•
.2
i
I OF 1
ORNL P
1373
...
.
.
.
*
...
o
-
7
Laminer
.
1
-
.:
:
.:!
i.
A5
SO
56
4 0
re
11:25 | 1.4 LE
MICROCOPY RESOLUTION TEST CHART
NATIONAL BUREAU OF STANDARDS - 1963
ORHA-P-137

ROUGH DRAF
CONF-65093
i
JUL 20 1965
PIASMA HEATING BY MEANS OF BEAM PLASMA
INTERACTION AT OAK RIDGE NATIONAL LABORATORY*
I. Alexeff W. D. Jones
R. V. Neidigh W. F. Peed
W. L. Stirling
...
:
.
-
-
Britann100
-
LEGAL NOTICE –
The report no vepared u w account of Goveromeat sponsored work. Neither the Valled
stal, OJT Wbg Commission, asr day pormon acting on behall of the Commission:
. Kamu Ly murranty or reprenotation, expronar.implied, with respect to the accu-
racy, complctadans, or w Worn of the information contalood to the report, or that the we
of any information, appantu, method, or process declorod la Wo roport may not Infringe
prinu! omod ripats; or
B. Asmmes ay liabudes ou respect to the use of, or for 'magos resulting from the
ON of any inform loa, 1ruratus, molbod, or procon discloud ip de roport.
As und to the abore, "prko action na heball of the Commisolo" Includes wy
ployoor contractor of the Commulon, or employee of such contractor, to the extent that
much nmployee or coatrictor of the Commisslon, or acaployee of such cootractor propares,
s provides access to, may taformados por surat to all employamal or coainct
wich the Commisslon, or be employment will such contractor.

-
-
..
•
PATENT CLEARANCE OBTAINED. RELEASA LU
THE PUBLIC IS APPROVES. PROCEDURES:
ARE ON FILE IN THE RECEIVING SECTION.
*Research sponsored by the U.S. Atomic Energy Comission under contrac
with the Union Carbide Corporation.
ů :
CULHAM PAPER - Igor Alexer?
I.
Introduction
The strong plasma-heating effects of intense electron beams have
ka
been studied at Oak Ridge National Laboratory for some time. For the
most part, the work has been carried out empirically. The electrons
have been injected axially into aniform and mirror fields. The primary
results of this work can be summarized as follows:
1. The transfer of energy from the beam to the plasma can have
efficiencies of at least several per cent.
2. When heating electrons in an ordinary magnetic mirrcr machine,
any resultant instabilities need not cause severe plasma losses.
3. The resultant hot-electron plasma can have an electron temp-
erature a factor of 10 higher in energy than the energy of
the electrons of the initial beam.
The resultant hot-electron plasma can have a B (ratio of
thermal pressure to magnetic pressure) close to unity.
5.
The hot-electron plasma apparently.can be stable.
In dense plasmas, the heating effect can be sufficiently inteje
to "burn out" or completely ionize the neutral gas filling the
system.
In this paper, we first discuss briefly the development of the
Oak Ridge beam-plasma work, employing the most important points of each
experiment. These early experiments include the Neidigh pressure-
gradient arc, and the initial hot-electron experiments performed in a
small mirror machine.?
-2-
We also report on further hot-electron work in which electrons
are heated and trapped by an electron beam directed along magnetic
lines of force but off axis in a large mirror machine. Here, the
trapped electrons precess out of the electron team and produce a large
cylindrical shell. This shell is of interest in that the hot elec-
trons can be studied in the steady state, yet are separated from the
electron beam source. This shell possibly can be developed into a
hot-electron blanket that shields a plasma from neutral-gas ir?!.ux.
Finally, we report on the "burnout" experiments, in which a small
volumz of plasma is kept at a high degree of ionization in the
presence of a surrounding neutral-ges source, by means of very high
steady-state rates of applied power. In the early experiments, power
was applied at the rate of one kilowatt per cubic centimeter of plasma.
In the resulting neutral-depleted plasma environment, charge-exchange
cooling of hot ions should be rather small.
Therfore, any ion-heating
process should become more evident, and this is found to occur.
II.
Pressure-Gradient Arc
The investigation of the bean-plasma interaction at Oak Ridge
National laboratory commenced with the study of the pressure-gradient
arc, or. Mode II, which was first discussed at the Geneva Conference in
1958. Tie basic construction of the device is shown in Fig. 1. The
apparatus is immersed in a uniform ragnetic field. Electrons from a
hot cathode are directed along lines of magnetic force into a hollow,
tubular anode. Note that in this apparatus, as in all other devices
to be discussed, the electron stream must move across the magnetic
field to reach the anode. We suspect that this cross-field electron
flow may be very important to the plasma-heating process. Gas is fed
into the hollow anode, while the rest of the apparatus is kept at a
high vacuum by means of high-speed pumps.
Under some conditions of operation—no pressure gradient, for
example the plasma produced in the hollow anode appears to be in a
relatively quiescent condition, and streams from the hollow anode
along magnetic field lines in a relatively well-defined column (Mode I),
as shown by the top picture in Fig. 2. Under other operating conditions,
generally reduced gas feed or increased power input, the plasma column
"blows up," as shown by the bottom picture in Fig. 2 (Mode II). Such
blow-up behavior has been(also observed in the experiments of honuuren, *
reglur and Dirawan and Fanelli
The region of fundamental interest is inside the hollow anode,
from which energetic ions are emitted with energies several times that
availabie , from the applied d-c potential. A typical energy spectrum
of argon ions is shown in Fig. 3. The maximum ion energy in this case
is about 2.5 times that which could be obtained directly from the
appliei d-c potential.
A simple, directAutrent mechanism for energetic-ion emission is
easily found if we consider a plasma that contains a group of energetic
electrons which is not conrined by the magnetic field.
The energetic
electrons promptly escape to the walls of the system, and the plasma
is left positively charged. Such a plasma then accelerates energetic
positive ions from its surface.
Note that in the Oak Ridge pressure-
gradient arc, electrons are i'ree to escape to the cathode along lines
of magnetic force. Thus, any process that can produce energetic or
hot electrons in an unconfined plasma, generally will lead to the
ejection of hot ions. If we make the assumptions that the energy dis-
tribution of the hot electrons is Maxwellian, and that the ions are
cold, then the maximum energy of ions emitted from the plasma is easily
shown to be about 7 k To, where k is Boltzmann's constant and T. is the
electron temperature.
The next question is why there should be a threshold for the
energetic-ion-ejection process, which both. the Oak Ridge workers and
others have observed. Either there may be a threshold at which power
from the electron beam may become coupled strongly to the plasma elec-
trons, or there may be a threshold at which an electron-cooling mechanism
becomes ineffective. In our own work ve suspect tžat the second thres-
in me I
hold mechanism is the dominant one. We postulate that, the plasma
electrons are cooled by radiation from the excited states of un-ionized
gas atoms in the plasma. Such an assumption is supported by the fact
that the transition in the Mode II discharge , occurs when the gardening
-5
reduced
decreasing the goo feel
effect of the neutral gas is removed by the complete or, burnout."
the-t-ionized gas.
The assumption that "burnout" is occurring inside the hollow anode
is supported by photoelectric data. In Fig. 4 one sees that as the
power into the hollow ancie of a discharge 18 increased, the light
along the axis begins to decrease. A further increase in power causes
the light to drop suddenly by several factors of 20. Other experiments
previously published also support the "burnout" model. Finally,
detaileủ calculations for which there is not sufficient room in this
publication, predict the energy of the ions emitted from the burned
out region. The minimum power, W requred to produce burnout in a
given volume of plasma, V, is given by the equation
2
V
Pex ve AE V
.
Here, n is the neutral gas density inside the apparatus before plasma
is formed, Oey is the excitation cross section, w. is the mean electron
velocity, and AE is the work lost per excitation.
If the above analysis of the Neidigh Mode II, ion-emitting arc is
correct, two conclusions are reached: First, the plasma electrons
inside the hollow anode are efficiently heated by some sort of bean-
plasma interaction; and second, the neutral gas in the hollow anode
can be highly ionized, or burned out by applying a sufficiently high
power f.nput to the system.
III. Early Hot-Electron Studies
In our earl.er work, we confined the secondary piasma of the
pressure-gradient arc between magnetic mirrors, as shown in the top
half of Fig. 5, and heated the electrons by means of the axial elec-
tron beam emerging from the hot cathode. The result was the hot-
electron plasma shown on the lower half of Fig. 5. Here, the electrons
have a temperature of about 100 kev, or about 15 times the energy of
the initial beam of 7 kev. The picture shown here was made, using the
Bremsstrahlung x-rays, by means of a pin-hole camera.
The decay of this plasma following the turn-off of the electron
beam 1s illustrated by the oscillogram of Fig. 6. Each vertical pip
represents an x-ray photos arriving in a scintillator detector. The
height of the pip represents photon energy. Horizontally 18 plotted
the time after beam turn-off. Note that the x-rays are still in evi-
dense after 500 milliseconds. This decay time corresponds to that
expected from the scattering of the electrocis out the mirrors by
collisions with the unionized 8.98 in the system.
Therefore, this
hot-electron plasma is said to be stable.
In summary, the various diagnostic techniques reveal the following
properties of the plasma:
1.
The maximum electron energy is over 1 Mev.
2. The average electron energy is about 100 kev, or many times
the energy of the electrons in the initial bean.
The density is about 104 hot electrons per cubic centimeter.
4.
The plasma B (ratio of thermal pressure to magnetic pressure)
18 at least 1/10.
5.
The plasma 18 usually stable during the decay.
The heating efficiency 18 quite high - a lower limit 18
several per cent.
The rate of plasma loss during the steady-state heating
process cannot be more than about 30 times greater than the
stable (single-particle, collision-dominated) 1088 during the
plasma decay, or otherwise the heating efficiency would have
to be greater than 100% in order to maintain the plasma.
The manner in which the preceding conclusions were obtained is to
be found 1n the open literature -- The Duyasoet flertall Hotel, OF 3A,
pp: 'A-689 - A-695, November 2; 1964 m so we need not discuss them here.
The important results of this set of experiments should be re-empha-
sized, however. Plasma electrons can be heated efficiently and without
abnormal losses, by an electron-beam plasma interaction, and the
resultant hot-electron plasma appears to be stable.
-8-
IV. Development of a Hot-Electron-Plasma Blanket
Since the hot-eleciron plasma appears to be stable, it possibly
would make an excellent blanket to surround a thermonuciear reaction
and protect it from the influx of neutral gas. Also, the stable prop-
erties of the hot-electron plasma might aid in stabilizing the
thermonuclear plasma. Finally, the hot-electron plasma is produced
without the rised of electronic tubes, and so perhaps can be extre-
pola ted to large sizes and densities comparatively easily.
A large-scale, hot-electron blanket was produced in the device
shown in Fig. 7. In this apparatus, the separation of the magnets.c
mirrors is 170 cm.
The mirror ratio on axis is 2:1.
The pressure-
gradient arc is located in one mirror throat and is placed 15 cm
off axis. At the midplane of the apparatus, the discharge is 22 cm
off axis. Several electrode configurations have been used. Typical
details are a hot-tantalum cathode 1.6 cm from the hollow, gas-fed
anode.
The anode is '15 cm long and has an inside diameter of 1.25 cm.
The operating voltage and current are typically 7.5 kV and 600 mA,
respectively. The gas used 18 generally hydrogen.
The hot electrons produced by the pressure-gradient arc are
trapped between the magnetic mirrors. Since the trapped electrons are
& considerable distance from the magnetic axis of the mirror machine,
they find themselves in a strong radial magnetic field gradient. Conse-
quently, the trapped electrons precess azimuthally, and form a plasma
cyclinder, or blanket.
The plasma blanket forved by the pressure-gradient arc is shown in
the photograh of Fig. 8. In this case, the camera was mounted 25 cm to
Pal
-96
the right of the machine's axis, as shown in tue photograph, and was
looking axially through a mirror coil. Observation tangentail to the
plasma cylinder produces the sharper definition on the right side of
the picture.
That the plasma blanket was produced by hot electrons is illustrated
in Fig. 9.
Here the camera was mounted 25 cm above the machine's axis
again looking axially through a mirror coil.
A tantalum plate lowered
into the midplane cuts off the right half of the plasma blanket and
shows that the charged particles are precessing clockwise.
This pre-
cession direction conforms to negative particles.
The temperature of the hot eiectrons in the plasma blanket has
been measured by two separate techniques, ard the agreement between
the two sets of measurements is within about a factor of 3. The elec-
tron temperature has been measured by studying the spectrum of the
emitted x-ray photons, and by measuring the scattering time of electrons
by gas atoms out the mirror loss cones.
These techniques are more
fully discussed in reference 2. The electron temperature varies from
50 keV to 150 keV. The electron density has been computed using the
total flux of x-ray photons emitted, as discussed in reference 2,
and by measuring the heating effect of the hot electrons on molybdenum
dust particies dropped into the plasma blanket. The electron density
obtained in early experiments was about 10 cm , but this densiiy
recently has been increased to at least 10° cm3..
The success of the plasma blanket experiments co far appears to be
encouraging.
The electron temperature is quite high, the electron
density is increasing steadily from experiment to experiment, and the
-10-
plasma appears to be stable. The next step in the blanket experiment
is to provide some method of filling the machine radially with hot
electrons.
Possibly multiple pressure-gradient arcs at various radii
will succeed.
-11-
V.
Burnout Experiments
A.
General Burnout Considerations
Simultaneously with the previously discussed hot-electron
experiments, a second set of experiments have been carried out at much
higher plasma densities. In these experiments, attempts are being made
to study in larger systems the almost complete ionization found inside
the hollow anode of the pressure-gradient arc. We believe that both
the hot-electron plasma and the highly ionized, "burned-out" plasma
are aspects of the same phenomena.
In the hot-electron case, we have
a few very hot electrons, while in the burned-out case we have many
moderately hot electrons. Given enough input power and a strong enough
confiring magnetic field, we hope eventually to produce a burned-out
hot-electron plasma.
The first step to producing a highly ionized plasma was the
pressure-gradient arc itself, now referred to as "Burnout I."
The
next step was to enlarge and to enclose the hollow anode with the
mirror coils. This was "Burnout II." Most of the work on which we
shall report was done with "Burnout IV," which is shown schematically
in Fig. D
With
Histrof its operating characteristics in hitreme
GOMIS
Some additional work has been carried on, in the much larger machine,
to be dracarred later
"Burnout V," show schematically in Fig. With distrofitomaper-
ating-characteristio&r-16-Table 3.
In the case of the larger Burnout v
machine, the studies are as yet in a very preliminary state.
Air
menina
kan
makabahan a
ontdaring
-12-
B.
General Properties of Burnout I
The most striking feature of the Burnout IV machine is the sudden
decrease in the emission of spectral light when the power input exceeds
the threshold value. As shown in Fig. It the spectraî light output
per unit volume near the central axis of the device decreases by about
a factor of 1000.
The burnout phenomena seems to spread from the central axis of the
machine to fill the entire cavity of the device as the power input is
increased. Energetically, this is quite feasible since detailed calcu-
lations show that the power required to maintain the burnout phenomena
on the axis of the machine is only a fraction of the power required to
initiate burnout, tore Appendix. 2).
One disturbing possibility suggested to us by Dr. Sweetman, of
Culham, is that the light in Burnout IV disappears not because the
plasma is fully ionized, but simply because there is no plasma at all
contained inside the apparatus. To check on this distinct possibility,
we replaced the posma catcher on the anti-cathode side of the apparatus
with a perforated box, as shown in Fig. 12. Plasma streaming out from
the mirror throat along magnetic field lines is carried into the box,
where it is neutralized and becomes a source of gas. By measuring the
pressure rise caused by plasma flow into the box, we could estimate .
how much plasma was escaping through the mirror throat. Within the
limits of experimental error, we found that all the gas entering the
center of Burnout IV appears to escape as ionized plasma. For this
ionizing efficiency to be present the plasma electron density inside
the device must be greater than about 1012 electrons per cm3.
-13-
A second striking feature of the burnout phenomena is the emission
of enormous amounts of radio-frequency energy from the plasma. Potentials
corresponding to fields of 300 volts per meter have been observed at one
meter distance. We assume that the radio-frequency energy is evidence
of strong electron heating.
The radio-frequency radiation is found to be concentrated into
four rather well-defined regions. We can tentatively identify these
four regions with the lon-cyclotron frequency, the ion-plasma frequency,
the electron-plasma frequency and the electron-cyclotron frequency.
From the in-plasma frequency and the electron-plasma frequency, we can
obtain two values for the plasma density. Both values lie between 2015
and 1073 electrons per cm3. These values are encouraging since a value
of 10t" is required to shield the center of the apparatus from hydrogen
molecules entering the plasma with thermal energy.
The above experimental observations of Burnout IV can be summarized
as follows: The spectral light emission of Burnout IV greatly decreases
on increasing the power input above the critical threshold; there is .
effectively 100% ionization of the neutral gas entering the center of
the unit in the burnout condition; and strong radio-frequency signals
are emitted that suggest an energetic plasma is produced with a density
greater than 10“ electrons per cm". These gross features suggest that
we do have a dense,
le o
1001
incident gas
atoms.
wow.. mowita American can
n
i
-14-
C.
Detailed Properties of Burnout IV
We have made many measurements on Burnout IV that help to clarify
what is going on 108i.de the device. However, these observations merely
supplement the fundamental observations discussed above, and do not
modify greatly ow. fundamental model.
The plasma potential on the axis of the device is found to be
approximately cathode potential, that is to say, the axis of the machine
is about 10 kV negative with respect to the surrounding wall. Such a
high radi.al potential produces a strong 7 x electric field in the
azimuthal direction, so we should not be surprised if the plasma appears
to rotate violently.
The plasma potential on the axis of the machine was inferred in the
following manner: First, we know from the studies discussed earlier
that essentially 100% of the gas entering the device leaves as ions
through the two mirror apertures. Thus, the effective ion current to
the anti-cathode plasma collector corresponds to half the input gas
flow of hydrogen or 1/2 x 0.15 cm3 sec-2 SIP. This corresponds to an
ion current of 0.61 A. Since the plasma collector is electrically
floating, so the ion and electron currents to the collector must be
equal. The heating effect of this curreät to the plasma collector was
measured to be 0.45 .5W. This power input corresponds to the 0.61 A of
current being accelerated through about 750 V.
If the plasma collector
were more negative than the plasma potential by 750 V, then the ions
would be accelerated into the collector, and the power dissipated would
be larger.
If the plasma collector were more positive than the plasma
potential by 750 V, then electrons would be accelerated into the
-15-
collector, and the same argument would apply.
Thus, the plasma
collector must be within £ 750 V of the plasma potential. Since the
plasma collector is observed to float at about -10 kV with respect to
ground, or at about cathode potential, we infer that the axial
potential of the plasma is about -10 kV 1 0.75 kV with respect to
ground.
Since a strong radial electric potential exists in the Burnout IV
machine, and since charge exchange cooling of hot ions has been reduced
strongly by the "burnout" process, then som: rotary plasma motion might
be expected. Evidence for a well-organized rotary motion in the pilasma
of Burnout IV was obtained using a simple charge-exchange technique.
A schematic is shown in Fig. 2. Neutral hydrogen atoms formed by
charge exchange of what gas is present in the plasma escape through a
pinhole and strike & glass plate coated with yellow Mo Og. This chemical
compound is not affected by molecular hydrogen. However, on being bom-
barded by atomic hydrogen or hydrogen ions, this molybdenum oxide is
first reduced to a blue sub-oxide, and then to grey metallic molybdenum.
One therefore would expect to see on the glass plate an outline of the
plasma produced by fast hydrogen atoms formed by charge exchange.
The actual picture formed by the fast hydrogen atoms striking the
molybdenum oxide are not at all what might be expected. A typical
picture is shown in Fig. 16.
Instead of an image of the plasma, one
finds a vertical streak.
This streak implies that the ratio of azimuthal
to axial energy is very high. Also, the streak extends only from the
image of the axis to the image of the edge of the plasma. This implies
first, that all the ion orbits must link the axis of the machine.
.
sostenem
Antenandrolone
inha

-16-
Thus, for example, a rotation of the plasma as a solid body about the
machine axis is permitted. A second implication is that some ion
orbits must be very large, since they both link, the axis of the machine
and extend to the edge of the plasma. This largest orbit diameter of about
_ cm in a Z kg magnetic field suggests that lons approaching
the outside edge of the plasma have an energy of 25 de K. Thus, the
charge exchange pictures suggest that a process by which ions become
hot 18 present.
A second way of observing energetic ions is to extract the ion
beam directly by means of a plasma container that has a step, as shown
in Fig. 15, the lon beam then impinges on a molybdenum-oxide glass
plate and leaves a streak, 8.8 occurred with charge-exchange neutrals.
If a transverse electric field is applied, the beam of ions is deflected,
as shown in Fig. 5. The results of the combined electric and mag-
netic field analysis reveals that the charged particles are, as expected,
deuterons, with an average energy of 2. kev.
Thus, again we find
evidence of hot ions in Burnout IV.
More detailed studies of the hot-ion populacion ob Burnout IV were
made using a small electrostatic energy analyzer that projected into
This
the plasma. Two groups of energetic ions were found. The first group,
lying at about 70 eV, suppotts the model predicting a maximum lifetime
for ions at this energy when immersed in a sea of neutral g
first group is obvicusly formed near the metal walls of the apparatus,
outside the burned-out region. The second group is found at an energy
of about 25 kev. If one accepts the model that burnout is indeed
present in the body of the p.lasma, then this second ion group can be
-17-
identified with those ions having a maximum lifetime between the short
lifetime group dominated by ion-ion scattering out the mirrors, and the
short lifetime group dominated by ion-neutral charge exchange. An
analysis similar to that given in reference shows that for 25 kev
lons to be present, the neutral density must be 10-3 times less than
the ion density. In any case, the electrostatic analyzer again shoris
that hot ions are present.
Further experiments that give only suggestive results are
tabulated below:
First, d-d neutron production shows that some ions in the device
are energetic.
However, we have not positively identified any neutrons
as being produced inside the plasma. ,
Second, attempts to observe the cut-off of microwave radiation in
the plasma to observe the plasma density seem to show an effect at wave
lengths corresponding to a few times 1072 electrons per cm3. Unfortu-
nately, the transmitted signal sometimes is increased rather than
attenuated at the critical density, and we do not yet know if we have
an instrumental effect, or an actual amplification occurring in the hot
plasma.
Third, attempts to study the penetration of a neutral beam of gold
atoms through the plasma show that the beam is indeed severely atten-
uated. However, sparking in the cathode circuit resulted in the
transmission measurement being an average over conditions with and
without plasma.
In any case, for continuous plasma operation, no
transmitted beam whatever was expected.
-18-
The energy distribution of the electrons in Burnout IV was
inferred from pinhole pictures made on stacked x-ray plates. Since
the image penetrates through several plates, we inferred that about
50 keV x-rays were present when about 7 keV was applied to the cathode
of the apparatus. At present, this is our only evidence for electron
heating, as the intense radio-frequency energy emitted by the Burnout IV
Jams the photon counters.
Thus, the supplementary experiments done on Burnout IV tend to
support the primary observations that one has a confined region of
plasma in which the neutral population has been reduced greatly and in
which sone energetic, ion-heating process occurs.
-19-
D. Properties of Burnout v
In the hope of learning the scaling laws governing the burnout-type
machines, we decided to build a larger version. Since the power required
to burn out a given volume of plasma scales directly as the volume, we
decided to keep the volume of the device small.
Therefore, to contain
interesting plasma densities and ion energies, we had to use strong
magnetic fields.
A schematic of the Burnout v device 1s shown in Fig. 26. The
magnetic field 18 30 kg in the midplane, and 18 maintained in steady
state operation. The mirror ratio 18 2:1. The total volume of the
plasma liner between the mirrors 18 about 15 liters.
Our experiments with this device so far lave been mostly adjusting
it for steady, efficient operation. The maximum power input to the
machine has been about 50 kW at 20 kW and 2.5 A. Burnout apparently
does occur in that the light emitted by the plasma streaming from the
end of the mirrors disappears at high power inputs. The light from
the center of the machine decreases, but does not disepper.r. We suspect
that the axis of the machine goes dark, but that a flowing shell still
remains, i.e., the burnout does not spread. At present, structural
problems have prevented our making a radial scan of the emitted light
.
.
.
a
to look for the flowing shell.
~..
Measurements made on the intense radio radiation emitted by the
-
Burnout v machine reveal, as in Burnout IV, four well-defined frequency
bands. Again, two frequency bands tentatively can be identified with
ion-cyclotron and electron-cyclotron rotation, and two with the ion and
the electron-plasma frequency.
The two bands assigned to the ion and

-20-
the electron-plasma frequency again give a density of several times
2012 electrons per cm3.
Microwave cut-off measurements were also attempted in Burnout v.
Here, a modulated oscillator was locked to a phase-sensitive detector
to provide maximum no.1, se rejection. For the machine in Mode I, cut-off
was observed for 3 cm microwaves, but not for 1.2 cm microwaves. Thus,
for the quiescent mode, the density lles between 1 x 101 and 8 x 1015
electrons per cms. However, for the burnout, active mode, the signal
increased in power by up to a factor of 10 over the empty cavity con-
dition. Later experiments showed that the microwave power lost in the
device corresponded to only a factor of 2-3, so perhaps we have observed
some amplification of microwave energy by an active, noisy plasma.
One very unusual observation was made in Burnout v with a probe
that measured the heating effect of the plasma. The results are shown
in Fig. 27. As the probe is inserted deeper into the plasma, the
power reaching the probe increases. However, at a given probe position,
we find that the power increases drastically as the magnetic field is
increased. This observation appears to be directly opposite to what
'
would be intuitively obviouc - that the stronger the magnetic field,
the better would be the plasma confinement.
Many other observations in Burnout v correspond to those in
Eurnout IV.
Charge exchange pictures using the sensitive MgO, films,
as previously discussed, also show the plasma ions to be rotating in a
highly organized fashion, as 1:a Burnout IV. Neutrons also are emitted
with a source strength of Oman, but most of these are found to
originate from the plasma collectors. No clear-cut identification of
· d-d reactions inside the plasma has yet been made.
21-
In general, the observations made on Burnout v tend to support
...020 made on earlier machines, although the Burnout v measurements
.:less complete.
The scale-up does seem to be successful in that
urge quantities of power can be fed into the machine in a steady
inte fashion, and that some of this power 18 being coupled to the
plasma.
down into consistentiemodligere
sic
u
m
REFERENCES
[]
[3]
[4]
[5]
NEIDIGH, R. V. and WEAVER, C. H., "Proc. 2nd UN Int. Conf.
PUAE (1958)
ALEXEFF, I., NEIDIGH, R. V., and PEED, W. F., Phys. Rev. 136
(1964) p. A689.
HARRISON, E. R., Nature 184 (1959) p. 245.
NEZLIN, M. V., Soviet Physics JETP, 14 (1962) p. 723.
DARWIN, H. W. and FUMELLI, M., S.R.F.C., EUR-CEA-FC-270,
Sept. 1964.
[67
ALEXEFF, I. and NEIDIGH, R. V., Phys. Rev. Letters 13 (1964)
p. 179.
ALEXEFF, I., NEIDICH, R. V., and SHIPJEY, E. D., Phys. Fluids 6
(1963) p. 450.
[7]
.
Figure Captions
Fig. l
Schematic Diagram of Apparatus for Production of Mode II
Arc in a Uniform Magnetic Field.
Fig. 2.
Top - Normal, Quiescent Plasma; Bottom - Ion Emitting Plasma.
Note in the lower picture the broad band of light which is
due to energetic ions ejected across the magnetic field.
The axial magnetic field runs left to right. In these pictures,
there is a plasma source at both ends of the plasma column.
Fig. 3.
The spectrum of energetic ions emitted by a Mode II arc as
detected by a gridded probe.
Fig. 4.
Radial light distribution inside the hollow anode of a pressure
gradient arc operating in nitrogen. An Abel inversion was used
to obtain light output per unit volume.
Fig. 5.
Visible hot-electron plasma and its X-ray image. Superimposed
outline shows magnet coils and electron-beam defining aperture.
Dumbbell-like ends on the image are caused by X-ray fluorescence
at the magnetic mirror throat.
Fig. 6.
The X-ray flux decay after turnoff. A smooth decay lasting well
over 0.5 sec is shown.
Fig. 7.
Schematic of the beam-plasma device producing the hot-electron
blan:et. The scale can be inferred from the 170 cm separation
of the magnetic mirrors.
Fig. 8.
Photograph of the hot-electron blanket (crescent) emerging
from the Mode II arc (bottom). The view corresponds to looking
in through the left-hand mirror in Fig. 7.
· Fig. 9.
Photograph of the hot-electron blanket being intercepted by a
paddle (top). Only the left-hand part of the blanket appears
showing that the blanket is formed of precessing negative
particles.
Fig. 10.
Schematic of "Burnout IV". The volume of plasma intercepted
by the electron beam was about 10 cm. A deuterium gas input
of 1/10 cm/sec is used to provide 10-3 torr between the mirror
and 10-5 torr in the cathode and anticathode regions. The
mirror field is 9 KG on the midplane and 13 kg in the mirrors.
A beam power input of 7 kw (10 kV at 0.7A) is applied.
rommene.
-
-
Fig. 12
The decrease in the light output per unit volume in Mode I and
Mode II for Burnout IV, Helium gas was used.
Fig. 12.
ers twica wietnams..dit podw.col.vidade -----
Schematic of apparatus used to measure the efficiency of ionization
in Burnout IV. Plasma is collected in the plasma catcher on the
right, neutralized, and the resultant pressure rise compared
with the known gas input. Within the limits of error (+ 25%)
all gas entering Burnout IV emerges as plasma.
h
ot
in more than
Fig. 13.
Schematic of technique used to obtain an image of the plasma
using charge-exchange neutral.s.
rig. 14.
Pinhole images of ion orbits. The pinhole forming the center
image was on the midplane. The other two were 1.25 cm on the
midplane. The other two were 1.25 cm off the midplane, but
on the axis.
Fig. 15.
Application of Electrostatic and Magnetic Analyzer to Burnout IV.
Part (a) is a midplane section; (b) shows the marks on the Moon
detector. Since the deflected mark shows no z spread, only length,
it, must have been made by monoenergetic ions (25 kev) passing
through the pinhole at different angles of incidence but all
lying in the midplane.
Schematic of Burnout v.
Fig. 16.
Fig. 17.
Power deposited on a grounded, water-cooled probe as a function
of distance into the plasma f'rom the cavity wall. Note that
as the magnetic field increases, the probe power increases,
while the neutron counting rate decreases.
ORNL-IN-OWO 26002
MAGNETIC FIELD
ON TRAJECTORIES
NEGATIVELY BIASED
TUNGSTEN FILAMENT
HEATED BY ION
BOMBAROMENT
FLOATING ELECTRODE
(ELECTRON REFLECTOR)

.
ARC DEFINING MOLE
ELECTRON SHEATH
ARC PLASMA

ION TRAJECTORIES
SAS FEED UNE
ENO WALLS AT
GROUND POTENTIAL-
-END WALLS AT
GROUNO POTENTIAL
arti
e
Schematic Diemen of Apparatus for Production of Made II Arc in Uniform Magnolle Field
.
.
.
. .
.
e, -..-
m
is dat mensen miniesos
ماممسناتنمنننهننهننمينوبنات.ة من
Ting
en
:**

When
,::.!
.
miinom
.
terenaine on
a
..
".
.
animiranih
ililib
Wings
Azeri
"SANTI MUSIM w
ith
VACL ASS.rifo
nro: 471?
*
...
..
.
..
30
..i
..
iii.
.
se
,"";
4
,
;.
......in
,
:
Y:
::
..
ci:
.
..:::...
.,.-.
TO
VA
viure
i..
:
T
.
heim
: mitt.
...:
.::
.
'i'..
N
...
:1;
...',ir
.
??
..
!!
I...
.
,
.
.
.
*
...
..
.
;
;::-
IN
Fig. 4.2. Top - Normal, Quioscone Plasmo; Bottom – lon Emitting Plasmo. Note in the lowor picture the
broad band of light which is due to energetię ions ojected across the magnetic field. The axial magnetic field runs
loft to right. In these picturos, there is a plasma source of both ends of the plasma column.
:
.
hay 2,

..; , : ; .. | ..;-
--4-4. -! ..
.*
ܙܪܐ-ff-;.-
܂
ܐܐ
ܨ-;;
ܝܝܐܪ
iiiftji
:;
n..f-
nn-
ܐ.... ܫܬܫܐܙܐܪܫܙܙܐ ܙܟܐ
-'-+-+-+-+- +--ܪ-1-
re reaching colleter
.ܢ-
.ܛ..ܝܼܲ
ܪ ܪ ܙ ܐ ܀
---;-;-ܙܝܙf
.܀ .4 .; .;
.ܕܫܰܚܪ
r'
ܙ'
;.i.7.;. it.;;.j.;.4..
ܕ ܙ - '
ܙ |
-
ܙ
' .
ܙ
•• •
ܪ
ܫܙܫܕܫܐܫܐܫܫܕܕܫܕ ••••
. ܙܫ .
• •• •
..
ܪܪܛܪ
܀ . ܐ .
ܕܲܐܲܪܼܪܼܲܛܼܲf-.
ܝ ܐܝ ܙ
-
f-f.+++-
---
-
-f--
;::-:-: '
. ܙ: 4
.
8 . ...
: .ܐ: ܐ : !ܐ ܙ ܪ
burdinice p
܂
ܘܙܫܙܝܫܙ-ܪ
܀.:.
.,܀܀ ،، ܐܼ ،
ܐ
ܝ
ܢܐ
. . .. ܐ• •• • • • •• •I --
--3-ܙܫܕܪ-;.ܨ ܫܫ. ܙ-;-:-܀
-. -4 . . ..
• •• ܂ • •
• • • •
•
;-3;
;
The time
܂
. . .܂ܫܢ .ܪܝ
. ........
ܟ ...ܝ܂ . .
eg
.
:
;
.
sae '. ܝ ܂ 328 ܂ ܂ ܩܘܐ
...
ܘ
ܘܐ _.
ܘ
in ensure er
܂
.
.
7-; ?
.
.
.
.
.
. . ܂

ܕܕܐ
ܝ
ܝ
:
ܘܢܐ
܆
.
܂
. .
.
. ܕ. ..… .:
.܂ 7
܂ '
܂
܀ ::: ;:i"" ,
':'.;
.o; ;;;!
- .q: ܙ. ܝ
"
܂ . ; ". ܛܼ
r ' .,.
, :
;
. .܀܀
. ܚ
ܙ
ܙ
;

LIONT DISTRIBUTION
MODE 2, VOLUMENICM
WATTO
| OAK RIDGE NATIONAL LABORATORY
280psgrii.
rini... :
LIONT INTENSITY
i
meninino
RADIUS (MILLIMETERS)
1
hominis

AL
.
VISIBLE PLASMA
ir

PAN
.
12
.5
.
-
T
NO
.:
X RAY ENERGY
20
.
.
..
.
AD
empresarios
.
IN
800 MILLE SECONOS
SMOOTH DECAY, LONG que
i.
imgarren
X.RAY IMAGE
e os
l'ic. 1, Visible bot-clectron plasma and its x-ray image. Super.
imansed outline shows magact coils and electron-bear defining
inc. Dumbbell-like ends on the image are caused by Iray
eccoco al tbe dagoetic-mirror throat.
d
Frings
Hang.6.

UNAISSIRED
CRYL Oil 65-5528
COILS
COILS
HOT ELECTRON BLANKET
SMOOC
FCATIVODE
proyecto
punim me to
GAS-
REFLEX DISCHARGE
LANT-CATHODE
170 PLAVPS
FIG. 7- BEAM PLASMA FACILITY
AXIAL CROSS SECTION
.་
་
་ ་བ
་ ，
-- ཀ་བ་ཀ་ཆབ༠，
•T，
་་་
.:.
：་:：
.
བ
༤
，
，， ༔
•，
" •
,，..
:
ཆབས་
•
•i!
•
རྒྱུ
•
་
:་,
་
་. ་
ས
་
་
：
-
• ་.
•.
.
་. ་་' ..
•， ，
，
.：，
་་
་
བ
• •
•.་
་
་
:: ....

ting &
Licenciamento prestations in
courtano
::::
.
22.4
:::
.
.
;
..
..::...*
..!
.
*
ri
4
!
119.7...
..1;&::,..
;..'.
.
:
3
.
::
.."
.
'.in.:
.:
.
.::
...
0201
more
.
i
.
7
..
..
-
..
..-
.
.
wwwamos
....
----



..I.
11.Ol.........
1
FILAMENT
COIL
COIL
INSULATED
SHIELD
UUOOLIPOIIIIII
WSULATED
SHIELD
PLASMA
...
0
00UITO
.
COIL
COU
DIFFUSION
PUMPED
DIFFUSION
PUMPED
GAS
FEED
avoin
Ting 10
..
.
.
.
-
-
-
-
ORT-om
64.11760
☆ ELECTRODE
(ELECTRICALLY
INSULATED)
AX ELECTRODE
(ELECTRICALLY
INSULATED)
MEASUREO GAS
ROW RATE

0011
1
D
.
.
.
.
-
-
.
.
.
-
.
.
.
.
.
-
.
.
-
LIINUNNIPIVNIM11122
.
.
.
.
NEGATIVELY
BIASED
FILAMENT
.
COIL
.
.
.
.
.
!
.
+
.
1
MASMA
OPEN
GRO
WI
PU1+1
COIL
...werneuron
//777777777
90°
.EU
BENO
REDHEAD ION GAGE
TO MEASURE
PRESSURE OURING OPERATION-
Fig. 4.2. Method of Measuring Plasma Pumping.
The insulotod AK Olectrode is more on the form of a
plasmo corcher with o 40% chance of catching all the
plasmo, Plasmo pumping rate is determined by re
producing, with a measured gas flow roto, tho prossure
rise in the electrode due to plasma.
Fig 12.
"
:
ONML -OWO 4- Ham
MAGNETIC
FIELO

{
PRIMARY
ELECTRON
BEAM
ION
XIS
PINHOLE
NEUTRAL -
MOO, PLATE
LENGTH OF M0203
Fig. 4. 5. Schematic Orogram Showing Rinhole Image
of lon Orbited in this mittenco, the image is formed by
nouirate. Wola shor the lon ontells must pass through
the oxis as gida tiho osymmomic image.
Ting 13
PHOTO 71335
...
Moro na
.a.,
"Loro.
progr
*
INCH
Fig. 4.6. Pinhole Images of lon Orbits. The pinholo
forming the contor image was on the midpleno. The
other two
off the midplano.
Fig. 14.
kitietoimintansin sa araw
mante wat die gerilim wwwe.
1.9.
ON NL-omi 4.11nI

PINHOLE

BURNOUT IV POSSIBLE NOW onoirs
al cm OEFLECTION
-
Molly
COATED GLASS
26 Mov O ONOIT THAU AXIS
GROUNOCO PLAT
- OEFLECTION
PLATE
USS THAN TANGENT-
TON MARK WITH
ELECTROSTATIC FIELD OF
ION MARK WITH
TUCTROSTATK
FKLO ON
CLECTROSTATKC NELO -
000/am
8 kov O
ION-
- PINHOLE
ELECTROSTATIC PLATES
- MANKS ON THE MOO, COATED GLASS
Fine Arte Application of Eloctrostatic and Magnetic Analyzer 10 Burnout IV. Port (n) in o midplano section;
(8) shows the marks on the Moo, dotector. Since the dellected mark shows no s spread, only longth, it must have
boon modo by mono-norgotic lons (25 kov) musing through the pinhalo of difforent angles of incidence but all
lying in the midplano.
Tag 15.
C
"Ournout IIIPlasmo Volume ~ 10 cm

II.
III.
I
.1.
I
I
.
12
!
!
!
!
!
.
1
.
L
.
17.11
►
I
1
1
.
!
I
!
2
!
!
I.!!
1
-
FILAMENT
COILY
COIL
10
I
.
1
1
1
1
1
1
1
!
!
!
T
10
1
1
1
1
1
PLASMA
.
LINSULATED
SHIELD
COIL
UHU
WA
IIIIIIIIIIIIIIII
III . 1 "
IIIIIIIIIII
100..1 III
O
oo..
I
UITO' IIIIIIIII
LINSULATED
SHIELD
GAS
FEED

Fing. 16
Jag 17
that
2222
Thing 17
urhoomail
natin Kashing P
netic field
.
.
10
1
c
.....
C
0
.
1
11
:
11
UOC
.
1
.
.
-
1
.
.
.
.
I
'
O
.
Contrate Cats/mund
1C
1
.
1
.
11
.
U
o wartime se ceara o
.
.
0
U
O
C
C
1
1
1
.
1
MILLIMETER
M VICT;.:: Goarns
C
.
C
1
0
0
11
1
1
1
1
1
C
MOV

END
:
DATE FILMED
11 9 / 65