Head-Tail Modes for Strong Space Charge

1. Head-Tail Modes for Strong Space Charge A. Burov FNAL AP&T Seminar, Jan 6&8, 2009

2. Head-Tail Modes • This talk is about transverse coherent oscillations in a bunched beam, when the space charge tune shift dominates over the lattice tune spread, the synchrotron tune and the wake-driven tune shift (typical for low and intermediate energy hadron rings): • To make the bunch unstable, its head and tail have to talk to each other. The head acts on the tail by means of the wake fields. The tail acts back on the head when they exchange position after a half of the synchrotron oscillations. It also can act on the head through the over-the-revolution wake field, which is not completely decayed. In my talk I will assume the synchrotron oscillations as a main reason for the tail-to-head action. This requires the synchrotron tune do not be too small. coherent tune shift space charge tune shift lattice tune spread synchrotron tune

3. Historical Remarks • Head-tail modes = transverse coherent modes of a bunched beam. • Without space charge, theory of head-tail stability was essentially shaped by Ernest Courant & Andrew Sessler, Claudio Pellegrini & Matthew Sands, and FrankSacherer 40-30 years ago. • With strong space charge, the problem was treated by MikeBlaskiewicz 10 years ago. He solved it analytically for square potential well, air-bag (hollow beam) longitudinal distribution, and no octupoles. There is no Landau damping in his model. • In this talk, I will present my solution for arbitrary potential well, arbitrary 3D distribution function and possible lattice nonlinearity. Structure of the head-tail modes and their Landau damping for strong space charge will be described (submitted to Phys. Rev. ST-AB).

4. Contents • Conventional (no space charge) head-tail theory • Coasting beam with strong space charge • Strong space charge for bunches • General equation for space charge head-tail modes • Modes for Gaussian bunch • Intrinsic Landau damping • Landau damping by lattice non-linearity • Vanishing TMCI • Summary

5. Wake fields • Wake fields are fields left behind by particles in a vacuum chamber. They are excited due to wall resistivity and variations of the chamber cross-section along the orbit. • Wake kicks are normally so small, that only their revolution-averages are needed. They are described by wake functions: • Wake fields are not Newtonian (Hamiltonian): while the leading particle 1 acts on the following particle 2, the reverse action is normally much smaller. • Due to the 2nd Law of Thermodynamics, any conservative Hamiltonian system in thermal equilibrium is stable. Since wakes are not Hamiltonian, they can drive coherent instabilities.

6. Chromaticity and mode coupling • For bunched beams, any two interacting particles exchange their position every half of the synchrotron period. 1st and 2nd particles acts one upon another by the same wake, exchanging their roles. • If the wake is so small, that its influence can be averaged over the synchrotron oscillations, for zero chromaticity the system becomes Hamiltonian, making the beam is stable. • This effective restoration of Hamiltonian does not happen for non-zero chromaticity. That is why for small wake and small enough chromaticity, the coherent growth rate is proportional to the chromaticity. This is called as weak head-tail instability. • If the wake is so high, that it cannot be averaged over the synchrotron period, the instability may happen even at zero chromaticity. This is called as transverse mode coupling instability (TMCI) .

7. Multi-turn wake and coupled-bunch instabilities • When wake field lasts so long that it does not vanish after the revolution, the bunch acts on itself by its multi-turn wake, what makes one more reason for instability. • For many bunches, this instability may happen when the wake does not decay for bunch-to-bunch intervals. In this case, the bunch self-action occurs through other bunches as through a relay link.

8. Conventional head-tail modes at small wake • When the wake is small, the conventional head-tail modes are described as: • Wakes are considered small, if the wake-driven coherent tune shift is much smaller then the synchrotron tune: • These eigenfunctions depends on 2 arguments and are described by 2 numbers (p,m). • Without wake, the modes are degenerate on the action (“radial”) number p , their spectrum is generated by azimuthal integers m : head-tail phase coherent tune shift radial number azimuthal number

9. TMCI • When the wake is not small, , the eigen-modes are linear combinations of the small-wake modes. • The instability may happen as a result of mode coupling, it is called as transverse mode coupling instability (TMCI), or strong head-tail instability.

10. Example: Conventional TMCI simulations (B. Salvant et al, HB2008 ) Simulations for TMCI, no chromaticity, no space charge, for CERN SPS MOSES and HEADTAIL mode spectra vs current Growth Rate vs Current

11. Coasting beam with space charge • Space charge separates coherent and incoherent frequencies. Incoherent frequencies are shifted down by: • In 1974, D. Möhl and H. Schönauer suggested to use the rigid beam approximation: • The assumption for the space charge term calculation is that any beam slice moves as a whole, there is no inner motion in it. This is a good approximation, when the space charge is strong enough: lattice wake space charge

12. Coasting beam with space charge (2) • So, the space charge works as a factor of coherence. Separating coherent and incoherent motion, it makes the lattice tune spread less and less significant. With strong space charge, all the particles of the local slice move identically, , so the rigid beam approximation is reasonable. • In this case, only far-tail particles can have their individual frequencies equal to the coherent one: • These resonant particles make a 2D surface at 3D action space. They are responsible for the Landau damping, proportional to their effective phase space density. coherent tune

13. Coasting beam: Landau damping (Burov, Lebedev, 2008):

14. Coasting Beam Thresholds (Burov, Lebedev, 2008) • Thresholds are determined by . For a round Gaussian beam: Octupole threshold Chromatic threshold

15. Head-tail with space charge (A. Burov, submitted to PRST-AB) • Rigid beam approximation is applicable when Using slow betatron amplitudes , the initial single-particle equation of motion writes as Here and are time and distance along the bunch, both in angle units. space charge tune shift lattice tune spread synchrotron tune wake space charge chromaticity

16. Solution of single-particle equation • After a substitution with a new variable , the chromatic term disappears from the equation, going instead into the wake term: • Let for the beginning assume there is no wake, W=0. Then: or

17. No-Wake Equation • The phase Ψruns fast compared with relatively slow dependence , so under the integral can be expanded in the Taylor series: After that, the integral is easily taken: • Using that and averaging over all the particles inside the slice, equation for the space charge modes follows (no wake yet): eigenvalue

18. No-Wake Equation (2) • This equation is valid for any longitudinal and transverse distribution functions. The space charge tune shift stays here as transversely averaged: form-factors definition

19. General Equation for space charge modes • With two wake terms (driving W and detuning D), the equation is modified as:

20. Zero Boundary Conditions • At far longitudinal tails, the space charge tune shift becomes so small, that local incoherent spectrum is covering the coherent line. This results in the decoherence of the collective motion beyond this point; the eigen-mode goes to zero after that: • A specific case is a square potential well (considered by Mike Blaskiewicz for the air-bag distribution in 1998). For this singular potential well, at its boundaries • With zero boundary conditions, the obtained equation for the space charge modes has a full orthonormal basis of solutions:

21. Full orthonormal basis of no-wake problem • At the bunch core, the k -th eigen-function behaves like or ; the eigenvalues are estimated as (zero wake) • If the wake is small enough, it can be accounted as a perturbation, giving the coherent tune shift as a diagonal matrix elements (similar to Schrödinger Equation in Quantum Mechanics):

22. Modes for Gaussian bunch • For the Gaussian distribution, , measuring eigenvalues in , the no-wake equation reduces to (no parameters!): Eigenfunctions for Gaussian bunch. The cyan dashed line is the bunch linear density, normalized similar to the modes. Asymptotic:

23. Coherent tune shifts Coherent tune shifts (eigenvalues) for the Gaussian bunch, in the units of : Each eigenvalue, except the first three, lies between two nearest squares of integers: Note the structural difference between this and no-space-charge coherent spectra. While the former is counted by squares of natural numbers, the latter is counted by integers.

24. Growth rates for Gaussian bunch and constant wake • Coherent growth rates for the Gaussian bunch with the constant wake as functions of the head-tail phase , for the modes 0 ,1, 2 and 3. • The growth rates reach their maxima at , where . After its maximum, the high order mode changes its sign at to the same sign as the lowest mode, tending after that to • The instabilities can be damped by Landau damping.

25. Intrinsic Landau damping • Landau damping is a mechanism of dissipation of coherent motion due to transfer of its energy into incoherent motion. The coherent energy is transferred only to resonant particles - the particles whose individual frequencies are equal to the coherent frequency. • How the resonant particles can exist, when the space charge strongly separates coherent and incoherent motion? • In the longitudinal tails, space charge tune shift  0, so the Landau energy transfer is possible at the tails. • It happens at the ‘decoherence point’ , where For Gaussian bunch

26. Intrinsic Landau damping (2) • At the decoherence point, the rigid-beam approximation is not valid any more, since single-particle local spectrum is covering here the coherent line. So any calculations for this point, based on the above space charge mode equation, are estimates only, valid within a numerical factor. • At the decoherence point, the coherent motion is transferred into incoherent. After M times of passing the decoherence point, the individual amplitude is excited by • Thus, the entire Landau energy transfer from the mode after M >>1 turns can be expressed as

27. Intrinsic Landau damping (3) • The power of the Landau energy transfer is calculated as • Since the space charge phase advance Ψ>>1, the sum over many resonance lines n can be approximated as an integral over these resonances. After transverse average and longitudinal saddle-point integration, it yields: • The mode energy dissipation is directly related to the Landau damping rate by with the mode energy numbers For Gaussian bunch

28. Intrinsic Landau damping (4) • This yields the damping rate • This damping does not assume any nonlinearity of the external fields, neither longitudinal (RF), nor the transverse (nonlinear lattice elements). • It does not depend on the chromaticity. The beam stability depends on the chromaticity, since the coherent tune shift does.

29. Landau damping by lattice nonlinearity • The nonlinearity modifies the single-particle equation of motion: • For , there is a point in the bunch , where the nonlinear tune shift exactly compensates the local space charge tune shift: • At this point, the particle actually crosses a resonance of its incoherent motion with the coherent one. Crossing the resonance excites the incoherent amplitude by leading to a dissipation of the coherent energy.

30. Landau damping by lattice nonlinearity (2) • This energy dissipation gives the Landau damping rate • For a round Gaussian bunch and symmetric octupole-driven nonlinear tune shift , • Total damping rate is a sum of the intrinsic rate and the nonlinearity-related rate. Since their space charge dependence is the same and numerical coefficients are comparable, the leading contribution is determined by a ratio of the synchrotron tune and the average nonlinear tune shift.

31. General equation for space charge modes • When both synchrotron tune and coherent tune shift are small compared to the space charge tune shift, then: • This is valid for any ratio between the coherent tune shift and the synchrotron tune. • The sought eigenfunction can be expanded over the full orthonormal basis of the no-wake modes ,

32. Method to solve the general equation • This reduces the problem to a set of linear equations:

33. TMCI For zero chromaticity, instabilities are possible due to • Over-revolution (or couple-bunch) wake (not considered here…) • Coherent tune shift large enough to make significant mode coupling (TMCI = transverse mode coupling instability). This may be expected when the coherent tune shift is comparable with the distance between the unperturbed modes, . However, this expectation appears to be incorrect: • The TMCI threshold is normally much higher than that.

34. Vanishing TMCI • While the conventional head-tail modes are numbered by integers, the space charge modes are numbered by natural numbers: • This is a structural difference, leading to significant increase of the transverse mode coupling instability: the most affected lowest mode has no neighbor from below. Coherent tunes of the Gaussian bunch for zero chromaticity and constant wake versus the wake amplitude. Note high value of the TMCI threshold.

35. TMCI threshold for arbitrary space charge A schematic behavior of the TMCI threshold for the coherent tune shift versus the space charge tune shift.

36. Summary • A theory of head-tail modes is presented for space charge tune shift significantly exceeding the synchrotron tune and the coherent tune shift. • A general equation for the modes is derived, its spectrum is analyzed. • Landau damping is calculated, both without and with lattice nonlinearities. • TMCI is shown to have high threshold due to a specific structure of the coherent spectrum of the space charge modes. • Couple-bunch instabilities for the space charge modes are the next step to analyze. • The results can be applied for any ring with bunched beam and significant space charge (Project X, GSI, CERN).