Transverse Beam Dynamics

C. Biscari

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6

 Outlook

Charged particle and magnetic fields

Linear approximation
Dipoles and quadrupoles
Lorentz equation
Horizontal plane
From Lorentz equation to Hill’s equation
From Hill’s equation to Twiss parameter definition
Matrix representation
Storage ring – periodic solution
Stability diagram
Chromaticity
Sextupole correction

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 2

 The accelerator from the particle point of view is a sequence of

Drifts – No external fields – Particles go straight

Magnetic fields – Particles are bent according to the magnetic rigidity
Electric fields – Particles go straight, gain or loose energy

Particle trajectory

B E B Reference orbit

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 3

 Transverse dynamics

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 4

 Single particle dynamics in magnetic fields

Reference system

y x : horizontal
s y : vertical
s : longitudinal along the trajectory
x

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 5

 Typical Magnetic Fields

• Normal: gap appears in the horizontal plane

• Skew: rotate around beam axis by p/2n

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 6

 Properties of Typical Magnets

Dipoles: used for guiding the particle trajectories

Bx = 0
By = Bo = constant

Quadrupoles: used to focus the particle trajectories

Bx = G y
By = -G x
G = constant

Sextupoles: used to correct chromatism and non linear terms

Bx = 2 S x y
By = S ( x2-y2)
S = constant

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 7

 Lattice in an accelerator
Lattice: Sequence of magnets interleaved with drifts (used for diagnostics, vacuum
pumping, Injection, extraction, etc)

The first step in calculating a lattice is to consider only the linear components of it
(quadrupoles and dipoles). Non linear effects and chromatic aberration corrections will
be evaluated later.
The trajectory of the reference particle (the particle with nominal energy and initial
position and divergence set to zero) along the optics is calculated.
All the other beam particles are represented in a frame moving along the reference
trajectory, and where the reference particle is always in the center.
Coordinate systems used to describe the motion are usually locally Cartesian or
cylindrical (typically the one that allows the easiest field representation)

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 8

 Magnetic rigidity
Remember:

v pc
 
c E Constant magnetic field: B
1
 
1  2
momentum p  mv r
total energy E  mc 2 q

kinetic energy K  E  mc 2

E2  mc    pc 
2 2 2 A charged particle (charge = q) will follow
a circle of radius r

Lorentz force: 𝐹𝐿 = 𝑞𝑣𝐵 𝑝

Centrifugal force: 𝐹𝑐𝑒𝑛𝑡𝑟 =
𝛾𝑚𝑜 𝑣 2
𝐵𝜌 =
𝐹𝐿 = 𝐹𝑐𝑒𝑛𝑡𝑟
𝜌
𝑞

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 9

 Equations of motion
Magnetic field representation (consider only normal terms)

𝑑𝐵𝑦 1 𝑑2 𝐵𝑦 2
𝐵𝑦 𝑥 = 𝐵0 + 𝑥+ 𝑥 +. .
𝑑𝑥 2! 𝑑𝑥 2
𝑑𝐵𝑥 1 𝑑 2 𝐵𝑥 2
𝐵𝑥 𝑦 = 𝑦+ 𝑦 +. .
𝑑𝑦 2! 𝑑𝑦 2

Concentrate in horizontal motion: only vertical fields

𝑑𝐵𝑦 𝑑𝑔
g𝑔 = , 𝑔′ = Gradient and its derivative
𝑑𝑥 𝑑𝑥

Let’s normalize to momentum

𝐵𝑦 (𝑥) 𝐵0 𝑔 𝑥 1 𝑒𝑔′
= + + +⋯
𝑝/𝑒 𝐵𝑜 𝜌 𝑝/𝑒 2! 𝑝/𝑒

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 10

 Equations of motion
Consider only linear terms:
𝐵(𝑥) 1 where 𝑘=
𝑔
=
1 𝑑𝐵𝑦
= + 𝑘𝑥 𝑝/𝑒 𝐵𝜌 𝑑𝑥
𝑝/𝑒 𝜌
In the ideal orbit, r = const dr/dt = 0
General trajectory: r = r+x, with x<<r
𝑑2 𝑚𝑣 2
𝐹 =𝑚 2 𝑥+𝜌 − = 𝑞𝐵𝑦 𝑣
𝑑𝑡 𝑥+𝜌
1 1 𝑥
Since x<<r Taylor expansion ≈ (1 − )
𝑥+𝜌 𝜌 𝜌

𝑑2
𝜌 =0
𝑑𝑡 2
𝑑 2 𝑥 𝑚𝑣 2 𝑥
𝑚 2− (1 − ) = 𝑞𝐵𝑦 𝑣
𝑑𝑡 𝜌 𝜌
𝑑2 𝑥 𝑚𝑣 2 𝑥 𝜕𝐵𝑦
𝑚 2− (1 − ) = 𝑞𝑣 𝐵0 + 𝑥
𝑑𝑡 𝜌 𝜌 𝜕𝑥

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 11

 Equations of motion
Pass from t to s as independent variable
𝑑𝑥 𝑑𝑥 𝑑𝑠 For vertical plane
= 1
𝑑𝑡 𝑑𝑠 𝑑𝑡 = 0, 𝑘 = −𝑘
𝜌2
𝑑2 𝑥 𝑑 𝑑𝑥 𝑑𝑠 𝑑 ′ 𝑑𝑠 𝑦 ′′ + 𝑘𝑦 = 0
= = 𝑥𝑣
𝑑𝑡 2 𝑑𝑡 𝑑𝑠 𝑑𝑡 𝑑𝑠 𝑑𝑡
𝑑𝑥 𝑑𝑣
= 𝑥 ′′ 𝑣 2 + 𝑣
𝑑𝑠 𝑑𝑠

Second term is zero 1

′′ 2
𝑣2 𝑥 𝑞𝑣𝐵0 𝑞𝑣𝑥𝑔 ′′
𝑥 +𝑥 2−𝑘 =0
𝑥 𝑣 − 1− = + 𝜌
𝜌 𝜌 𝑚 𝑚
Divide by v2 𝑦 ′′ + 𝑘𝑦 = 0
1 𝑥 𝐵0 𝑥𝑔
𝑥 ′′ − + 2 = +
𝜌 𝜌 𝑝 𝑞 𝑝 𝑞
1 𝑥 1
𝑥 ′′ − + 2 = − + 𝑘𝑥
𝜌 𝜌 𝜌
1
𝑥 ′′ + 𝑥 2 − 𝑘 = 0
𝜌

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 12

 Solution of equations of motion
Putting
1
Horizontal : 𝐾 = −𝑘
𝜌2
Vertical :𝐾 =𝑘

If K is constant we get the differential equation of harmonic oscillator with spring constant K
𝑥 ′′ + 𝐾𝑥 = 0
𝑥 𝑠 = 𝑎1 cos 𝜔𝑠 + 𝑎2 sin 𝜔𝑠
𝜔= 𝐾 (where now x can
represent both x or y)
At 𝑠 = 0 ⇒ 𝑥 = 𝑥0 , 𝑥′ = 𝑥′0

𝑥′0
𝑎1 = 𝑥0 𝑎2 = ,
𝐾

1
𝑥 𝑠 = 𝑥0 cos 𝐾𝑠 + 𝑥′0 sin 𝐾𝑠
𝐾
𝑥 ′ 𝑠 = −𝑥0 𝐾 sin 𝐾𝑠 + 𝑥′0 cos 𝐾𝑠

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 13

 Matrix formalism
We can write the equations in matrix formalism: coordinates at point s1 can be
obtained knowing the coordinates at s0

𝑥 𝑥
=𝑀
𝑥′ 𝑠1 𝑥′ 𝑠0

1
cos 𝐾𝐿 sin 𝐾𝐿
𝑀= 𝐾
− 𝐾 sin 𝐾𝐿 cos 𝐾𝐿

Example: Drift
Length: L 1 𝐿
𝑀𝐷𝑟𝑖𝑓𝑡 =
K=0 0 1

𝑥1 = 𝑥0 + 𝐿𝑥′0
𝑥′1 = 𝑥′0

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 14

 Matrix formalism
Focusing quadrupole:
Length L , K > 0

1
cos 𝐾𝐿 sin 𝐾𝐿
𝑀= 𝐾
− 𝐾 sin 𝐾𝐿 cos 𝐾𝐿

𝑥 ′0
𝑥1 = 𝑥0 𝑐𝑜𝑠 𝐾𝐿 + 𝑠𝑖𝑛 𝐾𝐿
𝐾
𝑥′1 = −𝑥′0 𝐾𝑠𝑖𝑛 𝐾𝐿 + 𝑥 ′ 0 𝑐𝑜𝑠 𝐾𝐿

Defocusing quadrupole:
Length L , K < 0

1
cosh 𝐾𝐿 sinh 𝐾𝐿
𝑀= 𝐾
𝐾 sinh 𝐾𝐿 cosh 𝐾𝐿

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 15

 Thin lens approximation
1
In parallel with optics cos 𝐾𝐿 sin 𝐾𝐿
𝑀= 𝐾
− 𝐾 sin 𝐾𝐿 cos 𝐾𝐿
L=> 0 KL constant
sin 𝐾𝐿
𝐾𝐿
-> 1 1 0
1 0 1
M= =
1 −𝐾𝐿 1 ± 1
𝑓= ≫𝐿 𝑓
𝐾𝐿

f positive or negative
depending on quad

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 16

 Dipole
Sector magnet:
Nominal particle trajectory is
perpendicular to dipole entrance
Horizontal plane: 𝐾 = 1 𝜌2 − 𝑘
Vertical plane: 𝐾 = 𝑘

𝑐𝑜𝑠𝜃 𝜌𝑠𝑖𝑛𝜗
If 𝑘 = 0, 𝐿 = 𝜌𝜃
𝑀𝐻 = 1
− 𝑠𝑖𝑛𝜃 𝑐𝑜𝑠𝜃
𝜌
1 𝜌𝜗
𝑀𝑉 =
0 1
Magnet with field index: 𝑘 ≠ 0

Exercise – write the matrix

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 17

 System of lattice elements: Drifts (MD), quads (MQ), bendings (or dipoles) (MB)
𝑥0 𝑥1
Starting with 𝑥′ The final position and divergence of the particle will be 𝑥′
0 1

𝑥1 𝑥0
𝑥 ′1 = 𝑀𝐷𝑛 ∙ 𝑀𝑄𝑛 ∙ 𝑀𝐷𝑛−1 ⋯ ∙ 𝑀𝐵1 ∙ 𝑀𝐷2 ∙ 𝑀𝑄1 ∙ 𝑀𝐷1 ∙ 𝑥 ′0

Or simpler
𝑥1 𝑥0
𝑥 ′1 = 𝑀(𝑠 ,
1 0𝑠 ) ∙ 𝑥 ′0

The mathematical representation of an accelerator lattice is a sequence of matrices

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 18

 Twiss parameters – Betatron tune
𝑥 ′′ + 𝐾𝑥 = 0 If K = constant /harmonic oscillator
𝑥 ′′ + 𝐾(𝑠)𝑥 = 0 If K varies with s: Hill’s equation
The solution of the Hill equation is given by:

𝑥 𝑠 = 𝜀 𝛽 𝑠 𝑐𝑜𝑠 𝜑 𝑠 + 𝜑0
e and φ0 integration constants
Inserting 1 in the equation of motion it can be shown that the phase advance is related to 
by
𝑠
𝑑𝑠
𝜑 𝑠 =
0 𝛽(𝑠)

In storage rings (length of circumference = L) beta is periodic

𝛽 𝑠 + 𝐿 = 𝛽(𝑠)

One complete turn: phase advance in one turn: Betatron Tune

𝟏 𝒅𝒔
𝑸𝒙,𝒚 =
𝟐𝝅 𝜷𝒙,𝒚 (𝒔)

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 19

 Solution of Hill’s equation
1 𝑑𝛽
𝛼=−
2 𝑑𝑠 𝑥= 𝜀𝛽 cos 𝜙
1 + 𝛼2 With ϕ = 𝜑 𝑠 + 𝜑0 and 𝛽 depending on s
𝛾=
𝛽
with

𝑑 𝛽 𝑑𝜙
𝑥′ = 𝜀 cos 𝜙 − 𝜀𝛽 sin 𝜙
𝑑𝑠 𝛼𝛼 𝑑𝑠

′
𝜀 𝑑𝜙
𝑥 = −𝛼 cos 𝜙 − 𝜀𝛽 sin 𝜙
𝛽 𝑑𝑠

𝑑2 𝛽 𝑑 𝛽 𝑑𝜙 𝑑 𝛽 𝑑𝜙 𝑑2𝜙 𝑑𝜙 2
𝑥’’= 𝜀 cos 𝜙 − 𝜀 sin 𝜙 − 𝜀 sin 𝜙 − 𝜀𝛽 2 sin 𝜙 − 𝜀𝛽 cos 𝜙
𝑑𝑠 2 𝑑𝑠 𝑑𝑠 𝑑𝑠 𝑑𝑠 𝑑𝑠 𝑑𝑠

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 20

 Substituting in the 2nd order equation, and equating to zero terms multipliying sin
and cos we get:
𝑑 𝛽 𝑑𝜙 𝑑2𝜙
−2 𝜀 − 𝜀𝛽 2 = 0
𝑑𝑠 𝑑𝑠 𝑑𝑠
Dividing by 𝜀 and differentiating 𝛽:

𝑑𝛽 𝑑𝜙 𝑑2𝜙
+𝛽 2 =0
𝑑𝑠 𝑑𝑠 𝑑𝑠

𝛽𝜙 ′ ′
=0
𝛽𝜙 ′ = 𝑐𝑡𝑒 = 1

′ 1 𝑠 𝑑𝑠
𝜙 = 𝜑 𝑠 = 0 𝛽(𝑠)
𝛽

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 21

• Beta function b x (s)Betatron
: oscillations
– Describes the envelope of the betatron oscillation in an accelerator

• Phase advance:
y (s) = ò 0
1 s
ds
b x (s)
• Betatron tune: number of betatron oscillations in one orbital turn
Nominal closed yorbit
(0 | C) Particles
ds oscillate around
R the closed orbit, a number
x =
BetatronQoscillation
2p
= ò /2p =is given by the betatron tune.
of times which
b xsquare
(s) of theá b xfunction
ñ
The by the emittance
represents the envelope of the betatron oscillations

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 22

 Twiss parameters
Amplitude of an oscillation
𝑥 𝑠 = 𝜀 𝛽(𝑠)

𝛽 𝑠 represents the envelope of all particle trajectories at a given position s in a storage ring

𝑥 𝑠 = 𝜀 𝛽 𝑠 𝑐𝑜𝑠 𝜑 𝑠 + 𝜑0
′
𝜀
𝑥 𝑠 =− 𝛼 𝑠 cos 𝜑 𝑠 + 𝜑0 + 𝑠𝑖𝑛 𝜑 𝑠 + 𝜑0
𝛽(𝑠)

1
𝛼 𝑠 = − 𝛽′(𝑠)
2
1 + 𝛼 2 (𝑠)
𝛾 𝑠 =
𝛽(𝑠)

a,  and  are the Twiss parameters

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 23

 Emittance
Inserting in x’ eq.
2
𝜀 = 𝛾 𝑠 𝑥 2 𝑠 + 2𝛼 𝑠 𝑥 𝑠 𝑥 ′ 𝑠 + 𝛽 𝑠 𝑥 ′ (𝑠)

e is a constant of motion, not depending on s.

Parametric representation of an ellipse in x,x’ phase space defined by alfa,
beta, gamma: Courant-Snyder invariant emittance ε
For a single particle, different positions in the storage ring and different
turns:

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 24

 Drawing particle position in different points of a
storage ring
alfa = 0;
beta = 1;
gamma = (1+alfa*alfa)/beta;
eps = 10; Same area, same emittance,
for i =1 : npoints
thdeg (i) = i*13; Different orientation
teta(i) = thdeg(i)*rad;
x(i) = sqrt(eps*beta)*cos(teta(i));
xp(i) = -sqrt(eps/beta)*( alfa*cos(teta(i)) + sin(teta(i)) );
hold on
end
grid on
plot(x,xp,'-r*')

10
xp beta = 1 10
8 xp beta = 4
8
6
6

4
4
alfa = -2 alfa = -2
2 2

0 alfa = 0 0 alfa = 0
-2 -2
alfa = 2 alfa = 2
-4 -4

-6 -6

-8
-8
x
x
-10
-10 -10 -8 -6 -4 -2 0 2 4 6 8 10
-10 -8 -6 -4 -2 0 2 4 6 8 10

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 25

 Liouville’s theorem
10
xp
8

-2

-4

-6

-8
x
-10
-10 -8 -6 -4 -2 0 2 4 6 8 10

Each particle, in absence of non conservative forces, has a constant invariant.

Under the influence of conservative forces the particle density in phase space is constant.
Magnetic fields of dipoles and quadrupoles are conservative:
In a beam the phase space is maintained constant
Beam size 𝑥𝑚𝑎𝑥 = 𝛽(𝑠)𝜀 𝑥′𝑚𝑎𝑥 = 𝛾(𝑠)𝜀

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 26

 Evolution of particle trajectory related to evolution
of Twiss Parameters
𝑥 𝑠 = 𝜀 𝛽 𝑠 𝑐𝑜𝑠 𝜑 𝑠 + 𝜑0

𝜀
𝑥′ 𝑠 = − 𝛼 𝑠 cos 𝜑 𝑠 + 𝜑0 + 𝑠𝑖𝑛 𝜑 𝑠 + 𝜑0
𝛽(𝑠)
Developing sin(a+b)

𝑥 𝑠 = 𝜀 𝛽 𝑠 (𝑐𝑜𝑠𝜑 𝑠 𝑐𝑜𝑠𝜑0 − 𝑠𝑖𝑛𝜑 𝑠 𝑠𝑖𝑛𝜑0 )

𝜀
𝑥′ 𝑠 = − 𝛼 𝑠 cos 𝜑 𝑠 𝑐𝑜𝑠𝜑0 − 𝛼 𝑠 sin 𝜑 𝑠 𝑠𝑖𝑛𝜑0 + sin 𝜑 𝑠 𝑐𝑜𝑠𝜑0 + cos 𝜑 𝑠 𝑠𝑖𝑛𝜑0
𝛽(𝑠)

𝑥0 1 𝛼0 𝑥0
Starting at 𝑠 0 = 0 𝜑 0 = 0 𝑐𝑜𝑠𝜑0 = , 𝑠𝑖𝑛𝜑0 = − 𝑥′0 𝛽0 +
𝜀𝛽0 𝜀 𝛽0

Substituting above
𝛽 𝑠
𝑥 𝑠 = cos 𝜑 𝑠 + 𝛼0 𝑠𝑖𝑛𝜑(𝑠) 𝑥0 + 𝛽(𝑠)𝛽0 𝑠𝑖𝑛𝜑(𝑠) 𝑥′0
𝛽0

′
1
𝑥 (𝑠) = (𝛼0 − 𝛼 𝑠 ) cos 𝜑 𝑠 − (1 + 𝛼0 𝛼 𝑠 )𝑠𝑖𝑛𝜑(𝑠) 𝑥0
𝛽(𝑠)𝛽0
𝛽 𝑠
+ cosφ s − 𝛼(𝑠) 𝑠𝑖𝑛𝜑(𝑠) 𝑥 ′ 0
𝛽0

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 27

 Particle coordinates evolution as a
function of Twiss parameters

These equations can be written under matrix formalism

𝛽 𝑠 → 𝛽𝑠 𝛼 𝑠 → 𝛼𝑠 𝜑 𝑠 → 𝜑𝑠

𝛽𝑠
cos 𝜑𝑠 + 𝛼0 𝑠𝑖𝑛𝜑𝑠 𝛽𝑠 𝛽0 𝑠𝑖𝑛𝜑𝑠
𝑥 𝑠 𝛽0 𝑥0
= ∙ 𝑥′
𝑥′ 𝑠 0
1 𝛽𝑠
𝛼0 − 𝛼𝑠 cos 𝜑𝑠 − (1 + 𝛼0 𝛼𝑠 )𝑠𝑖𝑛𝜑𝑠 cos𝜑𝑠 − 𝛼𝑠 𝑠𝑖𝑛𝜑𝑠 𝑥 ′ 0
𝛽𝑠 𝛽0 𝛽0

Knowing Twiss parameters in two points and phase advance in between it is possible to
define the particle position in the second point from the first one with no need of knowing
the elements in between

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 28

 One turn matrix
For one revolution:
𝛽 𝑠 = 𝛽0 𝛼 𝑠 → 𝛼0 𝜑 𝑠 → 2𝜋𝑄𝑥,𝑦

One turn matrix

cos2πQ + αo sin2πQ βo sin2πQ

MT 𝑠 = = cos2πQ ∙ 1 + sin2πQ ∙ J
−γo sin2πQ cos2πQ − αo sin2πQ

1 0 𝛼 𝛽
With 𝟏= 𝑎𝑛𝑑 𝐽 =
0 1 −𝛾 𝛼

It can be easily demonstrated that

𝑀𝑇 = 1

Knowing the one turn matrix at a certain location and the total phase advance,  and a are
defined at that location
Exercise: write the expression of  and a as a function of one-turn matrix terms

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 29

 N-turns Matrix

𝑀𝑁 = (1 ∙ 𝑐𝑜𝑠2𝜋𝑄 + 𝐽 ∙ 𝑠𝑖𝑛2𝜋𝑄)𝑁 = 1 ∙ 𝑐𝑜𝑠2𝜋𝑁𝑄 + 𝐽 ∙ 𝑠𝑖𝑛2𝜋𝑁𝑄

Motion bounded (stable) if the elements of MN are bounded:

𝑐𝑜𝑠2𝜋𝑄 ≤ 1
Or

𝑇𝑟(𝑀) ≤ 2

Stability condition
Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 30
 Transformation of Twiss Parameters
Start from two positions, so and s

𝑥 𝑥0
=𝑀∙ 𝑥′0
𝑥′
𝐶 𝑆
𝑀= ′
𝐶 𝑆′
Or
𝜀 = 𝛽𝑠 𝑥′2 + 2𝛼𝑠 𝑥𝑥 ′ + 𝛾𝑠 𝑥 2
𝜀 = 𝛽0 𝑥′20 + 2𝛼0 𝑥0 𝑥′0 + 𝛾0 𝑥02

𝑥0 −1 𝑥
′
𝑥0 = 𝑀 ∙
𝑥′

𝑥0 = 𝑆 ′ 𝑥 − 𝑆𝑥′
𝑥′0 = −𝐶 ′ 𝑥 + 𝐶𝑥′

Inserting in e and rearranging the terms on x and x’

𝛽 𝑠 = 𝐶 2 𝛽0 − 2𝑆𝐶𝛼0 + 𝑆 2 𝛾0
𝛼 𝑠 = −𝐶𝐶 ′ 𝛽0 + 𝑆𝐶 ′ + 𝑆 ′ 𝐶 𝛼0 − 𝑆𝑆′𝛾0
𝛾 𝑠 = 𝐶′2 𝛽0 − 2𝑆 ′ 𝐶 ′ 𝛼0 + 𝑆′2 𝛾0

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 31

 Transformation of Twiss Parameters

And can be written in matrix notation: given the Twiss parameters α, β, γ at any point
in the lattice we can transform them and calculate their values at any other point in
the ring if we know the transfer matrix.

𝛽 𝐶2 −2𝐶𝑆 𝑆2 𝛽𝑜
𝛼 = −𝐶𝐶′ 𝑆𝐶 ′ + 𝑆 ′ 𝐶 −𝑆𝑆′ 𝛼𝑜
𝛾 𝐶′2 −2𝑆 ′ 𝐶′ 𝑆′2 𝛾𝑜

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 32

 rms Emittance
The emittance is the area of the phase space occupied by all particles in a beam
Each particle has its own ‘invariant emittance’
rms emittance represents the beam characteristics, and is defined as:

𝜀𝑥 𝑟𝑚𝑠 = 𝑥 2 𝑥 ′2 − 𝑥 𝑥′ 2

Rms values behave the same for all distributions in linear systems
Most usual beam distributions are gaussian

X’

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 33

 Emittance and beam dimensions
• The emittance is the area of the phase space occupied by the particles
With the emittance and the Twiss parameters in a point of the
accelerator, the beam dimensions are obtained : sx,y e s’x,y
px
ex  x 2 x'2  xx'
2

s ' x = g xe x

x Ellipse area= pex

s x   xe x
x 2   xe x
x '2   xe x
x x'  a xe x

 x x  a x2  1
34
Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 34
 Effect of a quad on the phase space
The phase space orientation indicates if the beam trajectories are focused
or defocused (at rms values):

s0 s1 s2

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 35

 Example of betatron functions in a
storage ring - ALBA

Focusing quads above the line

Defocusing quads below,
Dipoles on both sides

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 36

 Emittance in linacs and transfer lines
Adiabatic damping:
The Courant-Snyder invariant emittance ε decreases if we accelerate the particle. This is
called “adiabatic damping”.
Particle with momentum p0:
𝑝02 = 𝑝𝑠0
2 2
+ 𝑝𝑥0 2
+ 𝑝𝑦0
The slope of the trajectory is
𝑝𝑧
𝑧 ′′ = (𝑧 = 𝑥 or 𝑦)
𝑝𝑠

Accelerating the particle: ps increases but pz does not change

′ ′
𝑝𝑧 𝑝𝑧 ∆𝑝𝑠
𝑧 + ∆𝑧 = = ≈ 𝑧′ 1 −
𝑝𝑠 + ∆𝑝𝑠 𝑝 1 + ∆𝑝𝑠 𝑝𝑠
𝑠 𝑝𝑠
And therefore
∆𝑝
∆𝑧 ′ = −𝑧′
𝑝

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 37

 In the phase space (z,z’):
𝜀
For each particle i: 𝑧𝑖 = 𝜀𝛽 cos 𝜑 + 𝜑0 𝑧′𝑖 = − sin 𝜑 + 𝜑0
𝛽

The emittance of the particle is:

𝜀 = 𝛽𝑧′2𝑖 + 𝛾𝑧𝑖2
The change in e applying Dz’
2 ∆𝑝 ∆𝑝
∆𝜀 = 2𝛽𝑧𝑖′ ∆𝑧𝑖′ = −2𝛽𝑧𝑖′ = −2𝜀 𝑠𝑖𝑛2 𝜑 + 𝜑0
𝑝 𝑝
Averaging over all particles:
∆𝑝 𝑑𝜀 𝑑𝑝
∆𝜀 = −𝜀 => =−
𝑝 𝜀 𝑝
𝑝0
𝜀 𝑝 = 𝜀0
𝑝
The emittance decreases in linacs as the beam is accelerated.
Normalized emittance is defined as the invariant part:
𝜀𝑛 = 𝜀𝛽𝛾
𝑣 1
Where now  and  are not the Twiss parameters, but: 𝛽 = 𝑎𝑛𝑑 𝛾 =
𝑐 1−𝛽2
In absence of other phenomena, the normalized emittance does not change during the
acceleration

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 38

 Off-Momentum Particles
Magnets are chromatic elements:
𝑝 ∆𝜌 ∆𝑝 ∆𝜃
Dipoles: 𝜌 = → = =
𝑒𝐵 𝜌0 𝑝0 𝜃0
𝐺 𝐺𝑒
Quads: 𝐾 = =
𝐵𝜌 𝑝
∆𝑝
∆𝜃 = −𝜃0
𝑝0
∆𝑝
∆𝐾 = −𝐾0
𝑝0
Dp/p<<1
Off-momentum particles are not oscillating around design orbit, but around a different closed
orbit (chromatic closed orbit). The displacement between the design and chromatic orbits is
regulated by the dispersion function D(s).
For particles with energy deviation the Hill’s equation has an extra term and is not
homogeneous:
′′
1 ∆𝑝
𝑥 +𝐾 𝑠 𝑥 =
𝜌 𝑝0
The solution is the sum of the solution of the homogenous equation + a term of dispersion:
∆𝑝
𝑥 = 𝑥𝐻𝑜𝑚 + 𝐷 𝑠
𝑝0

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 39

 Dispersion equation
The dispersion equation is:
′′
1
𝐷 (𝑠) + 𝐾 𝑠 𝐷(𝑠) =
𝜌

The dispersion in a storage ring is a closed orbit.

Example: Dipole
The solution of the dispersion equation is given by the homogeneous equation solution
plus a particular one of the non homogenous:

𝑠 𝑠 𝑠
𝐷 𝑠 = 𝐷0 cos + 𝐷′0 𝜌 sin + 𝜌 1 − cos
𝜌 𝜌 𝜌

𝐷0 𝑠 𝑠 𝑠
𝐷′ 𝑠 = − sin + 𝐷 ′ 0 cos + sin
𝜌 𝜌 𝜌 𝜌

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 40

 Or in matrix formalism, representing the particle coordinates including the
energy deviation:

𝑥 𝑥0
𝑥′ = 𝑀3𝑥3 𝑥′0
∆𝑝 𝑝 ∆𝑝 𝑝

𝐶(𝑠) 𝑆(𝑠) 𝐷(𝑠)

𝑀3𝑥3 = 𝐶′(𝑠) 𝑆′(𝑠) 𝐷′(𝑠)
0 0 1

𝑐𝑜𝑠𝜃 𝜌𝑠𝑖𝑛𝜃 𝜌(1 − 𝑐𝑜𝑠𝜃)

1
𝑀𝑠𝑒𝑐𝑡𝑜𝑟 = − 𝑠𝑖𝑛𝜃 𝑐𝑜𝑠𝜃 𝑠𝑖𝑛𝜃
𝜌
0 0 1

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 41

 Also quadrupole focusing properties depend on particle energy:
The focusing strength is

1 𝑑𝐵
𝐾=
𝐵𝜌 𝑑𝑥

Their implication on the dispersion function is of second order

e e e  Dp  Dp
k Dp   G   G   1  G  k0  Dk  error :Dk  k0
p p0  Dp p0  p0  p0

But they produce chromaticity

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 42

 Particles with different momentum have different betatron tunes
(The higher the energy the lower the tunes)
Chromatism is the spread in tune divided by the spread in momentum

DQx,y
å
1
x x,y = = - ki Li b x,i
Dp / p 4p i

And is given by the sum of the contributions of all elements in the ring

Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 43

 Chromaticity correction
• The natural chromaticity produced by quadrupoles is always
negative. x has maxima at focusing quads (QF, kx >0) and
minima at defocusing ones (QD). Horizontal chromaticity is
dominated by the focusing quads, and x at their positions.
• Chromaticity must be corrected to avoid large tune spread
(driving resonances or collective effects like head-tail
instability)

Sextupoles placed where D is not

zero, act on particles which are not
on the nominal orbit, but on
x = x0 + D Dp/p
The highest the energy deviation, the
largest x, the strongest the
focusing(defocusing) force

44
Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 44
 Sextupoles
Bx = m xy
• The field is quadratic :
By = m ( x2 - y2 )
1
2
• Normalized gradient: 2
1 By
d -3
m= [m ]
Br dx 2

• Kick: Dx' = -mx2, Dy' = 2mxy

• Change in tune due to sextupole:

1 Dp
DQx = b x,s0 mLsx con x = D
4p p

DQx 1
= b x,s0 mLsD
Dp / p 4p
45
Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 45
 Chromaticity correction
• Sextupole position are chosen so that the produced
chromaticity counteracts the natural one created by
quadrupoles
• Mainly where the dispersion is high and x e y are well
separated.
• Total chromaticity is then:

DQx
= - å b x,i ki Lq + å
1 1
xx = b x,i mLsD
Dp / p 4p i 4p i

Quadrupoles Sextupoles

46
Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 46
Accelerator Physics - UAB 2014-15 Transverse Dynamics C. Biscari - Lectures 4-5-6 47