RF Linac Structures

Nicolas Pichoff
CEA/DIF/DPTA/SP2A nicolas.pichoff@cea.fr

Outline q Basis o Advantage/disadvantage of linacs o RF Cavities q Beam dynamics though linacs o Energy gain o Linear motion o RMS matching

Why RF linacs ? Goal of an accelerator : Accelerate a wanted beam within the lower cost
wanted : particle, energy, emittance, intensity, time structure cost : construction, operation

Main competitors : RF linacs, Synchrotrons, Cyclotrons... RF linacs : Particles accelerated on a linear path with RF cavities.
Advantages : High current, high duty-cycle, low synchrotron radiation losses. Drawbacks : High room & cavities consumption, no synchrotron radiation damping

Main use of linacs : Low energy injectors, high intensity protons beam, high energy lepton colliders.

Linacs main applications Electrons High energy collider : No synchrotron losses High-quality e- beam for FEL : Strong focusing Medical/Industrial irradiation : Low energy Neutron sources : Material study Protons Synchrotron injectors : High intensity, high duty-cycle Neutron sources : High Power. Material study, transmutation, nuclear fuel production, irradiation tools, exotic nucleus production Heavy ions Nuclear physics research : High intensity, high duty-cycle Implantation : Semi-conductors Driver for inertial-confinement fusion

RF resonant cavity Goal : Give kinetic energy to the beam Basic principle
- Conductor enclosing a close volume, - Maxwell equations + Boundary conditions allow possible electromagnetic field En/Bn configurations each oscillating with a given frequency fn : a resonnant mode. The field is a weighted superposition of these modes. - The wanted (accelerating) mode is excited at the good frequency and position from a RF power supply through a power coupler, - The phase of the electric field is adjusted to accelerate the beam. Cavity Wave guide Power coupler RF power supply

Elements of mode calculation Boundary conditions : close to the surface Mode calculation : Electric field :
Ex : Drift Tube Linac (DTL) tank

r r E // = 0 r r B^ = 0 2 r r 2 r wn r w n = 2p f n En + 2 En = 0 c : speed of light c r r r r E (r , t ) = en (t ) En (r )

TM010 : f=352.2 MHz

TM011 : f=548 MHz

TM020 : f=952 MHz

Field time variation

r r r d 2en w2 + w2 en = ? - n E H n n dS n dt 2 em S 1 r r 1 d r r r 1 d rr + H E n n dS J (r , t ) E n (r ) dV e dt S e dt V 3 2

1) Joule losses in conductor 2) Energy exchange with outside

=-

w RF & en Q0 n

S : conductor surface S : open surface

w & = - RF en + S n (t ) e j (w RF t + j0 ) losses Qexn feed

3) Energy exchange with beam : Beam loading V : enclosed volume

= k n I (t )

Field time variation (cont) The last equation can be modelled by :

d 2 en w RF den 2 + + w n en = S n e j (w RF t + j0 ) + k n I (t ) 2 Qn dt dt
Which is a damped harmonic oscillator in a forced regime
With :
1 1 1 = + Qn Q0 n Qexn

the quality factor of the cavity is the filling time of the cavity is the RF source is the beam loading

t = 2

Qn w RF

S n e j (w RF t + j0 )
k n I (t )

RF definitions and properties

Per cavity
Cavity length : L Cavity voltage V0 :

V0 = E z (z ) dz

Dissipated power Pd : Mean power dissipated in conductor over one RF period Shunt impedance R :

R=

V02 2 Pd

Pd =

1 R V02 2

Transit time factor T (calculated latter) :

DWmax = qV0 T

DWmax : Maximum energy that can be gained by a particle in the cavity

Effective shunt impedance : RT 2 =

2 DWmax 2 Pd

RF definitions and properties

Per unit length
Cavity mean electric field E0 :

E0 =

V0 1 = E z (z ) dz L L

Shunt impedance per unit length Z :

Z=

2 E0 R = 2 Pd L

Pd =

1 2 Z E0 2

Dissipated power per unit length Pd over one RF period Maximum energy that can be gained per unit length by a particle with charge q in the cavity: Effective shunt impedance per unit length : ZT 2 =

DWmax = qE0 T
2 DWmax 2 Pd

Example of use of effective shunt impedance ZT2

19 MeV 45 MeV 85 MeV

234 MeV

375 MeV

The effective shunt impedance of the structures has been chosen to set the transition energy between sections for TRISPAL project (C. Bourra, Thomson).

Designing a cavity consists in :

Fitting the accelerating mode frequency with the RF frequency, Maximise the shunt impedance of the cavity
Rejecting the unwanted modes frequencies far from the RF frequency, Calculating the tuning system, Increasing the Q0 of the accelerating mode, Calculating the energy deposition geometry to define the cooling system (~20W/cm2 max), the temperature increase and the associated frequency shift, Matching the coupler to the accelerating mode, Damping the High Order Modes (HOM) considered as dangerous (having a frequency close to a multiple of the RF frequency), mainly excited by the beam itself and responsible of power losses or beam dynamics perturbations, Increasing the beam aperture to reduce beam losses, Reducing the peak electric field to reduce electron field emission, Reducing the peak magnetic field to avoid quenches (th. Nb max at 2K : 200 mT), Adjusting the cavity geometry to reduce the multipactor probabilities, Calculating and minimising the cavity deformation and the associated frequency shift through the actions of electromagnetic forces pressure (Lorentz forces), Introducing RF peak-up necessary for the field phase and amplitude control, Transmitting all these data to the guy calculating the low level RF control, ...

Various types of cavity : Tank RFQ (low energy ~50 keV-7 MeV)

DTL (medium energy ~5-100 MeV)

Various types of cavity : Coupled cavity CCDTL (medium energy ~5-100 MeV)

CCL (high energy ~80 MeV-2 GeV)

Various types of cavity : Superconducting Spoke (medium energy ~20-100 MeV)

Elliptical (high energy ~100MeV- 2 GeV)

One word on travelling wave cavity

These cavities are essentially used for acceleration of ultra-relativistic particles The longitudinal field component is :
n

E z (r , z ,t ) = En (r ) e j (w t - k n z )
j (w t - k n z )

En (r ) e

is a space harmonic of the field, given by the cavity periodicity

Particle whose velocity is close to the phase velocity of the space harmonic exchanges energy with it. Otherwise, the mean effect is null.

v p vjn =

w kn

The transit time factor and the particle phase

Energy gained by a particle in a cavity of length L :

DW = qEz (s ) cos(f (s )) ds

with :

f(s ) = f 0 + w t = f 0 +

w ds c s bz (s ) 0

Assuming a constant velocity : b w DW = qEz (s ) cos f 0 + (s - s0 ) ds bc with : V0 =

p

DW = qV0 T b cos f p
Cavity Voltage Synchronous phase

( )

Ez (s ) ds Ez (s ) sin (f (s )) ds f = arctan Ez (s ) cos(f (s )) ds 1 T = Ez (s ) cos(f (s ) - f ) ds V

p

Transit-time factor : 0 < T < 1

20 150 10

1 V0

Ez (s ) e

jf ( s )

ds

P
- 10 - 150 - 20 - 250

i l
1

RF
050 -50

f
s

=-3 0

fD =3 06
0 .8

El

t i

if l d

t i l

0 .6 0 .4

F ield m tudei plA Df= Df=Df =Df =Df = 120 Df= 1 60

ti i
0 .2 0

L
-0 .2 1

0 .8

Example 1 : The transit time factor in a one-cell cavity Fast particle : T @ 1

Electric Field

Field amplitude Field in the cavity with time Field seen by a non synchronous particle

Energy gain

Energy gain

DW = qV0 T b

( )

N
0 -0 .2

li
0 .2

E
0 .4

G i
0 .6

Example 1 : The transit time factor in a one-cell cavity Medium fast particle : T @ 0.85
Electric Field

Field amplitude Field in the cavity with time Field seen by a non synchronous particle

Energy gain

Energy gain

DW = qV0 T b

( )

Example 1 : The transit time factor in a one-cell cavity Slow particle : T @ 0.3
Electric Field

Field amplitude Field in the cavity with time Field seen by a non synchronous particle

Energy gain

Energy gain

DW = qV0 T b

( )

The synchronous particle - Linac design

The linac is designed with a hypothetical synchronous particle. Its phase in a cavity is called here the synchronous phase :
The absolute phase ji and the velocity bi-1 of this particle being known at the entrance of cavity i, its RF phase fi is calculated to get the wanted synchronous phase fsi, f i = ji - f si D W i = qV 0T cos f si the new velocity bi of the particle can be calculated from, if the phase difference between cavities i and i+1 is given, the distance Di between them is adjusted to get the wanted synchronous phase fsi+1 in cavity i+1. if the distance Di between cavities i and i+1 is set, the RF phase fi of cavity i+1 is calculated to get the wanted synchronous phase fsi+1 in it.
RF phase Particle velocity Distances Synchronous phase Cavity number fsi-1 i-1 fi-1 bsi-1 Di-1 fsi i fi bsi Di fsi+1 i+1 fi+1

Synchronism condition :

f si +1 - f si = w

Di + fi +1 - fi + 2pn b si c

Linac with coupled cavities (DTL)

Gaps have the same phase. Distances between them are adjusted for synchronism.

Field in cavities Particle synchronous with the field Particle not synchronous with the field

Its energy gain Its energy gain

 Linac with independently phased cavities (SCL)

The distance between the cavities is given. Cavities are phased to accelerate a given particle.

Field in cavities Particle synchronous with the field Particle not synchronous with the field Its energy gain Its energy gain

Choice of the synchronous phase

Acceleration condition : The field should accelerate the particle
1 0.5

DW > 0 qV0T cos f p > 0

-270 -180 -90 0 90 180 270 360

qV0 > 0 : f p [- 90, 90] qV0 < 0 : f p [90, 270]

-0.5

-1 -360

Stability condition : Late particles should gain more energy that early ones

dDW > 0 qV0T sin f p < 0 df p

qV0 > 0 : f p [- 180, 0] qV0 < 0 : f p [0,180]

1 0.9 0.8 0.7 qV0 (normalised) 0.6 0.5 0.4 0.3 0.2 0.1 0 -90 -60 -30 0 f () 30 60 90

qV0 > 0 : f p [- 90, 0]

qV0 < 0 : f p [90,180]

General equations of motion

r r r r r dp = q vB + E = F dt

+ +

pw = b wg mc
dt = ds b z c

bwc is the particle velocity along w direction g is the particle reduced energy, q and m its charge and rest mass. x and y are transverse directions, s is the abscissa along longitudinal direction z, x and y are called the particle slopes.

Fx dgb z x dgb x ds = mc 2 b = ds z Fy dgb z y dgb y = = 2 ds mc b z ds dgb Fz z = mc 2 b z ds

x + dgb z ds Fx x = gb z mc 2gb z2

These equation are non linear, coupled and damped. Each element (cavity, quadrupole ) contributes to the force.

Linear force
In the highest simplification level, the external force along direction w (x, y or j) can be considered periodic, linear, uncoupled and undamped over one period : d 2w Hill equation : k w (s + S ) = k w (s ) + k w (s ) w = 0 ds 2 Giving :

w(s ) = e 0 b wm (s ) cos (m (s ) + m 0 )

with : m0 and e0 constant, m (s ) = m 0 +

s0

b (s ) ,
wm

ds

and : b wm (s + S ) = b wm (s )

In the (w, w) phase-space, the particle is moving on an ellipse of equation :

g wm (s ) w 2 + 2a wm (s ) ww + b wm (s ) w2 = e 0
with : a wm (s ) = -

1+ a wm (s ) 1 db wm and g wm (s ) = b wm (s ) 2 ds

Courant-Snyder parameters.
s = m (s + S ) - m ( s )

The phase advance of the particle in a lattice is then :

Particle motion in a FODO Lattice

Particle Particle ellipse Phase-space trajectory Phase-space periodic looks
Foc. Quad. middle Foc. Drift. middle

Defoc. Quad. middle

Defoc. Drift. middle

Particle trajectory

Particle ellipse maximum size

RMS dimensions and Beam Twiss parameters

The rms dimensions of the beam are defined statistically as followed : rms size : rms slope : rms emittance :
~ w= ~ w =

(w (w -

)2 )2
)2

~ ~ ~ e w = w 2 w2 - (w - w ) (w - w

The beam Twiss parameters are then :

~ w2 bw = ~ ew

aw = -

(w -

w ) (w - w ~ e
w

~ w2 gw = ~ ew

g w w 2 + 2a w ww + b w w2 = 5 e w

RMS matched beam The beam is matched when :

b w = b wm a w = a wm g w = g wm
Matched beam Bigger input beam Smaller input beam Phase-space scanned by the mismatched beams
Phase-space trajectory Phase-space periodic looks

50% mismatched beam

Summary
Linacs are competitive for low energy, high current, high duty cycle beams or very high energy light-particles (e+-e-) colliders. Acceleration is generally done with RF resonant cavities, confinement with quadrupoles (except at very low energy). Cavities are pieces of metal (Cu or Nb) whose shape is optimised to accelerate the particles at the RF frequency with the higher efficiency (ZT2 as high as possible) and the lower cost. The choice of the accelerating electric field E is a compromise between the linac length reduction ( E) and the power dissipation ( E) . RF phases in cavities are adjusted with respect to a synchronous particle to accelerate the beam and keep it bunched (synchronous phase choice). Forces are linearised to calculate the beam matching to the structure.