Optical Diagnostics of High-Optical Diagnostics of High-Brightness Electron BeamsBrightness Electron Beams

Victor A. VerzilovVictor A. Verzilov

Synchrotrone TriesteSynchrotrone Trieste

ICFA AABD Workshop, Chia Laguna, Sardenia

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

Introduction (continue)

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

For high-brightness beams control of following parameters is essential

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)

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

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

Transverse planeOTR vs inorganic scintillators at a glance

OTR instantaneous emission linearity (no saturation effects) high resolution surface effect: thickness doesn’t

matter small perturbation to the beam

(small thickness) small radiation background (small

thickness) can be used in a wide range of

relatively low photon yield (limitation in pepper-pot measurements)

Scintillators (YAG:Ce, YAP:Ce, oth.)

high sensitivity no grain structure time response ~ 100ns conformance to HV radiation resistance

bulk effect

c/~

TR spatial resolution

J1 x x

2

xF 2

100

FWHM resolution is 2-3 times of the classical PSF

scales as ~ tails problem; mask can help high-resolution is experimentally confirmed

[CEBAF(4 GeV) SLAC (30 GeV)]

F x K1

x

J0 x x

1/, M 1

OTR resolution is determined by the angular acceptance

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.

Confirmed at ANL 220 MeV @ 0.8 nC30-40% discrepancy

Q=0.5nC

Instantaneous heating. TR case

N.Golubeva, V.Balandin TTF

dm

dE

cT

p

1

Temperature limits Si

Melting - 1683 ° Thermal stress –

1200° Al

Melting - 933 ° Thermal stress –

140-400°

Si: 1GeV @ 300um. For Al values ten times smaller

Heating by a bunch train

20um

50um

9MHz

1MHz

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@9MHZ Si @ 20 um1nC 1nC

90° Thompson scattering W.P.Leemans et al. LBNL

220

2

1

2

Noninvasive

Both transverse and longitudinal profiles

Synchronization Powerful laser Limited applicability

e-beam: 50 MeV@1.5nClaser: 50mJ@0.8m; 50-200fsphotons:30keV@105 ph/bunch

66m FWHM

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

DR intensity ~ e-a/and is strongly suppressed at wavelengths <a/

TR vs DR from a slit

Transition radiation

Diffraction radiation

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

Emittance measurement. Multislit vs quadscan

S.G.Anderson et all PRSTAB 5,014201(2002)

20

2

2 nI

IR

Measure ofspaces-chargedominance

Pepper-pot (multislit) Quadscan 3 screens

Widely used techniques

drift

High-brightness beam at “low energy”

Space-charge forces

LLNL 5MeV@50-300pC

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

Radiation from a bunch

),()()( TL FFF

spsptot IFNNINI )()1(

N

k

N

kj

rrnci jkeNN

F

/

1

1

All particles in a bunchare assumed identical. Noangular and energy spread.

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

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

F=0

0 <F< 1

F=1

Form-factor and bunch shape

F dz

z e i / c z2

z 1

cd

0

F cos z

c

For the normalized longitudinal distribution of particles in the bunch (z)

By inverse Fourier transform

)/sin()/()( TTLL FFF

1,)()( LFF

Symmetric bunch

Transverse coherence comes first. Unless the beam is microbunched.

1)0( xF

Bunch shape and form-factor

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.

Form-factors

Kramers-Kronig analysis

m 2

dxln f x / f

x 2 20

izci efezdz

/

0

F * f 2

c

zfd

cz m

cos1

0

If F() 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 transformReal part is the observable

R.Lai and A.J.Sievers NIM A397

Kramers-Kronig analysis.Experiment

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!

TESLA TDR

Polychromator

Single-shot capable Narrow bandwidth Discreteness

T.Watanabe et al. NIM A480(2002)315 Tokio University

1.6ps900fs

Results are consistent withstreak camera and interferometermeasurements

Hilbert -Transform spectrometer

pst )2.02.1(

M.Getz et al., EPAC98 TTF

020

2

222

4

2

s

sc dS

I

IReI

Josephson junction

Wide bandwidth More R&D is necessary

T= 4-78Kf= 100-1000GHz

I E t E t / c

2

dt

I I

cosc

d

Coupled to a frequency domain.

Fourier spectroscopy

Measurement in the time domainis a measurement of the autocorrelationof the radiation pulse.

Precise Established Time consuming

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.

Analysis in the time domain (TR case)

2)/(1)( ceg

2/,0

2/,/1)(

)()(22 2/

z

zzu

dyeyzuz y

2)/(22

2)/(

)1()2/(sin

)(2

2

cc

ece

I

A.Murokh,J.B.Rosenzweig et al

Filter function

Model bunch shape

Coherent spectrum

dc

sIsI

cos

Autocorrelation curve

TR. Finite-size screen

r

screen

The effect comes into play when the screen size is comparable or smaller than

The TR spectrum from a finite size target is a complex function of the beam energy, target extensions, frequency and angle of emission. 20,

inf, sin1 rkJII

r

r=20 mmd=0.05 rad

1mm2mm

Coherent diffraction radiation

Bunch length was measuredfor slit widths 0 to 10 mm.

Effect of the target finite size was proved.

M.Castellano et al. PRE 63, 056501 TTF

Coherent diffraction radiation.Result

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.

M.Castellano et al. PRE 63, 056501 TTF

225MeV @ 1nC

Electro-optic sampling (EOS)

Noninvasive Fast response ~40 THz Linearity&dynamic range Jitter dependent

Modulation of the polarizationof light traveling through a crystalis proportional to the applied electric field

El )/(

i irR

qE

2)(

Collective Coulomb fieldat R is nearly transverse

EOS Single-shot option

Single shot On-line Nearly jitter-free

Make use of a long pulse with a linearfrequency chirp

Bunch time profileis linearly encoded onto the wavelength spectrum

EOS Single-shot option.First prove

Resolution ≈ Chirp Pulse width ~300fs

~70 fs achievable

( )

0

0

fsps 5,1 0

1.72 ps

I.Wilke et al., PRL, v.88, is.2,2002 FELIX

e-beam: 46MeV@200pC0.5x4x4mm3 ZnTe crystallaser: 30 fs@800nm,chirp up to 20ps

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