ABSTRACT In JET C-wall L-mode plasmas with significant ion heating by NBI, higher electron stiffness and lower values of R/LTe are observed than in plasmas with pure ICRH heating, for the same electron heat flux [1]. A possible explanation is found in ETG modes and their correlation to Te/Ti . In order to study the level of ETG electron heat flux, a series of nonlinear gyrokinetic simulations are carried out using the GENE gyrokinetic code in the local limit [2]. Only separated electrons and ions scale simulations are carried out. This choice limits the direct comparison between simulations and experiments but nevertheless gives a strong indication of the amount of electron heat flux predicted to be related to ETG turbulence in the two experimental situations. EXPERIMENTAL OBSERVATIONS • C-wall, L-mode plasmas. B0~ 3.4 T, ICRH~ 3 MW, NBI~ 7 MW, Ip~ 1.8 - 3 MA (overshoot, ramp-up, ramp-down), ne,0~ (2 - 3)•1019 m-3, Te,0~ 5 keV, Ti,0~ 2.5 - 5 keV • R/LTe decrease in the NBI case is due to both a decrease in inverse critical gradient length and an increase in stiffness.This is in contrast with the reduction of ion stiffness observed in presence of NBI and shown in Fig.1a [2,3]. Figure 1: Effect of NBI heating on ions in L-mode JET discharges at ρtor=0.33 (left panel, reproduced from [2]) and on electrons at ρtor=0.33 (centre panel) and ρtor=0.5 (right panel) [1]. • One main difference between NBI and pure ICRH plasmas is lower values of Te/Ti. • Linear gyrokinetic simulations do not indicate strong effects of Te/Ti on TEM threshold [1]. • One possible effect due to Te/Ti is an increase of the heat flux carried by ETG modes, for which a stabilizing effect of τ = ZeffTe/Ti is expected [4]. GYROKINETIC SIMULATIONS Multiscale gyro-kinetic simulations necessary to study properly the ETG [6,7,8,9]. Such simulations demand exceeding computational resources. We carried out separate scale simulations. In the ETG simulations, we used adiabatic ions and included the measured external flow shear. This leads to ETG streamer saturation as can be seen in Fig. 2 and Fig. 3.

N. Bonanomi1,2, J. Citrin3,4 , P. Mantica1 and JET Contributors*

EUROfusion Consortium, JET, Culham Science Centre, Abingdon, OX14 3DB, UK

Impact of electron scale modes on electron heat transport in the JET tokamak

R/LTi

2 4 6 8 10

q i,gB

0

5

10

15

20

25

30

35

40

R/LTe

5 6 7 8 9

q e,gB

0

2

4

6

8

10

12

14

16

18

ICRHχs = 7

ICRH+NBI χs=0.5

ρtor = 0.33 ρtor = 0.33

ICRH+NBI χs ≈ 3

ICRHχs = 1.3

R/LTe

6 7 8 9 10

q e,gB

0

10

20

30

40

50

60

70

80

ICRH+NBI χs≈5

ICRHχs≈3.5

ρtor=0.5

• With nonlinear simulations we made a scan in R/LTe of the electron heat flux in order to make a comparison with experimental level of flux and stiffness. The results obtained at ρtor=0.5 for electrons and ions heat flux are shown in Fig. 6 and Fig. 7. Figure 7: Normalized electron heat flux vs R/LTe. Experimental flux (black triangles), GENE TEM/ITG flux (red squares), GENE ETG flux (green diamonds) and GENE TEM/ITG+ETG flux ( blue circles) for pure ICRH case(left) and ICRH+NBI case(right). Black lines indicate the experimental slope (stiffness). • Simulations indicate that non-diagonal terms, especially R/LTi, carry ~25% of the electron heat flux in the ICRH case and ~40% in the NBI case. • With ion scale modes we can reproduce only the ~50% of the experimental flux. Also, we can’t reproduce the experimental slope of the flux. • In both ICRH and NBI cases ETG modes are unstable and carry an important amount of the flux. CONCLUSIONS This work provide a comparison between experiment and gyrokinetic simulations for JET C-wall L-mode discharges with and without substantial ion heating provided by NBI. It indicates that a significant amount of electron heat flux (~25%) can be carried by non-diagonal terms. Using ion scale modes alone it’s difficult to reproduce the experimental slope and the experimental electron heat flux, reaching only the ~50% of the experimental values. Electron scale modes can help to reproduce the experimental fluxes in both ICRH and ICRH+NBI cases. Being more unstable in presence of substantial ion heating due to lower values of Te/Ti, ETG modes can help to explain why, in presence of NBI heating, we observe lower values of R/LTe and higher stiffness. ACKNOELEDGEMENTES The authors would like to thank Tobias Görler, Daniel Told and Frank Jenko for precious advice on GENE simulations. The authors are also grateful to D. R. Mikkelsen for assistance. This research used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No.DE-AC02-05CH11231. A part of this work was carried out using the HELIOS supercomputer system at Computational Simulation Centre of International Fusion Energy Research Centre (IFERC-CSC), Aomori, Japan, under the Broader Approach collaboration between Euratom and Japan, implemented by Fusion for Energy and JAEA. REFERENCES [1] N. Bonanomi et al., submitted to Nuclear Fusion [5] F. Jenko et al., Phys. Plasmas 7, (2000) [2] P. Mantica et al., Phys. Rev.Lett. 107, (2011) [6] N. T. Howard et al., Phys. Plasmas 21, (2014) [3] J. Citrin et al., Phys. Rev.Lett. 111, (2013) [7] T. Gorler, F. Jenko, Phys. Rev.Lett. 100, (2008) [4] F. Jenko et al., Phys. Plasmas 8, (2001) [8] R. E. Waltz et al., Phys. Plasmas 14, (2007)

R/LTe

5 6 7 8 9 10 11

γETG (R

/v th,e)

0

0.05

0.1

0.15

0.2

0.25τ=2.3τ=1.8τ=2.7

R/LTe

7.5 8 8.5 9 9.5 10 10.5

q e,gB

0

10

20

30

40

50

60

70

80

90

EXP, JET 78834

GENE, tot

GENETEM+ITG

GENE ETG

R/LTe

6.5 7 7.5 8 8.5 9 9.5

q e,gB

0

10

20

30

40

50

60

EXP, JET 78842

GENE, tot

GENE ETG

GENETEM+ITG

R/LTi

4 4.5 5 5.5 6 6.5 7 7.5

q i,gB

0

10

20

30

40

50

60

70

R/Lte= 10

R/Lte= 9

EXP, JET 78842

R/Lte= 8

R/Lte= 9

EXP, JET 78834

GENE

GENE

1 CNR - Istituto di Fisica del Plasma ‘P. Caldirola’, Milano, Italy 2 Università degli studi di Milano-Bicocca, Milano, Italy

3 CEA - Cadarache, St. Paul Lez Durance, France 4 FOM Institute DIFFER, P.O. box 6336, 5600 HH Eindhoven, The Netherlands

* See the Appendix of F. Romanelli et al., Proceedings of the 25th IAEA Fusion Energy Conference 2014, Saint Petersburg, Russia

Figure 2: Contour plots of the electrostatic potential of an ETG simulat ion obtained without external flow shear (left, long ETG streamers are not sa tura ted) and wi th external flow shear (right, the ETG streamers are saturated).

kyρe

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

ω (R

/vth

,e)

-1

-0.8

-0.6

-0.4

-0.2

0

γ (R

/vth

,e)

0

0.1

0.2

0.3

Adiabatic ionsKinetic ions: R/LTi = 0Kinetic ions: R/LTi = 4

Figure 5: Scan in ρeky of ETG linear growth rate and ETG linear growth rates vs R/LTe for different values of τ.

Figure 6: Normalized ion heat flux vs R/LTi. Experimental ICRH (black diamonds) and NBI case (purple triangles); GENE TEM/ITG flux for the ICRH case (blue circles) and for the NBI case (red squares).

• The experimental normalized ion heat flux is reproduced using fast ions and electromagnetic effects, confirming what found in [3].

time (R/vth,e)0 100 200 300 400 500 600

q e/(ρ2 e/R

2 v th,e

pe)

×104

0

1

2

3

4

5

6

time (a.u.) ×1040 0.5 1 1.5 2

q e/(ρe2 /R

2 v th,e

p e)

0

1

2

3

4

5

6

7 x103

Figure 3: Electron heat flux from ETG simulation obtained without external flow shear (left, unphysical amount of flux) and with external flow shear (right, the flux is comparable with the experimental ones).

γExB

0 0.01 0.02 0.03 0.04 0.05 0.06

q e (kW

/m2 )

0

5

10

15

20

25

30

35

40

45

50

EXP. JET 78842

GENE

Figure 4: Electron heat flux from ETG simulation obtained vs external flow shear.

• The ETG flux is less sensitive to shear flow than TEM/ITG flux. There is a slow reduction of the flux with external flow shear.

• A scan in γExB flow shear was made to establish its effect on ETG flux (Fig. 4)