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.
An international journal of news from the stellarator community
Editor: James A. Rome Issue 157 June 2017
E-Mail: James.Rome@stelnews.info Phone: +1 (865) 482-5643
On the Web at https://stelnews.info
Letter from the Editor:
Stellarator News has moved
Stellarator News (SN) began sometime before 1988 as “A
compendium of news, proposals, progress, and people in
the international stellarator community.” It was started by
the Fusion Energy Division of Oak Ridge National Laboratory
(ORNL) and hosted by ORNL for many years.
However, due to increased computer security at ORNL,
beginning in May 2017, I have moved SN to
https://stelnews.info, and am publishing SN independently.
SN has been published electronically every other month
when there are available contributions. As Editor, I assemble,
vet, clarify meaning, and format the content. The
English (and often the meaning) is professionally edited
by Margaret B. (Bonnie) Nestor. Her incredible editing is a
great reason to submit prepublication works to SN, even if
English is your native language. Rapid publication (a few
weeks from the due dates for issue submission) is a hallmark
of this publication.
SN accepts articles about theory, experiment, engineering,
proposals, and politics of relevance to the stellarator community.
Speculative works are welcomed, and the more
pretty color pictures that explain the content, the better.
Authors should try to make their submissions intelligible
to the non-specialist.
Stellarator News serves multiple purposes:
 Primarily it serves to coordinate efforts among the
worldwide stellarator researchers. SN provides an
opportunity to quickly convey information to the stellarator
community. It also publicizes meetings, such
as the International Stellarator Workshops and the
Coordinated Working Group Meetings, and publishes
their conclusions. In contrast to tokamaks (which
have just a handful of adjustable design parameters),
there is a wide variety of different viable stellarator
concepts, and their design, construction, engineering,
and performance are discussed herein with a view
towards coordinating and reducing duplication of
effort.
 SN informs other interested parties (other branches of
fusion research, program managers, elected officials)
about the above efforts to showcase the strength and
vibrancy of our community.
On the new Web site, I have scanned in the oldest issues
that were heretofore unavailable, and used OCR to make
all issues searchable, albeit in a kludgey way. Read the
instructions to get this feature to work. As always, the “In
this issue...” box will appear if you hover over the issue
name
SN is published in February, April, June. August, October,
and December. Submissions are due by the end of the preceding
month. Please send submissions to the Editor at
James.Rome@stelnews.info. Be sure to include a short
(few sentence) abstract for the “In this Issue” box on the
front page, and provide the author’s contact information.
James A. Rome, Editor of Stellarator News
E-mail: James.Rome@stelnews.info
In this Issue . . .
Stellarator News has moved
Stellarator News is now an independent entity at
https://stelnews.info. ............................................... 1
The bolometer diagnostic at Wendelstein 7-X
and its first results from the initial campaign
The sensitivity of the bolometer, which was originally
optimized for the higher radiation intensities of high
performance plasmas, was re-optimized to cope with
the low-density and low-heating-power limiter plasmas.
Strong radiation zones have been observed in
outer confined plasma regions, which were attributed
to low Z-impurity ion emissions. The radiated power
was observed in plasmas terminated by stopping the
heating power, or due to radiative collapse. The radiated
power fraction Prad/PECRH was studied vs ne
For a given heating power, it increases with line-averaged
density and decreases with increasing heating
power for a fixed ne. ............................................ 2
Stellarator News -2- June 2017
The bolometer diagnostic at
Wendelstein 7-X and its first
results from the initial
campaign
Introduction
In view of the intrinsic three-dimensionality of Wendelstein
7-X (W7-X), a multicamera bolometer system has
been planned to diagnose the radiation loss distribution.
The whole system consists of three subsystems — a bulk
bolometer system, a divertor bolometer system, and a supplementary
system dedicated to 3D effects. According to
their respective functions, they are positioned at appropriate
poloidal and toroidal locations between a triangular
and a bean-shaped plane, covering almost one complete
half of a field period [1]. The three subsystems will be
installed sequentially on the machine, following a
sequence according to the expected plasma scenarios and
the experimental programs.
For the first ten-week experimental campaign started on
10 December 2015 [2], two bolometer cameras had been
installed on a triangular plane to measure the bulk plasma
radiation. The plasmas were operated with five inboard
carbon limiters, and heated by electron cyclotron resonant
heating (ECRH) with a total maximum heating power of
about 4 MW. A maximum central Te(0) = 8 keV and Ti(0)
= 2 keV have been achieved and the line-integrated density
ne was typically less than 3  1019 m. During this
campaign, the design functions of all bolometer components
were checked and verified step by step. The sensitivity
of the resistive bolometer detectors, which were
originally optimized for higher radiation intensities of
high-performance plasmas, had to be re-optimized for the
low-density and low-heating-power limiter plasmas.
Diagnostic arrangements
The bolometer system in the first campaign consists of two
cameras with carefully constructed arrays of Au-foil resistive
detectors [3,4] — a horizontal bolometer camera
(HBC) with 32 channels and a vertical bolometer camera
(VBC) with two detector arrays of 24 channels each. The
detectors have a high responsivity to plasma radiation in
the VUV and soft-X range (300 nm–0.2 nm). They were
calibrated via an in situ Ohmic heating method [3] and
provide line-integrated signals. Figure 1 shows the positions
of the two cameras on W7-X and their lines of sight
(LoS). The LoS of the HBC span the whole cross-section
of the vacuum vessel, thus guaranteeing complete coverage
of the plasma cross section including the limiter
scrape-off layer. The horizontal elongation of the plasma
cross section at the triangular plane requires a larger viewing
angle for the VBC. This is realized by integrating two
separate detector arrays in this camera, with each having
its own aperture and orientation. Due to the limited port
space, the LoS have to be designed in a fan-shape and all
detectors (4-channel bolometer heads) have the same distance
to the aperture center in order to reduce the optic
incident angles and hence to achieve a maximum power
flux reception.
Fig.1. Setup of two bolometer cameras on W7-X viewing the plasma at the triangular cross section. The horizontal bolometer
(HBC) has 32 channels and the vertical one (VBC) consists of two sub-cameras with 24 channels each.
Stellarator News -3- June 2017
Because the metal-foil resistive bolometer also absorbs
microwave radiation, despite a low absorption coefficient
(0.1–1.0% depending on the roughness and the oxidation
of the surface), sufficient screening of the cameras from
microwave stray radiation is one of the most critical issues
in the bolometer design. Two technical solutions have
been found. First, a metal mesh (made of stainless steel)
with an opening smaller than half of the microwave wave
length is mounted in front of the detector array for microwave
screening. With a wire diameter of  = 90 m and an
opening of w = 0.24 mm, the metal mesh has a sufficiently
small microwave transmission factor of ~5%, while keeping
an acceptably high optic transmission factor of about
53%. Second, the inner surface of the detector housing is
coated by a microwave absorbing ceramic layer (TiO/
Al2O3) to reduce multi-path reflectance of the microwaves
coming into the camera housing. The efficiencies of these
techniques have been tested using a prototype of the W7-X
bolometer in a strong microwave background provided by
the MISTRAL chamber [6, 7]. The experimental results
are shown in Fig. 2. The ceramic coating led to a reduction
of the microwave impact by a factor of 10, which is further
suppressed by a factor of about 30 by the mesh screen. For
the W7-X bolometer in use, additional Cu-Be springs are
Fig. 2. Left: Test results of a bolometer prototype in a strong microwave background provided by MISTRAL in cases with
and without microwave suppression measures. A microwave absorber coating leads to a reduction of the microwave impact
by a factor of 10. A further reduction by a factor of 30 is achieved by using a metal mesh in front of the detectors. Right: The
location of the metal mesh screen and the ceramic coating in the design of the vertical bolometer camera.
Fig. 3. The structure of the camera head of the vertical bolometer. Two detector arrays are separated by an optic separator,
with a metal mesh mounted in front of the detector array for microwave screening. A Cu-Be spring welded on the aperture
holder is used as microwave sealing between the detector holder and the housing; the optic separator and the
detector enclosures (not shown) are coated with a ceramic TiO/Al2O3 layer for absorbing stray microwave radiation.
Stellarator News -4- June 2017
welded on the aperture holder to avoid microwave leakage
between the detector holder and the housing, as shown in
Fig. 3. For the purely ECR-heated plasmas during the
whole first experiment phase, no microwave perturbations
could be detected by the bolometer.
A UHV-compatible pneumatic-driven shutter has been
designed and implemented in the bolometer system.
Remote control of the shutter, which is necessary because
of restricted accessibility to the diagnostics during experiments.
was successfully demonstrated. In the first experimental
campaign, this shutter was used to protect the
detectors against contamination during wall conditioning
procedures. For the later long-pulse discharges in W7-X,
closure of the shutter allows researchers to perform in situ
calibration and to turn off measurement between experiment
segments.
In order to limit the smear-out effect originating from the
intersections of the LoS with the 3D flux surfaces in W7-
X, the toroidal extension of the lines of sight are restricted
in such a way that the displacement of the magnetic axis is
smaller than the spatial resolution in the poloidal cross
section. Thus, the toroidal dimension of the slit aperture is
associated with the required spatial resolution and signal/
noise ratio. The expected signal/noise ratio on average for
most plasma conditions of interest, e.g., for 10 MW ECRH
power, is larger than 500. The electronic system has a
noise level that corresponds to a signal induced by an
absorbed power of ~ 200 nW.
The W7-X bolometer system was designed with features
meeting the W7-X steady-state operation requirements
[5]. Some technical aspects relevant to long pulses (ten
minutes or longer) and high heating powers, e.g., thermal
drift suppression/correction, could not yet be assessed in
the first campaign.
First experimental results
Time traces of Prad and plasma parameters
During the first experimental campaign, discharges were
usually terminated in one of two ways: (1) turn-off of heating
constrained by the total input energy limit of 4 MJ due
to still incomplete shielding and hardening of in-vessel
components in this first campaign, and (2) radiative collapse
(RC) due to enhanced impurity radiation. It has been
observed by spectroscopic diagnostic (HEXOS) [7] and
pulse height analysis (PHA) [8] that the main intrinsic
impurity species in the plasma are carbon and oxygen, followed
by traces of chlorine, sulfur, fluorine, copper, and
occasionally iron.
Usually, after 20–40 min glow discharge conditioning, stationary
plasma density could be maintained and long discharge
duration became possible. Figure 4 shows the time
traces of a 6-s-long H2 discharge (#20160310.007). The
discharge was started with 1 MW of ECRH power [9],
which is later reduced to 0.6 MW in order to limit the total
injected energy to below 4 MJ. The plasma was maintained
in a stable state and actively terminated after the
4 MJ input energy limit was reached. The central electron
temperature Te(0) measured by the Electron Cyclotron
Emission (ECE) diagnostic [10] reached 5 keV at a lineintegrated
density, measured by a dispersion interferometer
[11], of around 1  1019 m2. The diamagnetic energy
measured by a diamagnetic loop [12] was ~60 kJ. The
total radiated power Prad, estimated by linearly volumescaling
the bolometer signals under the assumption of
toroidal symmetric radiation, is around 150 kW, i.e., about
25% of the 600 kW heating power. Line radiations from
C IV and O VI measured by HEXOS are shown in the
lowest panel of Fig. 4. It is noteworthy that, in a later
phase after turn-off of the ECRH, the O VI emission rises
sharply, while the C IV emission decreases gradually. The
bolometer measurements correlate more closely to the
time traces of the carbon impurities in this decay phase,
indicating that carbon could be the major radiator in this
discharge. After wall conditioning, oxygen concentration
Fig. 4. Time traces for a long (6 s) hydrogen discharge.
Shown are (from top to bottom) the ECRH power, the total
radiative power loss, the line-integrated plasma density,
the diamagnetic energy, the central electron temperature,
and spectroscopy signals from C IV and O VI.
Stellarator News -5- June 2017
should be low, thus making less contribution to the total
plasma radiation.
Figure 5 shows another typical H2 discharge
(#20160310.011), which was terminated by RC. The discharge
was started with 2 MW of heating power, which
was increased to 3.8 MW with a simultaneous hydrogen
gas puff at a later time point (t = 0.3 s). The plasma density
rose continuously before the second gas puff (t = 0.3 s)
occurred, probably because of outgassing from limiters or
wall elements [13]. The energy confinement (diamagnetic
energy) improved continuously until a third gas puff at t =
0.65 s driving the plasma into a thermally unstable state
(see the evolution of Prad). The central Te reached 8 keV
and doesn’t change very much during the discharge, while
the edge ECE channels (= 0.8) detect a clear drop of Te
at the edge after the sharp rise of Prad. The radiation fraction
increases from 22% in the 2 MW heating power case
up to ~37% before the thermal instability (t ~ 0.65 s). The
spectral lines of C IV (31nm) and O V (76nm), and the
bolometer channel signals (ch17 through  = 0 and ch9
through  = 0.75) are shown in the last two panels. The
O V-emission measurement clearly demonstrates the
importance of oxygen in contributing to this radiation
enhancement. The C IV emission level is maintained
despite the edge temperature change at t = 0.3 s, implying
that the C-ion emission zone lies outside this region, i.e.,
closer to the limiters ( ~ 1).
The bolometer measurements were in accordance with the
ECE results: the enhanced radiation, after gas puffing,
originates at the edge where the temperature dropped. The
third H2 puff at t = 0.65 s results in rapid increment of the
plasma density, accompanied with sharp increases of all
the radiation/emission related diagnostic signals collected
in Fig. 5. After this time point, the plasma develops into a
thermally unstable state (RC) in the sense that Prad quickly
reached the same level as the heating power and then
depleted the stored energy from the confinement region.
This is consistent with the diamagnetic loop measurement,
which shows correspondingly that the diamagnetic energy
dropped once Prad approached the heating power, and is
further supported by a limiter IR camera, which shows a
strongly reduced power load on the limiter in the plasma
collapse phase [14].
Fig. 5. Time traces of a typical H2 discharge terminated by
radiation (collapse starting at 0.68 s). From top to bottom:
ECRH power and total radiated power, line-integrated
plasma density, diamagnetic energy, electron temperatures
at the center and edge, the spectroscopic signals from O V
and C IV, and bolometer signals of two channels viewing
central and edge plasma regions. The first dashed vertical
line indicates the time point when the O V emission
increases and the edge Te ( = 0.85) decreases, while the
second indicates the time point when the central Te begins
to drop.
Stellarator News -6- June 2017
Radiation profiles
Assuming constant emissivity on the magnetic flux surface,
a 1D emissivity profile can be deduced from the
bolometer line integrals by means of Abel inversion. The
resulting profiles are shown in Fig. 6, which shows the
profile evolution for discharge #20160310.011 during the
collapse phase from 0.76 s to 0.79 s indicated by the two
vertical dashed lines in Fig. 5, using t = 0.6 s as reference.
At t = 0.6 s before the second gas-puffing, the emission
profile is peaked at  ~ 0.8. During the collapse process,
the peak shifts inward and eventually reaches the plasma
center leading to termination of the discharge.
The LoS of the bolometer channels in both HBC and VBC
allow for a tomographic reconstruction of a 2D radiation
distribution. Using a similar discharge [15]
(#20160309.007 to #20160310.011) as an example, we
show how the radiation distribution shrinks in 2D space
when the plasma begins to collapse. The inversion is based
on the Gaussian Process Tomography (GPT) method [16]
and carried out within the Minerva framework [17]. The
results are shown in Fig. 7, including a time point t = 0.5 s
prior to onset of the thermal instability, and two later time
points during the collapse process. The poloidal asymmetries
in radiation are not yet understood.
The radiative power loss factor
The total radiation under different plasma conditions has
been investigated for hydrogen plasmas. Figure 8 shows
the radiation loss fraction frad = Prad/PECRH as a function
of line-averaged density ne for different heating power
levels.
The data shown in Fig. 8 are taken from the discharges of
the last week of the first experiment campaign. Each point
is averaged over a 100-ms quasi-flattop of Te and lineintegrated
density. Associated with wall conditioning, the
data are scattered in a large range. Even so, they clearly
show two tendencies: frad, for a given heating power,
Fig. 6. 1D emission profiles obtained by Abel-inversion of
the horizontal bolometer line integrals in discharge
#20160310.011 for the flattop phase (t = 0.6 s) and during
radiative collapse (t = 0.76–0.79 s). The profile peak moves
inward when plasma collapses.
Fig. 8. Radiation loss factor in hydrogen plasmas as a function
of line-averaged plasma density at different heating
powers (black circles). The discharges are collected from
the last week of the first experimental campaign of W7-X.
Each point is averaged over a 100-ms quasi-flattop of Te
and ne. Cross colors indicate different heating powers
(blue: 0.5MW, green: 1.1MW, cyan: 2.0MW, red: 3.5MW).
The dashed lines are guide to the eye.
Fig. 7. Tomographic inversion based on bolometer signals for transient phases to radiative collapse in discharge
#20160309.007. The inner white line stands for the last closed flux surface defined by the limiter.
Stellarator News -7- June 2017
increases with line-averaged density ne, as indicated by
the dashed lines, and decreases with increasing heating
power for a fixed ne. For example, for a heating power
around 0.5 MW, frad lies in the range 30–60% (blue) and
decreases to 20–40% for 1.1 MW (green) and further
drops to 15–35% for higher power levels (2 MW in cyan
and 3.5 MW in red). More detailed analysis is under way.
D. Zhang, R. Burhenn, A. Alonso, B. Buttenschön, Y. Feng, L.
Giannone, M. Hirsch, U. Höfel, R. Lauber, M. Marquardt, K. Rahbarnia,
J. Svensson , G. A. Wurden, R. Brakel, O. Grulke, J.
Knauer, R. König, H. Laqua, S. Marsen T. Stange, T. Schröder,
H. Thomsen, G.M. Weir, A. Werner and the W7-X Team
Max-Planck Institut für Plasmaphyisk
17491 Greifswald, Germany
E-Mail: daihong.zhang@ipp.mpg.de
References
[1] D. Zhang, L. Giannone, et al., 34th EPS Conference On
Plasma Phy., Warsaw, Poland (2007).
[2] T. Klinger et al., Plasma Phys. Contr. Fusion 59, 1
(2017).
[3] K. F. Mast et al., Rev. Sci. Instrum. 62, 744 (1991).
[4] L. Giannone, K. Mast, and M. Schubert, Rev. Sci. Instrum.
73, 3205 (2002).
[5] D. Zhang et al., Rev. Sci. Instrum. 81, 10E134 (2010).
[6] S. Ullrich, H. J. Hartfuss, M. Hirsch, H. Laqua, Stellarator
News Issue 98, 2005.
[7] D. Zhang et al., 38th EPS Conf. on Contr. Fusion and
Plasma Phys., Strasbourg, P5.056 (2011).
B. Buttenschön et al., P4.012 in Proc. 43rd EPS Conf.
Plasma Phys. Contr. Fusion, Leuven, Belgium (2016).
[8] H. Thomsen et al., J. Instrum. 10, P10015 (2015).
[9] S. Marsen et al. in Proc. 43th EPS Conference Plasma
Physics, Leuven, Belgium, P4.002 (2016).
[10] M. Hirsch et al., in Proc. 43th EPS Conference Plasma
Physics, Leuven, Belgium, P4.007 (2016).
[11] P. Kornejew et al. to be published; J. Knauer, et al., 43th
EPS Conference On Plasma Physics, Leuven, Belgium,
P4.017.
[12] K. Rahbarnia et al., in Proc. 43th EPS Conference Plasma
Physics, Leuven, Belgium, P4.011 (2016).
[13] T. Wauters et al., in Proc. 40th EPS Conference Plasma
Physics, Leuven, Belgium, (2016).
[14] G. A. Wurden et al., Rev. Sci. Instrum. 2016.
[15] D. Zhang et al., in Proc. 43th EPS Conference Plasma
Physics, Leuven, Belgium, P4.015 (2016).
[16] J. Svensson, Nonparametric Tomography using Gaussian
Processes, JET Internal Report EFDA-JETPR(
11)24, 2011.
[17] J. Svensson, A. Werner, “Large Scale Bayesian Data
Analysis for Nuclear Fusion Experiments,” In Proc.
IEEE Workshop on Intelligent Signal Processing
(WISP) 2007.