Plasma-wave interaction in a helicon thruster

sunglowcitrineΠολεοδομικά Έργα

15 Νοε 2013 (πριν από 3 χρόνια και 8 μήνες)

98 εμφανίσεις

Plasma-wave interaction in a helicon thruster
IEPC-2011-047
Presented at the 32
nd
International Electric Propulsion Conference,
Wiesbaden,Germany
September 11{15,2011
Daniel Mart´ınez

and Eduardo Ahedo

Universidad Politecnica de Madrid,Madrid 28040,Spain
The absorption of large wave energy by the plasma is the first important process in
the discharge chamber of a helicon thruster.This paper presents first steps of ongoing
research on the development of a plasma-wave model consistent with the hydrodynamics of
the plasma inside the chamber.The standard plasma resonator model provides important
information about the energy fluxes inside the plasma column for different conditions.
There is a large axial energy flux that spreads the distribution of the absorbed energy map
far from the antenna location.This implies that the boundary condition at the front end
of the vessel requires further research.Other relevant aspects for a consistent plasma-wave
model are pointed out.
Nomenclature
B
0
= Applied magnetic field k

= transversal wavenumber component
E
= Induced electric field k

= longitudinal wavenumber component
B
= Induced magnetic field u = Power density
r
p
= plasma column radius S = Poynting’s vector
r
a
= antenna radius f,g = generic functions
r
w
= wall radius j = Current density
k
= wavenumber vector I = Electric current
I.Introduction
Space plasma thrusters based on helicon sources are a subject of intensive current research.
1{4
One of
the principal projects,and genuinely European,is HPH.COM(Helicon Plasma Hydrazine Combined Micro),
funded by the European Union within the 7th Framework Programme and conducted by a consortium of
15 institutions from 7 European countries.The main objective of the project is to design,test,optimize,
and develop an helicon-based plasma thruster in the range 50-100 watts.In Ref.5 we presented an overview
of the research carried out by our group for this project and for another one,financed by the Air Force
Office for Sientific Research,centered on plasma expansion and detachment in magnetic nozzles.So far,our
activity has covered the internal and external plasma dynamics inside and outside the source,except for
the plasma-wave interaction process,which is crucial for plasma generation and heating.This process is the
objective of the present paper.Section 2 details the physical processes and figures of merit of an helicon
thruster.Section 3 presents the plasma-wave interaction model.Section 4 presents and discusses results.
Conclusions are in Section 5.

MsE student,martinez.ruiz.daniel@gmail.com
y
Professor,eduardo.ahedo.upm.es (web.fmetsia.upm.es/ep2)
1
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
II.Physics of a helicon thruster
The main elements of a helicon thruster are (i) a feeding system that injects gas into a cylindrical vessel,
(ii) a radiofrequency(RF) antenna system that emits radiofrequency waves in the range 1-30 Mhz,and (iii) a
set of magnetic coils surrounding the vessel.The coils create a longitudinal field with three roles.Inside the
chamber,the magnetic field is quasi-axial and facilitates the propagation of helicon waves within the plasma
column.
6
Wave energy is deposited into electrons via plasma-wave resonance.Heated electrons ionize the
injected gas by bombardment.The second role of the magnetic field is to confine the plasma away from the
tube wall,thus avoiding energy losses and producing a highly collimated beam.For instance,Tysk et al.
7
find a 50 times radial decrease of the plasma density (for a 1000G field),instead of the 1.6 times decrease in
a free-acceleration unmagnetized presheath.Strong radial decreases of n
e
are also found by Gilland et al.
8
Outside the source,the magnetic field becomes divergent,thus creating a nozzle effect on the magnetized
plasma.
Beyond gaining insight in the plasma physics,a central aspect of our research has been to assess the
propulsive capabilities of an helicon thruster.In order to compete with other plasma thrusters the helicon
thruster must offer good figures for specific impulse,thrust efficiency,and thrust/weight ratio.The two first
figures-of-merit are directly related to the plasma discharge.First,since the helicon thruster is basically
an electrothermal accelerator,the specific impulse is proportional to the square root of the plasma internal
energy.Therefore a high plasma temperature is needed (a condition that is not required for an helicon
source for non-propulsive applications).Then,good thrust efficiency requires:(i) efficient wave-plasma
energy conversion from the antenna,(ii) near-total plasma ionization inside the vessel,(iii) efficient plasma
heating,with small energy losses to the walls,(iv) efficient conversion of internal energy into directed axial
energy,minimizing plume divergence,and (v) efficient plasma detachment.
In order to study the plasma discharge in a helicon thruster,it is convenient to distinguish the different
stages and processes.There are the production stage inside the plasma source and the acceleration stage
in the magnetic nozzle.Two types of processes take place inside the source:the resonant wave-plasma
interaction,leading to the deposition of wave energy into the plasma and the multiple phenomena governing
internal plasma dynamics.A choked plasma flow is expected around the tube exit.Two other distinguished
processes take place in the magnetic nozzle:the supersonic plasma acceleration and its magnetic interaction
with the thruster,and plasma detachment from the magnetic nozzle.The four main processes are coupled
but,in order to understand the main phenomena and parameters at play,some extra assumptions are
commonly made in order to treat each process independently.Subsequent studies and numerical codes will
look for a deeper and consistent coupling of the four independent models.
The internal radial dynamics for a magnetized plasma were studied in Ref.9.It is found that,within
the regime of interest and for the bulk of the plasma,the radial profile of plasma density is quasi universal
satisfying
n
e
(r,z)
n
0
= J
0
(
a
0
r
R
)
,(1)
with n
0
= n
e
(0,z) the density profile at the plasma axis,R the tube radius,J
0
a Bessel function,and
a
0
≃ 2.405 the first zero of J
0
.Eq.1 is valid except in the thin,inertial and Debye layers next to the
tube lateral wall,but these do not seem important for the analysis of the next section.The internal axial
dynamics,studied in Ref.10,determine n
e
(0,z) [see Fig.6 there].The r-averaged plasma density presents
a maximum where the plasma flow changes from backwards to forwards,and it decreases by a factor of two
at the tube ends.For a high-density plasma,of potential interest in a helicon thruster,effects of the induced
magnetic field modify both the total magnetic field in the chamber and the radial profile,Eq.1,of the
plasma density.
11
This regime is not desirable since helicon waves could not propagate in the plasma.
Plasma processes in the magnetic nozzle include supersonic acceleration,thrust transmission,plasma
detachment,and hypothetical formation of double layers.We have treated these processes in great detail in
several publications,the main ones being.
12{18
The the interaction of the rf wave emitted by the antenna with the plasma is the only process we had
not treated yet,since there was much theoretical research done on it.
6,19{21
Nonetheless,once we have
acquired a reasonable understanding of the internal plasma dynamics,is necessary to attempt advancing in
the plasma-wave interaction problem.There is still much work ahead before achieving a characterization of
the plasma-wave process that can be matched consistently with the plasma fluid dynamics.The rest of the
paper reports on our first steps in that direction.
2
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
O
max
k d
e
k d
e
O
trans
TG
HE
Figure 1.Dispersion relation for the infinite plasma case.
III.Formulation of the plasma-wave model
The most common analysis of the plasma-wave interaction assumes a normal wave with a time dependence
∝ (iωt) and to solve the Maxwell equations
∇×E = −iωB,∇×B = µ
0
[iω
¯
¯ϵ · E +j
a
],(2)
where the dielectric tensor
¯
¯ϵ(r) carries all the plasma information through its density n
e
(r) and collision-
frequency functions,and j
a
is the current density of the rf antenna.If the guide field is along 1
z
,the
dielectric tensor takes the form
¯
¯ϵ = ϵ
0



κ
1
−iκ
2
0

2
κ
1
0
0 0 κ
3



where the three scalar components have been normalized and their expressions are found in textbooks [for
instance,see p.111-112 of Ref.22].The ranges of frequencies of interest for helicon waves is
ω
pe
≫ω
ce
≫ω ≫ω
lh

e
,(3)
where ω
pe
=

e
2
n
e

0
m
e
is the plasma frequency,ω
ce
= eB
0
/m
e
is the electron gyrofrequency,ω
lh

eB/

m
e
m
i
is the lower hybrid frequency,and ν
e
is the effective collision frequency for electrons.
In that frequency region,both the ion motion and the displacement current are negligible and the
dispersion relation for the basic case of a uniform and innite plasma (i.e.n
e
= const) is
k
2
d
2
e
=
ω
±ω
ce
| cos θ| −(ω +iν
e
)
(4)
where k is the wavenumber,θ the angle between B and the wavenumber vector k,and d
e
= c/ω
pe
=

m
e
/(e
2
µ
0
n
e
) is the electron skin-depth.
Figure 1 plots Eq.(4),for ν
e
/ω = 0,in terms of the longitudinal and transversal components of the
wavenumber vector,k

= k cos θ and k

= k sinθ,respectively.The general solution are two pairs of modes,
consisting of long-wavelength helicon waves and short-wavelength Trievelpiece-Gould (TG) waves.If Eq.(4)
is solved for k

d
e
(k

d
e

ce
/ω,ν
e
/ω,ω
ce

pe
),one has that
(1) the two types of waves propagate for 2α < k

d
e
<

α/(1 −α),with α = ω/ω
ce
,
(2) only TG waves propagate for k

d
e
>

α/(1 −α),
3
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
Figure 2.Resonator with single loop antenna of radius r
a
placed at the middle of the chamber z
0
= L=2.
Azimuthal modes are not present since the antenna stands azimuthal symmetry,m= 0.
(3) no wave propagates for k

d
e
< 2α.
Since d
e
∝ n
−1/2
e
and ω
ce
∝ B
0
,plasma density and magnetic field determine the propagation character-
istics of a wave of given frequency ω.The range of k

d
e
where the helicon waves propagate is naturally the
propitious one for the plasma to absorb the rf emission.Since ν
e
/ω ≪1,helicon waves are weakly damped,
but TG waves are strongly damped,which makes them surface waves at the plasma edge.
A.The cylindrical plasma column
We focus now on our working case of a cylindrical,axially-uniform plasma column of length L and radius
r
p
.This column is surrounded by an antenna located at r = r
a
(r
a
> r
p
),and the two of them are encased
in a metallic cylinder of radius r
w
and length L,as sketched in Fig.2.This plasma resonator problem has
been treated extensively in the literature.The analytical studies for uniform plasma density of Shamrai and
Taranov
19,23,24
are particularly illuminating.Nonuniform plasmas have been treated by Arnush and Chen
21
and Cho and Liebermann
25
among others.Their modeling is followed here.
Let be n
e
(r) the plasma density and ν
e
(r) its collision frequency.First,the azimuthal and axial uniformity
of the plasma column allows a Fourier expansion in normal modes ∝ expi(lπz/L+mθ),for l and m integer
numbers.Second,since E
r
and E
θ
are zero at the axial ends (z = 0 and z = L) of the vessel some symmetries
apply in the l-modes.This yields the following expansions
19,25



E
r
(r,θ,z)
E
θ
(r,θ,z)
B
z
(r,θ,z)



=

m,l



E
r,lm
(r)
E
θ,lm
(r)
B
z,lm
(r)



· sin(lπz/L) · exp
imθ
(5)



E
z
(r,θ,z)
B
r
(r,θ,z)
B
θ
(r,θ,z)



=

m,l



E
z,lm
(r)
B
r,lm
(r)
B
θ,lm
(r)



· cos(lπz/L) · exp
imθ
(6)
The antenna current density takes the form
j
a
(r) = Iδ(r −r
a
)[1
z
s
z
(θ,z) +1
θ
s
θ
(θ,z)] (7)
where I is the current along the wire (in amperes,say) and the functions s
z
and s
θ
model the wire geometry.
These functions (expressed in m
−1
,for instance),are to be expanded in Fourier series as
s
k
=
1
L

l,m
I
k,lm
I
exp[i(lπz/L+mθ)],k = z,θ.(8)
For instance for the simple loop sketched in Fig.2,one has s
z
= 0 and
s
θ
(θ,z) = δ(z −z
a
) =
1
L
+
2
L


l=1
[
cos
(
l
πz
L
)
cos
(
l
πz
a
L
)
+sin
(
l
πz
L
)
sin
(
l
πz
a
L
)]
(9)
4
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
Next,we define phase-shifted fields
E
ϕ
= −iE
θ
,B
ϕ
= iB
θ
,(10)
and dimensionless magnitudes
ˆ
I
k
=
I
k
I
ℓ = l
πc

,ˆr =

c
,
ˆ
B =
Bc
µ
0

,
ˆ
E =
E
µ
0

.(11)
Observe that µ
0
Iω/c is the reference magnetic field for an antenna current I.
Then,the Maxwell equations for each normal mode (E
lm
,B
lm
) turn into the four differential equations
d
ˆ
E
ϕ
dˆr
=
(

2
ˆrκ
1

1
ˆr
)
ˆ
E
ϕ
+
ℓm
ˆrκ
1
ˆ
B
ϕ
+
(
1 −
m
2
ˆr
2
κ
1
)
ˆ
B
z
,(12)
d
ˆ
E
z
dˆr
=
κ
2

κ
1
ˆ
E
ϕ

(
1 −

2
κ
1
)
ˆ
B
ϕ

ℓm
κ
1
ˆr
ˆ
B
z
,(13)
d
ˆ
B
ϕ
dˆr
=
mℓ
ˆr
ˆ
E
ϕ
+
(
κ
3

m
2
ˆr
2
)
ˆ
E
z

1
ˆr
ˆ
B
ϕ
,(14)
d
ˆ
B
z
dˆr
=
(

2
+
κ
2
2
κ
1
−κ
1
)
ˆ
E
ϕ

mℓ
ˆr
ˆ
E
z
+ℓ
κ
2
κ
1
ˆ
B
ϕ


2
ˆrκ
1
ˆ
B
z
,(15)
and the algebraic linear equations
ˆ
E
r
=
κ
2
κ
1
E
ϕ

m
ˆrκ
1
B
z
+

κ
1
B
ϕ
,(16)
ˆ
B
r
= −ℓ
ˆ
E
ϕ
+
m
ˆr
ˆ
E
z
,(17)
where subscripts lm have been omitted from the variables.
The above set of differential equations have boundary conditions at r = 0,r = r
a
,and r = r
w
.At r = r
w
,
the electric field parallel to the wall is zero:
ˆ
E
ϕ
(ˆr
w
) = 0,
ˆ
E
z
(ˆr
w
) = 0.(18)
At r = r
a
,the current sheet generated by the antenna that produces jumps in the parallel magnetic field.
From Maxwell’s equations
ˆ
B
z
(ˆr
+
a
) −
ˆ
B
z
(ˆr

a
) = −s
θ
c
ω
= −ˆs
θ
,(19)
ˆ
B
ϕ
(ˆr
+
a
) −
ˆ
B
ϕ
(ˆr

a
) = is
z
c
ω
= iˆs
z
.(20)
At r = 0,only bounded modes are acceptable.The Taylor expansion of Eqs.(12)-(15) yields that there are
only two bounded modes.For m= 0,these modes satisfy
(
ˆ
E
ϕ
,
ˆ
E
z
,
ˆ
B
ϕ
,
ˆ
B
z
) = C
1
(ˆr,0,0,2) [1 +O(ˆr)] (21)
(
ˆ
E
ϕ
,
ˆ
E
z
,
ˆ
B
ϕ
,
ˆ
B
z
) = C
2
(0,2,ˆrκ
3
,0) [1 +O(ˆr)] (22)
Applying the principle of superposition,the general solution for modes satisfying conditions at ˆr = 0 and
ˆr = ˆr
a
is
ˆ
E
z
(ˆr) = C
1
ˆ
E
z1
(ˆr) +C
2
ˆ
E
z2
(ˆr) +
ˆ
I
θ
ˆ
E
z3
(ˆr) +
ˆ
I
z
ˆ
E
z4
(ˆr) (23)
and similarly for the other five field components.Then the boundary conditions at ˆr = ˆr
w
provide two
conditions,
A
1
[
C
1
C
2
]
= A
2
[
ˆ
I
θ
ˆ
I
z
]
,(24)
that determine (C
1
,C
2
) in terms of (
ˆ
I
θ
,
ˆ
I
z
).This procedure is followed for each (l,m)-mode.Resonance for
a particular (l,m)-mode takes place when
det(A
−1
2
A
1
) = 0.(25)
5
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
0
10
20
30
40
50
-200
-100
0
100
200
l
Er
0
10
20
30
40
50
-0.05
0
0.05
l
E
0
10
20
30
40
50
-2
0
2
4
l
Ez
0
10
20
30
40
50
-1
-0.5
0
0.5
1
x 10
-6
l
Br
0
10
20
30
40
50
-2
0
2
4
x 10
-7
l
B
0
10
20
30
40
50
-5
0
5
x 10
-6
l
Bz
Figure 3.Contribution of the first fifty longitudinal modes to each of the fields’ components at B
0
= 450 G.
The full 3D solution is obtained by solving all significant modes (l,m).For the single-loop antenna,with
m = 0 as the only longitudinal mode the solution is axilsymmetric.The relative weigh of the longitudinal
modes decays quickly after the first 5-7 modes and becomes negligible at l ∼ 30.Furthermore,the high-l
modes tend to compensate by pairs among themselves as shown in Figure 3.Also if the antenna is centrally
located (at z
a
= L/2),even l-modes are negligible.Figure 4 plots some solutions.
B.Energy absorption
Once the electromagnetic field are determined,we can proceed to determine the absorption of the wave
energy by the plasma.In a steady-state,one has
p
a
= ∇· S,(26)
where
p
abs
=
1
2
Re(E

· j) =
1
2
Re(E

·
¯
¯σ · E),S =
1

0
Re(E

×B),(27)
are the absorbed energy density,p
abs
and the Poynting or energy flux vector S,respectively,and the plasma
conductivity is
¯
¯σ ≃ iω
¯
¯ϵ.The fields E and B are the complex total fields,i.e.sum of all the (l,m) modes
and proportional to expiωt.
The absorbed energy density distribution is the main magnitude required by the plasma dynamics model.
Its integral over the plasma volume yields the total absorbed power,P
abs
which can also be computed from
the normal of the Poynting vector to the plasma column surface,
P
abs
= 2π

L
0
dz

r
p
0
drrp
abs
(r,z) ≡ −

L
0
dzS
r
(r
p
,z) (28)
Figure 5 plots,for a uniform plasma,the dependence of the absorbed power with the magnetic field and
the plasma density.Strong resonance peaks are observed when one of the two magnitudes is tuned.Notice
that the absorbed energy is proportional to the square of the antenna current,I
2
.The aim of the system
6
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
0
0.005
0.01
0.015
0.02
-200
0
200
r [m]
Er
0
0.005
0.01
0.015
0.02
0
0.05
0.1
r [m]
E
0
0.005
0.01
0.015
0.02
-2
0
2
r [m]
Ez
0
0.005
0.01
0.015
0.02
-20
-10
0
10
r [m]
Br
0
0.005
0.01
0.015
0.02
-200
0
200
r [m]
B
0
0.005
0.01
0.015
0.02
0
500
1000
r [m]
Bz
Figure 4.Radial profiles of electromagnetic waves for B
0
(in Gauss)=260 (blue lines) and 850 (red lines).
Short wavelength waves are the TG waves,most visible at high confinement.
23
Dotted vertical lines show the
interphase between plasma and vacuum and the antenna radial position.Forty longitudinal modes are used,
though even modes are zero.Resonator dimensions are L = 20cm,r
p
= 10mm,r
a
= 12mm,r
w
= 20mm.
would be to operate at one of the peaks,for instance at n
0
= 3.8· 10
18
m
−3
and B
0
≃ 300 Gauss.The figure
also plots the influence of collisional frequency in the absorbed power.
IV.Results and discussion
A.In uence of the density prole
The first study we wanted to carry out is to assess the differences between a column with uniform density
and a non-uniform column satisfying Eq.(1),which is the one we expect in most of the plasma column.
We will compare cases with the same average density and the average density for the profile of Eq.(1) is
0.44n
0
.Figure 6 shows the presence of similar resonances for the nonuniform case,although the behavior
is not so neat as in the uniform case.Further analysis of these results is required to discern whether there
is any relevant difference in behavior.Plasma collisionality is assumed proportional to plasma density (i.e.
dominated by electron-ion collisions).
Figure 7 plots the 2D distribution of the absorbed energy density (multiplied by r in order to weight
adequately cylindrical volumes) for the uniform and nonuniform cases at antiresonance regimes.Energy
absorption is more peaked (both radially and axially) near the antenna for the nonuniform plasma.The
difference in the axial profiles is the most surprising one.Stronger wave reflection at the back and front plates
are more relevant for the uniform plasma due to the higher axial transmission of energy when operating out
of resonance.
Figure 8 plots,for the two cases of Fig.7,the perpendicular component of the energy flux vector S
r
,
showing that most of the energy flux enters the plasma (negative flux according to the surface vector) around
7
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
0.02
0.025
0.03
0.035
0.04
0.045
0.05
0.055
0.06
0.065
0.07
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Magnetic field B
0
, T
Power absorption, W


 = 2.2 Hz
 = 0.5 Hz
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
5
x 10
19
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Plasma density, m
-3
Power absorption, W
Figure 5.a) Power transmision for different applied magnetic fields when I = 1 A and n
e
= 3:8  10
18
m
3
for the same geometry than previous figure.Red dashed line stands for low collisional frequency and blue
continuous line for high frequency of collisions.b) Power transmision for variation of plasma density for 350
Gauss confinement and the same other parameters as in a).Dashed line at b) represents the volume integration
of power density and continuous lines refer to energy flux through the control area.
0.02
0.025
0.03
0.035
0.04
0.045
0.05
0.055
0.06
0.065
0.07
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
B
0
, T
Pabs, W
Figure 6.Absorbed power versus magnetic confinement for uniform (blue) and non-uniform (red) density
profiles.
the antenna location.The axial profile of the absorbed power in sections z = const,
¯p
r,abs
(z) = 2π

r
p
0
drrp
abs
(r,z) (29)
is also plotted.The differences between these magnitudes means that the absorption of the wave energy by
the plasma does not take place preferentially near the antenna,but it is distributed rather uniformly axially.
Though,integration along z axis of both yields the same result.Therefore the Poynting vector has a strong
axial component inside the plasma,which is observed in Fig.9,and as a consequence boundary conditions
at the back and front plates are very relevant.
Moreover,radial profiles differences of power absorption between density cases are dependent on the res-
onance or antiresonance regime.As shown on Fig.10,resonance absorption turns out to be counterintuitive.
The radially varying plasma absorbs more power at the outer volume than an homogeneous column for the
pictured cases.On the other hand,antiresonance regimes show a more intense decay of the quantity of power
absorbed in the homogeneous case than in the nonhomogeneous one.The behaviour of the homogeneous
case is probably due to the stronger damping at the edge of the column meaning less deposition in the inner
region.
8
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
Figure 7.2D plasma power absorption map for B
0
=580 Gauss:(a) uniformand (b) nonuniformplasma.Same
geometrical values,average density of 3:8  10
18
m
3
and antenna current of 1A are used in both cases.
B.In uence of the antenna location
The resonator model shows strong wave reflection at both end plates.The case we want to analyze has an
open front end and the plasma flows through it.Therefore,the resonator boundary conditions at the front
end could be not adequate.This issue is quite important,taking into account the last results of the previous
subsection.Real boundary conditions ’downstream’ are still uncertain to us.Since the density profile is not
uniform axially,and density decays quickly once the plasma starts to accelerate supersonically,it is likely
that waves are reflected back below a critical density but this aspect has not been researched yet.
For now,we decided to apply the resonator model to a plasma column,with the front plate located very
far away from the antenna,in order to assess the differences in the region near the antenna.In particular
we consider a column of length L = 1m,while we keep the antenna near the back end at z
a
= 10cm.
Radial variation of density is considered from now on.Changes such as inserting odd longitudinal modes
(m= 0,1,2...∞) because of the symmetry loss and modification of the antenna’s current expresion through
Dirac delta series are made.Though,antenna’s current remains as 1 A for the following cases.
Wave number components from the dispersion relation of the uniform infinite plasma problem,Eq.(4)
are recalled.Both propagation components,k

and k

,depend on magnetic applied fields.Analysis of the
9
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
0
0.05
0.1
0.15
0.2
0
2
4
6
8
10
12
z, m
2 Pabs(r,z)r·dr & 2 r S·n
0
0.05
0.1
0.15
0.2
0
0.1
0.2
0.3
0.4
0.5
0.6
z, m
2 Pabs(r,z)r·dr & 2 r S·n
Figure 8.Power integration along the radius versus energy flow through each z section a) Homogeneous plasma
b) Radially-varying density plasma.Recall Fig.6 to notice lower absorption at (b) when using 580G.
r, m
z, m
0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
0
0.05
0.1
0.15
0.2
r, m
z, m
0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
0
0.05
0.1
0.15
0.2
Figure 9.Power absorption on the 2D plane together with Poynting’s vector direction.(a) Uniform plasma
(b) Non-uniform plasma.Notice the effect of the wall reflection on S.
equation leads to the fact that lower magnetic confinement implies more intense axial propagation of waves,
which is related to longitudinal component of the wave number,and viceversa.
Figures 11 and 12 show results for several confinement strengths.The first of them shows the previously
used longitudinal power comparison for the near-the-plate antenna.The other figure’s color diagram shows
the power density at each point of the theta-constant plane.Note that power density of the last diagrams has
not been multiplied by their corresponding radius unlike in Fig.9,in order to visualize fluxes and absorption
near the axis in an easier way.However,they are not volume-proportional as explained above.
Current lines get steeper as B
0
decreases.Even so,power absorption is spread along the z axis due to
the higher longitudinal component of the wave number (k

).As the magnetic regime increases,the lines
where power is deposited lean towards a longitudinal axis parallel orientation.Nonetheless,high magnetic
confinement regimes concentrate the power absorption at the transversal plane,where the antenna is located.
When integrating the power deposition along the volume conformed from z = 0 to z = 0.2,this quantity
over the total power absorbed reaches a percentage of ∼ 80% as the magnetic confinement is increased.See
Table 1.
At high confinement regimes power absorption for far distances in axial direction decreases and thus,
reflected fluxes become lower.Even for 0.11 Teslas (Fig.12.f),it bears a small contribution of the wall
reflection at z ∈ [0,0.2] and r ∈ [0,0.002] (area of interest),yet it is a negligible quantity (as radius is close
to zero).This confirms the conclusions drawn from the dispersion relation equation.
Nevertheless,these regimes of high B
0
are not the most important operation points as they do not achieve
high power absorption in order to use the source for space propulsion.Regimes close to high resonances may
10
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
0
0.002
0.004
0.006
0.008
0.01
0
5
10
15
r, m
2 Pabs(r,z)r·dz
0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
-100
0
100
200
300
400
500
r, m
2 Pabs(r,z)r·dz


Nonhomogeneous
Homogeneous
Figure 10.Power absorption profiles in the radial direction for homogeneous (red dotted line) and nonhomo-
geneous (blue line) plasma for (a) 580G (b) resonance regimes of 241G and 316G respectively.
0
0.2
0.4
0.6
0.8
1
-0.5
0
0.5
1
1.5
2
2.5
3
z, m
2 Pabs(r,z)r·dr & 2 r S·n
0
0.2
0.4
0.6
0.8
1
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
z, m
2 Pabs(r,z)r·dr & 2 r S·n
Figure 11.Power absorption (blue continuous line) and energy flux (red dotted line) along the z axis for a)
100 Gauss and b) 330 Gauss.Higher confinement produces a concentration of power absorbed at the position
of the antenna due to the dominant transversal wave number.Different fluxes from the antenna have to be
noticed as the plasma also varies and thus,the energy does not get in and out of the volume through the same
regions.
be preferred for these purposes.
B
0
100
280
330
500
650
1100
%P
23.97
33.67
30.66
21.96
51.4
81.38
Table 1.Percentage of power deposited from z = 0 to z = 20cm relative to the total absorbed power.
V.Conclusions
A standard plasma resonator model has been used in order to understand the propagation and absorption
of the helicon waves inside a plasma column.Resonance peaks in terms of magnetic confinement and plasma
density,known from other authors’ works,have been reproduced.The comparison of the response for a uni-
form plasma column and the nonuniform typical profile of a magnetized plasma show significant differences.
The analysis of the Poynting vector shows that,for homogeneous plasmas,energy flux penetrates the plasma
around the antenna location but then it spreads longitudinally.On the other hand,nonhomogeneous profiles
of density enhance absorption near the antenna and reduce the axial transmission.As a consequence,energy
absorption takes rather uniformly in the axial direction,mainly at uniform cases.The axial propagation
of the Poynting vector also means that the boundary condition at the front end can influence much the
response.
The study of the magnetic confinement variation confirms some facts obtained from the dispersion rela-
11
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
z, m


0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
200
400
600
800
1000
1200
1400
z, m


0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
200
400
600
800
1000
1200
1400
z, m


0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0
200
400
600
800
1000
z, m


0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
200
400
600
800
1000
1200
1400
z, m


0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0
100
200
300
400
500
600
700
800
900
1000
z, m


0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
0.01
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0
200
400
600
800
1000
1200
1400
1600
1800
Figure 12.Power absorption on the 2D plane together with Poynting’s vector direction for several values of
magnetic confinement a)100G b)280G c) 330G d)500G e)650G f)1100G.Color accounts for level of absorption.
White arrows show the direction and strength of the Poynting vector,energy flux.
tion.When front end is set at a great distance,for high B
0
fields,propagation of energy is directed in the
z-axis direction but power absorption is enhanced in the antenna plane.On the opposite side,low magnetic
fields enhance radial propagation and longitudinal absorption.Thus,good approximation of propulsion de-
vices by using a resonator are found only for large confinement values which diminish the effec of end plate
reflection.
Therefore,further research must be carried on the downstream boundary conditions before the plasma-
wave model can be matched to the fluid dynamics problem.This investigation can possibly be done taking
into account the axial variation of the density profile and looking for natural reflection of the waves below a
critical density.
There are other aspects of the problem that also merit consideration.First,wave effects on ions should
be assessed.Second,the cold plasma approximation needs to be checked,since in a helicon thruster the
electron temperature is expected to be of the order of several tens of eV.Third,the time scale separation
between the wave and fluid responses and nonlinear wave effects need to be studied.And finally,the presence
of hot electrons,observed in some experiments
7,26,27
needs to be worked out.
Acknowledgments
This work has been supported by the Gobierno de Espa˜na (Project AYA2010-16699).
References
1
Ziemba,T.,Carscadden,J.,Slough,J.,Prager,J.,and Winglee,R.,“High Power Helicon Thruster,” 41th Joint Propulsion
Conference,Tucson,AR,AIAA 2005-4119,American Institute of Aeronautics and Astronautics,Washington DC,2005.
2
Charles,C.,Boswell,R.,and Lieberman,M.,“Xenon ion beam characterization in a helicon double layer thruster,”
Applied Physics Letters,Vol.89,2006,pp.261503.
12
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011
3
Batishchev,O.,“Minihelicon Plasma Thruster,” IEEE Transaction on Plasma Science,Vol.37,2009,pp.1563–1571.
4
Pavarin,D.,Ferri,F.,Manente,M.,Curreli,D.,Guclu,Y.,Melazzi,D.,Rondini,D.,Suman,S.,Carlsson,J.,Bramanti,
C.,Ahedo,E.,Lancellotti,V.,Katsonis,K.,and Markelov,G.,“Design of 50WHelicon Plasma Thruster,” 31th International
Electric Propulsion Conference,Ann Arbor,Michigan,USA,IEPC 2009-205,Electric Rocket Propulsion Society,Fairview
Park,OH,2009.
5
Ahedo,E.,“Plasma dynamics in a helicon thruster,” Proceedings of EUCASS 2011,4-8 July 2011,Saint Petersburg,
Russia,paper 118,2011.
6
Boswell,R.,“Very efficient plasma generation by whistler waves near the lower hybrid frequency,” Plasma Physics and
Controlled Fusion,Vol.26,1984,pp.1147–1162.
7
Tysk,S.,Denning,C.,Scharer,J.,and Akhtar,K.,“Optical,wave measurements,and modeling of helicon plasmas for a
wide range of magnetic fields,” Physics of Plasmas,Vol.11,No.3,2004,pp.878–887.
8
Gilland,J.,Breun,R.,and Hershkowitz,N.,“Neutral pumping in a helicon discharge,” Plasma Sources Sci.Technol.,
Vol.7,1998,pp.416–422.
9
Ahedo,E.,“Parametric analysis of a magnetized cylindrical plasma,” Physics of Plasmas,Vol.16,2009,pp.113503.
10
Ahedo,E.,“Cylindrical model of a helicon-generated plasma,” 31th International Electric Propulsion Conference,Ann
Arbor,Michigan,USA,IEPC 2009-193,Electric Rocket Propulsion Society,Fairview Park,OH,2009.
11
Ahedo,E.,“Magnetic confinement of a high-density cylindrical plasma,” Physics of Plasmas (submitted),2011.
12
Ahedo,E.and Merino,M.,“Two-dimensional supersonic plasma acceleration in a magnetic nozzle,” Physics of Plasmas,
Vol.17,2010,pp.073501.
13
Ahedo,E.and Merino,M.,“On plasma detachment in propulsive magnetic nozzles,” Physics of Plasmas,Vol.18,2011,
pp.053504.
14
Merino,M.and Ahedo,E.,“Simulation of plasma flows in divergent magnetic nozzles,” IEEE Transactions Plasma
Science (to appear in August),2011.
15
Merino,M.and Ahedo,E.,“Plasma detachment mechanisms in a magnetic nozzle,” Proceedings of 47th Joint Propulsion
Conference,San Diego,CA,AIAA 2011-5999,American Institute of Aeronautics and Astronautics,Washington DC,2010.
16
Ahedo,E.and Merino,M.,“Current Ambipolarity and Plasma Detachment on a Magnetic Nozzle,” 32th International
Electric Propulsion Conference,Wiesbaden,Germany,IEPC 2011-050,Electric Rocket Propulsion Society,Fairview Park,OH,
2011.
17
Ahedo,E.and Mart¨ı¿
1
2
nez-S¨ı¿
1
2
nchez,M.,“Theory of a stationary current-free double-layer in a collisionless plasma,”
Physical Review Letters,Vol.103,2009,pp.135002.
18
Ahedo,E.,“Double-layer formation and propulsive assessment for a three-species plasma expanding in a magnetic nozzle,”
Physics of Plasmas,Vol.18,2011,pp.033510.
19
Shamrai,K.and Taranov,V.,“Resonance wave discharge and collisional energy absorption in helicon plasma source,”
Plasma Physics and Controlled Fusion,Vol.36,1994,pp.1719–1735.
20
Chen,F.,“Physics of helicon discharges,” Physics of Plasmas,Vol.3,No.5,1996,pp.1783–1793.
21
Arnush,D.and Chen,F.,“Generalized theory of helicon waves.II.Excitation and absorption,” Physics of Plasmas,
Vol.5,1998,pp.1239–1254.
22
Lieberman,M.and Lichtenberg,A.,Principles of plasma discharges and materials processing,Wiley-Blackwell,2005.
23
Shamrai,K.and Taranov,V.,“Volume and surface rf power absorption in a helicon plasma source,” Plasma Sources
Science and Technology,Vol.5,1996,pp.474.
24
Shamrai,K.P.,Pavlenko,V.P.,and Taranov,V.B.,“Excitation,conversion and damping of waves in a helicon plasma
source driven by an m=0 antenna,” Plasma Phys.Control.Fusion,Vol.39,1997,pp.505–529.
25
Cho,S.and Lieberman,M.,“Self-consistent discharge characteristics of collisional helicon plasmas,” Physics of Plasmas,
Vol.10,2003,pp.882–890.
26
Zhu,P.and Boswell,R.W.,“Observation of nonthermal electron tails in an rf excited argon magnetoplasma,” Phys.
Fluids B,Vol.3,1991,pp.869–874.
27
Chen,R.T.S.and Hershkowitz,N.,“Multiple Electron Beams Generated by a Helicon Plasma Discharge,” Physical
Review Letters,Vol.80,1998,pp.4677–4680.
13
The 32
nd
International Electric Propulsion Conference,Wiesbaden,Germany
September 11{15,2011