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 Fusion Energy Division, Oak Ridge National Laboratory
Building CR 5600 P.O. Box 2008 Oak Ridge, TN 37831-6169, USA
Editor: James A. Rome Issue 139 February 2013
E-Mail: jamesrome@gmail.com Phone (865) 482-5643
On the Web at http://www.ornl.gov/sci/fed/stelnews
Progress of high-temperature
experiments in LHD
Realization of high-temperature plasmas is one of the most
important issues in helical plasmas, which have an advantage
over tokamak plasmas for steady-state operation. In
the Large Helical Device (LHD), the heating capability
has been upgraded year by year and the high-temperature
regime has been successfully extended. In the LHD, three
negative-ion-based neutral beam injectors (NBs) produce
hydrogen neutral beams with a beam energy of 180 keV
and a total port-through power of 16 MW [1]. These negative
NBs are tangentially injected to the LHD plasma. A
positive-ion-based NB [positive neutral beam injection
(NBI)] with a low energy of 40 keV was perpendicularly
injected for ion heating. Since 2010, a second 40-keV perpendicular
NB has been operational in LHD and the total
port-through power for perpendicular NBI has thus
increased from 6 MW to 12 MW. An electron cyclotron
resonance heating (ECRH) system with eight gyrotrons
has been operated for preionization and plasma heating
[2]. Enhancement of the output power per gyrotron has
been planned in LHD and the replacement of the existing
gyrotrons with higher-power tubes is in progress. At present,
three 77-GHz gyrotrons with output power of more
than 1 MW each, are operational for plasma experiments.
LHD now has 28 MW of NB power and 3.7 MW of ECRH
power available for experiments.
Figure 1 shows (a) the radial profiles of ion temperature
Ti, electron temperature Te, and electron density ne in a
typical high-Ti discharge with a carbon pellet, and (b) the
progress of the achieved central ion temperature Ti0 in
LHD as the dependence of Ti0 on the density-normalized
ion heating power Pi/ne, where Pi is the volume integral of
the local NBI deposition power for ions evaluated using
the FIT3D code, and reff in Fig. 1(a) represents the effective
minor radius. Ion temperature of 7 keV at the plasma
center was successfully obtained and the achieved Ti0 has
increased with Pi/ne. High-Ti plasmas have been realized
with injection of a carbon pellet [3–5]. The kinetic
Fig. 1. (a) The radial profiles of Ti, Te, and ne in a typical
high Ti discharge, and (b) the progress of the achieved Ti
in the center of LHD.
In this issue . . .
Progress of high-temperature experiments in
LHD
The operational regime in high-temperature plasmas
has been successfully extended due to the upgraded
heating system and optimization of the discharge scenario
in the Large Helical Device (LHD). Transport
analysis of the high-ion-temperature plasma showed
that the ion thermal diffusivity and the viscosity were
reduced after the formation of the ion internal transport
barrier. .............................................................. 1
Stellarator News -2- February 2013
energy confinement improves by a factor of 1.5 after pellet
injection. In the high-Ti phase, a flat or hollow profile in ne
has been observed. This has different characteristics from
the PEP mode investigated in tokamaks [6–8].
This improvement was attributed to the upgrade of the NB
system, the optimization of the discharge scenario, and ion
cyclotron radio frequency heating (ICRF) wall conditioning.
It was found that intensive wall conditioning by a
series of ICRF minority-ion-heating discharges helped to
increase the ion temperature. In fact, there was no significant
increase of NBI power from 2010 to 2011 in LHD.
Figure 2 shows radial profiles of (a) Ti, (b) ne, and (c) the
ion-heating power normalized by ne before and after ICRF
conditioning. In Fig. 2, a99 is the averaged minor radius
inside which 99% of the stored energy is confined. In the
ICRF discharges, NBIs were not used and helium was
fueled as the working gas for minority heating. Before the
ICRF conditioning, a hollow ne profile was formed and Ti0
was below 6 keV. After 30 discharges of ICRF conditioning
with a pulse duration of 10 s, a lower electron density
with a parabolic profile and Ti0 exceeding 6 keV were
obtained. The partial pressures of both H2 and He during
the discharge were found to be decreased after ICRF wall
conditioning. This represents a decrease of outgassing
from the wall, namely a decrease in neutral particle recycling.
The increase of Ti0 after ICRF conditioning is considered
to be the result of an increase of NBI heating
power per ion at the core region due to the decrease of
edge density, as can be seen from Fig. 2(c), rather than an
improvement in the ion heat transport.
Figure 3 shows (a) radial profiles of Te and ne in a high-Te
discharge and (b) a map of simultaneously attained central
electron temperature Te0 and line-averaged electron density
for ECRH discharges. Highly accurate Te profiles
were obtained by the accumulation of the intensity of
Thomson scattered light at 17 times during fixed-condition
discharges with the three YAG lasers all injected together
[9]. Central electron temperature of 20 keV was successfully
achieved by centrally focused ECRH of 3.35 MW at
an electron density of 0.2  1019 m3. The plasma parameter
regime with regard to the electron temperature has been
successfully expanded in both low- and high-density conditions.
Dynamic transport analysis that takes into account the
slowing-down effect of NBI deposition is suitable when
the plasma parameters change transiently in the discharge
[10], as they do with the injection of a carbon [3–5]. Here
the temporal changes in ion heat transport and toroidal
momentum transport in high-Ti discharges are discussed.
Figure 4 shows the time evolution of (a) the line-averaged
electron density and the port-through NBI power, (b) the
NBI absorption power, (c) the NBI input torque, (d) Ti, (e)
the toroidal flow velocity Vφ, (f) χi/Ti
3/2 at the three posi-
Fig. 2. Radial profiles of (a) Ti, (b) ne, and (c) Pi/ne before
and after ICRF wall conditioning.
Fig. 3. (a) Radial profiles of Te and ne and (b) map of simultaneously
attained Te0 and ne for ECRH discharges.
Stellarator News -3- February 2013
tions, (g) the effective viscosity μφ consisting of the tangential
NB-driven rotation and the intrinsic rotation at reff/
a99 = 0.31, and (h) the dependence of the Prandtl number
Pr on Ti at reff/a99 = 0.31 in a typical high-Ti plasma. The
thermal diffusivity is normalized by Ti
3/2 to cancel the
gyro-Bohm dependence Here the Prandtl number is
defined as μφ/χi. The integrated absorption power and the
force in Figs. 4(b) and 4(c) were calculated taking account
of the slowing down of the energy and the velocity of NB
particles in the plasma [10]. The plasma was sustained by
three tangentially injected NBs and two perpendicularly
injected NBs with total-port-through power of 27 MW,
and a cylindrical carbon pellet (φ = 1.0 mm, l = 1.0 mm)
was injected at t = 4.57 s. In Ref. [3], the increment of Zeff
was shown to be ~1 just after carbon pellet injection,
decreasing to ~0.2 due to the formation of an impurity
hole. One line of the perpendicular NBs was modulated
for Ti measurement by charge exchange recombinant spectroscopy
(CXRS). After the pellet injection, Ti, Vφ, dTi/
dreff and dVφ/dreff, clearly increased in the core region,
indicating the formation of the ion internal transport barrier
(ITB). In the core region, χi was found to change more
slowly than in the peripheral region. The toroidal momentum
transport also improved with the reduction in thermal
diffusivity, and the Prandtl number (ignoring the intrinsic
torque) was close to unity despite the extensive change in
confinement during the discharge. Note that μφ and Pr
were underestimated in this analysis by a factor of 2 at
most, because the contribution of intrinsic rotation is in the
co direction and peaks at reff/a99 ~0.6 with magnitude
comparable to that driven by NBI [4]. However, the heat
and the momentum confinement degraded in the latter
phase of the discharge.
Figure 5 shows the flux-gradient relation between
(a) Qi/ne and dTi/dreff and (b) Pφ/nemi and dVφ/dreff at
reff/a99 = 0.31, where Qi is the ion heat flux and Pφ is the
toroidal momentum flux [5]. The slopes in the relation
between Qi/ne and dTi/dreff, and between Pφ/nemi, and
dVφ/dreff correspond to χi and μφ, respectively. As shown
in Fig. 5(a), dTi/dreff increased after carbon pellet injection
despite the small change of Qi/ne. This indicates that the
ion heat transport was improved, leading to the achievement
of 7 keV. However, the confinement improvement
was temporary and dTi/dreff gradually decreased after
dTi/dreff reached 13 keV/m. The toroidal momentum
transport was also improved by ion ITB formation. The
increase of dVφ/dreff with constant-momentum flux means
the decrease of μφ and implies an increase in intrinsic rotation.
However, the momentum transport went back to the
low-confinement branch in the latter phase of the dis-
Fig. 4. The time evolution of (a) ne, (b) the NBI absorption power, (c) the total force, (d) Ti, (e) Vφ, (f) χi/Ti
3/2, (g) μφ, and (h)
the dependence of Pr on Ti.
Stellarator News -4- February 2013
charge, similar to the ion heat transport. On the other hand,
a change in electron heat confinement was not observed in
the discharge [10]. The intrinsic rotation [5, 11] due to the
off-diagonal terms in the transport matrix should be investigated
for more qualitative and quantitative evaluation of
heat and momentum transport. Clear dependence of the
intrinsic torque on the ion temperature gradient was shown
in Ref. [5]. Detailed results of the heat and momentum
transport analyses, including off-diagonal terms, will be
reported in the near future.
This work was supported by NIFS grants, 11ULRR701,
11ULRR702 and ULRR703.
References
[1] Y. Takeiri et al., Fusion Sci. Technol. 58 (2010) 482.
[2] T. Shimozuma. et al., Fusion Sci. Technol. 58, (2010)
530.
[3] K. Ida et al., Nucl. Fusion, 49 (2009) 095024.
[4] K. Ida et al., Nucl. Fusion, 50 (2010) 064007.
[5] K.Nagaoka et al., Nucl. Fusion, 51 (2011) 083022.
[6] S. Sengoku et al., Nucl. Fusion, 25 (1985) 1475.
[7] B. J. D. Tubbing et al., Nucl. Fusion, 31 (1991) 839.
[8] L. R. Baylor et al., Nucl. Fusion, 37 (1997) 127.
[9] I. Yamada et al., Rev. Sci. Instrum 81 (2010) 10D522.
[10] H. Lee et al., Plasma Phys. Control. Fusion (in press).
[11] M. Yoshinuma et al., Nucl. Fusion, 49 (2009) 075036.
Hiromi Takahashi
Plasma Heating Physics Research Division
National Institute for Fusion Science
322-6 Oroshi-cho, Toki, Gifu, 509-5292, JAPAN
E-mail: takahashi.hiromi@LHD.nifs.ac.jp
Fig. 5. The dependence of (a) Qi/ne on dTi/dreff and (b) Pφ/
nemi on dVφ/dreff.