edge localized mode characteristics during edge localized...

Edge localized mode characteristics during edge localized mode mitigationby supersonic molecular beam injection in Korea Superconducting TokamakAdvanced Research

H. Y. Lee,1,2 S. H. Hahn,3 Y.-C. Ghim,4 J. G. Bak,3 J. H. Lee,3 W. H. Ko,3 K. D. Lee,3

S. H. Lee,3 H. H. Lee,3 J.-W. Juhn,3 H. S. Kim,3 S. W. Yoon,3 H. Han,3 J. H. Hong,1,2

J. H. Jang,1,2 J. S. Park,1,2 and Wonho Choe1,2,a)

1Department of Physics, Korea Advanced Institute of Science and Technology (KAIST), 34141 Daejeon,Republic of Korea2Impurity and Edge Plasma Research Center, KAIST, 34141 Daejeon, Republic of Korea3National Fusion Research Institute, 34133 Daejeon, Republic of Korea4Deparment of Nuclear and Quantum Engineering, KAIST, 34141 Daejeon, Republic of Korea

(Received 22 July 2015; accepted 2 December 2015; published online 28 December 2015)

It has been reported that supersonic molecular beam injection (SMBI) is an effective means of

edge localized mode (ELM) mitigation. This paper newly reports the changes in the ELM, plasma

profiles, and fluctuation characteristics during ELM mitigation by SMBI in Korea Superconducting

Tokamak Advanced Research. During the mitigated ELM phase, the ELM frequency increased by

a factor of 2–3 and the ELM size, which was estimated from the Da amplitude, the fractional

changes in the plasma-stored energy and the line-averaged electron density, and divertor heat flux

during an ELM burst, decreased by a factor of 0.34–0.43. Reductions in the electron and ion

temperatures rather than in the electron density were observed during the mitigated ELM phase. In

the natural ELM phase, frequency chirping of the plasma fluctuations was observed before the

ELM bursts; however, the ELM bursts occurred without changes in the plasma fluctuation

frequency in the mitigated ELM phase. VC 2015 AIP Publishing LLC.

[http://dx.doi.org/10.1063/1.4938505]

I. INTRODUCTION

The H-mode is an improved confinement mode in mag-

netically confined plasmas and is characterized by an edge

transport barrier (ETB) and ELMs.1 ELMs are well known

phenomena in which particles and plasma energy are quasi-

periodically ejected from the plasma edge.2 Since the large

heat loads produced by ELMs can cause rapid erosion of

plasma-facing components and divertors,3 ELM control, i.e.,

ELM mitigation or suppression, is an important issue for the

International Thermonuclear Experimental Reactor (ITER)

and future commercial fusion reactors. During the last couple

of decades, several techniques, such as resonant magnetic

perturbation (RMP),4,5 pellet injection,6 and impurity gas

seeding,7 have been developed for ELM mitigation or sup-

pression in tokamak devices. In Korea Superconducting

Tokamak Advanced Research (KSTAR) and HL-2A, ELM

mitigation by supersonic molecular beam injection (SMBI)

has been successfully demonstrated.8,9 SMBI10 is one of the

fuelling methods that can achieve deeper penetration of neu-

tral particles into plasmas than is possible with the conven-

tional gas-puffing, and has been utilized for fuelling

efficiency studies11,12 and ELM mitigation. Since SMBI is a

simple means of controlling ELM via cold gas injection to

the plasma, its system is much simpler than an RMP coil sys-

tem or a pellet injection system, so SMBI system would not

pose critical technical issues. SMBI can also be utilized for

low-cost fuelling, as well as for ELM control in fusion reac-

tors. The downside of SMBI is the plasma confinement deg-

radation due to the cold neutral gas injection.

In the previous study on KSTAR, ELM mitigation by

SMBI demonstrated the following characteristics: (1) an

increase in ELM frequency, fELM, by 2–3.5 times; (2) a

decrease in Da amplitude; (3) 0.2–0.4 s of SMBI influence

time; and (4) an increased level of transport and fluctuation

in the high frequency components after SMBI in the pedes-

tal region.8,9 After SMBI, the line-averaged electron den-

sity, �ne, increased by about 10% as shown in Fig. 2, thus, it

is necessary to investigate how the increase in electron den-

sity, ne, affects the ELM characteristics. Since SMBI

involves the injection of a cold neutral gas, the important

issues to be addressed are the change in global plasma pa-

rameters, divertor heat flux, qdiv, and pedestal plasma pro-

files, particularly ne, and the ion and electron temperatures,

Ti and Te, respectively, which are critical to understanding

the mechanism of ELM mitigation by SMBI. In this paper,

we report the effect of SMBI on the ELM mitigation and the

change in qdiv and the pedestal profiles during ELM mitigation

by SMBI. Since plasma fluctuation measurements are helpful

for discussing the ELM characteristics, plasma fluctuation

measurements are presented herein as well.

This paper consists of five sections. In Section II, the

experimental setup is described. Experimental results includ-

ing the characteristics of ELM mitigation by SMBI, pedestal

profiles, and fluctuation measurements are presented in

Section III. Extensive discussion is provided in Section IV,

and a summary follows in Section V.

a)Author to whom correspondence should be addressed. Electronic mail:

[email protected]

1070-664X/2015/22(12)/122512/9/$30.00 VC 2015 AIP Publishing LLC22, 122512-1

PHYSICS OF PLASMAS 22, 122512 (2015)

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

143.248.6.144 On: Thu, 07 Jan 2016 08:41:06

II. EXPERIMENTAL SETUP

Figure 1(a) shows the heating and diagnostics systems

available in KSTAR. KSTAR is a superconducting tokamak

whose main parameters are described in Ref. 13. In KSTAR,

H-mode discharges can typically be achieved by neutral

beam injection (NBI), whose maximum injection power is

3.5 MW.14 The toroidal rotation velocity (vt) and Ti are

measured using a charge exchange spectroscopy (CES)

system.15 Thomson scattering (TS)16 diagnostic system per-

forms ne and Te measurements, and Te measurement is also

acquired using an electron cyclotron emission (ECE) diag-

nostics system.17 For plasma fluctuation measurements,

a beam emission spectroscopy (BES)18 system and multi-

purpose soft X-ray (SXR) arrays19 with absolute extreme

ultra-violet (AXUV) photodiode (AXUV-16ELG) detectors

are available. The multi-purpose SXR array system has three

kinds of filters: a couple of Be foils, a pair of Ar Ross fil-

ters,20 and a blank opening (referred to as an AXUV bolome-

ter mode). One of the visible filterscopes21 measures the Da

emission, which was utilized as an indicator of ELM detec-

tion in this study. The plasma-stored energy (Wdia) is meas-

ured by a diamagnetic loop.22 Other diagnostic systems are

shown in Fig. 1(a). Post-discharge equilibrium reconstruc-

tion by an equilibrium fitting (EFIT) code23 is implemented

to calculate the plasma flux surfaces. The SMBI system is in-

stalled at the median C-port of KSTAR and the deuterium gas

pressure can be changed from 0.4 MPa to 2.2 MPa.24 Figure

1(b) shows the number of neutral particles injected by SMBI

as a function of SMBI pulse length during the 2013 and 2014

KSTAR campaigns, with an injection pressure of 1 MPa and

an injection gas temperature of 105 K after cooling.25 The

injected particle number linearly increases with the SMBI

pulse length. In this study, the SMBI was applied with a pres-

sure and pulse length of 1 MPa and 8 ms, respectively. The

number of neutral particles injected by a single SMBI pulse

was estimated to be about 3� 1020, and the averaged particle

throughput was approximately 3.8� 1022 s�1.

III. EXPERIMENTAL RESULTS

A. Typical time evolutions of plasma parameters

The experimental conditions were as follows: a toroidal

magnetic field BT¼ 2.5 T, elongation j� 1.6, major radius

R� 1.8 m, and minor radius a� 0.43–0.45 m. Figure 2 shows

the time evolutions of (a) the plasma current Ip, (b) �ne, (c)

Wdia, (d) the Da emission intensity, (e) the NBI power PNBI,

and (f) the SMBI sequence signal during the discharge

#10795. The plasma current was about 0.4 MA during the

time of interest. Two neutral beams were injected with a total

power of approximately 3.0 MW. The ELM frequency was

evaluated to be the reciprocal of the time interval between

the two peaks of the Da emission intensity that accompany

the abrupt changes in �ne due to ELM bursts. The SMBI was

applied at 5.5 s as shown by the green dashed line in Fig.

2(f). During the plasma start-up phase, the number of neutral

particles introduced by conventional gas puffing at the mid-

plane and the corresponding gas fuelling rate were approxi-

mately 8.1� 1020 and 5.5� 1020 s�1, respectively. SMBI

injects fewer particles, but the averaged particle throughput

was larger as its short pulse duration compared with the con-

ventional gas puffing in the plasma start-up phase. Although

the throughput of neutral particles by SMBI is large, the pen-

etration of neutral particles into the plasma core is expected

to be small in general due to the high pressure gradient at the

plasma edge. As shown, �ne is approximately 2.1� 1019 m�3

FIG. 2. Time evolutions of (a) Ip, (b) �ne, (c) Wdia, (d) Da emission intensity,

(e) NBI power, and (f) SMBI sequence signal (Shot #10795). Plasma profiles

are compared between the red and blue shaded regions.

FIG. 1. (a) Heating and diagnostic sys-

tems in KSTAR and (b) calibration

data of number of neutral particles

injected by SMBI, showing good line-

arity with pulse length.

122512-2 Lee et al. Phys. Plasmas 22, 122512 (2015)


143.248.6.144 On: Thu, 07 Jan 2016 08:41:06

before the SMBI, and rapidly increases by 10% (up to

2.3� 1019 m�3) after the SMBI. The line-averaged electron

density gradually decreases to 2.1� 1019 m�3 during the

mitigated ELM phase. The plasma-stored energy decreases

by about 15% from 200 kJ to 170 kJ just after SMBI and then

increases up to 190 kJ. The NBI modulation was also per-

formed to conduct CES measurement at the same time. The

reduction of Wdia was less than 5% due to the NBI modula-

tion in this discharge; however, the NBI modulation did not

change the ELM characteristics.

In the natural ELM phase (t< 5.5 s in Fig. 2), the time-

averaged fELM is approximately 80 Hz. After the SMBI, the

time-averaged fELM increases to about 210 Hz, and reduc-

tions in both the amplitude and the time interval of the Da

emissions were observed (5.5 s< t< 5.73 s in Fig. 2(d)).

Since the Da emission intensity does not reveal the ELM size

itself, we investigated the fractional changes in �ne (D�ne/�ne),

Wdia (DWdia/Wdia), and qdiv to estimate the ELM size.

Figure 3 shows the magnetic flux surfaces in the natural

(�5.4 s) and the mitigated (�5.6 s) ELM phases and the

location of the divertor probe, which was used for the analy-

sis of qdiv. The divertor probe was located near strike points

in both phases, thus we can estimate qdiv by ELM bursts.

Figures 4(a)–4(h) shows the time evolutions of the Da emis-

sion intensity, D�ne/�ne, DWdia/Wdia, and qdiv. The four left

figures (Figs. 4(a), 4(c), 4(e), and 4(g)) and the four right

figures (Figs. 4(b), 4(d), 4(f), and 4(h)) correspond to the nat-

ural mitigated ELM phases, respectively. In the mitigated

ELM phase, D�ne/�ne and DWdia/Wdia are significantly smaller

than they are in the natural ELM phase.

Figure 5 shows DWdia/Wdia as a function of D�ne/�ne dur-

ing a single ELM burst. In the natural ELM phase, the aver-

age D�ne/�ne and DWdia/Wdia by a single ELM burst are 4.4%

(2.7%–5.2%) and 6.5% (5.2%–8.3%), respectively, while in

the mitigated ELM phase D�ne/�ne and DWdia/Wdia decrease to

�1.9% (1.2%–3.3%) and �2.7% (1.3%–4.2%), respectively.

The reduction ratios of D�ne/�ne and DWdia/Wdia between those

two phases are approximately 2.3 and 2.4, respectively. The

divertor heat flux was evaluated by qdiv / TeCdiv using the

divertor probe measurement. Here Cdiv and Te are the parti-

cle flux and the electron temperature obtained from the

divertor probe measurements, respectively. The details of the

qdiv analysis are described in Ref. 26. The peak qdiv caused

by ELM bursts, qpeak, notably decreased during the mitigated

ELM phase. The average qpeak was reduced from

13.5 MW/m2 in the natural ELM phase to 4.6 MW/m2 in the

mitigated ELM phase, yielding a reduction ratio of qpeak of

approximately 2.9. These experimental results, which show

an increase in fELM and reductions in the Da emission inten-

sity, D�ne/�ne, DWdia/Wdia, and qdiv, provide a strong indication

of ELM mitigation by SMBI. It is noted that DWdia� fELM

remains more or less constant in both ELM phases. The

influence time of SMBI, sI, on ELM mitigation is determined

by the change in fELM, Da emission intensity, D�ne/�ne, DWdia/

Wdia, and qdiv, and it was determined to be approximately

0.23 s as shown in Fig. 2(d). After sI (t> 5.73 s in Fig. 2),

large ELMs appear again and fELM gradually decreases.

Figure 6 shows the time evolution of Te measured by

ECE near the plasma edge (R� 2.15 m). During the natural

ELM phase, Te linearly increases during inter-ELMs and

rapidly drops during the ELM bursts, which occur at

Te� 1.3 keV. During the mitigated ELM phase, Te is approx-

imately 1.0 keV and the Te variation caused by the ELM

bursts is smaller than that in the natural ELM phase. After

the mitigated ELM phase (t> 5.73 s), Te remains at about

1.0 keV.

B. Change in plasma profiles during ELM mitigationby SMBI

The plasma profiles in the natural (t¼ 5.3–5.4 s, red

region) and the mitigated (t¼ 5.65–5.7 s, blue region) ELM

phases are compared in Fig. 2. It should be noted that

t� 5.65 s is after the transient change in �ne but is within sI.

Discussion of the rapid profile changes that occur immedi-

ately after SMBI is left as a future work. Figures 7(a) and

7(c) show the radial profiles of ne and Te, respectively, during

the natural (averaged during t¼ 5.3–5.4 s) and mitigated

(averaged during t¼ 5.65–5.7 s) ELM phases. The red and

blue symbols correspond to the natural and mitigated ELM

phases, respectively. Here, ne was measured by TS, and Te

was measured by both TS (closed circles) and ECE (open

circles). The time integration used for both TS and ECE was

50 ms. The edge profiles were fitted with a modified tanhfunction,27 and are plotted as solid curves in Fig. 7. The aver-

age �ne values are about 2.05� 1019 m�3 and 2.15� 1019 m�3

at t¼ 5.35 s and t¼ 5.65 s, respectively. The difference in �ne

FIG. 3. Magnetic flux surfaces from EFIT equilibrium reconstruction before

SMBI (5.4 s) and after SMBI (5.68 s), and location of divertor probe (green

dot).



143.248.6.144 On: Thu, 07 Jan 2016 08:41:06

at those two time regions is approximately 5% and the meas-

ured ne profiles show no dramatic differences between the

natural and mitigated ELM phases. Te decreases by about

10%–20% at R> 2.05 m. Figures 7(b) and 7(d) show the

radial profiles of vt and Ti, respectively, during the natural (at

t¼ 5.395 s) and mitigated (at t¼ 5.685 s) ELM phases. Both

variables were measured by CES with a time integration of

10 ms. The edge profiles were also fitted using a modified

tanh function. Both Ti and vt decrease from the pedestal

region. At the pedestal top, both Ti and vt drop by 30% during

the mitigated ELM phase. In KSTAR, the Mach number,

which is the ratio of vt to the thermal speed, is typically

0.5–0.8 at the plasma core.28,29 The Mach number in KSTAR

is higher than that in JET, which is less than 0.65.30 It is

speculated that the large toroidal rotation may be related

to the low level of intrinsic error field compared to other

tokamaks. Details of low intrinsic error field are described in

Ref. 31. After SMBI, the reductions in Te, Ti, and vt are more

apparent than the change in ne at the plasma edge, which

strongly indicates that SMBI mainly contributes to the reduc-

tions of Te and Ti rather than that of ne at the plasma edge.

C. Fluctuation measurement

In some discharges, ELM precursor-like fluctuations

were observed. Figure 8 shows typical time evolutions of (a)

Ip, (b) �ne, (c) Wdia, (d) Te at R� 2.044 m, (e) the Da emission

FIG. 4. (a) and (b) are Da emission in-

tensity, (c) and (d) are �ne, (e) and (f)

are Wdia, and (g) and (h) are qdiv. (a),

(c), (e), and (g) correspond to natural

ELM phase, and (b), (d), (f), and (h)

denote mitigated ELM phase.

FIG. 5. Fractional change in Wdia as a function of fractional change in �ne.FIG. 6. Time evolution of Te measured by ECE at R� 2.15 m during (a) nat-

ural and (b) mitigated ELM phases.



143.248.6.144 On: Thu, 07 Jan 2016 08:41:06

intensity, (f) fELM estimated from the Da emission intensity,

(g) PNBI, and (h) the SMBI sequence signal, where ELM

precursor-like fluctuations are observed during the discharge

of #9082. BT and Ip were 2.3 T and 0.5 MA, respectively.

The injected neutral beam power was approximately

2.4 MW, and the SMBI was applied at 5.5 s, as shown by the

green dashed line in Fig. 8(h). Similar to the aforementioned

case, ELM mitigation was also obtained after the SMBI by

showing an increase in fELM by a factor of 2–3 and reduc-

tions in the Da emission intensity, D�ne/�ne, and DWdia/Wdia by

factors of 1/2–1/3. Edge localized mode mitigation is also

obtained after SMBI and sI was 0.3 s for this case. Although

the discharge condition is different from shot #10795, the

decrease of D�ne/�ne, and DWdia/Wdia and the increase of fELM

after SMBI were consistently observed in shot #9082. A sta-

tistical study of the above parameters including the divertor

heat flux under various discharge conditions will be helpful

for thorough understanding of the characteristics of ELM

mitigation by SMBI, which will be performed in the follow-

ing campaigns.

Figures 9(a) and 9(d) show the plasma radiation meas-

ured by the multi-purpose SXR array diagnostics in the

AXUV-bolometer mode during the natural and mitigated

ELM phases, respectively. Figures 9(c) and 9(f) are expanded

plots of Figs. 9(a) and 9(d) during t¼ 5.268–5.269 s and

t¼ 5.676–5.677 s in the natural and mitigated ELM phases,

respectively. We analyzed the AXUV-bolometer signal

whose line of sight passed through the plasma edge (indicated

by the thick red line in Fig. 9(g)). As shown in Figs. 9(c) and

9(f), the coherent fluctuations exist observed before the ELM

bursts. Figures 9(b) and 9(e) present the spectrograms of the

plasma fluctuation measured using the AXUV-bolometer dur-

ing the natural and mitigated ELM phases, respectively. In

the natural ELM phase, plasma fluctuations appear after just

1–2 ms of ELM bursts, and their frequencies (fpeak) are in the

range from 25 kHz to 30 kHz, as shown in Fig. 9(b). In the

inter-ELMs, fpeak decreases to 18 kHz and is not changed

by the next ELM burst. The fluctuation intensity slightly

increases from the generation until the extinction of the fluc-

tuation. These fluctuations continue for about 10 ms. During

the mitigated ELM phase, plasma fluctuations also appear

within about 1 ms of the ELM bursts with fpeak¼ 25–30 kHz,

but these fluctuations remain for a few milliseconds due to

FIG. 7. Radial profiles of (a) ne, (b) vt,

(c) Te and (d) Ti; Red and blue symbols

correspond to natural and mitigated

ELM phases, respectively. Closed and

open circles in (c) indicate measured

values by TS and ECE, respectively.

FIG. 8. Time evolutions of (a) Ip, (b) �ne, (c) Wdia, (d) Te at R� 2.044 m, (e)

Da emission intensity, (f) ELM frequency, (g) NBI power, and (h) SMBI

sequence signal (Shot #9082).



143.248.6.144 On: Thu, 07 Jan 2016 08:41:06

the high fELM. The frequency and intensity of the plasma fluc-

tuations do not change during the inter-ELMs. Figure 10

shows the power spectra during the plasma fluctuation phases

for the natural (t� 5.264 s) and mitigated (t� 5.6764 s) ELM

phases, whose time windows are marked by red and blue

bars, respectively, in Figs. 9(a) and 9(d). The peaks of the

power spectra were observed at f� 22 kHz and 30 kHz in the

natural and mitigated ELM phases, respectively. Although

the fluctuation intensity in the mitigated ELM phase is

smaller than that in the natural ELM phase, no significant dif-

ferences between the natural and mitigated ELM phases were

observed for other frequency ranges.

Figure 11(a) shows the time evolutions of the plasma

radiation measured using the AXUV-bolometer, whose line

of sight is described in Fig. 9(g), its spectrogram, and fpeak.

In the natural ELM phase, plasma fluctuations with an fpeak

at 25–30 kHz appear 1–2 ms after the ELM bursts, and fpeak

drops to 18 kHz before the next ELM bursts. On the other

hand, although plasma fluctuations also appear with

fpeak¼ 25–30 kHz in the mitigated ELM phase, the ELM

bursts occur immediately within 2–3 ms. The characteristic

of fpeak chirping was observed when the NBIs were turned

off (t� 5.60–5.62 s) and after the mitigated ELM phase

(t> 5.8 s). Figure 11(b) shows the time traces of fpeak before

the ELM bursts. Here t¼ 0 s represents the time of an ELM

burst, as estimated by the AXUV-bolometer measurement.

In the natural ELM phase, fpeak remains at 25–30 kHz for

2–3 ms after the generation of the fluctuation and decreases

to 18 kHz within 2–3 ms of the ELM bursts. In contrast, the

ELM bursts occur within 2–3 ms, so that fpeak chirping is not

observed in the mitigated ELM phase. This experimental

result suggests that the ELM bursts occur before the fpeak

chirping during the mitigated ELM phase, thus ELM bursts

during the mitigated ELM phase occur earlier than those dur-

ing the natural ELM phase.

The fluctuation measured by the AXUV-bolometer is a

combination of ne and Te fluctuations. Since continuum radi-

ation from the plasma, such as bremsstrahlung and recombi-

nation radiation, is proportional to ne2, ne fluctuation would

be dominant for the emission measured using the AXUV-

bolometer. We will perform further analysis on plasma

fluctuation by comparing BES and ECEI measurements in

the future.

IV. DISCUSSIONS

To discuss the effect of ne on fELM, we investigated the

change in fELM with gradually increasing ne by using conven-

tional gas-puffing. Figures 12(a) and 12(b) show the time evo-

lutions of �ne and the Da emission intensity, respectively,

during conventional divertor gas-puffing. The divertor gas-

puffing was implemented from 4 s to 7 s, and brought about a

gradual increase in �ne (from 3.4� 1019 m�3 to 4.3� 1019m�3).

The number of neutral particles injected during the divertor

FIG. 9. (a) and (d) Plasma radiation

intensities measured by AXUV-

bolometer, (b) and (e) their spectro-

grams corresponding to (a) and (d),

respectively, and (g) schematic view of

line of sight of AXUV-bolometer.

(a)–(c) were measured during natural

ELM phase, and (e) and (f) were meas-

ured during mitigated ELM phase.

FIG. 10. Power spectra at 5.264 s (natural ELM phase) and 5.6764 s (miti-

gated ELM phase).



143.248.6.144 On: Thu, 07 Jan 2016 08:41:06

gas-puffing was approximately 3.2� 1021 and the average fuel-

ling rate is about 1.1� 1021s�1.25 The particle throughput of

SMBI is 10 times greater than the gas fuelling rate of divertor

gas-puffing due to short pulse length of SMBI. In contrast to

SMBI, fELM slightly increased by less than about 10%, with

continuous conventional divertor gas-puffing as shown in Fig.

12(c), but we did not see a significant change in the Da inten-

sity. Figure 12(c) shows fELM as a function of �ne when SMBI

(#9081-#9084) and conventional divertor gas-puffing (#11121)

were employed. The change in fELM with increasing in �ne or

the conventional gas-puffing is only 8%–15%. On the other

hand, fELM significantly increases (fELMSMBI/fELM

0¼ 2.0–3.1)

during the mitigated ELM phase. This result suggests that

using SMBI more effectively changes in fELM than does

increasing of ne or using conventional gas-puffing.

Although Te is still smaller in the mitigated ELM phase

than it is during the natural ELM phase after sI as shown in

Figs. 6 and 8(d), �ne returns to its level during the natural

ELM phase after sI. Other plasma parameters, namely, R, a,

and q95, did not differ between the two phases; thus, the ped-

estal collisionality after the mitigated ELM phase may be

still larger than that in the natural ELM phase. Furthermore,

the fluctuation characteristics during inter-ELMs do not

change in the mitigated ELM phase as described in Section

III C. These results suggest that the ELM mitigation by

SMBI may not be due to a change in the ELM type, i.e.,

type-II ELM regime, which results from the increase in colli-

sionality. However, analysis with an absolutely calibrated

TS measurement is necessary for more definitive statements

on the observed ELM characteristics and the change of ELM

type. This analysis will be performed in our future work.

In contrast to the behavior of #10795 shown in Fig. 2,

we did not observe any reduction in Wdia after SMBI as one

of the examples is shown in Fig. 8. Figure 13 shows the time

evolution of Wdia before and after SMBI for discharges,

#9081-#9084. Despite the injection of cold neutral gas by

SMBI, Wdia remains constant after the injection event at

t¼ 5.5 s. Wdia only varies with the NBI power modulation (at

t¼ 5.4 s, 5.6 s, and 5.8 s in Fig. 8). This result suggests that

ELM mitigation can be achieved by SMBI without signifi-

cant confinement degradation. In contrast to the discharge

described in Section II, significant confinement degradation

was not observed after SMBI. Although it is difficult to

directly compare these results with previous studies since

magnetic geometry, Ip, and plasma heating power were dif-

ferent and the confinement properties may be changed after

SMBI,9 there are several viable interpretations of the change

FIG. 11. Time evolutions of (a) plasma

radiation intensity measured by the

AXUV-bolometer, spectrogram, and

frequency of ELM plasma fluctuation

(fpeak) from 5.35 s to 5.85 s. (b) Time

trace of fpeak before ELM bursts.

FIG. 12. (a) and (b) Time evolutions

of �ne and Da emission intensity,

respectively, during conventional di-

vertor gas-puffing. (c) fELM versus �ne;

red and blue symbols correspond to

natural and mitigated ELM phases,

respectively, and green symbol denotes

conventional divertor gas-puffing case.

FIG. 13. Time evolution of Wdia before and after SMBI for four discharges

(#9081-#9084).



143.248.6.144 On: Thu, 07 Jan 2016 08:41:06

in global confinement degradation. The pedestal parameters,

i.e., the density and temperature, and the magnetic geometry

can be key factors that affect particle deposition by SMBI,

the transport characteristics, and particle recycling at the

plasma edge. Therefore, a statistical study of the effects of

the above parameters on the change in global confinement

by SMBI is necessary with TS measurement data and will be

addressed in future work. The pedestal collisionality may

have an important consequence for ELM mitigation by

SMBI. The pedestal collisionality in this study is expected

to be relatively larger than that in ITER and future fusion

reactors. This effect of SMBI on ELM mitigation in lower

collisional plasmas relevant to ITER or fusion reactors

should be also investigated.

V. SUMMARY

In this paper, an experimental study on the ELM and

pedestal characteristics during ELM mitigation by SMBI in

KSTAR was reported. During ELM mitigation by SMBI,

fELM increased by a factor of 2–3 and the ELM size, i.e., qdiv,

decreased by a factor of 0.34–0.43 when the experimental

conditions were BT¼ 2.5 T, Ip¼ 0.4 MA, and j� 1.6.

Although no dramatic change in the ne profile was not

observed, Te and Ti decreased by about 10%–30% in the ped-

estal region during the mitigated ELM phase. Based on the

plasma fluctuation characteristics, the ELM bursts during the

mitigated ELM phase occur earlier than those during the nat-

ural ELM phase. Our experimental results demonstrate that

SMBI is significantly more effective than conventional di-

vertor gas-puffing during discharge for increasing fELM.

ACKNOWLEDGMENTS

The authors are grateful to Dr. S. A. Sabbagh and Dr.

Y.-S. Park for providing the EFIT reconstruction data and to

M. K. Bae for calculating the divertor heat flux. This work

was supported by the National Research Foundation of

Korea (NRF), which was funded by the Ministry of Science,

ICT (No. NRF-2014M1A7A1A03045089), and Project No.

PG1201-2.

1F. Wagner, G. Becker, K. Behringer, D. Campbell, A. Eberhagen, W.

Engelhardt, G. Fussmann, O. Gehre, J. Gernhardt, G. v. Gierke, G. Haas,

M. Huang, F. Karger, M. Keihacker, O. Kl€uber, M. Kornherr, K. Lackner,

G. Lisitano, G. G. Lister, H. M. Mayer, D. Meisel, E. R. M€uller, H.

Murmann, H. Niedermeyer, W. Poschenrieder, H. Rapp, H. R€ohr, F.

Schneider, G. Siller, E. Speth, A. St€abler, K. H. Steuer, G. Venus, O.

Vollmer, and Z. Y€u, Phys. Rev. Lett. 49, 1408 (1982).2H. Zohm, Plasma Phys. Controlled Fusion 38, 105 (1996).3G. Federici, A. Loarte, and G. Strohmayer, Plasma Phys. Controlled

Fusion 45, 1523 (2003).4T. E. Evans, R. A. Moyer, K. H. Burrell, M. E. Fenstermacher, I.

Joseph, A. W. Leonard, T. H. Osborne, G. D. Porter, M. J. Schaffer,

P. B. Snyder, P. R. Thomas, J. G. Watkins, and W. P. West, Nat.

Phys. 2, 419 (2006).5Y. M. Jeon, J.-K. Park, S. W. Yoon, W. H. Ko, S. G. Lee, K. D. Lee, G. S.

Yun, Y. U. Nam, W. C. Kim, J.-G. Kwak, K. S. Lee, H. K. Kim, and H. L.

Yang, Phys. Rev. Lett. 109, 035004 (2012).6P. T. Lang, H. Zohm, K. Buchl, J. C. Fuchs, O. Gehre, O. Gruber, V.

Mertens, H. W. Muller, and J. Neuhauser, Nucl. Fusion 36, 1531 (1996).7G. P. Maddison, R. V. Budny, P. Dumortier, S. Jachmich, A. Loarte, A.

M. Messiaen, P. Monier-Garbet, M. F. F. Nave, J. Ongena, J. Rapp, J. D.

Strachan, and JET EFDA contributors, Plasma Phys. Controlled Fusion

45, 1657 (2003).8W. W. Xiao, P. H. Diamond, W. C. Kim, L. H. Yao, S. W. Yoon, X. T.

Ding, S. H. Hahn, J. Kim, M. Xu, C. Y. Chen, B. B. Feng, J. Cheng, W. L.

Zhong, Z. B. Shi, M. Jiang, X. Y. Han, Y. U. Nam, W. H. Ko, S. G. Lee, J.

G. Bak, J. W. Ahn, H. K. Kim, H. T. Kim, K. P. Kim, X. L. Zou, S. D.

Song, J. I. Song, Y. W. Yu, T. Rhee, J. M. Kwon, X. L. Huang, D. L. Yu,

K. D. Lee, S. I. Park, M. Jung, S. Zoletnik, M. Lampert, G. R. Tynan, Y.

S. Bae, J. G. Kwak, L. W. Yan, X. R. Duan, Y. K. Oh, J. Q. Dong, the

KSTAR Team, and the HL-2A Team, Nucl. Fusion 54, 023003 (2014).9J. Kim, Y.-M. Jeon, W. W. Xiao, S.-W. Yoon, J.-K. Park, G. S. Yun, J.-W.

Ahn, H. S. Kim, H.-L. Yang, H. K. Kim, S. Park, J. H. Jeong, M. Jung, G.

H. Choe, W. H. Ko, S.-G. Lee, Y. U. Nam, J. G. Bak, K. D. Lee, H. K. Na,

S.-H. Hahn, P. H. Diamond, T. Rhee, J. M. Kwon, S. A. Sabbagh, Y. S.

Park, H. K. Park, Y. S. Na, W. C. Kim, J. G. Kwak, and the KSTAR con-

tributors, Nucl. Fusion 52, 114011 (2012).10L. Yao, New Developments in Nuclear Fusion Research (Nova Science

Publishers, 2006), pp. 61–87.11T. Mizuuchi, S. Kobayashi, M. Takeuchi, K. Mukai, N. Nishino, Y.

Nakashima, S. Yamamoto, H. Okada, K. Nagasaki, T. Minami, S. Oshima,

S. Kishi, H. Lee, K. Minami, Y. Takabatake, Y. Nakamura, K. Hanatani,

S. Konoshima, and F. Sano, J. Nucl. Mater. 415, S443 (2011).12V. A. Soukhanovskii, R. Maingi, C. E. Bush, R. Raman, R. E. Bell, R.

Kaita, H. W. Kugel, C. J. Lasnier, B. P. LeBlanc, J. E. Menard, S. F. Paul,

A. L. Roquemore, and the NSTX Research Team, J. Nucl. Mater.

363–365, 432 (2007).13G. S. Lee, J. Kim, S. M. Hwang, C. S. Chang, H. Y. Chang, M. H. Cho, B.

H. Choi, K. Kim, K. W. Cho, S. Cho, K. K. Choh, C. H. Choi, J. H. Choi,

J. W. Choi, I. S. Choi, C. J. Do, T. H. Ha, J. H. Han, J. S. Hong, K. H.

Hong, N. I. Hur, I. S. Hwang, K. H. Im, H. G. Jhang, Y. S. Jung, B. C.

Kim, D. L. Kim, G. H. Kim, H. S. Kim, J. S. Kim, J. Y. Kim, W. C. Kim,

Y. S. Kim, K. H. Kwon, M. C. Kyum, B. J. Lee, D. K. Lee, H. G. Lee, J.

M. Lee, S. G. Lee, H. G. Na, Y. K. Oh, J. H. Park, H. C. Ri, Y. S. Ryoo,

K. Y. Song, H. L. Yang, J. G. Yang, B. J. Yoo, S. J. Yoo, N. S. Yoon, S.

B. Yoon, G. H. You, K. I. You, W. Choe, D.-I. Choi, S. G. Jeong, D. Y.

Lee, Y. S. Bae, H. S. Kang, G. N. Kim, I. S. Ko, W. Namkung, J. S. Oh,

Y. D. Bae, Y. S. Cho, B. G. Hong, G. Hong, C. K. Hwang, S. R. In, M. H.

Ju, H. J. Lee, B. H. Oh, B. J. Yoon, S. Baang, H. J. Choi, J. Hwang, M. G.

Kim, Y. J. Kim, S. I. Lee, J. Yee, C. S. Yoon, K.-H. Chung, S. H. Hong,

Y. S. Hwang, S. H. Kim, Y. H. Kim, K. H. Chung, J. Y. Lim, D. W. Ha, S.

S. Oh, K. S. Ryu, Q. L. Wang, T. K. Ko, J. Joo, S. Suh, C. H. Choi, J. H.

Lee, Y. W. Lee, H. S. Shin, I. H. Song, J. Baek, I. Y. Han, Y. Koh, P. Y.

Park, C. Ryu, J. J. Cho, D. M. Hwang, Y. S. Kim, J. A. Schmidt, H. K.

Park, G. H. Neilson, W. T. Reiersen, R. T. Simmons, S. Bernabei, F.

Dahlgren, L. R. Grisham, S. C. Jardin, C. E. Kessel, J. Manickam, S. S.

Medley, N. Pomphrey, J. C. Sinnis, T. G. Brown, R. B. White, K. M.

Young, J. Schultz, P. W. Wang, L. Sevier, M. D. Carter, P. M. Ryan, D.

W. Swain, D. N. Hill, W. M. Nevins, and B. J. Braams, Nucl. Fusion 40,

575 (2000).14D. H. Chang, S. R. In, K. W. Lee, J. T. Jin, D. S. Chang, S. H. Jeong, B. H.

Oh, Y. S. Bae, Y. M. Park, and H. L. Yang, Fusion Eng. Des. 86, 244

(2011).15W.-H. Ko, H. Lee, D. Seo, and M. Kwon, Rev. Sci. Instrum. 81, 10D740

(2010).16J. H. Lee, S. T. Oh, and H. M. Wi, Rev. Sci. Instrum. 81, 10D528

(2010).17S. H. Jeong, K. D. Lee, Y. Kogi, K. Kawahata, Y. Nagayama, A. Mase,

and M. Kwon, Rev. Sci. Instrum. 81, 10D922 (2010).18Y. U. Nam, S. Zoletnik, M. Lampert, and A. Kov�acsik, Rev. Sci. Instrum.

83, 10D531 (2012).19S. H. Lee, K. B. Chai, S. Jang, W.-H. Ko, J. Kim, D. Seo, J. Lee, I.

N. Bogatu, J.-S. Kim, and W. Choe, Rev. Sci. Instrum. 83, 10E512

(2012).20R. T. Snider, I. N. Bogatu, N. H. Brooks, and M. R. Wade, Rev. Sci.

Instrum. 70, 316 (1999).21H. K. Na, S. Sajjad, J. M. Park, and M. Kwon, Fusion Eng. Des. 86, 66

(2011).22J. G. Bak, S. G. Lee, and E. M. Ka, Rev. Sci. Instrum. 79, 10F118 (2008).23Y.-S. Park, S. A. Sabbagh, J. W. Berkery, J. M. Bialek, Y. M. Jeon, S. H.

Hahn, N. Eidietis, T. E. Evans, S. W. Yoon, J.-W. Ahn, J. Kim, H. L.

Yang, K.-I. You, Y. S. Bae, J. Chung, M. Kwon, Y. K. Oh, W.-C. Kim, J.

Y. Kim, S. G. Lee, H. K. Park, H. Reimerdes, J. Leuer, and M. Walker,

Nucl. Fusion 51, 053001 (2011).



143.248.6.144 On: Thu, 07 Jan 2016 08:41:06

24Y. O. Kim, J. I. Song, K. P. Kim, Y. Chu, K. R. Park, H. T. Kim, H. K.

Kim, K. S. Lee, and Y. M. Kim, Fusion Eng. Des. 88, 1132 (2013).25J.-W. Juhn, private communication (2013;2014).26J. G. Bak, H. S. Kim, J. Kim, K. I. You, and S. H. Hong, J. Korean Phys.

Soc. 65, 1232 (2014).27R. J. Groebner, D. R. Baker, K. H. Burrell, T. N. Carlstrom, J. R. Ferron,

P. Gohil, L. L. Lao, T. H. Osborne, D. M. Thomas, W. P. West, J. A.

Boedo, R. A. Moyer, G. R. McKee, R. D. Deranian, E. J. Doyle, C. L.

Rettig, T. L. Rhodes, and J. C. Rost, Nucl. Fusion 41, 1789 (2001).

28S.-W. Yoon, J.-G. Kwak, Y.-S. Bae, S. G. Lee, B.-H. Park, and Y.-K. Oh,

in Proceedings of 25th IAEA Fusion Energy Conference (OV/3-4, St.

Petersburg, Russia, 2014).29Y. In, S. G. Lee, and H. H. Lee, private communication (2015).30P. C. de Vries, M.-D. Hua, D. C. McDonald, C. Giroud, M. Janvier, M. F.

Johnson, T. Tala, K.-D. Zastrow, and JET EFDA Contributors, Nucl.

Fusion 48, 065006 (2008).31Y. In, J. K. Park, J. M. Jeon, J. Kim, and M. Okabayashi, Nucl. Fusion 55,

043004 (2015).



143.248.6.144 On: Thu, 07 Jan 2016 08:41:06

edge localized mode characteristics during edge localized...

Documents