Indication of bulk-ion heating by Energetic particle driven Geodesic - - PowerPoint PPT Presentation

SLIDE 1

Indication of bulk-ion heating by Energetic particle driven Geodesic Acoustic Mode on LHD

T.Ido１, K.Ogawa１, A.Shimizu１, M.Yokoyama１, R.Seki１, C.Suzuki１, M. Isobe１, K. Toi１, D. A. Spong2, K.Nagaoka１, Y.Takeiri１, H.Igami１, T.Seki１, K.Nagasaki3 and LHD experiment group

1 National Institute for Fusion Science, Toki, 509-5292, Japan. 2Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA. 3Institute for Advanced Energy, Kyoto University, Uji, 611-0011,

Japan

NIFS M.Osakabe

1

SLIDE 2

CONTENTS

1. Background & Motivation 2. Experimental Observations

a. Brief description of the phenomena b. Bulk-ion behaviors c. Fast-ion behaviors d. Mode structures/ Frequencies

3. Possible candidate to explain the phenomena

a. Orbit topology change b. Enhancement of classical ion heating by the deformation of the energetic particle’s spectra with mode activities. c. Confinement improvement with the mode. d. Anomalous ion-heating with the mode.

4. SUMMARY

2

SLIDE 3

Background & Motivation

• Zonal Flow(ZF) and Geodesic Acoustic Mode(GAM) get much

interests recently, since it could be a knob to regulate turbulence in plasmas and to reduce anomalous transport. Moreover, additional heating by GAM is theoretically pointed

• ut (GAM channeling).*
• Energetic-particle(EP) induced GAMs are observed in several

toroidal devices, such as JET, DIIID and LHD**. Effect of GAMs

• n the energetic particle behaviors needs to be investigated.
• Recently, an influence of EP induced GAM on bulk-ions was
• bserved on LHD.

*M.Sasaki, et.al., PPCF 53(2011)085017 **H.L.Berk, et.al., Nucl. Fusion 46, S888, C.J.Boswell, et.al., Phys. Lett. A , 358, 154, R.Nazikian, PRL 101, 185002, G.Y.Fu, PRL 101, 185002, K.Toi: 22nd IAEA-FEC, EX P8-4, T.Ido : 23rd IAEA-FEC

3

SLIDE 4

a) b) c) d) e) f) g)

On LHD, increase of low energy neutrals are observed with up-sweeping n=0 modes by tangential NPA at low density plasmas.

• The mode was only excited during

counter-NB injection phase.

• Superposition of ECH seems to be

effective to excite the mode.

• The typical initial frequencies of the

mode are 50 -100kHz. The flux increase was associated with relatively large amplitude modes.

• No significant increase of Ha-signals

were observed.

=> Measured increase in neutral flux are due to the low energy ion behaviors.

• Typical slowing-down time is

estimated to be ~10[s] at the core

4

2 4 6 8 10 12 14 0.2 0.4 0.6 0.8 1 T

e()

ne() 

se/2

T

e[keV], n e [10 17 m

• 3], 

se/2[sec.]

r/a

SLIDE 5

Bulk-ion behavior

• Neutral flux close to the bulk-ion

energies are starts to increase after the mode excitation.

• The effective ion temperature also

starts to increase after the mode excitation.

5

Either confinement property of bulk-ions or bulk-ion heating mechanisms seem to be changed with the mode excitation.

1018 1019 1020 1021 1022 b)

0.9keV 1.3keV 2.4keV 1.9keV

NeutralFlux [a.u.]

• 3
• 2
• 1

1 2 3 a) B B

 [x10

• 6 T]

0.2 0.4 0.6 0.8 1 4.9 4.95 5 5.05 c) Ti

eff.

T

i

• eff. [keV]

time[sec] ~7.5ms Decay time = ~79.3ms 0.43keV 0.26keV

10

16

10

17

10

18

10

19

10

20

10

21

1 2 3 4 5 6 d) 4.9325s Neutral flux [a.u.] Energy[keV] 10

16

10

17

10

18

10

19

10

20

10

21

1 2 3 4 5 6 d) 4.9325s 4.9475s Neutral flux [a.u.] Energy[keV]

SLIDE 6

Fast-Ion Behavior

• Decay times of the increased flux were order of

10-100 ms.

 The fast-ions influenced by the mode activity were not promptly lost and were circulating on confined

• rbits.
• The decay times were smaller than the

expected slowing-down time of fast-ions and almost no-delays of decays were found at each energy.

 They were lost before they undergo slowing-down process.

• The decay times have energy dependence.

They became smaller as the energy decreases.

 By comparing the decay time with charge exchange cross section, the charge exchange loss was found to be the dominant loss process.

6

• 2.00
• 1.00

0.00 1.00 2.00

• 0.005

0.005 0.01 0.015 Mirnov (ch.=0) dB/dt [T/s] t-t0 [s]

a)

1016 1017 1018 1019 1020

b)

Before (-0.005~0[s]) Beginning (0~0.005[s]) Maximum (0.005~0.01[s]) End (0.01~0.015[s]) Neutral Flux ()[a.u]

• 2.00 1018
• 1.50 1018
• 1.00 1018
• 5.00 1017

0.00 5.00 1017 1.00 1018 0.00 50.00 100.00 150.00 200.00 c)

Beginning (0~0.005[s]) Maximum (0.005~0.01[s]) End (0.01~0.015[s])

 [a.u.] Energy [keV]

140keV 151keV

The charge exchange loss process produces the positive gradients in the energy spectra.  Source of instability drive.  Induces clump-hole formation in the energy spectra. The flux increase and its decay were observed with the mode activity.

SLIDE 7

MODE DE ST STRU RUCT CTUR URE AN E AND ITS D ITS F FREQ REQUE UENCY NCY

7

SLIDE 8

Evaluation of poloidal mode structure by HIBP measurement.

• The spatial structures are

consistent with the structures of the

• GAM. ( : m ~ 0, : m ~ 1)

Symmetry Asymmetry

Upper side Lower side

(The amplitudes are normalized by |Bp| to remove variation among events.)



e

n

Upper side Lower side



e

n

T.Ido, et.al.

8

SLIDE 9

Te dependence of the initial frequency of n = 0 mode

9 ,0 ,0

2

NB b

v f R   

85 kHz, where Eb = 175 keV,  = 0.35, R = 3.75m

7 4

GAM e i

f T T  

• 1. One has the Te 0.5-dependence of the mode frequency. (=> GAM)
• 2. The other mode has weak Te -dependence of the frequency. The

mode frequency is much larger than the usual GAM, and is close to the orbital frequency of the fast-ions produced by the NBI.

p

dB dt

T.Ido, et.al.

SLIDE 10

// , th ion

V k 

5 10 15 20v 0.2 0.4 0.6 f1

Slowing-down distribution

gGrowth > 0

The FI affects the GAM frequency

The GAM frequency is modified by FIs. (G.Y.Fu, PRL,101,185002 (2008) )

1 2 3 3 exp

exp

c EP

A v v P f e



                             Λ = 𝜈𝐶 𝐹

(T.Watari, et al, PoP, 13, 062504 (2006)

T.Ido, et.al.

2 3 3 2 3 3 3 3

exp exp

EP c c c

v v A v v v P f e v

 a 

                                     

The freq. of FI-driven GAM can be much higher than the ordinary GAM freq.

// , th ion

V k 

gGrowth < 0 gGrowth > 0

a = 6.8

𝜶 ∙ 𝑲𝑐𝑣𝑚𝑙

𝑕𝑓𝑝𝑒𝑓𝑡𝑗𝑑 + 𝑲𝑞𝑝𝑚𝑏𝑠𝑗𝑨𝑏𝑢𝑗𝑝𝑜 =0

𝜶 ∙ 𝑲𝑐𝑣𝑚𝑙

𝑕𝑓𝑝𝑒𝑓𝑡𝑗𝑑 + 𝑲𝐺𝐽 𝑕𝑓𝑝𝑒𝑓𝑡𝑗𝑑 + 𝑲𝑞𝑝𝑚𝑏𝑠𝑗𝑨𝑏𝑢𝑗𝑝𝑜 =0

10

SLIDE 11

Observed n=0 mode with weak Te-dependence is also identified as GAM

11

n=0 mode (Experiment) EGAM T.Ido, et.al. More detailed analysis based on numerical simulation can be found at H.Wang TH/P1-12

SLIDE 12

Mechanisms which might explain this phenomenon

• 1. Change of orbit topology by the GAM. activities
• 2. Enhanced ion-heating by the classical collision process due to

the deformation of fast-ion spectra by the GAM

• 3. Improvement in bulk-ion energy confinement with the mode
• 4. Enhanced ion-heating with the mode

12

SLIDE 13

Mechanisms which might explain this phenomena

• 1. Change of orbit topology by the GAM. activities
• 2. Enhanced ion-heating by the classical collision process due to

the deformation of fast-ion spectra by the GAM

• 3. Improvement in bulk-ion energy confinement with the mode
• 4. Enhanced ion-heating with the mode

13

These effects were examined by numerical calculations:  The first one was examined by orbit calculations with perturbed electrostatic potential by delta5d.

• The effect was small. Thus, it was ruled out.

 The second one was examined by TASK-3D calculation with an artificial spectral deformation, which reproduces the experimental observation.

• The effect was also small. Thus, it was ruled out.

SLIDE 14

Mechanisms which might explain this phenomena

• 1. Change of orbit topology by the GAM. activities
• 2. Deformation of fast-ion energy spectra by the mode and the

enhancement of ion-heating by classical collision process.

• 3. Improvement in bulk-ion energy confinement with the mode
• 4. Enhanced ion-heating with the mode

14

SLIDE 15

The effects of confinement improvement and enhanced ion- heating were evaluated by 0-D power balance model

• To investigate the possibility of these

effects, typical values of characteristic numbers, Pi (ion-heating power density) and E

• eff. (effective ion energy confinement

time) were examined in a simple power balance model with an assumption of constant ion–density.

1 2 b) b)

B

 rms[T]

B

rms [x10-6 T]

5 10 15 20 25 2 4 6 8 10 a)

ne(0) (Thomson) ne (fir)

T

e(0)

n

e(0), n e [x10 17m

• 3]

T

e(0) [keV]

0.2 0.4 0.6 0.8 1 4.9 4.95 5 5.05 b) T

i eff.

fitting T

i

• eff. [keV]

time[sec] ~7.5ms Decay time = ~79.3ms 0.43keV 0.26keV

𝑒𝑈𝑗

𝑓𝑔𝑔.

𝑒𝑢 = 𝑄𝑗 3𝑜𝑗 2 − 𝑈𝑗

𝑓𝑔𝑔.

𝜐𝐹𝑓𝑔𝑔.

Original value of Pi and E

• eff. can be

evaluated from:

(i) Ion heating power estimated by TASK3D: Pi= 170W/m3 => E

• eff. = 364ms .

(ii) Ion temperature decay: E

• eff. =~80ms => Pi= 774W/m3

50-150kHz

ne=ni assumed

15

c)

SLIDE 16

Evaluation of the enhanced ion-heating case and the confinement improvement case

• In the enhanced ion-heating case,

the observed ion temperature was reproduced by assuming 4.3[kW/m3]of ion heating power during the mode activity.

• In the case of improved

confinement case, the observed ion temperature behavior could not be reproduced even if the confinement time of infinitely large number was assumed(E=~10000[s]).

0.0 0.5 1.0 b) P

ion increase



E increase

E

• eff. [s]



E=10000s

0.0 2.0 4.0 6.0 a) P

ion increase



E increase

Pion [kW/m3]

0.0 0.5 1.0 4.90 4.95 5.00 5.05 c) P

ion increase



E increase

Ti

eff.[keV]

time [s]

The candidate of “confinement improvement” Was ruled out. Thus, “the ion heating enhancement” becomes most probable candidate.

16

SLIDE 17

Ion temperature behavior with the time integrated value of mode amplitudes

• Although clear response of the

flux increase was observed with the mode activity, the correlation between the mode amplitude and the temperature behavior was not clear.

• If we assume, the ion-heating

power is related to the mode amplitude, the stored energy would correspond to its time integrated value.

17

1 2

a) a)

B

 rms[T]

B

rms [x10-6 T]

0.5 1

4.9 4.95 5 5.05 b) T

i eff.[keV]

Ti

• eff. [keV]

time[sec]

10

• 1

10

10

• 16

10

• 15

10

• 14

10

• 13

Ti

eff [keV]

𝜀𝐶𝜄 2𝑒𝑢

𝑢0+Δ𝑢 𝑢0

SLIDE 18

SUMMARY

• On LHD, phenomena indicating the ion-temperature increase associated

with FI induced n=0 mode was observed by tangential NPA at very low density plasmas.

• Its poloidal mode structure was examined by the HIBP measurement, and

were consistent to the GAM.

• According to the Te-dependence of the frequency, mode can be classified

into two categories, i.e., one has Te1/2-dependence and the other has weak Te-dependence. The frequency of the latter mode are almost equal to the orbital frequency of fast-ions.

• Considering the positive gradient in the energy spectra of fast-ions, these

frequency features can be explained in the framework of GAM. Thus,

• bserved mode was identified to be GAM.
• Four candidates were proposed to explain the phenomena; (1)change of

measured ion spectra due to the orbit topology change, (2)increase of classical ion heating power by the deformation of energetic particle spectra, (3) improved energy confinement, and (4) enhanced ion heating with the mode activities. Among them, the enhanced ion heating with the mode activities becomes most probable candidate.

• But, the heating mechanism was left as an open question for the future

studies.

18

SLIDE 19

BAC ACKUP UP VI VIEW EWGRA RAPH PHS

19

SLIDE 20

Charge Exchange loss was seems to be significant at the discharge.

• Decay constants of the neutral flux show similar

energy dependence to the reactivity(scxv) for Charge eXchange (CX) loss with Hydrogen neutrals.

• If we assume the neutral density is ~2x1015[m-3],

the decay constants agree the CX-loss time, qualitatively.

• The evaluated neutral density is consistent to

the neutral density calculation by AURORA-code.

• The CX-loss is the major loss process of the

clump.

• Evaluated CX-loss time (10- 100ms) is much

smaller than the slowing-down time(order of 1s)

• f fast-ions

20

• The CX-loss of the fast-ions

diminishes the slowing-down feature of the clump.

• Positive-gradient in the energy

spectra can be formed by CX-loss.

10 20 30 40 50 0.0 0.5 1.0 1.5 2.0 2.5 0.0 20.0 40.0 60.0 80.0 100.0 b) b) 4.83 834s scx

cxv [m

[m3/s] /s] Neutral Flux Decay Const [s-1] scxv [x10-14 m3/s] Energy [keV]

• 5 10

17

5 10

17

1 10

18

1.5 10

18

2 10

18

4.8 4.85 4.9 4.95 5

#109032 65.2keV 75.8keV 50.9keV

NeutralFlux [a.u.] time[sec]

exp-fitting

SLIDE 21

Toroidal mode number is n=0

21

The initial frequency of the mode is much smaller than expected n=0 GAE frequencies (f>500kHz).

• 360
• 270
• 180
• 90

90 90 180 270 360 45 45 90 90 135 180 225 270 315 360 Toroidal/#94839@t=1.7983s Phase [deg.]  [deg.] n=0

• 360
• 270
• 180
• 90

90 90 180 270 360

• 180-135-90 -45

45 45 90 90 135 180 Poloidal/#94839@t=1.7983s Phase [deg.]  [deg.] m=1 m=2

SLIDE 22

Toroidal mode is zero. Phase for poloidal structure has a slope of unity, but ..

a phase jump was observed at around =0[deg]. This indicates a formation of standing-wave like structure.

22

• Standing wave is formed by the sum of

two modes having a same mode number and propagating opposite to each other.  In the case of m>3, we should

• bserve more than three nodes.

 In the case of m=1, it is difficult to have a node like structure with a slope of m=1. The m=2 is the most likely in this case.

• 360
• 270
• 180
• 90

90 90 180 270 360 45 45 90 90 135 180 225 270 315 360 Toroidal/#94839@t=1.7983s Phase [deg.]  [deg.] n=0

• 360
• 270
• 180
• 90

90 90 180 270 360

• 180-135-90 -45

45 45 90 90 135 180 Poloidal/#94839@t=1.7983s Phase [deg.]  [deg.] m=1 m=2

*Consistent to theoretical prediction for poloidal mode of Magnetic component of GAM; Zhou D, PoP, 14, 104502 (2007 )

SLIDE 23

Confinement improvement case

• Pion constant model -
• Energy confinement was assumed

to be improved to infinite number during the mode activity.

• Estimated temperature rise was

smaller than the experimental

• bservation.
• Charge-exchange loss time of

5~100ms for bulk ions with background neutrals (1014~ 1015m-

3) will also suppress the

confinement improvement.

0.0 0.2 0.4 0.6 0.8 1.0 a) 

E eff.=364ms



E eff.=80ms

P

ion [kW/m 3]

0.00 0.20 0.40 0.60 0.80 1.00 b) 

E eff.=364ms

E

eff.=80ms



E

• eff. [s]



E -> infinite(10 5 s)

0.0 0.2 0.4 0.6 0.8 1.0 4.90 4.95 5.00 5.05 c) 

E eff.=364ms



E eff.=80ms

Ti

eff.[keV]

time

Confinement improvement can not explain the measured temperature rise.

23

SLIDE 24

Enhanced ion-heating case

• Constant confinement time model -
• Ion heating power was assumed to

be increased during the mode activity.

• The amount of heating power was

determined so that the observed temperature rise was reproduced.

• The Pi enhancement factors of;

– 21 (Pi=3.6kW/m3) for i

eff.=364ms,

and – 5.6 (Pi=4.3kW/m3) for i

eff.=80ms.

are necessary to explain the

• bserved temperature rise.

0.0 1.0 2.0 3.0 4.0 5.0 a) 

E eff.=364ms

E

eff.=80ms

P

ion [kW/m 3]

0.00 0.20 0.40 0.60 0.80 1.00 b) 

E eff.=364ms



E eff.=80ms



E

• eff. [s]

0.0 0.2 0.4 0.6 0.8 1.0 4.90 4.95 5.00 5.05 c) 

E eff.=364ms



E eff.=80ms

Ti

eff.[keV]

time [s]

Indication of enhanced ion-heating power during the mode activity.

24

SLIDE 25

Combination of m=2/-2 mode might explain the

• bserved phase structure*.
• In the case of m>3, we should observe more than three nodes in

the poloidal phase diagram.

• In the case of m=1, it is difficult to have a node like structure with a

slope of m=1.

*Consistent to theoretical prediction of poloidal mode of Magnetic component of GAM; Zhou D, PoP, 14, 104502 (2007 )

Further investigation is necessary since it sometimes show m=1 structure without node.

• 1
• 0.5

0.5 1 1.5 2 0.5 1 1.5 2

m=+3/-3

A

+/A

• =0.0

A

+/A

• =0.1

A

+/A

• =0.2

A

+.A

• =0.3

A+/A-=0.4 A

+/A

• =0.5

A

+/A

• =0.6

A

+/A

• =0.7

A

+/A

• =0.8

A

+/A

• =0.9

A+/A-=1.0

Phase/2 Position/2 m=1 m=3

• 1
• 0.5

0.5 1 1.5 2 0.5 1 1.5 2

m=+2/-2

A

+/A

• =0.0

A

+/A

• =0.1

A

+/A

• =0.2

A

+.A

• =0.3

A

+/A

• =0.4

A

+/A

• =0.5

A

+/A

• =0.6

A

+/A

• =0.7

A

+/A

• =0.8

A

+/A

• =0.9

A

+/A

• =1.0

Phase/2 Position/2 m=1 m=2

• 1
• 0.5

0.5 1 1.5 2 0.5 1 1.5 2

m=+1/-1

A

+/A

• =0.0

A+/A-=0.1 A

+/A

• =0.2

A

+.A

• =0.3

A+/A-=0.4 A

+/A

• =0.5

A

+/A

• =0.6

A+/A-=0.7 A

+/A

• =0.8

A

+/A

• =0.9

A+/A-=1.0

Phase/2 Position/2 m=1

25

SLIDE 26

5 10 15 20v 0.2 0.4 0.6 f1

Slowing-down distribution

The EP affects the GAM frequency

The GAM frequency is modified by FIs. (G.Y.Fu, PRL,101,185002 (2008) )

1 2 3 3 exp

exp

c EP

A v v P f e



                             Λ = 𝜈𝐶 𝐹

(T.Watari, et al, PoP, 13, 062504 (2006)

T.Ido, et.al. 𝜶 ∙ 𝑲𝑐𝑣𝑚𝑙

𝑕𝑓𝑝𝑒𝑓𝑡𝑗𝑑 + 𝑲𝑞𝑝𝑚𝑏𝑠𝑗𝑨𝑏𝑢𝑗𝑝𝑜 =0

𝜶 ∙ 𝑲𝑐𝑣𝑚𝑙

𝑕𝑓𝑝𝑒𝑓𝑡𝑗𝑑 + 𝑲𝐹𝑄 𝑕𝑓𝑝𝑒𝑓𝑡𝑗𝑑 + 𝑲𝑞𝑝𝑚𝑏𝑠𝑗𝑨𝑏𝑢𝑗𝑝𝑜 =0

26

SLIDE 27

5 10 15 20v 0.2 0.4 0.6 f1

Slowing-down distribution

gGrowth > 0

The EP affects the GAM frequency

The GAM frequency is modified by FIs. (G.Y.Fu, PRL,101,185002 (2008) )

1 2 3 3 exp

exp

c EP

A v v P f e



                             Λ = 𝜈𝐶 𝐹

(T.Watari, et al, PoP, 13, 062504 (2006)

T.Ido, et.al.

2 3 3 2 3 3 3 3

exp exp

EP c c c

v v A v v v P f e v

 a 

                                     

The freq. of EP-driven GAM can be much higher than the ordinary GAM freq.

// , th ion

V k 

gGrowth < 0 gGrowth > 0

a = 6.8

𝜶 ∙ 𝑲𝑐𝑣𝑚𝑙

𝑕𝑓𝑝𝑒𝑓𝑡𝑗𝑑 + 𝑲𝑞𝑝𝑚𝑏𝑠𝑗𝑨𝑏𝑢𝑗𝑝𝑜 =0

𝜶 ∙ 𝑲𝑐𝑣𝑚𝑙

𝑕𝑓𝑝𝑒𝑓𝑡𝑗𝑑 + 𝑲𝐹𝑄 𝑕𝑓𝑝𝑒𝑓𝑡𝑗𝑑 + 𝑲𝑞𝑝𝑚𝑏𝑠𝑗𝑨𝑏𝑢𝑗𝑝𝑜 =0

27

SLIDE 28

Mechanisms which might explain this phenomena

• 1. Change of orbit topology by the GAM. activities
• 2. Deformation of fast-ion energy spectra by the mode
• 3. Improvement in bulk-ion energy confinement with the

mode

• 4. Enhanced ion-heating with the mode

28

SLIDE 29

Effect of electrostatic potential

• n the topology of particle orbit

29

Effect of ExB-drift on the deviation of particle

• rbit is small. Thus, it seems to be difficult to

claim this is responsible to the phenomena. delta5d

5 10 15 20 25 30

• 0.5

0.5 1 | | / |B| [a.u.] r

eff./a 99

ϕ 𝜔𝑂, 𝑢 = 𝜚0 exp − 𝜔𝑂 − 𝑐0 2 𝑏02 + 𝜚1 exp − 𝜔𝑂 − 𝑐1 2 𝑏12 𝑡𝑗𝑜 −𝜕𝑢

• 0.3
• 0.15

0.15 0.3

• 0.3
• 0.15

0.15 0.3



0=10kV/ 1=0kV



0=0kV/ 1=10kV



0=0kV/ 1=0kV

Z/a (R-R

0)/a

Start Point

SLIDE 30

The change of orbit topologies would change the Ti-profile.

30

The observed change seems to correspond to the increase of Ti central value rather than the change of the shape.

0.2 0.4 0.6 0.8 1

b)

n0 [x1016m-3] 0.5 1 1.5 2

1.0*(1-2)4+0.1 1.5*(1-2)4+0.1 1.5*(1-2)3+0.1 1.5*(1-2)+0.1

T

i [keV]

a)

0.02 0.04 0.06 0.08 0.1

c)

ne [x19 m-3]

ne[x1019m-3] 1 2 3 4 5 0.2 0.4 0.6 0.8 1

d)

Te [keV]

Te [keV] r/a

AURORA 10 1019

19

10 1020

20

10 1021

21

10 1022

22

10 1023

23

10 103 10 104 10 105 10 106 10 107 0.0 0.0 2.0 2.0 4.0 4.0 6.0 6.0 8.0 8.0 10.0 10.0 Neutral Spectra Neutral Spectra

1.79128+-0.00125s 1.79128+-0.00125s 1.79378+-0.00125s 1.79378+-0.00125s

1.0 x (1-(r/a)2)4+0.1 1.5 x (1-(r/a)2)4+0.1 1.5 x (1-(r/a)2)3+0.1 1.5 x (1-(r/a)2)+0.1

Experiment [a.u.] Experiment [a.u.] Calculation [a.u.] Calculation [a.u.] Energy[keV] Energy[keV]

Ti-fit=0.89 keV 1.23keV 1.30keV 1.26keV

SLIDE 31

Mechanisms which might explain this phenomena

• 1. Change of orbit topology by the GAM. activities
• 2. Deformation of fast-ion energy spectra by the mode
• 3. Improvement in bulk-ion energy confinement with the

mode

• 4. Enhanced ion-heating with the mode

31

SLIDE 32

When the energetic particle spectra was deformed, it might increase ion-heating power, classically.

Because, the population of energetic particles close to the Ec(~15Te ) would increase by the

• deformation. Thus, the ion-

heating rate might increase.

32



10

16

10

17

10

18

10

19

10

20

b)

N-flx@t=4.4175 N-flx@t=4.4225 N-flx@t=4.4275 N-flx@t=4.4325

Neutral Flux [a.u]

• 2.00 10

18

• 1.50 10

18

• 1.00 10

18

• 5.00 10

17

0.00 5.00 10

17

1.00 10

18

0.00 50.00 100.00 150.00 200.00

c)

@t=4.4225 @t=4.4275 @t=4.4325

 Energy [keV]

Te(0)=5keV Ec=75keV

a) b) Time-dependent analysis of NB heating power was performed based on TASK3D-code to evaluate the classical bulk-ion heating power by EPs . The possible increase of the heating power by the deformation was examined by introducing artificial deformation of the spectra in the calculation.

SLIDE 33

Increment of classical ion heating power by the deformation of the spectra is too small to cause observed increase of the ion temperature.

• Energetic particle spectra at t=4.94s with classical slowing-down spectra were evaluated by

TASK3D code.

• The spectra was artificially deformed to reproduced the observed clump-hole formation. The

increment of ion heating power by the deformation was examined.

33

• 3 10
• 3
• 2 10
• 3
• 1 10
• 3

1 10

• 3

2 10

• 3

3 10

• 3

50 100 150 200 b) Difference of fast-ion spectra [a.u.] Energy [keV] 1 10

• 3

2 10

• 3

3 10

• 3

4 10

• 3

a)

Slowing-down Artificially deformed

Fast ion spectra [a.u.]

r/a=~0.02

0.05 0.1 0.15 0.2 0.2 0.4 0.6 0.8 1 P

ion total

Pion

beam

P

e->i

P

ion deformation

P

ion [kW/m3]

r/a

This candidate is also ruled out.

SLIDE 34

Evaluated energy loss of EP by the deformation of the spectra reaches to 0.6[kJ/m3] at the core.

• Considering the duration time of the mode (~8[ms]), the

average power loss of EPs are ~75[kW/m3], which is larger than the necessary ion heating power (~4[kW/m3]).

0.2 0.4 0.6 0.8 0.2 0.4 0.6 0.8 1 25 50 75 100 Wbeam [kJ/m3] r/a Pbeam

loss [kW/m3] (tduration=8ms)

34

SLIDE 35



10

16

10

17

10

18

10

19

10

20

b)

N-flx@t=4.4175 N-flx@t=4.4225 N-flx@t=4.4275 N-flx@t=4.4325

Neutral Flux [a.u]

• 2.00 10

18

• 1.50 10

18

• 1.00 10

18

• 5.00 10

17

0.00 5.00 10

17

1.00 10

18

0.00 50.00 100.00 150.00 200.00

c)

@t=4.4225 @t=4.4275 @t=4.4325

 Energy [keV]

During the GAM bursting activity, the Clump-Hole formation was observed in high energy part, and the ion temperature increase was observed for the low energy part.

35

Low energy range High energy range

Neutral



Neutral



NBI (175keV)

• 2.00
• 1.50
• 1.00
• 0.50

0.00 0.50 1.00 1.50 2.00 4.415 4.420 4.425 4.430 4.435 Mirnov (ch.=0) dB/dt [T/s] time [s]

1016 1017 1018 1019 1020 1021 1 2 3 4 5 6 4.9325s 4.9475s Neutral flux [a.u.] Energy[keV] Ti

eff.=430eV

Ti

eff.=690eV