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.
Oak Ridge National Laboratory is managed by UT-Battelle, LLC, for the U.S. Department of Energy.
Published by Oak Ridge National Laboratory
Building CR 5600 P.O. Box 2008 Oak Ridge, TN 37831-6169, USA
Editor: James A. Rome Issue 144 August 2014
E-Mail: jamesrome@gmail.com Phone (865) 482-5643
On the Web at http://www.ornl.gov/sci/fed/stelnews
Direct generation of NBI
plasmas in TJ-II with lithiumcoated
walls
In contrast to the fully inductive scheme used in tokamaks,
plasma start-0up in stellarators is commonly achieved by
launching electron cyclotron resonant (ECR) or ion cyclotron
wave (ICW), thus creating a target plasma for highdensity
plasmas heated by neutral beam injection (NBI).
In large devices, however, direct start-up by unaided injection
of the heating beams has been proven to be feasible,
as reported in the Large Helical Device (LHD) and, only
for well-conditioned walls, in Wendelstein 7-AS (W7-AS)
[1, 2]. Microwave-assisted start-up, at 2.45 GHz, was also
recently reported in Heliotron-J [3], and ICW heating was
successfully applied in CHS [4]. In TJ-II, ECR injection at
53.2 GHz at the second harmonic is routinely used to produce
target plasmas for NBI coupling. The effect of initial
neutral pressure on the resulting delay of the gas breakdown
with respect to the nominal time of ECR injection
was addressed in a previous paper [5], with typical values
of 2–10 ms. A kinetic model, including atomic and molecular
processes for the injected H2, was successfully
applied to the results. Furthermore, the effect of wave
polarization on plasma start-up was also addressed [6].
The TJ-II stellarator has been operating at CIEMAT
(Madrid) since 1997, and different heating schemes and
wall condition scenarios have been applied for the generation
of hot plasmas [7]. Since 2007, a new kind of first
wall coating, produced by boron deposition followed by
lithiation, has been applied. This wall conditioning technique
was a key to achieving full-power NBI plasmas
under good density control conditions, not possible with
boron coatings alone [8], as well as transition to H mode
confinement. Since 2012, a liquid lithium limiter (LLL)
based on the Capillary Porous System (CPS) design
replaced one of the former carbon limiters, and very
recently the second C limiter has been replaced by another
LLL. Two NBI systems (Eb = 35 kV), delivering up to
600 kW each, are coupled to the machine in opposite
directions to compensate for induced currents in the
plasma. In the present experiments only one of the NBI
sources was used (co-direction). Due to the close coupling
to the vacuum vessel, neutral gas from the neutralization
box leaks into the plasma as the NBI is launched. In addition,
the interaction of the beam with the wall opposite to
it releases extra gas, depending on beam power and wall
conditions.
The magnetic configuration of TJ-II plasmas involves four
main magnetic field components, two of them (CC and
HX) produced by the central coil system [9]. For the magnetic
configuration used here, with BT = 0.95 T at the
magnetic axis, the steady-state values of the CC and HX
currents are10 kA and 4.8 kA, respectively. These currents
are produced in a 250 ms ramp-up time. The associated
toroidal electric fields induced are 2.8 and 1 V/turn,
respectively, corresponding to a total induced toroidal
field of 0.35 V/m, at the lower limit of those used for
ohmic break-down in tokamaks [10]. However, nested
magnetic surfaces are produced in TJ-II at times well
before (~200 ms) the plateau is reached.
Three gases were tried for the formation of the target
plasma: H2, He, and Ar. They were puffed during the
ramp-up phase, at different delays with respect to the NBI
In this issue . . .
Direct generation of NBI plasmas in TJ-II
with lithium-coated walls
NBI-heated plasmas have been obtained in TJ-II without
the help of any external power supply, such as
microwaves or RF sources. Delays as short as 6 ms
between the NBI launching time and plasma start-up
were achieved, with the required neutral gas directly
provided by the NBI source. Beam injection at the very
end of the field ramp-up and fully conditioned walls (B
+ Li coatings) were identified as prerequisites for very
consistently reproducible results. ........................... 1
Stellarator News -2- February 2012
time, yielding typical pressures at the vacuum chamber of
~ 105–104 mbar. NBI power scans were also tried, from
400 to 600 kW. Several H monitors, with submillisecond
time resolution and located at different toroidal positions,
were used to characterize the delay in start-up, together
with the other plasma diagnostics routinely used in TJ-II
operation.
First attempts to directly generate NBI plasmas were made
with an unconditioned first wall. Strong production of
hard X rays was seen, especially when gas was injected
some hundred milliseconds before the plateau phase at
1000 ms. Although chord-integrated electron densities on
the order of 1012–1013 cm3 were recorded by the interferometer,
no sustained plasma heating by NBI alone was
found. Figure 1 shows one of these cases. The results were
the same when H2 was replaced by He as the pre-filling
gas. Under these wall conditions, it was found that any
extra gas injected during the early phase of plasma breakdown
led to the production of strong high-energy X-ray
(HXR) emission without any improvement in the coupling
of the NBI energy to the target plasma. Thus, direct NBI
was used alone as both particle and energy source. A
sweep in NBI injection times during the ramp-up phase of
the CC and HX coil currents was performed, spanning the
last 200 ms of the current ramp-up. The results indicated
optimized conditions (low HXR generation, short delay
time to plasma breakdown) for NBI launch times between
960 and 980 ms.
This situation changed dramatically after full conditioning
of the TJ-II walls by boronization followed by lithiation.
Base pressures in the range of 1–2 107 mbar were readily
achieved after the conditioning, mostly due to the gettering
of residual water and oxygen by the deposited films.
Figure 2 shows one example of plasma start-up by NBI
with well-conditioned walls. The delay between the onset
of the H signal located on top of the NBI-wall interaction
area (beam dump) and the rise of the electron density is
negligible. Also, typical traces indicative of hot plasma
production were seen at times ~20 ms after the NBI time.
Central electron temperatures Te ~ 320 eV at chord densities
of ne ~ 4 1013 cm3 are routinely obtained for NBI
nominal powers of PNBI ~ 550 kW. Thomson scattering
profiles of the NBI-generated plasmas are basically identical
to those obtained earlier in target plasmas with ECR
heating (ECRH).
To date, this new scheme for plasma start-up has allowed
for 2-beam NBI operation at total nominal injected powers
up to 800 kW, chord-integrated densities up to 8 1013
cm3, and diamagnetic energy contents up to 4.5 kJ. Scans
in BT have led to successful plasma production up to BT
< 0.6 T, thus opening the possibility of physics studies
under conditions free from the ECRH resonant field constraint.
Finally, the toroidal electric field when the NBI gas
is actually injected (onset of H, top signal in Fig. 2) is at
least a factor of 5 lower than that produced earlier in the
ramp-up phase. Very recently, the use of the ohmic heating
circuit for inducing toroidal electric fields of magnitude
comparable to those actually required (less than 101 V/m)
has proven successful for plasma breakdown at times
within the nominal plateau time (t >1000 ms), possibly
helped by the remaining hot electrons accelerated during
the last stages of the current ramp-up.
F. L. Tabarés, E. Ascasibar, E. Blanco, F. Medina, I. Pastor,
D. Tafalla, and the TJ-II Team
Laboratorio Nacional de Fusion
Association Euratom/CIEMAT
Madrid
Fig. 1. Start-up with unconditioned walls.
Fig. 2. Start-up with fully conditioned walls.
Stellarator News -3- February 2012
References
[1] O. Kaneko. et al., Nucl. Fusion 39, 1087–91 (1999).
[2] W. Ott, E. Speth, and W7-AS Team, Nucl. Fusion 42,
796–804 (2002).
[3] S. Kobayashi, et al., Nucl. Fusion 51, 062002 (2011).
[4] O. Kaneko, et al., Proc. 13th Int. Conf. Plasma Physics
and Controlled Nuclear Fusion Research, Washington,
DC, 1990 (Vienna: IAEA) 2, 473 (1991).
[5] A. Cappa, et al., Nucl. Fusion 41, 363 (2001).
[6] A. Cappa, et al., submitted to Nucl. Fusion (2014).
[7] D. Tafalla, et. al., Fusion Eng. Design 85, 9.
[8] F. L. Tabarés, et al., Plasma Phys. Control. Fusion 50,
124051 (2008).
[9] C. Alejaldre, et al., Fusion Technol. 17, 131 (1990).
[10] R. Papoular, Nucl. Fusion 16, 37 (1976).