Experimental Observations and Simulations of Electron-Proton Instabilities in the Spallation Neutron Source Ring

1
Experimental Observations and Simulations of
Electron-Proton Instabilities in the Spallation
Neutron Source Ring

• S. Cousineau, A. Shishlo, A. Aleksandrov, S.
  Assadi, V. Danilov, C. Deibele, M. Plum
• ECLOUD 07, Daegu, S. Korea

2
SNS Accelerator Complex
Accumulator Ring
Collimators
Front-End Produce a 1-msec long, chopped, H-
beam
Accumulator Ring Compress 1 msec long pulse to
700 nsec
1 GeV LINAC
Injection
Extraction
RF
Liquid Hg Target
RTBT
2.5 MeV
1000 MeV
87 MeV
186 MeV
387 MeV
Ion Source
HEBT
SRF, b0.61
DTL
RFQ
CCL
SRF, b0.81
Target
Chopper system makes gaps
945 ns
mini-pulse
Current
Current
1 ms macropulse
1ms
3
SNS Ring Parameters

• Design ring parameters
• 1 GeV beam
• Intensity 1.4?1014 ppp
• Power on target 1.4 MW
• Working point (6.23,6.20)
• Ring circumference 248 m
• Space charge tune shift 0.15

• For eP instability mitigation
• a) All pieces of vacuum chamber coated
  with TiN
• b) Solenoids near the regions with high
  loss (collimation?)
• c) Clearing electrode near the stripper
  foil

4
First Neutrons on April 28, 2006

• Beam and Neutronics Project Completion goals were
  met
• 1013 protons delivered to the target
• The SNS Construction Project was formally
  Completed in June 2006
• We have officially started SNS operations, and
  are in the power ramp-up phase.

5
Beam Power Ramp-Up Timeline of Recent Events
An aggressive power ramp-up schedule has been
adopted
90 kW demonstration
60 kW demonstration
60 kW operation
1 GeV demonstration
30 kW operation
6
High Intensity Beam Studies
For the nominal neutron production conditions, no
instabilities have been observed so far. None
were predicted.
8.5?1013 ppp

• Dedicated, high intensity beam experiments have
  been performed.
• 8.4?1013 ppp (13.5 uC) of bunched beam have been
  accumulated.
• 9.5?1013 ppp (16 uC) of coasting beam have been
  accumulated.
• For these experiments, we have varied parameters
• Chopped or coasting
• RF on or off
• beam intensity
• chromaticity (natural or corrected)
• lattice tunes

7
Electron-proton instability for Coasting Beam

• Most of the experiments performed for coasting
  beam configuration
• Instability observed beginning at gt2?1013 ppp.
• Instability is observed in both planes
  vertical stronger.
• No instability is observed at natural
  chromaticity.

BPM trace for a 16 ?C (1?1014 ppp) beam
8
Intensity Scan of e-p Instability for Coasting
Beam
Turn-by-turn plot of frequency spectrum
Observations

• Instability gets faster with increasing
  intensity (40 turns for 16 uC case).
• Frequency spectrum is more sharply peaked at
  higher intensity.
• At highest intensity, frequency 79 MHz.

9
Calculation of Effective e-p Impedance
Evolution of the Dominant Harmonic
We can estimate the effective impedance of the
electron cloud, at different intensities
8 ?C
Formula works well above threshold, requires no
beam distributions information.
16 ?C
8 ?C beam
Re(Z)168 K?/m
16 ?C beam
Re(Z)1.9 M?/m
We have seen 3 types of instabilities in SNS
ring. e-P has largest impedance, by far (over 3
times larger for 16 uC case).
10
Coupling observed between transverse planes

• Instability is observed in both planes.
• Coupling is observed between the planes.

Vertical BPM signal
Horizontal BPM signal
Both fractional tunes observed in the betatron
spectrum of the horizontal data (Qx0.23, Qy0.2)
11
Split Tunes Case
(6.23, 6.20)
(6.23, 6.20)
Nominal Tunes (6.23, 6.20)
(6.24, 6.16)
(6.24, 6.16)
Split Tunes (6.24, 6.16)
Tune splitting had only a small effect on the
instability amplitude and frequency spectrum.
12
Instability for a chopped beam with no RF
In the latest set of high intensity experiments,
8.5?1013 ppp of chopped beam accumulated with no
RF on.
Gap is mostly full by extraction. Some structure
remains.
BCM signal at extraction
Vertical BPM signal
e-P instability is observed in both planes
(vertical BPM signal shown at right)
13
Longitudinal Position of Instability
Instability occurs at flat top, closer to front
of the beam, and moves backwards.
Integrated signal for one electrode
turn 300
Real space turn-by-turn evolution of instability
turn 300
turn 200
head
tail
turn 50
14
Beam Loss in the Region of Instability
For same turn number, wall current monitor shows
beam loss in region of high instability activity.
15
Frequency Content of Chopped Beam Instability
For the chopped beam with no RF, in frequency
space the excitation bands drift downwards. The
instability starts before end of beam
accumulation.
16
e-P Simulations with ORBIT code

• The parallel ORBIT electron-cloud
• Includes the interactions of electron cloud and
  proton beam in both directions (electrons act on
  protons and protons act on electrons).
• Describes the electron cloud build up and
  includes a secondary emission surface model
  (Furman and Pivi model).
• Uses PIC method for space charge for both proton
  and electron beam.
• Uses 3D space charge for the proton beam
• Allows an arbitrary number of localized electron
  cloud in the ring, up to the limiting
  computational ability of the parallel system.
• Allows e-cloud nodes in magnets.
• Model has been benchmarked with analytic model
  (Y. Sato) and PSR experimental data (A. Shishlo)

17
Simulation of Chopped Beam Case
We performed simulations of the chopped beam
case. Simulations done in two stages
Simulations by A. Shishlo
Stage 1 Accumulate distribution. No ECloud nodes.
Do it once only.
Stage 2 Store distribution, insert ECloud nodes.
Do multiple runs, varying e-node parameters.

• Stage 2 parameters varied
• Number of e-Cloud nodes in the ring (lt 4 for
  computational expense).
• Location of e-Cloud nodes
• Type of initial electron cloud (surface or
  volume)
• Proton loss rate (how electrons are ejected)

Computational statistics - 6?106 proton
macroparticles - 10,000 30,000 electron
macroparticles per electron node - 60 CPUs, 80
GFlops for 24 hours.
18
Simulations of chopped beam with No RF

• Case parameters
• 2 electron cloud nodes one in drift, one in
  dipole.
• Proton loss rate 1?10-4 protons per meter.
• SEM parameters TiN, 100 electrons per proton.
• Did not have good knowledge of energy
  distribution in beam, or evolution of beam in
  gap.

• Observations
• Instability seen right at the beginning of
  storage.
• Frequency content of instability is fairly
  narrow, possibly because of localized e-Cloud
  nodes.

19
Comparison of simulation and measurement real
space
The longitudinal location of strongest
instability is roughly the same between
simulation and measurement. Both show migration
toward tails.
Turn-by-turn evolution of beam centroid
Simulated
Measured
head
tail
head
tail
20
Frequency comparison of simulation with
measurement
We see narrower excitation frequency in the
simulation 20 65 MHz. Excitation frequency
content and extent is likely due the position and
localization of the two ECloud nodes. We see
the same drift of excitation bands to lower
frequency in both simulation and experiment.
Measured
Measured
Simulated
Simulated
21
Comparison of centroid oscillation
15 mm
Experiment
Centroid oscillation is larger in experiment than
simulation. Experiment 15 mm Simulation 2.2
mm We see huge beam loss in experiment, but
almost no beam loss in simulation.
Simulation
22
Electron cloud density

• In the simulation, complete neutralization of
  the beam in the gap occurs.
• Unfortunately, no electron collectors were
  available in the experiment to measure e-cloud.

23
Summary

• We have accumulated up to 8.4?1013 ppp of bunched
  beam, and 9.5?1013 ppp of coasting beam.
• No instabilities are seen for bunched beam, or
  for natural chromaticity beam.
• We see e-p instability for coasting beam and
  chopped beam cases with no RF
• Simulations of the chopped beam show reasonable
  agreement with experiment, though some
  differences exist.
• More accuracy in simulation might be gained by
  adding more e-cloud nodes, and having better
  knowledge of beam energy spread.