See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/37459268

Experimental test of damping models for n 1 toroidal Alfv n eigenmodes in JET

Article  in  Nuclear Fusion · June 2003

DOI: 10.1088/0029-5515/43/6/310 · Source: OAI

CITATIONS READS

24 30

5 authors, including:

Andre Jaun O. Sauter

Lifelong-learners École Polytechnique Fédérale de Lausanne
69 PUBLICATIONS   699 CITATIONS    674 PUBLICATIONS   8,739 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

tokamak physics View project

Neutronics computational support for the ITER Electron Cyclotron Heating (ECH) View project

All content following this page was uploaded by O. Sauter on 04 June 2014.

The user has requested enhancement of the downloaded file.

 submitted to Nuclear Fusion, October 2002

Experimental Test of Damping Models for n=1 Toroidal Alfvén

Eigenmodes in JET

D.Testa1,2, G.Y.Fu3, A.Jaun4, A.Fasoli1,2, O.Sauter1, and JET-EFDA contributors∗

[1] CRPP, Association EURATOM – Confédération Suisse, EPFL, Lausanne, Switzerland
[2] Plasma Science and Fusion Center, Massachusetts Institute of Technology, Boston, USA
[3] Princeton Plasma Physics Laboratory, Princeton, New Jersey, USA
[4] NADA VR – Euratom Association, Royal Institute of Technology, Stockholm, Sweden

Abstract
The damping rate of the n=1 TAE mode in Ohmic-heated JET plasmas has been measured
and compared with the prediction of the kinetic NOVA-K code, which includes electron and
ion Landau damping of the global shear-Alfvén wave field, collisional damping, and radiative
damping. It was found that the calculated damping rate is too small to account for the
measured value under these experimental conditions. A relatively weak dependence of the
damping rate on the normalised Larmor radius is observed experimentally.

PACS classification scheme: 52.35.Bj, 52.55.Fa.

∗
See annex 1 of J.Pamela et al., “Overview of Recent JET Results”, OV-1/1.4, Fusion energy 2002 (Proc. 19th
International Conference, Lyon, 2002) IAEA, to be published.
 submitted to Nuclear Fusion, October 2002

A major issue in future burning plasma experiments will be the control of the stability of the
alpha particle population (α’s). As resonant interaction between the α’s and Toroidal Alfvén
Eigenmodes (TAEs) [1] can cause radial redistribution of the α’s and may lead to net particle
losses [2], large experimental and theoretical efforts are undertaken to assess the TAE
stability in reactor relevant conditions.
This work focuses on damping mechanisms of n=1 TAEs. The understanding of damping
mechanisms and their scaling is important for validating theoretical models and predictions of
TAE stability in larger tokamak experiments such as ITER. Direct measurements of the mode
frequency and damping rate γ/ω, obtained with the JET dedicated AE diagnostic system [3-
6], are compared with the predictions of the kinetic code NOVA-K [7,8]. The experiment was
originally designed to test the NOVA-K model for the radiative damping, which results from
the coupling between a MHD TAE mode and kinetic Alfvén waves. For the parameters and
profiles realised in the experiments reported here, however, the radiative damping is not the
main damping mechanism and is much too small to account for the measured values. The
dominant damping mechanism in the NOVA-K calculation is the electron Landau damping
which is still a factor 20 smaller than the measured value.
The experiments presented here investigate the dependence of the damping rate for n=1
TAEs on the ion Larmor radius ρi, with the aim of assessing the validity and relevance of the
radiative damping model. The radiative damping rate, γ/ωRAD, is a strong function of the ion
Larmor radius through the kinetic parameter λ [7-10]:
1/ 2
mσ  3 Te  (1)
λ = 4 3/ 2  +  ρi .
rε m  4 Ti 
Here σ=(r/q)dq/dr is the magnetic shear, R=R0+r is the major radius (R0 is the position of the
magnetic axis and r the radial coordinate along the minor radius), q(r) is the safety factor
profile, Te and Ti are the electron and ion temperature, respectively. All the experimental
quantities entering the definition of λ, hence the calculation of γ/ωRAD, are computed at the
gap position rTAE for the TAE with toroidal mode n and poloidal mode numbers (m, m+1).
The parameter εm=5rTAE/2RTAE is a measure of the frequency width of the toroidal gap. In the
limit of small magnetic shear σ<<√8/π, the expression for γ/ωRAD reduces to [7-10]:

 γ  π2  π 3σ 2  (2)
= ε σ 2
× exp − 7 / 2  .
 ω  RAD 8
m
 2 λ
 submitted to Nuclear Fusion, October 2002

Two in-vessel antennas are used with opposite phasing to preferentially excite n=1 AEs. Due
to the extreme increase of γ/ ω for n=1 TAEs with increasing edge magnetic shear, elongation
and triangularity, discovered in previous experiments [4-6], a limiter plasma configuration
with low magnetic shear at the gap position, σGAP≈0.25<√8/π, is used. Similarly, the
elongation and triangularity are very low, 1.18≤κGAP≤1.23 and 0.01≤δGAP≤0.03 respectively,
and are held approximately constant during the limiter phase of the discharge. The q-profile
is monotonic, with q0=q(ψN=0)≈0.75÷0.85 and q95=q(ψN=0.95)≈2.65÷3.3, as obtained using
a magnetic reconstruction of the equilibrium constrained by the internal motional Stark effect
(MSE) and polarimetry measurements. Here ψN is the radial coordinate in units of the
normalised poloidal flux, ψN(r)=ψ(r)/ψ(r=a). These values are confirmed by the position of
the sawtooth inversion radius, as deduced from the electron cyclotron emission measurements
of Te. The electron density (ne) and temperature radial profiles show a typical L-mode shape
throughout the evolution of the discharge. The edge density and temperature are very low, of
the order of ne95/ne0≈Te95/Te0≈0.05. Diagnostic Neutral Beam Injection (NBI) heating up to a
power PNBI=3MW is used for the MSE and charge-exchange measurements of the q- and Ti
profile. However, due to the effect of the NBI ions on the n=1 TAE spectrum [11], the NBI
heating phase of the discharge is not considered here. The ion temperature is not directly
measured during the ohmic phase of the discharge, thus it is inferred from the measured Te
and neutron rate profile.
Figure 1 shows the main plasma parameters during the ohmic limiter phase of #52206,
representing the typical operating scenario for the experiments reported here. To be able to
discriminate between the radiative and other damping mechanisms, the magnetic field and
plasma current are ramped down at fixed q95 and edge shape in order to change ρi, hence λ,
affecting as little as possible the other background plasma parameters. Figure 2 shows the
radial profile of q-, ne and Te at various time points of interest during #52206. Figures 3 and 4
show respectively the shape of the poloidal cross-section of the plasma and the n=1 TAE gap
structure for #52206 at t=61s.
Figure 5 shows the measured, γ/ωMEAS, and computed damping rate for a n=1 TAE during the
ohmic phase of #52206. The gap position for n=1 TAEs, inferred using reflectometer and
ECE measurements, is located around ψN,GAP≈0.45, and the gap width ∆GAP≈εmrTAE/m [12] is
of the order of ∆GAP≈10cm. In addition to the radiative damping, γ/ωRAD, NOVA-K also
includes the electron Landau damping γ/ωELE and collisional damping due to trapped
electrons. The mode frequency decreases due to decrease in the magnetic field, and there is a
 submitted to Nuclear Fusion, October 2002

very good agreement between the measurements and the NOVA-K calculation. However, the
computed damping rate (mainly from electron Landau damping) is about a factor 20 smaller
than the measured one, with the radiative damping being even smaller.
Figure 6 shows the measured damping rate γ/ωMEAS versus the normalised Larmor radius at
the gap location ρ*iGAP=ρiGAP/a (where a is the plasma minor radius) for the database
collected during the ohmic limiter phase of the five discharges considered in this work. The
selection presented here is obtained with 2.65≤q95≤3 and 0.2≤σGAP≤0.3, and represents
approximately 70% of the complete data set, the remaining 30% including data obtained
during the NBI heating phase and at higher q95 and σGAP. The database covers a large range in
the density and temperature at the gap location, respectively 1≤neGAP(1019m-3)≤3 and
0.8≤TeGAP(keV)≤2.5, which, together with the variation in q95 and σGAP, is the origin for the
observed scatter in the data.
It is interesting to note that the main trend of the data indicates that the measured damping
rate is larger for smaller gyroradius. This is opposite to the scaling of the radiative damping
and other simple kinetic models. The experimental results show an approximate dependence
γ/ωMEAS≈1/ρ*iGAP for 2.5≤ρ∗iGAP(10-3)≤3.2, whereas at higher ρ∗iGAP, 3.2≤ρ∗iGAP(10-3)≤5,
γ/ωMEAS becomes practically independent on ρ∗iGAP. A similar trend is also obtained with
respect to λGAP. We have calculated the radiative damping rate using Eq.(2) for all these data
points and the value is about 20-50 times smaller than γ/ωMEAS, albeit increasing strongly with
ρ*iGAP and λGAP. This confirms the NOVA-K results shown in Fig.2 for three of these cases.
It is interesting to discuss the detailed measurements of the dependence of the n=1 TAE
damping rate upon ρ∗i and λ presented here in light of similar earlier work for ellipticity-
induced AEs on JET [13] and high-n TAEs in DIII-D [14,15]. First, the dependence of γ/ω
upon ρ∗i is similar to that reported for the earlier JET experiments [13], which, however, do
not show any correlation with λ, consistent with the data presented here, γ/ωRAD<<γ/ωMEAS.
However, here we observe an approximate dependence γ/ωMEAS≈1/λ for 2.5≤λ(10-2)≤3.5,
whereas γ/ωMEAS does not depend on λ for λ>0.04. We attribute this different scaling to the
limited amount of data points available in Ref.[13] and to the fact that the plasma parameters,
and particularly the edge elongation and magnetic shear, were not well matched between the
various discharges considered in these earlier studies. This could have possibly masked the
more subtle dependencies that can only be recognised by using better matched discharges and
a much larger number of data points. Second, in the DIII-D experiments [14,15] the damping
 submitted to Nuclear Fusion, October 2002

rate was not directly measured but inferred comparing the marginal stability limits for n=3-5
TAEs, assuming knowledge of the fast ion drive, i.e. of the fast ion distribution function and
its interaction with the mode. The observed stability thresholds for high-n TAEs were
accounted for by radiative damping. However, the intrinsic limitation of these indirect
estimates of the damping motivated the investigation presented herein, based on a direct
measurement of the mode damping rate. Further work is needed to resolve the discrepancy
between the theoretical damping models examined in this work and the experimental results.
The NOVA-K model has been tested on a number of other experiments, for instance giving a
good estimate of the stability threshold for α-driven TAEs in TFTR [7,8]. Thus, to improve
the understanding and modelling of the low-n TAE stability limits in JET, the missing
damping mechanism needs to be identified for the experimental conditions reported in this
work. To this aim, detailed comparisons with other models could be useful. As an example,
for one timeslice in the NOVA-K calculation, the frequency intersects with the Alfvén
continuum near the edge, due to the very low edge density in these experiments. In such cases
the continuum damping [16] should be included in the calculation of the global fluid
wavefield. Toroidal mode conversion provides for other mechanisms, which have previously
been found in agreement with the measurements under similar conditions [17,18].
Furthermore, NOVA-K predicts the radiative damping to increase with the toroidal mode
number and the ion Larmor radius. Thus the model needs to be further tested for TAEs with
higher toroidal mode numbers in conditions where a larger ρ*i is achieved. A new set of high-
n TAE antennas is being designed for installation on JET, with the aim to provide the data
required to validate code predictions for these modes.

The authors would like to acknowledge the contribution of the whole JET experimental team,
and particularly T.Bolzonella, D.Borba, C.Giroud, N.Hawkes, T.Hender, F.Milani,
V.Riccardo, S.Sharapov, P.de Vries, and K.-D.Zastrow. The authors would also like to thank
the Referees for their comments and suggestions, which have contributed to increase the
clarity of this Paper.
This work has been conducted under the European Fusion Development Agreement. D.Testa
and A.Fasoli were partly supported by DoE contract No. DE-FG02-99ER54563.
 submitted to Nuclear Fusion, October 2002

References.
1) C.Cheng, L.Chen, M.Chance, Ann. Phys. 161 (1985), 21.
2) K.Wong, Plasma Phys. Controlled Fusion 41 (1999), R1.
3) A.Fasoli et al., Phys. Rev. Lett. 75 (1995), 645
4) A.Fasoli et al., Phys. Plasmas 7 (2000), 1816.
5) D.Testa, A.Fasoli, Nucl. Fusion 41 (2001), 809.
6) A.Fasoli, A.Jaun, D.Testa, Phys. Lett. A265 (2000), 288.
7) C.Cheng, Phys. Rep. 211 (1992), 1.
8) G.Fu et al., Phys. Plasmas 3 (1996), 4036.
9) R.Mett, S.Mahajan, Phys. Fluids B4 (1992), 2885.
10) S.Sharapov, MHD Modelling of Alfvén Eigenmodes in Tokamaks, Proceedings of the 6th
Easter Plasma Meeting, Turin, 16-19 April 2000, to be published.
11) D.Testa, A.Fasoli, A.Jaun, Measurement of the damping rate of n=1 Toroidal Alfvén
Eigenmodes as a function of the neutral beam heating power and plasma β on JET,
submitted to Nucl. Fusion, July 2002.
12) G.Vlad, F.Zonca, S.Briguglio, Nuovo Cim. 7 (1999), 1.
13) W.Heidbrink, A.Fasoli, D.Borba, A.Jaun, Phys. Plasmas 4 (1997), 3663.
14) R.Mett, E.Strait, S.Mahajan, Phys. Plasmas 1 (1994), 3277.
15) E.Strait, W.Heidbrink, A.Turnbull, M.Chu, H.Duong, Nucl. Fusion 33 (1993), 1849.
16) F.Zonca, L.Chen, Phys. Rev. Lett. 68 (1992), 592.
17) A.Jaun, K.Appert, J.Vaclavik, L.Villard, Comput. Phys. Commun. 92 (1995), 153.
18) A.Jaun, A.Fasoli, W.Heidbrink, Phys. Plasmas 5 (1998), 2952.
 submitted to Nuclear Fusion, October 2002

Figure captions.
Figure 1. The main plasma parameters during the ramp-down phase of the magnetic field for
#52206, representing the typical operating scenario for the experiments reported here. The
data are evaluated at the gap position for the n=1 TAE, ψN,GAP≈0.45.
Figure 2. The radial profile of q-, ne and Te at various time points of interest during #52206.
Figure 3. The shape of the poloidal cross-section of the plasma for #52206 at t=61sec.
Figure 4. The structure of the n=1 Alfvén continuum for #52206 at t=61sec, computed using
the CSCAS and Helena equilibrium reconstruction codes.
Figure 5. The measured frequency and damping rate for a n=1 TAE during the ohmic heating
phase of #52206, compared with the radiative and trapped electron contribution, as computed
by NOVA-K. Note the large multiplicative factor between the measured damping rate and the
NOVA-K results.
Figure 6. The dependence of the measured damping rate for n=1 TAEs upon the normalised
Larmor radius at the gap position ρ∗iGAP during the ohmic limiter phase of the five discharges
considered here, which have a very similar q-profile within the uncertainty on the equilibrium
reconstruction. Note the decrease in γ/ωMEAS≈1/ρ∗iGAP for 2.5≤ρ∗iGAP(10-3)≤3.2, whereas
γ/ωMEAS becomes practically independent on ρ∗iGAP for 3.2≤ρ∗iGAP(10-3)≤5.
 submitted to Nuclear Fusion, October 2002

#52206: plasma parameters at n=1 TAE gap position

2
BΦ0(T)
1.5 Ip(MA)

0.5
2

1
neGAP(1019m−3)
TeGAP(keV)
0
3
σGAP
2 σ95
q95
1

0
59 59.5 60 60.5 61 61.5 62 62.5 63
time (sec)
Figure 1. The main plasma parameters during the ramp-down phase of the magnetic field for
#52206, representing the typical operating scenario for the experiments reported here. The
data are evaluated at the gap position for the n=1 TAE, ψN,GAP≈0.45.

D.Testa et al., Figure 1.

 submitted to Nuclear Fusion, October 2002

#52206: q, Te and ne radial profiles

4 t=59.05s
t=60.05s
3 t=61.05s
q−profile

t=62.05s
2 t=62.96s

1.5

1
Te(keV)

0.5

3
ne(1019m−3)

0
0 0.2 0.4 0.6 0.8 1
Figure 2. The radial profile of q-, ne and Te at various time points of interest during #52206.

D.Testa et al., Figure 2.

 submitted to Nuclear Fusion, October 2002

Pulse No: 52206 Plasma shape at 21s

2

1
Height (m)

-1
JG03.34-1c

2.0 2.5 3.0 3.5 4.0

Major radius (m)
Figure 3. The shape of the poloidal cross-section of the plasma for #52206 at t=61sec.

D.Testa et al., Figure 3.

 submitted to Nuclear Fusion, October 2002

#52206/t=61s: n=1 TAE continuum spectrum

2 . 00 xxx xx x
xxx xx x
xx xx x
xxx xx x x
xxx xx x
xxx xx xx
xxx xx x
xxxxxxxxxxxxxxxxxxxxx xxx xxx
xxxxxxxxxxxxxxxx xxx x x x
xxxxxxxxxxx
xxxxxxxxx xxx xx xx
xxxxxxxx
xxxxxxxx xxx xxx xxxx x
xxxxxxx xxxx x xxx x
xxxxxx xxx xx xxxx
xxxxxx xx xx
x
xxxxx
xxxxx
xxx x
xx xxxxxx
1 . 50 xxxxxx xxx
x x
xxxx x
xxxxxx xxx xx x
xxxxxx xxxxxxx x
xxxxx x
xxxxx xxx
xxxxx xxxxx xxx x
xxxxxx x x
xxxxxx xxxxx x
x
xxxxxxx xx
xxxx xx x
xxxxxxxx xxxxx xxx
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxx xx
normalized frequency

xxx x
xxx xxxx
xxxxxx
1 . 0 0 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
xxxxxx
xxxxx
xxxx
xxxxx
xxx x
xxxx x
xxxx
xxx x
xxxxxxxxxxxxxx xx
xxxxx xxx xxxxxxxx
xxx
xxxxx xxxxxx
q-profile (modelled) xxxxxxxxxxxxxxxxxx

0 . 50

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx normalized density profile (modelled)

xxxxxxxxxxxxxxxxx
xxxxxxxxxxxxxx
xxxxxxxxxxxx xxxxxxxxx
xxxxxxxxxxx xxxxxxx xxxxxxxx xxxxx
xxxxxxxxxx
xxxxxxxx xxxxxxxxxx xx
xxx xxx xx
xxxxxxxxx xxxxx xxx xx
x
xxxxxxx xxxxx xxx xx
xxxxxxx
xxxxxxx xxxxxxxxxx xx xxx
xxxxxx xx xx xx
xxxxxx
0 . 0 0 xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
0.00 0.20 0.40 0.60 0.80 1.00
normalized minor radius

Figure 4. The structure of the n=1 Alfvén continuum for #52206 at t=61sec, computed using
the CSCAS and Helena equilibrium reconstruction codes.

D.Testa et al., Figure 4.

 submitted to Nuclear Fusion, October 2002

#52206: NOVA−K model & measured n=1 TAE damping rate

120
fNOVA−K
frequency (KHz)

100 f
MEAS
f
TAE−GAP
80

60

1.5
γ/ω(%)

measured
1 100*γ/ωRAD (NOVA−K)
20*γ/ωELE (NOVA−K)
0.5

0.05
λGAP

59 59.5 60 60.5 61 61.5 62 62.5 63

time (sec)
Figure 5. The measured frequency and damping rate for a n=1 TAE during the ohmic heating
phase of #52206, compared with the radiative and trapped electron contribution, as computed
by NOVA-K. Note the large multiplicative factor between the measured damping rate and the
NOVA-K results.

D.Testa et al., Figure 5.

 submitted to Nuclear Fusion, October 2002

radiative damping: ohmic data (1.18<κGAP<1.23, q95<3)

3.5

3
n=1 TAE damping rate (%)

2.5

1.5

0.5
2 2.5 3 3.5 −3 4 4.5 5
ρ*(GAP) (10 )
Figure 6. The dependence of the measured damping rate for n=1 TAEs upon the normalised
Larmor radius at the gap position ρ∗iGAP during the ohmic limiter phase of the five discharges
considered here, which have a very similar q-profile within the uncertainty on the equilibrium
reconstruction. Note the decrease in γ/ωMEAS≈1/ρ∗iGAP for 2.5≤ρ∗iGAP(10-3)≤3.2, whereas
γ/ωMEAS becomes practically independent on ρ∗iGAP for 3.2≤ρ∗iGAP (10-3)≤5.

D.Testa et al., Figure 6.

View publication stats