Optical Diagnostics of HighBrightness Electron Beams - PowerPoint PPT Presentation

About This Presentation
Title:

Optical Diagnostics of HighBrightness Electron Beams

Description:

high resolution. surface effect: thickness doesn't matter ... high-resolution is experimentally confirmed [CEBAF(4 GeV) SLAC (30 GeV) ... – PowerPoint PPT presentation

Number of Views:30
Avg rating:3.0/5.0
Slides: 37
Provided by: Vic499
Category:

less

Transcript and Presenter's Notes

Title: Optical Diagnostics of HighBrightness Electron Beams


1
Optical Diagnostics of High-Brightness Electron
Beams
ICFA AABD Workshop, Chia Laguna, Sardenia
  • Victor A. Verzilov
  • Synchrotrone Trieste

2
Introduction
  • ID of a high-brightness beam
  • high charge per bunch (1 nC and more)
  • small transverse and longitudinal beam dimensions
  • extremely small normalized emittances
  • high peak current
  • space-charge effects in the beam dynamics
  • Two missions of beam diagnostics
  • Provide instruments for study of the physics
  • Assist in delivering high quality beams for
    applications

Every machine is as good as its diagnostics
3
Introduction (continue)
For high-brightness beams control of following
parameters is essential
  • Vertical and horizontal emittances
  • Transverse beam profile
  • Beam trajectory
  • Energy and energy spread
  • Bunch length
  • Longitudinal bunch shape
  • Charge per bunch
  • Current (peak and average)
  • Bunch-to-bunch jitter

Some of the parameters are measured by
traditional methods, others require specific
techniques and instrumentations
4
Specific requirements
  • Take into account space charge forces
  • Resolution from several millimeters to few tens
    of micrometers in both longitudinal and
    transverse plane
  • Large dynamic range both in terms of beam
    intensity and measuring interval
  • Non-invasive
  • Single-shot
  • Real time
  • Jitter-free and synchronized
  • Usual (stability, reliability ,etc)

5
Optical diagnostics and others
  • Optical diagnostics are based on analysis of
    photons generated by a beam in related processes
    or make use of other optical methods (lasers,
    etc.)
  • This talk reports the current status of optical
    diagnostics of high-brightness beams
  • Reasons
  • significant progress
  • make an essential part of available tools
  • impossible to cover everything
  • Other techniques
  • wire scanners
  • zero phasing
  • transverse rf deflection cavity
  • high-order BPM

6
Outline
  • Transverse and longitudinal profile measurements
    give the largest amount of information about beam
    parameters
  • Transverse plane
  • Spatial resolution is a key issue
  • Survival problem for intercepting monitors
  • Non-invasive methods
  • Emittance measurement issues
  • Longitudinal plane
  • Coherent radiation is a primary tool
  • Direct spectral measurements
  • Fourier transform
  • CDR vs CTR
  • Electro-optical sampling

7
Transverse planeOTR vs inorganic scintillators
at a glance
  • Scintillators (YAGCe, YAPCe, oth.)
  • high sensitivity
  • no grain structure
  • time response 100ns
  • conformance to HV
  • radiation resistance
  • bulk effect
  • OTR
  • instantaneous emission
  • linearity (no saturation effects)
  • high resolution
  • surface effect thickness doesnt matter
  • small perturbation to the beam (small thickness)
  • small radiation background (small thickness)
  • can be used in a wide range of g?????
  • relatively low photon yield (limitation in
    pepper-pot measurements)

8
TR spatial resolution
OTR resolution is determined by the angular
acceptance
  • FWHM resolution is 2-3 times of the classical PSF
  • scales as l/q
  • tails problem mask can help
  • high-resolution is experimentally confirmed
    CEBAF(4 GeV) SLAC (30 GeV)

9
Scintillator resolution
A.Murokh et al. BNL-ATF
Recent experiment at BNL expressed concerns about
micrometer-level resolution. Strong discrepancy
in the beam size compared to OTR and wire scans
was observed.
Q0.5nC
Confirmed at ANL 220 MeV _at_ 0.8 nC 30-40
discrepancy
10
Instantaneous heating. TR case
N.Golubeva, V.Balandin TTF
Si 1GeV _at_ 300um. For Al values ten times
smaller
  • Temperature limits
  • Si
  • Melting - 1683
  • Thermal stress 1200
  • Al
  • Melting - 933
  • Thermal stress 140-400

11
Heating by a bunch train
N.Golubeva, V.Balandin TTF
  • Two cooling processes contribute to the
    temperature balance
  • Radiation cooling temperature to the power of 4
  • Heat conduction depends on the thermal
    conductivity and temperature gradient

Si_at_9MHZ
Si _at_ 20 um
9MHz
1nC
20um
1nC
1MHz
50um
12
90 Thompson scattering
W.P.Leemans et al. LBNL
66?m FWHM
  • Noninvasive
  • Both transverse and longitudinal profiles
  • Synchronization
  • Powerful laser
  • Limited applicability

e-beam 50 MeV_at_1.5nC laser 50mJ_at_0.8?m
50-200fs photons30keV_at_105 ph/bunch
13
Diffraction radiation
  • Diffraction radiation is emitted when a particle
    passes in the proximity of optical
    discontinuities (apertures )
  • DR characteristics depend on the ratio of the
    aperture size to the parameter lg
  • DR intensity e-a/lg?and is strongly suppressed
    at wavelengths llta/g

14
TR vs DR from a slit
Transition radiation
Diffraction radiation
15
Effect of the beam size
  • Angular distribution depends on the relative
    particle position with respect to the aperture
    and can be used to measure the beam size
  • Strong limitation is a low intensity in visible
    and near infra-red
  • Energy and angular spread, detector bandwidth are
    interfering factors
  • Still has to be proven experimentally

A.Cianchi PhD Thesis
16
Emittance measurement. Multislit vs quadscan
S.G.Anderson et all PRSTAB 5,014201(2002)
High-brightness beam at low energy
Widely used techniques
  • Pepper-pot (multislit)
  • Quadscan
  • 3 screens

Space-charge forces
drift
Measure of spaces-charge dominance
LLNL 5MeV_at_50-300pC
17
Longitudinal plane
  • Small longitudinal bunches are crucial for many
    applications
  • Bunch lengths are on a sub-ps time scale
  • Conventional methods often do not work
  • Several new techniques have been developed
  • Coherent radiation has become a primary tool to
    measure the bunch length and its shape in the
    longitudinal plane
  • It is very powerful tool with nearly unlimited
    potential towards ever shorter bunches

18
Radiation from a bunch
All particles in a bunch are assumed identical.
No angular and energy spread.
19
Radiation zoo
  • Any kind of radiation can be coherent and
    potentially valuable for beam diagnostics
  • Transition radiation
  • Diffraction radiation
  • Synchrotron radiation
  • Undulator radiation
  • Smith-Parcell radiation
  • Cherenkov radiation
  • Nevertheless, TR is mostly common
  • Simple
  • Flat spectrum

20
Bunch form-factor and coherence
  • wavelength is much shorter than bunch dimensions
  • radiation is fully incoherent
  • particles emit independently
  • total intensity is proportional to N
  • wavelength is of the order of bunch dimensions
  • radiation is partially coherent
  • some particles emit in phase
  • increase in total intensity
  • wavelength is much longer than bunch dimensions
  • radiation is fully coherent
  • all particles emit in phase
  • total intensity is proportional to N2

F0
0 ltFlt 1
F1
21
Form-factor and bunch shape
Transverse coherence comes first. Unless the beam
is microbunched.
For the normalized longitudinal distribution of
particles in the bunch r(z)
By inverse Fourier transform
Symmetric bunch
22
Bunch shape and form-factor
Form-factors
Bunch shapes with the same rms bunch lengths
  • Although, in principle, the bunch shape can be
    retrieved from a measurement, be care, this could
    be ambiguously.
  • The bunch size, however, is recovered reliably.

23
Kramers-Kronig analysis
R.Lai and A.J.Sievers NIM A397
If F(w) is determined over the entire frequency
interval, the Kramers-Kronig relation can be used
to find the phase.
Both real and imaginary part of the form-factor
amplitude are to be known to recover the
asymmetry of the bunch shape.
By inverse Fourier transform
Real part is the observable
24
Kramers-Kronig analysis.Experiment
TESLA TDR
  • Spectral intensity has to be defined over a
    significant spectral range.
  • Errors are produced when asymptotic limit are
    attached to the data to complete the spectral
    range.
  • Front-tail uncertainty.
  • Analytical properties of the bunch shape function
    have to be taken into account.

Confirmed by recent SASE results!
25
Polychromator
T.Watanabe et al. NIM A480(2002)315 Tokio
University
1.6ps
900fs
  • Single-shot capable
  • Narrow bandwidth
  • Discreteness

Results are consistent with streak camera and
interferometer measurements
26
Hilbert -Transform spectrometer
M.Getz et al., EPAC98 TTF
Josephson junction
T 4-78K f 100-1000GHz
  • Wide bandwidth
  • More RD is necessary

27
Fourier spectroscopy
Coupled to a frequency domain.
Measurement in the time domain is a measurement
of the autocorrelation of the radiation pulse.
  • Precise
  • Established
  • Time consuming

28
Low-frequency cut-off
  • All experimental data suffer to a different
    extent from the low frequency cut-off.
  • There is a number of reasons which cause the
    cut-off detector band, EM waves transmittance,
    target size etc.
  • Data analysis usually consists in assuming a
    certain bunch shape and varying the size
    parameter for the best fit to undisturbed data.

29
Analysis in the time domain (TR case)
A.Murokh,J.B.Rosenzweig et al
Filter function
Model bunch shape
Autocorrelation curve
Coherent spectrum
30
TR. Finite-size screen
The effect comes into play when the screen size
is comparable or smaller than lg?
r20 mm qd0.05 rad
1mm
2mm
  • The TR spectrum from a finite size target is a
    complex function of the beam energy, target
    extensions, frequency and angle of emission.

31
Coherent diffraction radiation
M.Castellano et al. PRE 63, 056501 TTF
Bunch length was measured for slit widths 0 to 10
mm.
Effect of the target finite size was proved.
32
Coherent diffraction radiation.Result
M.Castellano et al. PRE 63, 056501 TTF
  • DR and TR results are consistent in a wide range
    of slit widths .
  • CDR can be successfully used for bunch length
    measurements.
  • Very promising for ultra-high power beams,
    because non-invasive.

225MeV _at_ 1nC
33
Electro-optic sampling (EOS)
Modulation of the polarization of light traveling
through a crystal is proportional to the applied
electric field
Collective Coulomb field at R is nearly transverse
  • Noninvasive
  • Fast response 40 THz
  • Linearitydynamic range
  • Jitter dependent

34
EOS Single-shot option
Make use of a long pulse with a linear frequency
chirp
Bunch time profile is linearly encoded onto the
wavelength spectrum
  • Single shot
  • On-line
  • Nearly jitter-free

35
EOS Single-shot option.First prove
I.Wilke et al., PRL, v.88, is.2,2002 FELIX
e-beam 46MeV_at_200pC 0.5x4x4mm3 ZnTe
crystal laser 30 fs_at_800nm,chirp up to 20ps
  • Resolution
  • Chirp
  • Pulse width
  • 300fs
  • 70 fs achievable
  • ( )

1.72 ps
36
Conclusions
  • Beam diagnostics has significantly advanced to
    meet specific requirements of high-brightness
    beams
  • Wide choice of available techniques from which
    one can select
  • Lack of suitable (simple and reliable)
    non-invasive methods for measurements in the
    transverse plane (near-future projects)
  • In the longitudinal plane CDR is likely OK
  • Difficulties with measurements at µm and sub-µm
    level in the transverse plane
Write a Comment
User Comments (0)
About PowerShow.com