STATUS AND PLAN OF GAMMA 10 TANDEM MIRROR PROGRAM T. Imai 1, M. Ichimura 1, Y. Nakashima 1, I. Katanuma 1, M. Yoshikawa 1, T. Kariya 1, R. Minami 1, Y. Miyata 1, Y. Yamaguchi 1, R. Ikezoe 1, T. Shimozuma 2, S. Kubo 2, Y. Yoshimura 2, H. Takahashi 2,, T. Mutoh 2, K. Sakamoto 3, M. Mizuguchi 1, M. Ota 1, H, Ozawa 1, K. Hosoi 1, Yaguchi 1, R. Yonenaga 1, Y. Imai 1, T. Murakani 1, K. Yagi 1, T. Nakamura, H. Aoki 1, H. Iizumi 1, T. Ishii 1, H. Kondou 1, H. Takeda 1, N. Ichioka 1, S. Masaki 1, T. Yokoyama 1, and GAMMA10 group 1 Plasma Research Center, University of Tsukuba, 305-8577 Tsukuba, Japan 2 National Institute of Fusion Science(NIFS),Toki, Japan 3 Japan Atomic Energy Research Institute(JAEA), Naka, Japan Recent progress and near future plan of GAMMA 10 efforts are presented. With high power plug electron cyclotron heating (ECH) up to ~ 400 kw, the ion confining potential of more than 2 kv was confirmed. The drift type low frequency fluctuations were suppressed by the positive radial electric field produced by plug ECH. It is found that the efficient EC heating on mirror devices from a strong B field side requires the minimization of the stray microwave in addition to the 100% X-mode excitation to avoid the enhancement of the ion loss. The development of a gyrotron, the key tool of these ECH experiments, has been made in collaboration with NIFS (National Institute for Fusion Science), More than 1.5 MW for more than 1s has been demonstrated at 77GHz. The plan of the boundary plasma research program with modification of GAMMA 10 is in progress. The new program includes the physics and technology studies of the divertor and SOL plasmas and PWI relevant to torus plasmas like ITER. The high heat flux experiments using the open end mirror throat has been started and we have obtained successful preliminary data, which include the heat flux of 8 MW/m 2. I. INTRODUCTION The GAMMA 10 shown in Fig. 1 is the world largest tandem mirror device (27 m), featuring the MW level high power heating systems (NBI, ICRF and ECH) and various diagnostics (Gold neutral beam probe, SX, Laser scattering, various spectroscopy, microwave and particle analyzers). The plasma confinement is achieved by a magnetic mirror configuration as well as positive and negative potentials at the plug/barrier region formed by ECH. The main plasma confined in the central cell of GAMMA 10 is produced and heated by ion cyclotron range of frequency (ICRF) waves. The typical electron density, electron and ion temperatures are about 2 10 18 m -3, 0.1 kev and 5 kev, respectively and most of the stored energy of this plasma is the contribution of the ions. 1 The control of an internal and/or edge transport barrier (ITB and ETB) is a key to improve confinement and edge plasma/wall heat load reduction. A radial electric field structure is said to play an important role in this barrier. Therefore, studies of effects of radial electric field structure on transport are crucial issues for fusion plasma research. Mirror devices having open magnetic-field lines provide advantages for the control of radial potential structures through the modification of axial particle-loss balance by end-plate biasing and/or by ECH. 2 Therefore, mirror-based systems enable the experimental study of the influence of the electric field shear or sheared flows on fluctuations and the associated anomalous cross-field transport in magnetized plasmas. This is one of the main research subjects of the GAMMA 10. High power ECH for high electron temperature which is the long term issue for mirrors is another. A gyrotron for ECH is the main tool for plasma heating and control. The development of the high power gyrotron is major hardware efforts in GAMMA 10, which have been carried out as the joint program with NIFS in collaboration with JAEA and TETD (Toshiba Electron Tubes& Devices Co.). More than 1 MW power has been obtained at 77GHz LHD gyrotron. 3 Fig. 1. GAMMA 10 tandem mirror device.
The new program to simulate the divertor plasma which utilizes the mirror advantages have been started, in addition to the core confinement studies as the mainframe 6 year work plan. Since the boundary plasma physics is key to sustain the steady-state fusion reactor plasma, the divertor plasma control and plasma wall interaction are urgent issues for ITER and fusion research. High heat flux generated at the open end of the GAMMA 10, which we call E-Divertor is relevant to the fusion reactor peripheral plasma. This is one of the uniqueness of the GAMMA 10 E-Div. The other frontier concept is Anchor Divertor simulator (A-Div.), where axisymmetric divertor coils are installed to make the separatrix configuration. Using these two divertor simulators, we could contribute to the various aspects of fusion divertor plasma issues from both physics and technology. This paper describes these GAMMA 10 recent efforts including future plan in the section II ~ IV and the conclusion is given in the section V. II. SUPPRESSION OF THE DRIFT TYPE FLUCTUATIONS BY ECH field, as seen in Fig. 3. In the range of 100-350 kw, qualitative profile is similar with P-ECH and the fluctuations are suppressed. Fig. 2. Time evolutions of the potential and its fluctuation before and after the P-ECH application. II.A. Potential formation and Suppression of the Fluctuations The intensive studies have been made on the potential formation and the resultant effect of radial electric field on the drift type fluctuations, particle transport and stored energy for the study of core plasma transport improvement and control, which are common issues for magnetic confinement. It was found previously that the plug ECH (P-ECH) efficiently produces the plug potential for the mirror axial ion confinement improvement. 4 The cross check of the high plug potential formation has been done using ELECA (End Loss Energy Component Analyzer) measurement. 5 The high plug potential of 2.3-2.5 kv has been confirmed by the ELECA with the 380 kw gyrotron power output, which expands the power scaling of the ion confining potential formation. The formation of the plug potential affects the potential on the axis of the central cell and it suppresses the drift type wave through change in radial electric field or potential profile. The Fig. 2 shows the typical time evolution of the potential fluctuation level and the central potential near the axis with P-ECH measured by the Gold neutral beam probe(gnbp). 6 It is clearly seen the increase of the central potential on the application of the P-ECH and the following decrease in the drift type fluctuation. The power dependence of the fluctuation level has been also investigated. The radial profiles of the electron density fluctuations, the potential and its fluctuations are measured by using a multi-channel interferometer and GNBP in the central cell. It seen that the potential profile changes from concave shape without P-ECH to the convex one with P-ECH, namely, from negative to positive electric Fig. 3. The radial potential profile before (ICRF only) and during P-ECH. II.B. Effect of ECH produced Radial Electric Field The figure 4 shows the dependence of the fluctuation level on the radial electric field, where the data with various plug ECH power are plotted including those without P-ECH. The radial electric field (E r ) is controlled by the plug ECH located near the final mirror throats. It is seen that the ECH produced positive electric field suppresses the fluctuations, while the fluctuations are excited in the state of the negative electric filed. Since the positive radial electric field is opposite to the diamagnetic direction, it is found that the change of the radial structure of the drift velocity and/or the drift velocity reduction by the EXB drift suppresses the drift type fluctuations, indicating the EXB flow and/or its shear flow take a key role in the fluctuation suppression and hence the anomalous transport. 7,8 The results also show the correlation of the particle flux driven by these fluctuations
and the stored energy change. We are preparing to get more accurate GNBP measurements of the E r by multichannel signal detectors to investigate which of them is the main force to suppress the fluctuations. is seen that ΔSX normalized by the SX just before the C- ECH (SX 0 ) increases in any polarization and the maximum is obtained in the case of 100% X-mode, as expected. The increase of the DM is obtained in case of the 100% X- mode only. Fluctuation Power (Rel. Unit) 1 0.1 w/o P-ECH 100 kw with P-ECH 100 kw w/o P-ECH 250 kw with P-ECH 250 kw w/o P-ECH 300 kw with P-ECH 300 kw 0.01-30 -20-10 0 10 20 30 Radial electric field (V/cm) Fig. 4. The dependence of the drift type fluctuations on the radial electric field, where various plug ECRH power discharge are included III. CENTRAL CELL ECH OPTIMIZATION AND ELECTRON TEMPERATURE MEASUREMENTS III.A. Central Cell ECH One more important issue of core mirror plasma is electron heating. It is often said to be difficult to obtain high electron temperature due to the high heat conduction loss along the axis. In the previous central cell ECH ( C- ECH) experiment in GAMMA 10, the electron temperature of ~500 ev estimated by SX measurement had been reported. 9 But it contains the ambiguity of the high energy component effects. To get efficient electron heating, the heating performance of C-ECH has been studied with various conditions. The Fig. 5 is the schematic view of the C-ECH configuration, using two mirrors (M1 and M2) and 28 GHz 500 kw gyrotron has been used for microwave source. Changing the position, polarization and power level of the C-ECH, heating performance and transport effect have been studied. The vertical position had been previously scanned and the best position was determined. 10 This time, the position of the absorption was scanned horizontally with movable M1 mirror The best position is on the axis as expected. In case of the far off axis heating, diamagnetic signal (DM) decreased much in spite of the soft X-ray signal (SX) increase. The dependence of the DM increase (ΔDM) and SX increase (ΔSX) on the X-mode (the extraordinary mode) ratio of the injection microwave has been investigated. The X-mode ratio is controlled by the polarizer installed on the transmission line. The results are shown in Fig. 6, where the C-ECH power was 100 kw. It Fig. 5. The two mirror antenna system (M1 and M2) used in GAMMA 10 ECH experiment. An ion sensitive probe (ISP) is placed at the central limiter shade on the position of 20.9 cm radius. The ISP is the electrostatic ion energy analyzer excluding the electron by the barrier making use of Larmor radius. The ion current of the ISP is also shown in Fig. 6 with closed circles. It is seen that the ISP currents increase linearly with O-mode (the ordinary mode ) ratio. The DM decreases with O-mode ratio even if the electron heating (SX increase) is observed. Since the some part of the hot ion power is lost through the electron drag, the DM increase is expected by the Te increase. The DM decrease in spite of the SX increase seems to be contradictory. The increase in the ISP current is indicative that the enhancement of the hot ion loss occurs with O-mode power. Fig. 6. The dependence of SX increase (closed diamond), diamagnetic signal (open circles) and Ion sensitive probe currents (closed circles) on the X-mode ratio. To see what happen in O-mode injection, the time behaviors of the SX profiles before and during the C-ECH
in case of the 100% X-mode and 0% X-mode are shown in Fig. 7. It is clearly seen that the axisymmetric electron heating occurs in the case of 100% X-mode (the left hand side figure) but not in the case of 0% (right hand side). The non-axisymmetric heating gives the non-axisymmetric potential profile and/or the plasma shift, which is considered to deteriorate the hot ion confinement through the neo classical ion loss and/or the CX loss. From the above results of the C-ECH, it is concluded that the axisymmetric heating is important and O-mode and stray microwave due to the diffraction must be minimized for the efficient ECH, since these power could be the source of the non-axisymetric heating. Fig. 7. The time evolution of the radial profiles of the SX intensities in the case of (a) 100% X-mode and (b) 0% X-mode. 1 0.8 0.6 #215655 DM 5 4 3 A Thomson scattering system for electron temperature measurements has been installed to determine the EC heated electron temperature, since it is difficult to estimate bulk Te by the soft X-ray measurement due to the high energy component. It is also important to see the electron temperature accurately for divertor simulator experiment. We plan to measure the electron temperature by using a Thomson scattering system in GAMMA 10 in collaboration with NIFS. In recent years, the direct electron heating experiments by central cell electron cyclotron heating (ECH) have been carried out. The highest electron temperature, about over 500 ev, was estimated by a soft x-ray measurement. We will be able to execute a crosscheck with the Thomson scattering system for more reliable data evaluation. A high power YAG laser (1064 nm, 2 J/pulse), a focusing lens (f = 2 m, φ50 mm), collection mirror (φ600 mm, R = 1200 mm), bundled optical fiber (input 2 3.5 mm, output φ3 mm, length of 5 m) and the NIFS 5 channel polychromator with avalanche photodiode are used in our system, where 90 degree Thomson scattering is employed. This system will be able to measure the electron temperature range from 0.02 to 10 kev, and its measureable radial range is about ±200 mm with space resolution of about Δd ~ 20 mm. The time resolution of the system is about 10 Hz. The schematic view of the GAMMA 10 Thomson scattering system is shown in Fig. 9. Noise reduction and the adjustment of system is almost finished. Recently we just began to obtain the scattering signals. 0.4 0.2 ECH SX 0 0 50 100 150 200 Time[ms] 250 2 1 Fig. 8. Time evolution of the diamagnetic signal (DM) and SX intensity (SX) with one mirror efficient antenna system. The new one mirror antenna has been installed, which has less diffraction loss compared with the two mirror antenna in Fig.5. The preliminary heating result is shown in Fig. 8. The DM increase is improved compared with the previous one, but there still seems to be hot ion loss. We are preparing the T e measurement by SX analysis and Thomson scattering measurements together with these optimization of the C-ECH. But the analysis and adjustment of SX and Thomson system to determine the T e are underway as in the next subsection and will give the data in near future. III.B. Thomson Scattering measurements Fig. 9 Schematic view of GAMMA 10 Thomson scattering system. IV. THE NEW DIVERTOR SIMULATION PROGRAM AND PRELIMINARY RESULTS A new program has been started to study the boundary plasma for developing the control physics of divertor plasma of torus systems, aiming at reducing the heat flux to the divertor plate and impurity back flow. In ITER
tokamak plasmas, the heat load to the divertor materials is expected to be more than 10 MW/m 2 during several hundreds sec. In order to solve such problems, a large number of so-called divertor simulators have been developed and carried out the divertor simulation experiments. 11-14 As the open end magnetic field is similar to the scrape off layer plasma of the magnetic divertor configuration of the torus systems, it is appropriate to use the open end section as the divertor simulator (E-divertor). In addition, divertor coils are planed to be installed on the anchor section (A-divertor) to study boundary plasma near the separatrix. Fig. 11. The ECH power dependence of the thermal flux density of three radial points at the axial position 30 cm from the mirror throat. The shaded region is the target of the heat flux more than 10 MW/m 2. (a) (b) Fig. 10. (a) Schematic view of the divertor simulator using the end cell region of the GAMMA 10. (b) The picture of the plasma flow near the mirror throat, where the divertor target will be placed. IV.A. E-Divertor As seen in Fig. 10 (a) the most simple application of the GAMMA 10 mirror plasma is to use the axial loss as the particle and heat flux simulator. 14 It is expected to produce the ITER relevant divertor plasma, making use of the advantage of a large-scale open magnetic field system with many plasma production and heating devices of the same scale of present-day fusion devices, such as radiofrequency (RF) wave, ECRH microwave and neutral beam injection as seen in Fig. 10 (a). 15, 16 Particle and heat flux is expected to be 10 22-10 24 /sm 2 and 1-10 MW/m 2 from the estimation of the bounce averaged Fokker-Planck code, 17 assuming the plug/barrier mid-plane density of 10 19 m -3 and the ion temperature of 100 ev. The investigation of plasma flow from the end-mirror exit of GAMMA 10 is carried out to examine its performance relevant to the divertor simulation studies. The picture of the end plasma flow observed from the port near the west mirror throat is shown in Fig. 10 (b). A simultaneous measurement of heat and particle fluxes has been carried out by using a set of calorimeter and directional probe installed at the west end-mirror exit. It is confirmed that plasma production with ICRF wave of 400 kw, ~0.2 sec and ECH of 300 kw, 0.025 sec has achieved the time-averaged heat flow of 1 MW/m 2 over the diameter range of ~8 cm. The peak heat-flux of 8 MW/m 2 was obtained during the ECH injection. It continues to increase with ECH power and is expected to be achieved up to the level of 10 MW/m 2 within the available power of ECH ( 450 kw), as seen in Fig. 11. This value almost corresponds to the heat load of the divertor plate of ITER. Angular dependence of ion-flux density and that of heatflux show the similar dependence in only RF plasmas, which indicates that the heat source is dominated by ions flowing out of the end-mirror exit. It is also observed that ion-flux density has a tendency to increase with the RF power. The above results give a clear prospect for attainment of the objectives (>10 MW/m 2, 10 23-10 24 H/s m 2 ) for divertor studies by building up heating systems to the end-mirror cell. IV.B. Future Plan and A-Divertor It is also planed to introduce the divertor coils on the GAMMA 10 (A-divertor). The conceptual view of the divertor section is shown in Fig. 12. By the installation of these coils, it is possible to study the parallel and cross field boundary plasma flow, simulating the toroidal
divertor plasma. The expected plasma parameters are one order less than those of the E-Div., but it complements the physics, which is not possible in the E-Div. experiments, such as the cross field flow effect and stochastic behavior of the particle flow. Plasma stability and magnetic field configuration of the A-Div. have been investigated and the arrangement of the new divertor coils has been studied to be consistent with the present system. TABLE I. Design parameter of 1MW Gyrotrons 28GHz Gyrotron 77GHz Gyrotorn for PRC(Tsukuba) for LHD(NIFS) E39200 E3988 Frequency 28GHz 77GHz Output Power 1MW 1.5MW 1.2MW 0.3MW Pulse Width 1s 2s 10s CW Output Efficiency 35% (W/O CPD) 50% (with CPD) Beam Voltage 80kV 80kV Beam Current 40A 60A MIG triode triode Cavity mode TE 8,3 TE 18,6 Mode Converter Built-in Built-in Output mode Gaussian like Gaussian like Output Window Sapphire CVD Diamond Collector W/O CPD Depressed Collector Weight 650kg 800kg IV.A. Development of 77 GHz gyrotron for LHD Fig. 12. The conceptual view of the planned anchor divertor section. The two divertor coils with opposite currents are planned to be installed. In the new program, by using the two simulators, we plan to challenge the following research; (a) Plasma-PFC interaction and impurity transport in the divertor region (surface physics under ELM-like heatflux environment). (b) Steady-state sustainment of detached plasmas and its physical mechanism (atomic-molecular processes in the divertor plasma, etc.). (c) Heat and particle control in pedestal and divertor region (SOL plasma physics). (d) Divertor pumping experiment using the large capacity cryopumping system. V. GYROTRON DEVELOPMENTS ECH is the key tool for potential control and electron heating. For the new program of GAMMA10, it is essential to upgrade the ECH power. For this purpose, MW power gyrotrons at 28 GHz and 77GHz have been started and the development is undergoing efficiently in the joint program of NIFS and University of Tsukuba with the collaboration of JAEA and TETD, 3 The design parameters of the most recent 77 GHz tube aiming 1.5 MW with the feedback of the previous studies and 28 GHz 1 MW tube are shown in the TABLE I. The major issues of the tube are the heat due to ohmic and dielectric loss, collector and diffraction loss. TE 18,6 cavity, synthetic diamond window and depressed collector, where the ITER gyrotron technologies are incorporated, are effective against these issues except the diffraction. In lower frequency tubes like 77 GHz or 28 GHz one, less convergence of the RF beam at the window as well as the diffraction (stray RF) are the key points due to larger wavelength. The first and second tubes demonstrated the 1 MW for 5 s. The long pulse operations were terminated in around 60 seconds at 300 kw due to out-gassing by the some internal component heating and their power was limited by α ( = v / v ) dispersion. In the third tube, internal mode converter and mirrors are carefully designed to optimize the shape minimizing the beam size at the window and the diffraction, which makes the edge field of the window to be 1/3 and the transmission loss from the the mode convertor to the window to be 3/4. The electron beam and cavity design of the third tube is also improved to increase the power above 1.5 MW. To suppress the effective α decrease in high current, the cathode area is enlarged by 30% and the laminar flow has been improved by the deeper angle of the cathode plane, which is expected to make beam quality better. The cavity Q is optimized for 1.5 MW with lower α and is 987. The maximum efficiency obtained is ~ 50% at 640 kw with spent electron beam energy recovery (CPD : Collector Potential Depression). After 1 MW short pulse test in Tsukuba gyrotron test stand, the tube has been tested in LHD ECH system. As the result, we have obtained the 1.5 MW for more than 1 s in the soft excitation mode. Fig. 13 shows the 1.5 MW, 1.5 s operation in the LHD gyrotron system. The output power decreases gradually, corresponding to the beam current decrease due to the cathode cooling. The maximum duration with 1.5 MW, 300 kw and 200 kw are 1.6 s, 7 min and 21 min, respectively, and the CW operation is limited by the outgas from MOU (Matching Optics Unit). A total power of more than 3 MW with 3 tubes has been
injected into the LHD plasma and good heating results have been obtained.18 parameters, which indicates the α decrease in the high current region. Fig. 14 The picture of the 28 GHz 1MW gyrotron for GAMMA 10, which demonstrated the 1 MW power Fig. 13 Time evolutions of waveforms in the 1.5 MW, 1.5 s. operation of the 77 GHz #3 gyrotron. P, V and I are output power, voltage and current. The subscripts, A, B and C denote anode, body and collector, respectively. II.B. Development of 28 GHz gyrotron for GAMMA10 The 1 MW, a few sec., 28 GHz gyrotron development for GAMMA 10 and other low B field device ECH source has been started, based on the experiences of both 28 GHz 500 kw and 77 GHz 1 MW tubes in close collaboration with JAEA and NIFS. The design parameters are shown in the TABLE I. The cavity mode has been chosen as TE8,3, from the consideration of the beam current density, which is ~3A/cm2 at 40A.19 The design of the inner mode converter is one of the key design point of the 28 GHz gyrotron, since the reduction of the stray RF in low frequency gyrotron is far more important than the other higher frequency tube. The mode conversion efficiency of ~ 95% has been obtained by taking the larger body diameter for larger mirrors than those of ITER, LHD and JT-60 gyrotrons. The experimental output performances aand the calculated efficiencies are shown in Fig. 15, where the two calculated efficiencies of α = 1.0 and 1.4 are plotted. An experimental output of more than 1 MW has been obtained with 40 A beam currents. The maximum efficiency of 40% is obtained with 16 A beam current but the efficiency experimentally obtained decreased with the beam current increase. The obtained power is consistent with the calculation taking the actual α value estimated from beam Fig. 15 Output power of 28 GHz 1 MW gyrotron vs. beam current. Closed and open circles indicate the experimental output and efficiency. The open squares and triangles are calculated efficiencies with a = 1.4 and 1.0. VI. CONCLUSIONS Recent progress and near future plan of GAMMA 10 are described. The recent efforts of the GAMMA 10 research has put emphasis on the physics study of the application of the plasma potential control and open end advantages. The drift type low frequency fluctuations were suppressed by the positive radial electric field produced by plug ECH. It is found that the efficient EC heating requires the minimization of the stray microwave in addition to the
100% X-mode excitation to avoid the enhancement of the ion loss. The development of a gyrotron, the key tool of these ECH experiments, has been made in collaboration with NIFS, More than 1.5 MW for longer than 1s has been demonstrated at 77GHz. The plan of the new mirror program with modification of GAMMA 10 is in progress. The new program includes the physics and technology studies of the divertor and SOL plasmas and PWI relevant to torus plasmas like ITER. The high heat flux experiments using the open end mirror throat has been started and obtained successful preliminary data, which include the heat flux of 8 MW/m 2. which gives the promising prospect to achieve ITER relevant heat flux with high B field, moderate ion temperature resembling the boundary condition of the reactor grade plasma. ACKNOWLEDGMENTS The authors thank the members of the GAMMA 10 group of the University of Tsukuba. This work is partially supported by NIFS Collaborative programs (NIFS04 KU- GM009, NIFS08KUGM030, NIFS07KUGM009, NIFS09 - KUGM032, NIFS09KUGM034 and NIFS09KUGM037). REFERENCES 1. S. Miyoshi et al., Plasma Phys. Controlled Nuclear Fusion Res. 1990 (IAEA, Vienna, 1991) Vol. 2, 539. 2. A. Mase et al., Nuclear Fusion, 31, 1725(1991). 3. T. Imai et al., Proc. 22th IAEA Fusion Energy Conf. (Geneva, 2008), IAEA/FT/P2-25. (2008). 4. T. Saito et al., J. of Plasma and Fusion Res., 81, 288 (2005). 5. K. Ishii, et. al., Nuclear Fusion, 30 (1990)1051. 6. K. Ishii, et. al., Physics of Fluids, B4 (1992) 3823. 7. M. Yoshikawa, et al., Fusion Science and Technology, 57 (2010) 312. 8. Y. Kishimoto et. al. Nuclear Fusion 40 (2000) 667. 9. T. Imai et al., Fusion Tech., 51, 2T (2007)208. 10. H. Shidara et al., Trans. of Fusion Sci. and Tech., 55, 2T, 131(2009). 11. D. M.Goebel et al., Nucl. Fusion, 28 (1988) 1041. 12. Y. Hirooka et al., J. Vac. Sci. Tech., A8 (1990) 1790. 13. N. Ohno et al., Nucl. Fusion, 41 (2001) 1055. 14. Y. Nakashima et. al., Proceeding of 19th Int, Conf. on Plasma Surface Interaction, SanDiego, May, 2010, P1-87. 15. T. Tamano, Phys. Plasmas, 2 (1995) 2321. 16. M. Ichimura et al., Phys. Plasmas, 8 No.5 (2001) 2066. 17. I. Katanuma et al., Phys. Fluids, 30 No.4 (1987) 1142. 18. H. Takahashi et. al., Fusion Science and Technology, 57 (2010) 19. 19. T. Kariya et al., Fusion Tech., 55, 2T (2009) 91.