Two-Stream Phenomena in CLIC G. Rumolo and D. Schulte for the ACE, 3 September 2008 * thanks to W. Bruns, R. Tomás, SPSU Working Team General introduction to two-stream phenomena: –Electron clouds –Fast ion instabilities Electron cloud in the CLIC damping rings –Simulations –Technical solutions under investigation Fastion code used to model fast ion instabilities –Description of the model –Simulations of the CLIC main linac conclusions and outlook

Principle of the multi-bunch multipacting.  Electrons can be generated via photoemission, rest gas ionization or beam loss at the chamber walls. They then multiply due to the seconday emission process Electron cloud build up: a multi-bunch process...

Tune shift The tune increases along a train of positively charged particle bunches because the bunches at the tail of the train feel the strong focusing effect of the electron cloud formed by the previous bunches. Electron cloud instability in rings  Coupled bunch phenomenon: the motion of subsequent bunches is coupled through the electron cloud and the amplitude of the centroid motion can grow.  Single bunch phenomenon: the motion of head and tail of a single bunch can be coupled through an electron cloud and give rise to an instability Tune shift along the train as well as instabilities affecting only the last bunches of long trains have been observed in several machines (SPS, KEKB-LER, Cesr-TA), clearly pointing to the electron cloud as source of these phenomena. Effects of the electron cloud on the beam:

The ions produced by gas ionization can be focused by the electric field of the following bunches and they accumulate in the vicinity of the beam (trapping condition) Two-stream instability The ion cloud can affect the motion of the bunches and an unstable ion-electron coupled motion can be excited Ions: residual gas ionization N e e - beam l ≈ 4  z CO + ion CO molecule

Electron cloud build up in the wigglers of the CLIC damping ring (simulations done with Faktor2)  The electron cloud in the wigglers can have high density values if The PEY is high enough (i.e., more than 0.01% of the produced radiation is not absorbed by an antechamber or by special absorbers), independently of the SEY The SEY is above 1.3, independently of the PEY Central densities for different PEYs and SEYs 33 3

Instability simulations to check beam stability (simulations done with HEADTAIL)  In case of electron cloud build up, we assume the „best case“ in the wigglers:  wig = 1.8 x m -3  The beam is strongly unstable * Vertical centroid motion* Vertical emittance evolution

Against the electron cloud.....  If there is electron cloud in the CLIC-DR, the beam becomes unstable! o Conventional feedback systems cannot damp this instability (wider band needed) o It is necessary to find techniques against the formation of the electron cloud  Several mitigation techniques are presently under study: Low impedance clearing electrodes Solenoids (KEKB, RHIC) -however only usable in field free regions! Low SEY surfaces  Grooved surfaces (SLAC)  NEG coating  New coatings presently under investigation (SPS) Carbon coatings, studied by the SPS Upgrade Working Team, seem very promising and a possible solution....

Evolution of CNe9 over 1 month Courtesy M. Taborelli from SPSU-WT 2h 3d 8d 15d 23d The maximum SEY starts from below 1 and gradually grows to slightly more than 1.1 after 23 days of air exposure. The peak of the SEY moves to lower energy. SEY of CNe

Measurements inside the SPS confirm the lab measurements! Perhaps this is the strategy to get rid of electron cloud issues ??...  However, need to check the PEY of these coated surfaces in order to fully validate their use for DRs as well!  Tests planned at ESRF (maybe also in Cesr-TA) NEG and CNe exhibit no electron cloud activity in the SPS !!!

FASTION (I) Multi-bunch code, ions and electrons modeled as macro-particles Ions of an arbitrary number of species are created at each bunch passage and propagated through the train Line can be: A simple sequence of FODO cells, with two kicks per FODO cell Described through an output PLACET Twiss file Described through an output MAD-X Twiss file Electromagnetic interaction: the ions are kicked by the passing bunches and the bunch macro-particles feel the effect of the ion field Bunch n+1

FASTION (II) Acceleration along the line can be taken into account Initial and final energy are given through the standard input file and a simple linear model of acceleration is applied The PLACET Twiss file gives the energy at all the specified s locations along the line The coordinates of the transverse phase space shrink like  -1/2 Ex. CLIC Main Linac

FASTION (III)  Input file  All the necessary parameters are passed through a simple ascii input file (N el, N ion, N bunch to be specified in the source) Number_of_ion_species: 2 Partial_pressures_[nTorr]: Atomic_masses: Ionization_cross_sections_[MBarn]: Number_of_electrons_per_bunch: 3.7e+9 Bunch_spacing_[ns]: 500.e-3 Normalized_horizontal_emittance_(rms_value)_[nm]: 660. Normalized_vertical_emittance_(rms_value)_[nm]: 10. Bunch_length_(rms_value)_[ns]: 0.12e-3 Initial_relativistic_gamma: Final_relativistic_gamma: Number_of_FODOs: 500 FODO_Length_[m]: 40. Phase_advance_per_cell_[degrees]: 70. The bunches can be accelerated through the line CO (or N 2 ) and H 2 O Possible change during acceleration

FASTION (IV)  Output files  The code transports the bunch train through the line, generates the ions at each interaction and propagates them Ion distributions in x and y are stored at an arbitrary time step (for instance, at the beginning) The phase space coordinates of 4 sample ions are stored at the same time step as above Snap-shots of the bunch by bunch centroids and emittances are saved at each time step Time evolution of the beam centroids and emittances (averaged over one third of the train) are stored 12N b /32N b /3NbNb IIIIII HeadTail

Application to CLIC Main Linac  We have carried out simulations of fast ion instability usingthe lattice file from PLACET (Daniel) and the following parameters  Beam energy from 9 GeV to 1.5 TeV over a length of 20 km.  Residual gas pressure from 10 to 50 nTorr for each species. Only two species of ions have been considered (CO and H 2 O)  Bunch population is 4 x 10 9  Bunch spacing is 500 ps  Number of bunches (N b ) is 311  Normalized emittances are 680,10 nm

Application to CLIC Main Linac (II) Through the twiss.dat file, both beta functions and beam energy at different locations in the main linac are passed to the FASTION code and used for tracking with the generated ions. Decreasing beam sizes along the Linac

Application to CLIC Main Linac (III) As ions are produced by residual gas ionization, the distribution of ions as generated at the first bunch passage follows the beam sizes along the line

Application to CLIC Main Linac (IV) Loss of trapping along the Linac shown by the extension of the tails of the ion distribution at the first and the last bunch passage at three different locations

Application to CLIC Main Linac (V) The beam is stable with a vacuum pressure of 10nTorr, even giving initial offsets of 0.1    to the different bunches (constant along the train or randomly distributed)

Application to CLIC Main Linac (VI) The beam is unstable with a vacuum pressure of 50nTorr, and both a coherent centroid motion and a small incoherent emittance growth appear along the line

Conclusions and outlook The code FASTION has been developed and applied to study the fast ion instability in the CLIC Main Linac. –Detailed lattice from PLACET –Effect of acceleration included –Evident loss of trapping along the Linac (stabilizing) Simulations show that: –The beam is stable with P=10 nTorr (no amplitude growth observed) –The threshold of instability lies between 10 and 50 nTorr The present model still does not include –The effect of the rf fields on the ions (expected not to be critical) –The effect of field ionization. This may become very important in the late stages of the ML and could significantly alter this picture A model of field ionization is currently under development to be included in FASTION