![Page 1: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/1.jpg)
Chaos and Emittance growth due to nonlinear interactions in circular
accelerators
K. Ohmi (KEK)SAD2006
Sep. 5-7. 2006 at KEK
![Page 2: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/2.jpg)
Emittance growth• External incoherent diffusion, radiation, I
ntrabeam etc.• Coherent motion, instability• Nonlinear diffusion• Nonlinearlity coupled to external diffusion
(noise)
1. Incoherent electron cloud2. Beam-beam limit3. Space charge limit
![Page 3: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/3.jpg)
Diffusive or not• Which system does not have emittance gro
wth?1. Integrable system2. System with two degrees of freedom
• Which system can have emittance growth?1. Nonintegrable system with three or more degrees of free
dom2. External diffusion: noise, radiation excitation… The exter
nal diffusion is amplified due to nonlinear interaction
![Page 4: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/4.jpg)
Nondiffusive system• Integrable system – needless to say• System with two degrees of freedom. Particles d
o not across torus layers.
System fall into global stochastic regime may be diffusive even for two degrees of freedom, but the diffusion is limited in the regime.
![Page 5: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/5.jpg)
Diffusion in three or more degrees of freedom
• Motion in a degree of freedom gives modulation
• Particles can get over KAM boundary through additional freedom.
![Page 6: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/6.jpg)
Round beam: Example of non-diffusive system
• Equal tune, no synchrotron oscillation -- equivalent to a system with 2 degrees of freedom (r-s).
2 2
2 21 ( ) ( , )2 2
x yP
p pH x y s U r z
2 22
2 ( ) ( , )2 2r P
p rp s U r zr
sincos' yxr pprp cossin'2
yx pprrp
(x,y)
ctsz
• H does not include , therefore p is a constant of motion.
• Trajectories on a torus r-pr-s• Poincare cross-section is mapped on two dimensional space.
![Page 7: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/7.jpg)
• Motion in r-pr space. max
min
2 2
2 2
1 1 22 2 2
r
r r r
E r EJ p dr drr
2
1
2 2
(2 )cos(2 )
rr
r
r J
J
12
J p d p
![Page 8: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/8.jpg)
Model- round beam interacting with electron cloud
• Transverse beam size depends on z; x(z), y(z).• Assume transverse Gaussian charge distribution. Applica
ble not only beam-beam but also electron cloud and space charge issues.
• Strong localized force x, y ~x, y2 2
2 2
2 20
exp2 2
( , ; ( ), ( ))2 2
x yex y
x y
x yu uNr
U x y z z duu u
( 0) exp : : ( 0)s U s x x
tyx zpypx ),,,,,( x
![Page 9: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/9.jpg)
The force for a round charge distribution• The force depends on z, because of r(z) of strong bea
m or electron cloud.
duuu
rrzrU
r
rer
0 2
2
2
22
exp))(;(
dzdrr
dzdzrU
zzrUF r
rr
er
r
rrz
2
2
2
2
2
2 2exp
212))(;())(;(
2
2
2exp112))(;(
r
err
rr
rrzrUF
• Example, LHC• L=26700 m, x=0.28, y=0.31, s=0.006
• x=y=8x10-9 m, x= x=100 m
• x=y=0.89 mm, z=0.13 m,
![Page 10: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/10.jpg)
Model of cloud• r (cloud) and tune shift• Ne L=1, 2, 4, 6 x1011. Interact at a point in the ring.
![Page 11: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/11.jpg)
Two degree of freedom
• 4 variable (x, px, s, H’). 1 integral for H’, Poincare cross section (certain s) 2 variable.
• Poincare plot, x-px plot at a certain s.
• When one more integral, J(x,px)=constant, the system is solvable. This relation gives a curve in x-px phase space.
![Page 12: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/12.jpg)
Poincare plot• r-pr =0.06, 0.13, 0.27, 0.38
Note r=0.89 mm
![Page 13: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/13.jpg)
Diffusion rate• Diffusion rates are very small compare than those
for 3 degrees of freedom, see later.• T0=89 s.
![Page 14: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/14.jpg)
Resonance overlap• Fourier expansion for r of U.
• Resonance position, Jr,R .
• Motion near the resonance position, pendulum motion, separatrix.
0
cosk rk
U U k
0( )r rr
UJJ
,( ) 2r RnJk
,r r Rr
J JI m
k
,
22 20
2
1 ( ) cos2
r R
r P kr J
UH k I s U
J
![Page 15: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/15.jpg)
Overlap condition, Chirikov criterion
• Resonance width
• Resonance separation
• Overlap condition, width>separation
120
24r kr
UJ UJ
120
, , 2
2( , 1) ( , )r R r Rr
UJ k n J k n
k J
20
2 2kr
Uk UJ
![Page 16: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/16.jpg)
Synchrotron oscillation and symplectic diffusion
• Add synchrotron motion, 3 degrees of freedom • The structures of tori are different for each z. • Are particles return the same torus after one synchrotron period.
s
z
???Are there different worlds for every z?
rpr
prr
22 22 2 2
2 ( ) ( , )2 2 2r z P
p rH p p z s U r zr
![Page 17: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/17.jpg)
Poincare plot for several z
Note r=0.89 mm
![Page 18: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/18.jpg)
Poincare plot for several z• Z=2, 4, 6, 10 cm
![Page 19: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/19.jpg)
Adiabatic invariant?
![Page 20: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/20.jpg)
Solvable or not• 6 variables (x,p,y,p,s,H’)• One integra H’, Poincare cross section at s.• 4 variables• When 2 integrals exist, the system is solvable. The soluti
on is represented by a surface in 3 dimensional space, r-pr-z.
• When a surface is not seen, the system is nonsolvable: i.e., emittance growth occurs.
Solvable
Nonsolvable
Blue: no synchrotron motion at z=0
![Page 21: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/21.jpg)
Separatrix crossing
0 0
( , , , ) cosr z z kl r zk l
U J J U k l
120
, , 2( 1) ( ) zr R r R
r
UJ l J lk J
• U depends on z. Fourier expansion for synchrotron phase.
• Resonance separation (narrower than n-n+1)
20
2
4 1klz r
Uk UJ
![Page 22: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/22.jpg)
Diffusion due to synchrotron motion
• z =0.006 =0.012
![Page 23: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/23.jpg)
Tune difference• 3 degrees of freedom
• p is not constant. Variation of p for small tune difference, x=0.285 & y=0.295.
2 22 2
( ) ( , )2 2
yxx y P
p yp xH s U r z
r
pr
p
2 2 22
2 cos 2 ( ) ( , )2 2r P
p r rp s U r zr
![Page 24: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/24.jpg)
• KAM for various p for equal tune,.
• p =1.2x10-9, 4.8x10-9, 2x10-8.
Note r=0.89 mm
![Page 25: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/25.jpg)
Motion in the phase space and tune space – example I
• Near integrable trajectory---nondiffusive
![Page 26: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/26.jpg)
Motion in the phase space and tune space - example II
• Chaotic trajectory --- diffusive
![Page 27: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/27.jpg)
Tune scan for 3 degrees of freedom
• Tune scan without synchrotron motion
0.05, 0.05
0.45, 0.45
![Page 28: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/28.jpg)
• Pumping machanism• Resonance• Separatrix crossing (~0)
![Page 29: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/29.jpg)
More
• 4 degrees of freedom – actual weak-strong model.
• Colliding beam – 4x2xN+- degrees of freedom• 3 degree of freedom+synchrotron motion. Modul
ation diffusion, stochastic pumping with separatrix crossing.
22 2 22 2 2
2 cos 2 ( ) ( , )2 2 2r z P
p r rH p p z s U r zr
![Page 30: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/30.jpg)
4 degrees of freedom• Tune scan with synchrotron motion, z =0.006.• Vertical emittance growth.• Resonance, m y=n is seen.• Emittance growth is large at cross points of resonances
0.05, 0.05
0.45, 0.45
![Page 31: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/31.jpg)
Importance of Lattice • Nonlinearity of beam-cloud interaction • Integrated the nonlinear terms with multiplying
function and cos (sin) of phase difference23 3 34 45 5 11 12 2 4: : : : : : : : : : : :: : : : : : : :e e e e e e e e e ...e nF U F F U FU F U UM
111 : :: :
1
e exp : ( ) : exp( : :)i
nFF
i Ti
U e M U
x
/ 2 / 21cos( )m m m
i ikx k J m
Nonlinear term should be evaluated with considering the beta function and phase of position where electron cloud exists.
Unphysical cancel of nonlinear term may be caused by simple increase of interaction point.
F: lattice transformation
![Page 32: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/32.jpg)
Beam-beam limit• 4 degree of freedom• Interaction during collision. If z~y, =1 rad,
4th-order term 4.2 2
2 2
2 20
exp2 2
( , ; ( ), ( ))2 2
x yex y
x y
x yu ur
U x y z z duu u
( ) ( , , , , )x x y y z z P T x yH J J J s U x p y p z
111 : :: :
1
e exp : ( ) : exp( : :)i
nFF
i Ti
U e M U
x
![Page 33: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/33.jpg)
Integrability near half integer tune Reduction of the degree of freedom.
• For x~0.5, x-motion is integrable. (work with E. Perevedentsev)
if zero-crossing angle and no error.
• Dynamic beta, and emittance
• Choice of optimum x
,0.5lim 0x
C yD
2 2,00.5
limx
xx
2
0.5limx
xp
1(crossing angle)+(coupling)+(fast noise)x
L
![Page 34: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/34.jpg)
X-px plot near half integer in x
• x=0.503, 0.510, 0.520, 0.540
![Page 35: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/35.jpg)
Crossing angle
• Calculate UT using Taylor map (Diff. algebra)
• Taylor coefficient ~ Fourier coefficient, Uklm.
0
cos( )T klm x y zklm
U U k l m
![Page 36: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/36.jpg)
4-th order Coefficients due to crossing angle• Short bunch z=3mm, x/z=1 at 2x15 mrad, original super KEKB.
![Page 37: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/37.jpg)
4-th order Coefficients as a function of crab sextupole strength, short bunch z=3mm, x/z=1
• H=K x py2/2, theoretical o
ptimum, K=1/xangle.• Clear structure- 220,121• Flat for sextupole streng
th- 400, 301, 040
![Page 38: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/38.jpg)
Crab crossing and crab waist
• Crossing angle induces synchro-beta and odd coupling resonance terms.
• Merit of Crab crossing is the absence of the terms.
• Crab waist reduces the odd coupling resonance term, but keeps the synchro-beta term.
• In the both method, Luminosity performance is improved.
![Page 39: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/39.jpg)
Space charge limit
• Similar as electron cloud, integrate the nonlinear interaction along the ring.
• UT: Gaussian?
111 : :: :
1
e exp : ( ) : exp( : :)i
nFF
i Ti
U e M U
x
( ) ( , , , , )x x y y z z P T x yH J J J s U x p y p z
![Page 40: Chaos and Emittance growth due to nonlinear interactions in circular accelerators](https://reader036.vdocuments.site/reader036/viewer/2022062323/56815e00550346895dcc43a1/html5/thumbnails/40.jpg)
Summary
Keyword of SAD
Dynamic aperture
Emittance