32nd EPS Conference on Plasma Phys. Tarragona, 27 June - 1 July 2005 ECA Vol.29C, P-5.026 (2005)
Spatial distribution of lost ripple-trapped suprathermal electrons.
F. Medina, M. A. Ochando, A. Baciero and J. Guasp
Laboratorio Nacional de Fusión. Asociación EURATOM/CIEMAT, 28040 Madrid, Spain
Introduction
The ECRH driven convective flux of ripple-trapped suprathermal electrons in low density
stellarator plasmas has been mentioned as being related to various features and phenomena:
the suprathermal feature and the burst-like phenomena in the ECE emissions [1], the
electron-root feature [2] and the phenomenon of electric pulsation [3]. In a previous
investigation on TJ-II plasmas [4], fast transitions in confinement regimes were found to be
associated to changes in the electron distribution function and to changes in the direct
convective losses at the heating region. There, low order rational surfaces located at the
gradient region seemed to be affecting the trapping/de-trapping rates. It was also reported [5]
that the plasma potential at the core region increased with the characteristic energy of the
electron suprathermal electron tail.
Here we present the spatial distribution of electron losses to the vacuum vessel and to the
graphite limiter detected through their bremsstrahlung emission in the range of the soft x
rays. Electron losses are monitored with the SXR tomography diagnostic, using thick Be
filters to reduce the contribution of plasma emission, and a planar Ge detector to obtain the
spectra [6], see Fig. 1 and 2.
Limiter A
Fig. 2. Schematic view of the Ge detector
Fig. 1. Schematic view of the tree
inner SXR cameras.
location and a typical spectrum in the range
of 2 keV to 10 keV
32nd EPS 2005; F.Medina et al. : Spatial distribution of lost ripple-trapped suprathermal electrons
2 of 4
Results
Three distinct losses regions were observed at the inboard side of the vessel wall. The
emissions coming from each of them behave differently with electron density, heating
method (ECR on-axis vs. off-axis) and Zeff. The experiments were carried out, in steady
state plasmas as well as in plasmas showing transitions between different confinement
regimes.
The SXR spectra in TJ-II at the plasma core usually present a suprathermal component,
with characteristic energies ranging from two to four times the central Te, when measured in
the range of 4 keV to 10 keV. This ECRH generated suprathermal electrons have now been
detected at the vacuum vessel and at the limiter, and their emission spectra show only the
suprathermal component. The thermal component that, at plasma core, dominates the spectra
in the range of 1 keV to 4 keV is absent.
Fast transitions in the plasma core, as observed by the ECE channels and central SXR
chords, have an instantaneous response in the signal levels of direct losses. An increase of
ECE signals is accompanied with a reduction in the plasma centre SXR emission, as a
consequence of a decrease of plasma density, and also a sharp increase in the monitors of
direct losses both, at the vacuum vessel and at the limiter (see Fig. 3).
Density
ECE_central
SXR_central
Density
Losses_iner wall
5
Losses_limiter
#12957
ECE_central
25
2,5
#13361
4
20
2
15
1,5
10
1
3
2
1
5
0
1000
0,5
0
1050
1100
1150
1200
0
1040
Time (ms)
1080
1120
1160
1200
1240
time (ms)
Fig. 3. Fast transitions at the plasma core as observed at (left) plasma wall and (right) the
limiter.
TJ-II has two equivalent limiters. By observing the direct losses on limiter A, while radialy
scanning the limiter C, there was observed no reduction in the monitor signals. This shows
that the electrons that originate the emissions do not circulate the torus, they are trapped in
the TJ-II helical magnetic field ripple (see Fig. 4).
32nd EPS 2005; F.Medina et al. : Spatial distribution of lost ripple-trapped suprathermal electrons
3 of 4
10
Magnetic Field at Axis
Direct losses to Limiter A
1,06
8
6
4
Limiter A
at -0.362 m (LCMS)
2
0
1,04
1,02
ECRH 1
1
Ge Detector & Limiter A
0,98
ECRH 2
SXR Tomography
Diagnostic
0,96
-0,365
-0,36
-0,355
-0,35
-0,345
-0,34
-0,335
0
20
40
Position of Limiter C
60
80
100
120
140
Toroidal Angle
Fig. 4. (left), Intensity of direct losses at limiter A while introducing the limiter C inside the
plasma. (right), Magnetic field at axis. Also shown the diagnostic and the microwave
injection ports positions.
As expected, the observed direct losses intensity scales inversely proportional with density:
an increase of density, at constant input power, means a decrease of power density and
therefore a reduction in the deformation of the electron distribution function. This is also
observed as a reduction of the characteristic energy in the spectra (see Fig. 5).
y = 0,0033245 * x^(-2,3665) R= 0,988
H
I
Density
Bolometer
y = 0,0055706 * x^(-2,2831) R= 0,95681
0,045
#12957
Losses_limiter
3
6
2,5
5
2
4
0,025
1,5
3
0,02
1
2
0,04
I losses
0,035
0,03
0,015
0,5
1
GasGas
pulsepulse
0,01
0,3
0,35
0,4
0,45
0,5
0,55
0,6
Densidad
de línea
Line density
0
1020
0
1030
1040
1050
1060
1070
1080
1090
1100
Time (ms)
Fig. 5. (left), Density dependence for the shot in Fig. 3 (left), the two confinements
regimes are treated separately. (right), Time evolution after a short pulse of gas
injection.
When microwaves are launched off-axis a two-fold increment of direct losses is observed.
Although off-axis would mean a reduction of the power density, because of a wider area of
power deposition, the possibility of microwave absorption away from magnetic axis could
explain the result. Also, the poloidal distribution of direct losses to the wall is greatly
modified (see Fig. 6). While three different bands are observed in on-axis heated plasmas, in
off-axis heated plasmas one band is absent and another is shifted. Variation in the electrons
pitch angle and the radial electric field could explain the existence of these bands and their
dependence with the heating conditions.
32nd EPS 2005; F.Medina et al. : Spatial distribution of lost ripple-trapped suprathermal electrons
4 of 4
ECRH off-axis
Fig. 6.
(left), Scheme:
the three regions
where direct
losses were
observed.
(right), Direct
losses for two
heating scenarios.
ECRH on-axis
The dependence of direct losses with Zeff is counterintuitive: an increase of plasma
collisionality should reduce the electron distribution function deformation and therefore the
direct losses. However, there are previous observations indicating that direct losses in TJ-II
are toroidally asymmetric [4].
Density
Bolometer
1,2
#12424
Losses_limiter
20
3
1
SXR_losses
2,5
15
0,8
Carbon
impurity
2
#12524
0,6
1,5
10
1
0,4
5
#12796
0,2
0,5
0
1100
0
1
1,5
2
2,5
Zeff
0
1120
1140
1160
1180
1200
1220
Time (ms)
Fig. 7. (left), Dependence of direct losses to the wall as a function of Zeff for three discharges
with identical average electron line density. (right), Time evolution after of sudden entrance of
carbon.
The increase of the observed direct losses could be the result of a double collisional process.
As a result of an increase of collisionality more electrons could be pumped away,
“untrapped”, from the magnetic wells close to the microwave injection ports, being then,
susceptible to be trapped again in the helical ripples where they are finally observed (see Fig.
4 right). Notice that, if this hypothesis were confirmed, the so far monitored direct losses
would be just a minor fraction of the total. To confirm this hypothesis a new detector will be
installed near the microwave injection ports.
References
[1] Romé M et al 1997 Plasma Phys. Control. Fusion 39 117
[2] Kick M et al 1999 Plasma Phys. Control. Fusion 41 A549
[3] Fujisawa A et al.1999 Plasma Phys. Control. Fusion 41 A561
[4] Ochando M A and Medina F et al 2003 Plasma Phys. Control. Fusion 45 221
[5] Medina F. et al Proc. 31st EPS (London 2004)
[6] Medina F. et al, RSI 1999