Realization of high T_i plasmas and confinement characteristics of ITB plasmas in the LHD deuterium experiments

Intensive wall conditioning using high power ECRH and/or ICRH with helium gas puffing has been conducted to produce a preferable condition for realizing high T_i plasmas in the LHD [45, 46]. Although the wall conditioning effect on reducing the wall recycling was found to depend on the total input RF energy, and not to depend on whether the RF source was ECRH or ICRH [46], the dependence of RF power on the wall conditioning efficiency has not been clarified. The gyrotrons equipped on the LHD can produce the output power of 1 MW each for several seconds and several hundred kW each for more than 10 s. We compared the wall conditioning efficiency using different RF power/pulse duration of 4 MW/2 s (8 MJ/shot) and 1 MW/40 s (40 MJ/shot). Figure 1 shows the dependence of the Dα emission intensity I_Dα in the wall conditioning discharge on (a) the number of the ECRH discharge conditioning (ECDC) shots and (b) the accumulated input ECRH energy ΣW_ECRH. The squares and circles represent the data for 4 MW/2 s and 1 MW/40 s, respectively. The ECRH plasmas for the wall conditioning were produced using 77 and 154 GHz gyrotrons in the presence of the magnetic field B_t of ~2.8 T. Helium was used as the working gas and the line-averaged-electron density n_{e_fir} was maintained around 1  ×  10¹⁹ m⁻³ during each discharge. The numbers of the wall conditioning discharges exceeded 30 for both cases. As can be seen from these figures, the reduction of I_Dα per conditioning discharge was significant for 1 MW/40 s. Also it was found that the reduction of the I_Dα scales with the accumulated energy of the injected RF. This means that the effect of the RF wall conditioning does not depend on the RF power but the accumulated RF energy in the present experimental condition. The contents of ions in the ECDC plasmas changed through the series of ECDC operation. The ion ratios of the ECDC plasma (n_H:n_D:n_He) were (0.07:0.51:0.42) for the ΣW_ECRH of 7.2 MJ, (0.03:0.19:0.78) for 250 MJ, and (0.04:0.09:0.87) for 510 MJ.

Zoom In Zoom Out Reset image size

Figure 2 shows the time evolution of (a) the port-through-NBI power P_NBI, (b) n_{e_fir}, (c) I_Dα, (d) the central ion temperature T_i0, and the radial profiles of (e) the electron density n_e, and (f) T_i for three plasmas produced in the different wall condition. Those plasmas were produced without wall conditioning, after the wall conditioning of 35 shots with 4 MW/2 s, and after the wall conditioning of 32 shots with 1 MW/40 s. The variable r_eff/a₉₉ in the horizontal axis of the figures 2(e) and (f) is the effective minor radius normalized by the averaged minor radius where 99% of the electron stored energy is confined. The plasmas were sustained using high power NBI under the magnetic configuration of R_ax  =  3.6 m and B_t  =  2.85 T with the toroidal magnetic field direction of counter clockwise (CCW), where R_ax is the magnetic axis position in vacuum. In order to inject the highest heating power possible, the tangential and perpendicular NBIs were operated using H and D beams, respectively. After the ECDCs, the Dα intensity was clearly reduced during the discharges and higher T_i was obtained compared with no ECDC case. The n_e profile was modified to the peaked shape from the hollow shape due to the application of the ECDC. Also, the lower I_Dα, the higher T_i, and more-peaked n_e profile were realized in the discharge after the 32-shots of 1 MW/40 s ECDC compared with the discharge after the 35-shots of 4 MW/2 s ECDC. This was due to the lower-recycling condition realized by higher energy input of ECDC, as shown in figure 1(b).

Zoom In Zoom Out Reset image size

The magnetic configuration is also the control knob of the plasma performance optimization because the heating profile, the plasma volume, the heat/particle transport, the MHD activities, and the other confinement characteristics were widely changed depending on the magnetic configuration in the LHD [47–53]. Figure 3 shows the radial profiles of (a) n_e, (b) T_e, and (c) T_i sustained using D gas puff and P_NBI ~ 20 MW with all D beams. The circles are the data for the magnetic configuration of R_ax  =  3.6 m/B_t  =  2.85 T (CCW) and the squares are the data for the inward-shifted configuration of R_ax  =  3.55 m/B_t  =  2.89 T (CCW). No wall conditioning was conducted in advance of these discharges. Although the T_e profile was similar between these two configurations, higher T_i with the steeper gradient in the plasma core was obtained for the inward-shifted configuration with higher n_e. The n_e profile in R_ax  =  3.6 m was hollow shape like that W/O ECDC shown in figure 2(e). On the other hand, the n_e profile was found to be peaked for the inward-shifted configuration even though the wall conditioning was not applied. The lower T_e/T_i and the peaked n_e shape contribute to reduce the linear growth rate of ITG/TEM. A more detailed configuration dependence of high-T_i plasma performance is shown in section 4.3.

Zoom In Zoom Out Reset image size

Turbulence transport theories predict that a certain amount of impurity contributes to stabilizing an ITG mode [54–56] due to the increase in the effective nuclear charge. Actually, high T_i plasmas have been obtained using impurity pellet injection [57, 58]. In the LHD, the quantity of the injected impurity can be actively controlled by changing the size of the impurity pellet [57, 58]. The optimum carbon density for minimizing the ion thermal diffusivity was found in the previous research in the LHD [58]. Here, the result of the impurity quantity optimization aiming for realization of higher T_i plasmas is shown. Figure 4 shows the dependence of T_i0 on the number of the carbon atoms injected using the impurity pellet injector. The magnetic configuration was fixed at R_ax  =  3.6 m and B_t  =  2.85 T (CCW). The data were obtained shot by shot using carbon pellets of different sizes. The optimum quantity of injected carbon for obtaining the higher-T_i plasma was found at around 0.7  ×  10²⁰ both for the plasmas W/D and W/O D through the pellet-size scan experiments.

Zoom In Zoom Out Reset image size

In the deuterium phase, we achieved the highest T_i ever obtained in the LHD experiments due to the several operational optimizations with the increased NBI power as mentioned in the sections above. Figure 5 shows the time evolution of (a) P_NBI, (b) n_{e_fir}, (c) the normalized magnetic fluctuation b_θ/B_t, (d) the neutron emission rate S_n, (e) the plasma stored energy W_{p_dia}, (f) T_i0, and (g) the radial profiles of T_i, T_e, and n_e of the highest-T_i plasma at the timing of the maximum T_i (t  =  4.85 s) in the LHD. The T_e and n_e data in 0.13 m  <  r_eff  <  0.53 m were scattered due to the stray light from the in-vessel components, thus the data are omitted in figure 5(g). Slightly inward-shifted configuration of R_ax  =  3.58 m/B_t  =  2.87 T (CCW) was chosen for the experiment. Also, intensive ECDC was conducted before the discharge. The plasma was sustained using high power NBI and the optimized-size carbon pellet was injected at t ~ 4.57 s. The central T_i was gradually increased after the additional NBI from t  =  4.6 s and reached the maximum value at t  =  4.85 s. The peaked T_i profile with the central value of 10 keV was successfully achieved in the discharge. The achievement of the T_i of 10 keV is a milestone toward realizing a helical reactor, which has an intrinsic advantage for steady state operation, because the T_i value is one of the important ignition conditions.

Zoom In Zoom Out Reset image size

The bursty MHD event, which is a so-called energetic ion driven resistive interchange modes (EICs), was observed during the discharge. The timing of the EIC occurrence is indicated by arrows and dashed lines in figures 5(c)–(f). The EICs are driven by the increased pressure gradient of helically trapped energetic ions, which are mainly generated by the perpendicular NBI [59–64]. The neutron emission rate dropped associated with the EIC event. This indicates the loss of the high energy ions from the plasma. Consequently, the plasma stored energy and the ion temperature both decreased. One of the control knobs of the EICs is ECRH. The threshold of the EIC excitation increases due to the decrease of the mode width, which is proportional to $T_{{\rm e}}^{-1/2}$ (β (dp_e/dr_eff)/p_e)^1/6 [61, 65, 66]. Thus the increasing T_e using ECRH is an effective technique to suppress EICs. The suppression of EICs using ECRH was reported in [61] due to the decrease of the mode width. On the other hand, degradation of the ion thermal confinement by ECRH due to the increase of T_e/T_i was also reported in several devices [67–71]. The increase in the ion thermal diffusivity χ_i with increase in the T_e/T_i during the stepwise ECRH superposition was also observed in the LHD [72]. The optimization study of ECRH for suppression of EICs with lesser degradation of ion thermal confinement will be conducted in order to realize stationary-sustained higher T_i plasmas.

The realization of the high T_i in the deuterium experiments is not only due to the increased NBI power and the operational optimizations but also to the improvement of the ion thermal confinement of the D plasmas compared with the H plasmas. Here, the comparative results of the confinement characteristics between plasmas W/O D and W/D are shown.

Figure 6 shows the time evolution of (a) P_NBI, (b) n_{e_fir}, (c) W_{p_dia}, (d) the central electron density n_e0, (e) the central electron temperature T_e0, (f) T_i0, and (g) the radial profiles of T_i at the timing of the maximum T_i for the typical high T_i plasmas W/O D and W/D. The magnetic configuration was R_ax  =  3.6 m/B_t  =  2.85 T (CCW) both for these discharges. The carbon pellet with cylindrical shape (ϕ  =  1.0 mm, l  =  1.0 mm) was injected at t ~ 4.57 s. The injected number of the carbon atom by the pellet is ~0.7  ×  10²⁰ and is optimum for obtaining high T_i plasmas as shown in figure 4. The increment of the electron density due to the injected carbon pellet is evaluated as ~1.4  ×  10¹⁹ m⁻³. The value is consistent with the increment of n_{e_fir} (~1.2  ×  10¹⁹ m⁻³) just after the carbon pellet injection shown in figure 6(b). In [21], the increment of Z_eff was shown to be ~1 just after carbon pellet injection, decreasing to ~0.2 due to the formation of the impurity hole with the strong outward convection of the impurity [73–75].

Zoom In Zoom Out Reset image size

The higher T_i plasma with higher n_e was realized for the plasma W/D. As shown in the figure 6(f), the T_i was higher in the plasma entire region for the plasma W/D, especially the increase in the T_i around the plasma centre was emphasized due to the steepened T_i gradient in |r_eff/a₉₉|  <  0.25 compared with the plasma W/O D. It was also found that the higher T_i was maintained for a longer time for the plasma W/D. That is the T_i0 continued to increase until 4.74 s for the plasma W/O D and 4.91 s for the plasma W/D. On the other hand T_e0 appeared to be slightly lower for the plasma W/D. The energy confinement time, which was evaluated as the ratio of the total kinetic energy to the total absorbed heating power from GNET [76–78] minus radiation power, was 49 ms at 4.74 s for #123130 and was 55 ms at 4.83 s for #133707. Please note that the ECRH wall conditioning with Helium gas introduced in section 3.1 was conducted prior to both discharges in figure 6. Thus a significant amount of He contaminated in the plasmas. The ratios of ion densities (n_H:n_D:n_He:n_C) are (0.556:0:0.407:0.037) at 4.74 s for #123130 and (0.169:0.293:0.532:0.005) at 4.83 s for #133707. The comparison of the plasma performance between H and D with higher ion purity is shown in section 4.3.

Here the plasma profiles and the ion thermal confinement are compared between the plasmas W/O D and W/D introduced in figure 6. Figure 7 shows the comparison of the radial profiles of (a)–(f) n_e, T_e, T_i, and the gradient, (g) T_e/T_i, and (h) the total ion heating power P_{i_tot} for the plasmas W/O D and W/D. The ion heating power calculated using GNET was slightly lower in the W/D case even though the total port-through NBI power was higher. In the operation W/D, the beam energy of the perpendicular NBI was increased as mentioned in section 2, leading to the enhanced deviation of the beam particles from the magnetic surfaces. This is one of the considered reasons for the decreased absorption power in the plasma W/D. In spite of the lower heating power, higher T_i with higher n_e was realized in the plasma W/D. The improvement of T_i and n_e were significant in the plasma central region. On the other hand T_e profiles were similar between these two plasmas and the value was slightly smaller in the central region for the plasma W/D, leading to the lower T_e/T_i in the entire plasma region except for the plasma edge in the case W/D.

Zoom In Zoom Out Reset image size

The ion species effect on the thermal confinement in hydrogen and helium plasmas were experimentally investigated in the LHD [79]. In [79], the ion thermal diffusivity was evaluated taking account of the gyro-Bohm factor, which is defined as $A_{{\rm i}}^{0.5}T_{{\rm i}}^{1.5}/(aZ_{{\rm i}}^{2}B_{t}^{2})$ [9, 11–15], where A_i is the ion mass and Z_i is the nuclear charge of each ion. In the present study, the contribution of hydrogen, deuterium, helium, and carbon were taken into account for the evaluation of the effective ion thermal diffusivity. Here, we assume that the ion temperature is same among the ions and define the effective ion thermal diffusivity χ_{i_eff} normalized by the gyro-Bohm factor as

$$\begin{align} \newcommand{\e}{{\rm e}} \displaystyle {\rm GB}\,{\rm normalized}\,{{\chi }_{{\rm i}\_{\rm ef\,f}}}=\frac{1}{\frac{{\rm d}{{T}_{{\rm i}}}}{{\rm d}{{r}_{{\rm ef\,f}}}}{{n}_{{\rm i}}}T_{{\rm i}}^{1.5}}\underset{k}{\mathop \sum }\,\frac{{{Q}_{k}}}{A_{k}^{0.5}Z_{k}^{-2}},\nonumber \end{align} \tag{ 1 }$$

where Q is the heat flux and the index k represents H, D, He, and C. Figure 8 shows the comparison of the radial profiles of (a) the effective ion thermal diffusivity normalized by the gyro-Bohm factor computed using equation (1), and (b) the maximum linear growth rate γ_max of ITG and TEM instability normalized by v_{th_H}/R_ax between the plasmas W/O D and W/D introduced in figures 6 and 7, where v_{th_H} is the proton thermal speed. The Q_k was calculated using GNET and T_i was obtained from CXRS measurement. The linear growth rates of ITG/TEM were calculated using GKV taking account of the effect of the multi-ion species [12]. The effective ion thermal diffusivity was found to be smaller in the whole plasma region for the plasma W/D. The GKV simulation showed that the dominant instability was ITG for the core region and was TEM for the plasma edge region both for the plasmas W/O D and W/D. In the outer core region of 0.6  <  r_eff/a₉₉ the tendency in the radial dependence appeared to be similar between the effective ion thermal diffusivity and the linear growth rates of ITG/TEM. However, the experimentally observed difference in χ_{i_eff} for r_eff/a₉₉  <  0.6 could not be explained only by the present linear simulation. Since the zonal flow generation is not taken into account in the linear simulations, non-linear calculations are required for more quantitative comparisons. Also, the stabilizing effect of the mean E_r shear, which is also not included in the present simulation model, may have impacts on the turbulent transport level. They will be addressed in future works.

Zoom In Zoom Out Reset image size

As mentioned in sections 4.1 and 4.2, higher ion thermal confinement was obtained in the operation W/D. However, those comparisons were complicated due to the contamination of the He resulting from the wall conditioning. Here, the plasma performance between H and D was compared with higher ion purity.

Figure 9 shows the n_e dependence of T_e0 and T_i0 in several magnetic configurations for (a), (b) H, and (c), (d) D plasmas, and the configuration dependence of (e) T_e0, and (f) T_i0. The helical coil current in these experiments was fixed as 11.4 kA (CCW) and the magnetic field strength on the axis was 2.82 T (R_ax  =  3.64 m)  ⩽  B_t  ⩽  2.89 T (R_ax  =  3.55 m). The plasmas were sustained using H-NBIs with H gas puff for H plasmas and D-NBIs with D gas puff for D plasmas. The baking and the glow discharge were conducted just before the experiments. Also the ECDC operations with He were not conducted in the series of the experiment. As a result, the ratios of the ion density n_H and n_D to the total ion density were over 0.73 and 0.91, respectively, for the target discharges. The total NBI port-through power was fixed as ~20 MW for all target plasmas here. Although the total NBI power was fixed, the power allocation of the tangential beam and the perpendicular beam was different between H and D plasmas. The port-through power of the tangential NBIs and the perpendicular NBIs were ~11 MW and ~9 MW for the H operation and were ~6 MW and ~14 MW for the D operation. The T_e0 tended to increase in the inward-shifted configuration both for the H and D plasmas. From the comparison of the configuration dependence between H and D, the T_e0 was systematically higher in the H plasmas. In contrast to T_e0, the T_i0 showed the strong dependence on the n_e and the magnetic configuration, especially in D operation. Higher T_i0 was realized with higher n_e in the inward-shifted configuration. Also the obtained T_i0 was found to be higher in D plasmas compared with H with the same magnetic configuration.

Zoom In Zoom Out Reset image size

Clear difference in the profile of the n_e and the impurity was also found between H and D plasmas depending on the magnetic configuration. Figure 10 shows the radial profiles of P_{i_NBI_abs}, n_e, n_C, and T_i in several magnetic configurations for (a)–(d) H, and (e)–(h) D plasmas. Here P_{i_NBI_abs} is the NBI absorption power for ion calculated using GNET. The line-averaged n_e of the data in the figures are ~1.2  ×  10¹⁹ m⁻³ (±8%) for H and ~1.1  ×  10¹⁹ m⁻³ (±8%) for D. Carbon pellets were not used in the operation. Thus the carbon in the plasma was originated from the carbon divertor of the LHD. For the D plasmas, the deviation of the trapped particle orbit from the magnetic surface is larger than the H plasmas due to the larger Larmor radius, leading to the lower heating efficiency of the NBIs. Consequently the ion heating power was similar between H and D with the same configuration of R_ax  =  3.55 m and the 3.6 m even though the port through power of the perpendicular NBIs, which mainly contribute to ion heating, was dominant for the D plasmas. The NBI heating power for ion also tended to increase in the inward shifted configuration. This is due to the decrease of the orbit loss of the NBI ions in the inward shifted configuration [77]. The flat or slightly hollow profile of the n_e and the hollow profile of the n_C so-called impurity hole was formed for H plasmas. The hollowness of the n_C profile became stronger in the outward-shifted configuration and was consistent with the previous experimental results [74]. On the other hand, the shape of the n_e and the n_C profiles of the D plasmas were found to depend on the magnetic configuration. Although the profiles of the n_e and the n_C of the D plasma were hollow for the R_ax  =  3.6 m case as with H case, those were peaked for the inward-shifted configuration of R_ax  =  3.55 m and were qualitatively different with the H case of R_ax  =  3.55 m. The intrinsic formation of the peaked n_C profile has never observed in the i-ITB plasmas in the LHD using hydrogen. The central T_i and the T_i gradient in the core region were larger in the inward-shifted configuration of R_ax  =  3.55 m compared with those of R_ax  =  3.6 m. Interestingly, high T_i, hollow n_C, and peaked n_e were simultaneously realized in the case of R_ax  =  3.57 m for the D operation although the bump was formed in the n_e profile around r_eff/a₉₉  =  0.8.

Zoom In Zoom Out Reset image size

Figure 11 shows the comparison of the radial profiles of the effective χ_i normalized by the gyro-Bohm factor between the H and D plasmas for the magnetic configurations (a) R_ax  =  3.55 m and (b) R_ax  =  3.6 m. The effective χ_i clearly reduced for the D plasmas in the plasma entire region both for the configurations of R_ax  =  3.55 m and 3.6 m. Especially the improvement of the ion thermal confinement in D plasma was significant in the inner half of the minor radius for the configuration of R_ax  =  3.55 m.

Zoom In Zoom Out Reset image size

The configuration dependence of the profile shape of n_e, n_C, and the ion thermal confinement is summarized in figure 12. Figure 12 shows the configuration dependence of (a) the peaking factor of the n_e profile, (b) the peaking factor of the n_C profile, (c) the ion temperature gradient, and (d) the effective χ_i normalized by the gyro-Bohm factor at the r_eff/a₉₉  =  0.4. The triangles and the circles are the data for H and D, respectively. Here the peaking factor of the radial profiles of n_e and n_C are defined as the ratio of the value at r_eff/a₉₉  =  0 to the value at r_eff/a₉₉  =  0.8. The line-averaged n_e of the data in the figures are 1.0–1.3  ×  10¹⁹ m⁻³. From the configuration dependence of the n_e peaking factor and the n_C peaking factor in the H operation, the n_e profiles were almost flat or slightly hollow shape and the hollowness in the n_C profile became stronger in the outward-shifted configuration as mentioned above. The T_i gradient became larger in the inward-shifted configuration. In the D operation, the peaking factor of the n_e and the n_C, and the T_i gradient were significantly increased in the inward-shifted configuration. These tendencies were qualitatively different from those of the H plasmas. The peaked n_C profile was also found in the high T_i plasmas W/D using a carbon pellet [80]. As can be seen in figure 12(d), the effective χ_i was found to be systematically smaller in the D operations than those in the H operations.

Zoom In Zoom Out Reset image size

In the previous sections, the difference in the ion thermal confinement of the ion-ITB plasmas between H and D was discussed. We also conducted the high T_e experiment using high power ECRH with H and D. Here the confinement characteristics of the high T_e plasmas associated with an e-ITB are compared between H and D.

Before the comparison of the thermal confinement of the e-ITB plasmas between H and D, the performance of the ECRH plasma depending on the ECRH injection direction is introduced. Figure 13 shows the time evolution of (a) the ECRH injection power P_ECRH and P_NBI, (b) n_{e_fir}, (c) the plasma current I_p, and (d) T_e0 and the radial profiles of (e) T_e and (f) n_e at 4.37 s for three different EC injection conditions. Those plasmas were produced using co ECCD (Co-ECCD), balanced injected ECRH, and counter ECCD (Ctr-ECCD), with the same P_ECRH of 3.4 MW. The magnetic configuration was R_ax  =  3.6 m/B_t  =  2.705 T and the direction of the toroidal magnetic field was clockwise (CW). Here 'Co-ECCD' means the ECCD injection to increase the rotational transform ι and 'Ctr-ECCD' means to decrease ι. The relation among the direction of B_t, the EC beam, and the driven current is summarized in table 1. The line-averaged-electron density was almost the same value of ~2  ×  10¹⁹ m⁻³ among the three discharges. Although the NBI was superposed on the ECRH plasmas from 3.5 s, this is the perpendicular injection thus the beam had no contribution as NBCD. Unfortunately the ι profiles were not measured in the three cases but the ι is considered to change as intended from the change in the I_p. As can be seen in figure 13(e), the T_e profile was clearly affected by the direction of the ECRH injection direction. Centre-peaked T_e profile and high T_e0 were obtained for the Co-ECCD case. On the other hand, the T_e profile was broad for the Ctr-ECCD case. For the balanced injection case, the T_e0 was between the value of Co-ECCD and Ctr-ECCD.

Zoom In Zoom Out Reset image size

This T_e profile change with the ECCD direction was found not to depend on the polarity of the confinement magnetic field. Figure 14 shows the comparison of the profiles between Co-ECCD and Ctr-ECCD for (a) T_e and (b) ι/2π under the magnetic configuration of R_ax  =  3.6 m/B_t  =  2.705 T with the toroidal field direction of CCW, which was opposite magnetic polarity to the operation of figure 13. The ECRH power and the line-averaged-electron density were 2.8 MW and ~2  ×  10¹⁹ m⁻³, respectively. The T_e profile was also changed with the ECCD direction and the dependence was the same as that shown in figure 13. That is, the centre-peaked T_e was realized for the Co-ECCD case. The ι profile was changed in the plasma central region depending on the ECCD direction. In the previous study, core temperature degradation was observed due to the stochastization of the confinement magnetic field [81, 82]. In those papers, the stochastization in the plasma core region was concluded to be originated from the weakened dι/dr at the low order rational surface of ι/2π  =  0.5 and the flattened temperature profile was appeared in the inner area from the position of ι/2π  =  0.5. In the present research, higher T_e was observed in the low ι shear case with Co-ECCD and the bifurcation point of the T_e profile was r_eff ~0.15 m, which is different from the ι/2π  =  0.5 position of r_eff ~ 0.3 m. Thus, the degradation of the core T_e in the Ctr-ECCD observed in the experiment is considered to be a different phenomenon from the stochastization of the magnetic field due to the weakened dι/dr at the position of ι/2π  =  0.5. We will continue the investigation for clarifying the mechanism of the T_e profile formation depending on the ECCD direction in order to establish the temperature profile control for the maximization and/or the optimization of the plasma performance.

Zoom In Zoom Out Reset image size

As introduced in the previous section, higher T_e with the strong e-ITB was realized using Co-ECCD in the LHD although the mechanism has not been clarified. In order to investigate the isotope effect on the thermal confinement of the e-ITB plasmas, the target plasmas for the comparison were chosen from the Co-ECCD condition. Figure 15 shows the radial profiles of (a) the absorbed ECRH power density p_ECRH and the volume integral P_ECRH, (b) n_e, (c) T_e, (d) the electron pressure n_eT_e, and (e) the electron thermal diffusivity χ_e for H and D with approximately the same n_{e_fir} ~2.4  ×  10¹⁹ m⁻³. The magnetic configuration was R_ax  =  3.6 m/B_t  =  2.705 T (CW) both for the H and D plasmas. The purity of the target ions are n_H/(n_H  +  n_D  +  n_He)  =  0.94 for the H plasma and n_D/(n_H  +  n_D  +  n_He)  =  0.81 for the D plasma. Unfortunately, one gyrotron had trouble in the D experimental phase and the gyrotron could not be operated in the experiments. Thus the total ECRH injection power became smaller in the D experiments. In spite of the decreased ECRH power for the D plasma, almost the same profiles of T_e and n_eT_e with H plasma were realized. Although the n_{e_fir} was fixed as ~2.4  ×  10¹⁹ m⁻³, the n_e profile was slightly different between H and D plasma. The χ_e was evaluated from the power balance analysis and was decreased in r_eff/a₉₉  <  0.45 both for the H and the D plasma due to the formation of the e-ITB. From the comparison between these two discharges, χ_e was clearly reduced in D plasma except for the plasma edge. The systematic data for the comparison of the global energy confinement of the e-ITB plasmas between H and D were also obtained with the n_{e_fir} of 1.5–4.7  ×  10¹⁹ m⁻³ and the injection ECRH power of 1–3 MW. The energy confinement in D plasmas was found to be statistically 10%–20% higher than H plasmas [83].

Zoom In Zoom Out Reset image size

The linear-microinstability simulation using GKV was also applied for the high T_e plasmas. Figure 16 shows the dependence of the linear growth rate of ITG and TEM on k_yρ_t at three different positions around the foot point of e-ITB for (a) H and (b) D, where k_y is the wave number of the eigenmode perpendicular to the magnetic field line and ρ_t is the ion thermal gyroradius. The same normalization with v_{th_H} was imposed to the linear growth rates both for the H and the D cases for the direct comparison of the absolute value. The GKV simulation showed the mixture of TEM and ITG depending on the radius. The linear growth rate of TEM/ITG significantly reduced in the D plasma. Note that both the slight difference of T_i gradient and the large difference of n_e gradient lead significantly different radial dependence of the linear growth rate, even though the T_e profiles are similar in the H and D plasmas. The normalized gradient T_e, T_i, and n_e at r_eff/a₉₉  =  0.6 are R_ax/${{L}_{{{T}_{{\rm e}}}}}$  =  8.87 for H and 8.81 for D, R_ax/${{L}_{{{T}_{{\rm i}}}}}$  =  4.64 for H and 6.80 for D, and R_ax/${{L}_{{{n}_{{\rm e}}}}}$  =  −7.34 for H and 0.124 for D. For further quantitative study, the non-linear simulation taking account of the effect of the mean E_r shear is required and will be carried out as future works.

Zoom In Zoom Out Reset image size