Verlag des Forschungszentrums Jülich
JUEL-2980
Finken, Karl Heinz
Interaction of pellets with hot plasmas
66 S., 1995
Introduction
In thermonuclear fusion research, pellet plasma interaction occurs in two important
areas: One is the so-called inertial confinement and the second one is the injection
of pellets in a magnetically confined plasma. In inertial confinement, a spherical
target of a few millimeters diameter filled with a deuterium-tritium mixture - the
pellet - is bombarded with a well focussed high power beam of either photons
(laser-beam, indirectly by laser beam initiated incoherent X-ray radiation) or
particles (electron-, light ion- or heavy ion-beam). The beam energy has to be in the
order of Mega-Joule for a time span of several nanoseconds. The beam heats the
outer surface of the pellet and forms an expanding plasma cloud. The expansion
establishes a reaction-force directed towards the pellet center and initiates a
compression wave into the pellet. To reach a positive energy balance, i.e. to
extract more fusion power from the pellet than has been put in by the beam, it is
necessary to compress the pellet center to a density of over one thousand times
the solid state density. Under this condition the inertia of the pellet matter keeps
the material together long enough for obtaining the desired burning rate. Inertial
confinement attracts much attention from the military side and, therefore, some
aspects are still classified. In the following, inertial confinement will not be
discussed.
The primary goal of the pellet injection into magnetically confined plasmas
is the fuelling of the discharge. Here, the pellets are hydrogen or deuterium ice
pieces with a volume of typically a few cubic-millimeters; the number of atoms of
such pellets is a substantial fraction of the total number of particles in the
discharge. Fuelling by peliet injection is complementary to the conventlonallvused
gas injection fuelling, The fuelling allows the stationary regulation of a discharge
density, if the device has a balanced particle sink. !n short pulse machines the sink
is either simply the wall, a pumped limiter or a pumped divertor. In long pulse
discharges like the planned ITER device, wall pumping eventually saturates, so that
active exhaust systems are necessary. Gas injection has the disadvantage, that the
ionization processes take place close to the plasma edge. This leads to a strong
reduction of the probability for the fuelling gas to penetrate into the central plasma.
The gas remains at the plasma boundary; this is equivalent to a iow fue!!ing
efficiency and a high neutral particle pressure at the plasma boundary. The low
fuelling efficiency in a fusion reactor means that an undesired high amount of
tritium gas has to be introduced into the machine; the tritium recycles in the
plasma boundary of the discharge and is extracted together with the helium ash.
The high recycling is connected with a high gas pressure at the plasma edge and
this high neutral density deteriorates the plasma confinement, This negative effect
can be avoided by injecting a pellet deeply into a fusion plasma. The influence of
the pellet injection on the plasma performance wi!! be discussed in chapter III. In
chapter II the effects directly associated with the penetration into the discharge are
treated.
The injection of pellets into a plasma leads to an enormous heat transfer
from the plasma to the pellet. For "normal" heat fluxes, the heat deposition to a
solid surface and the heat propagation into the solid is described by the well-known
heat equation. If, however, the incoming heat flux is so high that the surface
material is evaporated before the heat sufficiently penetrates into the bulk, a new
effect, the so-called 'gas shielding', is observed. In everyday life this effect is
known as the Leidenfrost phenomenon 11.1, L21 and it occurs if e.q. a droplet of
water falls on a hot plate. The water contracts to a ball so that the contact surface
to the hot plate is minimum. The calefaction leads to the formation of a vapor
cloud at the contact point, which lifts the droplet from the hot plate and inhibits
the heat transition by orders of magnitude.
If the shielding effect would be neglected and it would be assumed that the
water (1 mm thick) remains in contact to e.g. a 500 0 C plate, then the droplet
would evaporate within less than a second. The actual evaporation time of the
droplet is extended because during extreme heating the evaporation from the
surface leads to an isolating protection gas film between the plate and the droplet.
Some properties of the leidenfrost phenomenon are very specific and cannot be
generalized like e.g. the flow pattern of the gas layer supporting the droplet from
the hot plate. Other features like the protection cloud, however, are so general that
they are also of importance for the pellet ablation process in the plasma.
Neuerscheinungen
Schriften des Forschungszentrums Jülich
Ihre Ansprechperson
Heike Lexis
+49 2461 61-5367
zb-publikation@fz-juelich.de