in the inner limiter-configuration. The plasma current was ramped up athigh ramp-up rate of ∼0.5 MA/s against retarding loop voltage due to the current ramp-

EX/P1-38 6

FIG. 3: Time evolution of (a): Bt and Bv, (b): CS coil current and (c): plasma currentIp with and without superposed 28 GHz injection into Ohmically sustained plasma.

FIG. 4: Time evolution of (a): Ip and Bv, (b): loop voltage and (c): plasma shap-ing parameters of aspect ratio A, elongation κ and triangularity δ in the inner limiter-configuration.

up itself. The 54 kA current was finally obtained by ramping up Bv, and was sustainedfor 0.9 sec. The plasma shaping was almost kept at stable configuration for 1.3 sec. Themagnetic axis radius Rax and A were 0.67 m and 1.4 in a low aspect ST configuration,respectively. In Bv ramp-up experiments, the Ip of 66 kA was non-inductively attainedwith the 28 GHz injection. Hard X-ray intensities with more than 100 keV energy weremeasured. Electrons with the energy range measured by HX measurements were resonantin the off-axis configuration. The resonant area covered in the larger R region due to theDoppler shifted resonance. The driven current profile should be broader, if the energeticelectron orbits was taken into account.

5 Superposed 28 GHz and 8.2 GHz Injections

In the 54kA sustained plasma only by the 28 GHz injection, line-averaged density wasabout 5×1017, and still lower than the 8.2 GHz cutoff density. High-density plasmasbeyond the 8.2 GHz cutoff density were tried to attain by the 28 GHz injection forEBWCD experiments. Spontaneous density jump across the cutoff density was observedin the superposed 28 and 8.2 GHz injections. Figure 5 shows time evolution of Ip, line-averaged density, Hα intensity, Rax, minor radius a, β ∗

p (sum of average poloidal beta βp

and half plasma internal inductance li ) with and without the density jump. Flux signalsof inward and outward positions are also shown in the density jump case. The gas fueling

7 EX/P1-38

FIG. 5: Time evolution of (a): Ip, (b): line-averaged density, (c): Hα intensity, (d): Rax

and minor radius a, (e): β ∗p (sum of average poloidal beta βp and half plasma internal

inductance li) with and without the density jump. Flux signals of inward and outwardpositions are also shown in (f) in the density jump case.

was applied at t = 2.9 sec to increase the density. The operating parameters includingthe gas fueling were identical in both discharges. Hα intensity was kept constant, andRax and a were slightly decreased in the density jump case. Ip was once decreased, butwas recovered finally. β ∗

p was first decreased by the decrement of a, and secondarily bythe increment of Ip. The plasma was self-organized to be more stable shaping, and thenthe Ip was recovered in the high-density plasma. The flux signals showed inward shift ofplasma current distribution, suggesting from the decrement of Rax.

As superposed 8.2 and 28 GHz injected experiments, a 50 kA plasma was first startedand ramped up by the 28 GHz injection, and then the 8.2 GHz RF power was injectedinto the target plasma. Figure 6 shows time evolution of Ip, Bv, Rax, a, β

∗p and line-

averaged density. The 50 kA plasma was sustained in the superposed 28 and 8.2 GHzinjections by strong Bv at β ∗

p ∼1.5 where it was also stable in the density jump study. Inconstant Bv, the plasma was shifted outward and Ip was decreased to the 10 kA level. Bythe increased Bv, the minor radius a became small at Rax ∼0.6 m. Current driving loopvoltage was significantly observed at the Bv ramp-up, but the 50 kA current level wasnon-inductively kept later. The density began to increase spontaneously under the stableconfiguration at β ∗

p ∼ 1.5, but was not beyond the cutoff. High current non-inductiveEBWCD experiments are planned in spontaneous or gas fueling over dense plasmas underthe stable plasma shaping.

6 Summary

Kyushu University, University of Tsukuba and NIFS have begun bi-directional collabora-tive research on ECH/ECCD experiments using the high-frequency power tube, gyrotron.The 28 GHz ECCD effect was clearly observed in Ohmically heated plasmas with feedbackregulation of the CS coil current in 2nd harmonic inboard off-axis heating scenario. Theplasma current was fully driven only by the 2nd harmonic off-axis 28 GHz injection. The54 kA plasma current Ip was sustained for 0.9 sec in non-inductive current drive exper-

EX/P1-38 8

FIG. 6: Time evolution of (a): Ip and Bv, (b): Rax, a and β ∗p , (c): line- averaged density

in superposed 8.2 GHz injection to the 50kA 28 GHz target plasma.

iments only by the 28 GHz injection, Higher Ip of 66 kA was non-inductively obtainedby slow ramp-up of vertical field Bv using the 28 GHz ECH/ECCD. Density jump across8.2 GHz cutoff density was observed in superposed 28 GHz / 8.2 GHz injections. The 50kA plasmas were sustained by the 8.2 GHz injection into the 28 GHz target plasma if thestable plasma shaping was obtained. High current non-inductive EBWCD experimentsare planned in spontaneous or gas fueling over dense plasmas under the stable plasmashaping.

7 Acknowledgments

This work was performed with the support and under the auspices of the NIFS Collab-oration Research Programs (NIFS09KUTR046/ NIFS10KUTR048 / NIFS11KUTR059/NIFS11KUTR069). This research was partially supported by Ministry of Education, Sci-ence, Sports and Culture, Grand-in-Aid for Sc. Res. (B) [21360452].

References

[1] K. Hanada, et al., Proc. of the 22nd IAEA Fusion Energy Conference (2008),FT/P3-25.

[2] H. Idei, et al., IEEJ Transactions on Fundamentals and Materials, 53 (2012) 511.

[3] H. Idei, et al., Proc. of the 24rd IAEA Fusion Energy Conference (2012), EXW/P7-31.

[4] T. Imai, et al., Proc. of the 25rd IAEA Fusion Energy Conference (2014), FIP/2-2Rc.

[5] T. Kariya, et al., J. Infrared Millim. Waves 32 (2011) 295-310.

[6] H. Idei, et al., J. Plasma Fusion Res., SERIES, 9 112 (2010).

[7] H. Idei, et al., Nuclear Fusion, 46 (2006) 489-499.