All opinions expressed herein are those of the authors and should not be reproduced, quoted in publications, or
used as a reference without the author’s consent.
Editor: James A. Rome Issue 163 December 2018
E-Mail: James.Rome@stelnews.info Phone: +1 (865) 482-5643
On the Web at https://stelnews.info
Diagnosing the edge plasma in
Wendelstein 7-X with an alkali
atom beam
One key aim of the Wendelstein 7-X (W7-X) experiment
is to study plasma exhaust using an island divertor. In the
standard  = 1 configuration of this experiment, 5 independent
islands surround the confined plasma. Each of
these islands connects to itself after one turn in the toroidal
direction. The confined plasma is bounded by a separatrix—
the dividing surface between magnetic flux surfaces
in the core plasma, and the islands. The islands are intersected
by divertor plates where the plasma is neutralized
and the resulting gas can be pumped out. Due to the pumping
gap between the two divertor plates, not all magnetic
surfaces are cut and there is a region around the O-point
where field lines are connected neither to the divertor
plates nor to the core plasma. The situation is additionally
complicated by finite plasma pressure and manufacturing
tolerances that distort the ideal configuration; therefore, it
is essential to have a good measurement of the plasma
parameters in this complicated 3D environment. Such a
measurement is provided by reciprocating probes which
can briefly move various probe heads into the island
region and diagnose it in detail. However, the probes cannot
enter too deeply because they are damaged by the
plasma or the plasma is strongly perturbed by them. An
alternative technique is the thermal helium-beam diagnostic,
which is less perturbing but can also reach only the
outer part of this region. Deeper in the plasma, reflectometry
is the key diagnostic, but it has difficulties when the
density profile becomes too complex, especially when
local maxima appear.
To cover the important transition region from the confined
edge plasma to the island region, designers of W7-X foresaw
a lithium-beam diagnostic, which is used on many
tokamak experiments and was also successful on the predecessor
stellarator Wendelstein 7-AS [1,2]. As shown in
Fig. 1, such a diagnostic extracts a few-milliampere current
of lithium ions from a heated ceramic source and
accelerates the ions to energies of a few tens of kilovolts.
The few-cm-wide beam passes through a cell filled with
sodium vapor where lithium ions pick up an electron, and
thus a high-energy neutral lithium beam is injected into the
plasma. By using some deflection plates before the neutralizer,
the beam can be steered to different directions, or
with high enough deflection voltage, even prevented from
reaching the plasma.
The neutral beam can freely propagate through the magnetic
field of the fusion device. Beam atoms get excited by
the plasma particles and emit light at a characteristic
wavelength. The light intensity is proportional to the excitation
rate coefficient and the plasma density. As the rate
coefficient has only a slight dependence on temperature,
the emitted light is nearly proportional to the plasma density.
However, there is a catch! After excitation the atoms
reside in the excited state for some time before undergoing
a spontaneous transition with light emission. During this
time the atoms travel some distance which, due to the high
beam energy, is 2–3 cm in the case of lithium. This means
small and steep density structures are smeared along the
beam injection path. Deeper in the plasma, the beam atoms
are also ionized and leave the beam on a Larmor trajec-
In this issue . . .
Diagnosing the edge plasma in Wendelstein 7-X
with an alkali atom beam
An alkali beam diagnostic with a 50-keV sodium beam
measures the density profile and turbulence in the
edge and island region of Wendelstein 7-X. Efficient
optics and detectors enable time resolution in the 10
microsecond range. Outward propagating blobs (filaments)
are abundant, and their propagation is affected
by the island configuration. ..................................... 1
Stellarator News -2- December 2018
tory. The result is that the beam light decays and no measurement
is possible deep in the plasma. However, beam
ionization also helps because the beam light decay sets the
absolute magnitude of the plasma density. Using an appropriate
model of the excitation and ionization processes the
plasma density distribution can be determined from the
relatively calibrated beam light distribution. This method
also has high time resolution since the beam particles
travel with about 106 m/s velocity. They cover a typical
plasma edge in much less than a microsecond, creating a
snapshot of the plasma density even if plasma turbulence,
which is at frequencies typically below 1 MHz is considered.
The limitation is the collected light intensity and
background light, which call for large optical systems and
good detectors.
For W7-X, a port was dedicated to the lithium beam diagnostic
in the equatorial plane in the so-called banana shape
cross-section, through the center of the edge island chain
of the standard configuration. For the light observation, a
large-diameter port looking at the edge plasma was allocated.
A Eurofusion grant for this diagnostic was given to
the Wigner Research Center for Physics in Budapest, as
this group has extensive experience in this field [3].
Due to the expected steep and complicated density profile,
concerns were raised as to whether the Li-beam diagnostic
can provide the necessary spatial resolution. Fortunately,
the applicability of sodium as beam material was demonstrated
earlier by IPP researchers on ASDEX Upgrade [4],
and the necessary atomic physics data were also prepared
by Technical University Wien. Therefore, a sodium ion
source was built and the alkali beam injector [5] installed
on W7-X in the last weeks of the 2017 OP 1.2a campaign.
After considering various options for the optical setup it
was decided to avoid any complicated endoscope design
and build a 200-mm-diameter optical system on top of the
port. An endoscope with shutter and cooled window
would have limited the first lens size so that the collected
light intensity would have been the same as on top of the
port. In the chosen solution there is no need to protect the
window even in the steady-state phase of W7-X in OP 2.
Following earlier experience on tokamaks, a few percent
of the light was coupled to a CMOS camera in order to get
a direct observation of the beam shape in the plasma. Most
of the light is detected by avalanche photodiode detectors
[6] (APDs) with 500-kHz bandwidth amplifiers, so that
plasma turbulence can be measured. In order to keep good
radial resolution and to maximize light intensity, the light
collection area of one APD channel was set to 0.5 4 cm
(radial  toroidal). This could not be coupled effectively to
the circular detectors with lenses; therefore a fiber bundle
array was built with 0.6 5 mm input and 2-mm-diameter
circular output face for each of the 40 detector channels.
Outside the islands low light level was expected; therefore
14 MPPC detectors were also installed, providing a considerably
better signal-to-noise ratio—below 109 photon/
s. The whole detector system is built into a standard APDCAM-
10G framework from Fusion Instruments used on
many fusion devices. As this device transmits data on a
10-Gbit digital data link directly to PC memory, megahertz
rate sampling is possible even during the planned 20-
to 30-minute long discharges in OP 2.
A special problem was filtering of the beam light in order
to remove plasma radiation. At the expected high plasma
densities, Bremsstrahlung and impurity line radiation was
expected to be significant. Therefore a narrow filter was
needed. Interference filters have sharp cutoff in wavelength
when the passing light is nearly parallel, which can
be achieved only with a large-diameter filter, in case of the
W7-X alkali beam, 240 mm. Such a large filter could not
be manufactured; therefore it was designed using 4 quadrants
[Fig. 2(c)]. In contrast to lithium, sodium atoms radiate
at two close wavelengths separated by 0.6 nm.
Therefore a 1-nm-bandwidth filter was manufactured in
four quadrants by Andover Corp. and assembled in a
frame. To measure the remaining background, a fast HV
switch was built for the deflection plates, producing the
ability to switch the beam on-off at several 100 kHz; thus
even a fast modulated background can be subtracted [3]
(c)
(a)
(b)
Stellarator News -3- December 2018
and density profiles can be determined with about 20-
microsecond time resolution. Indeed, first measurements
in 2017 revealed excellent signal quality. However,
extremely high and modulated background was found in a
narrow radial region on the outboard side of the island.
This background was identified as the radiation of the
thermal sodium atoms escaping from the neutralizer cell,
excited and ionized at the far plasma edge. The strong
modulation is due to blobs (filaments) arriving at the
plasma edge, similar to tokamak observations. The fast
beam modulation proved to be essential here; the density
profile could be calculated even when the background
light was nearly 100 times the local beam light! For the
2018 campaign the neutralizer was switched to potassium
which turned out to be even more beneficial than sodium
as it requires lower temperatures. This change completely
removed the strong background and enabled measurement
to the far scrape-off layer. Although this extremely complex
diagnostic was not available due to technical problems
in some part of the OP 1.2b campaign it provided
excellent data in about half of the operation.
Although data analysis has just started, some interesting
observations can already be made. Perturbations similar to
blobs (filaments) in tokamaks are always observed with
the alkali beam diagnostic in the edge of W7-X. Correlation
with the video camera system shows that these perturbations
have a toroidal length of at least several meters and
a quasi-periodic structure in the poloidal direction. They
propagate radially outward at a few hundred m/s and propagation
is clearly affected by the presence of the island, as
they appear to go around the O-point. Strong periodic
modulation of the edge plasma with a few 100 Hz is often
observed in W7-X [7]. This is obvious in the alkali beam
data as well. Enlarging one such perturbation event, it also
becomes clear that the filament activity is also modulated
[Fig. 3(a) and (b)]. As expected in the standard configuration,
the density profile is strongly affected by the island.
(a)
(b)
(c)
(d)
(e)
Stellarator News -4- December 2018
In some cases it even shows a peak close to the O-point
[Fig. (c)–(e)].
The alkali beam diagnostic is ready for long-pulse operation
in OP.2. A 100-s-long measurement has already been
demonstrated. Some spectral measurements were also
done in 2018; they are being analyzed in order to explore
any possibilities arising from spectrally resolved measurements,
like charge-exchange spectroscopy or magnetic
field measurement from Zeeman splitting of sodium lines.
Sandor Zoletnik
Wigner RCP, Budapest
E-mail: zoletnik.sandor@wigner.mta.hu
References
[1] K. McCormick et al., Fusion. Eng. Des. 34–35, 125
(1997).
[2] S. Zoletnik et al., Phys. Plasmas 6, 4239 (1999).
[3] S. Zoletnik et al., Rev. Sci. Instrum. 89, 10D107 (2018).
[4] E. Wolfrum et al., J. Nucl. Mater. 390–391, 1110
(2009).
[5] G. Anda et al., Rev. Sci. Instrum. 89, 013503 (2018).
[6] D. Dunai et al., Rev. Sci. Instrum. 81, 103503 (2010).
[7] G. Wurden et al., “Quasi-continuous low-frequency
edge fluctuations in the W7-X stellarator,” P5.1077,
EPS 2018 Conf. (2018).