An Introduction to the ILC Damping Rings: An Introduction to the ILC Damping Rings Andy Wolski
University of Liverpool and the Cockcroft Institute
29 November, 2006
Overview of this talk…: Overview of this talk… Introduction
Functions of damping rings in a linear collider.
Overview of present ILC damping rings configuration.
“Global” design constraints.
Principles of operation
Timing issues.
Radiation damping.
Equilibrium beam sizes.
Beam dynamics issues
Vertical emittance and coupling correction.
Dynamic aperture.
Some collective (intensity-dependent) effects.
Technical subsystems
Injection and extraction kickers.
Damping rings are critical to linear collider performance: Damping rings are critical to linear collider performance Damping rings perform several important functions:
Reduce injected beam emittances by 6 orders of magnitude.
Remove jitter from the source, providing a highly stable beam for tuning of downstream systems.
Delay the beams from the source, allowing for feed-forwards.
ILC Damping Rings are large electron storage rings.
Circumference is 6.7 km.
Beam energy is 5 GeV.
Two rings (one for positrons, one for electrons) in a single tunnel.
Beam current is 400 mA.
Roughly 200 m of wiggler, to provide 25 ms damping time. Introduction
Slide4: Present ILC configuration has damping rings located “centrally” around the interaction region. Introduction
Slide5: Arc 1 (818 m) Arc 2 (818 m) Arc 3 (818 m) Arc 4 (818 m) Arc 5 (818 m) Arc 6 (818 m) short straight A (249 m) short straight B (249 m) short straight C (249 m) short straight D (249 m) long straight 1 (400 m) long straight 2 (400 m) injection extraction shaft/large cavern A shaft/large cavern C e+ 8 RF cavities 10 RF cavities wiggler wiggler wiggler small cavern 1 small cavern 2 IP wiggler Introduction
Damping ring designs are constrained by ILC parameters: Damping ring designs are constrained by ILC parameters An ILC “machine pulse” consists of one electron and one positron bunch train, extracted from the damping rings, accelerated in the main linacs and collided at the IP.
New bunch trains must simultaneously be produced by the sources, to replace the bunch trains extracted from the damping rings.
Machine pulse repetition rate is 5 Hz. Each bunch train consists of:
5782 bunches with 1010 particles per bunch.
Bunch separation of 189 ns in the main linacs.
Total bunch train length of 1.1 ms, or 328 km.
Positron and polarised electron sources produce bunches that are too large to generate much luminosity. We use damping rings to reduce the bunch emittances in the 200 ms between machine pulses.
The damping rings cannot be large enough to accommodate a full bunch train with bunches separated by 189 ns.
We “compress” the bunch trains in the damping rings by injecting and extracting individual bunches, to reduce the separation between bunches in the rings. Introduction
Damping ring designs are constrained by ILC parameters: Damping ring designs are constrained by ILC parameters Circumference Damping times Beam energy Lattice design RF voltage RF frequency Bunch train length Injection/extraction kicker performance Bunch charge Injected bunch emittances Extracted bunch emittances Machine pulse repetition rate Main linac RF frequency Beam instabilities Costs Availability and reliability Technical systems performance Introduction
Timing issues need to be carefully considered: Timing issues need to be carefully considered Principles of operation Consider a damping ring with h stored bunches, with bunch separation t.
If we fire the extraction kicker to extract every nth bunch, where n is not a factor of h, then we extract a continuous train of h bunches, with bunch spacing n×t.
An added complication is that we want to have regular gaps in the fill in the damping ring, for ion clearing. 1 5 2 3 4 6 1 2 3 4 5
Injection and extraction kickers must operate in ~ nanoseconds: Injection and extraction kickers must operate in ~ nanoseconds trajectory of stored beam trajectory of incoming beam preceding bunch following bunch empty RF bucket injection kicker 1. Kicker is OFF. “Preceding” bunch exits kicker electrodes. Kicker starts to turn ON. 2. Kicker is ON. “Incoming” bunch is deflected by kicker. Kicker starts to turn OFF. 3. Kicker is OFF by the time the following bunch reaches the kicker. Principles of operation
Radiation damping: Radiation damping Charged particles emit photons (synchrotron radiation) in bending magnets.
The majority of photons are emitted within a cone of angle 1/ around the instantaneous direction of motion of the particle.
For ultra-relativistic particles, is very large.
Particles lose longitudinal and transverse momentum in bending magnets.
Neglecting dispersion and chromaticity, the trajectory of the particle after emitting a photon is the same as it would be if no photon were emitted. particle trajectory closed orbit emitted photon bending magnet Principles of operation
Radiation damping: Radiation damping In an RF cavity, the particle sees an accelerating electric field parallel to the closed orbit: the RF cavities in a storage ring restore the energy lost by synchrotron radiation.
The increase in momentum of a particle in an RF cavity is parallel to the closed orbit. This leads to a reduction in the amplitude of betatron oscillations of the particle. particle trajectory closed orbit Principles of operation
Radiation damping: Radiation damping The combined effect of energy loss in the dipoles and energy gain in the RF cavities is to reduce the betatron amplitudes of particles over many turns around the ring.
The rate at which the betatron amplitudes are reduced is clearly related to the rate at which the particles lose energy in the dipoles.
A detailed analysis gives the result for the radiation damping time:
where E is the beam energy, U0 is the energy loss per turn, and T is the revolution period, and x is the decay constant for betatron oscillations:
Principles of operation
Radiation damping: Radiation damping Note that the emittance of a bunch of particles is given by the average of the square of the betatron amplitudes of particles in the bunch:
Therefore, the emittance decays with a time constant equal to half the radiation damping time: Principles of operation
Damping time in the ILC damping rings: Damping time in the ILC damping rings The damping time in the ILC damping rings is determined by the specifications on the injected and extracted emittances.
Injected positron emittance y(t = 0) = 0.01 m
Extracted positron vertical emittance y(t = 200 ms) = 20 nm
We find that the necessary damping time is:
To allow for effects such as quantum excitation (which slow the damping near equilibrium), we design for a damping time of around 25 ms. Principles of operation
Radiation damping: Radiation damping To calculate the radiation damping time in a storage ring, we need to know the energy loss per turn, U0. This can be found from classical electromagnetic theory. The result is:
where E is the beam energy, C is a physical constant, with value:
is the local radius of curvature of the trajectory of particles in the dipoles in the magnetic lattice, and the integral extends around the entire circumference. Principles of operation
Radiation damping in the damping rings: Radiation damping in the damping rings Let us make a crude estimate of the damping time in a 6.7 km, 5 GeV electron storage ring.
Assuming that the dipoles make up a fraction D 10% of the circumference, the energy loss per turn is:
and the damping time is:
This is much too long: the ILC damping rings need a damping time of 25 ms.
We need to use wigglers to provide extra bending field without increasing the circumference. Principles of operation
Damping wigglers reduce the damping time: Damping wigglers reduce the damping time A wiggler is simply a sequence of short, strong dipoles with adjacent dipoles bending the beam in opposite directions, so that the net deflection of the beam is zero.
The energy loss in a wiggler of length Lw and peak field Bw is:
To provide a damping time of 25 ms in a 6.7 km, 5 GeV storage ring with a wiggler of peak field 1.6 T, we need a total length of wiggler of 220 m. By = Bw sin(kzz) Principles of operation
Equilibrium beam sizes: Equilibrium beam sizes The emission of radiation leads to excitation of the emittances as well as damping.
A particle following the closed orbit has zero betatron amplitude. If this particle radiates a photon, it loses energy. If this occurs at a location where the dispersion is non-zero, there is a new closed orbit for the particle.
The particle will start to make betatron oscillations around the new closed orbit: there has been an increase in emittance of the beam. particle
trajectory closed orbit emitted photon bending magnet Principles of operation off-momentum closed orbit
Equilibrium emittances: Equilibrium emittances Excitation of the beam emittance by synchrotron radiation is a quantum effect, resulting from the emission of radiation in individual photons.
The equilibrium emittance of a beam is determined by the balance between radiation damping and quantum excitation.
The horizontal emittance is essentially determined by the dispersion generated by the main bending magnets.
Damping ring lattice designs generally use a style of arc lattice cell known as a “Theoretical Minimum Emittance” (or TME) cell, which achieves low dispersion in the bends.
With the large circumference of the ILC damping rings, it is possible (by using a large number of cells) to achieve low dispersion, and hence low emittance, using simple FODO arc lattice cells. Principles of operation
Theoretical Minimum Emittance (TME) arc cell: Theoretical Minimum Emittance (TME) arc cell Principles of operation
FODO arc cell: FODO arc cell Principles of operation
Vertical emittance: Vertical emittance The vertical emittance in a storage ring is generated by three effects:
The non-zero vertical opening angle of the synchrotron radiation (~ 1/) leads to a vertical “recoil” when a photon is emitted. This excites vertical betatron oscillations.
Typically, the vertical emittance generated by this effect is of the order of a few tenths of a picometer, around 10% of the emittance specification for the ILC damping rings.
Vertical orbit distortions lead to vertical dispersion, which generates vertical emittance in the same way as horizontal dispersion generates horizontal emittance.
The vertical orbit needs to be corrected at the level of a few tens of microns, to keep the vertical dispersion down to a few mm rms.
Skew quadrupole errors lead to the coupling of horizontal betatron motion into the vertical plane.
Vertical beam offset in the sextupoles (of the order of a few tens of microns) tends to dominate the sources of skew quadrupole field errors. Beam dynamics issues
Equilibrium emittances: Equilibrium emittances Beam dynamics issues Note: this table gives the geometric emittances, . When quoting ILC parameters, it is usual to give the normalised emittances, , since these quantities are conserved during acceleration. For the ILC damping rings at 5 GeV, 10,000.
Ultra-low vertical emittance in the KEK-ATF: Ultra-low vertical emittance in the KEK-ATF Y. Honda et al, “Achievement of Ultralow Emittance Beam in the Accelerator Test Facility Damping Ring,” Phys. Rev. Lett. 92, 054802-1 (2004). Beam dynamics issues
Ultra-low vertical emittance in the KEK-ATF: Measuring beam sizes of a few microns presents a technical challenge. In the ATF, this is achieved using a laser wire.
A laser beam is focused to a waist of a few microns, and scanned across the electron beam. Gamma rays from Compton scattering are detected: the rate at which the gamma rays are produced gives the electron beam intensity as a function of transverse position. Ultra-low vertical emittance in the KEK-ATF Beam dynamics issues
Dynamic aperture: Dynamic aperture The emittance of the injected beam is orders of magnitude larger than the extracted (“damped”) beam.
The positron source produces a beam with normalised emittance 0.01 m.
For the ILC damping rings, this means a typical beam size is around 6 mm.
The vacuum chamber can be made with a physical aperture large enough to accommodate the injected beam without losses…
…but achieving a large dynamic aperture is more difficult.
The dynamic aperture in a storage ring is limited by higher-order field components (sextupoles, octupoles…)
The effect of higher-order field components is to introduce nonlinear terms into the equations of motion for a particle moving through the lattice.
At some amplitude, the nonlinear oscillations of a particle moving through the lattice become unstable: the amplitude of the oscillations increases rapidly, and the particle hits the vacuum chamber. Beam dynamics issues
Dynamic aperture: example in a FODO lattice: Dynamic aperture: example in a FODO lattice A phase space portrait is produced by:
taking a set of particles with regular spaced over a range of betatron amplitudes;
tracking the particles over some number of turns;
plotting the phase space coordinates of every particle on every turn.
Phase space portraits are useful for giving a “rough and ready” picture of nonlinear effects (tune shifts and resonances). Horizontal phase space portrait (tune = 0.28) Beam dynamics issues
Dynamic aperture: example in a FODO lattice: Dynamic aperture: example in a FODO lattice The dynamic aperture depends strongly on the tune of the lattice. tune = 0.25 tune = 0.31 tune = 0.36 tune = 0.33 Beam dynamics issues
Dynamic aperture is typically dominated by sextupole magnets: Dynamic aperture is typically dominated by sextupole magnets Typically, the nonlinear dynamical effects in a storage ring are dominated by sextupoles.
The sextupoles are needed to correct the chromaticity (the change in tune with the energy of a particle).
To reduce the impact of the sextupoles on the dynamic aperture, the lattice has to be designed to keep the sextupole strengths as low as possible. x x x SF SF SD SD Beam dynamics issues
Dynamic aperture is computed from tracking studies: Dynamic aperture is computed from tracking studies Dynamic aperture plots often show the maximum initial values of stable trajectories in x-y coordinate space at a particular point in the lattice, for a range of energy errors.
The beam size (injected or equilibrium) can be shown on the same plot.
Generally, the goal is to allow some significant margin in the design - the measured dynamic aperture is often significantly smaller than the predicted dynamic aperture.
This is often useful for comparison, but is not a complete characterization of the dynamic aperture: a more thorough analysis is needed for full optimization. 5inj 5inj OCS: Circular TME TESLA: Dogbone TME Beam dynamics issues
Dynamic aperture: frequency map analysis : Dynamic aperture: frequency map analysis A more complete characterization of the dynamics can be carried out using Frequency Map Analysis.
Track a particle for several hundred turns through the lattice.
Use a numerical algorithm (e.g. NAFF; or interpolated Fourier-Hanning) to determine the betatron tunes with high precision.
Continue tracking for several hundred more turns.
Find the tunes for the second set of tracking data.
Plot the tunes on a resonance diagram; use a color scale to represent the change in tunes between the first and second sets of tracking data (the “diffusion rate”). Beam dynamics issues
Collective effects: Collective effects Particles are guided through an accelerator beamline using “external” fields from components such as magnets and RF cavities.
Particles in an accelerator will themselves generate fields that lead to interactions with:
the surrounding environment;
vacuum chamber
instrumentation and diagnostics
other particles in the beam.
As the number of particles in the accelerator increases, the interactions resulting from fields generated by the particles have greater impact, and can result in:
increases in emittance;
changes in betatron or synchrotron tune;
changes in the beam distribution;
coherent oscillations of whole bunches;
heating of components in the vacuum chamber. Beam dynamics issues
Collective effects: Collective effects Fundamentally, all “collective effects” are manifestations of electromagnetic interactions between particles in a beam, either directly, or mediated by vacuum chamber components, or by other particles in the beam pipe (ions or electrons).
Different types of collective effect are identified, depending on the circumstances in which the effect is observed, the dynamical mechanism that is involved, and its effect(s) on the beam.
Effects of concern for the damping rings include:
space charge
intrabeam scattering
Touschek effect
microwave instability
resistive-wall instability *
ion instabilities
electron cloud effects *
higher-order mode heating Beam dynamics issues
Wakefields: Wakefields The electromagnetic fields generated by a particle or a bunch of particles moving through a vacuum chamber are usually described as wakefields.
The electromagnetic fields around a bunch must satisfy Maxwell’s equations.
The presence of a vacuum chamber imposes boundary conditions that modify the fields.
Fields generated by the head of a bunch can act back on particles at the tail, modifying their dynamics and (potentially) driving instabilities. Wake fields following a point charge in a cylindrical beam pipe with resistive walls. (Courtesy, K. Bane) Beam dynamics issues
Coupled-bunch instabilities: Coupled-bunch instabilities We can describe the kick on the trailing particle (2) from the wakefield of the leading particle (1) in terms of a wake function (N0 is the bunch charge):
In a storage ring containing M bunches, we construct the equation of motion: betatron oscillations multiple turns multiple bunches Beam dynamics issues
Coupled-bunch instabilities: Coupled-bunch instabilities The equation of motion (from the previous slide) is:
We try a solution for the nth bunch of the form:
Substituting this solution into the equation of motion, we find an equation that gives us the mode frequency for a given mode number .
The imaginary part of gives the instability growth (or damping) rate. spatial (bunch number) dependence time dependence Beam dynamics issues
Coupled-bunch instabilities: Coupled-bunch instabilities In a coupled-bunch instability, the bunches perform coherent oscillations.
The mode number gives the phase advance from one bunch to the next at a given moment in time.
The examples here show the modes ( = 0, 1, 2 and 3) in an accelerator with M = 4 bunches.
Each mode can have a different growth (or damping) rate.
From A. Chao, “Physics of Collective Beam Instabilities in Particle Accelerators,” Wiley (1993).
Beam dynamics issues
Resistive-wall instability: Resistive-wall instability For the ILC damping rings, the resistive-wall wakefields are expected to lead to an instability with the fastest modes having growth times ~ 20 turns.
This is much faster than the synchrotron radiation damping rate, and close to the limit of the damping rates that can be provided by fast feedback systems.
The transverse resistive-wall wakefield for a chamber with length L and circular cross-section of radius b is given (for z<0) by:
Implications for the ILC damping rings are:
- beam pipe radius must be as large as possible to keep the wakefields small - note that the wakefield (and hence the growth rates) vary as 1/b3;
- beam pipe must be constructed from a material with good electrical conductivity (e.g. aluminum) to keep the wakefields small - note that the wakefields vary as 1/c Beam dynamics issues
Resistive-wall instability: Resistive-wall instability Resistive-wall growth rates in a 6 km ILC damping ring lattice: Linear scale:
All modes. Log scale:
Unstable modes only. Note: Revolution frequency 50 kHz. Synchrotron radiation damping time 25 ms. Beam dynamics issues
Electron cloud effects: Electron cloud effects In a positron (or proton) storage ring, electrons are generated by a variety of processes, and can be accelerated by the beam to hit the vacuum chamber with sufficient energy to generate multiple “secondary” electrons.
Under the right conditions, the density of electrons in the chamber can reach high levels, and can drive instabilities in the beam.
Important parameters determining the electron cloud density include:
the bunch charge and bunch spacing
the properties of the vacuum chamber surface (the “secondary electron yield”) Beam dynamics issues
Electron cloud effects: Electron cloud effects Simulations of electron-cloud build-up need to include all relevant effects (chamber surface, beam pattern, magnetic and electric fields etc.)
Depending on the SEY, peak cloud density can vary by orders of magnitude. Simulation of e-cloud build-up in an ILC damping ring, by Mauro Pivi, using Posinst. Beam dynamics issues
Electron cloud effects: Electron cloud effects Interaction between the beam and the electron-cloud is a complicated phenomenon. In the ILC damping rings, the dominant instability mode is expected to be a “head-tail” instability, which may appear as a blow-up of vertical emittance.
The effects are best studied by simulation. Various effects need to be taken into account, including the density enhancement (by an order of magnitude) that can occur in the vicinity of the beam during a bunch passage. Simulation of vertical emittance growth in a 6 km ILC damping ring in the presence of electron cloud of different densities. (K. Ohmi) Beam dynamics issues
Electron cloud effects: Electron cloud effects To avoid instabilities associated with electron cloud, we expect to need to keep the average electron cloud density below ~ 1011 m-3.
This will require keeping the peak SEY of the chamber surface below ~ 1.1, which will be a challenging task.
Presently, three main approaches to reducing the SEY are being investigated:
- Coating the aluminum vacuum chamber (peak SEY ~ 2) with a low SEY material, for example TiN or TiZrV.
- Cutting grooves in the vacuum chamber surface to “trap” and re-absorb low-energy secondary electrons before they can be accelerated by the beam.
- Using clearing electrodes.
Currently, an active research program is under way to find the most effective technique. Beam dynamics issues
Suppressing electron cloud with low-SEY coatings: Suppressing electron cloud with low-SEY coatings Coatings being investigated include TiN and TiZrV.
Achieving a peak SEY below 1.2 seems possible after conditioning.
Reliability/reproducibility and durability are concerns. Measurements of SEY of TiZrV (NEG) coating. (F. le Pimpec, M. Pivi, R. Kirby) Beam dynamics issues
Suppressing electron cloud with a grooved chamber: Suppressing electron cloud with a grooved chamber Electrons entering the grooves release secondaries which are reabsorbed at low energy (and hence without releasing further secondaries) before they can be accelerated in the vicinity of the beam. Beam dynamics issues
Suppressing electron cloud with clearing electrodes: Suppressing electron cloud with clearing electrodes Low-energy secondary electrons emitted from the electrode surface are prevented from reaching the beam by the electric field at the surface of the electrode. This also appears to be an effective technique for suppressing build-up of electron cloud. Beam dynamics issues
Injection and extraction kickers must operate in ~ nanoseconds: Injection and extraction kickers must operate in ~ nanoseconds trajectory of stored beam trajectory of incoming beam preceding bunch following bunch empty RF bucket injection kicker 1. Kicker is OFF. “Preceding” bunch exits kicker electrodes. Kicker starts to turn ON. 2. Kicker is ON. “Incoming” bunch is deflected by kicker. Kicker starts to turn OFF. 3. Kicker is OFF by the time the following bunch reaches the kicker. Technical subsystems
Injection and extraction kickers: Injection and extraction kickers Several different types of fast kicker are possible. For the ILC damping rings, the injection/extraction kickers are composed of two parts:
- fast, high-power pulser;
- stripline electrodes. Several technologies are possible for the fast, high-power pulser. The parameters for the ILC damping rings are very challenging, and pulser development is on-going.
The stripline electrodes are comparatively straightforward: they consist of two plates, connected to a high-voltage line, between which the beam travels.
The stripline design is fairly challenging, because of the need to provide a large on-axis field while maintaining field quality and physical aperture; and the need to match the impedance to the power supply. Technical subsystems
Injection and extraction kickers: Injection and extraction kickers Let us take a simplified model of the stripline electrodes, consisting of two infinite parallel plates. The beam travels in the z direction. We apply an alternating voltage between the plates:
From Maxwell’s equations, there are electric and magnetic fields between the plates: A particle traveling in the +z direction with speed c will experience a force: For an ultra-relativistic particle, 1, and the electric and magnetic forces almost exactly cancel: the resultant force is small. But for a particle traveling in the opposite direction to the electromagnetic wave, –1, and the resultant force is twice as large as would be expected from the electric force alone. x y z Technical subsystems
Injection and extraction kickers: Injection and extraction kickers Let us calculate the deflection of a particle traveling between a pair of stripline electrodes. Let us suppose that there is a voltage pulse of amplitude V and length 2L traveling along the electrodes, which consist of infinitely wide parallel plates of length L separated by a distance d:
The change in the (normalised) horizontal momentum of the particle is:
where E is the beam energy. In practice, we can account for the fact that the electrodes are not infinite parallel plates by including a geometry factor, g. Technical subsystems
Injection and extraction kickers: Injection and extraction kickers Stage 1: Leading bunch must exit kicker before voltage pulse arrives. Stage 2: Voltage pulse fills kicker as target bunch arrives. Stage 3: Voltage pulse fills kicker while target bunch is between striplines. Stage 4: Voltage pulse exits kicker before trailing bunch arrives. voltage pulse kicker target bunch Technical subsystems
Injection and extraction kickers: Injection and extraction kickers How large a kick is needed? Consider the extraction optics:
Assuming a distance from the kicker to the septum of 50 m, and a required beam offset from the reference trajectory of 30 mm at the septum, the required kick is: If we assume E/e = 5 GV, L = 30 cm, d = 20 mm and g = 0.7, we find that the required voltage pulse is 143 kV. This is not realistic for a ~ ns pulser! The solution is to use multiple pairs of striplines, each producing a ~ 10 kV kick. kicker quadrupole septum Technical subsystems
Injection and extraction kickers: Injection and extraction kickers Tests of DSRD fast pulser.
(A. Krasnykh) Output voltage 2.7 kV
Horizontal scale 1 ns/div Technical subsystems
Injection and extraction kickers: Injection and extraction kickers Tests of FID fast pulser at KEK-ATF.
(T. Naito) Technical subsystems
Final words: Final words Basic principles for the damping rings are well understood, but design constraints (including cost considerations) lead to a number of very challenging issues:
Beam dynamics
Dynamic aperture must be large to accept the injected positron beam.
Extremely low vertical emittance of 2 pm must be achieved on a routine basis.
Many and varied collective effects threaten the beam quality and stability.
Technical subsystems
Damping wiggler must have high field quality.
Injection and extraction kickers require further R&D to demonstrate rise and fall times, pulse amplitude, repetition rate, stability and reliability.
Vacuum system must have low impedance, and have good performance from point of view of ion and electron cloud effects.
The ILC damping rings are the subject of a highly active R&D programme, involving nearly 150 people from more than 20 different institutions.