Mon. Not. R. Astron. Soc. (2012) doi:10.1111/j.1365-2966.2012.21505.

The effects of strong quantizing magnetic fields on the cold emission

of electrons from ultramagnetized compact stellar objects

Arpita Ghosh and Somenath Chakrabarty

Department of Physics, Visva-Bharati, Santiniketan 731235, India

Accepted 2012 June 12. Received 2012 June 12; in original form 2012 February 29

ABSTRACT
Formalisms for both non-relativistic and relativistic versions of field emission of electrons in
the presence of a strong quantizing magnetic field, relevant for strongly magnetized neutron
stars or magnetars, are developed. In the non-relativistic scenario, where electrons obey the
Schrödinger equation, we have noticed that when Landau levels are populated for electrons in
the presence of a strong quantizing magnetic field the transmission probability exactly vanishes
unless the electrons are spin-polarized in the opposite direction to the external magnetic field.
On the other hand, cold electron emission under the influence of a strong electrostatic field
at the poles is totally forbidden from the surfaces of those compact objects for which the
surface magnetic field strength is 1015 G (in the eventuality that they may exist). Whereas
in the relativistic case electrons obey the Dirac equation, the presence of a strong quantizing
magnetic field completely forbids the emission of electrons from the surfaces of compact
objects if B > 1013 G.
Key words: atomic processes – dense matter – magnetic fields – relativistic processes – stars:
magnetars – stars: neutron.

out through the metallic surface (surface barrier). However, as soon

1 I N T RO D U C T I O N
There are three main kinds of electron emission processes from
metal surfaces: they are (i) thermionic emission, (ii) photoelectric
emission and (iii) cold emission or field emission.
The field emission or cold emission, which we have investigated
in the present article in the context of strongly magnetized neutron
stars or magnetars, is an electron emission process induced by a
strong external electrostatic field at zero or extremely low temper-
ature. Field emission can happen from solid and liquid surfaces,
or from individual atoms. It has been noticed that the field emis-
sion from metals occurs in the presence of a high electric field:
the gradients are typically higher than 1000 volts per micron and
the emission is strongly dependent upon the work function of the
material. Unlike thermionic emission and photo-emission of elec-
trons, the field emission process can only be explained by quantum
tunnelling of electrons, which has no corresponding classical ex-
planation. However, for a general-type surface barrier, this purely
quantum mechanical problem cannot be solved exactly and a semi-
classical approximation known as the Wentzel–Kramers–Brilloun
(WKB) approximation is needed to obtain tunnelling coefficients.
Now, to explain cold electron emission from metals one may as-
sume that, because of quantum fluctuation, electrons from the sea of
conduction electrons (degenerate electron gas) always try to tunnel
 E-mail: somenath.chakrabarty@visva-bharati.ac.in
as an electron emerges it induces an image charge on the metal
surface, which pulls it back and does not allow this emitted elec-
tron to move far away from the metal surface on the atomic scale.
Nevertheless, if some strong attractive electrostatic field is applied
near the metallic surface then, depending on the Fermi energy of
the electrons, the height of the surface barrier and the local work
function, the electrons may overcome the effect of the image charge
and become liberated. Since an external strong electric field is caus-
ing such emission and does not depend on the thermal properties of
the metal, the metal can even be at zero temperature and hence the
process is called field emission or cold emission.
The theory of field emission from bulk metals was first proposed
by Fowler and Nordheim in an epoch-making paper in the proceed-
ings of the Royal Society of London during 1928 (Fowler & Nord-
heim 1928; see also Stern, Gossling & Fowler 1929; Jensen 1995;
Forbes & Deane 2007; Liang & Chen 2008 for further discussion).
Fowler–Nordheim tunnelling is the wave-mechanical tunnelling of
electrons through a triangular-type barrier produced at the surface
of an electron conductor by applying a very high electric field.
Cold emission or field emission processes not only have signif-
icance in terrestrial laboratories but are also found to be equally
important when considering the electron emission processes from
cold and compact stellar objects, such as neutron stars. In the
case of a rotating neutron star, the existing large magnetic field,
≥1012 G (Shapiro & Teukolsky 1983) for conventional radio pulsars
or ≥1015 G for a magnetar inner field (Mereghetti 2009), produces


C 2012 The Authors

Monthly Notices of the Royal Astronomical Society 

C 2012 RAS
2 A. Ghosh and S. Chakrabarty
a strong electric field at the poles, approximately given by F ∼ 2 × field at the polar region, which is produced by the strong mag-
108 P−1 B12 V cm−1 , and this is parallel to B at the poles (Shapiro & netic field of rotating neutron stars. Taking this physical picture
Teukolsky 1983). Here P is the time period of the neutron star in s into consideration, in this article we have developed formalisms for
and B12 is the measure of magnetic field strength in units of 1012 . both non-relativistic and relativistic scenarios of field emission for
In proximity to the polar region of a strongly magnetized neutron electrons from the poles of neutron stars with 1010 ≤ B ≤ 1017 G.
star, the potential difference changes almost linearly with distance In the next section we study the effect of a strong quantizing mag-
from the polar cap, which is a region very close to the magnetic pole. netic field on the field emission of electrons for the non-relativistic
The repulsive surface barrier, in combination with this attractive case. In Section 3, we repeat the same calculation for the relativistic
potential, forms a triangular-type barrier at the poles. Therefore if scenario. In the last section we give our conclusions and review the
electron emission from the poles of neutron stars is of field emission future prospects of this work.
type, then the Fowler–Nordheim equation with proper modification
may be used to investigate the emission process. Now, the study of
plasma formation in a pulsar magnetosphere is quite an old astro- 2 E F F E C T O F S T RO N G Q UA N T I Z I N G
physical issue but still unresolved, in particular for the magneto- M AG N E T I C F I E L D O N C O L D E M I S S I O N :
spheres of strongly magnetized neutron stars/magnetars (Molofeev, N O N - R E L AT I V I S T I C S C E N A R I O
Malov & Teplykh 2004, 2005; Istomin & Sobyanin 2007). In the To develop the modified version of the field emission process for
formation of magnetospheric plasma, it is generally assumed that electrons in the non-relativistic scenario in the presence of a strong
there must be an initial high-energy electron flux from magnetized quantizing magnetic field, we have followed the basic calculation
neutron stars. Since the magnetic fields at the poles of neutron presented in the seminal Royal Society paper by Fowler & Nord-
stars/magnetars are strong enough, the emitted electrons flow only heim (1928). In the modified version, we assume a cylindrical co-
along the magnetic field lines. The flow of high-energy electrons ordinate system (ρ, θ , z) and constant magnetic field B along the
along the direction of the magnetic lines of force and their pene- positive z-direction, with the usual gauge for the vector potential
tration through the light cylinder is conventionally pictured using A = (B × ρ). Following Fowler & Nordheim (1928), we assume
current-carrying conductors. Naturally, if the conductor is broken that the triangular-shape surface potential is given by V(z) = C −
near the pulsar surface then the entire potential difference will de- Fz, which changes linearly with the z-coordinate. Here C is the
velop across a thin gap, called the polar gap. This is of course based constant surface barrier and F (absorbing the magnitude of elec-
on the assumption that above a critical height from the polar gap, tron charge e, we replace eF by F) is the driving field for electron
because of high electrical conductivity of the plasma, the electric emission from the poles. Consideration of a linear-type surface bar-
field F parallel to the magnetic field near the poles is quenched. rier potential no doubt has some historical importance. This type
Further, a steady acceleration of electrons originating in the polar of potential barrier was first used in the original work of Fowler
region of neutron stars, travelling along the field lines, will pro- and Nordheim. However, there are two other important reasons to
duce magnetically convertible curvature γ -rays. If these curvature consider it: since this is the first time the problem has been solved
γ -ray photons have energies >2me c2 (where me is the electron rest in the presence of a strong quantizing magnetic field, we have con-
mass and c is the velocity of light), then pairs of e− –e+ will be pro- sidered the simplest such triangular-type potential barrier in order
duced in enormous amounts with very high efficiency near the polar to obtain an analytical solution. Our intent is to solve the problem
gap. These produced e− –e+ pairs form what is known as magne- analytically in a way in which the physical meaning of the problem
tospheric plasma (Ruderman 1971; Ruderman & Sutherland 1975; is not lost. The other reason behind such a choice is that in the case
Michel 1982, 2004; Shapiro & Teukolsky 1983; Jessner, Lesch & of strongly magnetized rotating neutron stars/magnetars the electric
Kunzl 2001; Harding & Lai 2006; Diver et al. 2009). Cold emission field produced at the polar region is approximately constant for uni-
therefore plays a significant role in magnetospheric plasma forma- form magnetic field strength in that region and constant rotational
tion. In turn, the motion of charged particles in the magnetosphere period of the object. We believe that for the potential barrier at
in the presence of a strong magnetic field causes pulsar emission the poles both assumptions are approximately valid. Then, assum-
in the form of synchrotron radiation. Therefore the cold emission ing the conservative force-field relation dV/dz = −eF, we obtain a
process also indirectly affects the intensity of synchrotron radiation. triangular-type potential barrier at the poles. Under such a situation,
Furthermore, exactly solvable models with a simple-type tun- the Schrödinger equation satisfied by the electrons confined within
nelling barrier lead to equations (Fowler & Nordheim 1928; Stern the matter (in this case within the neutron star/magnetar crustal mat-
et al. 1929) that underestimate the emission current density by a ter) is given by (throughout the paper for the sake of convenience
factor of 1000 or more. If a more realistic-type barrier model is we assume natural units, i.e.  = c = 1)
used by inserting an exact surface potential in the simplest form of    
the Schrödinger equation, then a complicated mathematical prob- 1 1 ∂ ∂ψ 1 ∂ 2ψ ∂ 2ψ
− ρ + 2 +
lem arises with regard to the resulting differential equation. It is in 2m ρ ∂ρ ∂ρ ρ ∂θ 2 ∂z2
principle, therefore, mathematically impossible to solve the equa-  2 2 2 
ieB ∂ψ e B ρ
tion exactly in terms of the usual functions of mathematical physics, − + − E ψ = 0. (1)
2m ∂θ 8me
or in any simple way.
Moreover, to the best of our knowledge neither the non-relativistic For the electrons just liberated through quantum mechanical tun-
nor the relativistic versions of cold emission processes in the nelling, one has to consider the potential V(z) along with E. The
presence of a strong quantizing magnetic field, relevant for elec- relevant equation is given by
tron emission from the poles of strongly magnetized neutron    
1 1 ∂ ∂ψ 1 ∂ 2ψ ∂ 2ψ
stars/magnetars, have been properly investigated even with simple- − ρ + 2 +
type potential barriers. In the conventional pulsar model it is gen- 2m ρ ∂ρ ∂ρ ρ ∂θ 2 ∂z2
 2 2 2 
erally assumed that the emission of electrons and thereby the for- ieB ∂ψ e B ρ
− + − E + V (z) ψ = 0, (1a)
mation of the magnetosphere is mainly caused by a strong electric 2m ∂θ 8me
C 2012 The Authors

Monthly Notices of the Royal Astronomical Society 

C 2012 RAS
 Strong quantizing magnetic fields 3
where the energy eigenvalue E is given by equations (4) and (5) for Then we have, from equation (6),
two different physical situations. If we assume a separable solution
d2 f0
for equations (1) and (1a), satisfied by freely moving electrons mov- + yf0 = 0. (10)
dy 2
ing under the potential V(z), the wave functions can be represented
by Now, following equation (8) of Fowler & Nordheim (1928), the
solution for this equation is given by
ψ(ρ, θ, z) = φnρ ,m (ρ, θ)fν (z), (2)  
(2) 2 3/2
where for equation (1) the longitudinal part is of plane-wave type, f0 (y) = y 1/2 H1/3 y , (11)
3
whereas for equation (1a) we shall evaluate f ν (z) using the technique
(2)
discussed below. Since there is no potential associated with the where H1/3 (x) is the Hankel function of the second kind, of order
transverse motion for the electrons, the free transverse part of the 1/3 with argument x. As we shall see below, the quantity Q defined
wavefunction is given by in Fowler & Nordheim (1928) after equation (12) is much larger in
  our formalism in the presence of a strong quantizing magnetic field,
exp(imθ ) −1−|m| (| m | +nρ )!
φnρ ,m (ρ, θ) = ρ unless the direction of spin polarization for the emitted electrons
(2π)1/2 0 2|m| nρ ! | m |! is opposite to the direction of the external magnetic field. In the
   2 
ρ 2
ρ absence of an electron spin term or polarization along the direc-
× ρ |m| exp − 2 Lnρ ,m , (3)
4ρ0 2ρ02 tion of the magnetic field, the factor Q is virtually infinitely large
(Q ≈ ∞). We shall further show that this infinitely large Q will
where ρ 0 = (2/eB)1/2 is the Larmor radius, the radius of the lowest
make the transmission probability of electrons vanishingly small.
Landau orbit, and Lnρ ,m is the Laguerre polynomial. The energy
In our formalism, the factor Q is defined as
eigenvalue is given by  
2 C − w 3/2
pz2 Q = (2me F )1/2 , (11a)
E= + μB B(2nρ + ν ± ν + 1) (4) 3 F
2me
and is related to the Hankel function argument by the relation
without electron spin; with the inclusion of electron spin, it will be    
(2) 3
pz2 H1/3 exp − πi Q at z = 0. (11b)
E= + μB B(2nρ + ν ± ν + 1) ∓ μB B, (5) 2
2me
Now the transmission probability for electrons as defined in Fowler
where μB = e/2m, the Bohr magneton. For the motion of electrons
& Nordheim (1928) is given by
along the z-direction, the Schrödinger equation satisfied by f ν (z)
with and without potential V(z) can be obtained by averaging equa- | a |2 − | a  |2
D(w) = . (12)
tions (1) and (1a) respectively over the transverse wavefunction | a |2
φnρ ,m (ρ, θ). Then for newly tunnelled-out electrons in the lowest Since the transmission coefficient D is related to the Q factor by the
Landau level with nρ = ν = m = 0, moving along the z-axis under relation D ∼ exp (−2Q) and the field emission current for electrons
the influence of the potential V(z) and having no spin contribution, from the zeroth Landau level is related to D by the integral
we have  ∞
eB pz
1 d2 f0 R= f (w)D(w) dpz , (12a)
− + V (z)f0 = (E − μB B) f0 = wf0 . (6) 2π2 0 me
2m dz2
with f (w) the Fermi distribution function, then it is quite obvious
For the same kind of liberated electrons with spin polarization in
that in the electron field emission current the polarization effect will
the negative direction of the z-axis we have w = E and for the
come only from the transmission coefficient D(w). In the following,
polarization along the direction of magnetic field w = E − 2μB B.
we therefore focus our study on the investigation of the properties
On the other hand, for the electrons confined within the crustal
of the emission coefficient. For large Q, we have from equation (18)
matter of magnetars and moving freely we have V(z) = 0. The
of Fowler & Nordheim (1928) in our modified form
corresponding Schrödinger equation is given by  

[w(C − w)]1/2 4 C − w 3/2

d2 f0 D(w) ≈ exp − (2me F )1/2
+ wk2 f0 = 0, (7) C 3 F
dz2
where wk = (2mw)1/2 is the equivalent electron momentum along [w(C − w)1/2 ]
= exp (−2Q) . (13)
the z-axis and w can have three possible values as mentioned above. C
Then, following equation (10) of Fowler & Nordheim (1928), we Let us now analyse the argument part of the exponential. Following
can write down the solution of equation (7) for free electrons inside Fowler & Nordheim (1928), we put C = μe + wf , where μe is
the crustal matter in the form the electron Fermi energy and wf = wc × (B/Bc(e) )1/2 in eV is the
1   work function (see Ghosh & Chakrabarty 2011) for the emission
f0 = 1/2 a exp (iwk z) + a  exp (−iwk z) , (8)
wk of electrons along the direction of the magnetic field, where wc ≈
82.93 and Bc(e) ≈ 4.43 × 1013 G, the typical value of magnetic field
where a is the probability amplitude for electrons moving along strength at which the Landau levels for the electrons are populated
the positive direction of the z-axis (incident part) whereas a is the in the relativistic scenario. Again, as defined before, the quantity
corresponding quantity for left-moving waves (reflected from the w can have three possible values. To obtain an order-of-magnitude
surface barrier). Next, for the electrons just tunnelled out, we make estimate for the terms containing the Bohr magneton and work
the following coordinate transformation: function, we assume that the emitted electrons carry the maximum
 
C−w possible energy, i.e. the electron Fermi energy for the temperature
y= − + z (2me F )1/3 . (9)
F T −→ 0 limit. Then C − w = wf + β × μB B, where the parameter

C 2012 The Authors

Monthly Notices of the Royal Astronomical Society 

C 2012 RAS
4 A. Ghosh and S. Chakrabarty
β = 0, 1 or 2 for a spin polarization opposite to the direction of
the external magnetic field, in the conventional direction of spin
polarization in the presence of a strong magnetic field or with no
spin term or spin polarization along the direction of the magnetic
field, respectively. Taking into account the denominator F of the
argument, expressing in terms of the magnetic field strength as
defined at the beginning of the Introduction and using μB ≈ 5.79 ×
10−15 MeV G−1 , the numerical value for the Bohr magneton, we
have, approximately, from the expression for Q as defined above,
 3/2
−1/2
× 10−6
1/2
Q ≈ hf β × 0.5 + wc hf × 107 = QB + Qwf ,
(14)
where hf = B/Bc(e) , and QB and Qwf are the contributions from
the Bohr magneton or spin term and the work function part respec-
tively. It is quite obvious that the contribution from spin term QB
is extremely large for β = 1 or β = 2 and makes the transmis-
sion coefficient exactly zero. Physically, this means that if we do
not consider electron spin or assume an electron spin polarization
along the direction of the magnetic field, the electron transmission
coefficient and in turn the electron transmission current vanish ex-
actly, i.e. there will be no field emission under such situations. On
the other hand, the second term, the work-function part, unlike the
first term gives a finite contribution to the transmission coefficient.
The first term, i.e. the spin term, will make Q extremely high, even
if we do not assume that the emitted electrons are at the top of
their Fermi level. This particular term is therefore making the cold
emission probability of electrons from the poles of strongly mag-
netized neutron stars exactly zero for non-zero values of β. For the
transmission of spin-polarized electrons β = 0 and consequently
QB = 0, therefore Q = Qwf . In this situation the electron energy
eigenvalue obtained from equation (5) for nρ = ν = m = 0 is given
by
pz2
E= . (15)
2me
In this case there will be electron field emission with spins polar-
ized opposite to the direction of the external magnetic field. From
the above expressions for electron energy eigenvalue, it is quite
obvious that for our model of cold emission of polarized electrons
the rest of the mathematical formulation will be almost identical to
that of Fowler and Nordheim in the presence of a strong quantizing
magnetic field. We have further noticed that the values for trans-
mission coefficient D and transmission current R remain finite and
large enough for magnetic field strength ≤1015 G. Of course in our
model the magnetic field dependence of Q will come from the work
function wf and the field intensity F and only the polarized electrons
are allowed to tunnel through the surface barrier. Then, following
Fowler and Nordheim, we have obtained the cold electron emission
current from equation (12a) at T −→ 0 using Sommerfeld’s lemma.
While calculating the emission current numerically, we have used
the exact form of D as given in Fowler & Nordheim (1928) after
equation (16). In Fig. 1 we have shown schematically the variation
of cold electron current with the strength of electric field intensity
at the poles; this is represented by the solid curve. Since the electric
field at the poles is produced by a rotating magnetic field, we have
also shown in the same graph with a dashed curve the variation
of electron cold current with intensity of the magnetic field. Since
the electric field intensity varies linearly with the magnetic field
strength, the qualitative nature of the curves is almost identical. For
neutron stars with very low surface magnetic field, the produced
electric field, which acts as a driving force, is also small enough
and therefore the electron field emission current will be extremely
Figure 1. (i) The dashed curve shows the variation of electron field current
plotted along the left-hand vertical axis, with the neutron star magnetic field
plotted along the x-axis at the bottom and expressed in terms of critical
(e)
field Bc for electrons. (ii) The solid curve shows the variation of the same
quantity plotted along the right-hand vertical axis with the electric field in
V cm−1 produced by the rotating magnetic field of a magnetar plotted along
the horizontal axis at the top.
small, as shown in the curve. For neutron stars with moderate sur-
face magnetic field strength (B ∼ 1012 –1015 G), the field current is
quite high. Now for those objects with large surface magnetic field
(B  1015 G), the work function will also become large. There-
fore beyond some maximum value for electron field current, since
the work-function part dominates over the driving electric force at
the poles, the electron field current will decrease and will become
vanishingly small. This is the case for the objects with ultrahigh sur-
face magnetic field (1015 G). In the figure, we have not shown the
variation of electron field current with magnetic field strength (com-
puted at the surface) for a particular neutron star. Therefore each
magnetic field/electric field point corresponds to a particular type
of compact magnetized object, with surface magnetic field varying
from very low to ultrahigh. From the curves it is quite obvious that
for magnetars with surface magnetic field ∼1015 G the electron field
current is quite high and very close to the peak value. Unlike the
original work of Fowler and Nordheim, the tunnelling coefficient
does not follow an exponential law. However, the variation of cold
current with electric field strength can be obtained numerically. The
numerically fitted functional form is given by
1/2
R = 0.26F24 exp (−9.8 × F14 ), (16)
−2 −14
where R is the field current in A cm , F 14 = 10 F and F 24 =
10−24 F.
3 E F F E C T O F S T RO N G Q UA N T I Z I N G
M AG N E T I C F I E L D O N C O L D E M I S S I O N :
R E L AT I V I S T I C S C E N A R I O
In the relativistic scenario we have repeated the non-relativistic
calculation for the cold emission transmission coefficient. In this
section we have considered the emission of high-energy electrons
from the polar regions of strongly magnetized neutron stars with
magnetic field 1012 ≤ B ≤ 1017 G. The potential is introduced by
C 2012 The Authors
Monthly Notices of the Royal Astronomical Society 
C 2012 RAS
 Strong quantizing magnetic fields 5
hand in the Dirac equation using standard relativistic hadro-dynamic where as before a is the probability amplitude for electrons moving
technique. Following Fowler and Nordheim, here also we have along the positive direction of the z-axis (incident part), whereas
considered a triangular-type potential barrier at the polar region. As a is the corresponding quantity for left-moving waves (the part
in the previous case, we have considered a cylindrical coordinate reflected from the surface barrier). Assuming that the interface be-
system and the choice of gauge is the same for A. The radial part tween the crustal matter of strongly magnetized neutron stars and
of the upper component satisfies the equation the magnetosphere is at z = 0, the wavefunction and its derivatives
  must be continuous at z = 0 (Fowler & Nordheim 1928), i.e.
∂2 1 ∂
βλ + 2 + − k ρ R(ρ) = 0,
2 2
(17)    
∂ρ ρ ∂ρ [fν (0)]II = [fν (0)]I and fν (0) II = fν (0) I . (25)
where βλ2 = E 2 − m∗2 − pz2 + 2λk with λ = ±1 for up and down Using the relation Hl (u) = 2νHl−1 (u), we have
spin states respectively, k = eB/2, E is the energy eigenvalue for the  2
u
Dirac equation and m∗ = m + V(z) = m + C − Fz is the effective a + a  = N α 1/2 exp − 0 Hl (u0 ) (26)
electron mass, with m∗ = m for free electrons when V(z) = 0. The 2
solution of the above equation is given by and
   2
t u
R(ρ) = N exp − Ln (t), (18) iα 1/2 (a − a  ) = N exp − 0 [u0 Hl (u0 ) − 2lHl−1 (u0 )] , (27)
2 2
where Ln (t) is the Legendre polynomial of order n, t = k2 ρ 2 and N is where u0 = u(z = 0) = m∗ /F 1/2 . These two conditions may be
the normalization constant. For up and down spin states, the energy rearranged in the following form:
eigenvalues are given by E↑ = [pz2 + m2 + 2neB]1/2 and E↓ =
a + a  = X and a − a  = iY , (28)
[pz2 + m2 + 2(n + 1)eB]1/2 respectively. In general we may write
Eν = (pz2 + m2 + 2νeB)1/2 . Then, following the same averaging where X and Y are two real quantities. Hence it is straightfor-
technique as discussed in the previous section for the non-relativistic ward to verify from equation (12) that in the relativistic scenario
case, the wavefunction of electrons corresponding to the motion the transmission coefficient vanishes exactly. Therefore, from the
along the z-direction is given by analysis of this section, we may conclude that if the barrier in com-
bination with the external electrostatic driving force behaves like a
d2 fν
+ (E 2 − m∗2 − 2νeB)fν = 0. (19) scalar-type potential and is triangular in shape at the surface, then
dz2 relativistic electrons cannot tunnel through the surface barrier, what-
Using the transformation ever their kinetic energies and the strength of the external electric
m + C − uF 1/2 field.
z= , (20)
F
with the new variable being u, the above equation for f ν reduces to
4 CONCLUSION
d2 fν
+ (α 2 − u2 )fν = 0, (21) Our study of the non-relativistic scenario of cold electron emission
du2
in the presence of a strong magnetic field is believed to be the first
where α = (E2 − 2νeB)1/2 /F 1/2 , which is the well-known form
attempt in this direction. While obtaining the electron transmission
of differential equation for a one-dimensional quantum mechanical
probability in the non-relativistic scenario under the influence of a
harmonic oscillator. With α 2 = 2l, we have E2 = 2(νeB + lF) and
strong electric field at the poles, we have noticed that in our theo-
the solution is given by
 2 retical formalism emission is allowed if we take electron spin into
u account and also the electrons have conventional spin polarization,
fν,l = iÑ exp − Hl (u) = (fν )I (for example), (22)
2 i.e. one opposite to the direction of the external magnetic field. Even
if all these criteria are satisfied, at extremely high magnetic field
where Ñ is the normalization constant and H l (u) is the Hermite
strength, since the electron work function becomes large enough,
polynomial of order l. This spinor solution f ν,l is for those electrons
the transmission coefficient drops to zero. Since there are no stellar
that have already been liberated out through the surface into a
objects with surface magnetic field strength 1015 G, such vanish-
vacuum under the influence of electric field F (here liberated out
ingly small transmission coefficients will not be possible in reality.
from the crustal matter of strongly magnetized neutron stars or
A low charge density magnetosphere will therefore only be possible
magnetars into the magnetosphere through the polar region).
if super-exotic compact stellar objects with an ultra-strong surface
The equation satisfied by free electrons bound within the crustal
magnetic field exist in nature. In the case of a conventional magne-
matter by the barrier potential at the surface can be obtained by
tar or strongly magnetized neutron star, the electron field current is
putting V(z) = 0 and is given by
quite high, very close to the peak value, which is also obvious from
d2 fν Fig. 1. As a result there will be enough curvature gamma photons
+ α 2 fν = 0, (23)
dz2 produced by energetic electrons, which in turn produce an enormous
where in this case α = (E2 − m2 − 2νeB)1/2 is the free electron amount of e− –e+ pairs to form a normal magnetospheric plasma. In
momentum along the z-axis of energy E. Now, following the nota- our model the only difference from the non-magnetic or low mag-
tion of Fowler & Nordheim (1928), we express the solution for free netic emission cases is that the primary electrons are spin-polarized
electrons within the system, confined by the surface barrier V(z), in in the direction opposite to the direction of the external magnetic
the form field. As a consequence, for the conventional magnetar case the re-
sults obtained in Beloborodov & Thompson (2007) on the formation
1  
fν = 1/2 a exp (iαz) + a  exp (−iαz) = (fν )II (for example), of coronae of magnetars will not be seriously affected. Therefore,
α in the magnetar magnetosphere of a neutron star with B ≤ 1015 G,
(24)

C 2012 The Authors

Monthly Notices of the Royal Astronomical Society 

C 2012 RAS
6 A. Ghosh and S. Chakrabarty
the primary electrons are all spin-polarized along −B. Since the
electron emission current is almost zero for objects with ultrahigh
magnetic field strength (1015 G) then, if the electron emission
process in such exotic objects is dominated by cold emission, the
charge density of e− –e+ plasma in the magnetosphere will be ex-
tremely low. As a consequence there will be very weak synchrotron
emission in the radio waveband. The other possible mechanisms
by which e− –e+ plasma can be produced in the magnetospheres of
magnetars are (i) thermoelectric emission of electrons from the po-
lar region and (ii) photo-emission from the same region. However,
the work function in the polar region of a typical magnetar is several
GeV, whereas the temperature can be at most a few hundred MeV
for a young magnetar. Thermionic emission will therefore be sup-
pressed by the Boltzmann factor exp (−wf /kT), whereas in the case
of photo-emission the energy of the induced photon must be of the
order of GeV (γ -photons). In the polar region, if any such photons
exist to liberate electrons, they must have been produced as curva-
ture photons by high-energy electrons moving along the magnetic
lines of force. Since the possibility of such electrons is very rare,
the number of high-frequency photons is also vanishingly small; as
a result there will be almost no creation of secondary e− –e+ pairs
in the photo-emission process.
In conclusion, from our relativistic formalism of cold emission
of electrons we can state that relativistic electrons populating the
neutron star interior cannot be extracted from cold emission from the
poles of a neutron star, independently of the magnetic field strength.
Non-relativistic electrons with antiparallel spin can be extracted for
standard (observed) values of magnetic field strengths, but cannot
be extracted from the surfaces of objects with B  1015 G (in the
eventuality that such exotic objects can exist).
REFERENCES
Beloborodov A. M., Thompson C., 2007, ApJ, 657, 967
Diver D. A., da Costa A. A., Laing E. W., Stark C. R., Teodoro L. F. A.,
2009, preprint (arXiv:0909.3581)
Forbes R. G., Deane J. H. B., 2007, Proc. R. Soc. London, 463, 2907
Fowler R. H., Nordheim L., 1928, Proc. R. Soc. London, 119, 173
Ghosh A., Chakrabarty S., 2011, JA&A, 32, 377
Harding A. K., Lai D., 2006, Rep. Prog. Phys., 69, 2631
Istomin Ya. N., Sobyanin D. N., 2007, Astron. Lett., 33, 660
Jensen K. L., 1995, J. Vac. Sci. Technol., B13, 516
Jessner A., Lesch H., Kunzl T., 2001, ApJ, 547, 959
Liang S.-D., Chen L., 2008, Phys. Rev. Lett., 101, 027602
Mereghetti S., 2009, preprint (arXiv:0904.4880)
Michel F. C., 1982, Rev. Mod. Phys., 54, 1
Michel F. C., 2004, Adv. Space Res., 33, 542
Molofeev V. M., Malov O. I., Teplykh D. A., 2004, in Camilo F., Gaensler
B. M., eds, Proc. IAU Symp. 218, Young Neutron Stars and Their
Environments. Astron. Soc. Pac., San Francisco, p. 261
Molofeev V. M., Malov O. I., Teplykh D. A., 2005, Astron. Rep., 49, 242
Ruderman M., 1971, Phys. Rev. Lett., 27, 1306
Ruderman M. A., Sutherland P. G., 1975, ApJ, 196, 51
Shapiro S. L., Teukolsky S. A., 1983, Black Holes, White Dwarfs, and
Neutron Stars – The Physics of Compact Objects. John Wiley & Sons,
New York, p. 288
Stern T. E., Gossling B. S., Fowler R. N., 1929, Proc. R. Soc. London, 124,
699
This paper has been typeset from a TEX/LATEX file prepared by the author.
C 2012 The Authors
Monthly Notices of the Royal Astronomical Society 
C 2012 RAS