Advanced Beam Dynamics & Hamiltonian

Formalism

D
1. Introduction

Rick Baartman, TRIUMF

January 29, 2015

R. Baartman, TRIUMF 2015

 Introduction
~ y, z, t) and
Electrons are transported using electric and magnetic fields E(x,
~
B(x, y, z, t). These fields exert forces on charged particles according to

F~ = eE~ + e~v × B
~ (1)

e is the electron’s charge, and ~v its velocity.

We know F~ = d~
p
dt , so all the dynamics can be derived... Can they?

R. Baartman, TRIUMF 2015 1

There are problems with this approach.

• What kind of coordinate system to use?

• Grid size for field spec?

• Time step?

(Extreme) example: LHC beam is < 1 mm on average size (< 0.1 mm at

collision) but ring is 10 km across.

R. Baartman, TRIUMF 2015 2

 Electric field
Units are volts/metre or the like.

Kinetic energy gained by an electron

on the lower plate = Work done on the
particle = Force × Distance

F~ · d~l = eEs = e V
R

s s = eV.

Kinetic energy is given by ∆K = 12 mv 2

So equating, we get 12 mv 2 = eV, or,

r
2eV
v= . (2)
m

R. Baartman, TRIUMF 2015 3

 Example:
Across the plates, V = 30, 000 volts (usu. called 30 kV).

What is the energy and speed at which the electron hits the positive plate?

We find ∆K = eV = 1.6 × 10−19 coul × 30, 000 joules/coul =4.8 × 10−15 joule.

The mass is m = 9.1 × 10−31 kg.

The p
speed is p
v = 2eV /m = 2 × 4.8 × 10−15/9.1 × 10−31 metre/second = 1.0 × 108 m/s.

One third of the speed of light! So even with modest voltages, we are
approaching relativistic speeds.

R. Baartman, TRIUMF 2015 4

 Example again, (common nuclear units):
No nuclear/accelerator/particle physicist ever uses joules and kilogrammes for
units. Since energy and mass are equivalent by E = mc2, mass is expressed
in energy units. Further, since charge is quantized and with particles we
always deal with one or a few units of charge, a more natural energy unit is the
‘electron-volt’ (eV).

1 eV is the energy gained by a unit of charge falling through 1 volt. So

1 eV = 1.6 × 10−19 coul × 1 volt = 1.6 × 10−19 joule.

Back to our example:

∆K = eV = (1 e) × (30 kV) = 30 keV. m = 0.511 MeV/c2

q q
2×30,000
β = vc = mc
2eV
2 = 511,000 = 0.343.

R. Baartman, TRIUMF 2015 5

 Electrostatic steering:
Let’s operate the acceleration plates sideways on the beam:
How to calculate the angle θ?
F = eVs/s as before, but now F =
max where x is used to emphasize the
acceleration is in the x or transverse
direction.
(By convention, the particle’s main or longitudinal direction is z and the
transverse coordinates are (x, y).)
d2 x 2 d2 x ∆vx ax L
ax = dt2
= v0 dz2 . We want θ = v0 = v02
.

eVs ∆vx eVs L eVs L

ax = ms , so θ = v0 = mv02 s
= 2K s , where K is the kinetic energy.

R. Baartman, TRIUMF 2015 6

This can be simplified. Recall that the electron’s kinetic energy depends upon
the voltage of the place from which it came: K = eV0 where V0 is the potential
difference between the electron source (cathode) and the location at which
the calculation is being made (the steering plates in this case). Therefore,
eVs L
θ = 2eV 0s
, or,
Vs L
θ= (3)
V0 2s
Example: Say plates are 2 inches long, separated by 1 inch, there is 1 kV
across the plates, and the electrons passing through come from a cathode
which is at 30 kV.

Then
1 kV 2 in 1
θ= × = = 1.9◦. (4)
30 kV 2 × 1 in 30

In accelerator physics, one never uses degrees for angles. The common unit
is the milliradian (mrad), which is basically 0.001 radian. So in the present
case, θ = 33.3 mrad.

R. Baartman, TRIUMF 2015 7

Excursion: radians are dimensionless.

To think about: Why does the electron mass not occur in the formula for θ?

R. Baartman, TRIUMF 2015 8

 Relativistically...
We add energy eV (charge times voltage) to the rest energy mc2. Then

mc2 + eV V
γ= = 1 + (5)
mc2 511000 Volts

and p
2 2
1 + (βγ) = γ , so βγ = γ2 − 1 (6)
and r
1 1
β = 1 − 2, γ = p (7)
γ 1 − β2

Exercise 1: Re-calculate the previous case of 30 keV electrons to find

relativistically correct value of β.

R. Baartman, TRIUMF 2015 9

 Magnetic field:

F~ = e~v × B
~ (8)
This force:

• increases with v. This means that above some energy magnetic control of
the beam is more economical than electrostatic.

• is perpendicular to ~v . This means a magnetic field cannotR be used to do

work on the electrons and increase their energy. (∆K = F~ · d~l)

Units = Tesla (or, Weber per metre squared). 1 Tesla = 10,000 gauss. We
know that vB has same units as electric field. This means 1 Tesla metre/sec.
= 1 newton/coul.

R. Baartman, TRIUMF 2015 10

Typical room temperature magnets have fields on the order of 1 Tesla. This is
because the saturation field of steel is 2 Tesla. Superconducting magnets
have fields up to the order of 10 Tesla.

Excursion: Tesla and Volts

R. Baartman, TRIUMF 2015 11

 Particle inside a transverse magnetic field:
Clearly, the force is a centripetal force
and the electron will follow a circular
arc. What radius ρ?

Centrifugal force mv 2/ρ is balanced by the centripetal force evB.

Or, Bρ = mv/e = momentum/charge. This is a very important quantity, often
called ’magnetic rigidity’. We can calculate it non-relativistically as
√ p p
mv 2mK 2
2(mc /e)(K/e) 2V0(mc2/e)
= = = (9)
e e c c

Notice the dependence on only the source voltage and the mass-to-charge
ratio.

R. Baartman, TRIUMF 2015 12

But this is an approximation. Surprisingly, the formula Bρ =
momentum/charge is still correct relativistically.

Bρ = p/e = βγmc/e = 0.00170 Tesla-metre βγ = 17.0 Gauss-metre βγ (10)

p
Example: At 30 keV, βγ = (1 + 30/511)2 − 1 = 0.347, Bρ = 5.91 Gauss-m.
A field of 5.91 Gauss will bend with radius 1 m. A field of 1 Tesla will bend with
a radius of 0.59 mm.
Electron ‘Energy’ Momentum β Bρ
0.003 MeV 0.055 MeV/c 0.11 1.85 G-m
0.03 MeV 0.178 MeV/c 0.?? 5.9 G-m
0.3 MeV 0.63 MeV/c 0.78 21. G-m
3 MeV 3.47 MeV/c 0.989 116. G-m
30 MeV 30.5 MeV/c 0.99986 1018. G-m

R. Baartman, TRIUMF 2015 13

 Magnetic steerer:

Calculate θ. We know from geometry

θ = L/ρ:

L BL
θ= = (11)
ρ (Bρ)

Example: Let’s find how much a 300 keV electron deflects in traveling 1 metre
in the earth’s magnetic field (assumed 0.6 gauss).
0.6 Gauss×1 metre
θ= = 29 mrad.
21 Gauss-metre

This is LARGE. A 29mm deflection after 1m!

R. Baartman, TRIUMF 2015 14

 Simple DC accelerator

Put a hole in one plate: (from J.D.Jackson)

R. Baartman, TRIUMF 2015 15

Write Laplace equation in polar coordinates for azimuthal symmetry:

∂ 2Φ
 
1∂ ∂Φ
ρ + 2 =0 (12)
ρ ∂ρ ∂ρ ∂z

Expand near axis,

ρ2 d2Φ(0, z)
Φ(ρ, z) = Φ(0, z) − 2
+ ... (13)
4 dz

All we need is the function Φ(0, z); the on-axis potential.

Exercise 2: By direct substitution, find the next term in the expansion above.
Exercise 3: In the thin-lens approximation, find the focal length of a circular
aperture with asymptotic electric fields in the particle’s direction of zero on one
side and constant E0 on the other.
Excursion: What is the ‘thin lens approximation’?

R. Baartman, TRIUMF 2015 16

 Mathematica Code
(Reference URL: eqns. 3.105,3.106.)
2 2
 2 2

λ:=z + ρ − 1; R:=Sqrt λ + 4z ; Φ[ρ , z ] =
(Abs[z] + z)/2 + (Sqrt[(R − λ)/2] − Abs[z]ArcTan[Sqrt[2/(R + λ)]])/π;
ContourPlot[Φ[ρ, z], {ρ, −2, 2}, {z, −2, 2}, Contours → 100,
ColorFunction → “Pastel”]
Plot3D[−Φ[ρ, z], {ρ, −2, 2}, {z, −2, 2}, ColorFunction → “Pastel”]

R. Baartman, TRIUMF 2015 17

Can we reach any voltage we like by cascading this type of accelerator?

Answer: No (Why not? How do we get around this difficulty?)

R. Baartman, TRIUMF 2015 18

 Axially Symmetric Lenses – Einzel

Einzel lens is typically 3 electrodes, of which outer 2 are ground:

We can construct such an axially symmetric lens, using the math above
adding 4 circular apertures:

R. Baartman, TRIUMF 2015 19

R. Baartman, TRIUMF 2015 20
 Axially Symmetric Lenses – Solenoid

Let us try to find the motion equations

for this case starting from the Lorentz
force equation (1). A constant magnetic
field B in the direction z of the
particle motion. Paraxial particles have
components of momentum in the x and
y, and px, py  pz .

R. Baartman, TRIUMF 2015 21

Then

dpx
= evy B (14)
dt
dpy
= −evxB (15)
dt
dpz
= 0 (16)
dt

We set p~ = m~v , (come back to this later). Then we find vz is constant and
vx2 + vy2 is constant, and each varies harmonically. Thus particles execute
helical orbits.

For a finite length solenoid, this analysis is actually INCORRECT. Can you
think of a conservation law that it violates?

R. Baartman, TRIUMF 2015 22