Electron magneto-hydrodynamic waves bounded by magnetic bubble

V. P. Anitha, D. Sharma, S. P. Banerjee, and S. K. Mattoo

Citation: Phys. Plasmas 19, 082118 (2012); doi: 10.1063/1.4748573

View online: http://dx.doi.org/10.1063/1.4748573
View Table of Contents: http://pop.aip.org/resource/1/PHPAEN/v19/i8
Published by the American Institute of Physics.

Related Articles
Nonlinear dissipation of circularly polarized Alfvén waves due to the beam-induced obliquely propagating waves
Phys. Plasmas 19, 082317 (2012)
Verification and validation of linear gyrokinetic simulation of Alfvén eigenmodes in the DIII-D tokamak
Phys. Plasmas 19, 082511 (2012)
Non-local gyrokinetic model of linear ion-temperature-gradient modes
Phys. Plasmas 19, 082307 (2012)
Resonant wave-particle interactions modified by intrinsic Alfvénic turbulence
Phys. Plasmas 19, 082902 (2012)
Parametric decays in relativistic magnetized electron-positron plasmas with relativistic temperatures
Phys. Plasmas 19, 082104 (2012)

Additional information on Phys. Plasmas

Journal Homepage: http://pop.aip.org/
Journal Information: http://pop.aip.org/about/about_the_journal
Top downloads: http://pop.aip.org/features/most_downloaded
Information for Authors: http://pop.aip.org/authors
 PHYSICS OF PLASMAS 19, 082118 (2012)

Electron magneto-hydrodynamic waves bounded by magnetic bubble

V. P. Anitha, D. Sharma, S. P. Banerjee,a) and S. K. Mattoo
Institute for Plasma Research, Bhat, Gandhinagar 382428, India

(Received 8 June 2012; accepted 14 August 2012; published online 29 August 2012)
The propagation of electron magneto-hydrodynamic (EMHD) waves is studied experimentally in a
3-dimensional region of low magnetic field surrounded by stronger magnetic field at its
boundaries. We report observations where bounded left hand polarized Helicon like EMHD waves
are excited, localized in the region of low magnetic field due to the boundary effects generated by
growing strengths of the ambient magnetic field rather than a conducting or dielectric material
boundary. An analytical model is developed to include the effects of radially nonuniform magnetic
field in the wave propagation. The bounded solutions are compared with the experimentally
obtained radial wave magnetic field profiles explaining the observed localized propagation of
waves. VC 2012 American Institute of Physics. [http://dx.doi.org/10.1063/1.4748573]

I. INTRODUCTION impart various characteristics of the waves such as conserva-

tion of helicity9 and lack of observable nonlinearity8 of these
Whistler waves are electron magneto-hydrodynamic
waves even at very large amplitudes ðB~ > B0 Þ. Interest in
(EMHD) waves1,2 propagating in a magnetized plasma over
helicon waves stem mainly from the unusually high ioniza-
a frequency range xci < x < xce . Unlike the case of an elec-
tion efficiency and application in producing high density dis-
tromagnetic wave propagation in vacuum, where the dis-
charges.3–5 Several experiments have been carried out with
placement current j~d is responsible for the regeneration of
~ conduction currents proportional to excitation antennas designed for both m ¼ þ1 and m ¼ 1
wave magnetic field B,
modes, aimed at launching RH and LH polarized waves,
the first order E  B0 drift of the electrons, where E~ is the
~
respectively. A preferential excitation of R-Helicons, even
wave electric field and B0 ¼ Bplasma is the ambient field in
with antennas configured to excite L waves, has been
the plasma, play a dominant role in deciding the whistler
reported5 and continues to remain as one of the experimental
wave propagation in plasma. While j~d , in vacuum, can
puzzles.
assume appropriate directions to allow both right handed
Propagation of EMHD waves across the regions of van-
(RH) and left handed (LH) circular polarization of the wave,
ishing magnetic field, or a magnetic null, is observed to pro-
j~EB
~ for a whistler can only support right handed (RH) circu-
duce many interesting changes in the wave propagation.10–12
lar polarization, as it assumes a wrong direction3 to cause
These include the helicity reversal, drastic reduction in the
regeneration of B~ for LH propagation. In a plasma bounded
wave energy, and nonlinear effects that are otherwise absent8
by either conducting or dielectric radial boundaries;3–5 how-
because of force-free nature of whistler waves. These obser-
ever, the condition that the wave magnetic field must neces-
vations have been generally attributed to the departure from
sarily vanish at the boundary causes j~r to vanish resulting in
EMHD conditions around the localized null point where
parallel (to B0 ) currents j~jj to flow at the boundary. Due to
x > xce . The extent of the null point or the region where the
the finite inductance of the plasma, this j~jj causes some
x > xce holds good, however has been limited in compari-
charge pileup, which in turn builds up a radial space charge
son with the extent of the propagating structure. Understand-
distribution. The resulting electrostatic field changes the na-
ing of how such a null point influences a propagating EMHD
ture of the whistler wave and permits both RH and LH polar-
structure with dimensions larger than its extent remains far
izations to occur3 for these bounded waves, called Helicons.
from complete. All that is known is that the excited vortices
Experiments on excitation and propagation of whistlers
are EMHD vortices before the null region, and remain so
and Helicons in a plasma embedded in a uniform background
even after traversing the null point, but with a reversed helic-
magnetic field have been subjected to extensive laboratory
ity and reduced energy.
investigations.2–9 As for whistlers, excitation and propaga-
At this point, it needs to be emphasized that a null point
tion has been carried out in large plasma systems2,6–9 devoid
embedded in the background DC magnetic field of a plasma
of boundary effects. These experiments, apart from establish-
invariably has, associated with it, large magnetic field gra-
ing the formation of these EMHD waves through dispersion
dients prevailing around the point where B0 approaches zero.
and propagation characteristics, have also revealed their
The EMHD structure approaching the null point, from a
complex field topology. Whistlers are shown to be force-free
~ ¼ 0, where E~ is the electric field excited region of radially uniform magnetic field (where x < xce ),
with j  B0  neE
therefore passes through a transition region where there is a
by Hall effect. The force free nature has been projected to
considerable radial magnetic field gradient of scale length
comparable or smaller than the wave extent, and where the
a)
Present address: Department of Electrical and Computer Engineering, Uni- field values are still large enough to retain the EMHD condi-
versity of Alberta, Edmonton, Alberta, Canada T6G 2V4. tions valid. The influence of this transition region with

1070-664X/2012/19(8)/082118/8/$30.00 19, 082118-1 C 2012 American Institute of Physics

V
082118-2 Anitha et al. Phys. Plasmas 19, 082118 (2012)

strongly changing topology of the ambient magnetic field on

a propagating wave has not been analyzed critically. In other
words, it is not very clear if the vortex has already undergone
a characteristic change even before reaching the localized re-
gime where EMHD conditions are broken down. This transi-
tion region that the wave encounters prior to meeting the
null point may, owing to its complex field topology, have a
prolonged, nonlocal, influence on the propagating waves,
thereby modifying the wave characteristics even before the
wave reaches the null point. In conventional experimental
set ups, having small dimension null regions, it is, however,
not possible, even with miniaturized probes (5 mm  2 mm),
to probe this region with sufficient accuracy and carry out a
dispersive analysis of the propagating wave.
In our present observation, we unearth the above referred FIG. 1. Schematic of the set up in the large volume plasma device. The
transition region on a propagating wave, to understand the square compensation coils (a) and (b) are used to cancel the field compo-
nents By and Bx , respectively. The Helmholtz coil installed inside the system
underlying physical processes modifying the wave character- compensates the component Bz .
istics in greater detail. We have set up an arrangement in the
large volume plasma device (LVPD),13 where the transition
region discussed above has been reproduced and has been uniform field condition, a set of 10 external magnet coils
kept sufficiently large so that a complete dispersive analysis have been appropriately charged to produce a uniform
of the waves traversing this region can be carried out. We magnetic field in the range 4–20 G, in the z direction (device
reproduce the conditions similar to what a propagating wave axis).
excited as in Ref. 11, encounters before it reaches the null For experiments in the magnetic bubble configuration,
region. We term this region a magnetic bubble as it consists of the applied uniform field Bapplied as well as the Earths mag-
a central, reasonably uniform ambient field (200 mG) regime netic field (500 mG) is compensated with the help of three
surrounded by radial magnetic field gradients. The field additional Helmholtz coils.16 The Bx ; By compensation coils
remains uniform over an axial extent of 100 cm. The propaga- marked (b) and (a), respectively, in Fig. 1, installed outside
tion of linear EMHD waves (x < xce , B=B ~ 0 up to 0.5) the device and used to cancel the corresponding components
excited in this region has been investigated, and the results are of the Earths field, are fed with DC currents. The sequence
compared with their propagation in a uniform ambient field followed to produce the bubble configuration includes pro-
set up. The findings show transition of the EMHD wave from duction of plasma at uniform magnetic field (here, 4 G) and
a right handed whistler wave to a left handed helicon like then cancellation of this field component, just prior to the
mode up on switching from a uniform ambient field set up to EMHD wave excitation. The operation of this Bz compensa-
the magnetic bubble like configuration. The change of tion coil, installed inside the device, is in pulsed mode.
polarization is observed to be gradual, which further revealed Depending on the experiment, the field can be partially or
additional implications of the ambient magnetic field configu- fully compensated by varying the pulsed current applied to
ration, to the accessibility of—usually hard to excite—plane the Bz compensation coil to produce different values of com-
polarized helicon waves,14 constituted by a superposition of pensating field (Bcomp ) in the range 0–4 G. For complete field
right handed and left handed circularly polarized helicon compensation, the current through the Bz compensation coil
modes. The observed transition of the wave polarization from has a peak 380 A with typical rise time 250 ls and
right to left is accompanied by a drastic change in dispersion a decay of 5 ms.16 The magnetic field produced by this
properties of the excited waves that confirms the excitation of pulsed current has a rise time of 500 ls, followed by a
radially localized helicon like waves. The effective boundary plateau sustained for 200 ls. A typical radial vacuum field
is provided possibly by the magnetic gradients present in the profile of the bubble configuration obtained when the applied
radial direction rather than the material boundary of our magnetic field of 4 G is nearly compensated by switching the
device. The experimental observations and analysis by intro- compensating field on is shown in Fig. 2(a). With this field
ducing the effects of field gradient in the dispersion equation cancellation, an extended region is obtained in the experi-
are presented in this paper. mental volume where the field remains low, 200 mG, at
This paper is organized as follows. Experimental set up r ¼ 0, and uniform over an axial extent of 100 cm. Radial
is presented in Sec. II followed by observations discussed in gradients exist away from the axis (r ¼ 0).
Sec. III, and theoretical analysis based on the observed dis- The pulsed (5 ms) plasma discharge is produced using a
persion properties of the waves presented in Sec. IV. Sum- multifilamentary plasma source.15 The main diagnostics con-
mary and conclusions are presented in Sec. V. stitute Langmuir probes, magnetic probes for measuring the
diamagnetic response of plasma (diameter ¼ 12 cm, 4 turns),
and miniature magnetic probes (diameter 2 mm; length, 5 mm;
II. THE EXPERIMENTAL SET-UP
30 turns) for diagnosing the wave field perturbations. Fig. 3
The experimental set-up, shown in Fig. 1, is discussed shows temporal variation of the discharge parameters corre-
in detail in earlier reports.13,15,16 For experiments in the sponding to experiments in the uniform field case (left), i.e.,
082118-3 Anitha et al. Phys. Plasmas 19, 082118 (2012)

and the anode. Once the arc voltage is switched off at 5 ms Id

falls to zero in about 20 ls. The plasma confinement time is
1 ms, observable from the exponential decay of the density.
In the afterglow that initiates at the time when Id ¼ 0, the
floating potential has a sharp fall over 100 ls, followed by a
slow decay in the same manner as exhibited by the density.
This sharp fall is attributed to the loss of energetic primary
electrons once the discharge voltage is switched off.13
Once the discharge is turned on, the magnetic field
reduces gradually from its vacuum field value (Bapplied ) due
to plasma diamagnetism. To obtain the ambient field B0 in
the plasma, the method followed includes detecting and inte-
grating the diamagnetic loop voltage yielding the change in
field Bdia due to diamagnetism, which is subtracted from the
applied field giving B0 ¼ Bapplied  Bdia . Shown in Fig. 3(d)
is the temporal profile of B0 . In case when the field compen-
sation coils are not charged (left), there is a fast recovery,
over 100 ls, of B0 to the vacuum field in the afterglow
phase, that follows the loss rate of the energetic electrons.
The wave excitation experiments are carried out at about
400 ls in the afterglow plasma, where the background
plasma density 2:5  1011 cm3 and the temperature
1 eV show nearly uniform spatial profiles over the experi-
mental volume.13,15 For experiments in the bubble configura-
tion (right), the field compensation is carried out just prior to
the switch off of the discharge (at 4.9 ms) so that the desired
low field plateau (marked by a circle), obtained 500 ls later,
is devoid of energetic electrons. A period of 10 ls is cho-
sen within this 200 ls plateau, for EMHD excitation. The
value of B0 in this case is Bapplied  Bdia  Bcomp . As dis-
FIG. 2. (a) Radial variation of the ambient magnetic field B0 for the bubble
configuration, as measured in the experiment (*) and the functional fit given cussed in our earlier report,16 during this period, the magni-
by Eq. (10) (solid line) and (b) measured radial distribution of plasma den- tude and profile of the density (shown in Fig. 2(b)). And
sity in the bubble configuration during wave excitation. temperature are not affected by the field compensation. The
antenna (wire loop of 10 cm diameter) for the wave excita-
when the Bz compensation coil is not turned on, and in the tion is triggered during the 200 ls period of the low field
bubble configuration (right) where the compensation coil is plateau to produce small amplitude perturbations in the fre-
turned on at an appropriate time. Here, t ¼ 0 denotes the quency range 0.09–0.9 xce .
switching on of the arc voltage, applied between the filaments

III. THE EXPERIMENTAL RESULTS

The axial propagation of the excited waves is studied
using time of flight measurements of the measured field per-
turbations. Fig. 4 shows the propagation of wave perturba-
tions observed when B0 is uniform, ¼ 4 G (left). The
magnetic probe data were taken at various z locations (at
r ¼ 2 cm). The velocity of propagation estimated is
1:5  107 ms1 , which matches well with that calculated
for low frequency whistler waves. Although not shown here,
for a range of uniform 4–20 G, the phase velocity is found to
match well with the whistler velocity. Plots on the right cor-
respond to wave perturbations obtained in the bubble config-
uration. The phase velocity obtained in this case is
0:5  107 ms1 , which is about twice that expected for the
whistlers, estimated for the field B0 of 200 mG.
FIG. 3. Temporal profiles of Id (a), n (b), Vf (b), and B0 (d) for the cases Fig. 5 shows x as a function of kjj measured for different
with (right) and without (left) the Bz compensation. The arrow points to the
time (4.9 ms) when the Bz compensation coil is triggered. The plateau re-
ambient field conditions. The dashed curves represent the
gime encircled in curve (d) (right) denotes the 200 ls period in the afterglow theoretical dispersion relation for whistler propagation in an
plasma where the magnetic bubble configuration is formed. infinite cold plasma, given by1
082118-4 Anitha et al. Phys. Plasmas 19, 082118 (2012)

field and bubble field configurations, the polarization of the

waves in the two situations was measured by observing the
phase difference between the Bx and By perturbation. Shown
in Fig. 6 are the Bx and By perturbations for uniform field
(4 G) and the bubble geometry. For the former, the polariza-
tion, as expected, is found to be right handed, with By lead-
ing Bx . For the entire range (4–20 G) of the applied uniform
field value, the waves exhibit right handed polarization, as
long as the field is maintained uniform. Interestingly, a rever-
sal of polarization is observed for the waves excited in the
bubble field configuration, with a left handed polarization
indicated by Bx leading By .
In order to understand the nature of transformation of
the waves from right handed to left handed polarization in a
gradually changing magnetic field topology, experiments
were conducted by gradually increasing the compensation
field. Here, experiments were done in different plasma shots
applying different values of the compensating field in each
case, the two extreme situations being uniform B0 (4 G) and
the bubble configuration with B0 of 200 mG at the radial cen-
ter. Fig. 7 shows representative cases of wave polarization
FIG. 4. Temporal profiles of perturbed magnetic field at various axial loca- (left) for the reducing value of B0 . Plotted on the right in Fig.
tions z in cases where B0 is (left) 4 G with phase velocity 1:5  107 m/s and
(right), where this field is compensated using Helmholtz coils and phase ve-
7 is the observed phase difference D/ between the compo-
locity is 0:5  107 m=s. nents Bx and By with changing ambient field.
It is clear that the polarization change occurs gradually,
k2 c2 =x2 ¼ 1  ½x2p =xðx  xce Þ; (1) with a purely right handed (D/  90o ), wave excited at uni-
form field of 4 G, changing over to a left handed wave
(D/  70o ), in the bubble set up with B ¼ 200 mG at the
which is found to be in good agreement with the measured
axial location. We expect that a purely left handed mode
data for the cases with uniform ambient field (Figs. 5(a)–
with D/ ¼ 90o could have been possible with more effec-
5(c)). However, a clear disagreement can be seen for the
tive field cancellation, which was not attempted due to limi-
case with bubble configuration of the field (Fig. 5(d)).
tation in the pulse voltage applied to the Bz Helmholtz coil.
In order to resolve the observed ambiguity in the disper-
Interestingly, the wave excited at B  1:3 G is nearly plane
sion behavior of the wave in two cases, i.e., with uniform
polarized, or azimuthal standing waves, where both By and
Bx are found to be in same phase, and is similar to what is
reported in few earlier experiments using conventional
bounded set up.17,18
The radial profile of the wave magnetic field, Bz , recov-
ered typically for the field compensated case is plotted
in Fig. 8 for two cases of excitation frequencies. For

FIG. 5. Dispersion relation for the excited EMHD waves. The experimental FIG. 6. The polarization of the excited waves when the ambient field is uni-
k values (shown as circles) for excited frequencies x and function (1) (plot- form (top) and in the bubble configuration (bottom). A reversal of polariza-
ted as solid line) are shown for ambient magnetic field values, 20 G (a), 10 G tion from RH (top) to LH (bottom) is observed with the wave field By
(b), 4 G (c), and the field compensated configuration (d). leading Bx in the former case and lagging in the later case.
082118-5 Anitha et al. Phys. Plasmas 19, 082118 (2012)

FIG. 7. The wave fields (left) Bx and By for different

values Bcomp . The topmost plot with uniform B0
(with Bcomp ¼ 0) show RH wave polarization
which gradually changes to LH with increasing
Bcomp , and thus, increasing radial field gradient. The
phase difference as recovered for various values of
B0 is also plotted (right).

comparison, a typical profile obtained for whistler wave handed mode with a change of magnetic field configuration
when excited in B0 ¼ 4 G is also shown. The perturbation has from uniform radial profile to a bubble geometry. A propa-
a minimum at r ¼ 0, for the bubble configuration, in contrast gating L polarized wave below xce is unexpected as it is not
with a corresponding maximum obtained in case of the whis- supported theoretically in case of an unbounded plasma, for
tler. Further, this mode is observed to be well localized (with the parameters under consideration. The only kind of waves
in 15 cm) close to the axis region, dropping to vanishingly that can exhibit left polarization in this parametric regime
small values at much smaller r than the radial extent of the are helicon waves that are modes of a bounded system. The
whistler waves. other two factors that point towards the helicon like behavior
of the wave are vanishing wave field at the axis and its lim-
IV. DISCUSSION ited radial extension as compared to the whistler wave field.
In order to investigate the agreement of our observations
Observations discussed above show a gradual change in with helicon wave behavior, the xðkjj Þ dependence expected
D/ from a purely right handed mode to a predominant left for m ¼ 1 helicon waves has been generated, using the heli-
con dispersion relation,4

J1 ðTaÞ ¼ ðkjj aT=2aÞðJ2  J0 Þ; (2)

where Ji is the Bessel function, T is perpendicular wave vec-

tor, and a ¼ xx2p =kjj xc c2 . Fig. 9 shows the xðkjj Þ curves
obtained using the above relation, assuming B0 ¼ 200 mG
and m ¼ 1 for different values of radii assumed for the
boundary, a. The experimental xðkjj Þ data already discussed
(Fig. 5(d)) are re-plotted here with “þ” for comparison. It
can be seen that the solution (a) obtained using the antenna
radius (a ¼ 5 cm) does not agree with the experimental
observations. Upon using a range of a values, the curve is
found to modify and saturate approaching the curve (c) in
the range of a from 40 cm to 100 cm (device wall). Note that
the dispersion curves (c) show qualitative agreement with
experimental values (plotted with “þ”) which could be
improved further by choosing a value of a close to the exper-
imentally observed wave extent. For example, the curve
(b) plotted using a ¼ 15 cm, corresponding to the case of
x ¼ 6  105 rad s1 presented in Fig. 8, agrees well with the
experimental dispersion data.
FIG. 8. The wave magnetic field for bubble configuration with
x ¼ 2:5  106 rad=s (a), 0:62  106 rad=s (b) and for a representative case The above approach is, however, too approximate as the
of whistler wave in a uniform ambient magnetic field. observed radial extension a of the wave depends on the
082118-6 Anitha et al. Phys. Plasmas 19, 082118 (2012)

FIG. 10. Analytic wave magnetic field obtained from solutions of Eq. (3)
assuming uniform ambient field of 200 mG (a) and 250 mG (b), for same
FIG. 9. Comparison of the experimental dispersion data (þ) for the wave value of x and kjj .
excited in the magnetic bubble configuration, with solutions to Eq. (2) for
values of boundary a assumed as the antenna radius (a), observed wave
extent 15 cm (b) and the material boundary of the device (c).
@B
or ¼ ½ðj  rÞB0 þ ðB0  rÞj=en0 ; (5)
frequency x and an ambiguity exists in selecting a single @t
value of a for generating the full dispersion data comprising
where
a range of x values. Alternatively, the unbounded numerical
solutions of the standard radial equation for helicon waves4 r  B ¼ l0 j; (6)
 
@ 2 Bz 1 @Bz 2 m2 r  j ¼ 0; (7)
þ þ T  Bz ¼ 0; (3)
@r2 r @r r2
and
with an approximate value of the B0 , assumed uniform in
(3), are obtained and found to be finite up to the limits well j? ¼ en0 E  B0 =B20 : (8)
beyond the experimental radial extension of the wave, unless
a physical boundary a is imposed at these small r values. Note that accounting for the term ðj  rÞB0 for a monotoni-
Such unbounded solutions are shown for two values of cally varying B0 ðrÞ and using r  B ¼ l0 j results in the
B0 ¼ 200 and 250 mG in Fig. 10 which correspond to values condition,
of B0 at the center and at the typical radial extension of the
wave field in our experiment, respectively. We additionally B ¼ a1
c r  B; (9)
note that the radial scale length of the solutions is consider-
ably sensitive to the choice of B0 . where ac ¼ aR þ iaI is a complex quantity having finite
Hence, the standard Helicon model assuming a radially implications on the existence of the wave in the regions of
uniform ambient field B0 , chosen for the above analysis, is larger rB. The form of Eq. (5) additionally suggests that the
not readily applicable to the present set up where a radial wave field must decay radially over a spatial scale equivalent
variation of the ambient field exists. In order to include the to that of the radial growth in the ambient field B0 .
confining effect of radially varying B0 in the model, we Considering now that for the present experiment the
attempt to obtain a modified version of Eq. (3) in the follow- radial dependence of B0 is suitably represented by the
ing analysis. approximation,

V. THE MODIFIED HELICON MODEL B0 ðrÞ ¼ A1 þ A2 r4 ; (10)

Since the helicon waves in their simplest form are
obtained from where A1 ¼ 2  105 T and A2 ¼ 1:5  103 T cm4 (as
plotted in Fig. 2 where the dots correspond to experimental
@B values of B0 ), the ratio of two terms in the right hand side of
¼ r  E ¼ r  ðj  B0 Þ=en0 (4)
@t the Eq. (5) can be approximated as
082118-7 Anitha et al. Phys. Plasmas 19, 082118 (2012)

FIG. 11. Radial profile of the wave magnetic

field as obtained from the experiment (þ), and
from the numerical solutions of the modified ra-
dial equation Eq. (13) (solid line). Plots (a)–(f)
correspond to the increasing values of x pre-
sented in Fig. 5(d), respectively.

ðj  rÞB0 @B0 =@r jr values where the applied magnetic field grows stronger in
 : (11)
ðB0  rÞj B0 kjj jjj the experiment. This is in contrast with the solutions of Eq.
(3) (Fig. 10) which, for same set of parameters, are still non-
Further, in the limit of small m, the Eq. (7) implies zero for the values of r that well exceed the system dimen-
kjj jjj  kr jr , hence sions. In order to verify this experimentally, a full radial scan
of wave magnetic field Bz ðrÞ was carried out for various fre-
ðj  rÞB0 @B0 =@r quencies x. The correction produced by the introduction of
 : (12)
ðB0  rÞj B 0 kr radial variation of the applied magnetic field is evident as the
experimentally observed wave field profiles (plotted with
Clearly, for kr sufficiently larger than ð@B0 =@rÞ=B0 or when “þ” in Fig. 11) which are localized in a region of smaller ra-
the wave magnetic field is well localized with in the region dius show a good agreement with the solutions of Eq. (13).
of small gradient, as in the present case, retaining just the The deviation of numerical solutions from the experimental
term ðB0  rÞj in Eq. (5) and using the analytical form of B0 observations at larger r, where the magnetic field vanishes in
given by (10) introduces reasonable effect of finite radial the experiment at a certain radial location, can be attributed
variation of the applied magnetic field, producing a modified to the absence of the term ðj  rÞB0 which is excluded in
version of the original Eq. (3), obtaining the Eq. (13). The contribution ðj  rÞB0 in the
  wave magnetic field must reduce as 1=r with the radius r for
@ 2 Bz 1 @Bz 2 m2 the present set up considering that the current density j /
þ þ T ðrÞ  2 Bz ¼ 0; (13)
@r 2 r @r r B1
0 and the analytical form of applied magnetic B0 ðrÞ given
by Eq. (10).
where T(r) is function of r given by As discussed in Sec. I, the existence of an m ¼ 1 left
handed circularly polarized mode involves inductive
T 2 ðrÞ ¼ a2 ðrÞ  kjj2 (14) effects associated with dominant parallel current at the
boundary, a mechanism different from that behind the
and usual m ¼ þ1, right handed circularly polarized helicon
x l0 en0 wave. The above analysis indicates a reasonable possibility
aðrÞ ¼ : (15) that the stronger ambient magnetic field at the larger radii
kjj B0 ðrÞ
is capable of suppressing the perpendicular current density
j? ¼ en0 E  B=B2 and providing a current closure at a
The asymptotic solutions of the modified model for the wave
critical field value via stronger parallel currents generated
field are obtained using the measured values of Bz and
in accordance with r  j ¼ 0.
@Bz =@r at the axis, and treating kjj as parameter to solve
Eq. (13) as an eigen value problem, imposing the boundary
VI. SUMMARY AND CONCLUSIONS
condition that wave field must vanish at larger radii.
The numerical solutions for the values of kjj measured in To summarize, we presented results of experiments with
the experiments are presented in Fig. 11 (solid line), which propagation of EMHD waves in an axially extended low
show the tendency to vanish within the relevant range of r magnetic field region surrounded by stronger field at the
082118-8 Anitha et al. Phys. Plasmas 19, 082118 (2012)

boundaries providing a limiting radial boundary for the where the constituent left and right circularly polarized
waves. A gradual shift in the polarization of the excited waves must satisfy the boundary condition at the common
waves is observed with a change in the field configuration radial boundary are hard to construct. This makes plane
from uniform B0 to bubble like, which indicates a transition polarized modes to be generally inaccessible in the cases of
where a right handed wave gives way to a left handed one. conventional conducting or dielectric physical boundaries.14
The excited waves show characteristics of bounded waves In this aspect, we like to point out that the possible role of a
where measured magnetic field perturbations are found to be magnetic boundary, as detected in the present set up, in
radially localized. Dispersion relation of the observed waves allowing the two constituent modes to coexist with different
give reasonable agreement with the conventional helicon dis- radial boundaries is an interesting effect which needs to be
persion only when the value of the boundary a is assumed explored in further detail.
much less compared to the boundary corresponding to the
device wall. An analytical model for the existence of helicon 1
R. A. Helliwell, Whistlers and Related Ionospheric Phenomena (Stanford
like waves with finite boundary effects introduced by the University Press, 1965).
2
radially changing ambient magnetic field is developed. The R. L. Stenzel, J. Urrutia, and C. Rousculp, Phys. Fluids B 5(2), 325
radial solutions of modified bounded model, obtained using (1993).
3
F. F. Chen, Phys. Plasmas 3(5), 1783 (1996).
vanishing wave-field as radial boundary condition, are com- 4
F. F. Chen, Plasma Phys. Controlled Fusion 33(4), 339 (1991).
pared with the experimentally obtained radial wave magnetic 5
M. Light and F. F. Chen, Phys. Plasmas 2(4), 1084 (1995).
6
field profiles explaining the propagation of waves with wave R. L. Stenzel, J. M. Urrutia, and C. L. Rousculp, Phys. Plasmas 2(4), 1114
field localized close to the axis of the device. The analysis (1995).
7
R. L. Stenzel, Phys. Fluids 19, 857 (1975).
indicates that finite boundary effects can be generated by the 8
J. M. Urrutia and R. L. Stenzel, Phys. Plasmas 3(7), 2589 (1996).
radial magnetic field gradients where the dispersion charac- 9
C. L. Rousculp, R. L. Stenzel, and J. M. Urrutia, Phys. Plasmas 2(11),
teristics of the wave tend to change strongly with the radius 4083 (1995).
10
R. L. Stenzel and J. M. Urrutia, Phys. Rev. Lett. 81(10), 2064 (1998).
r. The propagation of the wave in the experiments may also 11
M. C. Griskey and R. L. Stenzel, Phys. Plasmas 8(11), 4810 (2001).
be limited to small radial values owing to conventional heli- 12
J. M. Urrutia, R. L. Stenzel, and M. C. Griskey, Phys. Scr. T84, 117
con model being modified to have a complex coefficient a in (2000).
13
certain parameter regimes of the wave propagation with radi- S. K. Mattoo, V. P. Anitha, L. M. Awasthi, G. Ravi, and LVPD Team,
Rev. Sci. Instrum. 72, 3864 (2001).
ally increasing magnetic field. Existence of left handed cir- 14
F. F. Chen, Plasma Sources Sci. Technol. 6, 394 (1997).
cular and plane polarized modes, that are unlikely in cases of 15
L. M. Awasthi, G. Ravi, V. P. Anitha, P. K. Srivastava, and S. K. Mattoo,
a whistler wave, indicates the presence of a magnetic bound- Plasma Sources Sci. Technol. 12, 158 (2003).
16
ary which is further capable of providing attributes necessary S. P. Banerjee, V. P. Anitha, G. Ravi, and S. K. Mattoo, Phys. Plasmas 13,
092503 (2006).
to produce inductive effects that are essential to the existence 17
J. H. Kim, S. M. Yun, and H. Y. Chang, IEEE Trans. Plasma Sci. 24, 1364
of m ¼ 1 left handed circularly polarized helicon like (1996).
18
waves. The analytical solutions for a plane polarized wave A. R. Ellingboe and R. W. Boswell, Phys. Plasmas 3, 2797 (1996).