REVIEW OF SCIENTIFIC INSTRUMENTS 87, 025105 (2016)
The upgraded Large Plasma Device, a machine for studying frontier basic
plasma physics
W. Gekelman, P. Pribyl, Z. Lucky, M. Drandell, D. Leneman, J. Maggs, S. Vincena,
B. Van Compernolle, S. K. P. Tripathi, G. Morales, T. A. Carter, Y. Wang, and T. DeHaas
Department of Physics and Astronomy, University of California, Los Angeles, California 90095, USA and
Institute of Geophysics and Planetary Physics, University of California, Los Angeles, California 90095, USA
(Received 16 September 2015; accepted 10 January 2016; published online 9 February 2016)
In 1991 a manuscript describing an instrument for studying magnetized plasmas was published in
this journal. The Large Plasma Device (LAPD) was upgraded in 2001 and has become a national
user facility for the study of basic plasma physics. The upgrade as well as diagnostics introduced
since then has significantly changed the capabilities of the device. All references to the machine still
quote the original RSI paper, which at this time is not appropriate. In this work, the properties of
the updated LAPD are presented. The strategy of the machine construction, the available diagnostics,
the parameters available for experiments, as well as illustrations of several experiments are presented
here. C 2016 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4941079]
I. INTRODUCTION
This original Large Plasma Device (LAPD) machine1 and
its upgrade were both designed to accommodate a variety
of experiments, which could be rapidly reconfigured. The
machine is presently part of a user facility for basic plasma
physics research. An experiment can run from several days to
several weeks. When it is completed, the current experiment
must be removed from the device within several hours and
the next one set up within a day. This involves removing
all unneeded probes, antennas, and other specialized devices
and replacing them with new ones. Since the vacuum can
never be broken, vacuum interlocks and pumpdown stations
are required.
The original LAPD had a 10 m long plasma column.
The source was a barium oxide (BaO) coated cathode, which
provided a pulsed DC discharge. The plasma length enabled
the study of Alfvén waves, which have parallel ((to B0z)
the background magnetic field) wavelengths on the order of
meters. The machine made possible the study of shear Alfvén
wave (SAW) cones2 and their propagation in the kinetic3
and inertial2 regimes. Landau damping of the waves was
observed when the wave phase velocity equaled the electron
thermal velocity.4 When Alfvén waves reflected from the ends
of the machine, field line resonances were observed.5 Large
amplitude Alfvén waves in which electron heating and density
perturbations were also studied.6 Processes related to space
plasmas such as mode conversion of whistlers to lower hybrid
waves on a density striation7 and the interaction of lower
hybrid waves with striations8 all took advantage of the 10 m
long plasma column.
In 2000, the LAPD device was upgraded using a Major
Research Instrument (MRI) award from the National Science
Foundation, funds from the Office of Naval Research as well
as matching funds from UCLA. The device is housed in a
newly constructed building on campus, which was designed
to accommodate large machines that need a great deal of
0034-6748/2016/87(2)/025105/19/$30.00
power and cooling water. The new machine, also called the
LAPD (the original one was de-commissioned), has an 18 m
long plasma column. The guiding principle for the design
was to make the machine as flexible as possible. The key
was not to envision experiments that were improvements
on what was already going on rather to engineer it for the
largest possible parameter space and diagnostic access to
accommodate experiments that had not been thought of yet.
The upgraded LAPD is shown in Figure 1.
A. Vacuum system
The LAPD is a linear device. The vacuum chamber is 24.4
m long and consists of two end chambers (2 m long 1.5 m
diameter), a main section which consists of 10 cylindrical
vacuum vessels made from 3/8 in. stainless steel with end
flanges rolled from 1.5 in. by 3 in. stainless steel bar, and
O-ring grooves cut into them with a CNC lathe. An additional
vacuum chamber to house the second plasma source is 1 m
long.
The tungsten inert gas (TIG) welding technique was used
on the central seam. A total of 360 circular ports were cut into
the working chambers on a large lathe. 6 in. OD, 1 in. wall
stainless steel pipe was used to make the circular ports. First,
end flanges with “O” ring grooves were welded on to the pipe.
This in turn was vacuum welded on to each chamber. Every
port is identical to allow diagnostics to be interchangeable
between them. The working port diameter is 4 in. and can
accommodate adapting flanges to KF-40 or KF-50. This limits
the size of probes that can be introduced into the vacuum
system to about 3.5 cm in diameter. It was recognized that
larger objects such as antennas had to be introduced into the
plasma. To this end, larger rectangular ports were installed
2 m apart along the device axis. Each of the “octoports” (thus
named as there are eight of them surrounding the chamber) has
eight rectangular ports 4 in. wide and 12 in. long. There are
eight octoports along the machine for a total of 64 rectangular
87, 025105-1
© 2016 AIP Publishing LLC
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-2
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 1. (a) Composite photograph of the LAPD device. Shown on the left is a pump-out chamber, which contains the BaO cathode and two of the four 2200
l/s, turbo-molecular pumps. The magnets and supports fill most of the scene and the chamber containing the lanthanum hexaboride second cathode is on the
right. The entire machine is 24.4 m end-to-end. (b) Machine drawing showing the two types of ports as well as the connections to the power supplies. The
yellow magnets are from the CCT tokamak at UCLA (and before that the ACT tokamak at PPPL). The purple magnets were constructed at UCLA. An additional
magnet is used to provide a magnetic field at the second LaB6 cathode.
ports. The flange drawings for the circular and rectangular
ports are given in Figure 2.
The octoports were fabricated in two stages. First,
rectangular holes were plasma cut into the vacuum vessels and
then finished with grinders. Ports could not be welded directly
to the large rectangular holes as the vessels would deform
under pressure. Instead, rings fabricated from stainless steel
that fit snugly over the chambers were fabricated. Eight faces
were milled into each ring and “O” ring grooves were cut into
the flats with rectangular holes. They were then slid over the
chambers and welded into place (see Figure 3). Needless to say
this had to be done before the end flanges on each chamber
were welded into place. The end flanges were rolled from
1.5 × 3 in. 304 stainless steel bar and welded into rectangular
cross section rings. “O” ring grooves were cut into them on
a large lathe and ¾ in. diameter holes at 22.5◦ intervals were
drilled into them so that the vacuum chambers could be bolted
together. The bolts were insulated with G-10 sleeves so the
chambers could be electrically isolated if necessary. One ring
was milled to have a flat “O” ring surface and its partner had
a 0.5 in. wide, 0.375 in. deep “O” ring groove. Square cross
section Viton “O” ring material was used for a vacuum seal.
Most of the circular ports are used for diagnostics probes
(Langmuir probes, magnetic pickup probes, electric dipoles,
fiber optics, etc.), microwave interferometers, and optical
measurements. Many of the probes are mounted on unique ball
valves,9 which allow for positioning of the probe at any angle
with respect to the normal to the chamber wall. The probes can
be attached to computer driven probe drives, which accurately
position them (to within 1 mm) in the machine. The probes are
aligned using a transit mounted in front of a large window at
the end of the device using carefully placed fiduciary markings
inside and outside the machine.
The barium oxide plasma source does not tolerate exposure to oxygen (or any electronegative substance); therefore,
the probes have to be placed in vacuum interlocks and pumped
down to pressure of 3 × 10−6 Torr before being introduced
to the vacuum system. This is accomplished using portable
FIG. 2. (a) Dimensions for adapting a plate and limiting size for introducing probes/antennas for one of the 64 rectangular ports. The holes are tapped for ¼-20
bolts. (b) Relevant dimensions for one of the LAPD circular flanges. All dimensions are in inches.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-3
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 3. Dimensions of stainless steel “rings” that became octoports. The
rings were 4 in. thick and made by rolling and welding 6 in. × 4 in. 304
stainless steel flatbar. They are then machined to have 8 flat faces, each with a
4.5 in. × 13.88 in. rectangular port (see Figure 2). Matching rectangular holes
were cut in the circular vacuum vessel walls and the “rings” were welded in
place.
pumping stations, which use cryogenic pumps or small turbo
pumps to evacuate the interlock. A photograph of such a station
at work is shown in Figure 4. When larger objects (antennas,
electron beams, ion beams, and additional cathodes) are
introduced, they are placed in “boxes” mounted on the
machine. Rectangular gate valves are bolted to extensions on
the octoports and in turn the boxes are bolted to the valves.
A port capable of mounting a cryopump or a turbo with an
additional valve is mounted on the side of the pumpdown
box. A drawing how the boxes are mounted, the specialized
rectangular gate valve and cryopump access are shown in
Figure 5.
FIG. 5. Right is a cutaway view of the vacuum vessel with octoport adaptor
box mounted on it. The figure on the right is a side view. The adaptor boxes
and gate valves are permanently mounted on the vacuum vessel. The adaptor
boxes are mounted with antennas, beams, etc., as needed and then pumped
down. Each of the vacuum vessels and magnets are mounted on carts. The
carts are on rails, which were useful for alignment purposes. Once the system
was aligned the carts were locked in place using steel angle. A single purple
magnet OD is 60.25 in. and ID = 48 in., thickness = 3.25 in. The magnets
are cooled with 200 psi chilled (55-60 ◦F) water drived through manifolds
using an 80 HP pump. A second pump provides 60 psi chilled water to cool
the shrouds inside the vacuum chambers, power supplies, etc. The flow of
circulated water is roughly 750 gal/min.
The vacuum is sustained by four 2200 l/s magnetically
levitated turbo-molecular pumps backed by 50 CFM mechanical pumps. Initial pump down is accomplished using a 150
CFM Surgavac pump, which is valved off when the other
pumps take over. The pumps are installed on the two larger
end chambers (2200 l/s, Maglev). These chambers have ports,
which are useful for electrical power feedthroughs to the
cathode, and mounting gauges and windows for viewing and
image recording. One end chamber houses the BaO cathode
structure, and the second has a large gate valve for viewing or
the installation of an ion beam.
The vacuum is monitored with residual gas analyzers
(RGAs) and the partial pressure of impurities recorded in every
data set. The system base pressure is 5 × 10−7 Torr.
B. Plasma sources
FIG. 4. Cryopump shown on the right attached to a small pumpdown port.
The probe is in a KF-40 nipple. When the pressure in the pumpdown port is
less than 3 × 10−6 Torr, the pumpdown valve is closed and the chamber valve
is opened. The pumpdown station can be let up to air after valving it off. A
residual gas analyzer (RGA) is part of the pumpdown system. It is used to He
leak test all the vacuum seals and identify impurities.
The LAPD has two plasma sources. One is a BaO
cathode capable of making a 60 cm diameter background
plasma. The densities attainable from the BaO cathode are
1010 ≤ n ≤ 2 × 1012 cm−3. The upper density is determined
by the emissivity of the cathode. An emission current of 2
A/cm2 corresponds to a total discharge current of 5.6 kA,
easily attainable in this device. Discharge voltages between
the anode and cathode can range from 40 to 70 V. The BaO
⃗ forces present
cathode was designed to withstand the J⃗ × B
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-4
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 6. (a) CAD model of the BaO cathode and the inside of the vacuum
vessel. The circular ports are 32 cm apart center-to-center. The cathode is
60 cm in diameter. Several of the magnets (yellow) and part of the vessel
wall are shown cut away to view the interior of the chamber; (b) side view of
a hydrogen plasma (discharge current, ID = 5 kA, discharge voltage, VD = 75
V). A microwave horn used to launch waves at the upper hybrid layer and a
probe are also visible; (c) end-on view of a neon plasma. Circular ports and
a set of rectangular ports are visible.
given background magnetic field (up to 2 kG) and emission
current (up to 10 kA). The current emitted from the cathode is
of no concern as it is magnetic field-aligned; however, the eight
current feeds (45◦ apart), as well as the radial current flowing
in the cathode substrate, are at right angles to the background
field. The emitted electron current is collected on an anode
50 cm away. The molybdenum anode (50% transparent) is
also secured against the magnetic forces. The cathode design
and coating methodology is described in a previous RSI
publication10 as well as the transistor switch and capacitor
bank that powers it.11 Figure 6 is a computer generated cutaway model of the BaO cathode inside the LAPD device.
Figures 6(b) and 6(c) are photographs of hydrogen and neon
plasmas, respectively.
Two years ago, a second cathode was installed to provide
a hot, dense inner plasma if needed. The second plasma source
was fabricated from lanthanum hexaboride and is similar to
one described in a previous RSI paper.12 The cathode is square
and made of four tiles held together in a tongue and groove
arrangement. It is surrounded by heat shields; it sits within a
three-sided oven. The LaB6 cathode must be heated to 1850 ◦C
for maximum emission, which requires about 50 kW of DC
power. The second cathode had to be designed to allow it to be
removed from the machine when not in use. It also had to be
movable off machine center while hot so that newly introduced
probes could be aligned. To install this cathode, an additional
vacuum chamber had to be added to the train of 10 chambers
described above. It was installed before the end chamber on the
side opposite the BaO source. A magnet close to the cathode
had to be fabricated and placed around the chamber housing
the LaB6. As the inner diameter of the vacuum chamber is 1 m,
the heat flux to the wall is considerable. Water-cooled copper
shrouds had to be placed along the walls for 1.5 m in front
and back of the cathode. The “holding” box is 36 in. long and
FIG. 7. (a) A CAD model of the high-emissivity LaB6 cathode and the
“holding tank” can be moved in and out of during machine operation. When
not in use, the cathode is kept in the tank with a small current (50 A) in
the heater element. This keeps the cathode and heater at a temperature of
about 100 ◦C and prevents them from absorbing impurities. For clarity, the
cooling shrouds on the inside walls are not shown. (b) is a drawing showing
the placement of the chamber and holding tank at the end of the LAPD.
has a large rectangular gate valve and its own vacuum pumps.
This allows for removal of the cathode if repairs are necessary,
while the machine is running. The box also has water-cooled
walls so the cathode can be heated to 900 ◦C to allow it and its
assembly to outgas before placing it into the vacuum system.
Figure 7 shows a model of the second cathode as well as a
drawing of the section of the device containing it. Figure 8
contains photographs of the holding box for the cathode and
the 1.5 in. diameter ball screw drives to move it in and out.
Although the cathode and high current feedthroughs are heavy
(200 lbs), the system is carefully balanced and can be moved
in and out with a simple drill motor with mm precision.
Figure 8 is two views of the chamber and connections for
the second plasma source shown in two views. The manifold
supplies water to cooling shrouds which line inside of the
cooling box, the transfer section (colored yellow in Figure 7),
and the vacuum vessel in which the cathode is placed. The end
chamber and the vacuum chamber to the left of the holding
tank in Figure 7 have cooling shrouds in them as well.
The LaB6 chamber has its own turbo-molecular and
roughing pump. It has its own RGA to monitor outgassing
when it is heated. The heater power is about 50 kW (500 A
at 100 V DC). If there is a problem with the cathode, the gate
valve separating the holding tank from the rest of the system
can be closed and the LaB6 system brought up to air. After the
repairs, it can be re-introduced to the vacuum system.
The power system for this cathode mirrors that for the
main BaO cathode. The heater is brought up under computer
control, typically at 2 A/min. The pressure in the holding tank
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-5
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 8. The chambers constructed to house and deploy the movable LaB6 cathode. The cathode is moved in/out of the main vacuum vessel using two screw
drives, with bearings within the black cylinders, are mounted on the vertical Al plate. The two high current (400-500 A) conductors for the heater are attached
to insulated feedthroughs. The cylinders with 3/4 in. Cu wire insulated with ceramic move through double O-ring sliding seals. The gap between the O-rings is
attached to flexible tubes, which are pumped to 50 mTorr. (All sliding seals on the LAPD work in this manner.) A third smaller feedthrough is attached to the
pulsed discharge power supply.
as well as impurities is monitored, and if the pressure rises to
3 × 10−6 Torr, the heater current stops rising or is decreased.
When the cathode has been under vacuum with a small current
through it, this takes about 4 h. If the system has been in air,
or components replaced it can take 2 days. The anode may
be placed as close as 32 cm from the cathode for a standard
DC discharge, or as far away as 12 m if a current carrying
plasma or flux ropes are desired. The LaB6 cathode has its own
discharge capacitor bank (4 F), which is charged by a 100 A,
500 V power supply. A transistor bank much like that used for
the BaO cathode11 is used to switch the discharge voltage. The
second plasma source may be triggered before, during or after
the BaO plasma: they are completely independent. The LaB6
cathode is not sensitive to ion bombardment and discharge
voltages can be as high as 200 V. Higher discharge voltage
results in higher density as well as hotter electrons. Electron
temperatures of 12 eV are attainable. The electrical ground of
the two pulse circuits is isolated from one another and can be
configured to include a relative bias from one another if there
is an experimental requirement.
C. Probe drives
No matter how well a plasma physics device is constructed, it is useless without comprehensive diagnostics.
The LAPD was constructed so that probes can be used to
measure nearly every quantity of interest in a basic plasma
physics experiment. If the plasma was ten times denser
(n > 3 × 1014 cm3) or 10 times hotter (100 eV) probes would
vaporize. The diagnostics problem then becomes similar to
that in a fusion device where laser and microwave scattering,
spectroscopy, charge exchange, and other methods are the
rule. These are expensive, sometimes difficult to interpret and
cannot make measurements at tens of thousands of locations.
Drive mechanisms are required to move probes throughout
the plasma volume. The LAPD has a number of 2D probe
drives and 3D drive has just been completed. An example of
a drive that can position a probe in a plane transverse to the
background magnetic field is shown in Figure 9.
The inherent probe motion is in arcs at any given radial
position. It is far easier to analyze the data if they are acquired
on a Cartesian plane. Any pair of (x,y) positions can be reached
by a combination of motions of r,θ. The user specifies a grid
of points, which are requests for probe motion. Such a grid is
shown in the upper right of Figure 9. The positioning accuracy
is 0.5 mm, which determines the smallest motion. The number
of points on a plane 40 cm on a side with this resolution
is 6400, the largest possible. Moving a probe to this many
locations would make data run impossibly long as there is a
20% overhead on the time of a run due to probe motion. A
typical transverse data plane has 900-2000 positions. The time
it takes is governed by the plasma repetition rate (1 Hz) and
the number of data acquisitions (DAQs, one per plasma pulse)
at each location. Detailed transverse planes can take 12 h or
more. These are usually overnight runs.
A second probe drive capable motion in an orthogonal x,z
plane is shown in Figure 10 along with a photograph of the
x,y drive discussed above.
The stepping motor systems are controlled by one of
the computers that comprise the DAQ. The power supply
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-6
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 9. A schematic drawing of the probe in the center and extreme positions (θ = ±31.5◦). At the end of the probe shaft is a box with SMA vacuum feedthrough
connectors for the probe electronics. The angular motion feedthrough makes the probe motion possible. The system was designed so the probe can get to the
edge of the plasma at any radius. Upper left: Grid of 900 x-y positions the probe will move to during a run. This is generated by the data acquisition system.
Note the views in the inset of this figure are from the opposite vantage point of the drawing.
for the motors sends pulses, which step the motors by small
increments ensuring that the spatial resolution of the data is
limited by the size of the probes. The motors also provide a
holding torque, also with the use of current pulses. Often the
experiments on linear waves involve highly amplified signals,
which are easily contaminated by pickup from these pulses.
The DAQ turns the motors off during data acquisition. The
probes are aligned with a transit positioned to look through the
machine end window at the position x = y = 0, the machine
central axis. The transit uses pre-positioned markers to ensure
it is correctly centered. The alignment is good to 0.5 mm.
When probes have to be aligned when the LaB6 cathode is
in use, the cathode is moved to the side, even if it is at
1800 ◦C, using the method previously described. After the
probe alignment, the cathode is repositioned. Once the probe
is positioned at the center of the machine, the DAQ can move it
to positions on the grid. The probe drive can also be controlled
with a wireless controller, which makes it convenient to move
the probe from anywhere in the lab, or watch its motion
through a window. The probes have ¼ or ¾ in. shafts and as
they enter the machine they can droop due to the shaft’s weight.
This was carefully calibrated outside the machine using a grid
and the motion software accounts for the probe droop.
D. Data acquisition
Two, independent, computerized systems manage normal
operation of the upgraded LAPD. The first is a housekeeping
system that monitors the health and operational parameters
of the device itself. The second is a DAQ system, which
oversees automated experimental parameter variation, probe
motion, signal digitization, and data storage. Both systems
are custom programmed using the National Instruments, LabVIEW software development platform. Both of these systems
are described in more detail below and are schematically
displayed in Figure 11.
1. Housekeeping system
This system is based on a dedicated computer (the
housekeeper) that is connected to a wide variety of sensors,
digitizers, and control systems. The primary purpose of this
system is to protect the LAPD and infrastructure from damage.
Secondary tasks include control of numerous operational
parameters, and reporting the instantaneous state of a key
subset of all such parameters to other lab computers, mobile
devices, and the main DAQ system upon request. For security
reasons, the housekeeper and its Ethernet-controlled devices
communicate using a private (non-routed) subnet.
2. Data acquisition system
This system comprises a central data acquisition computer, independent-Ethernet-controlled devices, and more
advanced systems that require ancillary, dedicated computers.
The central computer is responsible for producing stored data
files for user-defined sequences of programmed tasks. Each
automated sequence is referred to as a “datarun” and is defined
by the devices used; one (or more) configurations for each
device; and, the order in which devices areaccessed and data
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-7
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 10. Photograph of three probes drives used on the LAPD device. Shown are 2 x-y drives. One is pictured in a horizontal position (along the x axis) and
the second in an extreme up position (θ = +31.5◦). The extreme positions are shown in the drawing in Figure 9. An x-z probe drive is also pictured (z is along
⃗ = B0z k̂). The probes require a stepper motor for each axis. The motors have shaft encoders which guarantee that their absolute
the background magnetic field B
position during the run is known after they are aligned. The probes move in and out of the system through a double “O” ring seal. The space between the “O”
rings is pumped with a mechanical pump attached to all the probes through a manifold.
stored. Dataruns are designed to be highly flexiblefor the
acquisition of data without regard for restricted paradigms
ofdata storage. This places the flexibility on the experimental,
ratherthan the analysis end—as required by a dynamic
experimentalenvironment. This flexibility is facilitated by
using a customizedscripting language. The central computer
maintains a relationaldatabase that organizes device configurations for each datarun and stores the operational state of
the LAPD for any time requested in a datarun. Dataruns
can last for a single LAPD discharge (1 s) or even 24-h
and can record data with any combination of high spatial
resolution, automated parameter variation (wave frequencies,
bias voltages, laser-wavelengths, etc.) Data are stored using
a platform-independent, scalable, and self-describing format:
Hierarchical Data Format, version 5 (HDF5) developed by the
HDF Group (https://www.hdfgroup.org/).
E. Machine procedures
The LAPD runs continuously 24 h a day, every day at a
1 Hz repetition rate. If there is no air leak or other reasons
to open the machine, it will run for about 4 months. At this
time, the BaO cathode coating will nearly be sputtered off
from continuous ion bombardment. The machine is then let
up to air and the cathode recoated. This procedure is described
in an earlier publication.10 In short, the old cathode coating is
washed with water, then ethanol, and then scraped off and then
sanded and cleaned with ethyl alcohol. Worn out parts, such
as springs or small heat shields, are replaced. This takes about
1–2 days. The cathode is then re-coated and reinstalled. The
molybdenum wire mesh anode is acid washed at every opening
to have a fresh surface devoid of any deposits that might
affect its local resistivity. The machine is pumped out and the
BaO conversion procedure begins. The partial pressures of
all impurities are carefully monitored as the heater current is
raised. For the first two days of the conversion, the cathode is
kept above 150 ◦C and heating strips heat the machine walls
to about 100 ◦C. This removes most water from the system.
The chamber pressure and partial pressures are monitored
by the housekeeping computer, which controls heater power
supply during conversion. After the water is abated, the
cathode temperature is increased further, while maintaining
a maximum limit on the total pressure of 7 × 10−6 Torr. The
conversion process takes about 6 days.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-8
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 11. Schematic of the LAPD data acquisition and housekeeping systems.
The LaB6 cathode does not need to be converted. When a
new cathode and heating element is installed, they are slowly
brought to their final temperature with a second housekeeping
computer. The newly installed structure takes several days to
outgas. The two cathodes are not heated at the same time
as impurities from the LaB6 cathode can poison the BaO
coating. Once the LaB6 cathode is “clean,” it can be cooled
again and used as needed by reheating it. To prevent the LaB6
cathode from accumulating impurities while under vacuum
it is maintained at about 100 ◦C (requiring 300 W of heater
power).
The BaO cathode alone can make plasma with density 2-3 × 1012 cm3 with an electron temperature of 5 eV.
Figure 12(a) shows the discharge current for a He plasma
with a background magnetic field of 1 kG. The machine
runs well at fill pressures from 2 × 10−5 to 5 × 10−4 Torr.
Mass flow controllers (MFCs) allow several gases to mixed
in any ratio. Typical discharge currents are 2-8 kA. Also
shown is the density obtained with a microwave interferometer
7.48 m from the cathode. The timing starts (t = 0) when the
TABLE I. Range of plasma parameters in LAPD.
F. Plasma properties
With the possibility of using one or two sources, simultaneously or independently the range of available parameters in
the LAPD is quite broad. If the magnetic field is kept constant
along the device axis, it can range from 50 G to 2.0 kG. At
lower fields, the cathodes are kept in a background field of
300 G as their emission drops in lower fields. The column
is 18 m in length and 60 cm in diameter for the BaO source,
20 cm in diameter for LaB6. The range of parameters available
is given in Table I for a He plasma with fill pressure (corrected
for the He mass) of 1 × 10−4 Torr.
Table I is representative for several values of the density
and magnetic field. All these quantities have ranges. For
µ L V
example, the Lundquist number, S = 0 η0 A , depends upon
the density, magnetic field, and temperature. L0 is the length
of the device (L0 = 18 m), and the resistivity is taken to be
the classical Spitzer value. At 1 kG, S = 1.8 × 105 during the
LaB6 discharge of plasma and can be 100 times smaller in the
late afterglow. There are 2000 ion gyroradii across the BaO
plasma at 2 kG and only 54 when the field is 50 G.
Parameter
Max density
Max Te
Max Ti
λ P (Alfven)
VA
δ = ωcpe
δI =
c
ω pi
cS (ion sound)
Rci
D/Rci
fpi (ion plasma)
Lundquist
number
Reynolds
number
Plasma β
τie (equilibrium
time)
BaO source
only 60 cm dia
Lab6 source
only 20 cm dia
2.0 × 1012
3.0 × 1013
cm3
6 eV
0.5 eV
2.0 m
7.7 × 107 cm/s
3.7 mm
Conditions
cm3
12 eV
5.0
0.5 m
2 × 107 cm/s
0.1 mm
32 cm
8.3 cm
1.24 × 106 cm/s
1.4 mm
429
148 MHz
2.4 × 105
1.7 × 106 cm/s
4.6 mm
44
575 MHz
3.3 × 105
1.8 × 104
1.8 × 105
0.008
900 µs
0.24
120 µs
1.0 kG shear wave
1 kG
1 kG, He
1 kG, He
1 kG
u = VA, L = 18 m
B = 300 G
250 G
Max n,T
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-9
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
electron temperature in the core can be as high as 12 eV.
When the LaB6 cathode is in the machine, it serves as the
termination of the plasma column. When only the BaO source
is present, the plasma column may be terminated by a movable
grid (which can be electrically floated or biased). Without the
end grid, the field lines diverge and the plasma is terminated
on the wall of the end chamber.
The LAPD (and LaB6) plasmas are highly reproducible.
After changing the gas or magnetic field, the machine can take
15 min to settle down. After that every discharge current and
voltage is identical to within 2% to the previous one. This
can be maintained for weeks at a time. The experiment on
current sheet tearing, described below as well as many others
in which volumetric data were acquired, has gone on for over
one million “shots” and can take 2-3 weeks of continuous data
acquisition.
II. EXAMPLES OF RESEARCH
FIG. 12. (a) The density obtained with a 60 GHz microwave interferometer
at a distance dz = 7.48 m from the source. There are six such interferometers
along the machine axis. He plasma, B0z = 1 kG, VDischarge = 38 V, ID (max)
= 4.4 kA. (b)-(e) Density and temperature profiles with both the BaO
and LaB6 cathodes present and emitting. Hydrogen plasma, B0z = 500 G
P = 5 × 10−5 Torr. (b) Lineout of the density at y = 0. The higher density in the
center, as well as the higher electron temperature, is shown in insets d and e.
The BaO plasma is 60 cm in diameter, only the central region is shown here.
The LaB6 discharge current was ID = 1.88 kA, and voltage VD = 100 V. The
BaO current was ID = 5.34 kA and discharge voltage VD = 35 V Hydrogen
plasma, B = 0.5 kG.
discharge current is 1 kA. A master trigger is generated at
this time, which goes out to several dozen optical trigger
boxes around the device. There are two time periods during
which experiments can be executed. The plasma builds up and
fills the device in 4 ms and then flat tops. Experiments done
during the main discharge, which can be on for up to 15 ms,
are done at a higher electron temperature (2-6 eV) and rms
density fluctuations of order 2%. There are always some higher
energy electrons present (at the 1% level) which originate in
the discharge. Experiments can also be done in the plasma
afterglow, which starts when the discharge is terminated. The
electrons become colder as a function of time in the afterglow,
and the density fluctuations are well below 1%. The LAPD
usually runs at a 1 Hz operation. To get a reliable discharge
every shot, a 1 A background plasma is switched on 50 ms
after the plasma is terminated and left on until the next plasma
pulse starts. The density associated with the 1 A discharge is
approximately 5 × 109 cm3. The second LaB6 cathode may be
switched on at any time during the BaO discharge or afterglow.
Figures 12(d) and 12(e) show density and temperature profiles
when both cathodes were present in the system and emitting.
In this case, the background magnetic field was B0z = 250 G
during an experiment of Alfvén wave propagation in a high β
plasma. The highest density achieved to date from the LaB6
plasma is 4 × 1013 cm3. This plasma is fully ionized. The
The LAPD was designed to be versatile and make
it easy to switch from one experiment to another. The
number of available ports, high repetition rate, quiescent, and
reproducible plasmas allows for the design of very different
experiments which impact the fields of space plasma physics,
solar physics, fusion studies, and plasma astrophysics. This
is the reason that NSF/DOE made it the first basic plasma
physics research facility in the nation in the year 2000. It
continues to serve the plasma physics community to this day.
The following is a list of some of the topics that have been
explored on the LAPD:
(1) Remediation of mirror trapped energetic particles.
(2) Effect of electric fields and rotation on plasma confinement.
(3) Dynamics and formation of filamentary structures.
(4) Linear and nonlinear Alfvén wave physics.
(5) Alfvén wave propagation in multiion species plasmas.
(6) Whistler waves ducting, chirping, and generation by
electron beams.
(7) Interaction of laser generated plumes with magnetized
plasmas.
(8) Magnetic field line reconnection and the generation of
Quasi-Seperatrix Layers (QSLs).
(9) Production of collisionless Alfvénic shocks.
(10) Generation of Alfvén waves by particle beams.
(11) Flux rope studies.
(12) Tearing mode studies.
(13) Langmuir turbulence and electron phase space holes.
(14) Effect of turbulence on energetic ions.
(15) Electron heat transport in magnetized plasmas.
(16) Role of nonlinear Alfvén waves in MHD turbulence.
(17) Interaction of whistler and lower hybrid waves with
density striations.
(18) Plasma edge turbulence.
(19) Shear Alfvén wave propagation in non-uniform density
and magnetic fields (gap modes).
Publications on these and many others may be found on
the facility website, http://plasma.physics.ucla.edu/bapsf. The
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-10
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 13. Magnetic vectors in two transverse planes, which are 2.6 m apart at time δt = 762.1 µs after the current sheet is switched on. At this time, the current
density has not risen to its peak value of −11 A/cm2. The current density Jz calculated from the total magnetic field is negative (pointing upward in this figure)
as the plasma current in the sheet is carried by electrons. Current “field lines” are also shown as purple lines. Reproduced with permission from Gekelman et al.,
“Experimental study of the dynamics of a thin current sheet,” Phys. Scr. (to be published). Copyright 2016 IOP.
following are results from several of these experiments that
highlight the versatility of the LAPD.
A. Tearing of a current sheet
A longstanding issue in plasma physics is the stability
of sheets of current in magnetoplasmas. Tearing modes
in plasmas have been a great interest for both the fusion
community as they can lead to disruptions in tokamaks13
and to the plasma physics community as the process is of
importance in the earth’s magnetotail.14 Tearing mode theory
has been around since the 1970s.15 When a current sheet
tears into filaments, they can interact with one another if the
magnetic field of one is large enough to affect the next via
⃗ force. The currents are then interacting magnetic
the J⃗ × B
flux ropes. This is a fully three-dimensional process as
demonstrated in Particle in Cell (PIC) simulations.16 Tearing
has been experimentally observed before in a current sheet
with unmagnetized ions.17 Both theory and simulation indicate
that in order to tear, the sheets must have a large transverse
aspect ratio, at least 10.
In this experiment, a sheet was formed in the background
plasma (He, Te = 4 eV, Ti = 1 eV, B0z = 200 G, n = 1 × 1012
cm−3). The second, LaB6 cathode, was masked with a narrow
slot (1 cm thickness, 20 cm height). The cathode was biased
with respect to a second anode and produced a current sheet
8 m long. A three-axis B-dot probe was used to measure
the magnetic field as a function of time on 5 planes, which
were 64 cm apart. Data were acquired at 3888 locations on
each plane with 4 mm separation between data points and 10
“shots” at every location. Data were acquired at 12 288 time
steps, δt = 0.16 µs.
The current sheet tears very quickly, in approximately an
Alfvén transit time along the length of the sheet. The tearing of
the current sheet is apparent in Figure 13. Here, the transverse
magnetic vectors are shown in two planes 2.4 m apart and
at τ = 762.1 µs after the sheet is switched on (the current is
switched on at τ = 0). The transparent yellow surface is an
isosurface of current density (Jz = −0.3 A/cm2). Two large
islands have formed at the ends of the sheet and several small
islands in the center.
A current sheet can be very dynamic, and one of the
first things that come to mind is the role of magnetic
field line reconnection. This process can cause ion jetting,
plasma flows, electron heating, etc., when magnetic field
energy is converted into other forms.18 The electrons and
ions in the current sheet do get hotter and plasma flows
occur, but is this the result of reconnection alone? We first
explored reconnection in this experiment and later compared
it with other forms of energization such as resistive heating.
When reconnection occurs several things happen. There is
a change in the local magnetic topology, induced electric
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-11
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 14. A vector map of the total electric field at δz = 1.28 m from the stat of the current sheet and at a fixed time. The field pattern is vertical at δz = 0.064 m
and is tilted at this location. (b) is the time evolution of both components of the electric field at the location marked by a dot in (a).
fields, which come about from the changing magnetic field
appear and energy is released. The reconnection process
is accompanied by the generation of a QSL.19 QSLs were
identified in this experiment. At early times, during tearing,
the value of Q was low (Q ≈ 6) dictating a modest level of
reconnection. This was much smaller than in experiments
in which flux ropes carrying large currents interacted (there
Q > 200). To further analyze the data, it was necessary to mea⃗ = −∇Vp − ∂ A⃗ throughout the
sure the total electric field, E
∂t
volume.
This was accomplished by calculating the vector potential
from the magnetic field data and the electrostatic component
from the gradient of the 3D plasma potential. The plasma
potential was measured with an emissive probe capable of
working in high-density plasmas.20 The inductive electric field
is generated by magnetic field line reconnection. An example
of the electric field measured is shown in Figure 14. Fig. 14(a)
is a vector map of the total electric field on a plane 1.28 m
from the origin of the current sheet. At this position, the sheet
has tilted. Fig. 14(b) shows the two components of the electric
field at one location on the transverse plane. The inductive
electric field is always much smaller than the space charge
field. The center of the current sheet is negatively charged
with respect to the plasma around it.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-12
Gekelman et al.
The reconnection rate, Ξ, is the integral
of the inductive

⃗ ⃗l, where d⃗l which
electric field along field lines, Ξ = − Ed
is the incremental distance along the local magnetic field
was evaluated. While the current shear was tearing it was
the largest, of order 1 V, in the region of highest Q. This
voltage drop is a measure of the reconnection rate and waves
driven by this can cause electron heating21 and anomalous
scattering.22 The measured reconnection rate does not agree
with the Sweet Parker estimation by two orders of magnitude.
This is not surprising as the reconnection is fully three
dimensional in this experiment. Reconnection is said to cause
particle energization, flows, heating, and, of course, changes
in magnetic field topology. This experiment incorporates a
multiplicity of processes that reflect what often occurs in
nature. Heating is not due to reconnection alone. There are
current driven instabilities and waves present. The flux ropes
generated by the tearing mode interact with one another, while
the entire current sheet rotates and oscillates at low frequency
as evident in Fig. 14(b). Waves at higher frequencies have
been detected and will be explored. It is not possible at this
time to separate the contribution from reconnection or plasma
instabilities that have nothing to do with reconnection to the
enhanced plasma parallel resistivity. A detailed publication on
this experiment has been accepted and will soon be published
in Physica Scripta.23
B. Transport experiments with a temperature ring
Recently, the LaB6 cathode has been used to create a ringshaped elevated temperature filament in a cold background
plasma for transport studies. The cathode is masked leaving
only a ring of LaB6 exposed and is discharged at low voltages
(<20 V) with respect to a far away anode. The low energy
electron beam thermalizes within a few meters. The ring
shaped source thus produces a hollow cylindrical heated region
embedded in a cold, background plasma. Intermittent pressure
profile collapses were observed in the experiment. The profile
collapses were associated with drift Alfvén waves, which
Rev. Sci. Instrum. 87, 025105 (2016)
were driven unstable by the large pressure gradients. The
drift waves enhance the cross-field transport, moving heat,
and density radially outwards, thereby collapsing the profile.
Once the profile is collapsed, the plasma enters a quiescent
period, with no drift waves, during which the profile gradually
steepens up due to the continued heating of the plasma by
the ring-shaped source. Once the pressure profile gradient
reaches a threshold steepness, the next collapse occurs.
These intermittent collapses were identified as avalanches in
magnetized plasmas.24 The 2D dynamics of an avalanche event
were measured in great detail. Figure 15 shows the radial
pressure profile during an avalanche event.
Panel (a) of Fig. 15 shows the profile before the
collapse happens with no fluctuations at the edge. Panel (b)
demonstrates the onset of drift wave activity and panel (c)
shows the final stage of the collapse when heat and density
has been moved considerably outwards in the radial direction.
C. Observation of “chirped whistler waves”
A 10 cm diameter energetic electron beam has been
constructed for use on the LAPD. The electron beam source is
based on a smaller LaB6 cathode.25 The source, with a series of
grids, is separated by ceramic insulating spacers on the front
side. The 10 cm LaB6 disk is biased negative with respect to
the machine wall (0 < Vbias < 5 kV). Emission currents can
be as high as 10 A and can be varied by changing the heating
power for the LaB6 disk. The energetic electron beam has
been used to study interactions between energetic electrons
and whistler waves of relevance to magnetospheric physics.26
Whistler waves and their interaction with energetic electrons
are thought to be a main driver of the dynamic variability of
the Earth’s radiation belts.
Using the energetic electron beam, the pitch angle
scattering by whistler waves through resonant processes
was verified in the laboratory.26 It was shown that whistler
waves, launched by an antenna, deflect energetic electrons
at the predicted resonant energy through the relationship
FIG. 15. Ion saturation current measurements show the radial and azimuthal structures of the heated region during the avalanche event. (a) Before onset of
fluctuations, (b) early stages of the profile collapse, and (c) late stages. Reprinted with permission Van Compernolle et al., Phys. Rev. E 91, 031102(R) (2015).
Copyright 2015 American Physical Society.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-13
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
mechanism. The energetic electron beam experiments at
LAPD now allow us to test under controlled conditions the
leading theories on non-linear whistler wave excitation. A
manuscript on linear excitation of whistlers by the energetic
electron beam is currently in preparation.
D. Fast ion studies using an intense ion source
FIG. 16. (a) Beam current and voltage, (b) magnetic field fluctuations, (c)
spectrogram of magnetic field fluctuations showing a clear discrete rising
tone, and (d) zoom in on time trace during part of the rising tone. Reprinted
with permission Van Compernolle et al., Phys. Rev. Lett. 114, 245002 (2015).
Copyright 2015 American Physical Society.
ω − kPvbeamP = Ωe . More recently, the excitation of both linear
and non-linear whistler waves by an energetic electron beam
is being studied. For the first time in the laboratory, discrete
frequency chirping whistler waves were excited.27 Figure 16
shows an example of a discrete rising tone. The frequency
sweep rate df/dt is on the order of 7 MHz/µs. During the
frequency chirp, the beam current and beam voltage are
constant. A zoom-in on the time trace, shown in panel (d),
clearly shows the increase in frequency during a portion of the
rising tone.
Frequency chirping whistler waves have been seen in
the Earth’s magnetosphere for decades, known as chorus
waves, but open questions remain as to the precise excitation
An intense ion beam source (20 keV, 10 A) has been built
for conducting a variety of laboratory plasma experiments
relevant to interaction of fast-ions (speed ∼Alfvén-speed)
with magnetized plasmas. These experiments can impact
topics related to controlled thermonuclear fusion, laboratory,
and space plasmas. The ion source is mounted to the side
opposite to the BaO cathode on the LAPD and it provides
flexibility for beam injection at a variety of pitch angles
into the plasma (see Fig. 17). Moreover, the beam energy
can be varied to access sub-Alfvénic and super-Alfvénic
regimes of the beam propagation. The LAPD provides an
ideal platform for performing these experiments due to its
diagnostic capabilities, three-dimensional measurement of
plasma parameters, and large size of the plasma in which
waves can be destabilized by a nearly collisionless ion-beam
(mean-free-path for Coulomb collisions of fast ions is much
greater than the length of the ambient plasma).
The ion source28 consists of three subsystems: (1) a
plasma source, (2) an ion-beam extractor, and (3) a cylindrical
vacuum chamber (95 cm long, 71 cm diameter). The source
for the beam is based on a compact LaB6 cathode58 (nmax
∼ 1013 cm−3, Te ≈ 7 eV, Bmax = 0.1 T) in a chamber which
is 21 cm diameter, 25 cm long to produce a quiescent ion
beam. The beam extractor consists of three multi-aperture
molybdenum grids (8 cm × 8 cm size, 57% transparency).
The spacing between grids was optimized to minimize the
beam divergence (0.8◦–1.5◦). The vacuum chamber of the ion
source has multiple vacuum-baffles and three turbo molecular
pumps (2500 l/s combined pumping speed) for minimizing the
neutral gas fraction in the ion beam, which in turn reduces the
neutral gas load and loss of fast-ions due to charge exchange
FIG. 17. Photograph displaying the ion source attached to the LAPD. A flexible bellows connects the LAPD to the ion source (see the large end-flange of the
LAPD on the left side). The ion source is mounted on wheels and can be rotated in a horizontal plane around a pivot located under the bellows. This facilitates
beam-injection at a variety of angles into the LAPD. The compact LaB6 plasma source (used for the ion beam generation) is on the right side in this photograph.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-14
Gekelman et al.
with neutrals. To date, spontaneous excitation of lower hybrid,
ion cyclotron, drift, and shear Alfvén waves by the ion beam
has been observed.29
Future experiments utilizing this ion source on the LAPD
are expected to further explore beam-driven Alfvén waves in
a multiple-mirror configuration, fast-wave generation through
anomalous Doppler-shifted cyclotron resonance with the ion
beam, and Cherenkov radiation of waves by the ion beam.
E. Destruction of magnetic-mirror-trapped hot
electron ring by shear Alfvén waves
The magnets in the LAPD device are divided into ten
sections along the axial direction and are controlled by ten
independent programmable power supplies (each capable of
delivering 0.5 MW), making possible a variety of magnetic
field configurations. One of the simplest is a magnetic mirror
within the plasma column. This makes the study of mirror
trapped electrons possible provided that a high-energy electron
ring can be created in the mirror.
The interest in this problem is the protection of satellites
and the current world-wide communication, internet, and
banking system. In 1962, an exo-atmospheric nuclear test
(code-named Starfish Prime) produced an intense energetic
charged particle flux,30 which was trapped by the Earth’s
magnetic field for years.31 One third of the entire satellite fleet
on the Low Earth Orbit (LEO) at the time was destroyed as a
result.32 In addition, the earth’s radiation belts are constantly
fed by geomagnetic storms and coronal mass ejections directed
toward the Earth. They are often awash with dangerously
energetic electrons and protons, posing severe threats to the
widely spreading satellite technology.33 There is a great deal of
current interest in concepts that can lead to artificial mitigation
of the trapped highly energetic charged particle flux.34 The
topic has come to be known as “radiation belt remediation.”35
A recent series of experiments36 conducted on the LaPD
have clearly demonstrated, for the first time in a controlled
environment, that a shear Alfvén wave can effectively de-trap
energetic electrons confined by a magnetic mirror field.
The experiment was performed in the quiescent LaPD
afterglow plasma. A symmetric magnetic mirror field was
Rev. Sci. Instrum. 87, 025105 (2016)
established near the center of the chamber. An X-mode
(Emicrowave||y) high power microwave pulse was introduced
radially into the center of the mirror section. A trapped
energetic electron population (>100 keV) was generated
by the 2nd harmonic electron cyclotron resonance heating
(ECRH) of the microwave pulse and formed a hot electron ring
due to the grad-B and curvature drift. Shear Alfvén waves with
arbitrary polarization of choice were launched externally by
a Rotating Magnetic Field (RMF) antenna37 (δB/B0 ≈ 0.1%)
from outside the mirror trap. The experimental setup is shown
in Fig. 18.
Hard X-rays (100 keV–3 MeV) were produced when hot
electrons were lost from the magnetic mirror trap and struck
the chamber wall or other metallic objections, as shown in
Fig. 19. An intense X-ray burst was generated when the
hot electron ring was irradiated by a shear Alfvén wave.
Fig. 19(a) shows an overlay of 19 traces, each measured
with a 100-cycle Alfvén wave (0.87 ms duration) launched at
different starting times. The observation of X-ray bursts after
30 ms, when the ECRH was shut off, clearly demonstrated
that a population of fast electrons persisted in the mirror trap,
and was de-trapped by application of the SAW.38 Further Xray tomographic measurements showed that the de-trapped
electrons were lost from both the radial and axial directions
of the magnetic mirror. A close-up look at the x-ray signal,
such as the case shown in Fig. 19(b), revealed that the electron
loss rate was modulated at the Alfvén wave frequency. The
modulation continued remarkably for at least another 50 wave
periods after the removal of the Alfvén wave. Such modulation
is a strong indication that the distribution function of the nearly
collisionless hot electrons is deeply modified by the shear
Alfvén wave and hints that the role of the Alfvén wave plays
in the de-trapping process is catalytic in nature.
F. Turbulence, transport, and flows in LAPD
The large diameter of the LAPD plasma provides a
distinct boundary between the center of the plasma, where
the density is constant for about 60 cm and the plasma edge
or gradient region. This allows for studies of the plasma edge
not possible in a narrow plasma column. Understanding the
FIG. 18. (Left) A schematic of the experiment studying the destruction of a mirror-trapped hot electron ring by shear Alfvén waves. (Right) A schematic plot
⃗ 0z at the center of the mirror, with measured Alfvén wave B
⃗ ⊥ vectors over-plotted. A population of fast electrons
of the experiment on a plane transverse to B
was generated by pulsed ECRH at 2.45 GHz and trapped by the mirror field. They formed a hot electron ring in the mirror due to grad-B and curvature drift.
Reprinted with permission from Wang et al., Phys. Rev. Lett. 108, 105002 (2012). Copyright 2012 American Physical Society.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-15
Gekelman et al.
FIG. 19. (Top) Overlay of 19 X-ray flux traces, each measured with a
100-cycle Alfvén wave launched at different time delays marked by an arrow
on the top. The ECRH was on from t = 0 to 30 ms, but only after about 20 ms
were there sufficient high-energy electrons to produce a measurable X-ray
flux. A population of fast electrons persisted after the termination of the
ECRH and could be de-trapped by application of the SAW to produce X-ray
bursts. (Bottom) X-ray flux and By of the Alfvén wave, measured just after
the Alfvén wave driver was shut off. The electron loss is clearly modulated
at the Alfvén wave frequency. Note that the time scale is different from that
of the top panel. X ray tomographic experiments39 revealed that electrons
were scattered out of the loss cone and “lit up” the remote LAPD anode.
Reprinted with permission from Wang et al., Phys. Rev. Lett. 108, 105002
(2012). Copyright 2012 American Physical Society.
radial transport of heat and particles at the edge is key towards
the path to fusion. In many ways, the LAPD edge plasma is
like the edge of tokamak plasmas and because probes may be
used in the LAPD it is a natural place to investigate turbulent
transport.
Suppression of turbulent transport by sheared flow has
been documented in a range of experiments and simulations.40
However, we still lack a complete, quantitatively correct
theoretical model of transport suppression by sheared flow.
This theoretical understanding is essential in the development
of a predictive capability for turbulent transport, a capability
that is critical in ensuring the success of future experiments
such as ITER. Experiments performed on LAPD have
documented in detail the response of turbulence and turbulent
transport to externally controlled flow and flow shear.
Azimuthal flow is driven in LAPD through biasing either
the vacuum chamber wall or an annular limiter relative to
the plasma source cathode. Cross-field currents are driven
(carried by ions due to Pedersen conductivity), leading to
⃗ torque and azimuthal rotation. In the case of biasing
J⃗ × B
the vacuum chamber wall, H-mode-like behavior is observed,
with suppression of turbulent particle transport and steepening
of the edge density profile.41 Transport is reduced from Bohmlike levels to classical if the wall bias is above a threshold value
(a factor of 100 reductions in particle diffusion coefficient39).
A more detailed examination of the impact of flow shear on
turbulent transport was enabled through the introduction of
an annular limiter, which brings a biasable surface closer to
the plasma edge. Biasing the limiter relative to the cathode
provides the ability to vary the edge flow and flow shear
continuously. As the LAPD plasma spontaneously rotates in
the ion diamagnetic direction and biasing tends to drive flow
in the electron diamagnetic direction, zero shear and zero
Rev. Sci. Instrum. 87, 025105 (2016)
FIG. 20. (a) Gradient scale length versus shearing rate. (b) Particle flux
normalized to no-shear flux as a function of normalized shearing rate. Filled
symbols represent points with flow in the ion diamagnetic direction. Inset:
Measured turbulent particle flux versus gradient scale length. Reprinted
with permission from Schaffner et al., Phys. Rev. Lett. 109 135002 (2012).
Copyright 2012 American Physical Society.
flow states are accessible as well as flow reversal. Figure 20
shows the measured density gradient scale length and turbulent
particle flux as the flow shear is varied continuously in the edge
of LAPD. The density gradient steepens (decreases) as shear
is increased, indicating a reduction in cross-field transport.
Consistent with this, the measured turbulent particle flux drops
monotonically with increasing shear.42 The shearing rate on
the axis is normalized to the turbulent autocorrelation time, as
measured with zero flow shear; this is taken as a proxy for the
eddy decorrelation (or “turn-over”) time. Substantial changes
in both and particle flux occur for normalized shearing rates
of order unity.
LAPD experimental data have been compared to a number
of analytical theoretical models of shear suppression.43 The
ability to continuously vary the edge flow shear in LAPD
allowed the collection of data for both the weak (γs τac < 1) and
strong (γs τac > 1) shearing regimes. The data were fit to two
functional forms motivated by theoretical models developed
for these two regimes. While these functional forms do fit
the data reasonably well, the fit coefficients obtained are
not a good match to theoretical predictions, suggesting that
new models may be needed to explain LAPD data. Future
work will focus on comparison of LAPD data to numerical
simulation (e.g., using the GENE code) and to more recently
developed analytical models of shear suppression of turbulent
transport.44
G. LAPD operation using multiple ion species
The LAPD can also be used for studies involving multiple
ion species. This is possible because the gas feed system
supports five gasses fed through the system using five separate
MFCs. The MFCs, once set, provide a constant gas flow
regardless of changes in the ambient air temperature. Using
one or more gasses, the pressure (partial pressure) will not vary
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-16
Gekelman et al.
Rev. Sci. Instrum. 87, 025105 (2016)
for months. The provision for multiple ion species allows the
study of basic plasma physics experiments in a wide variety
of environments, including tokamak D-T edge plasmas, the
solar wind, and planetary magnetospheres. Presently, any
combination of H2, He, Ne, or Ar can be produced. Because
of differences in ionization potentials and cross sections, the
ratios between partial gas fill pressures (here measured with
commercial, quadrupole residual gas analyzers) do not reflect
the concentration ratios of the plasma ions. Wave propagation
characteristics often provide the most accurate measures of
the ion concentration ratios.45–47 In the LAPD, a technique
was developed using shear Alfvén wave propagation in a
two-ion-species plasma48 to determine the ion concentration
ratio.
In order to understand the diagnostic technique, it is first
necessary to review the basic propagation characteristics of
the shear Alfvén wave in a plasma with two ion species.
The parallel wavenumber for these
waves
) in the limit of
(
ck⊥
large perpendicular wavenumber ωpi > 1 for these waves48
√
k ∥ ∝ ε ⊥.
When two ion species are present, ε ⊥ can be written as
ε⊥ ∝
ω2 − ω2ii
,
ω2 − ω12 ω2 − ω22
(1)
where ωii the well-known ion-ion hybrid frequency.49
ε ⊥ is positive (hence propagating waves) in two frequency bands: ω < ωci1 and ωii < ω < ωci2, “1” and “2”
label the heavier and lighter ion species, respectively. These
propagation bands are measured using antenna-launched
magnetic perturbations in LAPD and presented in Fig. 21(a).
The measurements are performed in a plasma with equal
concentrations of helium and hydrogen ions. The waves are
excited in the linear regime with a current impulse to the
antenna, thus launching all frequencies simultaneously. The
upper and lower propagation bands as well as the propagation
gap are all present. Since ωii determines the lower boundary of
the upper propagation band, a direct measurement of this cutoff
frequency also provides the diagnostic for the concentration
ratio of the two ion species. The ratio of the two ion densities
can be expressed using a dimensionless reformulation of the
definition of the ion-ion hybrid frequency,
)
( m ) (( ω )2
ii
1
−
1
m2
ω ci1
ρ12 =  ( ) 2 (
(2)
) ,
m1
ω ii 2
−
m
ω
2
ci1
with “2” labeling the light ion.
As an example, Eq. (2) is used to compute the helium/hydrogen concentration ratio for four cases, which are
presented in Figure 21(b). The figure displays the power
spectra obtained at r = 5 cm (r = 0 is the center of the machine)
and the same axial location (z = +528 cm) as in Figure 21(a).
With the frequency axis scaled to the ion cyclotron frequency
ω
of the heavier species, ωci−He
, the value of the scaled ionω ii
ion hybrid frequency, ω , is found by noting the lower
ci1
bound of the upper propagation band in each power spectrum.
These frequencies are identified by the corresponding colored
triangles on the horizontal axis and are used in Eq. (2) to obtain
FIG. 21. (a) Antenna-launched wave magnetic field amplitude as a function
of plasma radius and frequency. The upper and lower propagation bands are
both evident and separated by a clear propagation gap. Frequencies are scaled
to the helium cyclotron frequency; waves are detected at z = +528 cm for
a H+-He+ plasma of equal ion densities ρ 12 = nnHe
= 1. (b) Fourier power
H
spectra of wave magnetic fields for four values of the helium/hydrogen ion
concentration ratio. These values are determined by identifying the scaled
ion-ion cutoff frequency (marked with a color-coded) triangle for each spectrum and the use of Eq. (2). Another propagation gap becomes evident near
ω ii = 2ω ci1, when the ion-ion hybrid frequency lies below this harmonic. In
all cases, the electron density is held fixed at 1 × 1012 cm−3 by monitoring the
real-time output of a microwave interferometer. Reprinted with permission
from S. T. Vincena, W. A. Farmer, J. E. Maggs, and G. J. Morales, Phys.
Plasmas, 20, 01211 (2013). Copyright 2013, AIP Publishing LLC.
the values shown in the numerical inset in this figure. It should
be noted that, in the present experiment, when ωii lies below
the second harmonic of helium a narrow propagation gap is
also observed in the vicinity of this harmonic. This additional
gap is an effect associated with ion temperature, which results
in Bernstien mode-like features.
H. Basic studies of electron heat transport
The great length, and relatively large radial dimension of
the LAPD, allows for the three-dimensional study of electron
heat transport in temperature filaments that have a “free”
boundary at one end, that is, the filament ends in the plasma
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-17
Gekelman et al.
FIG. 22. (Top panel) The experimentally observed temperature profile at
early times (black curve) is accurately modeled by classical transport (red
dashed curve). (Bottom panel) The temperature filament produced by classical transport before the growth of drift waves is azimuthally symmetric with
the steepest temperature gradient located at about r = 2.5 mm. Reprinted with
permission from Pace et al., Phys. Plasmas 15, 122304 (2008). Copyright
2008, AIP Publishing LLC.59
and not on a material object. The physical situation essentially
models to a great degree the behavior of a temperature filament
surrounded by an infinite, colder, and magnetized plasma. In
the experiments, a temperature filament is created by injecting
a low-voltage electron beam by negative biasing of a single
crystal (3 mm diameter) LaB6 source. The beam, injected
at voltages below ionization potential, acts as an ideal heat
source as it deposits its energy into the background, afterglow
plasma.
Initially, temperature transport in the filament proceeds
at rates predicted by the classical theory of heat transport
based upon Coulomb collisions. The temperature profile
produced by classical transport is shown in Figure 22. The
temperature gradients associated with this profile are unstable
Rev. Sci. Instrum. 87, 025105 (2016)
to the growth of drift-Alfvén waves. When the drift Alfvén
waves excited by the temperature gradient in the edges of
the filament grow to sufficient amplitude, the system enters a
state of “anomalous” transport.50 The transition from classical
to anomalous transport occurs when the power spectrum of
temperature fluctuations in the gradient region of the filament
changes from a line spectrum to a broadband spectrum. The
broadband spectrum is exponential in nature and has been
experimentally demonstrated to arise from structures in the
time signals of temperature fluctuations, namely, Lorentzian
shaped pulses.51 These features are a signature of chaotic
dynamics.52
One technique for distinguishing between chaotic and stochastic dynamics is a permutation-entropy analysis. Structure
in time signals can be detected by determining the probability
of permutations in the amplitude ordering of consecutive
signal values. The probability of these amplitude orderings is
called the Bandt-Pompe (BP) probability.53 The BP probability
is used to compute the Jensen-Shannon statistical complexity,
C, and the normalized Shannon entropy, H, to analyze data in
the CH plane.54 Such an analysis was applied to temperature
fluctuation time signals obtained in the temperature filament
experiment.55 In addition, the dynamics in the experiment are
explored by using a chaotic E × B advection model,56 whose
results are compared to experimental data. Figure 23 shows
the results of the permutation entropy analysis. Data from the
experiment and the chaotic advection model are gathered from
16 azimuthal locations (shown as black “dots” in Figures 23(b)
and 23(c)) and a permutation entropy analysis is performed.
The CH plane locations of the model (red triangles) and data
(blue crosses) fall in the chaotic region of the CH plane
(lilac shaded region in Fig. 23(a)). The permutation entropy
analysis indicates that the plasma dynamics producing the
temperature fluctuations observed in the filament experiment
are chaotic. This finding is important for understanding
electron heat transport in plasmas and for developing transport
models.
FIG. 23. Data from a chaotic advection model (b) and the temperature filament experiment (c) are obtained at the 16 locations indicated by the “black dots.”
The locations in the CH plane (a) of both data sets are in the chaotic region (purple shading). Reprinted with permission from J. E. Maggs and G. J. Morales,
Plasma Phys. Controlled Fusion, 55, 085015-085022 (2013). Copyright 2013, IOP Science.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-18
Gekelman et al.
I. Summary and conclusions
The large plasma device has been significantly upgraded
in 1999-2000 and improvements in the machine and diagnostics continue to this day. The machine was designed and
constructed by the first seven authors of this paper. The LAPD
has two plasma sources, which produce an 18 m long plasma
column with a quiescent and highly ionized, magnetized
plasma column. The machine is capable of 1 Hz operation
and runs round the clock for up to 4 months at a time. It
was designed to be versatile and, therefore, can accommodate
a variety of very different experiments, such as the recent
observations of Alfvénic shocks.57 This machine is now the
heart of a Basic Plasma Science user Facility (BaPSF) funded
by NSF/DOE. The success of BaPSF for the past 15 years can
be attributed, in part, to the flexibility of the device which was
built in to the design. This makes possible the wide variety
of very different studies done on the device. This variety is
reflected in the publications, which may be downloaded from
http://plasma.physics.ucla.edu/bapsf.
ACKNOWLEDGMENTS
This work was supported by research grants from the
Department of Energy and the National Science Foundation
for support of the Basic Plasma Science Facility. The authors
would like to thank Dr. Chuck Roberson for support as well as
the Office of Naval Research to supplement the construction
funds provided by the MRI award. We would also like to
thank Prof. R. Peccei, who was Vice Chancellor of Research
at UCLA when the machine was constructed, for securing
University matching funds and support. We also wish to thank
Mr. E. Deitl for his expert vacuum welding during machine
construction.
1W.
Gekelman, H. Pfister, Z. Lucky, J. Bamber, D. Leneman, and J. Maggs,
“Design, construction, and properties of the large plasma research device—
The LAPD at UCLA,” Rev. Sci. Instrum. 62, 2875–2883 (1991).
2W. Gekelman, D. Leneman, and J. Maggs, “Experimental observation of
Alfvén wave cones,” Phys. Plasmas 1, 3775–3783 (1994).
3D. Leneman, W. Gekelman, and J. Maggs, “Laboratory observations of
kinetic effects on Alfvén waves,” Phys. Rev. Lett. 82, 2673 (1999).
4S. Vincena, W. Gekelman, and J. Maggs, “Shear Alfvén waves in a magnetic beach and the roles of electron and ion damping,” Phys. Plasmas 8,
3884–3896 (2001).
5C. Mitchell, S. Vincena, J. Maggs, and W. Gekelman, “Laboratory
observation of Alfvén resonance,” Geophys. Res. Lett. 28, 923–926,
doi:10.1029/2000GL012165 (2001).
6W. Gekelman, S. Vincena, N. Palmer, P. Pribyl, D. Leneman, C. Mitchell,
and J. Maggs, “Experimental measurements of the propagation of large
amplitude shear Alfvén waves,” Plasma Phys. Controlled Fusion 42,
B15–B26 (2000).
7J. F. Bamber, J. E. Maggs, and W. Gekelman, “Whistler wave interaction
with a density striation: A laboratory investigation of an auroral process,” J.
Geophys. Res. 100, 23795, doi:10.1029/95JA01852 (1995).
8S. Rosenberg and W. Gekelman, “Electric field measurements of directly
converted lower hybrid waves at a density striation,” Geophys. Res. Lett.
25, 865, doi:10.1029/98GL00382 (1998).
9D. Leneman and W. Gekelman, “A novel angular motion feedthrough,” Rev.
Sci. Instrum. 72, 3473 (2001).
10D. Leneman, W. Gekelman, and J. Maggs, “The plasma source of the large
plasma device at UCLA,” Rev. Sci. Instrum. 77, 015108 (2006).
11P. Pribyl and W. Gekelman, “24 kA solid state switch for plasma discharge
experiments,” Rev. Sci. Instrum. 75, 669 (2004).
Rev. Sci. Instrum. 87, 025105 (2016)
12C.
M. Cooper, W. Gekelman, P. Pribyl, and Z. Lucky, “A new large
area lanthanum hexaboride plasma source,” Rev. Sci. Instrum. 81, 083503
(2010).
13P. H. Rutherford, “Nonlinear growth of the tearing mode,” Phys. Fluids 16,
1903 (1973).
14J. F. Drake et al., “Formation of electron holes and particle energization
during magnetic reconnection,” Science 299, 873 (2003).
15H. Furth, J. Killeen, and M. Rosenbluth, “Finite-resistivity instabilities of a
sheet pinch,” Phys. Fluids 6, 459 (1963).
16W. Daughton et al., “Role of electron physics in the development of turbulent magnetic reconnection in collisionless plasmas,” Nat. Phys. 7, 539
(2011).
17W. Gekelman and H. Pfister, “Experimental observations of the tearing of
an electron current sheet,” Phys. Fluids 31, 2017 (1988).
18R. L. Stenzel, W. Gekelman, and N. Wild, “Magnetic field line reconnection
experiments. 4. Resistivity, heating and energy flow,” J. Geophys. Res.
87(A1), 111, doi:10.1029/JA087iA01p00111 (1982).
19P. Démoulin, “Extending the concept of seperatricies to QSLs for magnetic
reconnection,” Adv. Space Res. 37, 1269 (2006).
20M. Martin, J. Bonde, W. Gekelman, and P. Pribyl, “A resistively heated CeB
6
emissive probe,” Rev. Sci. Instrum. 86, 053507 (2015).
21W. Gekelman, R. L. Stenzel, and N. Wild, “Magnetic field line reconnection
experiments,” Phys. Scripta T2/2, 277 (1982).
22W. Gekelman, R. L. Stenzel, and N. Wild, “Magnetic field line reconnection
experiment. 3. Ion acceleration, flows and anomalous scattering,” J. Geophys. Res. 87, 101–110, doi:10.1029/JA087iA01p00101 (1982).
23W. Gekelman, T. DeHaas, B. Van Compernolle, W. Daughton, P. Pribyl, S.
Vincena, and D. Hong, “Experimental study of the dynamics of a thin current
sheet,” Phys. Scr. (to be published).
24B. Van Compernolle, G. J. Morales, J. E. Maggs, and R. D. Sydora,
“Laboratory study of avalanches in magnetized plasmas,” Phys. Rev. E 91,
031102(R) (2015).
25B. Van Compernolle, W. Gekelman, P. Pribyl, and C. M. Cooper,
“Wave and transport studies utilizing dense plasma filaments generated with a lanthanum hexaboride cathode,” Phys. Plasmas 18, 123501
(2011).
26B. Van Compernolle, J. Bortnik, P. Pribyl, W. Gekelman, M. Nakamoto, X.
Tao, and R. M. Thorne, “Detection of resonant electron pitch angle scattering
by whistler waves in the large plasma device,” Phys. Rev. Lett. 112, 145006
(2014).
27B. Van Compernolle, X. An, J. Bortnik, R. M. Thorne, P. Pribyl, and W.
Gekelman, “Excitation of chirping whistler waves in a laboratory plasma,”
Phys. Rev. Lett. 114, 245002 (2015).
28S. K. P. Tripathi, P. Pribyl, and W. Gekelman, “Development of a radiofrequency ion beam source for fast-ion studies on the large plasma device,”
Rev. Sci. Instrum. 82, 093501 (2011).
29S. K. P. Tripathi, B. Van Compernolle, W. Gekelman, P. Pribyl, and W.
Heidbrink, “Excitation of shear Alfvén waves by a spiraling ion beam in
a large magnetoplasma,” Phys. Rev. E 91, 013109 (2015).
30W. N. Hess, “The artificial radiation belt made on July 9, 1962,” J. Geophys.
Res. 68, 667, doi:10.1029/jz068i003p00667 (1963).
31H. S. Ghielmetti, N. Becerra, A. M. Godel, H. Heredia, and G. Roederer,
“Enhancement of the x-ray intensity at balloon altitudes in the South American Anomaly,” Phys. Rev. Lett. 12, 388 (1964).
32D. G. Dupont, “Nuclear explosions in orbit,” Sci. Am. 290, 100 (2004).
33D. N. Baker, S. G. Kanekal, X. Li, S. P. Monk, J. Goldstein, and J. L. Burch,
“An extreme distortion of the Van Allen belt arising from the Halloween
solar storm in 2003,” Nature 432(7019), 878 (2004).
34C. J. Rodger, M. A. Clilverd, T. Ulich, P. T. Verronen, E. Turunen, and N.
R. Thomson, “The atmospheric implications of radiation belt remediation,”
Ann. Geophys. 24, 2025 (2006); X. Shao, K. Papadopoulos, and A. Sharma,
“Control of the energetic proton flux in the inner radiation belt by artificial
means,” J. Geophys. Res. 114, A07214, doi:10.1029/2009ja014066 (2009).
35U. S. Inan, T. F. Bell, J. Bortnik, and J. M. Albert, “Controlled precipitation of radiation belt electrons,” J. Geophys. Res. 108(A5), 1186,
doi:10.1029/2002JA009580 (2003).
36Y. Wang, W. Gekelman, P. Pribyl, and K. Papadopoulos, “Scattering of
mirror-trapped fast electrons by a shear Alfvén wave,” Phys. Rev. Lett. 108,
105002 (2012).
37A. Gigliotti, W. Gekelman, P. Pribyl, S. Vincena, A. Karavaev, X. Shao, A.
S. Sharma, and D. Papadopoulos, “Generation of polarized shear Alfvén
waves by a rotating magnetic field source,” Phys. Plasmas 16(9), 092106
(2009); A. V. Karavaev, N. A. Gumerov, K. Papadopoulos, X. Shao, A. S.
Sharma, W. Gekelman, Y. Wang, B. Van Compernolle, P. Pribyl, and S.
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40
025105-19
Gekelman et al.
Vincena, “Generation of shear Alfvén waves by a rotating magnetic field
source,” ibid. 18, 032113 (2011).
38Y. Wang, W. Gekelman, P. Pribyl, and K. Papadopoulos, “Enhanced loss of
mirror trapped fast electrons by a shear Alfvén wave,” Phys. Plasmas 21,
055705 (2014).
39Y. Wang, W. Gekelman, and P. Pribyl, “Hard x ray tomographic studies of
the destruction of an energetic electron ring,” Rev. Sci. Instrum. 84, 053503
(2013).
40P. W. Terry, “Suppression of turbulence and transport by sheared flow,” Rev.
Mod. Phys. 72, 109 (2000).
41J. E. Maggs, T. A. Carter, and R. J. Taylor, “Transition from Bohm to
classical diffusion due to edge rotation of a cylindrical plasma,” Phys.
Plasmas 14, 052507 (2007); T. A. Carter and J. E. Maggs, “Modifications of
turbulence and turbulent transport associated with a bias-induced confinement transition in the large plasma device,” ibid. 16, 012304 (2009).
42D. A. Schaffner, T. A. Carter, G. D. Rossi, D. S. Guice, J. E. Maggs,
S. Vincena, and B. Friedman, “Modification of turbulent transport with
continuous variation of flow shear in the large plasma device,” Phys. Rev.
Lett. 109, 135002 (2012).
43D. A. Schaffner, T. A. Carter, G. D. Rossi, D. S. Guice, J. E. Maggs, S.
Vincena, and B. Friedman, “Turbulence and transport suppression scaling
with flow shear on the large plasma device,” Phys. Plasmas 20, 055907
(2013).
44G. M. Staebler, R. E. Waltz, J. Candy, and J. E. Kinsey, “New paradigm for
suppression of gyrokinetic turbulence by velocity shear,” Phys. Rev. Lett.
110, 055003 (2013).
45M. Ono, M. Porkolab, and R. P. H. Chang, “Parametric decay into ion
cyclotron waves and drift waves in a multi-ion species plasma,” Phys. Fluids
23, 1656 (1980).
46A. M. Hala and N. Hershkowitz, “Ion acoustic wave velocity measurement
of the concentration of two ion species in a multi-dipole plasma,” Rev. Sci.
Instrum. 72, 2279 (2001).
47G. W. Watson, W. W. Heidbrink, K. H. Burrell, and G. J. Kramer, “Plasma
species mix diagnostic using ion–ion hybrid layer reflectometry,” Plasma
Phys. Controlled Fusion 46, 471 (2004).
48S. T. Vincena, W. A. Farmer, J. E. Maggs, and G. J. Morales, “Investigation
of an ion-ion hybrid Alfvén wave resonator,” Phys. Plasmas 20, 012111
(2013).
49S. J. Buchsbaum, “Resonance in a plasma with two ion species,” Phys. Fluids
3, 418 (1960).
Rev. Sci. Instrum. 87, 025105 (2016)
50A.
T. Burke, J. E. Maggs, and G. J. Morales, “Experimental study of
fluctuations excited by a narrow temperature filament in a magnetized
plasma,” Phys. Plasmas 7, 1397 (2000); “Experimental study of classical
heat transport in a magnetized plasma,” ibid. 7, 544 (2000).
51D. C. Pace, M. Shi, J. E. Maggs, G. J. Morales, and T. A. Carter, “Spontaneous thermal waves in a magnetized plasma,” Phys. Rev. Lett. 101, 035003
(2008).
52J. E. Maggs and G. J. Morales, “Generality of deterministic chaos, exponential spectra, and Lorentzian pulses in magnetically confined plasmas,”
Phys. Rev. Lett. 107, 185003 (2011); “Exponential power spectra, deterministic chaos and Lorentzian pulses in plasma edge dynamics,” Plasma Phys.
Controlled Fusion 54, 124041 (2012).
53C. Bandt and B. Pompe, “Permutation entropy: A natural complexity measure for time series,” Phys. Rev. Lett. 88, 174102 (2002).
54O. A. Rosso, H. A. Larrondo, M. T. Martin, A. Plastino, and M. A. Fuentes,
“Distinguishing noise from chaos,” Phys. Rev. Lett. 99, 154102 (2007).
55J. E. Maggs and G. J. Morales, “Permutation entropy analysis of temperature
fluctuations from a basic electron heat transport experiment,” Plasma Phys.
Controlled Fusion 55, 085015 (2013).
56M. Shi, D. C. Pace, G. J. Morales, J. E. Maggs, and T. A. Carter, “Structures
generated in a temperature filament due to drift-wave convection,” Phys.
Plasmas 16, 062306 (2009).
57C. Niemann, W. Gekelman, C. G. Constantin, E. T. Everson, D. B.
Schaeffer, A. S. Bondarenko, S. E. Clark, D. Winske, S. Vincena, B.
Van Compernolle, and P. Pribyl, “Observation of collisionless shocks in
a large current free laboratory plasma,” Geophys. Res. Lett. 41, 7413,
doi:10.1002/2014GL061820 (2014); D. B. Schaeffer, E. T. Everson, A. S.
Bondarenko, S. E. Clark, C. G. Constantin, S. Vincena, B. Van Compernolle,
S. K. P. Tripathi, D. Winske, W. Gekelman, and C. Niemann, “Laser-driven,
magnetized quasi-perpendicular collisionless shocks on the large plasma
device,” Phys. Plasmas 21, 056312 (2014); C. Niemann, C. G. Constantin,
D. Schaeffer, A. Tauschwitz, T. Weiland, Z. Lucky, W. Gekelman, E. T. Everson, and D. Winske, “A high energy Nd;glass laser facility for collisionless
laboratory astrophysics,” J. Instrum. 7, P03010 (2012).
58The original source used a rf plasma. The plasma was dense enough
for beam extraction but the RF interfered with data collection and was
abandoned.
59D. C. Pace, M. Shi, J. E. Maggs, G. J. Morales, and T. A. Carter, “Exponential
frequency spectrum and Lorentzian pulses in magnetized plasmas,” Phys.
Plasmas 15, 122304 (2008).
Reuse of AIP Publishing content is subject to the terms at: https://publishing.aip.org/authors/rights-and-permissions. IP: 128.97.43.17 On: Tue, 09 Feb 2016 18:16:40