Criterion of sheath formation in magnetized low pressure plasma

R. Moulick, S. Adhikari, and K. S. Goswami

Citation: Physics of Plasmas 24, 114501 (2017);

View online: https://doi.org/10.1063/1.4994261
View Table of Contents: http://aip.scitation.org/toc/php/24/11
Published by the American Institute of Physics

Articles you may be interested in

Rogue waves in space dusty plasmas
Physics of Plasmas 24, 113701 (2017); 10.1063/1.5000027

Solitary structures in an inhomogeneous plasma with pseudo-potential approach

Physics of Plasmas 24, 114502 (2017); 10.1063/1.4993812

Electrostatic solitary waves in an electron-positron pair plasma with suprathermal electrons

Physics of Plasmas 24, 102905 (2017); 10.1063/1.5000873

Avalanches driven by pressure gradients in a magnetized plasma

Physics of Plasmas 24, 112302 (2017); 10.1063/1.5001321

Interplay between Alfvén and magnetosonic waves in compressible magnetohydrodynamics turbulence

Physics of Plasmas 24, 102314 (2017); 10.1063/1.4997990

Effect of magnetic quantization on ion acoustic waves ultra-relativistic dense plasma

Physics of Plasmas 24, 112301 (2017); 10.1063/1.4996601
 PHYSICS OF PLASMAS 24, 114501 (2017)

Criterion of sheath formation in magnetized low pressure plasma

R. Moulick,1,a) S. Adhikari,2 and K. S. Goswami2
1
Lovely Professional University, Phagwara, Punjab 144411, India
2
Centre of Plasma Physics-Institute for Plasma Research Sonapur, Guwahati, Assam 782402, India
(Received 5 July 2017; accepted 18 October 2017; published online 6 November 2017)
A criterion of sheath formation is obtained for magnetized low pressure plasmas. The criterion
includes the effect of both collision and the magnetic field. The collision parameter is defined by
the ratio of the ionization length to the ion mean free path. The ionization frequency is assumed to
be constant. The condition obtained is consistent with the electrostatic case. In the absence of the
magnetic field and the collision, it retrieves the Bohm criterion at the sheath edge. For an
electrostatic case in the absence of ion neutral collision, the Bohm criterion determines the ion
entry speed within the sheath. However, the presence of collision limits the validity of the criterion
to a threshold value of the collision parameter. In the magnetized scenario, the validity is found to
be dependent on the magnetic field angle besides the collision parameter. Even in a collisionless
scenario, the validity is in question depending on the field angle. A critical collision parameter is
found for a definite magnetic field strength beyond which there is no more angle dependency on
the validity criterion. The effect of the magnetic field on the space charge deposition is highlighted.
Published by AIP Publishing. https://doi.org/10.1063/1.4994261

d
The non-neutral region near the boundary of a confined ðni tz Þ ¼ Si ¼ Zne ; (1)
dz
plasma has been drawing considerable interest for many
years.1,2 The ions enter the region with a typical velocity known where ni and tz are the density and z-directional velocity of
as the Bohm velocity. The criterion as originally described puts the ions. Z is the ionization frequency (assumed constant),
a condition on the ion velocity that must be satisfied in order to and ne is the electron density.
form a stable sheath near the plasma wall.3 Attempts are made The momentum equation of the ions is given by
to find out the kinetic theory of the Bohm criterion in order to
seek a generalized criterion for the more general electron-ion t
d~ d/ ^
mi ni tz ¼ ni e k þ ni eð~ ~Þ  mi~
tB tSi  i ni mi~
t:
distribution.4 With the advancement in time, many other dz dz
attempts have been made to find out modifications to the condi- (2)
tion under different plasma scenarios. The presence of negative
The first and second terms on the right-hand side of the
ions, dust particles, and more species of positive ions greatly
Eq. (2) are the standard electrostatic and magnetic force
affects the condition.5–9
In this brief communication, a generalized condition for terms. Third and fourth terms correspond to momentum
the sheath formation for magnetized plasma is found out. The transfer due to ionization and collision with neutral atoms. i
ion-neutral collision is considered along with the tilted mag- is the ion-neutral collision frequency. The other symbols
netic field. A schematic of the problem is shown in Fig. 1. have their usual meaning. It is important to note that under
Let D be the distance between the undisturbed plasma and the assumption of low pressure plasma with Ti  Te , the dif-
the wall. The problem finds its importance in many situations fusion term on the right hand side of Eq. (2) is neglected.
including fusion reactors, processing plasmas, and space plas- However, for the case where Ti is significant enough, the
mas (where the surface of space stations may be tilted with term must be retained.
respect to the earth’s magnetic field10). The motivation is to The velocity vector ~ t can be resolved into its compo-
understand the basic nature of the magnetized sheath forma- nents along x, y, and z directions to yield the following set of
tion under collisional conditions. equations:
The plasma is assumed to consist of electrons and singly
dtx Si tx
charged positive ions. The magnetic field is tilted at an angle tz ¼ ty xz   i tx ; (3)
a with respect to the vertical x-axis. Governing equations are dz ni
the fluid continuity and momentum equation for ions. The dty Si
tz ¼ ðtz xx  tx xz Þ  ty  i ty ; (4)
coupled equations are closed by Poisson’s equation. The prob- dz ni
lem considered is one dimensional with three components of
dtz e d/ Si
the ion velocity. The ionization function is considered to be tz ¼  ty xx  tz  i tz : (5)
constant. The plasma is steady, homogeneous and parallel to dz mi dz ni
the wall.
Here, xx and xz are the ion cyclotron frequencies along
The continuity equation for the ions is
x and z directions. The expressions are given by xx ¼ eB
mi
x

a)
Email: [email protected] ¼ eB eBz eB0
mi cosðaÞ and xz ¼ mi ¼ mi sinðaÞ.
0

1070-664X/2017/24(11)/114501/6/$30.00 24, 114501-1 Published by AIP Publishing.

114501-2 Moulick, Adhikari, and Goswami Phys. Plasmas 24, 114501 (2017)

 
dni ni eE ty
¼ X þ xx ; (8)
dz tz mi tz tz
dne ne eE
¼ ; (9)
dz Te

where X ¼ i þ 2Z nnei is the effective local collision frequency.

The variation of the space charge r is
   
dr d ne ni ni ty
¼ ðni  ne Þ ¼ eE  þ x x þ X :
dz dz Te mi t2z tz tz
(10)

Starting from a region within the plasma where the space

charge density is nearly zero, the formation of the sheath
demands the condition

dr
> 0: (11)
dz
FIG. 1. Schematic of the problem. D is the distance of the wall from the Which upon simplification yields
unperturbed plasma.
!   
tX t2 ty xx
Electrons being Boltzmann are described by the follow- 1  z2 <1þ : (12)
tz tc tz X
ing equation:
  Condition (12) is the criterion for the sheath formation. In
e/
ne ¼ n0 exp ; (6) the inequality (12), there are two characteristic velocities,
Te
given by
n0 is the equilibrium density of the charged species and Te is rffiffiffiffiffi
eE ni
the electron temperature. Regarding the Boltzmann consider- tX ¼ and tc ¼ tB : (13)
mi X ne
ation of the electrons, it should be noted that the relation
appears as a consequence of the absence of the magnetic The characteristic speeds are the functions of not only elec-
field, collision with the neutrals, and particle generation in tron temperature but also space. The speed tc takes into
the momentum balance equation for electrons. Even in the account the space charge deposition and is equal to the usual
absence of latter two, the presence of the magnetic field alone Bohm velocity in a quasi-neutral region (i.e., tc ¼ tB ). Thus,
may be sufficient to cause the deviation from the Boltzmann this strongly represents the space charge variation and serves
behavior.11 In the present communication, the electron neu- as the characteristic velocity for the sheath formation more
tral collision is assumed to be insignificant (due to the small strongly than the Bohm velocity. The characteristic speed tc
cross-section) and the momentum transfer due to the ioniza- was described and formulated in Ref. 15 for the first time.
tion is small as compared to electrostatic and pressure gradi- The sheath condition (12) can be written as
ent forces. Regarding the magnetic field, there are instances " #
   
where the electrons are assumed to be strongly magnetized, tz tc tc ty xx tz
þ þ > 1: (14)
fluid velocity is assumed to be small compared to the thermal tc tX tz tz X tc
velocity, and this retains the Boltzmann behavior. Moreover,
this keeps simplicity in the problem and helps to concentrate Condition (14) is more complex than the simple Bohm crite-
more on the ion dynamics.12–14 rion for sheath formation. As it is obvious, the middle term
Poisson’s equation is used for closure of the set of gov- in the expression is due to the magnetic contribution. Again,
erning equations. at the location of the sheath formation, the following two
conditions are satisfied:
d2 / e
¼  ðni  ne Þ: (7) dne dni
dz2 0 <0 and < 0: (15)
dz dz
The attempt in this paper is to find a criterion which will be
valid for both collisional and magnetized plasma scenarios. These two conditions lead to
The electrostatic analysis of the collisional sheath formation ty eE
is performed in Refs. 15 and 16. In the presence of the mag- E>0 and Xþ xx < : (16)
tz mi tz
netic field, ion and electron density variations along the z-
direction can be explicitly written as The second condition in Eq. (16) states
114501-3 Moulick, Adhikari, and Goswami Phys. Plasmas 24, 114501 (2017)

  
tz t y xx 1
1þ < 1: (17) ni tz ¼ ne c;
tX tX X 4
qffiffiffiffiffiffi
Thus, combining Eqs. (14) and (17), we say that sheath 8Te
c¼ pme is the random electron velocity at the wall.
forms under the general condition Corresponding to B ¼ 0:1 T; 2 T; and 5 T for a ¼ 30 and
    
tz t y xx tz K ¼ 1, we have the normalized wall coordinate as nw
1þ <1< ¼ 0.5456, 0.3661, and 0.3431. On the other hand for B
tX tX X tc
"     # ¼ 2T; a ¼ 30 ; and K ¼ 1, 100, and 1000, nw ¼ 0.3661,
tc tc ty xx tz 0.2197, and 0.1455, respectively.
 þ þ :
tX tz tz X tc In order to solve the governing equations, the initial val-
(18) ues are to be determined near the midplane. This is done by
solving the equations in power series. The following are
The condition (18) is the generalization of condition (20) in used for this purpose:
Ref. 15. Condition (17) shows that the magnetic contribution
is interwoven within the electric contribution. The bracketed Ns ¼ Ns0 þ Ns1 n2 þ Ns2 n4 þ   
term on the extreme left is always greater than unity. Hence,
for the validity of the condition, tz must be less than tX . This Vs ¼ Vs1 n þ Vs2 n3 þ   
is exactly the condition obtained for electrostatic sheath by Cs ¼ Cs1 n þ Cs2 n3 þ   
Valentini.15 It is interesting to note that despite the fact that
this condition is derived from the conditions which are inde- g ¼ g1 n2 þ g2 n4 þ   
pendent of collision, the validity of the condition is subject
to collisionality. At very high collision, tz is greater than tX . The quantities N; V; C; and g represent the density, velocity,
The reason behind this is the ratio XE (which affects tX ), flux, and potential for any species s ðelectron or ionÞ. The first
which is reduced with the increase in collision. Thus, after a order coefficients are given by
definite value of the collision parameter K, the condition is
Ni0  Ne0
reversed. On plotting the two velocities over a wide range of C1 ¼ bNe0 ; g1 ¼ ;
2a0
collision, it is seen that the cross over point lies somewhere
at K ¼ 260. It should however be noted that in the regime C1
x1 ¼ ; and Ne1 ¼ g1 Ne0 :
of constant collision frequency tz ! tX as collision is Ni0
increased.17 Let us designate the general condition (18) as
Along with these, there are three more equations from which
condition <1 and condition >1, for the two parts which are
the central ion density and the first order coefficients of the
less than and greater than unity, respectively.
other two velocity components (u and t) can be determined.
The set of governing equations has to be normalized in
order to analyze the criterion. The scaling parameters are 3
Ni0 2
 ð1 þ cx a0 t1 Þ Ni0  2a0 b2 ¼ 0
chosen as follows:
u1 ðbNe0 þ Ni0 x1 Þ  cz Ni0 t1 ¼ 0
z tx ty tz e/
n¼ ; u¼ ; t¼ ; x¼ ; g¼ ; t1 ðbNe0 þ Ni0 x1 Þ þ cz Ni0 u1 ¼ cx Ni0 x1 :
L cs cs cs Te
rffiffiffiffiffi
ni ne Te Z Here,
N i ¼ ; N e ¼ ; cs ¼ t B ¼ ; L ¼ kni ¼ ;
n0 n0 mi cs
L L L
kni is the ionization length. The normalized wall coordinate cx ¼ xx ; cz ¼ xz and b¼ Z:
cs cs cs
can be written as nw ¼ D=L. Therefore, the domain in the
normalized coordinate is 0  n  nw or 0  nnw  1. The default parameters of the calculation are as follows:
It is often seen that the collision is modeled for the con-
stant cross-sectional regime than constant mobility. The dif- Te ¼ 1 eV; mi ¼ 40 AMU;
5 1
ference in results, due to two regimes of collision at a high Z ¼ 1  10 s ; and n0 ¼ 1  1016 m3
value of the collision parameter, is highlighted in Ref. 5.
Under a constant mean free path, the parameter of collision Figure 2 shows the evolution of terms in the condition (18)
can be defined as with the collision parameter. The analysis is done at the wall.
L LHS1 and RHS1 stand for the terms in condition (17) (taking
K¼ ; tz
ki tX on the other side of the inequality) and LHS2 and RHS2, cor-
respond to the condition (12). At the value K ¼ 260, the
K is the inverse ratio of the Knudsen number. Increasing the change in the condition is clearly visible. Below K ¼ 260,
value of K implies shortening ki and hence serves the pur- LHS1 < RHS1, and beyond this collision limit, LHS2 < RHS2.
pose of the collision parameter. If the right wall is assumed These two conditions are with condition <1 and condition >1,
to be floating and the left wall to be the unperturbed plasma, respectively. Thus, the figure clearly illustrates the shift of
then at the right wall, the condition of flux balance must be the sheath condition with increasing collision. K ¼ 260 may be
maintained. Therefore, considered as the critical collision parameter for the said
114501-4 Moulick, Adhikari, and Goswami Phys. Plasmas 24, 114501 (2017)

collision parameter, both the conditions are valid in the region

of interest of the plasma. However, beyond the critical value,
one of them shifts beyond the region of interest. Therefore,
beyond the critical value, only one condition practically domi-
nates over the whole plasma (i.e., condition >1 since the other
lies beyond the point of interest).
In an attempt to find the limits of the Bohm criterion in
collisional plasma, Valentini and Kaiser18 found a threshold
collision parameter beyond which the sheath is mainly cre-
ated by the collision and ionization. The Bohm treatment is
valid between a velocity interval which becomes narrower
with an increase in collision and disappears at the threshold.
In order to investigate the same for magnetized plasma, the
spatial rate of variation of space charge density can be writ-
ten as

FIG. 2. Test of different terms in the general condition over a wide range of d
ðNi  Ne Þ ¼ B þ C þ M; (19)
collision at the position of the wall for B ¼ 2 T and a ¼ 45 . LHS1 and dn
RHS1 are but the same terms.
where
parameter regime. However, for other regimes of parameters, !
the critical value will be different. For the low angle, the critical N i F x2
B¼ 2 1
collision value is low; however, for the low magnetic field, the x Vc2
critical collision parameter is relatively high. The figure also
depicts the dominance of the conditions based on the collision 2Ne þ KNi
C¼ :
regime. Hence, collision may be said to govern the phenome- x
non of sheath formation in the magnetized plasma. And
In Fig. 3, the variation of the ion and electron density is
  
shown for the parameter set K ¼ 100, B ¼ 2 T, and a ¼ 45 . t xx
The two points on the graph designate the validity of the two M ¼ Ni 2 :
x Z
conditions, namely, condition <1 and condition >1. It is
qffiffiffiffi
interesting to note that condition <1 is valid beyond the point Here, Vc ¼ Ni
(normalized characteristic speed) and F
Ne
shown, whereas condition >1 is valid till the point shown.
Thus, the validity of the two conditions overlaps in space. ¼ (normalized electric field) with E0 ¼ ekTeni .
E
E0
The spatial overlapping zone for the conditions becomes thin The normalized electric field varies according to the fol-
with increasing collision, ultimately leading to a sharp bound- lowing equation:
ary value at the critical collision parameter. Till the critical
dF Ni  Ne
¼ ;
dn a0
 2 r
ffiffiffiffiffiffiffiffiffiffiffiffi

where a0 ¼ kLD and kD ¼ 0 Te
n0 e2 is the Debye length. In

this paper, the figures are obtained using a0 ¼ 2:305  105 .

The term B describes the contribution of the electric field
and ion inertia to the formation of sheath, justifying Bohm’s
treatment. The contribution of B appears towards sheath for-
mation only if B > 0. The Bohm criterion does not hold at
all if B  0 throughout the plasma. C includes the collision
and ionization for the space charge variation in front of the
sheath, and M is the magnetic contribution towards sheath
formation. Figures 4(a), 4(b), and 4(c) show the spatial varia-
tion of each term together with the variation of the magnetic
field angle for different collision parameters K.
For an electrostatic treatment, the threshold to the forma-
tion of sheath is given by the collision parameter only. The
paper of Valentini and Kaiser18 finds the threshold parameter
to be K ¼ 25.6 for a definite value of a0 . In contrast, when the
FIG. 3. Spatial test of the general condition for K ¼ 100 at B ¼ 2 T and magnetic field is present, the terms show the variation based
a ¼ 45 . on the magnetic field angle too. Thus, even for K ¼ 1, at
114501-5 Moulick, Adhikari, and Goswami Phys. Plasmas 24, 114501 (2017)

a ¼ 30 , term B is below the zero level. This, in other words,

states that the Bohm criterion is not satisfied for this regime
of parameters. The contribution of collision and ionization
plays a moderate role; however, the dominance is shown by
the magnetic field. At an even less angle, the dominance
would be more and highly prominent. Thus, in this case, lim-
its to the Bohm criterion are provided not only through the
threshold collision but also the threshold field and the field
angle. For B ¼ 2 T and K ¼ 1, a ¼ 32 acts as the threshold
angle above which the Bohm criterion would be satisfied and
below which the sheath formation is magnetic field domi-
nated. For this regime of parameter, any small rise in colli-
sion is enough to cause a large deviation from the Bohm
criterion. Again, for B ¼ 2 T and K ¼ 10, a ¼ 49 becomes
the threshold. The dependency of the threshold condition is
however limited to a certain collision parameter. For exam-
ple, for B ¼ 2 T and a ¼ 89 , K ¼ 17.8 happens to be the
threshold collision parameter, above which the Bohm crite-
rion loses its significance and B < 0 throughout the plasma.
Thus, K ¼ 17.8 is the critical collision parameter for B ¼ 2 T
above which the angle dependency is removed and the Bohm

FIG. 4. (a) Spatial variation of B, C, and M with the magnetic angle for
K ¼ 1. (b) Spatial variation of B, C, and M with the magnetic angle for FIG. 5. (a) Velocity variation of B, C, and M with the magnetic field angle
K ¼ 100. (c) Spatial variation of B, C, and M with the magnetic angle for for K ¼ 1. (b) Velocity variation of B, C, and M with the magnetic field
K ¼ 1000. angle for K ¼ 100.
114501-6 Moulick, Adhikari, and Goswami Phys. Plasmas 24, 114501 (2017)

the plasma. Comparison of the figures with Ref. 15 shows

that the nature of the variation is governed by the collision
parameters, but the early spatial deposition of space charges
is due to the magnetic field. This is, in fact, obvious because
the magnetic contribution is interwoven within the colli-
sional contribution in the criterion of sheath formation. Thus,
the effect of the increasing magnetic field is to widen the
sheath. There is experimental evidence which proves that
sheath is expanded in the source region with the increase in
the magnetic field under collisionless conditions.19,20 The
widening is again dependent on the field angle. The sheath is
wider at a low angle than the higher. Also, according to Fig.
6, at the low magnetic field and low collision, widening is
somewhat indifferent to the field angle.
In summary, the condition obtained for the magnetized
collisional sheath formation is found to be more complex
than the simple electrostatic one. The spatial zone of the
FIG. 6. Variation of space charges for different angles, collision parameters, validity of the conditions is highlighted. For the low and
and magnetic fields. moderate value of the collision, the conditions overlap; how-
ever, beyond a collision limit, i.e., K ¼ 260, condition <1 is
valid beyond the wall. This means that beyond K ¼ 260, the
validity of condition >1 dominates over the whole plasma.
Threshold parameter sets are obtained for the validity of
the Bohm criterion, and the effect of the magnetic field is
highlighted. It is found that beyond a certain critical collision
parameter, the threshold conditions are no more dependent
on the magnetic field angle. The effect of magnetic field
strength is to deposit the space charges early in the space
coordinate. However, the nature of the variation of space
charges in front of the wall is determined by the status of the
collision. There are many instances besides the magnetic
fusion where the wall interacts with a tilted magnetic field.
The present findings may be useful for studying such scenar-
ios including space plasmas, fusion plasmas, and laboratory
processing plasmas.
FIG. 7. Variation of space charges for different collision parameters at an
angle of 45 and different magnetic field strengths.
1
K.-U. Riemann, J. Phys. D: Appl. Phys. 24, 493 (1991).
2
N. Hershkowitz, Phys. Plasmas 12, 55502 (2005).
criterion is never satisfied for any angle. Thus, in the presence 3
D. Bohm, The Characteristics of Electrical Discharges in Magnetic
of the magnetic field, a set of threshold parameter regimes Fields, 1st ed. (McGraw-Hill Book Company, Inc., 1949).
4
exists rather than a single parameter threshold. Similarly, for S. D. Baalrud and C. C. Hegna, Plasma Sources Sci. Technol. 20, 25013
another magnetic field strength, threshold conditions will be (2011).
5
R. Moulick, M. K. Mahanta, and K. S. Goswami, Phys. Plasmas 20, 94501
given by another regime of parameters. (2013).
Figures 5(a) and 5(b) show the velocity variation of B, 6
R. Moulick and K. S. Goswami, Phys. Plasmas 21, 83702 (2014).
7
C, and M with the magnetic angle at different collision 8
R. Moulick and K. S. Goswami, Phys. Plasmas 22, 43701 (2015).
parameters. R. Moulick and K. S. Goswami, Phys. Plasmas 22, 33510 (2015).
9
G. C. Das and P. Kalita, J. Phys. D: Appl. Phys. 37, 702 (2004).
The effect of the magnetic field is to increase the space 10
R. H. Cohen and D. D. Ryutov, Phys. Plasmas 2, 2011 (1995).
charge deposition deeper into the plasma. Figure 6 shows the 11
R. N. Franklin, J. Plasma Phys. 78, 21 (2012).
12
variation of space charge with the distance at three different P. C. Stangeby, Nucl. Fusion 52, 83012 (2012).
13 
J. Kovačič, T. Gyergyek, and M. Cerček, Eur. Phys. J. D 54, 383 (2009).
angles for given values of the magnetic field and the colli- 14
R. Chodura, Phys. Fluids 25, 1628 (1982).
sion parameter. 15
H.-B. Valentini, Plasma Sources Sci. Technol. 9, 574 (2000).
16
In Fig. 7, the space charge deposition is shown to vary H.-B. Valentini and D. Kaiser, Plasma Sources Sci. Technol. 23, 015004
for three different values of collision parameters at an angle (2014).
17
R. N. Franklin, J. Phys. D: Appl. Phys. 36, 2821 (2003).
of 45 and given magnetic field strengths. 18
H.-B. Valentini and D. Kaiser, Phys. Plasmas 22, 53512 (2015).
From above figures, it is clear that the effect of the mag- 19
B. Singha, A. Sarma, and J. Chutia, Rev. Sci. Instrum. 72, 2282 (2001).
20
netic field enhances the space charge deposition deeper into B. Singha, A. Sarma, and J. Chutia, Phys. Plasmas 9, 683 (2002).