Studies of the outer-off-midplane lower hybrid wave launch scenario for plasma start-up on the TST-2 spherical tokamak

Establishment of an efficient central solenoid (CS) free tokamak plasma start-up method may lead to an economical fusion reactor. CS-free start-up using lower hybrid (LH) waves has been studied on the TST-2 spherical tokamak. Plasma current of about a quarter of CS-driven discharges has been obtained fully non-inductively using the outer-midplane and top LH launchers. Recently, an outer-off-midplane LH launcher was developed to achieve higher plasma current by optimizing for core absorption and minimal fast electron losses. Using the (outer-)off-midplane launcher, fully non-inductive plasma current start-up up to about 8 kA was achieved. Coupled ray-tracing and Fokker–Planck simulation was performed on equilibria reconstructed with an extended MHD model. It was found that the experimentally observed plasma current was in reasonable agreement with the numerical simulation. The simulation predicted appreciable orbit losses for the off-midplane launcher driven discharge at the present parameters, which was consistent with the experimentally observed x-ray radiation characteristics. The simulation showed that the current density was saturated for the present off-midplane launcher discharges and higher density and higher LH power was necessary to achieve higher plasma current.


Introduction
Establishment of a central solenoid (CS) free tokamak plasma start-up method may help realize a fusion power plant that requires high field and/or low aspect ratio [1,2].CS-free start-up methods have been studied extensively in spherical tokamaks (ST) for which elimination of the CS is considered essential.A tokamak plasma can be generated without using the CS with methods such as coaxial helicity injection, merging compression and rf wave current drive [3].For rf start-up, waves in the electron cyclotron range of frequencies [4][5][6][7][8] and lower hybrid (LH) range of frequencies [9][10][11][12][13] have been used successfully to generate a tokamak plasma.
LH current drive is attractive for its high efficiency.Long pulse tokamak operation has been realized with non-inductive plasma current sustainment using LH waves in EAST [14] and Tore Supra [15].Applicability of LH waves for CS-free tokamak start-up has been studied in conventional tokamaks including JT-60U [16].Application of LH waves to ST plasmas is more challenging because the accessible high-density limit decreases at low magnetic field strength.However, working LH start-up scenarios have been developed in STs including Globus-M [9] and TST-2 [10][11][12][13].In the TST-2 experiment, plasma current start-up up to about a quarter of the ohmic discharges has been achieved using the outer-midplane and the top LH launchers.
Numerical analysis of the LH start-up plasmas has been performed using a coupled ray-tracing and Fokker-Planck simulation [17,18].It was found that kinetic modification of the MHD equilibrium by fast electrons was substantial, and equilibrium reconstruction based on extended MHD was needed to accurately describe the LH start-up plasmas [19].The simulation showed that, for the outer-midplane launch scenario, plasma current could be driven only very near the inner limiter where the parallel phase velocity slowed down sufficiently to interact with bulk electrons.For the top launch scenario, robust absorption was predicted above sufficiently high plasma current due to poloidal wavenumber up-shift.However, strong x-ray radiation was observed during the top launch experiment that was more than an order of magnitude greater than what was expected from core fast electrons [20].This suggested that fast electrons were lost through radial transport driven by the LH waves, which produced strong thick-target radiation from the limiters.Furthermore, for both the top and outer-midplane launch scenarios, plasma current was simulated to be driven only very near the last closed flux surface (LCFS) where fast electron confinement time was expected to be limited.To overcome these difficulties, an outer-off-midplane LH launcher was recently fabricated and installed in TST-2 [13].The initial results showed higher core electron temperature, low x-ray radiation intensity and good density control for the new (outer-)off-midplane launcher compared to the previously installed outer-midplane and top launchers.
In this work, detailed analysis of the off-midplane launcher driven discharges as well as the top and outer-midplane launcher driven discharges were performed to understand current drive characteristics with different phase space interaction regimes.The paper is organized as follows.The numerical models used to analyze the experiment is introduced in section 2. The experimental setup is described in section 3. Experimental results including extended MHD equilibrium reconstructions are presented in section 4, and numerical simulation results of LH current drive are presented in section 5. Discussions are given in section 6, and conclusions in section 7.

Numerical model
The LH driven plasma was analyzed using the equilibrium reconstruction code based on the extended MHD model [19].The extended MHD model considered a two-component plasma that consisted of bulk MHD fluid and kinetic fast electrons.A parametrized orbit-averaged (bounce-averaged) electron distribution function was fitted together with the parameters of the bulk free functions to obtain the MHD equilibrium that best fitted the magnetic and kinetic measurements.
The electron distribution function was parametrized as follows: for and The subscripts 1 denote values at (R, The orbit labels are the total energy E = m e v 2 /2, the magnetic moment µ = m e v 2 ⊥ /(2B), the toroidal angular momentum per charge ψ * = ψ + (RB ϕ /Ω e )v ∥ (Ω e = −eB/m e is the algebraic angular cyclotron frequency) and the parallel velocity sign σ = v ∥ /|v ∥ |. ψ(E; R 1 ) determines the center of the fast electron current.∆ψ as defined above sets the radius R 1 to be at two e-folding length from the current center.The normalization N was determined to give the specified fast electron current I fast .The parallel velocity distribution is a mostly constant plateau extending from E min = 9T e1 to E max = m e c 2 /(2N 2 ∥ ).The perpendicular distribution is Maxwellian with the temperature T e1 that was set to be the measured bulk electron temperature.R 2 determines the inner edge of the current profile that was set to be at the inner limiter radius.σ 0 = −1 for the present TST-2 experimental setup.Only the fast electron current center R 1 and the total fast electron current I fast are the fitted parameters.
For the top launch scenario, the electrons were assumed to be accelerated on the high-field side of the magnetic axis, around R 1 : For the outer-midplane and (outer-)off-midplane launch scenarios, the acceleration point was assumed to switch to the lowfield side, at radius R 3 , above a threshold energy E thr : The threshold energy was set to be a half of the launched LH wave parallel phase velocity; this was E thr = 2.6 keV for outermidplane launch and E thr = 0.64 keV for off-midplane launch.
The subscripts 3 denote values at (R, Z) = (R 3 , 0).R 3 > R axis (R axis is the magnetic axis) was determined such that ψ was continuous at E = E thr .Current drive simulation was performed using a raytracing code GENRAY [21] coupled with a Fokker-Planck solver CQL3D [22].CQL3D solves a bounce-averaged (orbitaveraged) Fokker-Planck equation.In the present study, we have assumed zero orbit width for the Fokker-Planck simulation.CQL3D gives an estimation of orbit loss by assuming a loss region for trapped particles where the analytically calculated orbit width is greater than the distance to the LCFS.We have used this estimation to understand qualitatively the orbit losses of different launch scenarios.It should be noted that such a zero orbit width model cannot capture the difference between electron acceleration at the low-field side and highfield side of the magnetic axis, as occurs for outer-midplane launch and top launch scenarios.

Experimental setup
TST-2 is a ST (R 0 = 0.36 m, a = 0.23 m, B t0 < 0.3 T) located at the University of Tokyo [23].Three LH launchers are installed in TST-2 to investigate fast electron generation under various wave launch scenarios as shown in figure 1; the outermidplane launcher (parallel refractive index N ∥ = 5.5) [10], the top launcher (N ∥ = 4.9) [11] and the outer-off-midplane launcher (N ∥ = 13) [13].Four tetrodes are available to generate rf waves at 200 MHz.Capacitively-coupled combline antennas are used to excite LH waves efficiently in this frequency range [10,11,13].A combline antenna is an array of straps fed only at the two ends for power input and output.The antenna acts as a transmission line in vacuum and radiates LH waves in the presence of plasma.When the condition for LH wave radiation is not met, the input power simply comes out of the exit port; the excited spectrum barely changes for a wide range of plasma conditions without any control.For the present experiments, each of the launchers was fed by one tetrode that can deliver up to ∼100 kW.Electron cyclotron waves at 2.45 GHz (<5 kW) was used for pre-ionization of the plasma.
The electron density and temperature profiles were measured using a tangential Thomson scattering diagnostic [24].A radially viewing diode detector with a Be filter was used to measure the overall soft x-ray radiation power between 1 and 10 keV.The line-averaged density was monitored with a 104 GHz interferometer along a radial chord at the midplane.Tangential ion doppler spectrometer was used to measure the ion temperature profile.Magnetic probes were used for direct rf wave measurement in the scrape-off-layer (SOL) [25].

Non-inductive LH start-up experimental results
Discharges with matched vacuum vertical field and LH power waveform were produced to compare the current drive characteristics of the three launch scenarios.The overview of the discharge is shown in figure 2. For all three launch scenarios, the discharge was started with the outer-midplane launcher.The launcher was switched to the top launcher or the offmidplane launcher at 32 ms.At the plasma current flat-top around 60 ms, each of the discharges were driven solely by one of the launchers.The loop voltage (b) was slightly negative showing these are fully non-inductive discharges.The evolution of the plasma current (c) was similar between the three launch scenarios since the vertical field waveforms were matched.The soft x-ray radiation in the 1-10 keV range (e) is from fast electrons since the bulk electron temperature is less than 0.1 keV.The x-ray radiation is considered to be from thick-target radiation from the limiters [20], and was the strongest for the top launch case.The density was somewhat lower for the off-midplane launch case (d) partially due to the different wall condition.Attempt to produce a better matched off-midplane launcher discharge at higher density with additional gas puff was not successful, possibly due to limited LH power.The measured electron profiles are shown in figure 3. The electron temperature was the highest for the off-midplane launch case as was previously found [13].
LH power turn-off was introduced at the flat-top of the discharge (62 ms) to investigate the heating characteristics of the three launch scenarios.Figure 4 shows the plasma response around LH power modulation.It can be seen that the x-ray radiation intensity dropped promptly after LH power turn-off, followed by slow decrease of the plasma current and associated small increase of the loop voltage.The fast time response is the primary reason why we consider the x-ray radiation during the LH pulse to be dominated by thick-target radiation (1-10 keV) by fast electrons (f ) total LH power.All three discharges were started from the outer-midplane launcher.Black solid curve: outer-midplane launcher only, red dashed curve: switched to top launcher at 32 ms, blue dash-dotted curve: switched to off-midplane launcher at 32 ms.The vertical lines at 60 ms shows the timing of the Thomson scattering measurement.[20].Although the off-midplane launcher was designed to minimize fast electron losses [13], we see here fast drop of x-ray radiation for the off-midplane launch case as well.This is likely due to the lower plasma current realized here compared to the design target.This point will be discussed further in section 5 Ion temperature response to LH power modulation is shown in figure 5.The ion temperature dropped substantially 0.5 ms after the LH power turn-off for the outer-midplane launch case as well as the off-midplane launch case.The response of the bulk electron temperature to LH power modulation for the discharges matched to figure 4 is shown in figure 6.The electron temperature modulation was appreciably smaller than the ion temperature modulation.This indicates that the primary heating channel of bulk ions is direct heating by LH waves and not collisional heat transfer from the electrons.The result is consistent with the previous finding [26].
Although LH waves are absorbed only by electrons in the linear regime, non-linear processes such as parametric instabilities can result in ion heating [27].To investigate the spatial profile of LH wave field and parametric instabilities, rf magnetic probe measurement was performed.The results are shown in figure 7. The frequency spectra for the outerlimiter probe (at d = −0.1 m in figure 7) oriented for measurement of the vertical component of the magnetic fluctuations is shown in figure 8.The pump signal shows the intensity of the waves excited directly by the LH launchers.The pump wave intensity was the strongest at the outer midplane for the outer-midplane launch case.For the top launch case, the pump wave intensity was weak overall, which was consistent with the expectation that the LH waves propagate mostly on the high-field side where we presently lack fluctuation measurements.The pump wave intensity at the outer midplane for the off-midplane launch case was similar in magnitude to the outer-midplane launch case, with the probe above the midplane measuring slightly stronger intensity compared to that below.This was also consistent with the upward shift of the launch location for the off-midplane launcher compared to the outer-midplane launcher.The lower side-band (LSB) is produced through non-linear parametric instabilities with ion cyclotron quasimodes and electron Landau quasimodes [25].LSB intensity was especially strong around the midplane for the outer-midplane launch case reaching 30% of the pump intensity.Such strong parametric instability may well provide a few kW of ion heating power necessary to produce the observed ion temperature modulation (figure 5).
Extended MHD equilibrium reconstruction [19] was performed at 60 ms of the matched discharges (figure 2).The global parameters of the reconstruction are shown in table 1.The fast electron content was substantially higher for the offmidplane launch case compared to the outer-midplane and top launch cases.The reconstructed poloidal fluxes are shown in figure 9.The plasma was limited by the inner limiter and the outer gap was around 15 mm for all three cases.The reconstructed toroidal current profiles along the midplane are shown in figure 10.The off-midplane launch case had the lowest current density in the core.This is consistent with the global equilibrium characteristics; higher fast electron content tends to produce a hollow current profile, high elongation and low l i equilibrium.
Discharges optimized for the outer-midplane and top launchers were also produced for comparison of the three launch scenarios around their best performance.The overview of the discharges is shown in figure 11.The main difference of these discharges from the low current ones (figure 2) is the rf Red circles: LH on phase at 55.5-60.5 ms, black downward triangles: LH off phase at 62.25-62.75ms (about 0.5 ms after LH power turn-off at 62 ms).Ion temperature modulation was strongest for the outer-midplane launch case and weakest for the top launch case.
Figure 6.The electron temperature response to LH power modulation for the discharges matched to figure 4 for the outer-midplane launcher.Red circles: LH on phase (60.4 ms, two sets of discharges), black downward triangles: 0.4 ms after the LH power turn off (62.4 ms), blue squares: 0.9 ms after the LH power turn off (62.9 ms).7).The probe was oriented to measure the vertically (poloidally) polarized component of the magnetic fluctuations.The vertical line shows the pump frequency.Black solid: outer-midplane launcher, red dashed: top launcher, blue dash-dotted: off-midplane launcher.2) and the optimized high current discharges (figure 11).Ip is the (total) plasma current.The fast electron current I f is shown as a fraction of the total plasma current.βp is the bulk poloidal beta, l i is the normalized internal inductance and qa is the edge safety factor.

Launch scenario
Ip power.Strong x-ray radiation was observed for the top launch case as in the lower current case.The plasma current driven by the outer-midplane launcher was 15 kA, which was somewhat (a few kA) lower than what can be obtained under the best wall and antenna conditions.The top launcher driven current of 21 kA is close to the maximum that is considered to be limited due to the fact that plasma current is driven only very near the LCFS [18].The measured electron profiles are shown in figure 12.The global parameters of the extended MHD equilibrium reconstruction of the optimized high current discharges are shown in table 1.The fast electron current fraction was 100% for the high current cases.The reconstructed poloidal fluxes are shown in figure 13.The reconstructed toroidal current profiles along the midplane are shown in figure 14.Both the outer-midplane launch and top launch cases The extended MHD reconstruction of the toroidal current profile along the midplane for the matched discharges (figure 2).Black solid with plus symbols: outer-midplane launcher, red dashed with crosses: top launcher, blue dash-dotted with triangles: off-midplane launcher.The current profile was the most hollow for the off-midplane launch case.
were shown to have high elongation and very hollow current profiles.These results are consistent with what has been reported previously [18,19]

Numerical simulation of LH current drive
Coupled ray-tracing and Fokker-Planck simulation was performed for the optimized outer-midplane and top launcher discharges (figure 11) and the off-midplane launcher discharge (figure 2, blue dash-dotted).The launched toroidal refractive index spectra for the three cases are shown in figure 15.
The simulated current profiles are shown in figure 16.The ray trajectories in phase space are shown in figure 17.The top launch case had broad wavenumber spectrum (note N ∥ = c/v ∥ ) that resulted in the highest current density among the three launch scenarios.However, strong wave refraction limited the minor radius the LH waves could propagate up to.For the off-midplane launch case, strong bulk electron absorption was realized due to high electron temperature and high parallel refractive index (low phase velocity).It should be noted that the fast electrons have rather low energy for this off-midplane launch scenario, and current drive efficiency is not expected to be very high.This means sufficiently high power needs to be coupled to the plasma to obtain high plasma current with this off-midplane launcher.The simulated driven plasma current is plotted versus the equilibrium (experimental) plasma current in figure 18.The off-midplane launcher was originally designed to drive current in the 20-30 kA range [13], but such plasma current is yet to be achieved.However, current drive analysis consistent with the experimentally obtained equilibrium profiles shows that, in fact, the off-midplane launcher was performing as expected theoretically.The possible reasons for the performance lower than the design phase prediction are the difference in the equilibrium assumed in the simulation, and limited coupled LH power in the experiment.This point will be discussed further in section 6.
The off-midplane launcher was designed to minimize fast electron losses.Experimentally, presence of fast electron losses were indicated from the x-ray measurement (figure 4).The orbit loss power estimated by CQL3D was 2.6 ± 0.1 kW for the outer-midplane launch case, 7.2 ± 0.9 kW for the top launch case and 3.6 ± 0.1 kW for the off-midplane launch case; the simulated orbit loss was not particularly small for the off-midplane launcher.The zero orbit width model used here is not expected to give quantitatively accurate prediction of orbit loss.On the other hand, considering the agreement of the driven current shown in figure 18, it is likely that the simulated fast electron energy and location of LH wave absorption are in the right ballpark.Therefore, it is likely that appreciable population of electrons have energy high enough to be lost through orbit diffusion for the presently obtained (low current) off-midplane launcher driven discharge.

Discussion
Comparison of the three launch scenarios with matched parameters showed that the newly developed off-midplane    launcher had characteristics as was designed; high core electron temperature with reduced fast electron losses compared to the top launch scenario.However, the off-midplane launcher is yet to outperform the previously installed outer-midplane and top launchers.Figure 19 shows the simulated dependence of the off-midplane launcher driven plasma current on density based on the discharge shown in figure 2. Each of the symbols corresponds to an independent simulation performed The LH ray trajectories in phase space for the optimized outer-midplane and top launcher discharges (figure 11) and the off-midplane launcher discharge (figure 2, blue dash-dotted).The ordinate is the parallel phase velocity (= resonant electron velocity) normalized to the light speed and the abscissa is the normalized minor radius.The hatched region on the low velocity side shows the region less than three time the thermal velocity where absorption by bulk electrons becomes strong.The hatched region on the high velocity side shows the accessibility limit.The top launch case had broad spectrum whereas the off-midplane launch case had good bulk interaction due to low phase velocity.for fixed profiles and rf power.The driven plasma current is predicted to increase with density.In steady-state LH current drive experiments, the driven current usually decreases with density since it is limited by rf power.For LH startup, however, the driven current is often limited due to saturation of the distribution function for a given wave spectrum [17].The present off-midplane launcher case is in the latter regime where the driven current increases with density.On the other hand, the simulated driven current started to saturate above 10 18 m −3 which indicates that higher rf power is needed.Figure 20 shows the simulated dependence of the plasma current on LH power at the central density of 10 18 m −3 .The simulated driven current increased further with rf power.With 80 kW coupled to the plasma with on-axis density of 10 18 m −3 , the driven current was predicted to reach 18 kA (for 8 kA equilibrium profile).The increase of the plasma current with rf power is small due to the saturation of the distribution function stated above.However, the predicted maximum plasma current is close to that of the target plasma for which the off-midplane launch scenario was originally optimized where driven current in the range of 20-30 kA was predicted [13].
We have not yet been able to couple sufficiently high power from the off-midplane launcher due to strong antenna plasma interactions that led to arcing.Stable rf pulse could be obtained by shifting the plasma position away from the antenna, but this resulted in low coupling efficiency that limited the net power into the plasma to less than 30 kW.We plan to install a private limiter and move the launcher closer to the plasma to increase coupling as well as to minimize antenna plasma interactions.

Conclusions
The (outer-)off-midplane launch LH start-up scenario that has good core absorption and minimal fast electron losses was studied in detail.The outer-midplane, top and offmidplane launcher discharges with similar plasma current was produced.The equilibrium reconstruction based on the extended MHD showed the highest fast electron content for the offmidplane launch case, indicating good fast electron confinement.Optimized high current discharges were performed for the outer-midplane and top launch scenarios for comparison with the off-midplane launch scenario through numerical simulation of current drive.The experimentally obtained plasma current was consistent with the simulated driven plasma current for all three cases.For the off-midplane launch case, good absorption due to higher electron temperature and high parallel refractive index was confirmed.Although the off-midplane launcher was designed to minimize fast electron losses, x-ray radiation intensity responded promptly to LH power modulation that indicated finite orbit loss.The orbit loss power estimated by numerical simulation was, in fact, comparable to the optimized outer-midplane launch case at the presently achieved low plasma current.
The plasma current achieved with the off-midplane launcher was limited to 8 kA that was lower than the design target of 20-30 kA.Increasing the plasma density as well as the LH power was predicted to more than double the driven current.The launcher (radiating straps) and the limiter position will need to be modified to couple higher LH power at higher density.
Direct ion heating by LH waves and parametric instabilities were observed during the off-midplane launcher pulse which may have reduced the current drive power.Ion heating and parametric instabilities for the off-midplane launch case were weaker than the outer-midplane launch case but stronger than the top launch case.More work is required to know quantitatively the impact of non-linear wave interactions during the LH start-up discharges.

Figure 1 .
Figure 1.The poloidal cross-section of TST-2 showing the poloidal locations of the top, off-midplane and outer-midplane launchers.The flux surfaces are for a typical top launcher driven discharge.

Figure 2 .
Figure 2. The time traces of the matched LH start-up discharges for three launchers.(a) Toroidal field (b) loop voltage (c) plasma current (d) line-averaged density (e) intensity of soft x-ray radiation (1-10 keV) by fast electrons (f ) total LH power.All three discharges were started from the outer-midplane launcher.Black solid curve: outer-midplane launcher only, red dashed curve: switched to top launcher at 32 ms, blue dash-dotted curve: switched to off-midplane launcher at 32 ms.The vertical lines at 60 ms shows the timing of the Thomson scattering measurement.

Figure 3 .
Figure 3.The electron (a) density and (b) temperature profiles measured by the Thomson scattering diagnostic at 60 ms of the matched discharges shown in figure 2. Black upward triangles: outer-midplane launcher, red squares: top launcher, blue downward triangles: off-midplane launcher.The off-midplane launch case had the highest electron temperature in the core.

Figure 4 .
Figure 4.The high time resolution traces of the matched discharges (figure 2) around the LH power modulation.(a) Plasma current (b) loop voltage (c) soft x-ray radiation intensity (1-10 keV) (d) LH power.Black solid curve: outer-midplane launcher, red dashed curve: top launcher, blue dash-dotted curve: off-midplane launcher.X-ray radiation intensity decreased promptly at the LH power turn-off (vertical lines).

Figure 5 .
Figure 5.The ion temperature (CIII) response to LH power modulation for the matched discharges (figure 4) driven by (a) the outer-midplane launcher (b) the top launcher and (c) the off-midplane launcher.The abscissa is the tangency minor radius.Red circles: LH on phase at 55.5-60.5 ms, black downward triangles: LH off phase at 62.25-62.75ms (about 0.5 ms after LH power turn-off at 62 ms).Ion temperature modulation was strongest for the outer-midplane launch case and weakest for the top launch case.

Figure 7 .
Figure 7.The poloidal profile of rf magnetic field fluctuations for the matched discharges (figure 2, 61-62 ms) driven by (a) the outer-midplane launcher (b) the top launcher and (c) the off-midplane launcher.The abscissa is the poloidal distance d along the limiters where the probes reside.d = 0 is the outer-midplane, d < −0.45 m is on the bottom limiter and d > 0.40 m is on the top limiter.Black upward triangles: pump wave intensity integrated over 199.5-200.7 MHz, red downward triangles: lower side-band (LSB) wave intensity integrated over 180.0-199.5 MHz.The LSB intensity for the outer-midplane launch case was strong, reaching 30% of the pump intensity.

Figure 8 .
Figure8.The frequency spectra measured by the outer-limiter magnetic probe (at d = −0.1 m in figure7).The probe was oriented to measure the vertically (poloidally) polarized component of the magnetic fluctuations.The vertical line shows the pump frequency.Black solid: outer-midplane launcher, red dashed: top launcher, blue dash-dotted: off-midplane launcher.

Figure 10 .
Figure10.The extended MHD reconstruction of the toroidal current profile along the midplane for the matched discharges (figure2).Black solid with plus symbols: outer-midplane launcher, red dashed with crosses: top launcher, blue dash-dotted with triangles: off-midplane launcher.The current profile was the most hollow for the off-midplane launch case.

Figure 11 .
Figure 11.The time traces of the optimized LH start-up discharges at high plasma current for outer-midplane launch (black solid) and top launch (red dashed).(a) Toroidal field (b) loop voltage (c) plasma current (d) line-averaged density (e) soft x-ray radiation intensity (1-10 keV) (f ) total LH power.The top launch case was started with the outer-midplane launcher and switched to the top launcher at 32 ms.The vertical lines at 70 ms shows the timing of the Thomson scattering measurement.

Figure 12 .
Figure 12.The electron (a) density and (b) temperature profiles measured by the Thomson scattering diagnostic at 70 ms of the optimized high current discharges shown in figure 11.Black upward triangles: outer-midplane launcher, red squares: top launcher.

Figure 13 .
Figure 13.The extended MHD reconstruction of the poloidal fluxes for the optimized high current discharges (figure 11).Black solid: outer-midplane launcher, red dashed: top launcher.The LCFS is shown with the thick curve.Limiter boundary is also shown with the black curve.

Figure 14 .
Figure 14.The extended MHD reconstruction of the toroidal current profile along the midplane for the optimized high current discharges (figure11).Black solid with plus symbols: outer-midplane launcher, red dashed with crosses: top launcher.

Figure 16 .
Figure16.The simulated driven toroidal current profiles for the optimized outer-midplane and top launcher discharges (figure11) and the off-midplane launcher discharge (figure2, blue dash-dotted).The uncertainty is indicated by the hatched region obtained by modifying the input density profile within experimental errorbars.

Figure 17 .
Figure17.The LH ray trajectories in phase space for the optimized outer-midplane and top launcher discharges (figure11) and the off-midplane launcher discharge (figure2, blue dash-dotted).The ordinate is the parallel phase velocity (= resonant electron velocity) normalized to the light speed and the abscissa is the normalized minor radius.The hatched region on the low velocity side shows the region less than three time the thermal velocity where absorption by bulk electrons becomes strong.The hatched region on the high velocity side shows the accessibility limit.The top launch case had broad spectrum whereas the off-midplane launch case had good bulk interaction due to low phase velocity.

Figure 18 .
Figure 18.The simulated driven plasma current versus the equilibrium (experimental) plasma current.Black upward triangle: outer-midplane launch, red square: top launch, blue downward triangle: off-midplane launch.

Figure 19 .
Figure 19.The simulated driven plasma current dependence on density for the off-midplane launcher.The lowest density is the experimental point.27 kW LH power.

Figure 20 .
Figure20.The simulated driven plasma current dependence on the off-midplane launcher power at 10 18 m −3 .The coupled power in the experiment was ≲30 kW, presently.

Table 1 .
The global parameters of the extended MHD reconstructions for the matched discharges at low current (figure