Space Charge and High Intensity Studies on ISIS

Space Charge and High Intensity Studies on ISIS

C M Warsop

Reporting the work of

D J Adams, B Jones, B G Pine, C M Warsop, R E Williamson

ISIS Synchrotron Accelerator Physics

and: S J Payne, J W G Thomason,ISIS Accelerator Diagnostics, ISIS Operations, ASTeC/IB.

Contents

• Introduction

• Main Topics

1 - Profile Monitor Modelling

2 - Injection Painting

3 - Full Machine Simulations

4 - Half Integer Losses

5 - Image Effects & Set Code

• Summary

Introduction

• ISIS Spallation Neutron Source ~0.2 MW

- Commissioning Second Target Station

- Now ramping up operational intensity

- ISIS Megawatt Upgrade Studies started

• Will summarise our programme of Ring High Intensity R&D

- Underpins the work above (& has wider applications)

- Aim to understand intensity limits of present and upgraded machines

- Experimentally verify simulation and theory on ISIS where possible

- Broad: covers diagnostics, experiments, simulation, theory

0 2 4 6 8 10

0

0.5

1

1.5

2

2.5

Time (ms)

Beam

Inte

nsity p

pp x

1e13 /

Beam

Loss

Beam Intensity ppp x 1e13 Beam Loss

Circumference 163 m

Energy Range 70-800 MeV

Rep. Rate 50 Hz

Intensity 2.5x1013 → ~ 3.0x1013 protons per pulse

Mean Power 160 → ~ 200 kW

Losses Mean Lost Power ~ 1.6 kW (≤100 MeV)

Inj: 2% (70 MeV) Trap: 5% (<100 MeV)

Acceleration/Extraction: 0.1 – 0.01%

Injection 130 turn, charge-exchange paint injected beam of ~ 25 mm mr

Acceptances horizontal: 540 mm mr with dp/p 0.6% vertical: 430 mm mr

RF System h=2, frf =1.3-3.1 MHz, peak Vrf=140 kV/turn

h=4, frf =2.6-6.2 MHz, peak Vrf=80 kV/turn

Extraction Single Turn, Vertical

Tunes Qx=4.31, Qy=3.83 (variable with trim quads)

The ISIS Synchrotron

• Profile measurements essential for space charge study

- This work: Modelling & experiments to determine accuracy

- Overlaps with diagnostics R&D work - S J Payne et al

• Residual gas ionisation monitors

- Detect positive ions in 30-60 kV drift field

• Two main sources of error:

(1) - Drift Field Non-Linearities

(2) - Beam Space Charge

• Modelled dynamics of ions with

- CST Studio™ for fields

- “In house” particle trackers

ELECTRODE

DETECTOR

1. Profile Monitor Studies ~ 1

Rob Williamson, Ben Pine, Steve Payne

Potential from CST

y

zx

Φ(x,y)

Φ(y,z)

Introduction

Drift Field Error

2D Tracking Study

3D Tracking Study

1. Profile Monitor Studies ~ 2Rob Williamson, Ben Pine

(xs, ys)

xd

Particle Trajectory

- Field error distorts trajectories

- Measured position xd=F(xs,ys)

For given geometry find:

- Averaged scaling correction

- More complicated in 3D case

- Longitudinal fields – new effects

- Detected ions from many points

- Scaling corrections still work

- Ideas for modifications

Φ(x,y)

Φ(y,z) Blue: Trajectory of particles entering detector

Red: Origin of particles entering detector

Black: Transverse section of beam at given z

Trajectories as a function of z along beam

Space charge field distorts trajectories

Space Charge Error

-40 -20 0 20 40

-100

-75

-50

-25

0

25

50

-40 -20 0 20 40

-100

-75

-50

-25

0

25

50

-40 -20 0 20 40

-100

-75

-50

-25

0

25

50

1. Profile Monitor Studies ~ 3Rob Williamson, Ben Pine, Steve Payne

dd

sc

VE

Ex

1

Profile widths parabolic simulation

0.11

0.12

0.13

0.14

0.15

0.16

5.0 25.0 45.0 65.0 85.0 105.0

Electrode voltage / kV

Det

ect

ed 9

0% W

idth

/m

Theory Simulations Experimental data

S J Payne

Ion Trajectories (2D)

increasingdE

90% Width vs Vd

Sim & Meas & Theory

k vs Width

Sim (3D) & Meas

Width vs Vd-1

Simulation (3D)

• Increase in given percentage width

• Also - for “normal” distributions

• So can correct a profile for space charge

• Confirmed experimentally & in 2D/3D simulations

Simple calculation: trajectory deflection

1%%

dxx VkW

1 dV

%% xx Wk

Width vs Vd-1

Measurement

D

SCptDpm V

pKwKw

,,

• Good understanding of monitors

- Correction scheme: good to ±3 mm

• Experimental verification

- Many checks and agrees well

- Final checks needed: EPB monitor

• Monitor Developments (S J Payne)

- Multi-channel, calibration, etc

- Drift field increase and optimisation

• Seems to work well

- See next section …

1. Profile Monitor Studies ~ 4

Summary Rob Williamson, Ben Pine

Basic correction scheme

- drift field and space charge

- for near-centred, “normal” beams

3D simulation: original, “measured” and corrected profile

angular acceptance of detector, reduces errors to ± 3 mm

Injection Septum

Vertical Sweeper

Injection Dipoles

Foil Injected Beam

Closed Orbit

Dispersive Closed Orbit

2. Injection Painting ~ 1 Bryan Jones, Dean Adams

• ISIS Injection

- 70 MeV H- injected beam: 130 turns

- 0.25 μm Al2O3stripping foil

- Four-dipole horizontal injection bump

- Horizontal: falling B[t] moves orbit

- Vertical: steering magnet

• Studies of injection important for:

- ISIS operations and optimisation

- ISIS Megawatt Upgrade Studies

- Space charge studies

• Want optimal painting

- Minimal loss from space charge, foil

• Start is Modelling-Measuring ISIS

Injection Studies: Aims and Background

2. Injection Painting ~ 2Bryan Jones, Dean Adams

• Direct measurement of painting

- Use “chopped” beams

- Low intensity (1E11 ppp); less than 1 turn

- Inject chopped pulse at different times

- Least squares fit to turn by turn positions

- Extract initial centroid betatron amplitude

Injection Painting Measurements

-500 -400 -300 -200 -100 00

20

40

60

80

100

120

140

160

Time Before Field Minimum (us)

Ce

ntro

id E

mitt

anc

e (

pi m

m m

r)

• Profiles measured on RGI monitors

- Corrections as described above

• Plus other data …

- Injected beam, sweeper currents, …

00

2

2cos2

exp cocon zznQnQnQn

Az

• Compare Measurement-Simulation

- Normal anti-correlated case

- Trial correlated case

• Change vertical sweeper to switch

- Reverse current vs time function

-0.4 -0.2 0

Time (ms)

Simulation and Measurement: Normal Painting

2.5x1013 ppp2.5x1012 ppp

-0.3ms

-0.2ms

-0.1ms

-0.3ms

-0.2ms

-0.1ms

2.5x1013 ppp2.5x1012 ppp-0.3ms

-0.2ms

-0.1ms

-0.3ms

-0.2ms

-0.1ms


Horizontal

Vertical

Painting

anti-correlated

Horizontal Profile Vertical Profile

Key - Measured (corrected) - Simulation (ORBIT)

Not the final iteration, but pretty good agreement

2.5x1013 ppp2.5x1012 ppp-0.3ms

-0.2ms

-0.1ms

-0.3ms

-0.2ms

-0.1ms

2.5x1013 ppp2.5x1012 ppp-0.3ms

-0.2ms

-0.1ms

-0.3ms

-0.2ms

-0.1ms

Anti-correlated Correlated


Horizontal

Vertical - correlated Vertical Profile

Simulation and Measurement: Painting Experiment

Vertical ProfilePainting

Vertical - anti-correlated

Key - Measured (corrected) - Simulation (ORBIT)

• Follows expectations … [ran at 50 Hz OK!]

• Plan to develop and extend to study

- other painting functions: optimal distributions

- emittance growth (during & after injection)

- foil hits & related losses

Injection Simulation Details

- ORBIT multi-turn injection model

- Painting: H - Dispersive orbit movement; V - Sweeper Magnet

- Injection bump, momentum spread and initial bunching

- 2D transverse (with space charge)

- 1D longitudinal (no space charge yet)

3. Machine Modelling ~ 1 Dean Adams, Bryan Jones

(x,x’) (y,y’)

(x,y) (dE, phi)

Turn 9 Turn 39 Turn 69 Turn 99 Turn 129

Example: Normal anti-correlated case 2.5E13 ppp

ORBIT

Longitudinal Studies ~ work in progress

- TRACK1D - works well - basis of DHRF upgrade (C R Prior)

- Now working to model in detail in ORBIT (1D then 2.5D)

- Collaborating on tomography (S Hancock, M Lindroos, CERN)

3. Machine Modelling ~ 2

TRACK1D ORBIT 1D

Dean Adams

Tomography trialsComparisons and trials at 0.5 ms after field

minimum on ISIS for ~ 2.5x1013 ppp

(real data!)

Full Machine Modelling in ORBIT ~ work in progress

• Simulation of full machine cycle 2.5D – some reasonable results

- time variation of loss

→ reproduces main loss 0 - 3 ms

• Collimators now included

~ space variation of loss

→ good results (normal ops & Mice target)

3. Machine Modelling ~ 3 Dean Adams

Loss vs Time

Simulation

MeasurementB

LM s

igna

l*

* some energy dependence

Spatial LossLo

st P

artic

les

Incr

ea

se in

ten

sity

Importance for the ISIS RCS

• Transverse space charge - key loss mechanism

- Peaks at ~0.5 ms during bunching ΔQinc~-0.4

- In RCS is 3D problem: initially study simpler 2D case

4. Half Integer Losses ~ 1Chris Warsop

3.8 3.9 4 4.1 4.2 4.3 4.4 4.5Qx3.4

3.5

3.6

3.7

3.8

3.9

4Qy

1 ms

10 ms

Incoh Q Shift

• First step: envelope equation calculations

- ISIS large tune split case: independent h and v

- Get 8/5 “coherent advantage” (e.g. Baartman)

- Numerical solutions confirm behaviour

20 40 60 80 100 120

0.25

0.5

0.75

1

1.25

1.5

x

6 7 8 9f

10

20

30

40

amplitude20 40 60 80 100 120

0.25

0.5

0.75

1

1.25

1.5

x

6 7 8 9f

2

4

6

8

amplitude

20 40 60 80 100 120

0.25

0.5

0.75

1

1.25

1.5

x

6 7 8 9f

0.5

1

1.5

2

amplitude

20 40 60 80 100 120

0.2

0.4

0.6

0.8

1

1.2

y

6 7 8 9fy

5

10

15

20

amplitude

20 40 60 80 100 120

0.2

0.4

0.6

0.8

1

1.2

x

6 7 8 9fx

0.5

1

1.5

2

2.5

amplitude

1D 2DEnvelope Amplitude Frequency Amplitude Frequency

Envelope

Horizontal

Vertical

(Qh,Qv)=(4.31,3.83)

ORBIT 2D Simulation Results

- 5E4 macro particles; ~RMS matched waterbag beam

- Tracked for 100 turns; driven 2Qv=7 term

390 395 400 405 410 415 420 425 430 435 4400

500

1000

1500

2000

2500

3000

3500

Frequency of Dipole Motion: Horizontal

100 x azimuthal frequency

Am

plitu

de

340 345 350 355 360 365 370 375 380 385 3900

500

1000

1500

2000

2500

3000

Frequency of Dipole Motion: Vertical


Am

plitu

de

780 790 800 810 820 830 840 850 860 870 8800

1

2

3

4

5

x 104 Frequency of Envelope Motion: Horizontal


Am

plitu

de

680 690 700 710 720 730 740 750 760 770 7800

1

2

3

4

5

6

7

8

x 104 Frequency of Envelope Motion: Vertical


Am

plitu

de

0 2000 4000 6000 8000 10000 12000 14000 16000 18000-10

0

10

20

30

40

50

60

70x (3,1), x-y (2,2) and y (1,3) envelopes - 2nd Order

x (2,0)

x-y (1,1)

y (0,2)

390 395 400 405 410 415 420 425 430 435 4400

500

1000

1500

2000

2500

3000

3500



Am

plitu

de

340 345 350 355 360 365 370 375 380 385 3900

500

1000

1500

2000

2500

3000



Am

plitu

de

780 790 800 810 820 830 840 850 860 870 8800

2

4

6

8

10

12



Am

plitu

de

680 690 700 710 720 730 740 750 760 770 7800

2

4

6

8

10

12

14

16

18



Am

plitu

de

0 2000 4000 6000 8000 10000 12000 14000 16000 18000-10

0

10

20

30

40

50

60

70

80


x (2,0)

x-y (1,1)

y (0,2)

390 395 400 405 410 415 420 425 430 435 4400

500

1000

1500

2000

2500

3000

3500



Am

plitu

de

340 345 350 355 360 365 370 375 380 385 3900

500

1000

1500

2000

2500

3000



Am

plitu

de

780 790 800 810 820 830 840 850 860 870 8800

2

4

6

8

10

12



Am

plitu

de

680 690 700 710 720 730 740 750 760 770 7800

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2



Am

plitu

de

0 2000 4000 6000 8000 10000 12000 14000 16000 18000-20

0

20

40

60

80

100


x (2,0)

x-y (1,1)

y (0,2)

6x1013 ppp

5x1013 ppp

7x1013ppp

(x,x’) (y,y’)

(x,y) (εx,εy)

4. Half Integer Losses ~2Chris Warsop

Envelope Frequencies Incoherent Q’s EnvelopesHorizontal Vertical

Turn 100

ORBIT 2D Simulation Results

- Repeat similar simulations, but driven by representative 2Qh=8 & 2Qv=7 terms

- If allow for BF and energy is compatible with loss observation on ISIS

4. Half Integer Losses ~3Chris Warsop

Driven both planes 2Qh=8 & 2Qv=7

0 1 2 3 4 5 6 7 8 9 107

7.5

8

8.5

9

Env

Osc

/Tur

n

Coherent Envelope Frequency

0 1 2 3 4 5 6 7 8 9 103

3.5

4

4.5

Inco

h O

sc/T

urn

Minimum Incoherent Frequency (Peak Q Depression)

0 1 2 3 4 5 6 7 8 9 1060

80

100

120

140

Circulating Protons (x1E13)

RM

S E

mitt

ance

(pi

mm

mr) RMS Emittance on Turn 100

Horizontal

Vertical

Coherent Limit

Incoherent Limit

Emittance Growth

Questions important for real machines …

• What causes εrms growth?

Mis-match, non stationary distributions,

driving terms from lattice, … ?

• Can we minimise it?

• Do codes give good predictions?

- can they predict emittance growth & loss?

Have compared ORBIT with theory

- to see if behaviour follows models

Study of Halo & Future Work

Vertical (YN, YN')

Simulation

Theory [*]

Incr

easi

ng Inte

nsi

ty

7.00 x1013 ppp 7.25 x1013 ppp

8.00 x1013 ppp 7.50 x1013 ppp

8.50 x1013 ppp 7.75 x1013 ppp

4. Half Integer Losses ~ 4

• Comparison of halo structure with theory

- ORBIT: Poincare routines: AG ISIS Lattice; RMS Matched WB; quad driving term; large tune split;

- Theoretical model: Smooth, RMS equivalent KV,quad driving term; “small tune split” (equal) [*Venturini & Gluckstern PRST-AB V3 p034203,2000]

- Main features agree …

Chris Warsop

• Next

- Check number of particles migrating into halo …?

- Introduce momentum spread (then extend to 3D)

- Comparison with ISIS in Storage ring mode~ trials now underway

Normalised vertical phase space

Developing a space charge code “Set"

(1) Model and Study Rectangular Vacuum Vessels in ISIS

- implement the appropriate field solvers

- study image effects: rectangular vs elliptical geometry

(2) Develop our own code

- allow us to understand operation and limitations

- develop and enhance areas of particular interest

- presently 2D: will extend …

- plus use of ORBIT, SIMBAD, TRACKnD, etc

5. Images and Set Code ~ 1 Ben Pine

View inside ISIS vacuum vessels

Field Solver Benchmarking: Set solver vs CST Studio

5. Images and Set Code ~ 2

(xc,yc)=(0,0)

(xc,yc)=(5,5)

(xc,yc)=(15,0)

Set solver and CST agree to <0.1%

Ben Pine

ρ(x,y) Ф(x,y) Relative Error Ф(x,y)

Comparisons of Set with ORBIT

5. Images and Set Code ~ 3 Ben Pine

(x,x’) (y,y’)

(x,y)

(x,x’) (y,y’)

(x,y)

Incoherent Tune Shifts• ISIS half integer resonance (as above)

- ~ RMS matched WB beam, 2Qv=7 term etc

- Track for 100 turns; vary intensity

• Good Agreement - where expected

- Incoherent tunes, envelope frequencies

- evolution of εrms, beam distributions

ORBIT Set

Distributions on turn 100

ORBIT Set

Set: Dipole Tune Shift and Next Steps

5. Images and Set Code ~ 4Ben Pine

Coherent tune shifts from Set

• Next Steps

- Are now modelling closed orbits with images

- See expected variations in orbit with intensity

- evidence of non linear driving terms …

- planning experiments to probe images …

• Coherent Dipole Tune Shift in Set

- Expect some differences between ORBIT & Set

- ORBIT - just direct space charge (as we used it)

- Set - images give coherent tune shift

• Making good progress in key areas

- experimental study (collaboration on diagnostics)

- machine modelling and bench marking

- code development and study of loss mechanisms

• Topics covered

- Current priorities: Space charge and related loss, injection.

- Next: Instabilities, e-p, …

• Essential for ISIS upgrades

• Comments and suggestions welcome!

Summary

Acknowledgements:

ASTeC/IB - S J Brooks, C R Prior ~ useful discussions

STFC e-Science Group ~ code development

ORNL/SNS ~ for the use of ORBIT

Space Charge and High Intensity Studies on ISIS

Documents

profile monitor studies

beamspace charge field

space charge study

work of d j adams

isis accelerator diagnostics

high intensity studies

isis operations

rob williamson