1 Overview of JT-60U Progress Towards Steady-State Advanced Tokamak

1 OV/1-1 Overview of JT-60U Progress towards Steady-state Advanced Tokamak

S. Ide and the JT-60 Team Naka Fusion Research Establishment Japan Atomic Energy Research Institute Naka-machi, Naka-gun, Ibaraki, 311-0193 Japan e-mail contact of main author: [email protected]

Abstract. Recent experimental results on steady state advanced tokamak (AT) research on JT-60U are presented with emphasis on longer time scale in comparison with characteristics time scales in plasmas. Towards this, modification on control in operation, heating and diagnostics systems have been done. As the results, ∼ 60 s Ip flat top and an ∼ 30 s H-mode are obtained. The long pulse modification has opened a door into a new domain for JT-60U. The high normalized beta (β N)of2.3 is maintained for 22.3 s and 2.5 for 16.5 s in a high β p H-mode plasma. A standard ELMy H-mode plasma is also extended and change in wall recycling in such a longer time scale has been unveiled. Development and investigation of plasmas relevant to AT operation has been continued in former 15 s discharges as well in which higher NB power ( 10 s) is available. Higher β N ∼ 3 is maintained for 6.2 s in high βp H-mode plasmas. High bootstrap current fraction (f BS)of∼ 75% is sustained for 7.4 s in an RS plasma. On NTM suppression by localized ECCD, ECRF injection preceding the mode saturation is found to be more effective to suppress the mode with less power compared to the injection after the mode saturated. The domain of the NTM suppression experiments is extended to the high βN regime, and effectiveness of m/n=3/2 mode suppression by ECCD is demonstrated at βN ∼ 2.5 - 3. Genuine center-solenoid less tokamak plasma start up is demonstrated. In a current hole region, it is shown that no scheme drives a current in any direction. Detailed measurement in both spatial and energy spaces of energetic ions showed dynamic change in the energetic ion profile at collective instabilities. Impact of toroidal plasma rotation on ELM behaviors is clarified in grassy ELM and QH domains.

1 Introduction The major objectives of the JT-60U project are to establish scientific basis for ITER and to explore new plasma domains leading to realization of an attractive fusion reactor. Toward realization of a steady state (SS) tokamak fusion reactor, it is essential to increase the fraction (fBS) of the bootstrap (BS) current relative to the total plasma current (Ip); a value fBS 70% is typically required [1], and the normalized beta (βN); βN ∼ 3.5 is required. This is so-called advanced tokamak (AT) concept. Development of AT relevant plasmas, one with weak magnetic shear (“high βp plasma”) and one in reversed magnetic shear (“RS plasma”), has been pursued. Formation of the internal and/or the edge transport barriers, that is the H-mode pedestal, (ITB and ETB) is one of effective methods achieving high parameters. Towards high βN, MHD instabilities are to be overcome. Non-inductive current drive and current profile control are key issues towards SS operation. Heat and particle handling is a key for divertor feasibility. We have made effort to investigate and integrate these segments towards our objectives. Furthermore recent modification in control system on the JT-60U facilities enables us to explore plasmas of longer pulse duration. In this paper, recent JT-60U experimental result after the 19th IAEA Fusion Energy Conference [2] are reviewed with emphasis on the extension of pulse duration in comparison with characteristics time scales in plasmas.

2 Extension of JT-60U pulse duration and machine status As described above, investigation of plasma behavior in longer time scale in comparison with various characteristics time is an important issue in the AT research. The characteristics time in JT-60U plasmas extends from the energy confinement time (τE ), order of several hundreds 2 OV/1-1 milli seconds, to wall saturation time (τwall), order of several tens of seconds. Among the characteristics times, one of the most important is so-called the current relaxation time (τR), a time scale of the toroidal current density profile (j(ρ), where ρ is the normalized flux radius) 2 to saturate. In this paper, τR is defined as µ0 <σ>a/12, where <σ>is the volume average of plasma conductivity and a is the plasma minor radius [3]. Although τR varies T depending on the electron temperature ( e), the ef- 400 fective charge (Zeff ) and so on and also location where external current is supplied, it ranges around a few to a few tens of seconds in JT-60U plasmas. 300 Therefore conventional JT-60U NB pulse length of 10 s may not be long enough in view of investi- 200 gation in j(ρ) or plasma characteristics which are closely related to the change in j(ρ). In order to en- 100 able research in such a long time scale, modification total input energy (MJ) on control in operation, heating and diagnostics sys- 0 tems of the JT-60U facility was done without ma- 1 10 100 jor hardware upgrade. As the results, the maximum duration (s) pulse length of a discharge is extended from 15 s to FIG. 1: Progress in the total input energy. Closed 65 s, the maximum duration of both parallel P-NB square corresponds to the data with open divertor that is limited by carbon bloom. Colsed circle cor- (four units) and N-NB (negative-ion source NB) in- responds to the data with the W-shape divertor but jections are extended to 30 s (formerly 10 s). Pulse before the long pulse modification that is limited by the heating duration. Open circle corresponds length of the perpendicular P-NB (seven units) is to the newly achieved result after the long pulse unchanged (10 s) except one that is used for the modification. charge exchange recombination spectroscopy (CXRS), but they can be injected arbitraly in 30 s from the beginning of the first P-NB injection. Therefore six P-NB units can be injected for 30 s. The maximum duration of RF (ECRF and LHRF) pulses are extended to 60 s. Since no major modification is done on the hardware, there are several limitations. Since the capacity of the poloidal coil power supply (motor generator) and tolerance of feeders against joule heat are unchanged, sustainable Ip, triangularity (δ), height of the divertor X-point and so on are limited. The toroidal magnetic field at the machine center (Bt0) is also limited to 3.3 T for about 30 s and 2.7 T for 65 s. The NB power per unit is reduced from about 2.5 MW to 2 MW. Heating systems require conditioning in order to extend their pulse lengths. The pulse length of the P-NBs is fully extended as planned. The pulse length of N-NB has reached up to 25 s with injection power of 1 MW. The ECRF pulse has reached 28 s and the total input energy reached 8.4 MJ using four Gyrotrons in series. A grill mouth made of carbon is newly placed

E43173 Ip flat top with divertor ~60s (a) 10 Ip 0.5 PNB (MA) 5 (MW) 0.0 0 50 I (b) P F 0 EC (kA) -50 (MW) 1.5 1.0 n (c) e 1.0 0.5 V (V) (1019 m-3) 0.5 l 0.0 0.0 01020304050 60 70 time (s)

FIG. 2: Waveforms of a JT-60U 65 s discharge (E43173). (a) the plasma current (I p) and the positive and negative NB powers (PNB), (b) the Ohmic heating coil (“F-coil”) current (IF) and the ECRF power (PEC, a.u.). (c) the line averaged electron density (n¯ e) and the surface loop voltage (V). 3 OV/1-1 at the LHRF luncher. It yet has been under conditioning, so far the maximum injection power 1.3 MW (about 60% of the previous value) and the total input energy of 16.3 MJ (before carbon mouth placed it was 10.9 MJ). Owing to the extension of heating period, the total input energy of 350 MJ has been achieved (Fig. 1). A 65 s discharge with Ip = 0.7MA ﬂat top of ∼ 65 s is obtained (Fig. 2). And a 30 s NB heated plasma has been obtained with Ip up to 1.4 MA.

3 Towards steady state sustainment of AT relevant plasmas In JT-60U, large efforts in extending various AT relevant plasmas had been continued within 10 s of heating pulse. Towards development of the ITER hybrid operation scenario [4], the long pulse experiments will largely contribute. The purpose of the hybrid scenario is to maximize ﬂuence in a year, therefore extending pulse length with help of non-inductive current drive keeping as high fusion power as possible is desired. Steady state scenario in ITER and steady state reactor is determined as a discharge fully sustained with non-inductive current drive, the bootstrap current and externally drive current(s). In practical or economical view, fBS is expected to be large enough, 50% in ITER and 70% in a reactor. Towards these steady state development, conventional NB operation, in which pulse length is shorter but higher power is available, is contributing. In those AT domain plasmas, the target or resultant current proﬁle has important meaning. The major issue here is the MHD instabilities. Towards the integration of the present developing AT relevant scenarios to ITER or reactor relevant situations, compatibility of such AT plasmas to the divertor feasibility is also a major issue.

3.1 Extension of high βN sustainment One of the major objectives in the extension of a JT-60U pulse length is to extend duration of high βN sustainment. We had succeeded in sustaining βN = 2.7 for 7.4 s [5]. By optimizing high βp ELMy H-mode plasma of Ip = 0.9 MA at Bt (the toroidal magnetic field at the plasma major radius) = 1.6 T and q95 (the safety factor at the 95% toroidal flux) ∼ 3.2, βN = 2.3 is successfully maintained for 22.3 s (Fig. 3) [6]. It corresponds to 13.1τR. A little shorter sustained duration of 16.5 s but higher βN = 2.5 is also achieved. The evolution of βN was carefully optimized in order to reach as high as possible but avoiding the neo-classical tearing mode (NTM) to appear, since the sustainable power is not enought to raise βN with the NTM existing. Although localized ECCD has been proved effective on stabilizing an NTM [7], the ECRF power that can be sustainable for 20 s is too low to apply to this duration, therefore it is not applicable. By the optimization no distinct NTM is observed for these durations. As showninthefigure, βN gradually decreases in the later phase, this is attributed to the decrease in confinement. The reason why the confinement is getting worse can be attributed to the increase in the wall recycling which appears as an increase in the Dα line brightness (Fig. 3 (c)). The

βN=2.3 for 22.3s 1.0 E44092 15 Ip 10 PNB 0.5 P-NB (MW) (MA) (a) N-NB 5 0 0 β 2 N=2.3 Te,i β Te Ti 5 N 1 (b) (keV) 0 0 19 (c) ne/10 2 Dα (m-3) 1 (arb) 0 0 0 5 10 15 20 25 30 time (s) FIG. 3: Waveforms of a high βN long susteinment discharge (E44092). (a) the plasma current (I p) and the positive and negative NB powers, (b) the normalized beta (β N) and the electron and the ion temperatures (T e and Ti). (c) the line averaged electron density (n¯ e) and the intensity of the Dα line. 4 OV/1-1 wall recycling will be discussed later in Sec. 5.1. Although the confinement gradually degrades, 2 a factor H89PβN/q95 (here H89P is the confinement improvement factor to the L-mode scaling (H89P) [8]), which is a figure of merit of the fusion performance, 0.4 is kept for 22.3 s. It 2 is noted that H89PβN/q95 ∼ 0.4 corresponds to the ITER standard ELMy H-mode scenario 2 with Q = 10, while H89PβN/q95 ∼ 0.3 corresponds to the ITER steady state scenario with Q = 5. In these discharges, fBS is ∼ 35-40%. Considering these parameters, these discharges can be categorized as the ITER hybrid operation. It should be stressed that although change in recycling slightly affects the confinement characteristics, no significant phenomenon has occurred even in such a very long duration relative to τR. These results are encouraging to the ITER hybrid operation. E042883 βN=3 for 6.2s 3 β P Not only the extension towards time axis, ef- N 2 20 1 PEC PN-NB PP-NB (MW) 0 (a) 0 β 19 fort to extend the domain into higher N has ne/10 (b) Dα 2 (m-3) (au) also been carried out. Since higher power is 0 3 Ip 1 q95 required, this experiment has been carried out (MA) (c) 0 2 (d) with conventional 10 s P-NB injection set- f 10 ting in which the maximum total P-NB power (kHz) 0 3456789101112 can be 27.5 MW and that of N-NB can be 4- time (s) 5 MW. In this experiment, NTMs are to be FIG. 4: Waveforms of a high βN = 3 discharge (E42883). avoided as well. For this, the experiment was (a) the normalized beta (βN) and the positive and negative NB powers, (b) the line averaged electron density q95 carried out at rather low of below 3. Due (n¯e). (c) the intensity of the Dα line. (d) the magnetic to the low q95 but q0 is kept around unity, q = fluctuation, no clear activity is observed during β N =3. 1.5 or 2 rational surfaces, at which NTMs can occur mostly, shift quite outwards at which the pressure gradient can be low enough so as NTMs do not to occur. Also broadening of heating profile was intended. As the result, βN = 3 is sustained for 6.2 s (Fig. 4) without distinct NTM.

3.2 Extension towards steady state operation with high fBS Nearly Full CD As mentioned before, in ITER E44104 20 5 (e) fBS P (a) steady state scenario, around 10 NB P 4 8.3s NNB (MW) 3 50% or higher is expected. With 0 q 12s f q 3 (b) 2 this range of BS, various pro- 2 1 1 N ﬁle can be consistent. That is p 0 0 1 q 1 V (c) (f) from one with quite ﬂat pro- loop 8.3s 0 ) 0.8 2

(V) j j TOT file in the core region to one that -1 0.6 BD 3 0.4 has very high q0.Aflat q profile D div (d) j 2 (MA/m 0.2 BS 1

q ∼ (a.u.) with 0 1.5-2.5 is preferably 0 0 5 7 9 111315 0 0.2 0.4 0.6 0.8 1 accepted, in view of MHD sta- time (sec) bility, alpha particle confinement, FIG. 5: Typical waveforms of a weak shear ELMy H-mode discharge with large bootstrap current fraction under nearly full non-inductive fBS, easier external j(ρ) control current drive: (a) injected P-NB and N-NB power, (b) normalized beta and so forth. In JT-60U AT rele- (βN) and poloidal beta (βp), (c) surface loop voltage and (d) deuterium vant plasma of this kind has been recycling emission at the divertor (Dα). (e) Temporal evolutions of q β profile shows current profile became stationary. (f) bootstrap, beam developed based on the high p driven and total current density profiles (jBS, jBD and jtot). ELMy H-mode [9]. By extending the pulse length of N-NB, this domain of operation is investigated [10]. Although the fraction (fCD) of the non-inductively driven current to the plasma current is not 100%, fCD > 90% with fBS ∼ of 45% is maintained for 5.8 s (2.8τR)atβN ∼ 2.4 (Fig. 5). Owing to the off-axis BS current, the q profile is almost flat but slightly reversed in the core region. The minimum in the q profile (qmin) is just around 1.5 as shown in the figure. Since Ip is not fully driven non-inductively qmin slightly decreases. However no clear NTM is observed during the 5.8 s period, probably due to that qmin stays just around 1.5. 5 OV/1-1

5.1s 1 E43046 20 12 (a) (e) 6.3s Ip P 10 10.8s NB 12.4s

(MA) 8 (MW)

0 0 q 6 2 (b) 1 4 N p 0 2 H 10 4 (c) 89 (f) 8 2 6

(keV) 4 3 div i 6.3s (d) D 2 T 2 10.8s 1 12.4s (A.U.) 0 0 4681012 0 0.2 0.4 0.6 0.8 1 Time (sec) r/a FIG. 6: Typical waveforms of a reversed shear ELMy H-mode discharge with large bootstrap current fraction under nearly full non-inductive current drive: (a) plasma current (I p) and injected NB power (PNB), (b) normalized beta (βN: solid curve) and poloidal beta (βp: dotted curve), (c) H factor (H89p) and (d) deuterium recycling emission at the divertor (Dα). Temporal evolutions of (e) q profile and (f) ion temperature profile (T i) show the current and the pressure profiles became stationary. In a reactor design, fBS of 70% or higher is expected. In order to raise fBS an RS configuration is expected to be benefitial. In JT-60U, high fBS of 80% was sustained for 2.7 s in a high confinement RS plasma (Ip = 0.8MA, Bt = 3.4 T) with fCD ∼ 100% [11]. By optimizing the similar plasma a high fBS of ∼ 75% is successfully sustained for 7.4 s with very high confinement of HH98(y,2) = 1.7 (Fig. 6) [10]. The duration corresponds to ∼ 2.7 τR. Continuous off-axis heating to maintain the ITB and the optimization of injection of on-axis counter (to the Ip direction) parallel P-NB to avoid disruption are found to be the keys for the sustainment. In the discharge, the duration is limited by the NB pulse length (10 s operation). No trial for further extension with 30 s NB pulse has been carried out yet.

3.3 MHD studies to access higher beta As described in Subsection 3.1, in positive shear (PS) discharges with q0 close to unity, the NTM is one of the most critical modes that prevent sustainable βN from approaching to the ideal MHD limit. On JT-60U, we have demonstrated that localized ECCD is effective to suppress NTMs [7]. By further study it is found that ECRF injection before the mode fully develops (early injection) is more effective than the injection after the mode gets fully developed (late injection) for the first FIG. 7: Dependence of the m/n = 3/2 mode time (Fig. 7) [12, 13]. Decrease in the current density stabilization on the EC power, magnetic fluctuation amplitude. Closed circles indicate a at the mode island as the mode grows and compensa- case where ECRF injected during the mode is tion of the lost current by ECCD are confirmed by the growing, while open triangles indicate a case Mortional Stark Effect (MSE) measurement [14]. ECRF injected after the mode saturated.

On the other hand among RS plasmas, some discharges disrupt at low βN, ∼ 1orevenlower. This disruption should be avoided by a reliable operation. We have studied these disruptions at low βN, and reported that the double tearing mode (DTM) could be a cause of them [15, 16]. By improved measurements and analysis in temperature and magnetic ﬂuctuations and the q proﬁle (or the equilibrium) it is found that this low βN disruption can also be triggered by mode coupling of the surface MHD instability driven by peripheral large plasma current and the internal mode in the RS region at the rational surface whose safety factor corresponds to the 6 OV/1-1 mode number of the surface mode [17]. Both a surface mode and a internal mode can trigger to interact with each other.

3.4 Divertor compatibility of AT plasmas 2.0 In ITER or a reactor, it is required to raise the electron density near or even above the Greenwald den- 1.5 RS sity (nGW), in order to attain enough fusion reaction. H98(y,2) H 1.0 High radiation loss is also required in order to reduce H-mode β High p (a) the heat ﬂux to the divertor. Compatibility of AT 0.5 plasmas with high density and high radiation loss has been investigated in both RS plasma and high βp H- 1.0

mode plasma with a weak positive shear [18]. In the rad 0.8 HH f RS plasmas, high confinement of 98(y,2) = 1.3 is 0.6 achieved at the high density above nGW, that is fGW (b) 0.4 (= n¯e/nGW) = 1.1, even with NB fuelling only. The 0.60.7 0.8 0.9 1.0 1.1 1.2 n /n total radiation loss is enhanced to the level greater e GW than 90% of the net heating power with high confine- FIG. 8: (a) HH98(y,2) and (b) radiation fraction as a function of n¯e/nGW. Squares : RS plasma. ment (HH98(y,2) = 1.1) at high density (fGW = 1.1) Circles : high βp H-mode plasma. diamonds : by injecting seed impurity Ne together with D2 gas ELMy H-mode plasma. Open and closed sym- into the RS plasmas. In the high βp H-mode plasmas, bols show new (after the last IAEA meeting) and HH old (before the last IAEA meeting) data. Double high confinement ( 98(y,2) = 0.96) is maintained at lines show the data with impurity seeding. high density (fGW = 0.92) with high radiation loss fraction (frad ∼ 1) by utilizing high-field- side pellets and Ar injections. These results are summarized in Fig. 8. In these plasmas, the high fGW is obtained due to a peaked density profile inside the ITB. The strong core-edge parameter linkage is observed in the high βp H-mode plasmas with pellets and Ar injections, as well as without Ar injection, where the pedestal βp is enhanced with the total βp. On the other hand, the pedestal βp is kept at small value in the RS plasma, indicating that confinement improvement is mainly attributed to the strong ITB. The radiation loss profile in the main plasma is peaked due to the impurity accumulation in both plasmas.

3.5 Development of real-time current profile control Control of safety factor profile (q(ρ)) is essential for stable sustainment of AT plasmas. There- fore, a real-time q(ρ) control system has been developed in JT-60U. This system enables real time evaluation of q(ρ) by MSE diagnostic and control of CD location by adjusting the parallel refractive index N of LH waves through the change of phase difference (∆φ)ofLHwaves between multi-junction launcher modules. Real time estimation of q(ρ) that is evaluated from the MSE measurement has been demonstrated for the first time. Real time feed back control with the LHRF system is under going.

Progress in sustainment of high βN and high fBS is summarized in Fig. 9. As shown in the figure, βN > 2 is remarkably extended beyond 10 s owing to the long pulse modification. Although the experiments have been carried out with 10 s NB pulse, progress of high fBS ∼ 75% is also eminent. As described earlier, how much the duration is extended relative to τR is a key issue. A ratio of sustained duration to τR is plotted in Fig. 10 for typical discharges. For high βN discharges, sustained duration has exceeded far beyond τR. For high fBS discharges, now the discharges have been entering into a domain where the current profile is about to diffuse. 7 OV/1-1

4 100 after 19th IAEA before 19th IAEA 80

N 3 (%) β

BS 60 2 40

sustained 1 27.5MW sustained f available PPNB 20 after 19th IAEA 14MW 12MW before 19th IAEA 0 0 0 510 15 20 25 30 0246810 duration (s) duration (s) FIG. 9: Left ﬁgure; progress of sustainment of high βN, sustained βN is plotted against sustaining period. Upper hatched belt indicates the ITER hybrid and steady state domain, while lower one indicates the ITER reference H-mode domain. Right ﬁgure; progress of sustainment of high f BS, sustained fBS is plotted against sustaining period. Closed circles indicate the results obteined before the last IAEA conference, while open circle represents the result after the conference. 100 3 80

60 2 β fBS N 40 1 20 0 0 1 10 τ duration / R FIG. 10: The sustained fBS (circles) and βN (squares) against sustained duration normalized to τ R Closed symbols correspond to the typical data obtained before the last IAEA conference while open symbols correspond to the new results.

4 Towards understanding of conﬁnement and transport in AT plasmas and more general issues As described at the beginning, recent JT-60U experiments have focused largely on extension of AT relevant issues towards time-axis. However, at the same time we have continuously kept investigation on other issues that are related to AT development as well and on more general plasma physics.

4.1 Current hole studies 3 4 (b) (a) j +j +j It has been investigated if there ex- 2 EC BD BS 3 j +j j

] BD BS+ OH 2 ]

jtot 2 2 ists some mechanism to clamp the 1 jOH current density at zero level in the 0 1 jtot

j [MA/m 0 -1 j [MA/m current hole. Though the nearly zero jEC+jBD+jBS+jOH -2 jOH -1 jBD+jBS toroidal current in the central re- t = 5.5 s t = 6.2 s -3 -2 gion (a ’current hole’) is sustained 0 0.2 0.4ρ 0.6 0.8 1 0 0.2 0.4ρ 0.6 0.8 1 for several seconds in the JT-60U FIG. 11: Radial profiles of current density in the current hole ex- tokamak [19], it has not been clear periments. The inductive toroidal electric field is negative in (a) and positive in (b). Here, jtot (solid line with a shaded belt) denotes whether the current drive source measure current density, jOH is calculated inductive current den- such as inductive toroidal elec- sity, jEC is calculated EC-driven current density, jBD is calculated beam-driven current density and jBS is calculated bootstrap current tric field (Eφ) and non-inductively density. The sum of jOH + jBD + jBS + jEC is shown by a dotted driven current (jNI) remains at zero line with error bars. level or some mechanism works to clamp the current density at zero level against the current drive source. Two kinds of experiments were performed to investigate responses to Eφ and jNI separately [20]. In the first experiment, Eφ was changed transiently, with keeping jNI inside the current hole as small as possible, by changing non-inductive current (bootstrap and EC-driven 8 OV/1-1 currents) outside the current hole. In Fig. 11, the sum of calculated jOH, jEC, jBD and jBS and measured jtot are compared for (a) negative Eφ and (b) positive Eφ cases. In both cases, the calculated current density is dominated by inductive current and is largely negative in (a) and is largely positive in (b). The measured current density, however, remained nearly zero. In the second experiment, EC current drive inside the current hole was attempted in the co- and counter-directions to the plasma current during the quasi-stationary period with sufficiently small Eφ. In both directions, the EC current drive did not change the current inside the current hole and the current hole was maintained. From these results, it has been shown experimentally for the first time that the current hole is maintained by some mechanism to clamp the current density at zero level once when the current density becomes at zero level in the central region. Some simulation results show that resistive MHD instabilities take place in the current hole and they work for the current clamp [21]. In our experiments, however, no MHD instabilities with a high frequency (1-100 kHz range) were observed. Though mini collapses with longer intervals ( 0.1 s) were observed in some discharges, it was found that these collapses are not a cause of current clamp in the current hole. Model simulation on current hole formation and sustainment has been carried out [22]. Results on formation and sustainment of a current hole that are consistent with the experiments are obtained.

4.2 Conﬁnement of energetic ions

3 α 100 befor ALE Anomalous transport of particles before ALE (a) 2.5 after ALE after ALE 10 E by collective instabilities induced by NNB ) 2 (a.u.) n -3

F 1 the energetic ions and/or MHD in- m 1.5

18 4 (b) stabilities can cause serious prob- 1 n (10 2 / F

0.5 n lems in a fusion reactor. In JT- F 0 energetic ion density 0 60U, behavior of high energy ions 0 0.2 0.4 0.6 0.8 1 0 50 100 150 200 250 300 350 400 r/a energy (keV) and their inducing instabilities have FIG. 12: Left: Energetic ion density profiles before ALE at t = been investigated by utilizing N-NB 4.642 s and after ALE at t = 4.654 s, which are estimated from the [23,24]. Newly installed energy ana- neutron emission profile measurement. The inversion of the energetic ion profiles by ALE can be seen at ρ ∼ 0.45. Right: Energy lyzer using natural diamond detector spectrum of (a) neutral particle fluxes before and after ALE, (b) the enables measurement on the energy fraction of enhanced neutral particle fluxes due to ALEs. distribution of the energetic ions [25]. Multi channel neutron emission detector brings information of spatial distribution of the energetic ions. It is found that in a weak shear plasma with higher βh (the beta of the high energy ions) a bursting mode called abrupt large-amplitude event (ALE) redistributes energetic ions from the core region to the outer region of the plasma (Fig. 12 Left). Neutral particle flux in limited energy range (100 – 370 keV) is found to be enhanced by the ALEs (Fig. 12 Right). This indicates that energetic ions in this energy range are redistributed. The energy corresponds to the N-NB ions that induce the modes.

4.3 Electron transport Transient transport experiments are performed [26]. This research is carried out in collabora- tion with the National Institute for Fusion Science (NIFS), Toki, Japan, in order to compare experimental results between a tokamak (JT-60U) and a helical device (Large Helical Device: LHD) to ﬁnd out common physics in a torus plasma or in more general sense. The dependence of χe on the electron temperature (Te) and the electron temperature gradient (∇Te) is analyzed by an empirical non-linear heat transport model. In an OH plasma and low power NB heated L- and H-mode plasmas, two different types of non-linearity of the electron heat transport are observed from cold/heat pulse propagation utilizing pellet injection and ECRF injection. It is 9 OV/1-1 found that χe depends not only on Te but also on ∇Te in JT-60U, while χe depends only on Te in LHD.

4.4 Disruption and runaway electron generation mitigation Fast plasma shut-down and mitigation of disruption are key issues for a safety operation. Pufﬁng noble gas and intense hydorogenic gas is found to be effective to reduce divertor heat load and suppress runaway electrons generation [27].

4.5 Innovative operational concept development In order to reduce the construction cost and the weight of a tokamak fusion reactor, it is ben- eficial to remove a massive center solenoid (CS). In such a case, it is necessary to develop a scheme to initiate and ramp up Ip without CS. In the last IAEA conference, we reported that Ip could be initiated and ramped up without the CS (F-coil in JT-60U), by RF (ECRF and LHRF) and induction from poloidal coils only, and high confinement high fBS plasma was obtained by injecting NB in such a plasma [28,29]. However, since some of the poloidal coils (’VT-coil’ that is used to control the triangularity) have turns inside the torus the situation was not complete CS-free. Further optimization with the inside turns of VT-coil disconnected, it is demonstrated that Ip of about 100 kA can be formed with ECRF and flux injection by the outboard coils only (genuine CS-less condition) [30].

5 Pedestal, SOL and divertor studies towards particle/heat handling The extended heating phase of a discharge has unveiled phenomenon that had not been experi- enced when the heating duration had been limited within 10 s. One of the most distinct ones is saturation in the wall retention; the wall recycling rate keeps increasing to unity. The increase in the recycling can affect plasma performance. The divertor pumping plays an important role in preventing wall saturation from degrading performance. Understanding of ELMs is a key to mitigate particle/heat load to the divertor for a steady-state operation. Characteristics of ELMs have been investigated with emphasis of the plasma toroidal rigid rotation.

5.1 Extension of high recycling ELMy H-mode -1.0 One of other important issues in long pulse operation is change in wall recycling that has a longer time scale than the current diffusion. A typical effect of increase in wall recycling can be seen, for example, in Fig. 3. The normalized beta gradually de- Minor radius (m) D creases (Fig. 3 (b)). At the same time the intensity of the α -1.5 brightness gradually increases (Fig. 3 (c)). This link between 3.0 3.5 D Major radius (m) the confinement and α intensity is often observed in JT-60U FIG. 13: Magnetic configurations long pulse discharges. Therefore, the degradation of confine- around the divertor of a high ( 0.36 ) and a low ( 0.27 ) trianglerity ment is though to be attributed to the increase in recycling. Af- plasma, shown by a solid and a bro- ter sufficient wall conditionig, typically glow discharge of sev- ken line, respectively. eral to more than ten hours, base level of the Dα intensity becomes much lower and does not increase much. However, sooner or later the Dα intensity, that is the wall recycling, is expected to become large enough to disturb desired confinement. The uncontrollable continuous increase in the wall recycling in a discharge like one shown in Fig. 3 can be attributed to a fact that such a plasma configuration is not suitable for divertor pumping in JT-60U. The divertor section of JT-60U is so-called ’W-shaped’ divertor as shown in Fig. 13 [31]. The pumping slots are located at the end of the both wings of the dome. It is necessary to set divertor legs closer to the slots in order to make pumping efficient, since neutral pressure at the slot that should be high enough 10 OV/1-1 for pumping does not increase if the legs are far off even by several centimeters. For a discharge like E44092 in which higher βN is intended, it is necessary to raise the triangularity for higher pedestal height. However, in such a plasma the hit point of the inner divertor leg tends to move upwards in JT-60U. Furthermore, even if the legs are close enough to the slots, pumping is not efficient for plasmas with low fueling such like E44092. Due to these reasons, it is very difficult to make divertor pumping efficient in a discharge like E44092 and to demonstrate that divertor pumping can suppress recycling low enough to maintain good confinement. In order to make the effectiveness of divertor pumping clearer, experiments are carried out in a high recycling ELMy H-mode plasma [32]. 19 E043918 19 E043860 3.0x10 0.12 3.0x10 0.12 Pressure ( Pa ) (a) Pressure ( Pa ) (c)

2.0 0.08 2.0 0.08 ) ) -3 -3

1.0 0.04 1.0 0.04 line-averaged line-averaged ( m electron density ( m electron density 0.0 0.00 0.0 0.00 8x1022 8x1022 (b) (d) Injected 6 Injected 6 Pumped Pumped 4 4 Wall Wall 2 Retention 2 Retention Particles ( number ) Particles ( number ) 0 0 5 10 15 20 25 30 5 10 15 20 25 30 Time ( sec ) Time ( sec ) FIG. 14: (a) line-averaged electron density, neutral pressure in the main chamber, (b) number of injected particles ( neutral beam + gas ), pumped ( divertor-pumping ), and retained in the wall for a high triangularity ( ∼ 0.36 ) plasma (w/o divertor pumping). (c) and (d) for a low triangularity ( ∼ 0.27 ) plasma (w divertor pumping). Shaded regions indicate periods of wall-pumping rate of ∼ 0.

In Fig. 14 shown is a comparison between a discharge without (left hand side) and with (right hand side) divertor pumping. When the wall is almost saturated, in other words when the wall retention is almost saturated (shown with a shade belt), without divertor pumping the electron density and the neutral pressure in the main chamber keep increasing, while with pumping the electron density stays almost constant and the pressure decreases. It should be noted here that the saturation in wall retention 2 becomes observable in JT-60U only after the ex-

19 21 tension of the pulse length. As shown in Fig. 14, ) / s x10 19 the saturation does not occur within 15 s which x10 1

was conventional pulse length in JT-60U. As ( x 10 plotted in Fig. 15, as the divertor pumping rate dt / i increases the incremental rate of the total elec- N d trons (dNe/dt, whiere Ne is the number of the 0 0 1 2 3 21 4 5 total electrons in the plasma) monotonously de- Pumping-Rate ( x 10 / s ) creases. This means that with sufﬁcient pump- FIG. 15: Increase rate of number of ions in the core ing, the electron density can be controlled even plasma as a function of divertor-pumping rate under the condition of wall-pumping rate of ∼ 0 ( open cir- if the wall is saturated. cles) and positive wall-pumping rate ( solid square ). 5.2 H-mode pedestal, ELMs and divertor compatibility Mitigation of particle/heat load to the divertor is a critical issue for a steady-state operation. Towards mitigation of ELM impact, characteristics of various ELMs have been investigated, especially on energy lost at each ELM pulse and impact of toroidal plasma rotation on ELMs by utilizing co and counter parallel NBs. The energy loss for the grassy ELM has been studied to investigate the applicability of the grassy ELM regime to ITER [33]. The grassy ELM is characterized by the high frequency periodic collapse up to ∼ kHz, which is ∼ 15 times 11 OV/1-1

50ms (e) 100 2 faster than that for type I ELM. A diver- E43942 Type I ELM (120Hz) (a) p~1.6 (b) 1 (a) (c) (d) tor peak heat ﬂux due to grassy ELMs is 50 /s) 0 5550 5600 2 2 E43941 small type I ELM and large grassy ELM (440Hz) p~1.6 less than 10% of that for type I ELMs. 1 (b)

ph/sr/m 0 19 0 2 4800 4850 This smaller ELM heat flux is caused E43942 grassy ELM (880Hz) p~1.6 (c) 1 by narrower radial extent of the collapse -50 0 4200 4250 (km/s) rotation Toroidal intensity (10 intensity 2 E43940 grassy ELM (1500Hz) of temperature pedestal. The domi- D p~1.5 1 (d) -100 0 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 nant ELM energy loss for grassy ELMs 5850 5900 Time (ms) R-Rsep (m) seems to be conductive loss, and its ra- FIG. 16: (a)-(d) Time evolution of D α signal during plasma tio to the pedestal stored energy was toroidal rotation scan (q95 ∼ 4.9 and δ ∼ 0.59). (e) Toroidal rotation profiles mapped into outer midplane measured with 0.4–1%, which is smaller by a factor charge-exchange recombination spectroscopy. Shaded region of about 10 than that for type I ELMs shows area of top of Ti pedestal. Plasma rotation profiles were changed by using different combination of NBs: (a) of 2-10%. ELM size and type can be 2CO+2perp+2NNB, (b) 2CO+3perp+1NNB, (c) 2CO+5perp changed from type I ELM to high fre- and (d) 1CO+1CTR+5perp. quency grassy ELM as increased CTR plasma rotation (Fig. 16). The complete ELM suppression (QH-mode) has been achieved using counter and perpendicular NBIs, when the plasma position was optimized. The existence of the edge fluctuations localized pedestal region may reduce the pedestal pressure, and therefore the QH-mode can be sustained for long time up to 3.4s (∼ 18 τE). A transient QH phase was also observed during the CO-NB injection phase with almost no edge toroidal rotation, which is a similar condition in ITER. Collaborative experiments between JET and JT-60U has been performed for the first time to compare the H-mode pedestal and ELM behaviour in the two devices [34]. The size and shape are adjusted to be as close as possible between the devices, except the aspect ratio (∼ 15% different). Contrary to expectations, a dimensionless match (the normalized Ramor radius, the normalized collision frequency and so on) between the devices has been quite difficult to achieve. The toroidal rotation profile and the magnitude of the toroidal ripple field in the large bore plasma used in JT-60U for the similarity experiments are identified as the main difference with JET. When ripple induced fast ion losses were reduced by factor of 2 by replacing Positive ped NBIs with Negative-ion based NBIs, higher pe providing a good match to full power JET H-modes was achieved in JT-60U.

5.3 SOL plasma and plasma wall interaction Understanding of parallel and perpendicular transport during an ELM in the SOL region is important. It is found that radial velocity of the SOL plasma expansion is 1-3 km/s at the low- field-side, which is larger than that at the high-field-side. Analysis of the first wall tiles shows that hydrogen and deuterium retention in divertor region is mostly < 0.04 in (H+D)/C ratio that is less than most of observation in other tokamaks probably due to high wall temperature (573 K) in JT-60U [35].

6 Summary With extension of JT-60U pulse length in addition to the continuous effort in conventional pulse length, successful progress has been made on development and understanding of AT relevant plasmas and their issues towards steady state operation. Sustainable duration of high βN that is comparable to the ITER advanced operation domain has been extended to 22.3 s with βN = 2.3 (or to 16.5 s with more favorable βN = 2.5). The duration corresponds to about thirteen times τR. During this very long time span no significant phenomenon is observed. This indicates ro- bustness of a high βN operation. Duration of higher βN (= 3) sustainment which is rather closer to the reactor domain has been extended to 6.2 s. DUration of nearly full-current drive with 12 OV/1-1 high fBS ∼ 45% is extended to 5.8 s. High bootstrap current fraction RS plasma was demonstrated to be maintained for 7.4 s. Long pulse (< 30 s) high recycling ELMy H-mode plasma was obtained. Effectiveness of divertor pumping was studied on that plasma and demonstrated. Compatibility of AT relevant plasmas with divertor was investigated. High confinement and high radiation were realized at high density region in both weak and reversed magnetic shear plasmas. Injecting ECRF in prior to an NTM became saturated was found to be more effective to stabilize the mode than injecting ECRF after the mode got saturated. Genuine CS-less tokamak plasma start-up was demonstrated. It was shown that no current could be driven by any scheme (inductively by toroidal electric field, non-inductively by external current drivers such as ECCD and N-NB current drive) in the current hole region. Detailed measurement of energetic ions in both real and energy spaces showed spatial redistribution of energetic ions res- onating with the collective mode that is induced by these energetic ions themselves. Energy lost at each ELM was found to be smaller in Grassy ELM domain than that in Type I ELM domain. Impact of the toroidal rotation on both Type I and Grassy ELMs were found.

Acknowledgement

The authors would like to acknowledge all people who have contributed to the JT-60U project. Reference [1] M. Kikuchi, Nucl. Fusion 30, 265 (1990). [2] T. Fujita and the JT-60 Team, Nucl. Fusion 43, 1527 (2003). [3] D. R. Mikkelsen, Phys. Fluids p. 333 (1989). [4] ITER Technical Basis, no. 24 in ITER EDA DOCUMENTATION SERIES (IAEA, Vienna, 2002). [5] A. Isayama, et al., Nucl. Fusion (2003). [6] T. Suzuki et al., this conference, EX/1-3 . [7] A. Isayama, et al., Plasma Phys. & Controlled Fusion 42, L37 (2000). [8] ITER Physics Expert groups on Conﬁnement and Transport, Conﬁnement Modelling and Database, and ITER Physics Basis Editors, Nucl. Fusion 39, 2137 (1999). [9] Y. Kamada and JT-60 team, Nucl. Fusion 41, 1311 (2001). [10] Y. Sakamoto et al., this conference, EX/4-3 . [11] T. Fujita, et al., Phys. Rev. Lett. 87, 085001 (2001). [12] K. Nagasaki, A. Isayama, S. Ide, and JT-60 Team, Nucl. Fusion 43, L7 (2003). [13] K. Nagasaki et al., this conference, EX/7-4 . [14] T. Oikawa et al., this conference, EX/P5-15 . [15] S. Takeji, et al., Nucl. Fusion 42, 5 (2002). [16] Y. Ishii, M. Azumi, and Y. Kishimoto, Phys. Rev. Lett. 89, 205002 (2002). [17] M. Takechi et al., this conference, EX/P2-32 . [18] H. Takenaga et al., this conference, EX/6-1 . [19] T. Fujita, et al., Phys. Rev. Lett. 87, 245001 (2001). [20] T. Fujita et al., this conference, EX/P4-3 . [21] J. A. Breslau et al., Phys. Plasmas 10, 1665 (2003). [22] N. Hayashi et al., this conference, TH/1-6 . [23] K. Shinohara, et al., Nucl. Fusion 41, 603 (2001). [24] M. Takechi, et al.,in19th IAEA Fusion Energy Conference Lyon, 2002 (IAEA, Vienna, 2003), pp. EX/W–6. [25] M. Ishikawa et al., this conference, EX/5-2Rb . [26] S. Inagaki et al., this conference, EX/P2-12 . [27] M. Bakthiari et al., this conference, EX/10-6Rb . [28] Y. Takase, et al.,in19th IAEA Fusion Energy Conference Lyon, 2002 (IAEA, Vienna, 2003), pp. PD/T–2. [29] S. Shiraiwa, et al., Phys. Rev. Lett. 92, 035001 (2004). [30] Y. Takase et al., this conference, EX/P4-34 . [31] N. Hosogane, et al., J. Nucl. Mat. 266-269, 296 (1999). [32] T. Nakano et al., this conference, EX/10-3 . [33] N. Oyama et al., this conference, EX/2-1 . [34] G. Saibene et al., this conference, IT/1-2 . [35] T. Tanabe et al., this conference, EX/P5-32 . THE JT-60 TEAM H.Akasaka1), N.Akino1), T.Ando30), K.Anno1), T.Arai1), N.Asakura1), N.Ashikawa28), H.Azechi32), M.Azumi1), M.Bakhtiari33), L.Bruskin17), A.Chankin16), C.Z.Cheng34), S.Chiba1), T.Cho43), Ming-S Chu10), B.J.Ding38), N.Ebisawa1), T.Fujii1), T.Fujita1), T.Fukuda32), A.Fukuyama21), H.Funaba28), H.Furukawa39), M.Furukawa40), X.Gao16), P.Gohil10), N.N.Gorelenkov34), Y.Gotoh39), L.Grisham34), S.Haga39), K.Hamamatsu1), T.Hamano39), K.Hanada23), K.Hasegawa1), H.Hashizume41), T.Hatae1), A.Hatayama19), T.Hayashi1), N.Hayashi1), S.Higashijima1), T.Hino14), S.Hiranai1), Y.Hirano29), H.Hiratsuka1), Y.Hirohata14), J.Hobirk25), A.Honda1), Masao Honda1), Mitsuru Honda21), H.Horiike32), K.Hoshino1), N.Hosogane1), H.Hosoyama39), H.Ichige1), M.Ichimura43), K.Ida28), S.Ide1), T.Idehara9), H.Idei23), Y.Idomura1), K.Igarashi39), S.Iio42), Y.Ikeda1), T.Imai43), S.Inagaki28), A.Inoue39), D.Inoue43), A.Isayama1), S.Ishida1), Y.Ishii1), K.Ishii39), M.Ishikawa33), Y.Ishimoto33), K.Itami1), Sanae Itoh23), Satoshi Itoh23), K.Iwasaki39), K.Kajiwara10), S.Kajiyama12), S.Kakinoto43), Y.Kamada1), A.Kaminaga1), K.Kamiya1), K.Kashiwa1), K.Katayama23), T.Kato28), M.Kawai1), Y.Kawamata1), Y.Kawano1), T.Kawasaki23), H.Kawashima1), M.Kazawa1), M.Kikuchi1), H.Kikuchi39), K.Kikuchi39), A.Kimura21), H.Kimura37), Y.Kishimoto21), S.Kitamura1), A.Kitsunezaki35), K.Kiyono1), K.Kizu1), N.Kobatake12), S.Kobayashi21), Y.Kobayashi21), K.Kodama1), Y.Koide1), S.Kokubo27), S.Kokusen39), M.Komata1), A.Komori28), T.Kondoh1), S.Konishi21), S.Konoshima1), S.Konovaliv16), A.Koyama21), M.Koyanagitsu37), H.Kubo1), T.Kubo32), Y.Kudoh1), K.Kurihara1), G.Kurita1), M.Kuriyama1), Y.Kusama1), N.Kusanagi39), L.L.Lao10), S.Lee11), J.Li16), X.Litaudon3), A.Loarte6), J.Lonnroth7), T.Luce10), T.Maekawa21), K.Masaki1), T.Matsuda1), M.Matsukawa1), T.Matsumoto1), Y.Matsunaga39), M.Matsuoka26), K.Meguro39), A.Mikhailivskii20), K.Mima32), M.I.Mironov2), O.Mitarai22), Y.Miura1), Y.Y.Miura32), N.Miya1), S.Miyamoto32), N.Miyato1), Y.Miyo1), K.Mogaki1), Y.Morimoto37), A.Morioka1), S.Moriyama1), M.Murakami31), M.Nagami1), Y.Nagasaka12), K.Nagasaki21), Y.Nagase23), S.Nagaya1), Y.Nagayama28), O.Naito1), N.Nakajima28), K.Nakamura23), Y.Nakamura28), T.Nakano1), Y.Nakashima43), M.Nakatsuka32), Y.Narushima28), R.Nazikian34), S.V.Neudatchin20), H.Ninomiya1), M.Nishikawa23), K.Nishimura28), N.Nishino13), T.Nishitani1), T.Nishiyama1), N.Noda28), K.Noto39), K.Oasa1), T.Obuchi28), I.Ogawa9), H.Ogawa1), Y.Ogawa40), T.Ohga1), N.Ohno27), K.Ohshima39), A.Oikawa1), T.Oikawa1), M.Okabayashi34), N.Okamoto27), K.Okano4), F.Okano1), J.Okano1), K.Okuno37), Y.Omori1), S.Omori1), A.Onoshi37), T.Ono27), Y.Ono40), H.Oohara1), T.Oshima1), Y.Oya40), N.Oyama1), T.Ozeki1), V.Parail7), B.J.Peterson28), G.D.Porter24), G.Rewoldt34), A.Sagara28), G.Saibene6), T.Saito43), Y.Sakamoto1), M.Sakamoto23), A.Sakasai1), S.Sakata1), T.Sakuma39), S.Sakurai1), T.Sasajima1), M.Sasao41), M.Sato1), F.Sato39), K.Sawada36), M.Sawahata1), M.Seimiya1), M.Seki1), J. P.Sharpe15), T.Shibahara27), K.Shimada1), R.Shimada42), K.Shimizu1), M.Shimizu1), A.Shimizu23), M.Shimono1), K.Shinohara1), S.Shinozaki1), H.Shirai1), S.Shiraiwa40), M.Shitomi1), S.Sudo28), M.Sueoka1), A.Sugawara39), T.Sugie1), K.Sugiyama27), H.Sunaoshi1), Masaei Suzuki39), Mitsuhiro Suzuki39), S.Suzuki1), T.Suzuki1), Yoshio Suzuki33), Yutaka Suzuki39), M.Takahashi39), K.Takahashi39), S.Takamura27), S.Takano39), Y.Takase40), M.Takechi1), N.Takei8), T.Takeishi23) , H.Takenaga1), T.Takizuka1), H.Tamai1), N.Tamura28), T.Tanabe27), Y.Tanai39), J.Tanaka21), S.Tanaka21), S.Tanaka40), T.Terakado1), M.Terakado1), H.Terakado39), K.Toi28), S.Tokuda1), T.Totsuka1), Y.Toudo28), K.Tsuchiya1), T.Tsugita1), Y.Tsukahara1), K.Tsuzuki1), T.Tuda1), T.Uda28), Y.Ueda32), K.Uehara1), T.Uehara39), Y.Ueno21), Y.Uesugi18), N.Umeda1), H.Urano1), K.Urata39), M.Ushigome40), K.Ushigusa1), K.Usui1), M.Wade10), M.Wakatani21), S.Wang5), T.Watari28), M.Yagi23), Y.Yagi29), H.Yagisawa39), J.Yagyu1), H.Yamada28), T.Yamamoto1), Y.Yamamoto21), Y.Yamashita39), K.Yamazaki28), K.Yamazaki28), H.Yamazaki39), K.Yatsu43), K.Yokokura1), I.Yonekawa1), Hajime Yoshida14), Hidetoshi Yoshida1), N.Yoshida23), Hidetsugu Yoshida32), M.Yoshida33), A.Yoshikawa37), H.Zushi23)

1) Japan Atomic Energy Research Institute, Japan 23) Kyushu University, Japan 2) AF Ioffe Physical-Technical Institute of the Russia, Russia 24) Lawrence Livermore National Laboratory, USA 3) Association Euratom-CEA, France 25) Max-Planck-Institut fur Plasmaphysik, Germany 4) Central Research Institute of Electric Power Industry, Japan 26) Mie University, Japan 5) Chinese Academy of Sciences, China 27) Nagoya University, Japan 6) EFDA Closed Support Unit, Germany 28) National Institute for Fusion Science, Japan 7) Euratom/UKAEA Association, UK 29) National Institute of Advanced Industrial Science and Technology, Japan 8) Fellow of Advanced Science 30) Nippon Advanced Technology Co.Ltd., Japan 9) Fukui University, Japan 31) Oak Ridge National Laboratory, USA 10) General Atomics, USA 32) Osaka University, Japan 11) High Energy Accelerator Research Organization, Japan 33) Post-Doctoral Fellow 12) Hiroshima Insitute of Technology, Japan 34) Princeton Plasma Physics Laboratory, USA 13) Hiroshima University, Japan 35) Research Organization for Information Science & Technology, Japan 14) Hokkaido University, Japan 36) Shinshu University, Japan 15) Idaho National Engineering and Environmental Laboratory, USA 37) Shizuoka University, Japan 16) JAERI Fellow 38) Southwestern Institute of Physics, China 17) Japan Society of the Promotion of Science Invitation Fellowship 39) Staff on loan 18) Kanazawa University, Japan 40) The University of Tokyo, Japan 19) Keio University, Japan 41) Tohoku University, Japan 20) Kurchatov Institute, Russia 42) Tokyo Institute of Technology, Japan 21) Kyoto University, Japan 43) University of Tsukuba, Japan 22) Kyushu Tokai University, Japan