Argon–seeded detachment during ELM control by RMPs in KSTAR

In this study, we demonstrate argon-seeded discharges that exhibited a detached divertor during the full suppression and mitigation of edge-localized modes (ELMs) by an International Thermonuclear Experimental Reactor-like, three-row resonant magnetic perturbation (RMP) configuration in KSTAR. During the ELM suppression phase, the peak heat flux on the divertor target was successfully reduced from 1.6 MW m−2–0.5 MW m−2 via argon seeding. Further, the ion saturation current densities corresponding to the particle fluxes on both targets were reduced by more than 50%. During the RMP grassy-ELM regime, a further reduction to 0.1 MW m−2 in the divertor heat load was successfully achieved. A highly localized radiation zone near the X-point was also observed during divertor detachment. The calculated degree of detachment based on the two-point model increased to levels of approximately 3 and 2.3 for the outer target and inner target cases, respectively. These results provide valuable information regarding the effect of mid-Z impurities on RMP-detachment-compatible discharges.


Introduction
The realization of nuclear fusion power faces many scientific challenges, such as sustainable discharges with high confinement as well as controlled heat loads toward divertors. One of the key challenges is to simultaneously achieve these requirements compatible with future fusion devices, * Author to whom any correspondence should be addressed.
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
including the International Thermonuclear Experimental Reactor (ITER).
The ITER aims for inductive H-mode burning plasma as its baseline scenario [1]. The H-mode is characterized by high confinement with a steep gradient and pedestal at the plasma edge. However, H-mode plasmas are often accompanied by periodic heat and particle bursts or edge-localized modes (ELMs) [2][3][4]. Such ELM crashes can cause severe damage to the divertor material, resulting in a threat to sustainable operation, and therefore research to control the ELM phenomenon has been actively explored. Various techniques for ELM control have been proposed, including the application of resonant magnetic perturbations (RMPs) [5], ELM pacing with pellets [6], impurity gas seeding [7], impurity powder injection [8,9] and supersonic molecular beam injection [10]. Among these techniques, RMPs are considered to be the leading strategy by demonstrating complete suppression of ELM crashes [11][12][13][14]. Furthermore, KSTAR has demonstrated divertor heat flux control during ELM crash suppression using ITER-like three-row RMP configuration [15].
While maintaining the desired ELM phase, the control of steady-state divertor heat loads by divertor detachment is simultaneously required in high-performance future tokamaks, including ITER. To achieve this operational regime, experimental and theoretical studies on the compatibility of RMPinduced ELM control with detachment have been conducted [16,17]. Also, impurity-seeded detachment, as well as ELM suppression, was experimentally achieved [18]. Recently, small grassy-ELM regimes have been investigated for highperformance operation with metal walls because the ELM acts as an impurity exhaust mechanism [19][20][21][22]. However, sufficient studies have not been conducted to ascertain whether impurity-seeded detachment is compatible with RMP-induced ELM control or which impurity species are suitable for seeding to achieve radiative detachment in ITER operations. Furthermore, impurity-seeded radiative divertors tend to interfere with RMP-driven ELM suppression. Therefore, it is necessary to investigate and address this possibility through dedicated experiments.
In this study, we report the first demonstration of argonseeded discharges that exhibit a detached divertor during full ELM suppression and mitigation by an ITER-like three-row RMP configuration in KSTAR. During the ELM suppression phase, the peak heat and particle fluxes on the outer target were successfully reduced by argon seeding, and a sign of detachment was observed in the tangentially reconstructed two-dimensional bolometer images. In addition, in the small ELM phase, which is considered a favorable regime for impurity exhaust, further reduction in heat and particle fluxes on the targets was observed compared with the ELM suppression phase.
Among various impurity species, we carried out argonseeded experiments. Neon has low radiation efficiency in both low and high temperatures. Nitrogen has high radiation efficiency in the divertor region, but unfavorable compounds such as ammonia are produced by chemical reactions. On the other hand, argon has a high radiation efficiency in the wide temperature range and is also non-reactive. From that point of view, argon is considered a promising candidate for the simultaneous enhancement of core and divertor radiation.
The experiment demonstrated here was motivated by previous argon trace-amount experiments during RMPs [23]. In these experiments, it was shown that RMPs mitigated the strong core accumulation of argon ions, while the radiated power in the divertor region increased compared with that for the case without RMP. These experimental observations suggest the possibility that the application of RMPs can act favorably for detachment through enhanced divertor radiation.
In the next section, we present the details of the experiment and analyses as follows. In section 2, the experimental setup is explained along with detailed schemes for argon seeding. Section 3 presents the detailed experimental results with various diagnostics. Finally, a discussion and conclusions are presented in section 4.

Discharge with argon seeding in the presence of RMP-driven ELM control
To investigate the compatibility of divertor detachment and RMP-driven ELM control, an argon-seeding experiment was performed at KSTAR. Figure 1 shows the temporal evolution of the operational plasma parameters during the n = 1, +90 • phasing RMP configuration for ELM control.
The target discharge was an H-mode plasma with I p = 0.5 MA and B T = 1.8 T (clockwise direction). Neutral beam injection was used as the main heating source and injected in a stepwise way with a maximum power level of 5 MW at 1.5 s. RMP coil currents I RMP started to ramp up at 4.0 s, and reached the maximum current of 2.0 kA/turn on three-rows. Argon gas was injected from 6.0 s to the end of the discharge through the gas pipeline connected to the midplane at a constant rate. Prefilled and additional deuterium fuel gases were introduced into the plasma from the midplane and divertor, respectively. Langmuir probe arrays on the divertor targets were used to measure the ion saturation current density, and a divertor infrared television (IRTV) system with high resolution was used to measure the heat flux on the outer target [24]. Tangentially reconstructed two-dimensional radiated power distributions were obtained using an infrared imaging video bolometer (IRVB) [25,26].
As RMP was applied, the D α signal showed that ELM was fully suppressed from 4.5 s to 6.6 s, and mitigated from then to the end of discharge, as shown in figure 1(f ). Normalized beta β N and safety factor q 95 remained nearly constant during the RMP phase with values of about 2 and 5, respectively. As argon was introduced to the plasma from 6.0 s, the electron density increased slightly; however, the stored plasma energy W MHD and β N did not exhibit much difference.

Experimental results of argon-seeded detachment during RMPs
During RMP-induced ELM control, noticeable experimental evidence for a reduced divertor heat load was observed from the measured data, such as IRTV, Langmuir probes, and IRVB. Furthermore, we estimated the time trace of the degree of detachment (DoD) based on the two-point model.
After the RMP application, the line-averaged electron density <n e > was decreased by about 20% compared to the before RMP case and then ELM crashes were suppressed. During argon seeding, <n e > was increased to the level of the before RMP case and the ELM suppression phase continued to 6.6 s and changed to the ELM mitigation phase. In the experiment, the position of the strike point was controlled to remain at the Langmuir probe arrays on the divertor. During the RMP  Figure 2 shows the time evolution of the heat flux on the outer target obtained by the divertor IRTV. As shown in figure 2(a), the position of the peak heat flux started to move outward during the RMP current ramp-up phase after 4.0 s and then returned to the initial position after 4.5 s. The maximum heat flux just before argon seeding during RMPs (at 5.8 s) was about 1.5 MW m −2 and reduced to 0.5 MW m −2 during the ELM suppression phase (at 6.6 s).
During the ELM suppression phase, the heat flux continuously reduced by 70% to that of before the argon seeding case (see 5.8 s versus 6.6 s in figure 2(b)), and then the core radiation continued to increase from 0.2 MW to 0.6 MW owing to the argon accumulation inside the plasma. The ELM suppression phase was suddenly changed to the ELM mitigation phase after 6.6 s, and during the ELM mitigation phase, the height of the D α peaks decreased by about a factor of 3 to that of the ELM phase, as seen in figure 1(f ). Since the RMP suppression condition is sensitive to the plasma density [27], it seems to change the ELM mitigation phase as the density increases, as shown in figure 1(c) after argon injection. The heat flux was significantly reduced by approximately 90% compared to that of before the argon seeding case (5.8 s versus 6.8 s in figure 2(b)).
The ion saturation current density j sat corresponding to the particle flux behavior during argon seeding also showed a similar tendency to the heat flux. Figure 3 depicts the inter-ELM target j sat profiles at the targets measured using the Langmuir probe arrays. The position of the strike point was controlled to be located on the divertor target where the Langmuir probe array was installed. During the ELM suppression phase, the particle flux continuously decreased by approximately 50% compared with that before argon seeding (5.8 s). When the ELM suppression phase was changed to the ELM mitigation phase, j sat on both targets was further reduced by approximately 60% compared to that of before the argon seeding case, as shown in figures 3(a) and (b). Figure 4 shows the time trace of DoD defined as jsat,2PM jsat,m = C×<ne> 2 jsat,m , where j sat,2PM is the extrapolated 'attached' divertor ion saturation current density based on the two-point model and j sat,m is the measured divertor ion saturation current density [28]. Following the two-point model, j sat,2PM is expressed as C × <n e > 2 and the normalization constant C was obtained experimentally from the attached phase.
Here, C × <n e > 2 was calculated for the RMP phase before argon seeding (at 5.5 s). After argon seeding for 6.0 s, the DoD at both targets increased simultaneously during the ELM suppression phase with similar levels. When the ELM suppression phase turned into the mitigation phase immediately after 6.6 s, the DoD started to rapidly increase, and then remained at approximately 3 and 2.5 for the outer and inner target cases, respectively. During the ELM mitigation phase, the DoD of the outer target was significantly greater than that of the inner target. In KSTAR, the upstream density at the detachment onset is smaller at the outer target than that of the inner target [29]. Such opposite in/out detachment asymmetry compared to other devices was attributed to the strong deuterium molecule accumulation near the outer target related to the divertor geometrical effects, and consequent strong momentum loss by plasma-neutral interactions, resulting in larger DoD at the outer target compared to that at the inner target.   Clear signs of detachment were observed in the 2D radiation distribution (figure 5). Before the argon seeding, locallyconcentrated radiation occurs near both targets, which infers the attached state. After argon seeding, the region of the most intense radiation shifted progressively from the inner target plate to the X-point. Finally, near the end of the discharge (7 s), radiation near the inner and outer targets was significantly reduced, while the strongest radiation zone was located inside the X-point compared to the case before seeding. It can be inferred that there is a clear temperature  drop near the targets, which is a sign of detachment by argon seeding. As argon seeding continued until the end of the discharge, increased radiation at the core plasma was observed. Figure 6 shows the temporal evolution of the radiated power divided into four regions. It is observed that the radiated power in the main plasma inside the separatrix (purple) immediately increased after argon puffing at 6 s, which became the dominant radiation owing to the accumulation of the argon impurity. On the other hand, the power in the inner divertor region (green) decreased after argon seeding. This is because the radiated power near the inner target decreased owing to a decrease in electron temperature near the target, suggesting a sign of detachment by the argon seeding. In the case of the outer divertor region (orange), there was no decrease in radiated power, despite the reduced radiation near the outer divertor. This is because the most intensive radiation zone shifted from the inner divertor region toward the X-point. As a result, the radiated power in the outer target region seems to increase. However, the radiated power near the outer target clearly reduced during the divertor detachment, as shown in figure 5. Although it is difficult to separate the radiated power emitted by each impurity species, as shown in figure 6, it can be inferred that a substantial amount of argon gas penetrated the plasma as the radiated power increased rapidly after argon seeding. In addition, it had a favorable effect on the reduced heat load on the divertor targets.

Discussion and conclusion
In this study, argon impurity-seeded detachment compatible with RMP-induced ELM control discharge was explored in KSTAR. The main results are summarized as follows: • Argon-seeded detachment during the RMP-driven ELM suppression phase was successfully achieved. When the ELM suppression phase changed to the mitigation phase (small ELM) during argon seeding, further reduction in heat and particle fluxes on the divertor targets occurred compared with the ELM suppression phase. Because the small-ELM regime may have a favorable effect in terms of impurity exhaust in metallic wall devices, our results provide valuable experimental data for the investigation of small-ELM scenarios with impurity seeding. • During the application of RMPs with argon seeding, the DoD based on the two-point model indicated that the DoD at both targets gradually increased to similar levels after argon seeding. When the ELM suppression phase turned into the mitigation phase, the DoD started to rapidly increase, and then remained at approximately 3 and 2.5 for the outer and inner target cases, respectively. • The most intense radiation zone moved from the inner divertor region to the X-point during argon seeding. Furthermore, the radiated power near the targets decreased during detachment.
The experimental results presented in this study demonstrate the possibility of compatible discharge with mid-Z impurity-seeded detachment and RMP-driven ELM control, which is essential in future tokamaks with high performance, including ITER. In future work, experimental studies on impurity-seeded detachment with long-lasting RMP-driven ELM suppression are planned. To maintain the ELM suppression phase during impurity gas seeding, the delicate control of the relevant plasma parameters to meet the ELM suppression window will be explored.