Electron Cloud Induced Beam Dynamics at CesrTA. Kiran Sonnad

Transcription

1 Electron Cloud Induced Beam Dynamics at CesrTA Kiran Sonnad

2 List of Contributers Experiments Simulation Laurel Bartnik* Lillian Pentecost$ Mike Billing* Mauro Pivi& Mike Forster* Kiran Sonnad % John Flanagan Robert Holtzapple# Kiran Sonnad*,% Data Analysis Robert Holtzapple Lillian Pentecost Kiran Sonnad Sam Tucker# * CLASSE Cornell University $ Colgate University # California Polytechnic State Univ % KEK & SLAC Special Thanks to K. Ohmi, Gerry Dugan, Brian Heltsley Nate Rider, Robert Meller and Cesr Operators

3 About CesrTA The Cornell Electron Storage Ring (CESR) was colliding electrons and positrons in the CLEO detector. Since 2008, the storage has been reconfigured as a test facility to study damping rings. The detector region has a series of damping wigglers installed. The CesrTA program has involved understanding electron clouds in positron storage rings, tuning the machine for lower beam emittances, understanding other collective effects such as intrabeam scattering and fast ion effects. The program has made valuable contributions toward the conceptual design report (CDR) of the ILC. Several advances in modeling/simulating the system under different conditions have been made and compared against experiments. These same simulations programs, once validated at CesrTA can be used to study designs of future facilities.

4 The Witness Bunch Experiment Feed back... Positron Train Witness + Bunch e Use the train to generate the electron cloud Observe the behaviour of the witness bunch Alter the properties of the witness bunch - position behind the train - charge - feed back (on/off) - emittance, chromaticity

5 Instrumentation used for our studies The X-ray beam size monitor (xbsm) Captures synchrotron radiation emitted from a bunch onto a detector. A beam position monitor (BPM), that consists of four buttons that pick up signal from the beam A typical signal produced on the xbsm with an estimated beam size

6 Simulation method: CMAD *The accelerator is divided into several Segments which contain electron cloud. *The electron cloud of each segment is collapsed onto a 2D mesh. *The beam is divided into several slices represented by a series of 2D meshes. *The beam passes through the cloud slice by slice. *Both the species are evolved during this process. Program originally developed by Mauro Pivi

7 Computation Method *The computation is distributed over several processors using MPI routines. *Each processor handles one or more slice-cloud interaction. *Once the beam distribution of the slices, belonging to a processor is evolved, they can proceed to the next interacting point independent of slices from other processors. *The beam particles from all slices/processors collected and distributed once per turn. The computations are performed at the NERSC supercomputers located in Berkeley. Program originally developed by Mauro Pivi

8 Table of Parameters: Experiments and Computation Physical Parameters Circumference Energy Bunch Length Emittance Chromaticity Tunes Bunch Charge Bunch Spacing m 2.085GeV 1.22cm 2.6nm(x) 20pm(y) ~ (x) 9.628(y) nc 14ns Simulation Parameters number of IPs (x,y) extent Xσ Extent in z 8Xσ # of macro e # of macro e # of grid cells 128X128 # bunch slices 96 # of processors 96

9 Observed Tune Shifts using spectrum analyzer 98ns behind train *Shift in tune Between 1st bunch In train and witness Bunch. *Tune Shift occurs mostly in y in the bottom figure.. *Several other lines appear along with a large tune shift. 14ns behind train Conditions: 45 Bunch train, 14 ns spacing 0.5mA/bunch or 1.28nC/bunch. Witness bunch at same charge as bunch in train.

10 Tune Shift between Simulations with and without electron cloud Motion in x Motion in y Electron cloud at 1e11/m3 Same bunch charge as previous slide; Motion in x Motion in y Electron cloud at 1e12/m3 We see a tune shift only in y as in experiments

11 Amplitude of oscillation of beam centroid in simulations Cloud density = 1012/m3 x-sigma = 0.1 mm y-sigma = mm Centroid motion is self excited. and is noisy because of the small amplitude of oscillation. We must redo the calculations with an initial beam offset for a cleaner fft and better tune shift calculations.

12 Estimation of Cloud density from tune Shift. Cloud density can be estimated by 2ΔQ y ρ =γ r ec β y from: K Ohmi - PAC 2001 Electron cloud density can be estimated using the above formula. In experiments, only the tune shift is known. In simulations, cloud density and tune shift are known

13 Beam size measurements, 30 bunch train feedback on + witness bunch feedback on/off Note: The train is under feed back in both cases. Thus, multi bunch effects are minimized Witness bunches just behind the train have a much larger beam size in the absence of feedback. Significant beam size expansion is seen even with feedback. We dont have an explanation for the first bunch expansion!

14 Beam Size measurements 45 bunch train (feed on) + witness Bunch (feedback on/off) Longer train, less charge per bunch. We do not see the lead bunch blow up in this configuration. The results are otherwise similar to previous case. The hump near witness bunch 65 is most likely not due to electron cloud.

15 Tune shift of all witness bunches for the 45 bunch train case The Hump in horizontal tune peak corresponds to the Hump in vertical beam size of previous slide. The appearance of the large peaks in the vertical tune corresponds to the transition in beam size increase Of the witness bunch in the previous slide

16 Sample of oscillation amplitude of bunch with and without feedback Measurement from xbsm, Beam under feedback 20th bunch of 45 bunch train BPM data of witness Bunch 33 behind 30 Bunch train NOTE: The conditions and location of the two do not match so they are not corresponding data sets. The number must be regarded as Typical values.

17 Effect of variation of chromaticity e+45 Bunch Train Vary Chromaticity I=0.6mA/bun 4/19/14 File:51025 WB60 XQ1 = -135 File:51109 WB60 XQ1 = -100 File:51114 WB60 XQ1 = -50 File:51045 WB54 XQ1=-135 File:51119 WB54 XQ1=-100 File:51124 WB54 XQ1=-50 File:51065 WB49 XQ1=-135 File:51129 WB49 XQ1=-100 File:51134 WB49 XQ1=-50 File:51088 WB46 XQ1=-135 File:51139 WB46 XQ1=-100 File:51144 WB46 XQ1= v Average Vertical Beam Size ( m) ns spacing Place 0.6mA witness bunch at number 46, 49, 54, and 60. Larger vertical beam size with feedback on. Typically, Lower chromaticity provides larger beam size e+45 Bunch Train Vary Chromaticity Witness Bunches 54 and 60 I=0.6mA/bun 4/19/14 Bunch Number e+45 Bunch Train Vary Chromaticity Witness Bunches 49 and 46 I=0.6mA/bun 4/19/14 Average Vertical Beam Size ( m) 45 v 80 v Average Vertical Beam Size ( m) File:51063 WB49 XQ1=-135 FDBK On File:51061 WB49 XQ1=-135 FDBK Off File:51127 WB49 XQ1=-100 FDBK On File:51125 WB49 XQ1=-100 FDBK Off File:51132 WB49 XQ1=-50 FDBK On File:51130 WB49 XQ1=-50 FDBK Off File:51086 WB46 XQ1=-135 FDBK On File:51084 WB46 XQ1=-135 FDBK Off File:51137 WB46 XQ1=-100 FDBK On File:51135 WB46 XQ1=-100 FDBK Off File:51142 WB46 XQ1=-50 FDBK On File:51140 WB46 XQ1=-50 FDBK Off Bunch Number File:51023 WB60 XQ1 =-135 FDBK On File:51021 WB60 XQ1 =-135 FDBK Off File:51107 WB60 XQ1 =-100 FDBK On File:51105 WB60 XQ1 =-100 FDBK Off File:51112 WB60 XQ1 =-50 FDBK On File:51110 WB60 XQ1=-50 FDBK Off File:51043 WB54 XQ1=-135 FDBK On File:51041 WB54 XQ1=-135 FDBK Off File:51117 WB54 XQ1=-100 FDBK On File:51115 WB54 XQ1=-100 FDBK Off File:51122 WB54 XQ1=-50 FDBK On File:51120 WB54 XQ1=-50 FDBK Off Bunch Number 58 60

18 Effect of Variation of current in witness bunch e+45 Bunch Train Vary Witness Bunch Current 45 bunch train at I=0.6mA/bun 4/19/ File:51063 WB49 I=0.6mA FDBK On File:51061 WB49 I=0.6mA FDBK Off File:51170 WB49 I=0.75mA FDBK On File:51168 WB49 I=0.75mA FDBK Off File:51175 WB49 I=1.0mA FDBK On File:51173 WB49 I=1.0mA FDBK Off File:51086 WB46 I=0.6mA FDBK On File:51084 WB46 I=0.6mA FDBK Off File:51180 WB46 I=0.75mA FDBK On File:51178 WB46 I=0.75mA FDBK Off Bunch Number e+45 Bunch Train Vary Witness Bunch Current Witnes Bunches 54 and 60 4/19/14 Bunch Number Place 0.6, 0.75, 1mA witness bunch at number 46, 49, 54, and 60. Larger vertical beam size with feedback on. Higher witness bunch current provides larger beam size in most cases. 45 v 0 Average Vertical Beam Size ( m) 20 e+45 Bunch Train Vary Witness Bunch Current Witnes Bunches 46 and 49 4/19/14 v 80 14ns spacing Average Vertical Beam Size ( m) File:51025 WB60 I=0.6mA File:51150 WB60 I=0.75mA File:51155 WB60 I=1.0mA File:51045 WB54 I=0.6mA File:51160 WB54 I=0.75mA File:51165 WB54 I=1.0mA File:51065 WB49 I=0.6mA File:51172 WB49 I=0.75mA File:51177 WB49 I=1.0mA File:51088 WB46 I=0.6mA File:51182 WB46 I=0.75mA v Average Vertical Beam Size ( m) File:51023 WB60 I=0.6mA FDBK On File:51021 WB60 I=0.6mA FBCK Off File:51148 WB60 I=0.75mA FDBK On File:51146 WB60 I=0.75mA FDBK Off File:51153 WB60 I=1.0mA FDBK On File:51151 WB60 I=1.0mA FDBK Off File:51043 WB54 I=0.6mA FDBK On File:51041 WB54 I=0.6mA FDBK Off File:51158 WB54 I=0.75mA FDBK On File:51156 WB54 I=0.75mA FDBK Off File:51163 WB54 I=1.0mA FDBK On File:51161 WB54 I=1.0mA FDBK Off Bunch Number

19 Effect of varying cloud density, chromaticity And bunch current in simulations. *Increase in beam size with increased cloud density and a transition to rapid beam size growth rate. *Increase in beam size growth rate with increased bunch charge. *Increase in beam size growth rate with decrease in chromaticity.

20 First order Head-Tail Motion from Simulations Cloud density = 1e11/m3 Cloud density = 1e12/m3 Perform a linear fit over the centroid of all the bunch slices for every turn. Motion of bunch centroid is removed in the fit. Amplitude of head-tail oscillation (m=1 mode) increases by An order of magnitude with increase in cloud density by a factor of 10. Note: The head tail motion is self induced by the electron cloud.

21 Second order Head-Tail Motion, from simulations Cloud density = 1e11/m3 Cloud density = 1e12/m3 Remove the centroid motion and orientation of bunch, and perform a parabolic fit over all the bunch slices for every turn. The amplitude of this mode increases by an order of magnitude When the cloud density is increased by a factor of 10

22 Turn by Turn variation of linear slope: varying cloud densities The magnitude of the slope is obtained from linear fit for every turn. e The amplitude of oscillation of slope clearly increases with increased cloud density. Spectrum shows strong synchrotron side bands, and a peak at betatron tune only for higher cloud density.

23 Total Charge in Bunch Turn by Turn, Varying Electron Cloud Density Initial Bunch Population = 0.8e10 Positrons

24 Conclusion We worked on a series of experiments designed for observing single bunch dynamics induced by electron clouds on positron beams. Simulations were performed using the program CMAD. We observed emittance growth with increased cloud density, decreased chromaticity, increased bunch current. Simulations showed a similar behavior when the above parameters were varied. Simulations showed that head tail motion was induced from electron clouds and is directly co-related with increased emittance growth. Significant emittance growth was observed even when bunch oscillation was suppressed with feedback.