Notes from PlasmaX

Dominik Perfi Sta

nczak

May 8, 2015

These notes will not be 100% comprehensive, as Im making them mainly for my own use. However, if
you spot any mistakes, feel free to catch me on the forums and Ill fix any mistakes.

Week 1. Description of the plasma state, with Paolo Ricci

Didnt start making notes until 1.5 so Ill be skimming the earlier topics.

1.1

Plasmas in nature and laboratory

Plasma - the 4th state of matter. Heat stuff up to 11400K (= 1eV) and gases begin being ionized.
The Sun is a miasma of incandescent plasma1
Lightning is plasma (ionized air)
Plasma diplays
Nuclear fusion - cant really get there without turning stuff into plasma
The word plasma comes from greek , which means moldable substance or jelly, though
it was mentioned on the forums that it might mean living thing... which is really fitting when you
think about it
A brief history:
1920s-1930s: ionospheric plasma research (for radio transmission) and vacuum tubes (Langmuir)
1940s: MHD plasma waves (Alfven)
1950s: research on Magnetic Fusion. Geneva UN conference on uses for atomic energy which
dont kill people
Fusion experiments: L-1, TFTR, JET, ITER tokamaks; W7-X stellarator at MPI in Germany; the
NIF inertial fusion facility in US
The Earths magnetosphere; van Allen belts
Jets - space plasmas
Lots of industrial applications

1.2

Rigorous definition of plasma: Debye length

A plasma is a globally neutral ionised gas with collective effects

The following parameters classify plasmas:
Debye length
Distance over the potential of a charged particle decreases by a factor 1/e due to screening by other
charged particles
1

https://www.youtube.com/watch?v=sLkGSV9WDMA

r
De =

0 Te
e2 n0

(for electrons)
Solved in lecture by a statistical approach which assumed n 34 3De ND  1 (for a Debye sphere; in
the lecture n3De was used, which relates to a Debye cube. Theres not much difference between them,
a factor of 4). ND means the number of particles inside a sphere (or cube, following the lecture) of
radius equal to the Debye length. The condition means theres plenty of particles to screen our test
particle. This also assumed that binary interactions between particles were weak ( e
Te  1)

1.3

Plasma definition: frequencies and parameters

Plasma frequency Assume a plasma of same density of ions and electrons. Displace electrons by
x. They begin to exhibit harmonic oscillations (for x not too large). Newtons 2nd law gives
n0 e 2
d2 x
+
x = 0
dt2
0 me
Can define plasma frequency
s
pe

vth,e
n0 e 2
=
0 me
De

where vth,e denotes the thermal speed of electrons

Collision frequency
The frequency of coulomb collisions between particles
coll

n0 e 4
3
1620 m2e vth,e

Size of plasma has to be much larger than its Debye length (or theres no quasineutrality)

1.4

Particle motion in a static uniform magnetic field . Plasma magnetic properties

Larmor radius - particles gyrate around the guiding center at this distance
mv

|q| B
Cyclotron frequency
c

v
|q| B
=

Particle rotation direction on their helical trajectory

q > 0 (by default): left hand rotation with respect to B
q < 0 (electrons): right hand rotation
Magnetic moment
2
|q| e 2 mv
Ekin
=
=
2
2B
B
(direction opposite to B) is an adiabatic invariant for every particle; doesnt change under slow
changes of factors involved in the equation for . However, it will change through heat exchange,
which usually operates on slower timescales than magnetic field changes (see 2c).
Plasmas are diamagnetic (they reduce externally applied magnetic fields) (because of direction
of )

|| IA =

1.5

Particle motion in given electromagnetic fields: the drifts

Static and uniform E and B fields. Particles under Lorentz force which can be decomposed as:
Parallel direction:
m

dvk
= qEk
dt

Uniform acceleration
Perpendicular direction:
m

dv
= q(E + V B)
dt

The many drifts in a plasma:

E B drift
Perpendicular component averages out over gyroperiod
ve =

E B
B2

This is a motion of the guiding center which is superposed over the gyromotion
Does not depend on charge, neither in magnitude nor in direction (but gyromotion direction
does)
Guiding center moves over lines of constant electrostatic potential (the drift does not change
the particle energy!)
A generalization of this drift for any force:
vF =

F B
qB 2

For a gravitational force (say, space plasmas), this depends on charge. Separates positive and
negative charges. Polarizes the plasma, creating a E field and an E B drift
Curvature drift
B field curved, particle follows the B field - this happens through a centrifugal force
Fc =
This causes a drift:
vd =

mvk2
2
RB

RB

2
mv
Fc B
=
2 (RB B)
qB 2
qB 2 RB

~
Gradient drift B B)
Happens in changing (spatially) magnetic fields
v B =

mv 2
(B B)
2qB 3

 A derivation so complicated, it deserved a separate appendix. As particles gyrate, they move

between regions of smaller and bigger B. This causes a drift in a direction perpendicular to both
the B field and the gradient of its value We consider a small variation in B and expand B in a
taylor series around B0 .
Then we use that expansion to solve m dv
dt = qv B, plugging in our expansion for B.
We also decompose the velocity: an average v0 and a small perturbation. v0 is the solution to
the equation for constant magnetic field B0 .
We neglect the cross product of the two small perturbations and average over a gyroperiod.
We use our knowledge of the solution for the static magnetic field (gyration in the plane perpendicular to B) to deal with the perpendicular velocities (x and y in this decomposition under
the assumption that B is along z).
The drift velocity is the perturbation described by the formula above for an arbitrary geometry
of the problem.

1.6

Plasma confinement based on single particle motion. Magnetic mirrors, stellarators, tokamaks

How do you confine a plasma?

Charged particles follow helical trajectories along B field. This confines them in the perpendicular
direction. What about the parallel one?
Can use open field lines. Take two circular coaxial electromagnets.

Can use closed field lines. Closed geometries. Example: tokamaks (toroidal), stellarators.

 The magnetic mirror geometry is neat for particles really close to the axis. B is maximum (field
density increases) near the electromagnets
force in the axial direction is
Fz = |B|
vk has to vanish at Bmax so that the kinetic energy is just composed of the perpendicular component
of velocity
Particle reflection condition

2
v
Bmin
>
Bmax
+ vk2

2
v

This means that particles in the loss cones in phase space (marked red; those which dont satisfy
the inequality) cannot be confined in the mirror!
Neat example: the Earths magnetic field is a magnetic mirror!

 What about closed magnetic field lines? Can those deal with loss cones?
B is not homogeneous! Curved! Has curvature and gradient drifts!
For a purely toroidal field, positively charged particles drift towards the bottom, while negatively
charged ones drift towards the top. This polarizes the plasma and introduces the EB drift outwards,
sending the plasma crashing into the major radius wall.
A solution: a poloidal magnetic field to short circuit the charge accumulation. Either:
Drive a current through the plasma Tokamaks
Get rid of axial symmetry Stellarators

Week 2: Kinetic description of plasmas, with Paolo Ricci

2.1

2a) From single particle to kinetic description

Kinetic description of plasma. A (relatively?) complete description of plasma which covers both the
particles and the fields evolving over time.
The usual diagram for a plasma description, seen often in simulations:
(a) Take Newtons equations using electric and magnetic fields for all particles at all times (use
Lorentz force)
(b) Use positions and velocities to compute charge and current densities. Charge density given as sum
over particles of their charges, localized through use of Dirac delta functions. Current density similar, but multiplied by particle velocity vectors inside the sum.
(c) Take charge and current density, plug them into Maxwell equations, calculate E and B fields at
positions

(d) Take calculated E and B fields and apply them as forces to particles. Repeat cycle until bored or
simulation returns segmentation fault.
But real plasmas involve on the order of 102 1 particles for a fusion plasma. Too much strain on our
computational abilities. Impractical. We use a distribution function:
f (r, v, t)drdv = number of particles at time t, in phase space volume drdv located at r, v. We have a
separate distribution function fi for every species
Total number of particles NS given by integral of distribution function over all positions and velocities
(which covers all the phase space)
Number density of particles ns given by integral over all velocities for a given location r
R
Average velocity given by n1s vfi (r, v, t)dv
Examples of distribution functions
Maxwell-Boltzmann distribution function, for three dimensions
F0 (v) = n0 (

1
2vthermal

)3/2 exp(

v2
2
2vthermal

In 1D, only the normalization of the distribution changes from the 3D case:
F0 (v) = n0 (

1
2vthermal

)1/2 exp(

v2

)
2
2vthermal

Monoenergetic beam in 1D
F0 (v) = n0 (v v0 )
Two counterstreaming beams in 1D (two-stream instability!)
F0 (v) =

n0
[(v v0 ) + (v + v0 )]
2

Conservation of particles number

If there are no sources or sinks, we have the following condition for conservation of number of particles
dfs
= 6D (ufs )
dt
where we introduce the six-dimensional nabla operator because whos gonna stop us
6D = (
u=(

d d d d
d
d
d d
, , ,
,
,
)=( ,
)
dx dy dz dvx dvy dvz
dr dv

Flongrange + Fshortrange
dr dv
F
,
) = (v,
) = (v,
)
dt dt
ms
ms

Long range forces - collective interactions. Short range forces - binary collisions (between individual
particles, like youd have in a gas). Plugging these back into the particle conservation equation:
Flongrange + Fshortrange
dfs
d
d
= (vfs )
[
fs ]
dt
dr
dv
ms
Boltzmann equation We can improve on the previous equation. Start out with the expanded particle
conservation equation:
Flongrange + Fshortrange
dfs
d
d
= (vfs )
[
fs ]
dt
dr
dv
ms

 In the phase space approach, velocity is treated as a completely independent variable than v (though
d
s
you could consider one as a derivative of the other). Thus dr
(vfs ) = v df
dr
long range force can be decomposed into electric field independent of v, and the v B term d
s
perpendicular to v. Thus, dv
[Flongrange fs ] = Flongrange df
dv
Plugging in:
Fshortrange
dfs Flongrange dfs
d
dfs
= v

(
fs )
dt
dr
ms
dv
dv
ms
Can be rewritten as:
Fshortrange
dfs
dfs Flongrange dfs
d

fs )
+v
+
=
(
dt
dr
ms
dv
dv
ms
Term on the right is called a collision operator ( df
dt )c .
And we get the Boltzmann equation:
dfs
dfs
qs
dfs
dfs
+v
+
(Elongrange + v Blongrange )
= ( )c
dt
dr
ms
dv
dt

2.2

2b) Coulomb collisions in plasmas. Bonus module.

We use Boltzmann equation and look into the short range interactions
An electron with charge e approaches a positive ion (assumed immobile) with charge Ze. Electron trajectory changes. ve - initial electron velocity b - impact parameter, shortest distance between extrapolated
line of initial electron trajectory and ion position
Ze2

Coulomb interaction energy

40 b2 1
Kinetic energy
me ve
(similar to one so collision interaction is important)
b

Ze2
= b/2
40 me ve2
2 4

e
Coulomb cross section: /2 = b2/2 = (4Z
2 2 4
0 ) me ve
Collision frequency: /2 = ni ve 2 = f racni Z 2 e4 (40 )2 m2e ve3
Is this a correct estimate? Do collective small angle deflections matter in a plasma? How can we take
the interaction with many particles into account properly? Average over all phase space somehow?
Take the electron-ion collision again. Denote - angle between initial and final electron velocity.
Particles interact through Coulomb force. Angular momentum and energy - conserved (if electron is
e
much lighter than ion, m
mi  1).

tan(/2) =

b/2
Ze2
=
b
40 me ve2 b

b/2 - impact parameter at which collision deflects electron by 90 .

Cumulative effect for many collisions? Imagine electron moving towards ion cloud.


2 
Due to symmetry we take hve i = 0 but ve
6= 0 . So magnitude could change, but there will
be no preferred direction. stands for parallel to initial velocity.

2  Z
d ve
= dbni ve 2b
dt
(we integrate over all possible impact parameters
2
ve
= ve2 sin 2 = ve2 tan /22 [1 + tan /22 ]2

Plugging into the integral:

2 
Z D
(b/2 /b)2 b
d ve
3
dbb
= 8ni ve
dt
(1 + (b/2 /b)2 )2
0
We neglect quantum effects (thus integrating from 0) and integrate up to Debye length as coulomb interactions are screened beyond it. Finally, we get:

2 
d ve
D
= 8ni ve3 b2/2 ln
(if D  b/2
dt
b/2
Following section may have some 4s swapped for s.
Note that electrons do not lose much energy as me  mi . Basically reflected balls from a wall. Thus
2
=0
ve (vke ) + 0.5ve

And


d vke
D
= 4ni ve2 b2/2 ln
dt
b/2

We define the coulomb logarithm:

ln ln

D
In most plasmas equals 15 to 25
b/2


d vke
= ei ve
dt
Collision frequency of electrons against ions:
ei = 4ni b2/2 ve ln = ni ei ve
Whereas
ei = 4b2/2 ln
Can compare
b2/2
/2
=
1
ei
4b2/2 ln
Much smaller than 1! So small angle deflections dominate over large scale deflections!

2.3

2c) Collisional processes in plasmas

2.3.1

Slowing down of an electron beam


d vke
ni Z 2 e4 ln
= ei ve =
dt
420 m2e ve3

We could use this to calculate how an electron beam slows in a plasma. Assume a Maxwellian distribution
of electron velocities with mean velocity ue  vthermal,e in 1D:
fe (v) = n0 (

 1
due
= ei vke =
dt
n0

me
2vthermal,e

)1/2 exp(

me (vke ue )2
)
2
2vthermal,e

Z
ei vke fe (vke )dvke ' hei i ue (if ue  vthermal,e )

The average collision frequency between electrons and ions ei is

2
ni Z 2 e4 ln
ei =
12 ( 3/2) 2 m(e 1/2)Te( 3/2)
0

There are also collisions between electrons coming from the beam and electrons in the plasma:

2
ne e4 ln
hei i ne
ee =

(
(
(
Z 2 ni
12 3/2) 2 me 1/2)Te 3/2)
0

2.3.2

Plasma resistivity

Take a cloud of ions and electrons. Apply electric field E. Ions will move in direction of E, whereas
electrons will move in opposite direction. E then drives a current in a plasma - charges are moving!
We neglect the slow and heavy electrons and focus on electron movement. From Newtons second law:
due
= ene E + Rei
dt
is the collision term we have just calculated. This slows down the current.
me ne

Rei

Rei = me ne hei i (ue ui ) (assuming ue  vth,e )

After a transient, well reach steady state operation and

d
dt

=0

The current can be depicted as j = ne e(ue ui )

Thus:
e2 ne E = me hei i j
E=

me hei i
j j
e2 ne

By comparison with Ohms law we can define the plasma resistivity:

me hei i
2me Ze2 ln

=
3/2
e 2 ne
12 3/2 20 Te
The bigger the temperature, the lower the resistivity. Unlike in metals. Its also independent of density!
The contributions of increasing the number of carriers and increasing the number of collisions cancel each
other out exactly.

2.3.3

Overview of plasma collision frequencies

Electron - ion collision frequency nuei =

Electron - electron collision frequency nuei
Ion - ion collision frequency nuii .
Ions gain energy when you fire an electron beam into a plasma (could be heated this way?).
me ve = mi vi
0.5mi |vi |2 =


m2 
m2e
|ve |2 e ve 2 
2mi
2mi

(as we can ignore the change in parallel electron velocity)

Rate of exchange of energy (between species! This equalizes the temperatures between electrons and
ions!):

ni Z 2 e4 me ln
me
Z
hE i =
hei i
3/2
mi
3 22 mi Te
0

The electrons have a similar, very fast rate of collisions with each other and with ions. The rate of
collisions between ions happens 40 times slower, and then the rate of energy exchange is 40 times slower
than that. At a similar rate to that of energy exchange is the rate of ions colliding with electrons.

2.4
2.4.1

2d) Vlasov equation

Derivation from Boltzmann equation
dfs
qs ~
dfs
dfs
dfs
+v
+
(E + ~v B)
= ( )c
dt
dr
ms
dv
dt

If we can assume that the number of particles in a Debye cube is REALLY HIGH: n3D , so that
( df
dt )c = 0, then the Vlasov equation holds:
dfs
dfs
qs ~
dfs
+v
+
=0
(E + ~v B)
dt
dr
ms
dv
E and B here represent the long range interactions. The charge density is computed as indicated before,
integrating out all the velocities. The currents are likewise obtained by summing over the species and
calculating the average velocities at each positions.
2.4.2

Conservation laws for the Vlasov equation

dfs
dfs
qs ~
dfs
+v
+
(E + ~v B)
=0
dt
dr
ms
dv

This satisfies the following conservation properties:

Number of particles - we can integrate the Vlasov equation over all positions and velocities. Intes
grating dfdts gives us dN
dt . The second term gives us, by means of Gauss (divergence) theorem and
pushing the boundaries out to infinity, where fs should decay to zero, zero. In the third term we
have a velocity divergence. Since no particles have infinite velocities2 , we can once more use the
s
divergence theorem (in velocity space!!!) to eliminate the third term and we reach dN
dt = 0. Particles
are conserved.
2

Einstein says hi.

 Momentum, which we calculate as the sum of particle and field momenta. No actual derivation is
given except for the formula:
Ptot =

Z
ms

Z
dr

dv vfs + 0

dr(E B) = const

Total energy can be decomposed into energy of particles and energy of field.
Z
Z
XZ
1
1
2
E=
dr dv ms v fs +
dr(0 E 2 + B 2 /0 ) = const
2
2
s
Entropy, as given by information theory
S=

XZ

Z
dr

dvfs ln fs = const

This is because collisions are neglected by the Vlasov equation. Therefore it is time-reversible!
2.4.3

Interpretation of Vlasov equation

fs has incompressible motion (in phase space) - it can be considered as an incompressible fluid
(moving in phase space - its going to obey Liouvilles theorem)
As seen by particle along orbit
dfs
fs
fs
F fs
fs
fs
qs
fs
=
+v
+

=
+v
+
(E + v B)
dt
t
r
ms v
t
r
ms
v
And the funny thing is, thats just the Vlasov equation itself. Thus along particle orbits, fs = 0.
2.4.4

(Formal) solutions of Vlasov equation

If cj is a constant of motion, then any distribution function being a function of any number of constants
of motion cj is a solution.
This is unlike the Boltzmann equation, where only the Maxwellian distribution was a stationary solution.
It can be really difficult to find constants of motion, as well. It seems to be implied that practical
solutions rely on numerical methods - that way you need not specify constants of motion for a formal
solution.

2.5

The two stream instability!

We make this our testing ground for the Vlasov equation. The situation is two beams of electrons moving in
opposite directions in 1D. Spoiler alert - WERE ACTUALLY GOING TO MESS AROUND WITH THE
MATLAB CODE FOR THIS KIND OF SIMULATION IN THE HOMEWORK, THIS IS AWESOME.
Ahem.
2.5.1

Simplifications

We take the Vlasov equation and Maxwells equations

dfs
dfs
qs ~
dfs
+v
+
(E + ~v B)
=0
dt
dr
ms
dv

E=
0

B=0
dB
E=
dt
dE
)
dt
We simplify the situation. Set B = 0 for an electrostatic situation.
B = 0 (j + 0

dfs
dfs
qs ~ dfs
+v
+
E
=0
dt
dr
ms
dv

E=
0
B=0
E=0
which implies
E =

=
0
We also assume that ions are stationary in the background with density n0 (electrons move at a much
faster time scale). This means we only have to solve Vlasov for the electron motion and can replace fs by
f for brevity:
~ df
df
df
eE
+v

=0
dt
dr ms dv
Z
e
e
=
f dv n0
0
0
2.5.2

Linearisation

Were going to be analysing small perturbations from equilibrium. This means basically expanding the
quality of interest in a short Taylor series:
g = g0 (equilibrium) + g1 (perturbation, g1  g0 )
In our case, f0 being the initial distribution functions which is isotropic over all position space (thus
no dependence on r) and f1 being our small perturbation:
f (r, v, t) = f0 (v) + f1 (r, v, t)
= 1 (r, t) as we can set 0 = 0 since its constant anyway
E = E1 (r, t)
We plug these into the Vlasov equation:
f0 + f1
(f0 + f1 ) eE1 (f0 + f1 )
+v

=0
t
r
me
v
Simplifying and neglecting E1

f1
v

as small times small:

f1
f1 eE1 f0
+v

t
r
me v

Also plugging in E1 = :

e
1 =
0

Z
f1 dv

Thus:
f1
f1
e 1 f0
+v

t
r
me r v
Z
e
1 =
f1 dv
0
We now apply Fourier analysis to f1 :
Z
Z
f1 (r, v, t) = dk d f1 (k, v, ) exp i(k r t)
This is neat because:
The time derivative is simple:
f1
(r, v, t) =
t

Z
dk

d(i)f1 (k, v, ) exp i(k r t)

The spatial derivative is also pretty simple:

Z
Z
f1
(r, v, t) = dk d(ik)f1 (k, v, ) exp i(k r t)
r
So we go back to the Vlasov equation and plug in the expressions above:
f1
f1
e 1 f0
+v

=0
t
r
me r v
Z

Z
dk

d[(i + ik v)f1 +

ie1
f0
k
] exp i(k r t) = 0
me
v

This can be true only if all the coefficients vanish:

ie1
f0
(i + ik v)f1 +
=0
k
me
v
1
f0
e
k
f1 =
me k v
v
Plugging this result into the Fourier transform of the Poission equation above:
Z
e
1 =
f1 dv
0
Z
Z
0
k f
e
e2 1
v
k 2 1 =
f1 dv =
dv
0
0 me
kv
Which then implies
1 k 2 [1 +

e2
0 me k 2

0
k f
v
dv] = 0
kv

We denote the bulky part as the dispersion function

e2
D(, k) 1 +
0 me k 2

0
k f
v
dv
kv

Which we can find the roots of, and the solutions (values of and k) gives us the normal modes of our
plasma.
Theres a singularity at = k v. This is when particles match velocities with wave velocities in the
plasmas... this may or may not be connected to Landau damping. This topic will not be further developed
in the course.
2.5.3

Getting to the two-stream instability

The two-stream instability is thermodynamically weird. The velocity distribution is non-maxwellian. Its
just two sharp peaks (low entropy). Could there be intrinsic modes in the system that restore thermodynamic equilibrium (high entropy)?
We take a 1D system. The derivation above is general (plenty of vectors brought to you by yours truly).
f = f0 (vx ) = f (u)
e2 1
D(, k) = 1 +
0 me k

df0 du
du ku

Im pretty sure theres a mistake in the lecture and a 1/k factor was lost there (it should be 1/k 2 ).
For two counter streaming beams:
n
[(u v0 ) + (u + v0 )]
2
The distribution function luckily avoids the singularity. We calculate the dispersion function and
arrive at
f0 (u) =

D(, k) = 1

ne2
1
1
me [
+
]
20
( kv0 )2 ( + kv0 )2
2

2 = ne )
(note that the plasma frequency pops up: pe
0 me
This is a 4th order polynomial in the nominator. The function has two vertical asymptotes at = kv0
and a horizontal one at 1.
Depending on the parameters, if D( = 0, k) 0, theres 4 real roots. The modes will be oscillatory
instead of exponentially growing.
Otherwise, if D( = 0, k) < 0, theres 2 real roots corresponding to oscillations and 2 complex roots
corresponding to exponential explosions.

D( = 0, k) = 1

2
pe
2
< 0 k 2 v02 < pe
k 2 v02

Unstable modes are those that have sufficiently long wavelengths. Thats all we can tell analytically.
ITS PARTICLE IN CELL TIME!

2.6
2.6.1

2f ) Kinetic plasma simulations. The Particle in Cell method

The various time scales in a plasma

1010 s - electron cyclotron motion

107 s - ion cyclotron motion
105 s - microturbulence
103 s - fast global instabilities
101 s - slow global instabilities

 1s - energy confinement time

103 s - gas equilibrisation
Its extremely hard to simulate all of these at once! Accurate modelling of cyclotron motion (say, 1014
timestep) would mean observing fast global instabilities after 101 1 timesteps. Thats a lot of timesteps.
2.6.2

Simulation approach - Particle in Cell (PIC) method

One could solve the entire Vlasov equation by separating the whole 6D phase space into a grid and solving
the partial differential equation. Possible. But usually, one uses the particle in cell method.
f is constant along particle trajectories. This translates directly into Newtons equations for a single
particle:
dr
=v
dt
dv
q
= (E + v B)
dt
m
In the PIC (Particle in Cell) method, we approximate the distribution function by a discrete sum
f (r, v, t) '

f (r r0 , v v0 )

The f s are called superparticles; they have compact support - theyre zero everywhere but on a
small domain in phase space aroundR rR0 and v0 .
We introduce the integral I =
drdvf which is done over the small domain of the particle.
At all times, the superparticles satisfy
X
f (r, v, t) '
f (r r (t), v v (t))

If we have that

where

dr
= v
dt
q
dv
= (E + v B )
dt
m
Z Z
1
Ef dr dv
E =
Id
Z Z
1
B =
Bf dr dv
Id

and we have some initial conditions

r (t = 0) = r0
v (t = 0) = v0
Then the distribution function defined as the sum of all superparticles (their distribution functions)
also solves the Vlasov equation.3
3

This is a complicated way of saying that we can use particles to approximate the motion of the plasma - see Birdsall and
Langdon.

2.6.3

Practical PIC

In one dimension, we have to:

(a) Solve Poissons equation to get the potential and field E
This is done by discretizing space and time. We get discrete approximations to charge densities,
potentials and fields.
Derivatives are discretized using central differences.
(b) Evaluate electric fields acting on superparticles
E = Ej if |xj x | x/2
(c) Apply fields to superparticles. Solve for the motion of superparticles by a numerical algorithm
(Eulers or Runge-Kuttas)
(d) Assign charge to discrete locations on grid. We sum the charges of superparticles in the j-th cell (in
xj1/2 < x < xj+1/2 ) and divide by the cell size.
On a closing note: PICs are awesome, theyre used in all kinds of fields. Electromagnetic fields,
gravitational fields... go learn them. :)