USE OF IMPURITY PELLETS TO CONTROL ENERGY
DISSIPATION DURING DISRUPTION
G. PAUTASSO, K. BUCHL, J.C. FUCHS, 0. GRUBER, A. HERRMANN,
K. LACKNER, P.T. LANG, K.F. MAST, M. ULRICH, H. ZOHM,
ASDEX UPGRADE TEAM
Max-Planck-Institut fur Plasmaphysik,
Garching, Germany
ABSTRACT. Injection of impurity pellets has been shown to be a successful method of reducing
thermal and mechanical loads during disruptions. The evolution of the quenching plasma after pellet
injection in ASDEX Upgrade is described and the requirements of such a method for mitigating
disruptions in future devices are discussed.
1. INTRODUCTION
Plasma disruptions constitute a major operational
and technological obstacle to developing a tokamak
fusion reactor. A number of plasma scenarios can lead
to an MHD-unstable current profile and sudden dete-
rioration of the plasma energy confinement, i.e. to the
so-called thermal quench [l]. The present design spec-
ifications for ITER [2] prescribe that, during thermal
quench, power densities of the order of lo5 MW/m2
can be deposited on the divertor, which causes consid-
erable erosion of the target plates. In addition, halo
currents would exert forces of the order of lo4 tonnes
[l] on the vessel and its components, which makes the
design of these structures a critical issue.
Present efforts in disruption studies are aimed a t
recognizing disruption precursors, developing proce-
dures for suppressing unstable MHD modes, iden-
tifying the largest thermal and mechanical loads
expected and setting up strategies for mitigating
these loads on the machine during disruptions.
Injection of impurity pellets into a plasma that is
irreversibly going to disrupt is one of the mitigation
strategies and is a possible method of reducing the
high power load on divertor plates during thermal
quench. The idea of radiating most of the thermal
energy of the plasma before it is naturally conducted
from the plasma core through the SOL onto a lim-
ited divertor plate area is gaining increasing attention
because of its simplicity and its preliminary success
in JT-6OU [3]. Since then, similar experiments have
been conducted on several other tokamaks and have
produced some positive results [4-71.
This paper reports recent and well documented
impurity injection experiments in ASDEX Upgrade
that show a significant reduction of the thermal load
on the divertor plates and a reduction of the mechan-
ical loads on the vessel components in disruptions of
elongated, single-null plasmas. There is also a discus-
sion of how future experiments should be conducted
in order to make them significant for an ITER-like
scenario.
2. EXPERIMENTAL SET-UP
The experiment consisted in injecting neon pellets
with a velocity of 560 m/s into ohmically heated and
neutral beam heated (NBH) plasmas. Pellets with
nominal dimensions of 1.65 x 1.65 x 1.75 mm3 and
containing about 1.7 x lozo neon atoms were used.
Large pellets and medium velocities were chosen here,
on the basis of preliminary calculations, to pene-
trate up to the plasma centre and rapidly cool down
the plasma. Pellet penetration depths were estimated
from the temporal evolution of visible radiation and
SXR data. The neon pellets penetrated close to and
beyond the plasma axis in, respectively, the NBH dis-
charge and ohmic discharges, which are described in
the following.
All these pellets cooled down the plasma in less
than 1 ms and triggered the current quench. No
attempt was made to trigger the injection of the pellet
on the basis of some disruption precursor. Neverthe-
less, the time delay with which the pellet enters the
plasma after the pellet injector has been triggered is
of the order of 10 ms. This makes our present injec-
tor already adequate for use in conjunction with some
pre-disruption characteristic.
During the experiments we made measurements
of, inter alia, the radiated power (bolometer, with
1 ms time resolution), the power deposited on the
divertor plates (thermography, 0.25 or 1.3 ms), the
halo currents (shunts, 0.2 ms) and the forces on the
passively stabilizing loop (PSL) and vessel supports
(strength gauges, 1 ms).
NUCLEAR FUSION, Vol. 36, No. 10 (1996) 1291

PAUTASSO et al.
(a>
No. 5864
Plasma Current
I- 1 Thermal Energy
Power div.in
~
1.97 1.974 1.978 1.982 1.986 1.99
Time (s)
0.6 MA
0
50 k~
0
20 m
0
20 M w
0
40 MW
0
200 v
0
CO)
No. 5698
Plasma Current
20 MW
Power div.out
t
I * . I 0 . 1 . . 1 . I 1 . 8 I . . . I . * . I . I . I I I . . I
1.832 1.836 1.84 1.844 1.848
Time (s)
FIG. 1. T i m e histories of the plasma current, plasma thermal energy, power deposition o n inner and outer divertor plates,
radiated power and loop voltage for: (a) a DL disruption in a n ohmic plasma (shot 5864) and (b) a disruption in a n ohmic
plasma following injection of a neon pellet at t = 1.8356 s (shot 5698).
3. ENERGY BALANCE DURING
DISRUPTIONS
In disruptions without injection of impurity pel-
lets we observe that up to 100% of the plasma ther-
mal energy, E t h , can be deposited on the divertor
plates during the thermal quench within 1 to 2 ms;
the energy deposition on the divertor, over the whole
disruption, can amount to 30% of the total (thermal
and magnetic) plasma energy [8]. A reduction of the
energy deposited on the divertor plates is observed
in shots where impurities were already puffed in dur-
ing the discharge to reduce the energy deposited. In
these cases, it is observed that the power deposited is
reduced to 50% of the thermal energy during thermal
quench, and down to 10% of the total plasma energy
during the whole disruption.
Figure 1 shows time histories of the plasma ther-
mal energy, of the power load on the inner and outer
divertor plates and of the radiated power for an ohmic
density limit (DL) disruption without neon (shot
5864, Fig. l (a)) and a disruption triggered by injec-
tion of a neon pellet (shot 5698, Fig. l(b)). In shot
5864 the attainment of the density limit leads to a
few minor disruptions and to the final thermal quench
at t = 1.974 s. While the plasma looses its thermal
energy, an equivalent amount of energy is deposited
on the divertor plates within 3 ms from the onset of
the thermal quench. Table I summarizes the energy
balance during the disruption for shot 5864 and for
the other shots described below. The electromagnetic
energy deposited in the conductors as induced cur-
rents during the current quench phase is of the order
of magnitude of the error affecting the measurements
of the radiated energy, of the energy flux to the plates
and of the plasma thermal energy, and is not calcu-
lated here.
After injection of the neon pellet at t = 1.8356 s
into an ohmic plasma with a typical thermal energy of
55 kJ (shot 5698, Table I) we observe the suppression
1292 NUCLEAR FUSION, Vol. 36 , No. 10 (1996)