Latex Notes
Latex Notes
1 Introductory Remarks
This lecture is intended to be a brief introduction to what I consider to be the
principal ”must know” characteristics of plasma. It is, in no way, intended to
be a comprehensive discussion of the topic. For more advanced introductions
to plasma physics, there are several good resources: eg. Introduction to Plasma
Physics (F. Chen), Plasma Physics (R, Goldston, P. Rutherford). There are
also free online lecture notes of Intro to Plasma courses: R. Fitzpatrick at UT
Austin and R. Parker at MIT (linkable on the pdf of this document).
2 Plasma Characteristics
When gas becomes ionized it becomes a plasma. Typically, what we consider to
be a plasma is actually not fully ionized. In many cases, only a small fraction of
the gas is ionized. These are called (not surprisingly) weakly ionized plasmas,
as opposed to fully ionized plasmas (deep in the sun or inside a magnetically
confined fusion device). The degree of ionization is determined by the Saha
Equation:
ni T 3/2 −Ui /kB T
≈ 2.4 × 1021 e (1)
nn ni
Where ni and nn are the density of the ions and the neutrals in [m−3 ], T is the
gas temperature in Kelvin, kB is Boltzmann’s constant and Ui is the ionization
energy, that is, the energy required to remove the outermost electron. As a
comparison, at standard temperature and pressure, nitrogen has a degree of
ionization of:
ni
≈ 10−122 . (2)
nn
As the temperature starts rising to the order of Ui (that is, to around a few
thousands degrees K), the ionization becomes non-negligible and the gas be-
comes a plasma.
1
3 Review of basic mechanics equations
Disregarding magnetic forces, the basic equation of motion of a given particle of
mass m1 and electric charge q1 when it comes a distance r1,2 to another charged
particle of mass m2 and charge q2 is given by the equation:
" #
Gm 1 m2 q 1 q2
m1~a = ΣF~ = F~G + F~E = − 2 + 2 r̂ (3)
r1,2 4π0 r1,2
where F~G is the gravitational attraction (hence the minus sign) and F~E is the
electrical force. G and 0 are the gravitational constant and the permittivity of
free space respectively. Assuming particle 1 is an electron and particle 2 is a
Deuterium isotope, then the ratio between the forces is:
FE
= 1.1 × 1039 , (4)
FG
therefore, for laboratory plasmas, gravitational forces can be disregarded and
we can focus only on electric and magnetic forces, otherwise called the Lorentz
Force. Note that gravity IS important for astrophysical plasmas due to the low
degree of ionization and size of the systems.
For a particle of mass m and charge q moving with a velocity ~v through an
electric and magnetic field of magnitudes E ~ and B~ respectively, the equation of
motion of the particle is:
h i
F~ = m~a = q E ~ + ~v × B
~ (5)
This is the equation we will use when analyzing the mechanics of individual
particles in the plasma.
2
F~ = ~
eE
~
E
x x̂ x
Figure 1: Moving the center of mass of the electrons with respect to the ions
creates a restoring force
3
Figure 2: The pressure difference between two vertical positions is related to
the weight of the slab between them.
Not surprisingly this is the electron plasma frequency of the system. Analo-
gously, for an ion pof charge Ze and mass mi , the ion plasma frequency can be
defined as: ωpi = (Z 2 e2 ni )/(mi 0 ). Let’s look at it in a little more detail: If
you look at the thought experiment, for the same displacement, the total charge
in each slab will increase as you increase the electron number density ne , hence
the force is stronger and our oscillation is faster. Also, for the same field, the
acceleration on electrons is greater than that of ions because the same force
~ is excerpted on such disparate masses. This explains the inverse relation
(eE)
on mass.
4
where n is the number density of the gas.
If the column is subject to gravity, the pressure at h1 should differ from that
at h1 + ∆h just by the pressure exerted by the weight of the slab between the
2 heights:
But the mgh numerator in the exponent of the RHS of Equation 18 is simply
the potential energy of the particles at that position (for example, the potential
energy of an electron in a potential Φ: P.E. = −eΦ). So Equation 18 can be
generalized as:
−P.E.
n(~x) = n0 e kT (19)
assuming the potential energy is a function of a generalized position ~x. This is
a powerful equation and we will come back to it in later sections.
Now that we have the dependence of density on position (as a proxy of the
potential energy) in a thermalized gas, let’s explore its dependence on velocity.
We can assume that the dependencies are separable. That is,
where f and g are functions of only position and velocity respectively. Going
back to the column of thermalized gas, we will now explore a different question:
What is the relationship between the number of particles that cross vertically
upwards at 2 different planes h1 and h2 as shown in Figure 3? It is clear that not
all particles that cross h1 will reach h2 since some will not have enough energy.
This is the reason that it’s more tenuous at higher P.E. In order to reach, the
particles will have to have a vertical velocity ~v greater than a minimum value ~u
such that: 1/2mu2 = mg∆h. In other words:
5
Figure 3: The particles crossing plane h2 upwards must have had a vertical
velocity of at least u (where mg∆h = 1/2mu2 ) as they crossed plane h1
where N(h2,v>0) is the number of particles crossing plane h2 upwards per unit
time, and N(h1,v>u) is the number of particles crossing plane h1 upwards per
unit time with vertical velocity greater than u.
We can look now at the number of particles crossing both planes with the same
velocity restriction: v > 0, that is, we can compare N(h2,v>0) and N(h1,v>0) .
As stated in Equation 20, the density can be separated between the positional
dependence and the velocity dependence. Since we’ve imposed the same velocity
restriction (v > 0):
N(h2,v>0) n(h2) −mgh −1/2mu2
= = e kT = e kT (22)
N(h1,v>0) n(h1)
Finally, we can substitute Equaiton 21 into Equation 22 to get:
N(h1,v>u) −1/2mu2
= e kT (23)
N(h1,v>0)
N(v>u) −1/2mu2 −1/2mu2
= e kT → N(v>u) ∝ e kT (24)
N(v>0)
where we have used, again, the separable nature of the density. That is, the
positional dependence cancels out. At any height, the number of particles cross-
2
ing the vertical plane upwards per unit time is proportional to exp(− 1/2mu
kT ).
Equation 24 is independent of position and should apply everywhere in space.
We can now introduce the probability distribution function: f (v) such that
f (v)dv will be the probability that the particles have a velocity between v and
v + dv. Equation 24 sets a constraint on f (v). But the number of particles
crossing a given plane per unit time, T , with speed v > u are not just:
Z ∞
N(v>u) 6∝ f (v)dv (25)
u
6
as one could intuitively think, since this would not take into account that within
a unit time T , particles with velocity v will only cross the plane if they’re within
a distance vT beneath the plane (slow ones have to be closer than fast ones).
For a given velocity v, the number of particles a distance vT below the plane
would be ∝ (vT )f (v), therefore, for all v > u we have the equation:
Z ∞
−1/2mu2
N(v>u) ∝ vf (v)dv ∝ e kT (26)
u
R∞
Using the normalization: −∞
f (v)dv = 1, Equation 26 leads to:
r
m − mv2
f (v) = e 2kT (27)
2πkT
7
Figure 4: Maxwell-Boltzmann distribution function of Argon gas at different
temperatures. The x̂-axis is proportional to the speed of the atoms. Note that
the area under the curve must be 1.
p
vpeak = 2kT /m. There is, therefore, a characteristic speed of the particles
which we call the thermal speed defined as:
r
kT
vt ≡ (33)
m
Now,
p the electrons and
p ions will have their own thermal speeds given by: vte =
kTe /me and vti = kTi /mi . How do these speeds typically compare?
Let’s say an electron and an ion are getting energy from an electric field (which
is often the case) for a given amount of time t. The momentum gained by both
particles is the same: me ve = mi vi = eEt. so ve /vi = mi /me 1. This
disparity translates to vte and vti hence, if energy transfer between species is
low, Te /Ti ∝ mi /me 1. Note that even if the particles have enough time
to reach thermal equilibrium, which is more common in magnetically confined
plasmas, Te = Ti still leads to vte vti .
8
or [length]/[time] in vt . Therefore, we can immediately deduce a characteristic
length called the Debye length:
q
kT
s
vt m kT 0
λD ≡ =q 2 = (34)
ωp q n q2 n
m0
More generally, the Debye length for a single species ion of charge Ze is defined
as: s
k0
λD = (35)
e2 (ne /Te + Z 2 ni /Ti )
which can be derived from a more detailed analysis of Poisson’s equation to be
explored later. Nonetheless, when we can take the ions as stationary (particu-
larly in weakly ionized cold plasmas), the Debye length is effectively taken as
the electron Debye length:
r
kTe 0
λD = . (36)
e2 ne
To get a more intuitive picture of what the Debye length is related to, we can
go back to the thought experiment where we now have a picture of an electron
that is subject to a simple harmonic oscillator system. If we were to follow the
motion of the electron in this simple picture, it would follow a harmonic motion
of the form:
x = A cos (ωpe t) (37)
where I have disregarded any phase and I still haven’t determined it’s amplitude.
How can we determine the amplitude A of oscillation? If we take the time
derivative of Equation 37, we can find the velocity of the electron:
But we know that the speed of the electrons is around vthe (of course, this is a
characteristic speed), so we can use that as the constraint and we have
9
Figure 5: The charge density of a system with a point charge Q immersed in a
plasma
where n0e and n0i are the electron and ion densities far from the point charge
(Φ → 0) where quasi-neutrality prevails, so n0e = Zn0i . Assuming eΦ kT :
eΦ
ne ≈ n0e 1 + (43)
kTe
ZeΦ
ni ≈ n0i 1 − (44)
kTi
Z 2 enoi Φ
en0e Φ
ρp ≈ −e n0e + − Zn0i + (45)
kTe kTi
e2 Φ n0e Z 2 noi
ρp ≈ − + (46)
k Te Ti
e2 n0e z 2 noi
Q 1 Q
∇2 Φ(r) = + Φ(r) − δ(0) = 2 Φ(r) − δ(0) (47)
k0 Te Ti 0 λD 0
where we have used the definition of Debye length from Equation 35. We can
10
Figure 6: The plasma shields the point charge and lowers the electric potential
close to the charge
before we solve this, note that if the second term in the LHS were not there, we
would recover the equation for the electric potential of a point charge: Φc (r) =
Q
4π0 r . The solution of Equation 48 is of the form:
Qe−(r/λD )
Φ(r) = (49)
4π0 r
Figure 6 illustrates the effect the presence of the plasma has on the potential
as compared to a point charge in empty space. In the presence of a positive
charge, the electrons will quickly try to shield it, but since they are moving so
fast (vte ) and in all directions, there is a region close to the charge where the
electrons will escape (due to their own inertia) and not completely shield it.
This region, where the electric fields are not completely shielded, is called the
sheath and its length is of the order of the Debye length.
11
System ne [m−3 ] Te [eV ] ωpe [s−1 ] λD [m]
Interstellar gas 106 1 105 10
Solar Wind 107 10 105 10
Van Allen belts 109 102 106 1
Ionosphere 1011 10−1 107 10−2
Solar Corona 1013 102 108 10−3
Candle flame 1014 10−1 109 10−4
Neon lights 1015 1 109 10−4
Gas Discharge 1018 2 1011 10−5
Process Plasma 1018 102 1011 10−4
Fusion Experiment 1019 103 1011 10−4
Fusion Reactor 1020 104 1012 10−4
Lightning 1024 3 1014 10−8
Electrons in metal 1029 10−2 1016 10−12
12
Figure 7: The AM spectrum is well below the ≈ 10M Hz ωpe of the ionosphere,
leading to their reflection. FM waves, at higher frequency, penetrate it.
13
as the distance traveled by an electron (on average) before it collides with an-
other electron.
Every electron has an associated volume in the system ∆V = 1/ne , that is, ∆V
is the volume that is occupied by each electron. Therefore, we can define the
electron-electron cross section, σee , with the equation ∆V = σee λmf p , as shown
in Figure 8. Therefore, using Equations 50 and 33:
Not surprisingly, the larger the temperature, the closer the electrons can ap-
proach. Putting Equations 54 together with 52 and 51, we get the final result:
p e2 2 ne e4
νe ∝ ne Te /me = 1/2 3/2 (55)
Te me Te
since we know that many assumptions have been made, we’ve disregarded the
constants.
Since the resistivity, η, is proportional to the collisional frequency, we get the
very important result:
η ∝ Te−3/2 (56)
that is, the plasma becomes a better conductor as the temperature goes up.
This dependance is of great importance in astrophysical plasmas as well as in
tokamak plasmas.
As a point of comparison, the resistivity in a metal is well known to increase
with temperature (contrary to the case in plasmas).
14
Figure 9: An electron with speed vthe colliding with a stationary one can be
viewed in the center of mass frame and the distance of closest approach can be
derived from conservation of energy.
6 Magnetized plasmas
Finally, we’ll do a small introduction to what happens when we incorporate
effects of magnetic fields on the plasma. As shown in Equation 5, the force of a
particle which is moving in a magnetic field is of the form:
~ → ~a = q (~v × B)
F~ = m~a = q~v × B ~ (57)
m
Suppose a positively charged particle of mass m and charge q is moving in
~ pointing
the plane of the paper with velocity ~v and there is a magnetic field B
into the paper. As shown in Figure 10, the force, hence the acceleration of the
particle is always pointing towards a center of motion and the particle draws a
circular orbit in the plane. From Equation 57, the magnitude of the acceleration
is a = qvB/m. But we know from kinematics that if a particle is rotating around
a fixed point, the acceleration must be centripetal and the magnitude should
be:
v2 qvB vm
a= = →r= (58)
r m qB
If the particle that is rotating is an electron (ion) with speed vte (vti ) then the
radius of rotation is called the electron (ion) gyro-radius or Larmor radius and
the equations are as follows:
me vte mi vti
ρe = , ρi = (59)
eB ZeB
Finally, we can figure out the frequency of rotation of an electron or ion that is
rotating at thermal speeds: vt = ωc ρ. These frequencies are very important in
magnetized plasma physics and are called electron and ion gyro-frequencies (or
15
Figure 10: Trajectory of a positively charged particle moving with a velocity ~v
where there is a magnetic field pointing into the page.
cyclotron frequencies):
If the particles are not confined to the plane perpendicular to the magnetic
fields but can move in three dimensions, the particles move freely in the direc-
tion parallel to the magnetic field but are confined to move in circular orbits
perpendicular to the fields, therefore, they trace spiral orbits around the mag-
netic fields, as shown in Figure 11.
16
electron
Ion
Figure 11: In three dimensions, particles follow spiral trajectories around mag-
netic fields.
17