PFC/JA-84-3
Energy Confinement of High Density
Pellet-Fueled Plasmas in Alcator C
M. Greenwald, D. Gwinn, J. Parker, R. Parker, S.
Wolfe, M. Besen, F. Camacho, S. Fairfax, C. Fiore, M.
Foord, R. Gandy, C. Gomez, R. Granetz, B. LaBombard,
B. Lipschultz, B. Lloyd, E. Marmar, S. McCool, D.
Pappas, R. Petrasso, P. Pribyl, J. Rice, Y. Takase,
J. Terry, R. Watterson
Plasma Fusion Center, Mass. Institute of Technology
Cambridge, MA
02139
S. Milora, D. Schuresko
Oak Ridge National Laboratory, Oak Ridge, Tenn.
March 1984
This work was supported by the U.S. Department of Energy Contract
No. DE-AC02-78ET51013. Reproduction, translation, publication, use
and disposal, in whole or in part by or for the United States government is permitted.
By acceptance of this article, the publisher and/or recipient acknowledges the U.S. Government's right to retain a non-exclusive,
royalty-free license in and to any copyright covering this paper.
- 1 -
ENERGY CONFINEMENT
OF
HIGH
DENSITY
PELLET-FUELED
PLASMAS
IN ALCATOR C
M.
Greenwald,
S.
Wolfe,
Foord,
R.
Gwinn,
B.
S.
F.
Besen,
Gandy,
Lipschultz,
Petrasso,
Terry, R.
D.
M.
C.
Milora(i),
Gomez,
Lloyd,
P. Pribyl,
Watterson'
S.
Camacho,
E.
J.
R.
D.
Parker,
Granetz,
Marmar,
Rice,
J.
Fairfax,
S.
B.
McCool,
Parker,
M.
LaBombard,
B.
D.
Schuresko(i),
R.
Fiore,
C.
Y.
Pappas,
R.
Takase,
J.
Plasma Fusion Center
Massachusetts Institute of Technology
Cambridge, MA 02139
Abstract
A series of pellet-fueling
out on
the Alcator
penetrate to
C tokamak.
within a
experiments
has been carried
High
hydrogen
few centimeters
speed
pellets
of the magnetic
axis,
raise the plasma density and produce peaked density profiles.
Energy confinement
is
observed
to
discharges fueled only by gas puffing.
values of electron density,
increase
over
similar
In this manner record
plasma pressure,
and Lawson number
(nT) have been achieved.
(i)
Address:
37830
Oak Ridge National Laboratory, Oak Ridge, Tenn.
-
puf-
is by gas
tokamaks
fueling
of
The standard method
2 -
fing, which supplies particles to the plasma edge in the form
of neutral
densities,
Fueling of
not
may
which
density profiles
broad
with
optimal
to
respect
the mechanism which carries part-
Secondly,
energy transport.
be
for the
relatively
produces
fueling
Edge
following reasons.
problematic
then becomes
plasmas
density
particle
the
source is concentrated at the discharge boundary.
large high
into
far
penetrate
cannot
line-integral
large
plasmas with
neutrals
Since
atoms.
icles up the density gradient to the center of the plasma is
it
thus
not understood,
carries
particles
in Alcator C,
is
energy confinement
plasma
the
into
responsible for anomalous energy loss.
may be occurring
into
Finally, there is the possibility that the
reactor regimes.
mechanism which
extrapolate
to
difficult
is
is
core
Some of these problems
where at high plasma densities
considerably
worse than what
be
would
expected from the TE - ne scaling observed at lower densities
1)1.
(Fig.
Injection of high speed frozen hydrogen pellets has been
proposed as
and in
an alternate
recent years
method
pellet
for
fueling
injectors
fusion devices
capable
of
fueling
3
current generation of tokamaks have been developed .
ments on pellet
and the
effects
being studied
on
describe effects
penetration
of
several
on
devices 4 ,5,6,
of pellet
,
the
Experi-
into plasmas have been performed
fueling
pellet
2
fueling
on
plasma
7
.
properties
In this letter
energy
confinement
are
we
in
-
3 -
Alcator C.
The pneumatic injector used in the experiments described
here was
designed
built at
MIT.
at
It
the
Oak
fires
pellets with velocities
appropriate changes
Ridge
four
National
Laboratory
independently-timed
between
8
in operating
and
9 x 104
procedures
it
tubes
prevents the
and
helium
radius).
80 and
tral
3
for
propellant
Each pellet contains 6
2 x 1014 cm-
provision
x
hydrogen
cm/sec.
With
can also
fire
deuterium pellets at slightly reduced velocities.
with guide
A beamline
differential
from
and
reaching
pumping
the
plasma.
1019 particles corresponding to <ne> =
in Alcator C (16.5 cm minor radius and 64 cm major
Standard operation is with a toroidal field between
120
kG and
plasma currents
electron temperatures
plasma densities
are
are
in the
from 400
to 800 kA.
from 1400 to
range
2
to
Cen-
2000 eV.
6 x 1014
Target
cm-
3
,
line
averaged.
The pellets last for 100 to 150
of temperature
discharge.
and
density
that
prevail
inside
the
Alcator
At the pellet's nominal velocity this corresponds
to penetration of 8.5 to 13 cm.
magnetic axis
plasma.
psec under the conditions
but
do
Penetration
shielding model 8 .
scattering are
is
in less than 500 Usec.
a multi-chord
deposit
their
fuel
in rough agreement
Density
initially
The pellets do not reach the
profiles
hollow,
as
deep
inside
with the
measured
but become
the
neutral
by
Thomson
centrally
peaked
Scattering measurements and those of
interferometer
show
profiles
with
peak
to
-
average ratios
of
about
4 -
two.
In
contrast,
profiles
without
pellet injection are flatter, with peak to average ratios of
1.2 to 1.4
(Fig.
2(a)).
By injecting more than one pellet it
is possible to double or triple the density without disrupting the
discharge.
Line
averaged
densities
up
to
1
cm- 3 have been achieved with central densities near 2
x
Following injection
to
the
density
falls,
original value in 50 to 150 msec (Fig.
time appears
to
increase
with
returning
3).
1015
x
1015.
the
The density decay
background
density
and
is
longer for deuterium than for hydrogen.
A sharp temperature decrease accompanies the density rise
(Fig.
3).
This is due to the dilution of hot plasma electrons
and ions with cold gas from the evaporating pellet.
temperature
profiles
emission (ECE)
x-ray pulse
are
measurements,
height
profiles regain
In
from
Thomson
analysis.
scanning Fabry-Perot
injection.
determined
The
their
Gaussian
addition,
unchanged although
the
considerably (Fig.
2(b)).
the
shape
width
magnitude
of
cyclotron
scattering,
ECE
interferometer
electron
Electron
soft
a
fast
instrument,
shows
that
within
of
and
250
these
the
temperature
psec
after
profiles
temperature
is
drops
The electron temperature recovers
to the preinjection level in 15 - 40 msec.
The ion temperature
as measured by neutron rate and Doppler broadening of impurity
lines shows
similar
electrons and
100 -
200 eV.
behavior,
overshooting
the
recovering
as
pre-injection
This is consistent
ion coupling at higher densities.
with the
quickly
as
temperature
the
by
improved electron
-
Total plasma current is
the long
skin
by
the current
profile.
expected
in
The loop voltage increases by about
.5
is
little
the
that
change
is
field
electric
pre-
to or below its
Calculations of field
40 msec.
diffusion suggest
no
change
volt immediately after injection then falls
vious value in 15 -
of
in
there
because
temperature profile,
the electron
Because
few percent.
a
seen
and
time
perturbed by pellet
only slightly
falling
typically
injection,
5 -
and current
is
perturbation
larger in the interior of the plasma but lasts no longer than
While this transient lasts,
that of the surface fields.
is high and
ohmic heating power
reheating
shown
netism and
most of
for the rapid
is responsible
that is observed.
A =
of
The behavior
loops is
the
in
p + li/2,
Fig .
3.
calculations
of
the
seen
change
changes in ap
not
of ap obtained
in
this
calculations of
Measurements
magnetic
after
li/2.
way
total
plasma
values
good
in
of
and
is
time
agreement
energy.
diamag-
indicate
injection
that
due
to
histories
with
Energy
pellet fueled plasmas has been as high as 80
responding
plasma
diffusion
pellet
The
are
Be
measured by a set of
kinetic
content
of
kJ with a cor-
ap of .5 at Ip = 750 kA.
MHD activity
usually continue
is
to
altered
sawtooth
by
pellet
but
with
injection.
increased
Plasmas
period
and
amplitude.
Sawteeth periods up to 50 msec have been observed
compared to
2
-
4
msec
seen
with
gas
fueling.
The
large
amplitude of sawteeth seen by the soft x-ray arrays is likely
-
due to
peaked
densities
density
and
impurity
m =
and
m =
large
injection.
6 -
These
2
oscillations
profiles.
At
3 MHD oscillations
are
probably
accompany
related
density threshold behavior previously reported
A variety
very high
9
to
the
.
of methods has been employed to calculate the
confinement properties
of
pellet
fueled
discharges.
The
TRANSP code provides an excellent method for analysis of time
dependent phenomena 1 0 ,'1 .
files
as
a
function
surface voltage,
tions,
electron
and neutral
of
TRANSP
and
ion
transport.
particle confinement
ties as well
as
time
the
ficient for
TRANSP,
along
solves
the
eneryy
Outputs
times,
experimental
and
magnetic
balances,
of
the
particle
correcting
are
rates
When
be
balances
which
data
voltage
simple
kinetic
similar calculations
using
calculation
ap data.
can
be
can be
is
insuf-
by
for
and allowing for the change in plasma energy, dU/dt.
from this
using
dI/dt
Results
compared
In addition,
and
diffusivi-
calculated
loop
and
equa-
energy
particle
neutron
can
current
diffusion
code
and
values.
confinement
models,
with plasma
thermal
beta values
compared to
standard profile
Given temperature and density pro-
with
simulations
have been performed using a 0-D time dependent model and time
slice modeling
with the ONETWO code
from GA1
2
.
In all cases
the confinement times quoted are determined after the plasma
has reheated.
It is clear that the consistent increase seen in plasma
energy at nearly constant input power is the result of improved
- 7
TE
improvement of
some increase
between gas
is
with density,
seen.
fueled
The
injection
in the saturated regime
issue
fueled
is the
comparison
The differ-
plasmas.
1 where plasmas which have been fueled
are
seen
the TE - ne
rollover can
expect
to
have
significantly
better
While these discharges show some weak saturation
confinement.
they follow
We
discharges.
even
crucial
pellet
and
ence can be seen in Fig.
by pellet
fueled
pellet
in
energy confinement
be
curve
explained
by
to
ion
higher
and
densities
losses
at
1
x
the
neoclassical
(Chang-Hinton).
Improved energy confinement at high densities allowed us
to reach record
levels
of plasma pressure and Lawson number,
ne(o)TE (Fig.
4).
Average pressures of
achieved with
peak
pressures
in the range .6 -
.9
x
over
8
1.6 atmospheres
atmospheres.
1014 sec/cm3 were measured,
the Lawson criterion for thermalized breakevenl
3
.
nfl
were
values
in excess of
These values
were reached at ion temperatures near 1500 eV, giving numbers
for nTT above 1017 eV of thermonuclear
sec/cm 3 .
Additionally, record levels
neutron production,
1 -
2 x 1013 /sec,
were
All of these parameters were achieved simultaneously
measured.
with 1.6 MW of ohmic heating power.
While the results of these pellet
injection experiments
are clear, the physics underlying the results is not.
nations of
gories.
the
improved
confinement
fall
into
three
Explacate-
The first attributes the improvement directly to the
peaked density profile.
The
second line of reasoning blames
- 8 -
the poor confinement seen in high density discharges on the gas
fueling
itself.
The high edge neutral density associated with
strong gas puffing leads to lower edge temperatures through a
variety of
processes
these
thus
Edge
current densities.
large in
and
to
lower
fluctuations
4
dischargesl
.
edge
are
Finally,
pressures
known
edge
to
fueling
detrimental to the plasma
in another way as well.
plasma does
allow
not
seem
since classical
particles
to
mechanisms
into
the
hollow profiles,
center
then an
hollow
are
at
density
incapable
a
rate
confinement
and all
it
that
need
to
is observed.
be
may
of
Other
investigated,
and
transporting
sufficient
to
5
avoid
work.
and, we may
for the degradation
explanations
meanwhile
be
Since the
anomalous process must be at
is also responsible
very
profiles
At least one such mechanism has been identified1
speculate that
be
and
may
in
exist
experiments
are
continuing.
In summary,
plasmas with
pellet
high
fueling
densities
experiments
and
peaked
have
produced
profiles.
Energy
confinement
is better than in comparable discharges fueled by
gas puffing.
Very high values of plasma pressure were reached
and the
Lawson criterion
for thermalized
breakeven was exce-
eded.
Acknowledgements:
The authors wish to thank PPPL for providing the TRANSP
code and
assistance;
for useful
discussions;
GA for providing
and
to
the
ONETWO;
entire
Peter Politzer
Alcator
technical
-9-
and support staff.
This work
is supported by U.S.
number DE-AC02-78ET51013.
Department
of Energy contract
-
10 -
References:
Fairfax, et.
1S.
Conference on
Research 2
al.,
Proceedings of the 8th International
Fusion
Nuclear
Controlled
and
Physics
Plasma
IAEA, Brussels, Vol. 1, pg. 439 (1980).
U.S.
DOE,
Proceedings
of
the
Fusion
Fueling
Workshop,
CONF-771129, (1978).
1491
3S.
L.
Milora,
4S.
L. Milora, C. A. Foster, et. al., Nuclear Fusion, 20,
Journal
of
Fusion
Energy
1,
15
(1981).
(1980).
5 M.
Greenwald, et. al., Proceedings of the llth European
Conference on
Aachen, Vol. 1, pg.
6S.
Fusion
Controlled
and
Physics
Plasma
-
EPS,
7 (1983).
L. Milora, G. L. Schmidt, et. al., Nuclear Fusion 22,
1263 (1982).
7
L.
Lengyel,
Max-Planck Institut
Fur Plasma Physik report,
IPP 1/213 (1982).
8
P. B. Parks,
Foster,
Nuclear Fusion,
(1977).
17, 539
9
R. J. Turnbull, C. A.
R. Granetz, Phys. Rev.
1 0 R.
Lett.,
49,
658
(1982).
Hawryluk, Physics of Plasmas Close to Thermonuclear
Conditions,
Proceedings of the course held in Varenna,
Italy, Vol.
sion of the European Communities,
1 1 D.
McCune,
1 2 W.
W.
Bull.
Am.
Pfeiffer,
et.
Lawson,
Proc.
Phys.
al.,
Soc.
General
1,
pg.
27,
19 (1979).
971
Atomic
Commis-
(1982).
Report,
GA
A16178 (1980).
13
j. D.
14
R.
Watterson,
et.
al.,
Phys.
Bull.
Soc.
Am.
B,
70,
Phys.
6
Soc.
(1957).
26,
885
(1981).
15B.
Coppi and C.
Spight,
Phys. Rev Lett.,
41,
551 (1978).
-
11 -
Figure Captions
1.
Energy
confinement
times
shown by the solid circles.
These points follow the
scaling law only below densities of 2
pellet fueled
discharges
are
fueled discharges
from gas
shown
x
1014 cm- 3 .
by
These clearly show higher confinement times.
T a ne
Data from
open
the
are
circles.
The solid curve
is the energy confinement time calculated by using neoalcator
scaling for the electron heat diffusivity and 1
x
Chang-Hinton
neoclassical ions (BT = lOT, Ip = 750 kA).
2(a).
Electron density profiles before and just after pellet
injection.
2(b).
The pellet was injected at 315 msec.
Temperature
jection.
profiles before and just after pellet
in-
Although the temperature falls due to the influx of
cold electrons
and
ions
from
the
pellet
the
shape
of
the
temperature profile is unchanged.
3.
Typical pellet fueled discharge.
Asingle deuterium pellet
is injected into a deuterium discharge at 390 msec.
4.
Values of the Lawson product,
ne(o) TE are shown as a func-
tion of density for pellet fueled plasmas.
i
0
50-
0
0
000 OD
40-
o
Oo
00
0
40 -0
3
0
0
00
20.4I
10
0 Gas fueled
0 Pellet fueled
0
12
3
4
(1GUR
5
6
7
8 9
3)
FIGURE 1
a"Z24
i
i
i
co
- --
00
0
E
--
0
(C A01jt) e
FIGURE
2
4
Ai.
x10]
-3
cm
0
1.5
Te (o)
keV
0
500
Ip
kA
0
V
2
VL
0
2.0
POH
MW
0
0.3
rOp
0
Neutrons
0
100
200
300
t (Ms)
FIGURE 3
400
500
600
Confinement Parameter
I
I
I
I
I
I
I
I
0.9
Pellet fueled
* Gas fueled
o
0.8
0
0.7
'E
0
0
0 og C
0.61-
0
0
Brea keven
000
0.5
0
OoO
SO
0.4
0
0
7
0
0.3
0
T
001
0
0.2
0*
0.1
eel?,*
I
1
_
I
I
2
3
_
I
I
I
I
I
4
5
6
7
8
ii(101 4 cri 3 )
FIGURE 4
8527