The Many Faces of Shear Alfve N Waves

PHYSICS OF PLASMAS 18, 055501 (2011)
The many faces of shear Alfvén wavesa)

W. Gekelman,b) S. Vincena, B. Van Compernolle, G. J. Morales, J. E. Maggs,
P. Pribyl, and T. A. Carter
Department of Physics and Astronomy, University of California, Los Angeles, California 90095, USA
(Received 21 December 2010; accepted 1 February 2011; published online 1 June 2011)
One of the fundamental waves in magnetized plasmas is the shear Alfvén wave. This wave is
responsible for rearranging current systems and, in fact all low frequency currents in magnetized
plasmas are shear waves. It has become apparent that Alfvén waves are important in a wide variety
of physical environments. Shear waves of various forms have been a topic of experimental research
for more than fifteen years in the large plasma device (LAPD) at UCLA. The waves were first
studied in both the kinetic and inertial regimes when excited by fluctuating currents with transverse
dimension on the order of the collisionless skin depth. Theory and experiment on wave propagation
in these regimes is presented, and the morphology of the wave is illustrated to be dependent on the
generation mechanism. Three-dimensional currents associated with the waves have been mapped.
The ion motion, which closes the current across the magnetic field, has been studied using laser
induced fluorescence. The wave propagation in inhomogeneous magnetic fields and density
gradients is presented as well as effects of collisions and reflections from boundaries. Reflections
may result in Alfvénic field line resonances and in the right conditions maser action. The waves
occur spontaneously on temperature and density gradients as hybrids with drift waves. These have
been seen to affect cross-field heat and plasma transport. Although the waves are easily launched
with antennas, they may also be generated by secondary processes, such as Cherenkov radiation.
This is the case when intense shear Alfvén waves in a background magnetoplasma are produced by
an exploding laser-produced plasma. Time varying magnetic flux ropes can be considered to be low
frequency shear waves. Studies of the interaction of multiple ropes and the link between magnetic
field line reconnection and rope dynamics are revealed. This manuscript gives us an overview of the
major results from these experiments and provides a modern prospective for the earlier studies of
shear Alfvén waves. VC 2011 American Institute of Physics. [doi:10.1063/1.3592210]
I. INTRODUCTION AND THEORY as stellar winds, extragalactic jets, and quasar clouds is given
by Jatenco-Pereira.9
Alfvén waves are of fundamental importance in space
In 1942, Hannes Alfvén proposed the existence of
plasmas, astrophysical plasmas, and in earthbound efforts to
hydrodynamic waves.10 This was met with skepticism since
achieve thermonuclear fusion in confined magnetized plas-
it was believed that any conducting fluid would immediately
mas. None other than Alfvén conjectured Alfvén waves to
short circuit wave electric fields. As plasma lore has it, phys-
exist in the solar corona in 1945,1 and many others have dis-
icists suddenly reversed their thinking when Enrico Fermi
cussed them in this context since. Recently their motion in
agreed with Alfvén at a seminar at the University of Chi-
the corona was observed in detail by imaging the velocity of
cago.11 As with many discoveries, the importance of these
a FeXIII line.2 They were first seen in the solar wind in
waves in space, solar, astrophysical, and finally magnetic
19663 and are abundant there.4 The solar wind fluctuations
fusion plasmas took decades to become apparent. Alfvén
seem to be filamentary with large correlation lengths along
waves were not observed immediately after their existence
the magnetic field5 (>1.5 106 km) and shorter ones across
was predicted. The original fluid description gives a disper-
the field (105 km). The earth’s aurora is replete with Alfvén
sion relation for the shear (or torsional) wave of frequency x
waves, and they have also been observed to be filamentary,
and field-aligned wavenumber kjj
the structures, sometimes as small as the electron skin
depth.6 Alfvén waves in the aurora have been seen in associ- x B
¼ VA ¼ pffiffiffiffiffiffiffiffiffiffiffiffi ; (1)
ation with skin depth size density striations.7 One may also kjj 4pnM
consider time varying magnetic flux ropes as low frequency
shear waves. The importance of these waves in the auroral where B is the magnetic field strength, n the plasma density
region cannot be understated. Measurements taken by the and M the ion mass. The waves exist only below the ion cy-
Polar satellite indicate that the downward flux of Alfvén clotron frequency and in simple MHD descriptions propagate
waves has enough energy to power the aurora.8 A review of along the magnetic field. The other branch of these waves is
the import of Alfvén waves in astrophysical processes such the fast or compressional wave. These propagate at the
Alfvén speed along and across the magnetic field and exist
a)
Paper AR1 1, Bull. Am. Phys. Soc. 55, 20 (2010). above and below the ion cyclotron frequency. To launch
b)
Invited speaker. shear Alfvén waves in an experimental device several factors
1070-664X/2011/18(5)/055501/26/$30.00 18, 055501-1 C 2011 American Institute of Physics

V
Author complimentary copy. Redistribution subject to AIP license or copyright, see http://php.aip.org/php/copyright.jsp
055501-2 Gekelman et al. Phys. Plasmas 18, 055501 (2011)
must be taken into consideration. The low frequency of the x2pe x2pi
waves coupled to the requirement that ions must be magne- e? ¼ 1 þ : (4)
x2ce x2 x2ci
tized pose a requirement that the ion Larmor orbit be small
0
enough and the plasma be dense enough for the waves to fit Here Z ðnÞ is the derivative of the plasma dispersion func-
in the experimental device. If the plasma is very dense the tion,17 with n ¼ pffiffi2xk v and ve is the electron thermal velocity.
jj e
parallel wavelength is short, but unless the plasma is hot the
For cold ions and for a wave frequency not too close to
waves will be damped by collisions. High temperature trans-
the ion gyro-frequency xci , the dispersion relation takes the
lates to larger gyroradii and therefore larger plasma sources
form
and raises survival issues for internal diagnostic probes. In
2
the 1940s laboratory plasmas suitable to study these waves 0 V x2
did not exist, and the first observation was made in mer- Z ðnÞ A2 1 2 n2 ¼ k? 2 2
d ; (5)
ve
2 xci
cury.12 One of the first experiments in which Alfvén waves

were glimpsed in an ionized gas was by Bostick and Lev- where d ¼ c xpe is the electron skin depth or electron iner-
ine.13 In this experiment standing waves, in the compres- tial length.
sional branch were observed in a pulsed toroidal plasma. This expression has two important limiting cases. For
Other early experiments14 involved pulsed collisional plas- n1 the wave is called the inertial shear Alfvén wave, and
mas and verified that the wave propagates at the predicted the dispersion relation (5) reduces to
velocity, but these lacked detailed measurements of the wave
electric or magnetic field. From the 1980s to the present, ba- 2 !
x
sic science groups in Australia, Japan, and in the United VA2 1
x2 xci
States have done experiments on Alfvén waves. In these ¼ : (6)
experiments some fundamental properties of the shear kjj2 2 d2 Þ
ð1 þ k?
Alfvén wave have been explored and their direct relevance
to the spacecraft measurements previously discussed has The inertial limit occurs when the Alfvén velocity is much
emerged. At the same time a great deal of research on Alfvén higher than the electron thermal speed; the wave is well out-
waves in thermonuclear plasmas occurred. The greater part side the electron distribution function. The other limiting
of the waves studied was bounded cavity modes launched by case is the opposite, n21, the electrons are hot, and the
antennas close to the chamber wall. More recently a great wave phase velocity is within the main body of the electron
deal of attention has been paid to toroidal Alfvén eigenmo- distribution function. This mode is referred to as the kinetic
des (TAE) modes in magnetically confined fusion plasmas.15 Alfvén wave. The real part of the dispersion relation is
These are shear waves with a dispersion relation affected by
x2 2 x2 2 2
the toroidal geometry of the background magnetic field of ¼ V A 1 þ k q
? s ; (7)
the Tokomak plasmas in which they are generated. TAE kjj2 x2ci
waves have been observed in several fusion experiments and
are worrisome if they adversely affect the plasma confine- where qs ¼ cs =xci is the ion sound radius, with cs the ion
ment where they to grow to large amplitudes. The dispersion sound speed.
relation16 for shear waves in a warm plasma with ion temper- In both of these cases the shear wave has a parallel electric
ature much less than the electron temperature, TI Te is field, which is important as it can lead to particle acceleration.
more complicated than Eq. (1). If the value of n is not in these two extreme limits, the wave is
The regime of transverse scale lengths small compared termed to be in the intermediate regime, and the full plasma
to the ion skin depth is of interest in problems related to elec- dispersion relation must be used. This intermediate regime is
tron acceleration, ion heating, and current modulation and significant in that the wave phase velocity is close to the elec-
has a fast or compressional mode that is evanescent. Thus its tron thermal velocity, and Landau damping can be strong.
contribution to the shear mode can be separated. This results These expressions for the dispersion relation in the vari-
in a highly accurate and compact dispersion relation ous regimes are only part of the story because they describe
a single wave number. Since any real signal originates at a
1
x pffiffiffiffiffi c2 k?
2 2 source, a spectrum of wave numbers must be considered to
kjj ¼ e? 1 2 ; (2) ascertain its spatial structure. Equation (1) is for a plane
c x ejj
wave propagating exactly parallel to the confinement mag-
where ejj and e? are the parallel and perpendicular (to the netic field. In this idealized case the wave’s magnetic and
confinement magnetic field) components of the dielectric electric fields are both at right angles to the direction of prop-
tensor, respectively. Here c refers to the speed of light and agation. The wave carries plasma current, but how it closes
k? is the transverse wavenumber. is not apparent from Eq. (1). What is more realistic is the
For collisionless, hot electrons consideration of waves associated with currents of finite
! dimensions, e.g., currents whose transverse dimension is
kD2 0 x comparable to the electron inertial length or ion sound gyro-
ejj ¼ 1 2 Z pffiffiffi ; (3)
2kjj 2kjj ve radius. If the radius of the current channel is r0, and the cur-
rent density is j0 at the channel center, the wave magnetic
while for a single species of cold ions, field is in the azimuthal direction and given by18
055501-3 Many faces of shear Alfvén waves Phys. Plasmas 18, 055501 (2011)
ð1
4pr0 j0 J1 ðkr0 Þ The behavior in the kinetic regime does not exhibit the
Bh ðr; z; xÞ ¼ J1 ðkrÞeikjj ðkÞz dk; (8) characteristics of the cone given in Eq. (8). Instead the wave
c 0 k
is confined to a maximum angle of propagation given by18
in which the confinement magnetic field is in the z direction. x 2 pffiffiffiffiffi
tan hmax ¼ be ; (10)
In this expression the relationship between k and k\ is medi- xci 332
ated by the dispersion relation which in practice may require where be is the electron beta. The magnetic field pattern
the inclusion of collisional effects. exhibits radial oscillations within hmax which evolves into a
For the case of the inertial wave19 the spatial structure 1/r behavior at large radial positions.
predicted by Eq. (8) is governed by the ratio p ¼ rd0 . For large Figure 1 shows the spatial pattern in an x-z plane from
values of p (narrow current channels) the theory predicts that Eq. (5), here the magnetic field is in the z direction and the
the parallel wave current lies within a cone, which emanates center of the current filament is located at x ¼ 0.
from the edge of the exciter disk. Further away from the
source the cone spreads at an angle h given by
II. EARLY ALFVÉN WAVE EXPERIMENTS
rffiffiffiffiffiffi " 2 #12 There were no laboratory plasma sources adequate to
r r0 me x x
tan h ¼ ¼ 1 : (9) perform Alfvén wave measurements at the time of their pre-
z M xci xci
diction. The first theory was for a conducting fluid in which
the particle’s gyroradii are assumed to be zero, and the con-
Outside of the cone the wave current is zero, and the wave ductivity assumed to be infinite. The initial experiments
magnetic field decays as 1/r. The wave current is confined were done in liquid metals. Lundquist20 used mercury, which
within the cone, which expands radially with distance is a liquid at room temperature. Waves on the surface of
along the magnetic field away from the current source. Far stirred magnetized column were observed on the surface
from the source the wave magnetic field in the center of with a mirror. Lehnert21 used sodium that is 34 times more
the cone vanishes. The magnetic field is in some sense conductive than mercury but has to be heated (Tmelt ¼ 98 C)
similar to that of a current carrying wire; however, here to the liquid state. The experimental apparatus is shown in
the “plasma wire” expands along its axis. Unlike the field Fig. 2. The sodium column was “twisted” in the azimuthal
of a wire the pattern exhibits a radial diffraction pattern far direction by the stirrer. At higher conductivity the magnetic
from the source. field is frozen into the fluid, and the shear (or torsional)
FIG. 1. (Color) Theoretical patterns of one component, By, of the Alfvén wave in the kinetic and inertial regimes. The waves propagate from left to right.
FIG. 2. From Lenhardt (1952). Liquid Sodium apparatus for studying

Alfvén waves. The key elements are a heater on the bottom (21) and a stirrer,
which, oscillated at 30 Hz (15) and a double probe that measured potential dif-
ference on the surface (20). The sodium was threaded by a uniform vertical FIG. 3. From Lenhardt. The cylinder is the Na column. When the liquid is
magnetic field of 0.1-1T. The height of the sodium column was 10 cm. moved in the azimuthal direction the magnetic field entering the bottom is
twisted sideways.
Alfvén wave emanates from the stirrer. This is illustrated in
the second figure from Lenhardt, Fig. 3. probes picked up 10 G oscillations, which exhibited delay times
Using a fluid theory, which included the fact that the in accord with the Alfvén propagation speed. The experiment
waves are bounded in the radial direction and collisional was highly collisional, and no wave structure was measured.
damping due to the finite conductivity of the medium, Len- In yet another early experiment by Sawyer et al.,23
hardt predicted the potential difference on the surface probes hydrodynamic waves were observed in a linear discharge of
and compared it to the experimental results. This is shown in 60 cm long and 15 cm in diameter. There was an axial mag-
Fig. 4. The conductivity of a highly ionized plasma is orders netic field of 500 G. The plasma was produced with a 200 kJ
of magnitude higher than that of liquid sodium, and as soon capacitor bank in relatively high (0.1–10 mm) pressure deu-
as plasmas became available the laboratory study of Alfvén terium. Magnetic loop probes 3 mm in diameter, electrostati-
waves switched media. The first measurement in laboratory cally shielded, and set in quartz tubes were used to measure
plasma was by Bostick and Levine in 1952.4 In this simple the azimuthal, radial, and axial component of the magnetic
experiment a gas in a toroidal tube was ionized with a pulse field. A self-generated helical wave was measured and
applied to a coil wrapped around it. There was a background detailed plots of current distribution were presented. The
toroidal magnetic field of 400 G. A floating Langmuir probe wave magnetic field was proportional to p1ffiffiq, and the wave
recorded oscillations of order 104 Hz. The oscillations were seemed to travel at the Alfvén speed. The resulting wave
interpreted as a standing MHD wave and satisfied the rela- was deduced to be a mix of hydrodynamic waves traveling
tionship f ¼ 2LpBffiffiffiffiffi
4pq
ffi, where L is the circumference of the torus in the axial and azimuthal directions. Although the wave
and q is the mass density. No waves were seen when the to- physics is quite complex, this is one of the first experiments
roidal magnetic field was zero. The diagnostics were not where detailed probe measurements were made.
good enough to distinguish between a shear and a compres- Wilcox et al. performed a better experiment, which
sional wave (it was probably the latter). was executed in a linear device.24 The plasma was produced
In another early experiment by Jephcott22 which was by a high-pressure (100 l, n0 ¼ 3.5 1015 cm3) discharge
similar to that of Bostick, an antenna wound around a torus between the ends of the device using a switch and a capacitor
was used to ionize a gas (Id ¼ 10 kA, td ¼ 200 ms). A back- bank. The wave was launched by applying RF from a ringing ca-
ground toroidal magnetic field of 3–14 kG was present. Two pacitor between the cathode of the plasma source, and the wall
external coils were used to produce an oscillating magnetic of the device after the plasma was formed. The wave phase ve-
field transverse to the background magnetic field. Magnetic locity was measured with a magnetic pickup probe as a function
into account in theoretical comparisons. If one raises the

temperature to circumvent these effects then internal probes,
which are the best means of getting detailed spatial information,
tend to evaporate. One might think that lowering the magnetic
field is the path to shorter wavelengths. There are two problems
with this. First of all shear waves exist below the ion cyclotron
frequency, which is directly proportional to B. When the wave
frequency is low collisional effects become worse. The ion gyro-
radius is inversely proportional to B and one must have a larger
diameter plasma column, ideally many ion gyroradii across.
Experiments which measured the wave dispersion,
where the plasma column was narrow enough to require Bes-
sel function solutions in the radial direction, were done by
Wilcox,27 Swanson et al.,28 and Müller.29 Starting in 1980,
Amagishi and his colleagues did experiments on Alfvén
waves at the TPH arcjet30 at Nagoya University in Japan for
about ten years. The arcjet produced a 2 m long, 15 cm diam-
eter (full width at half maximum ¼ 6 cm) helium plasma
(n ¼ 5 1014 cm3, Te ¼ Ti ¼ 4 eV, B ¼ 2.5 kG). At these
densities the Alfvén wavelength is of order 40 cm (f ¼ 390
kHz, fci ¼ 950 kHz), but the plasma was not completely ion-
ized (70%) and the collision rate vei was higher than the
wave frequency by a factor of ¼ 104. In another experiment
the compressional and shear waves were both observed using
FIG. 4. Theory and experimental predictions by Lenhardt (1952). The ordi- a Stix type coil31 having a spectrum of wave numbers32 with
nate is the potential difference due to the wave electric field and the abscissa
the inverse of the background magnetic field. The excellent agreement veri- the plasma selecting parallel wavelengths that could fit in the
fied the existence of the shear wave. machine. The waves were detected by a small (5 mm diame-
ter, 100 turn) magnetic probe. The wave modes were bounded
of the background magnetic field using the time delay between in the radial direction as the arc plasma diameter was about 6
the applied sinusoid and the received signal on a magnetic probe. cm. Other experiments33 by this group observed mode conver-
It agreed with the predicted Alfvén speed. Wave attenuation was sion of fast waves launched by a strap antenna near the wall
also measured, and this roughly agreed with theory. of the machine, to shear waves on the Alfvén resonance layer
As time elapsed, laboratory experiments on Alfvén (position on the radial density gradient where the axial phase
waves and detailed comparison with theory improved. velocity matched that of the shear wave). These and other
Another experiment by Jephcott and Stocker25 was carried experiments have been reviewed by Gekelman.34
out in a 2-m long, linear device in an argon plasma. The The theory outlined previously indicates that narrow
wave was launched by applying an oscillating waveform, current channels fluctuating below fci will excite cone-
generated in an LC circuit, to a ball electrode 50 cm from shaped wave field structures. In the experiments outlined
one end of the device. The time delay between two magnetic thus far the plasma column was too narrow and the plasma
pickup probes 50 cm apart measured the phase velocity of too collisional to observe them. One of the first clear experi-
the waves. Double Langmuir probes were used to calculate mental observations of these cone like structures was made
the density at the position of the magnetic probes, and the elec- by Borg et al.35 in the afterglow of a pulsed plasma where
tron temperature (n ¼ 1 1015 cm3, Te ¼ 2.3 eV). The ion the magnetic field and density was high (Baxis ¼ 0.5T, ne
temperature was estimated from a spectroscopic measurement ^ 1 1013cm3) and the electron temperature was Te^ 10 eV.
of Doppler broadening (Ti ¼ 0.2 eV). The conductivity was esti- A small rectangular antenna 2 cm high and 8 cm long was
mated from electric probe measurements. The plasma was col- aligned with the toroidal magnetic field and placed in the
lisional; the ion neutral collision frequency was much greater center of the plasma. The wave was detected by magnetic
than the wave frequency (vei ¼ 4.5 106 Hz, fwave ¼ 125 kHz). pickup probes located on the opposite side of the torus that is
The magnetic probes could be moved radially and the azi- 180 away from the exciter. The spatial measurement along
muthal component of the wave magnetic field, Bh(r) was meas- with theory of the predicted pattern is shown in Fig. 5. What
ured. The data were compared to predictions of a theory of is clear is that the wave magnetic field vanishes at the center
bounded Alfvén waves by Woods,26 which included the effect of the current channel. The experiment was in the inertial re-
of neutral collisions. The comparison was excellent. gime with VVtheA ’ 3, but since it was collisional the fine cone
Experiments continued for the next thirty years, which structure due to large k\ could not be seen.
for the most part took place in arc type plasmas that were
collisional. Plasmas of this sort allowed one to operate at
14 3 III. THE LARGE PLASMA DEVICE
high density (n 10 cm ) and therefore relatively shorter
wavelengths k / pffiffin . Higher density plasmas, which are
1
The original 10 m long large plasma device (LAPD) at
3
cold, become collisional ðvei / nT 2 Þ, and this must be taken UCLA (Ref. 36) has been in operation since 1990 and was
Possible magnetic field geometries are uniform, linearly

increasing in z, mirror fields, multiple mirrors, and cusps.
The plasma is produced by a DC discharge between a
Barium oxide coated cathode38 and mesh anode. The bulk of
the plasma carries no net current and is therefore quiescent.
The plasma parameters are: He, Ar, Ne, H, Xn, dn n 3%,
1010cm3 ne 2 1012cm3, 0.05 Te 10 eV, Ti
1 eV. The plasma is pulsed at 1 Hz with typical plasma dura-
tion of 15 ms. The machine can run continuously for approx-
imately four months before the cathode has to be cleaned
and recoated. There are 450 access ports serviced by portable
pump-down stations, which allow introduction of probes,
antennas, additional cathodes, ion beams, electron beams,
etc., while the machine is running. Probes go through spe-
cially designed ball valves39 that allow computer controlled
FIG. 5. Measurement by Borg et al. (Ref. 25) of the magnetic field of a
shear Alfvén wave launched by an antenna at r ¼ 5 cm. B ¼ 0.8 T, f ¼ 550 probe drives to accurately move them on transverse planes
kHz, in a toroidal plasma. The solid lines are theory with N the number of throughout the device.
times the wave circled the torus.
IV. ALFVÉN WAVES FROM NARROW CURRENT
FILAMENTS
upgraded to a 22 long meter long device in 2001. The present The LAPD plasma is much less collisional
LAPD plasma column is 60 cm in diameter and 18m long. (C ¼ vxcoll 1) than the plasma in the experiment by Borg
The axial magnetic field is produced by 90 solenoidal mag- et al. (C ¼ vxcoll 29). A fluctuating current filament in the
nets (Fig. 6) and may be varied over the range 300G B02 LAPD was used to generate shear waves and the Alfvén
2.5kG. Ten power supplies generate current for the mag- cone patterns were verified and compared to theory with col-
nets therefore the axial magnetic field profile is variable. lisions and Landau damping.40 Another early measurement
FIG. 6. (Color) Plasma production in the LAPD device. An oxide-coated cathode heated to 900 C is placed 55 cm from a molybdenum mesh anode. A tran-
sistor switch (Ref. 37) is used to apply a voltage between the cathode and anode, the current is supplied by a 4 F capacitor bank. A bank of 90 magnets supplies
the axial DC magnetic field, B0z. The upper right insert is a photograph of a He plasma through an end window located radially near the wall of the machine.
V. THE MORPHOLOGY OF SHEAR WAVES

LAUNCHED FROM VARIOUS ANTENNAS
A key element of the shear wave, often overlooked in
the MHD description, is the closure of the wave current. For
plane waves this issue is swept away by declaring that the
currents close at infinity. This is not the case with waves
associated with current filaments. Figure 8 shows the data of
Alfvén wave currents and was acquired in a plane with one
axis parallel to the background magnetic field.42 The figure
is highly compressed, dz ¼ 7m, dr ¼ 9 cm. The waves spread
across the background magnetic field by 0.3 or about 11 cm
over the 18-m LAPD column. Electrons, which move along
the magnetic field, carry the parallel wave current but as the
electrons are highly magnetized (rce 100lm), the cross-
field current is carried by ions via the ion polarization drift.
FIG. 7. (Color) The wave magnetic field of a shear Alfvén wave launched This has been verified by tracking the ion motion using
by a small mesh grid in a plane transverse to the background magnetic field laser-induced fluorescence.43 The ion distribution function
B0z. An image of the grid antenna to scale is shown, although the grid is on was measured with a fiber optics probe through which a
a plane dz ¼ 1.54 m (1 parallel wavelength) from the data plane shown. The
laser beam from a tunable dye laser was injected; a small
background magnetic field is B0z ¼ 1.1 kG, xxci ¼ 0:77, Helium, Te ¼ 10 eV,
n ¼ 2.0 1012 cm3. lens and second fiber were used to detect the LIF signal.
About 1 mm3 of the plasma was imaged, and the probe
moved to a series of locations in the same data plane that
of the structure of Alfvén waves due to a skin depth size cur- the wave fields were measured. Different components of the
rents measured the perpendicular wavenumber of the wave distribution function transverse to the field were obtained
magnetic field.41 Figure 7 shows the magnetic field of a shear by rotating the probe. The ion distribution function was
wave in a plane launched by applying a tone burst of oscillat- recorded as a function of time at each position. The drift ve-
ing potential to a circular mesh grid. A small DC potential is locity of the ions was obtained by measuring the displace-
! !
applied to the grid during the tone burst so that the oscillat- ment of the peak of f ð r ; v? ; tÞ. Figure 9 shows the magnetic
ing potential modulates the electron current to it at the shear field of the shear wave at two instants of time as well as the
wave frequency. The magnetic field is governed by Eq. (5). ion drift displayed as red arrows. The wave pattern is differ-
In Fig. 7 one can clearly see that Bh reverses in the radial ent from that shown in Fig. 7, because in this case the
direction, outward from the center of the current channel. waves were launched with a helical antenna. Here there are
From the data k\; 1.26 cm1, kjj ¼ 0.41 cm1, the wave is in two parallel current channels, which rotate about one
the kinetic regime. another in time. The non-random part of the ion motion is a
FIG. 8. (Color) The measured current

system of a shear Alfvén wave (in the ki-
netic regime, b mMe ¼ 6:8). The current
source has a 0.5 cm radius. The centers
of the current vortices are located 1.25 m
k
apart in z, which is 2jj . The pattern is
invariant with rotation about its center
along the x-axis. Note that the data is
acquired on a plane in the LAPD device,
the picture is highly compressed along z.
FIG. 9. (Color) Magnetic fields shown as streamlines and background Ar ion drift (red arrows). In (a) the ions drift in the E ~ B ~ direction and in (b) taken 76ls
later the ions are seen to close the current via the ion polarization drift. The experimental conditions: n ¼ 1.3 1012cm-3, Te ¼ 3.6 eV, Ti ¼ 0.8 eV, B0z ¼ 1200 G,
fwave ¼ 36 kHz. The experiment was in the kinetic regime.
combination of the E ~ B~ drift (vEB ¼ cE? 2 ) and the ion these cases is the cross-field group velocity, which changes
B0 ð1- Þ
dE?
c
polarization drift (vpol ¼ xci ð1- 2 Þ dt ) where x ¼ xxci . sign as in Eqs. (11) and (12)
The wave fields were on the order of 0.5 G in the center
of Fig. 9(a) and 0.1 G at the edges of the “O-points.” The vgr? xk? d2 12
¼ 1 þ ½k? d
2 ; ðVA > ve Þ (11)
wave fields are stronger in the center because the magnetic vgrjj VA
fields of the oppositely directed axial currents, jz, in the two 2 !32
channels add. Twice each cycle the ions drift from one current vgr? xk? q2s x 2
~ B ~ drift ¼ 1 þ ðk? qs Þ ; ðV A < v e Þ
channel to another, closing the current. Since the E vgrjj VA xci
is measured directly the wave electric field can be calculated
(12)
and was E\ ¼ 75.6 mv/cm, at a location dz ¼ 66 cm from the
exciter. This was compared to the theoretical value
Here d is the electron inertial length and qs the ion sound
E? VA ð1x 2 Þ
Bwave ¼ c 2
1 : with agreement to within 16%, the gyroradius. The change in phase velocity between the two
2 þk? q2s Þ2
ð1x
regimes was verified39 in measurements during the LAPD
largest uncertainty was in the measurement of k\. The parallel
when the electron temperature is high (kinetic regime) and in
wave electric field was not measured directly but can be esti-
ik k q2
the early afterglow plasma when the density has not changed
jj ? s
mated from Ejj ¼ ð1 2 Þ E? . In this experiment the parallel
x significantly but the electron temperature is lower by a factor
electric field was about 1% of the perpendicular electric field. of about 10 (inertial regime). The two regimes were also stud-
This is not large but can have important consequences in ied in the LAPD with a specialized antenna44 shown in Fig.
space where the parallel wavelengths are large and the plasma 10 designed to produce shear waves with a fixed k\ (Fig. 11).
parameters non-uniform along the direction of propagation. The morphology of the shear wave depends on how it is
Experiments have been done studying the waves in both launched. An antenna designed to launch a shear wave with
the inertial and kinetic regimes.42 One marked difference in arbitrary polarization was constructed using two perpendicular
FIG. 10. (Color) (left) Schematic dia-

gram of the Iowa University spatial
waveform antenna for launching shear
waves with fixed k\. The antenna has 48
elements that can be biased with individ-
ual waveforms for phase control in the x
direction as shown in the left image.
(right) Measured data for one compo-
nent, (By), of the wave is shown on the
right. The perpendicular wavelength for
this case is 2.5 cm. For each launched
value of k\ the parallel wavelength can
be measured and the dispersion relation
plotted (Ref. 45). This is shown in Fig.
11 for a wave in the inertial regime.
space plasma community motivated the initial studies. An

overview of the topic can be found in a report by Hase-
gawa.49 In the earth’s plasma these field line resonances are
observed as low frequency oscillations.50,51 FLRs have been
observed at the onset of auroral substorms,52 and there is
some suggestion that there may be a casual relationship
between the two.53 Spacecraft cannot map out the structure
of the waves, but the spectra derived from the time history of
magnetic measurements on spacecraft clearly reveal the dis-
crete modes.54 This is shown in Fig. 14 where data from the
THEMIS satellite was used55 to acquire magnetic field data
in magnetopause boundary layer.
The first observation in a laboratory plasma to identify
standing Alfvén waves56 was done in the previously men-
tioned arcjet. A peak in the amplitude of Bh was observed at a
position several centimeters off the device axis. The location
2 2
of the peak matched the radial position having kx2 V 2 ¼ 1 xx2 ,
jj A ci
where x is the applied angular frequency, and n ¼ n(r).
The LAPD device is well suited to study field line resonan-
ces. The first version of the LAPD device was 10 m long and a
spectrum of standing waves was introduced by applying a cur-
rent pulse to a field aligned antenna two electron skin depths in
diameter. The spectral content of the pulse extended to the cy-
clotron frequency. The experiment was repeated at ten different
ambient magnetic fields, and spectra such as that in Fig. 15
(left) were recorded.57 p The frequencies of the FLRs are pre-
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
dicted to be: fn ¼ VA = ð4L2 =n2 þ VA2 =fci2 Þ with the second
term in the denominator arising from the ion cyclotron correc-
tion in Eqs. (6) or (7). The allowed modes have kn ¼ 2L n where
FIG. 11. (Color) Measured phase velocity and damping of a shear wave in L is the length of the plasma column. In this approximation the
the inertial regime where k\de is the relevant scale length. Here de ¼ 0.61cm, k\ terms are neglected. Figure 15 indicates that waves up to the
B0z ¼ 2.3 kG. The wave frequency was 380 kHz (xxci ¼ 0:43). The theoretical
dispersion relation, including damping, is drawn as a solid black line. The eighth harmonic were detected with no signal measured above
dash curves are theory bounds reflecting uncertainty in plasma parameters. the ion cyclotron frequency as expected. The Q of the resonant
system can be obtained from the spectrum, and it is 3.8 for the
lowest harmonic (n ¼ 2) and 10.2 for the highest (n ¼ 8).
multi-turn loops.46 The experimental setup is shown in Fig.
The perpendicular wavenumber of the n ¼ 1 mode was
12. The rotating magnetic field antenna (RMF) coils are
determined by a Bessel function decomposition of the spatial
driven by independent RF amplifiers capable of driving more
magnetic field Fig. 15 (left) after interpolating the data into
than 1 kA in each loop (for high power wave studies). The ki-
polar coordinates.58 The value of hk? i was determined to be
netic shear Alfvén wave was studied in the kinetic regime for
hk? i ¼ 0.55 6 .025cm1. The measurement was compared to
typical plasma parameters: (n ¼ 2.3 1012cm3, Te
a theoretical prediction of the wavelength by Streltsov59 with
¼ 6 6 1eV, Ti ¼ 1 6 0.5 eV, B0z ¼ 1kG, He).
agreement within experimental uncertainties. When the
In this experiment the wave had a maximum amplitude of 4
initial wave is localized, it spreads across the magnetic field
G, dBBwave 0:4%. Figure 13 shows isosurfaces of the helical
0 due to its finite k\ and could not survive successive reflec-
wave currents derived from the measured magnetic field. In this
tions. The plasma parameters in the ionosphere vary consid-
case the shear wave has two side by side current channels, with
erably. If one follows a field line, close to the poles, the
the current as always closing due to the ion polarization drift.
magnetic field is higher and the plasma is colder and denser,
A Green’s function technique similar to the one
while the opposite is true near the equator. The waves there-
described in Sec. I was used to model the RMF antenna47
fore go from the inertial to the kinetic and back to the inertial
and a 3D MHD code48 was used to study the antenna proper-
regime along their trajectory. The perpendicular group veloc-
ties above the ion cyclotron frequency. The agreement with
ity changes sign during the trip [Eqs. (11) and (12)] and in a
experimental measurement was excellent in both cases.
sense the average wave can remain close to a field line. The
spread and then refocusing of a shear wave in a large gradi-
ent was observed in an experiment when the waves propa-
VI. FIELD LINE RESONANCES
gated in a plasma with a large gradient in the plasma beta
Alfvénic field line resonances (FLRs) occur when the along the background magnetic field.60
waves are bounded by reflectors of some sort, and a standing When the shear wave due to a current filament propa-
wave is established, much as that on a guitar string. The sub- gates across the magnetic field it may encounter density and
ject has nearly a 30-year history and problems facing the temperature gradients. This was observed in an experiment
FIG. 12. (Color) Experimental setup (a). The RMF antenna is aligned so the rotation axis of the induced magnetic field at the center of the antenna is parallel
to the direction of the background magnetic field B0. The data are collected with the use of 3-axis magnetic pickup probes (1mm cube). (b) The magnetic field
of the wave measured on a plane transverse to the background field 33 cm from the antenna. Data was acquired at 13 448 spatial locations and in figure (b) ev-
ery 4th vector was drawn.
on the LAPD in which the electron beta, be ¼ M 8pnTe

m B20z varied ing, the generation of an axial magnetic field component,
by a factor of four across the outer edge of the plasma col- and the change from linear to left-hand circular polarization.
umn.61 The center of the plasma was kinetic, and the perpen- Stix62 detailed the cyclotron damping of Alfvén waves in the
dicular group velocity was outward; there was a frequency case of a uniform background magnetic field; experimen-
dependent radial position at which vgr\ ¼ 0, and the wave tally, some of the earliest works on Alfvén waves in nonuni-
became inward propagating (backward) in the outer (inertial) form magnetic fields were performed by researchers63–65
region. The difference in radial propagation velocities caused studying the excitation of radial eigenmodes of a cylindrical
pileup of wave energy on a ring of radius between plasma. The wave propagated axially into a region of
1 be 2 which agreed with theory. Spatially localized decreasing magnetic field to the point where the wave fre-
accumulation regions such as this could result in large ampli- quency matched the local ion-cyclotron frequency—the so-
tude waves, representing sources of electron acceleration due called ‘‘magnetic beach.’’ Later, the propagation of shear
to Landau damping. Alfvén waves and ion heating were also of great interest in
the study of mirror machines. For example, radio frequency
heating was studied in the THM-2 mirror machine66 using a
VII. MAGNETIC BEACH
Stix-type coil exciter, and in the Phaedrus-B tandem mirror67
Due to their long parallel wavelengths and nearly field- using a rotating field antenna68,69 which selectively excited a
aligned propagation, shear Alfvén waves typically encounter global, m ¼ 1 eigenmode of the shear Alfvén wave and
spatially varying regions of ambient plasma parameters. Per- was found to efficiently transfer energy to the ions.
haps the most important of these is the variation of the back- The magnetic beach scenario was later revisited by Vin-
ground magnetic field. When the local ion cyclotron cena et al.149 They used a small circular mesh to broadcast
frequency becomes comparable to the oscillation frequency shear Alfvén waves into a decreasing magnetic field as
of the shear wave, additional effects arise such as ion heat- shown in Fig. 16. Detailed measurements of the wave
FIG. 13. (Color) Isosurfaces of current

density at s ¼ 4.6ls after the wave
(xxci ¼ 0:54) is launched by the RMF
antenna described in the text. The surfa-
ces begin 33 cm to the right of the
antenna and end approximately 8 m
away. Two rotating counter-propagating
helical current channels can be seen
flowing in the z direction along B~0z . As
time advances the currents rotate in a
left-handed sense. The outer surface rep-
resents a current density of 0.25 A/cm2
and the inner surface a current density of
0.5 A/cm2. Red denotes current flow in
the positive z direction and the blue
surfaces current in the negative z direc-
tion. Some representative magnetic field
vectors are also shown.
cates the location where the wave frequency equals the local
cyclotron frequency of the helium ions. The wave is effi-
ciently absorbed here, as in Fig. 17. In addition to the previ-
ously discussed radial ion polarization and parallel electron
currents, the streamlines show a pronounced azimuthal com-
ponent. This current arises from the relative slippage in
Er B0 drift velocities between the ions and electrons near
the ion cyclotron frequency and leads to an observed axial
component of the wave magnetic field.
VIII. ALFVÉN WAVE MASER

Several theoretical studies70–72 have investigated the
concept of a natural Alfvén wave maser formed in the plas-
mas surrounding the earth. Some theories consider the maser
cavity formed by the magnetospheric plasma bounded by the
highly conducting, conjugate ionospheres. These relatively
narrow regions play the role of reflectors. The nonequili-
FIG. 14. Distribution of oscillation frequencies obtained from the estimation brium, active medium that excites the wave has been associ-
of magnetopause motion by spline interpolation (Plaschlke et al., 2009).
ated with a population of energetic ions73 in the radiation
belt. Another class of Alfvén maser activity is considered to
magnetic fields showed that in addition to the ion cyclotron occur at the lower altitudes. Due to the spatial variation of
damping, the wave parallel electric fields lead to significant the ionospheric plasma parameters, it has been identified in
transfer of wave energy to the electrons via (equal) electron various studies55,74–76 that an Alfvén resonator is naturally
Landau damping and electron-ion Coulomb collisions. Fig- formed between the lower E region of the ionosphere and a
ure 17 displays the measured wave fields and WKB solutions region in the topside ionosphere at an altitude of about 104
to the wave equation that delineate the damping effects by km. This higher region acts as a partial reflector that allows
ions and electrons. The best agreement with the data required amplified signals to propagate into the magnetosphere.
all three damping mechanisms. The increase in the wave am- Observations by satellite77 and rocket78 suggest that signals
plitude for the ion damping only (red) curve is due to the consistent with resonant amplification of Alfvén waves are
slowing down of the parallel group velocity as the wave naturally generated and corroborate ground-based measure-
approaches the cyclotron frequency. This effect is over- ments79,80 of the phenomena. The amplified signals corre-
whelmed when all three damping mechanisms are included spond to waves that undergo multiple reflections within the
(blue curve). resonator. A modeling study81 has proposed that the strong
The magnetic field measurements were also used to cal- fluctuations associated with the partially standing waves in
culate the volumetric wave currents, as shown in Fig. 18. the ionospheric resonator are responsible for driving the out-
This figure displays a snapshot in time of the shear wave as flow of source ions from the ionosphere to fuel the magneto-
it propagates from left to right into the magnetic beach. spheric plasma.
Shown is the magnitude of the azimuthal magnetic field in Various nonequilibrium sources have been considered to
the x-z plane, and streamlines of wave current calculated excite the ionospheric Alfvén resonator. These include the
using Ampere’s law and the azimuthal symmetry of the natural magnetospheric convection,55,82 electron beams,83
wave magnetic field. The magenta surface on the right indi- lightning discharges,84,85 and powerful HF radio signals.86 In
FIG. 15. (Color) (left) Spectrum of frequencies in the exciter pulse (dotted line), and observed spectrum measured by an in-situ magnetic pickup probe. (right)
Measured magnetic field for the lowest order mode n ¼ 1, showing the FLR is confined to the center of the device.
FIG. 16. Side view of the original LaPD

showing the magnitude of the confining
axial magnetic field (at r ¼ 0) for the
magnetic beach experiments. Also indi-
cated is the region in which detailed
measurements were obtained. The wave
travels from a region of high electron
Landau damping to a region of ion cy-
clotron absorption.
laboratory studies87,88 performed in the LAPD device, an transparent mesh anode that permits the amplified shear
equivalent Alfvén resonator has been realized in which the Alfvén waves to propagate into a large plasma column. In
nonequilibrium source leading to spontaneous amplification this region, where the magnetic field is uniform, the unique
is a field-aligned current. properties of the maser signal can be exploited to study a
The LAPD studies have demonstrated that a steady-state wide variety of interactions of relevance to magnetic con-
Alfvén wave maser can be realized in the laboratory under finement research and space plasma investigations.
controlled conditions. The configuration shares the essential An example of maser behavior near threshold conditions
elements of lasers and maser devices that operate at optical is shown in Fig. 19. The bottom panel of this figure shows
and microwave frequencies, namely, a resonant cavity with a the temporal behavior of the spontaneous fluctuations of the
partially reflecting boundary and an active medium that pro- transverse component of the magnetic field when the plasma
vides amplification. Because of the peculiar properties of current barely exceeds a threshold value for a given strength
shear Alfvén waves, these elements take a form different of the confinement field. The threshold value corresponds to
from the optical cavities and inverted atomic populations the condition when the return current in the device results in
typically found in conventional lasers and masers. The reso- an electron drift velocity that exceeds the phase velocity of
nant cavity used in LAPD consists of a cathode and a semi- the shear mode (approximately given by the Alfvén speed).
The top panel displays an expanded view of the temporal
behavior (single shot) of the flare event over a small interval
of 0.1 ms. It shows that the flaring shear mode is nearly
monochromatic and remarkably coherent, in particular since
it grows out of spontaneous ambient noise. For plasma cur-
rents well-above threshold the maser signal achieves a
steady-state whose length is determined by the discharge
duration.
The Alfvén maser in the LAPD produces a highly coher-
ent shear mode with dx/x 5 103 excited spontaneously
from ambient noise when a threshold plasma current is
exceeded. There is no need for external pumping to trigger
the maser action. The amplitude of the magnetic fluctuations
associated with the noise-grown signal attains a level
exceeding 1.5% of the confinement magnetic field. The fre-
quency of the maser mode is continuously tunable by chang-
ing the strength of the confinement field. The cathode-anode
separation and the dispersion relation of the shear mode
determine its specific value. For the particular experimental
arrangement in the LAPD this corresponds to x ¼ 0:6xci . A
FIG. 17. (Color) Comparison of measured to theoretical wave magnetic quantitative comparison of the measured threshold plasma
field as the shear Alfvén wave propagates into a decreasing magnetic filed. current, over a factor of 2 variation in confinement field,
Lower axis is in cm and upper is the ratio of wave frequency to local helium shows that the amplification mechanism is the electron drift
ion cyclotron frequency. The red curve shows on ion cyclotron damping in
velocity associated with the natural return current in the
the model while the blue further incorporates electron Landau damping and
electron-ion Coulomb collisions. The electron damping channels are equally plasma. The maser action exhibits a wide range of temporal
effective. The antenna is located at z ¼ 0 cm. variability. Near threshold it consists of flare events whose
FIG. 18. (Color) Shear Alfvén wave

currents (streamlines) and magnetic field
magnitude (colored plane) at one instant
of time. The wave propagates from left
to right until it encounters the magenta
surface at the right where the wave fre-
quency equals the local ion cyclotron
frequency (see also Fig. 16). Parallel
currents are carried by electrons, radial
currents are due to ion polarization
drifts, and azimuthal currents result from
the slippage of the electron and ion
E B drifts near the cyclotron
frequency.
statistics displays a non-Gaussian tail. For conditions well In summary, it has been demonstrated that Alfvén wave
above threshold the maser behavior reaches a steady state masers can be realized in the laboratory and their operation
whose duration is determined by the lifetime of the plasma can be quantitatively predicted by a differential equation
discharge. Mode transitions, from azimuthal mode number model.89 The maser is not only a useful laboratory tool for
m ¼ 0 to m ¼ 1, occur and are associated with slow changes investigating a wide variety of problems related to Alfvénic
in the radial density profile and with changes in plasma pa- interactions,90 but may also be an important process in space
rameters. Under steady-state conditions the radiation resist- and astrophysical plasmas. A discussion of such perspectives
ance associated with the Poynting flux emanating from the has been given by Boswell.91
resonant cavity modifies the discharge current. In spite of the
relatively large level of the signals, the related magnetic fluc-
IX. ALFVÉN GAP MODES
tuations do not exhibit a significant nonlinear distortion. This
may be expected because, at the maser frequency of 0.6 Xci, When a shear wave propagates in a rippled magnetic
the higher-harmonic modes are non-propagating. However, field, one with periodic variation, an interesting effect
ion saturation signals that are phase coherent with the maser occurs; the waves can be observed only in certain propaga-
fluctuations indeed exhibit strong nonlinearity. tion bands. The effect is analogous to Bragg scattering in a
crystal in which standing waves are set up by absorption and
re-radiation of an electromagnetic wave by the valence elec-
trons in a crystal lattice. The first paper by to discuss the gap
modes in a toroidal paper was published in 1978 by Pogutse
and Yurchencko.92 In a plasma the variation of the magnetic
field results in a variation of the Alfvén velocity, which in
93
turn causes a variation of the index
of refraction

2
16pni mi c 2pz
N2 ¼ 2
1 þ 2g cos : (13)
ðBmax þ Bmin Þ Lm
Here Lm is the mirror length, Bmax and Bmin the maximum and
minimum values of the magnetic field, ni the ion density, mi
the ion mass, and g the modulation amplitude g ¼ ððBBmax Bmin Þ
max þBmin Þ
.
Figure 20 shows the experimental setup. Four magnetic
mirrors, created along the length of the LAPD and a small
disk antenna40 or a field aligned “blade” antenna, made a fluc-
tuating current filament and launched a shear Alfvén wave.
The magnetic field of the waves was measured on a number
of transverse planes and lines along the z axis ofthe machine.
Figure 21 displays the frequency response clearly show-
ing the Alfvén gap where there is little wave activity (for an
infinite system of mirrors there would be none). The gap
FIG. 19. Alfvén wave maser signal near threshold shows bursty behavior.
width for an infinite number of mirror cells is given by
VA
Inset: blow up of a single shot trace of Bx is highly coherent. Above thresh- Df ¼ gfBragg ¼ g 2L M
, where VA is the average Alfvén speed
old the maser signal reaches steady state. which was measured by probes that detected the delay
FIG. 20. (Color) Experimental magnetic field setup of the magnetic mirrors in the LAPD. Here g ¼ 0.26 and Lm ¼ 3.5 m. The plasma parameters: n ¼ 1 1012
cm3, Te ¼ 6 eV, Ti ¼ 1 eV, He. The waves were launched at 50kHz f 180kHz.
between launch and reception at two axial locations. Figure become coupled to the electrostatic drift wave. The resulting
21(right) illustrates how the gap modes develop after the mode is known as the drift-Alfvén wave. It displays simulta-
Alfvén waves have had time to reflect from the ends of the neous fluctuations in density and transverse magnetic field.
device. The gap spectral width was found to increase with These modes can be driven unstable by electron Landau
the mirror depth, M as predicted by theory. This experiment damping and/or collisions. When they reach large amplitude,
is related to TAE modes in fusion devices.94–96 In a burning they play a key role in the cross-field transport of energy and
plasma these modes could be destabilized by energetic alpha particles. Drift-Alfvén waves have been observed in density
particles. Fast ions expelled by these instabilities have dam- and temperature striations produced under controlled condi-
aged vacuum components in past experiments97 and are a tions in the LAPD.98–100
worry for experiments on the ITER device. The anomalous electron heat transport induced by drift-
Alfvén waves has been documented in experiments on the
temporal and spatial evolution of an electron temperature fil-
X. DRIFT-ALFVÉN WAVES AND TRANSPORT
ament. The filament is created by injecting low voltage (less
In large magnetized plasmas in which pressure gradients than the ionization voltage) electrons into the afterglow he-
are present the diamagnetic drift causes the Alfvén wave to lium plasma of the LAPD using a small (3 mm) source.
FIG. 21. (left) The wave energy density as a function of f/fBragg. The width of the dip at 1.05 agrees well with the prediction for Df. (right) Time history of the
running cross-covariance between the wave field and the current of the antenna that launched the wave at g ¼ 0.25. The grey and white contours show the
wave magnetic field. The first white curve indicates the time it takes for the wave to travel towards the detector probe one-way, the second curve the time it
takes for the wave to return to the probe after being reflected from the anode/cathode for the first time. The standing wave and “gap” mode appears after this.
the random occurrence of Lorentzian-shaped pulses in the

time signal.103 The Lorentzian temporal shape observed in
the time signals arises from structures in electron tempera-
ture created from the flow driven by drift-Alfvén waves. Fig-
ure 24 shows the contour of temperature in a simulation104
consisting of two waves, one with m ¼ 1 and one with m ¼ 6.
The wave induced flows to create complex structures whose
temporal signals exhibit Lorenzian shaped pulses. There is a
strong indication that the exponential power spectrum and
FIG. 22. A schematic of the experimental setup used in creating and study- Lorentzian pulses are a universal feature of E B driven
ing the temperature filament showing the location of the electron beam and edge turbulence, as they are observed in a limiter experiment
the probes used to diagnose the experiment. at the edge of the LAPD plasma column involving density,
and not temperature fluctuations at a much larger perpendic-
Since heat conduction along the magnetic field is much ular scale.
stronger than across, the small size of the source limits the
XI. NONLINEAR ALFVÉN WAVES
length of the filament so that it fits entirely within the LAPD
plasma column. The experimental set up is shown in Fig. 22. Nonlinear effects associated with Alfvén waves are im-
It is observed that initial temperature transport is gov- portant in laboratory, space, and astrophysical plasmas.
erned entirely by classical processes (Coulomb collisions), Large amplitude Alfvén waves can modify the background
but that later in time there is a transition to anomalous trans- plasma density through ponderomotive forces,105 which arise
port.101 This transition in the nature of transport occurs as due to spatial gradients in the wave energy density. Field-
the result of the growth of drift-Alfvén waves due to the aligned density cavities in the presence of strong magnetic
pressure gradient associated with the spatially localized fila- fluctuations have been observed in the magnetosphere by the
ment.102 The drift-Alfvén waves driven by the gradient are FAST satellite.106 Alfvén waves are in turn affected through
observed96 to have azimuthal wave numbers ranging from interactions with field-aligned density structures. Field-
m = 1 to m 6, as shown in Fig. 23. Convective E ~ B ~ flows aligned density cavities can act as waveguides for Alfvén
driven by these waves create structures in the temperature, waves,107 channeling Alfvén wave energy. In magnetic con-
and, once a threshold amplitude is crossed (vDrift > x/k\), finement fusion devices, nonlinear processes associated with
the behavior of the plasma flow becomes chaotic leading to multiple Alfven eigenmodes can lead to a significant
enhanced transport. enhancement of fast ion transport.108 It has recently been
The key signature of the transition in transport behavior proposed that Alfvén waves with frequency below the cyclo-
is found in the power spectrum of the fluctuations. The tron frequency can through a nonlinear process resonantly
power spectrum changes from a line spectrum observed dur- heat ions,109 which could explain anomalous ion tempera-
ing the early classical transport stage when the drift-Alfvén tures in magnetic confinement experiments.110
waves become linearly unstable to a broadband spectrum in Nonlinear interactions between Alfvén waves are respon-
the anomalous transport phase when the waves achieve large sible for the cascade of energy in magnetohydrodynamic
amplitude. The broadband spectrum is linear in a log-linear (MHD) turbulence.111,112 In the incompressible MHD limit,
plot indicating that the spectrum is exponential in nature. It nonlinear interactions are only possible among counter-propa-
has been shown that the exponential spectrum arises from gating shear waves, and an anisotropic energy cascade
FIG. 23. (Color) (top panel) Contours of fluctu-

ations in ion saturation current show a steady
progression in wave number of the drift-Alfvén
waves driven by the pressure gradient. (bottom
panel) The temporal signal obtained at the loca-
tion marked in the first contour plot indicates
that the frequency of the signal is nearly con-
stant even though the m-number of the mode is
changing dramatically, indicating that the ei-
gen-frequencies of this system are nearly
degenerate.
FIG. 25. (Color) (top) Measured magnetic fluctuation spectrum during weak
multimode emission (dominant m ¼ 1 mode). To the right is a line plot of the
power spectrum at t ¼ 8 ms. (bottom) Measured magnetic and density (ion sat-
uration current) fluctuation spectrum during strong multimode emission.
bation then scatters the Alfvén waves, resulting in the crea-

FIG. 24. (Color) (top panel) Contours of temperature produced by convection of sidebands. The first observation of this interaction
tion in the presence of two drift waves shows the development of structure was made during simultaneous, spontaneous emission of
from the initially Maxwellian temperature distribution. (middle panel) Lor-
entzian-shaped decreases in temperature produce near the center of the fila-
two frequencies (two azimuthal modes, m ¼ 0 and m ¼ 1) by
ment (r ¼ 1.85 mm) produced by inward transport. (bottom panel) the Alfvén wave maser. Multimode spontaneous emission is
Lorentzian-shaped increases in temperature produced in the outer regions of often observed early in time (mode-hopping as the plasma
the filament (r ¼ 3.85 mm) due to outward transport. profiles evolve),92 but in some circumstances the multimode
emission continues for longer periods of time. Figure
25(left) shows measured magnetic field and ion saturation
results.113,114 As this cascade proceeds, perpendicular scales current power spectra during two discharges, the first with a
are reached where dispersion and kinetic damping become dominant single mode (m ¼ 1) and the second with sustained
important, and a cascade of kinetic Alfvén waves is multimode emission (both m ¼ 0 and m ¼ 1 are emitted
expected,115 and evidence for the transition to a KAW cas- simultaneously). The observed magnetic fluctuation spec-
cade is observed in solar wind measurements.116 With the trum in the multimode case contains many sidebands around
inclusion of dispersion and compressibility, a greater range of the MASER emission frequencies and a low frequency fluc-
nonlinear wave-wave interactions are possible among shear tuation at the beat/sideband separation frequency (and har-
Alfvén waves. This includes decay instabilities, such as para- monics). The response at the beat frequency is a non-
metric117 and modulational118 decay, as well as interactions resonant quasimode, having wave number consistent with
between co-propagating waves. three-wave matching rules, but inconsistent with the disper-
Studies of large amplitude shear Alfvén waves and sion of linear plasma waves.
wave-wave nonlinear interactions have been carried out in This beat wave interaction has been used in LAPD to
the LAPD.90 Large amplitude shear Alfvén waves are gener- couple to linear plasma modes. An experiment was per-
ated in LAPD using two sources: (1) the resonant maser formed in which two Alfvén waves were broadcast along a
cavity defined by the anode and cathode in the source filamentary density depletion on which a coherent unstable
region89,90 and (2) loop antennas. Using these wave sources, mode was present.119 The beating Alfvén waves were found
the interaction between two co-propagating shear Alfvén to resonantly drive the instability when the beat frequency
waves has been studied. In these experiments, the two was tuned to match the unstable mode frequency. More
waves, launched at slightly different frequencies are interestingly, it was found that when the beat frequency was
observed to beat together and nonlinearly drive a density tuned to resonantly drive nearby (in frequency) damped
fluctuation at the beat frequency. The driven density pertur- modes, the original unstable mode was suppressed.
XII. ALFVÉN WAVES FROM DENSE EXPANDING experiment was interpreted as the signature of reconnec-
PLASMAS tion.132 The data in that experiment was interpreted in terms
of the Sweet-Parker model133 (which is two-dimensional and
The expansion of a dense plasma into a magnetized predicts the correct reconnection rate if multiplied by a large
background plasma is of interest in a variety of contexts. It is situation dependent number), a turbulent resistivity 25 times
of importance in the study of coronal mass ejections the classical Spitzer resistivity was assumed. It was conjec-
(CME)120 and high altitude explosions.121 In the past there tured that this could be due to ion acoustic turbulence. Ion
have been chemical releases in the ionosphere,122 in which acoustic turbulence and enhanced resistivity were observed
barium was released (AMPTE (Ref. 123) mission). The bar- in a reconnection experiment over 25 years ago.134 Another
ium rapidly became photoionized and expanded across the high energy colliding lpp experiment was designed to model
geomagnetic field. Experiments at UCLA in which a dense young stellar objects and studied collimated flows produced
laser produced plasma (lpp) expanding into the background in the interaction.135 Plasma jets and shocks were observed.
LAPD plasma have studied the formation of a diamagnetic These experiments are complicated, the diagnostics are diffi-
cavity124 and generation of Alfvén waves in the background cult to set up, and “campaigns” consist of several shots.
plasma.125 A discussion of the relevance of these experi- These experiments had no background plasma or background
ments to space plasmas has been reported.126 Figure 26 magnetic field although in astrophysical situations they are
shows the setup for laser target experiments in the case usually present.
where two targets are struck simultaneously. We first briefly The magnetic field data in the two-target LAPD experi-
discuss the case of a single expanding lpp. When the dense ment was acquired with a three axis differentially wound
lpp plasma is created in vacuum the electrons cannot stream magnetic probe 1 mm in diameter over weeks of continual
away along the existing ambient magnetic field because they operation at a rep rate of 1 Hz. The magnetic probe was com-
are held back by the ion space charge. The presence of a puter controlled and stepped through a series of planes trans-
dense enough background plasma allows electrons to move verse to the background magnetic field. Magnetic fields
away in localized streams and be replaced by drifting elec- associated with the dense plasmas and wave propagation in
trons from the much larger background plasma. The initial the background plasma are digitized (14 bit, 100 MHz) and
burst of electrons generates the emission of whistler and acquired at 16 000 spatial locations. An average of 10 shots
lower hybrid waves for s 400ns (Ref. 127). The lower fre- was stored at each location. The reconnection event at s
quency current system that is set up subsequently becomes ^ 400 ns, shown in Fig. 28, is accompanied by a burst of
that of shear Alfvén waves. Figure 27 illustrates the current current induced along the background magnetic field
systems and Alfvén wave formation process.128 (I ¼ 30A Ielec-sat). This is not the result of fast electrons
Note that initially the current system is coaxial. The escaping the bubbles but is due to the inductive electric field
electrons streaming away (red) are surrounded by those in a generated in this first reconnection event. The lpp current/
return current (blue). At later times the current system return current systems follow the initial reconnection event
becomes complicated because the lpp is moving across the and develop on an Alfvénic time scale. These currents
magnetic field in the x direction and shedding current as it (s ¼ 5.25 ls) are shown in Fig. 29. The system is dynamic,
goes. The magnetic field shown on the plane at dz ¼ 1.3 m is and the currents are observed to merge, break up into two or
associated with a shear wave that is just beginning to form. four channels, and re-merge. Over time scales
Note the morphology of the wave bears a stringing resem- 1ls s 40ls, the reconnection is that of the magnetic
blance to that generated by the RMF antenna (Fig. 12). fields of shear Alfvén wave currents. The magnitude of cur-
The interaction of the current systems from multiple rents of the lpp
Alfvén
waves varies in space and time and is
laser-produced plasmas brings an additional phenomenon of the order of
J~wave
0:2A=cm2 . The magnetic field dur-
into play, magnetic field line reconnection.129 Colliding plas- ing one such event is shown in Fig. 30. The induced electric
mas are relevant to astrophysical situations.130 Consider two field may be calculated using
targets simultaneously struck by laser beams (Fig. 26). The ð ~ 0
~ J ð~
r Þ
lpp’s jet across the magnetic field and collide but the relevant ~ind ¼ 1 @ A ;
E ~ ¼1
A dV 0 ; (14)
part of the experiment of concern here is the interacting c @t c r ~
j~ r 0j
V0
current systems. The experiments were done in helium ðB0z
¼ 600 G; fci ¼ 229 kHz; n ¼ 3 1012 cm3 ; Te ¼ 4 eVÞ. The where J is the volumetric current evaluated from the mag-
laser power incident on each target was 5 1010 W/cm2. In netic field data. The induced electric field is on the order of
the initial stage, instead of a single diamagnetic bubble there 2.5 V/m during the bursts of reconnection. If one ignores all
are two. The expelled magnetic field is in the axial direction, the other relevant terms in Ohm’s law, an extreme oversim-
but as the diamagnetic currents are helical there is a field plification, the resistivity is g ¼ 1.25 103X m, 32 times
component in the transverse direction. When they collide the higher than the classical Spitzer resistivity gSpitzer ¼ 4
anti-parallel transverse field components are forced together 105X m. This is a limit one must view with suspicion.
and reconnection occurs. This is shown in Fig. 28. In the
case of colliding lpp’s formed by more powerful lasers (100-
XIII. MAGNETIC FLUX ROPES
200J) at the Rutherford laboratory,131 the transverse fields at
the bubble edge are on the order of 1 MG. Jetting of plasma A magnetic flux rope is a bundle of magnetic field lines
from the edge of the reconnection region in the Rutherford with pitch varying with radius. One may alternatively
FIG. 26. (Color) Two laser beams (shown in red although they are 1 micron wavelength and not visible, s¼ 8 ns, 1.5 J) strike a pair of carbon targets in the LAPD
device which contains the LAPD background plasma (diameter indicated), 60 cm in diameter. The port spacing along the background magnetic field is 32 cm.
FIG. 27. (Color) Magnetic field data at t ¼ 390 ns after the target is struck. Solid planes are current density obtained from J~ ¼ 4p
c
! B.~ Red/blue represents
electrons going in the –z/þz direction. Vector plot shows the perpendicular magnetic field. The target is shown to scale at its appropriate location relative to the
z ¼ 6 cm plane. The laser is incident along the x axis.
FIG. 28. (Color) Magnetic field measured 400 ns after the targets are struck by two 1-J lasers. The data plane is 5 cm away from the targets. The spatial extent
of the data is Dx ¼ 22 cm, Dy ¼ 23 cm. (a) Transverse magnetic field from a head-on perspective. A magnetic X point is clearly visible in the center. The bub-
bles are moving towards each other and flux is annihilated in the center. (b) Another view of the same data showing that most of the magnetic field is in the z
direction. The background magnetic field points downwards.
consider a rope as a spiraling current system. To create a flux flux ropes may be considered to be Alfvén waves. The sun is
rope, one needs current along a background magnetic field, host to arched magnetic flux ropes136 which can erupt to
Bz, large enough to produce a significant Bh. Current systems become coronal mass ejections. CME’s are flux ropes that
in magnetoplasmas are rarely steady state, and once they emerge the sun and extend outwards for large distances.137
change on time scales less than the ion cyclotron period the They can break away from the sun to become plasmoids,138
FIG. 29. (Color) Three-dimensional currents in the two-target lpp experiment derived from the volumetric magnetic field data set. The targets are drawn in the
background for reference. The dominant Alfvén wavelength is kP ¼ 2.62m. The current closes by ion polarization drift every half wavelength, the currents
close at dz ¼ 1.31 meters from the targets at s ¼ 5.25 ls after the targets are struck.
which can strike the Earth. It is also possible that they remain the ropes were emitted from two LaB6 cathodes. The experi-
rooted in the Sun when their leading edge arrives at Earth. ment generated a volumetric data set (12 data planes perpen-
An early experiment on the interaction of low-current ropes dicular to B0z ¼ 270 G, dz ¼ 64 cm, dx ¼ dy ¼ 0.3 cm, 32
without a background plasma139 observed that they spiraled 400 spatial locations, dt ¼ 0.64 ls, time digitized 6.55 ms).
around one another due to J~ B ~ forces and then merged. The ropes twisted about themselves, writhed about each
Over distance and time the currents became field aligned in other, and were also kink unstable. The flux rope collisions
accord with Taylor’s conjecture.140 When conditions are led to multiple reconnection sites in space and time. As there
such that the flux ropes violently bash into one another, was a strong background magnetic field, the reconnection
which can happen, for example, if they are kink unstable, occurred in the absence of a magnetic null point just as in
field line reconnection can occur in the space between them. the lpp experiment. Solar physicists have introduced the con-
This was recently seen in an experiment done on the RSX cept of a quasi-separatrix layer143,144 (QSL) to study recon-
device at Los Alamos141 where the magnetic topology, nection in complex geometries and no magnetic null points.
reconnection rate, and resistivity (larger than classical as The volumetric magnetic field data set enabled the evalua-
usual) were measured. In the same year an experiment at tion of the QSL associated with the reconnection event. The
UCLA on colliding flux ropes142 was done on the LAPD QSL region has a simple interpretation. Two closely spaced
device. field lines, which enter the QSL, wind up at very different
The UCLA flux rope experiment, however, had a back- spatial separations at finite distances along the current chan-
ground plasma capable of supporting Alfvén waves (He, n nel. Outside the QSL neighboring field lines remain close to
^ 2.5 1012 cm3, Te^ 5 eV, Ti^ 1 eV). The currents of each other. The QSL can be found by moving the footprint
FIG. 30. (Color) Magnetic field lines of a shear Alfvén waves produced 1.6 ls after the collision of two laser-produced plasmas. The targets are several meters in
the back of the picture. At this instant of time there are two strong and two weak current channels. The background magnetic field (not shown) is out of the plane
of the image. The inductive electric field is calculated from the time varying magnetic field and rendered as sparkles on the field lines (the brighter the sparkle the
higher the field). The largest induced fields, parallel to B0z, are of order 2.5 V/m. The reconnection region is in the center.
of a field line a small distance from its original position (x,

If N >> 1 then the field lines used to generate it are said to
y) on a plane (z ¼ z0) and measuring where it winds up (X,
be on a QSL. The concept of the QSL has been extended to
Y) on a plane far away (z ¼ z1).
encompass situations in which there is a guide field.145 The
Consider the quantity
slip squash factor Q is defined by
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi N2
u" 2 2 2 2 # Q ¼

: (16)
u @X @X @Y @Y
N¼t þ þ þ : (15)
Bz ðz0 Þ
@x @y @x @y
B ðz Þ
z 1
FIG. 31. (Color) (a) Magnetic field lines started at the edge of the two, 2.5 cm diameter flux ropes. The current in each rope is 30A. The 270 G guide field is
included in the calculation. The grid lines occur at dx ¼ 2 cm, dz ¼ 30 cm. Note the field lines twist about themselves, and rotate about one another. The ma-
genta surface is the QSL ¼ 200 surface which is also a flux tube. The yellow lines are the transverse magnetic field of the ropes dz ¼ 1.3 meters from their
source. Careful inspection of transverse field lines shows there are “X” type structures in the center. The blue arrows are electron current density. Note that at
the right current is flowing backwards, i.e., electrons are going towards the source of the ropes. (b) Lower right corner X point topology dz ¼ 6.6 m from start
of flux ropes. t ¼ 2.5 ms after the flux ropes are switched on.
The squashing degree may be used as a measure of 3D become close to one another in the course of their writhing
reconnection.146 In the case of this experiment the guide field about and when reverse current sheets develop. The flux
is much stronger than the field due to the currents; it does not ropes field lines and QSL are shown in Fig. 31.
vary in z, and Q ! N2. The minimum value of interest is The QSL surface shown in Fig. 31 indicates a region
Q ¼ 2, much larger values denote more intense reconnection (colored magenta) snaking between the field lines at the edge
events. The maximum value of Q peaks when the flux ropes of the ropes; somewhere within that region magnetic field
FIG. 32. (Color) Magnetic field lines of

three flux ropes (background field of 300
G included). Note the scale difference
between the transverse (Dx ¼ 13.2 cm)
and axial distances (Dz ¼ 4.79 m). The
flux ropes twist around one another and
collide in space and time. The reflective
surface is placed on the bottom to aid in
perspective.
FIG. 33. (Color) Three-dimensional picture (anaglyph) of three flux ropes and the resultant QSLs formed (Q ¼ 20). To view this requires red/blue 3D glasses,
which can be purchased in a host of specialty shops. The speckled pattern in the back provides a reference at infinity.
FIG. 34. (Color) Shear wave Hydra.
line reconnection occurs. To identify the location(s) one and even for Alfvén waves the world was a simpler place.
must examine the detailed magnetic field topology. The right The advances continue and as they do so the complicated and
inset in Figure 31 shows the transverse magnetic field in a sometimes beautiful world of Alfvén continues to unfold.
plane dz ¼ 6.6 m from the source of the flux ropes. Viewed This publication is accompanied by stereo slide pairs
in time field lines enter the “X” point, break, and reconnect. and stereo movies, which may be downloaded from the
It is also possible to estimate the resistivity
Ð by comput- Physics of Plasmas website (http://pop.aip.org/shear_alf-
ing the time rate of change of the helicity: V A ~ B
~dV and the ven_waves).150 There is an accompanying text file with fig-
magnetic energy.147 ure captions as well as instructions for viewing.

2

dH
ACKNOWLEDGMENTS

ð

dt
1
g

; with M ¼ B2 dV: (17) The authors would like to thank our collaborators: Craig

dM
8p Kletzing, Bill Heidbrink, Andrew Collette, Roger McWil-
2M
dt liams, Hans Boehmer, Dennis Papadopoulis, Yuhou Wang,

Eric Lawrence, Y Zhang, and Shu Zhou, for their valuable
The relevant quantities were calculated from the volumetric contributions. We also appreciate the expert technical sup-
data set.148 The volume-averaged resistivity varies in time port of the LAPD staff, Zalton Lucky, Marvin Drandell, and
and jumps to a minimum of five times the Spitzer resistivity Mio Nakamoto. The operation of the LAPD at the Basic
when the flux ropes collide and there is a QSL. When there Plasma Science Facility (BaPSF) is supported by DOE/NSF.
is no reconnection the QSL disappears. Partial support for some aspects of the work was provided by
The next study involved three flux ropes produced using ONR.
three emitters 2.5 cm in diameter arranged in a triangle. A
volumetric data set was acquired (32 400 spatial locations, 1
H. Alfvén, Mon. Not. R. Astron. Soc. 107, 211 (1947).
2
volume 13.2 cm 13.2 cm 9.6 m, dx ¼ dy ¼ 3 mm, dz S. Tomczyk, S. W. McIntosh, S. L. Keil, P. G. Judge, T. Schad, D. H.
Seeley, and J. Edmondson, Science 317, 1192 (2007).
¼ 63.9 cm, dt ¼ 0.64 ls, Tacquis ¼ 6.55 ms). The data at each 3
P. J. ColemanJr., Phys. Rev. Lett. 17, 207 (1966).
location is averaged over 12 plasma shots. Light fluctuations 4
M. L. Goldstien, D. A. Roberts, and W. H. Matthaeus, Ann. Rev. Astro-
recorded with a fast diode are used to synchronize the data phys. 33, 283 (1995).
5
stream to the motion of the ropes. The flux ropes at one W. H. Matthaeus, M. L. Goldstein, and D. A. Roberts, J. Geophys. Res.
95 (20), 675 doi:10.1029/JA095iA12p20673 (1990).
instant of time are shown in Fig. 32. 6
V. M. Chmyrev, S. V. Bilchenko, O. A. Pokhotelov, V. A. Marchenko, V.
There are more degrees of freedom in this case, and one I. Lararev, A. V. Strelsov, and L. Stenflo, Phys. Scr. 38, 841 (1988).
7
expects that more than one QSL can occur. This is indeed P. J. Louran, J. E. Wahlund, T. Chust, H. de Feraudy, A. Roux, B. Hol-
the case as seen in Fig. 33, which is an anaglyph of the three back, P. O. Dovner, A. J. Eriksson, and G. Holmgren, Geophys. Res. Lett.
21, 1823 doi:10.1029/94GL01424 (1994).
ropes shown in Fig. 32 as well as two QSLs. The rotational 8
J. R. Wygant A. Keiling, C. A. Cattell, M. Johnson, R. L. Lysak,
motion of the ropes is reflected in an optical signal M. Temerin, F. S. Mozer, C. A. Kletzing, J. D. Scudder, W. Peterson,
(f 4kHz). A vacuum ultraviolet spectrometer tuned to a C. T. Russell, G. Parks, M. Brittnacher, G. Germany, and J. Spann,
303 nm He II line (40 eV) recorded oscillations interpreted J. Geophys. Res. 106, 675 (2000).
9
Y. Jantenco-Pereira, Phys. Scr. T 60, 113 (1995).
as electron temperature fluctuations, perhaps due to succes- 10
H. Alfvén, Nature 150, 405 (1942).
sive reconnection events. These experiments point to the fact 11
C.-G. Fålthammar, IEEE Trans. Plasma Sci. 25, 409 (1997).
12
that multiple interacting flux ropes will have multiple QSLs S. Lundquist, Nature 164, 145 (1945);S. Lundquist, Phys. Rev. 76, 1805
(1945).
and the associated reconnection will be bursty and occur at a 13
W. H. Bostick and M. Levine, Phys. Rev. 94, 671 (1952).
host of spatial positions. It is certainly three-dimensional. 14
W. Gekelman, J. Geophys. Res. 104, 14417 doi:10.1029/98JA00161
(1999).
15
XIV. SUMMARY AND CONCLUSIONS W. Heidbrink, Phys. Plasmas 15, 056107 (2008).
16
S. Vincena, G. J. Morales, and J. E. Maggs, Phys. Plasmas 17, 52106
Since the prediction and discovery of shear Alfvén (2010).
17
waves it has been realized that they touch or even dominate B. Fried and S. D. Conte, The Plasma Dispersion Function (Academic,
San Diego, CA, 1961).
nearly every area of plasma physics from Astrophysics to 18
G. J. Morales and J. E Maggs, Phys. Plasmas 4, 4118 (1997).
fusion. This overview has concentrated on a number of 19
G. J. Morales, R. Loritsch, and J. E. Maggs, Phys. Plasmas 1, 3765 (1994).
20
aspects of the wave, which has emerged in recent carefully 21
S. Lundquist, Nature 164, 145 (1949).
controlled experiments. A detailed review of the wave as B. Lehnert, Phys. Rev. 94, 815 (1952).
22
D. F. Jephcott, Nature 183, 1652 (1959).
observed in space is left to other monographs. What appears 23
G. A. Sawyer, P. L. Scott, and T. F. Stratton, Phys. Fluids 2, 47 (1959).
as a variety of very different waves, in different circumstan- 24
J. W. Wilcox, F. Boley, and A. De Silva, Phys. Fluids 3, 15 (1960).
25
ces (uniform plasmas, non-uniform plasmas, density stria- D. F. Jephcott and P. M. Stocker, J. Fluid Mech. 13, 587 (1962).
26
tions, magnetic field gradients, multiple mirrors, flux ropes, L. C. Woods, J. Fluid Mech. 13, 579 (1962).
27
J. W. Wilcox, F. Boley, and A. De Silva, Phys. Fluids 3, 15 (1960).
dense expanding plasmas…) are all different manifestations 28
D. Swanson, R. Gould, and R. Hertel, Phys. Fluids 7, 269 (1964).
of the shear wave. This is exemplified in Fig. 34. 29
G. Müller, Plasma Phys. 16, 813 (1974).
30
Experimental sources, diagnostics and theory have Y. Amagishi and A. Tsushima, Plasma Phys.Controlled Fusion 26, 1489
(1984).
become increasing sophisticated over the years and in all 31
T. Stix, Waves in Plasmas (AIP, New York, 1992), p. 12.
likelihood will continue to do so. Thirty years ago data at 32
Y. Amagishi, M. Inutaki, T. Inutake, and T. Akitsu, Jpn. J. Appl. Phys.
best were x-y plots, simulations, if any, were one dimensional 20, 2171 (1981).
33 68
Y. Amagishi, Phys. Rev. Lett. 64, 1247 (1990);Y. Amagishi, J. Phys. Soc. Y. Yasaka, R. Majeski, J. Browning, N. Hershkowitz, and D. Roberts,
Jpn. 59, 2374 (1990). Nucl. Fusion 28, 1765 (1988).
34 69
W. Gekelman, J. Geophys. Res. 104, 14417 doi:10.1029/98JA00161 D. R. Roberts, N. Hershkowitz, R. P. Majeski, and D. H. Edgell, AIP
(1999). Conf. Proc. 190, 462 (1989).
35 70
G. G. Borg, M. H. Brennan, R. C. Cross, L. Giannoue, and I. J. Donneley, S. V. Polyakov, V. O. Rapoport, and V. Yu. Trakhtengerts, Fiz. Plazmy
Plasma Phys. Controlled Fusion 27, 1125 (1985). 9, 371 (1983).
36 71
W. Gekelman, H. Pfister, Z. Lucky, J. Bamber, D. Leneman, and J. E. P. P. Belyaev, S. V. Polyakov, V. O. Rapoport, and V. Yu. Trakhtengerts,
Maggs, Rev. Sci. Instrum. 62, 2875 (1991). Geomagn. Aeron. 24, 2 (1984).
37 72
P. Pribyl and W. Gekelman, Rev. Sci. Instrum. 75, 669 (2004). P. P. Belyaev, S. V. Polyakov, V. O. Rapoport, and V. Yu. Trakhtengerts,
38
D. Leneman, W. Gekelman, and J. E. Maggs, Rev. Sci. Instrum. 77, Izv. Vyssh. Uchebn. Zaved. Radiofiz. 33, 408 (1990).
73
015108 (2006). V. Yu. Trakhtengerts and A. Ya. Feldstein, J. Geophys. Res. 96, 19363
39
D. Leneman and W. Gekelman, Rev. Sci. Instrum. 72, 3473 (2001). doi:10.1029/91JA00376 (1991).
40 74
W. Gekelman, D. Leneman, and J. E. Maggs, Phys. Plasmas 1, 3775 B. J. Thompson and R. L. Lysak, J. Geophys. Res. 101, 5359
(1994). doi:10.1029/95JA03622 (1996).
41 75
W. Gekelman, S. Vincena, D. Leneman, and J. Maggs, Plasma Phys. Con- R. L. Lysak, J. Geophys. Res. 104, 10017 doi:10.1029/1999JA900024
trolled Fusion 39, A101 (1997). (1999).
42 76
D. Leneman, W. Gekelman, and J. E. Maggs, Phys. Rev. Lett. 82, 2673 O. A. Pokhotelov, D. Pokhotelov, A. Streltsov, V. Khruschev, and M.
(1999). Parrot, J. Geophys. Res. 105, 7737 doi:10.1029/1999JA900480 (2000).
43 77
N. Palmer, W. Gekelman, and S. Vincena, Phys. Plasmas 12, 072102 M. Grzesiak, Geophys. Res. Lett. 27, 923 doi:10.1029/1999GL010747
(2005). (2000).
44 78
C. Kletzing, S. Bounds, J. Martin-Hiner, W. Gekelman, and C. Mitchell, N. Ivchenko, G. Marklund, K. Lynch, D. Pietrowski, R. Torbert,
Phys. Rev. Lett. 90, 035004 (2003). F. Primdahl, and A. Ranta, Geophys. Res. Lett. 26, 3365 doi:10.1029/
45 1999GL003588 (1999).
C. Kletzing, D. Thucks, F. Skiff, S. Bounds, and S. Vincena, Phys. Rev.
79
Lett. 104, 095001 (2010). P. P. Belyaev, T. Bosinger, S. V. Isaev, and J. Kangas, J. Geophys. Res.
46 104, 4305 doi:10.1029/1998JA900062 (1999).
A. Gigliotti, W. Gekelman, P. Pribyl, S. Vincena, A. Karavaev, X. Shao,
80
A. S. Sharma, and D. Papadopoulos, Phys. Plasmas 16, 092106 (2009). F. Z. Feygin, A. K. Nekrasov, K. Mursula, J. Kangas, and T. Pikkarainen,
47 Ann. Geophys. 12, 147 doi:10.1007%2Fs00585-994-0147-8 (1994).
A. V. Karavaev, N. A. Gumerov, K. Papadopoulos, X. Shao, A. S.
81
Sharma, W. Gekelman, P. Pribyl, Y. Wang, B. Van Compernolle, and S. C. C. Chaston, J. W. Bonnell, C. W. Carlson, J. P. McFadden, R. E.
Vincena, “Generation of shear Alfvén waves by a rotating magnetic Ergun, R. J. Strangeway, and E. J. Lund, J. Geophys. Res. 109, A04205
field,” Phys. Plasmas (in review). doi:10.1029/2003JA010053 (2004).
82
48
A.V. Karavaev, N. A. Gumerov, K. Papadopoulos, X. Shao, A. S. R. L. Lysak and Y. Song, J. Geophys. Res. 107, 1160 doi:10.1029/
Sharma, W. Gekelman, A. Giglotti, P. Pribyl, and S. Vincena, Phys. Plas- 2001JA000308 (2002).
83
mas 17, 012102 (2010). N. G. Lekhtinen, G. A. Markov, and S. M. Fainstein, Izv. Vyssh. Uchebn.
49
A. Hasegawa and C. Uberoi, The Alfven Wave (Technical Information Zaved. Radiofiz. 38, 312 (1995).
84
Center, US Department of Energy, Springfield, VA, 1982). H. Fukunishi, Y. Takahashi, M. Sato, A. Shone, M. Fujito, and Y. Wata-
50
K. Takahashi and R. McPherron, Planet Space Sci. 32, 1343 (1984). nabe, Geophys. Res. Lett. 24, 2973, doi:10.1029/97GL03022 (1997).
85
51
M. J. Engebretson, L. J. Zanetti, T. A. Potemra, and M. H. Acuna, A. I. Sukhorukov and P. Stubbe, Geophys. Res. Lett. 24, 829
Geophys. Res. Lett. 13, 905 doi:10.1029/GL013i009p00905 (1986). doi:10.1029/97GL00807 (1997).
86
52
M. R. Lessard, M. K. Hudson, J. C. Samson, and J. R. Wygant, J. Geo- V. Yu. Trakhtengerts, Geomagn. Aeron. 29, 3 (1989).
87
phys. Res. 104, 12361 doi:10.1029/1998JA900117 (1999). J. E. Maggs and G. J. Morales, Phys. Rev. Lett. 91, 035004 (2003).
88
53
J. C. Samson, D. D. Wallis, T. J. Hughes, F. Creutzberg, and T. J. Hughes, J. E. Maggs, G. J. Morales, and T. A. Carter, Phys. Plasmas 12, 013103
J. Geophys. Res. 97, 8495 doi:10.1029/91JA03156 (1992). (2005).
89
54
H. J. Singer, W. J. Hughes, and C. T. Russell, J. Geophys. Res. 87, 3519 G. J. Morales and J. E. Maggs, Phys. Plasmas 13, 052109 (2006).
90
doi:10.1029/JA087iA05p03519 (1982). T. A. Carter, B. Brugman, P. Pribyl, and W. Lybarger, Phys. Rev. Lett.
55
F. Plaschke, K. H. Glassmeier, H. U. Auster, O. D. Constantinescu, W. 96, 155001 (2006).
91
Magnes, V. Angelopoulos, D. G. Sibeck, and J. P. McFadden, Geophys. R. Boswell, Nature (London) 425, 352 (2003).
92
Res. Lett. 36, LO2104 doi:10.1029%2F2008GL036411 (2009). O. P. Pogutse and E. I. Yurchencko, Nucl.Fusion 18, 1629 (1978).
93
56
A. Tsushima, Y. Amagishi, M. Inutake, Phys. Lett. 88A, 457 (1982). Y. Zhang, W. W. Heidbrink, H. Boehmer, R. McWilliams, G. Chen, B.
57
C. C. Mitchell, S. Vincena, J. E. Maggs, and W. Gekelman, J. Geophys. N. Breizman, S. Vincena, T. Carter, D. Leneman, W. Gekelman, P.
Res. 28, 923 (2001). Pribyl, and B. Brugman, Phys. Plasmas 15, 012105 (2008).
94
58
C. C. Mitchell, J. E. Maggs, and W. Gekelman, Phys. Plasmas 9, 2909 C. Uberoi, Phys. Fluids 15, 1673 (1972).
95
(2002). D. A. D’Ippolito and J. P. Goodbloed, Plasma Phys. 22, 1091 (1980).
96
59
A. Strelsov and W. Lotko, J. Geophys. Res. 101, 5343 doi:10.1029/ C. Z. Cheng, L. Chen, and M. S. Chance, Ann. Phys. 161, 21 (1984);C. Z.
95JA03762 (1996). Cheng and M. S. Chance, Phys. Fluids 29, 3695 (1986).
97
60
C. Mitchell, “Alfven Eigenmodes and Field Line Resonances,” Ph.D. dis- H. H. Duong, W. W. Heidbrink, E. J. Strait, T. W. Petrie, R. Lee, R. A.
sertation UCLA, 2001. Moyer, and J. G. Watkins, Nucl. Fusion 33, 749 (1993).
98
61
S. Vincena, W. Gekelman, and J. E. Maggs, Phys. Rev. Lett. 93, 105003 J. E. Maggs and G. J. Morales, Phys. Plasma 4, 290 (1997).
99
(2004). A. T. Burke, J. E. Maggs, and G. J. Morales, Phys. Plasmas 7, 1397 (2000).
100
62
T. H. Stix, Phys. Fluids 1, 308 (1958). S. Vincena and W. Gekelman, Phys. Plasmas 13, 064503 (2006).
101
63
T. H. Stix and R. W. Palladino, in Proceedings of the 2nd UN Interna- A. T. Burke, J. E. Maggs, and G. J. Morales, Phys. Rev. Lett. 81, 3659
tional Conference on the Peaceful Uses of Atomic Energy, Geneva, 1958, (1998); A. T. Burke, J. E. Maggs, and G. J. Morales, Phys. Plasmas 7,
Vol. 31, p. 282. 544 (2000).
102
64
F. I. Bowley, J. M. Wilcox, A. W. DeSilva, P. R. Forman, G. W. Gordon, J. R. Penano, G. J. Morales, and J. E. Maggs, Phys. Plasmas 7, 144
W. H. Hamilton, and C. N. Watson-Munro, Phys. Fluids 6, 925 (1963). (2000).
103
65
J. A. Lehane and F. J. Paoloni, Plasma Phys. 12, 823 1970. D. C. Pace, M. Shi, J. E. Maggs, G. J. Morales, and T. A. Carter, Phys.
66
D. G. Swanson, R. W. Clark, P. Korn, S. Robertson, and C. B. Wharton, Plasmas 15, 122304 (2008).
104
Phys. Rev. Lett. 28, 1015 (1972). M. Shi, D. C. Pace, G. J. Morales, J. E. Maggs, and T. A. Carter, Phys.
67
A. Breun, P. Brooker, D. Brouchous, J. Browning, G. Butz, J. Conrad, E. Plasmas 15, 122304 (2008).
105
Dales, J. Ferron, R. Goulding, N. Hershkowitz, T. Intrator, C. Litwin, R. C. C. Chaston, C. W. Carlson, R. E. Ergun, and J. P. McFadden, Phys.
Majeski, S. Maessick, B. Nelson, L. Peranich, H. Persing, J. Radtke, D. Scr. T 84, 64 (2000).
106
Roberts, G. Severn, D. Sing, E. Wang, D. A. D’Ippolito, J. R. Myra, and C. C. Chaston, J. W. Bonnell, C. W. Carlson, and J. P. McFadden, J. Geo-
G. L. Francis, in 11th IAEA International Conference on Plasma Physics phys. Res. 108, 8003 doi:10.1029/2002JA009420 (2003).
107
and Controlled Nuclear Fusion Research, Kyoto, Japan, 13–20 November E.V. Mishin and M. Forster, Geophys. Res. Lett. 22, 1745 doi:10.1029/
1986, Vol. 2 of Nuclear Fusion Supplement 1987, pp. 263–271. 95GL01442 (1995).
108
E. D. Fredrickson, N. N. Gorelenkov, R. E. Bell, J. E. Menard, A. L. ham, M. Tatarakis, Z. Najmudin, W. Rozmus, R. G. Evans, M. G. Haines,
Roquemore, S. Kubota, N. A. Crocker, and W. Peebles, Nucl. Fusion 46, A. E. Dangor, and K. Krushelnick, Phys. Rev. Lett. 97, 255001 (2006).
132
S926 (2006). P. M. Nielson, L. Willingale, M. C. Kaluza, C. Kamperidis, S. Minardi,
109
L. Chen, Z. H. Lin, and R. White, Phys. Plasmas 8, 4713 (2001). M. S. Wei, P. Fernandes, M. Notley, S. Bandyopadhyay, M. Sherlok, R.
110
D. A. Gates, N. N. Gorelenkov, and R. B. White, Phys. Rev. Lett. 87, J. Kingham, M. Tatarakis, Z. Najmudin, W. Rozmus, R. G. Evans, M. G.
205003 (2001). Haines, A. E. Dangor, and K. Krushelnick, Phys. Plasmas 15, 092701
111
R. H. Kraichnan, Phys. Fluids 8, 1385 (1965). (2008).
112 133
C. S. Ng and A. Bhattacharjee, Astro Phys. J. 465, 845 (1996). E. N. Parker, J. Geophys. Res. 62, 509 doi:10.1029/JZ062i004p00509
113
J. V. Shebalin, W. H. Matthaeus, and D. Montgomery, J. Plasma Phys. (1957).
134
29, 525 (1983). W. Gekelman, R. L. Stenzel, and N. Wild, J. Geophys. Res. 87, 101
114
P. Goldreich and S. Sridhar, Astro Phys. J. 485, 680 (1997). doi:10.1029/JA087iA01p00101 (1982).
115 135
G. G. Howes, W. Dorland, S. C. Cowley, G. W. Hammett, E. Quataert, C. D. Gregory, J. Howe, B. Loupias, S. Myers, M. M. Notley, Y. Sakawa,
A. A. Schekochihin, and T. Tatsuno, Phys. Rev. Lett. 100, 065004 (2008). A. Oya, R. Kodama, M. Koenig, and N. C. Woolsy, AstroPhys. J. 676,
116
S. D. Bale, P. J. Kellogg, F. S. Mozer, T. S. Horbury, and H. Reme, Phys. 420 (2008).
136
Rev. Lett. 94, 215002 (2005). B. Roberts, Physics of Magnetic Flux Ropes, Geophysical Monograph
117
A. Hasegawa and L. Chen, Phys. Rev. Lett. 36, 1362 (1976). Vol. 58, edited by C. T. Russell, E. R. Priest, and L. C. Lee (AGU Wash-
118
J. V. Hollweg, J. Geophys. Res. 99, 23431 doi:10.1029/94JA02185 ington, DC, 1990), p. 113.
137
(1994). J. Chen, R. A. Howard, G. E. Brueckner, R. Santaro, J. Krall, S. E. Pas-
119
D. W. Auerbach, T. A. Carter, S. Vincena, and P. Popovich, Phys. Rev. waters, O. C. St. Syr, R. Schwenn, P. Lamy, and G. M. Simnett, Astro-
Lett. 105, 135005 (2010). Phys. J. Lett. 490, L191 (1997).
120 138
B. C. Low, J. Geophys. Res. 106, 25142 doi:10.1029/2000JA004015 J. T. Gosling, Physics of Magnetic Flux Ropes, Geophysical Monograph
(2001). Vol. 58, edited by C. T. Russell, E. R. Priest, and L. C. Lee (AGU Wash-
121
S. Colgate, J. Geophys. Res. 70, 3161 doi:10.1029/JZ070i013p03161 ington, DC, 1990), p. 343.
139
(1965). W. Gekelman, J. Maggs, and H. Pfister, IEEE Trans. Plasma Sci. 20, 614
122
P. A. Bernhardt, Phys. Fluids B 4, 2249 (1992). (1993).
123 140
G. Haerendel, G. Paschmann, W. Baumjohann, C. W. Carlson, Nature J. B. Taylor, Phys. Rev. Lett. 33, 1139 (1974).
141
320, 720 (1986). T. Intrator, X. Sun, G. Lapenta, L. Dorf, and I. Furno, Nat. Phys. 5, 521
124
M. A. Van Zeeland and W. Gekelman, Phys. Plasmas 11, 320 (2004);A. (2009).
142
Collette, W. Gekelman, A. Collette, and W. Gekelman, Phys. Rev. Lett. E. Lawrence and W. Gekelman, Phys. Rev. Lett. 103, 105002
105, 195003 (2010). (2009).
125 143
M. A. Van Zeeland, W. Gekelman, S. Vincena, and G. Dimonte, Phys. C. H. Mandrini, P. Démoulin, L. Driel-Gesztelya, B. Schneider, G.
Rev. Lett. 87, 105001 (2001). Cauzzi, A. Hoffan, Sol. Phys. 168, 115 (1996).
126 144
W. Gekelman, M. A. Van Zeeland, S. Vincena, and P. Pribyl, J. Geophys. E. R. Priest and P. Démoulin, J. Geophys. Res. 100, 23443 (1995);
Res. 108, A781 doi:10.1029/2002JA009741 (2003). P. Démoulin, Adv. Space Res. 27, 1269 (2006).
127 145
S. Vincena, W. Gekelman, M. A. Van Zeeland, J. Maggs, and A. Collette, V. S. Titov, G. Horning, and P. Démoulin, J. Geophys. Res. 107, 1164
Phys. Plasmas 15, 072114 (2008). doi:10.1029/2001JA000278 (2002).
128 146
M. A. Van Zeeland, W. Gekelman, S. Vincena, and J. Maggs, Phys. Plas- V. S. Titov, T. Forbes, E. Priest, Z. Mikic, and J. Linker, Astro Phys. J.
mas 10, 1243 (2003). 693, 1029 (2009); P. Démoulin, Adv. Space Res. 37, 1269 (2006).
129 147
W. Gekelman, A. Collette, and S. Vincena, Phys. Plasmas 14, 062109 M. Berger, Geophys. Astrophys. Fluid Dyn. 30, 79 (1984).
148
(2007). W. Gekelman, E. Lawrence, A. Collette, B. Van Compernolle, P. Pribyl,
130
R. Ruytov, R. P. Drake, J. Kane, E. Laing, B. A. Remington, and W. M. M. Berger, and J. Campbell, Phys. Scr. T 142, 014032 (2010).
149
Wood-Vasey, AstroPhys. J. 581, 821 (1991). S. Vincena, W. Gekelman, and J. Maggs, Phys. Plasmas 8, 3884 (2001).
131 150
P. M. Nielson, L. Willingale, M. C. Kaluza, C. Kamperidis, S. Minardi, M. See supplementary material at http://dx.doi.org/10.1063/1.3592210 which
S. Wei, P. Fernandes, M. Notley, S. Bandyopadhyay, M. Sherlok, R. J. King- contains stereo image pairs, captions, and instructions for viewing.

The Many Faces of Shear Alfve N Waves

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Many Faces of Shear Alfve N Waves

Uploaded by

Copyright:

Available Formats

PHYSICS OF PLASMAS 18, 055501 (2011)

The many faces of shear Alfvén wavesa)

1070-664X/2011/18(5)/055501/26/$30.00 18, 055501-1 C 2011 American Institute of Physics

FIG. 2. From Lenhardt (1952). Liquid Sodium apparatus for studying

into account in theoretical comparisons. If one raises the

Possible magnetic field geometries are uniform, linearly

V. THE MORPHOLOGY OF SHEAR WAVES

FIG. 8. (Color) The measured current

FIG. 10. (Color) (left) Schematic dia-

space plasma community motivated the initial studies. An

on the LAPD in which the electron beta, be ¼ M 8pnTe

FIG. 13. (Color) Isosurfaces of current

VIII. ALFVÉN WAVE MASER

FIG. 16. Side view of the original LaPD

FIG. 18. (Color) Shear Alfvén wave

the random occurrence of Lorentzian-shaped pulses in the

FIG. 23. (Color) (top panel) Contours of fluctu-

bation then scatters the Alfvén waves, resulting in the crea-

of a field line a small distance from its original position (x,

FIG. 32. (Color) Magnetic field lines of

FIG. 34. (Color) Shear wave Hydra.

dt liams, Hans Boehmer, Dennis Papadopoulis, Yuhou Wang,

You might also like