Holger Schlarb - indico.desy.de

Holger Schlarb for the optical synchronization team

Femtosecond Synchronization

Motivation

Synchronizationsystem

Komponenten des optischen Synchronisationssystem

bei FLASH

Femtosecond synchronization

2

24.03.2010, Groemitz, “Beschleuniger Seminar”

Holger Schlarb, MSK, DESY

31km

Motivation: shorter electron bunch length

Towards shorter electron pulse duration 1 ps = 10-12 s

= 1000 fs

= 300 m/c0

Synchrotron light sources …BESSY z = 10mm ~ 20 ps

Low-alpha Multi Bunch Hybrid Modus ~ 2 ps

Linac driven colliders … SLC z = 2 mm ~ 6 ps

ILC =0.3 mm ~ 1 ps

Free Electron Lasers … European XFEL z = 20 um ~ 60 fs

Short pulse operation modes: FEL pulses < 5 fs

Circ = 240 m

t

t

31km~ 10-3


3



Motivation: Pump-probe experiments

Classical setup:

Variable delay

pump

probe

Probe = flash

Pump

Shot pulses fs ps

Knowledge of time delay is crucial!

Same source

Pump pulse initiate reaction, probe pulse records current state.

Atomic / Molecular Physics// Solid state dynamics

FELs: two different pulse sources: FEL beam and optical laser

Comments: 1) measured at experiment => post analysis

2) for small event rates => not possible


4



Motivation: seeding with higher harmonics

Ti:Sa Laser

Gas cell 800nm30nm

undulatorSeparation

100kW>1 GW

Experiment

Electron beam

Temporal overlap

New class of experiments:

• Electron beam is seeded or manipulated using external laser

• Timing of FEL pulse defined by now by seed (laser)

Requirements:

• Temporal overlap between electron bunch & seed pulse essential ~ 10-50 fs

• Particular high demands on synchronization to pump-probe laser ~ 1-5 fs

Gain 104

Generic layout of seeding a single pass FEL

Transfer line ~ 30-100m, << 10 fsTi:Sa Laser

t

t


5



Motivation: Peak current & arrival time jitter

RF

gun

photo cathode laser

booster

magnetic

chicaneacc. modules undulator

target

pump-probe laser

Compression process & FEL resonance condition

• stability of peak current I/I < 1-10 %

• control of arrival t < 10-50 fs

• FEL wavelength fluctuations E/E < 10-4

Increased acceleration RF stability requirements

Phase stability (rel.) ~ 0.05 …0.005 (~ H … ~ 1 MHz)

Amplitude stability ~ 0.05% …0.005% (~ H … ~ 1 MHz)

seed laser

z

E

z

E

z

E

z

E

Acc-RF

Ipeak ~ 10 A Ipeak ~ 1000 A


6



Motivation: Lasers for FELs

Generic layout of single pass FELs

injector pre-linac main linacchicanechicane

Ph

oto

-ca

thod

e

La

ser

hea

ter

EO

E-S

AS

E

Op

tica

l re

pli

ca EO

See

dS

eed

Few

cy

cle

lase

r

Pu

mp

-pro

be

Pla

sma

la

ser

Ma

nip

ula

tio

n

Ma

nip

ula

tio

n

RF RF

Master Laser

Oscillator

>> 100 m but < 10fs


7



Motivation: summary

Accelerator technology advanced from

picosecond bunch duration few 10 fs bunches

Facility length range from ~100 m up to 3.5 km (European XFEL)

Stability of FELs puts stringent demands on RF requirements

Optical laser systems become an integral part of FEL facilities for the beam generation,

beam diagnostics,

beam conditioning/seeding and

for user experiments

Changed the requirements on synchronization from ps to fs precision


8



Synchronization

Several scheme on the market!

1) RF distribution

~f ~ 100MHz …GHz

3) Carrier is optically + detection

~f ~ 200 THz

2) Carrier is optically

~MZT

f ~ GHz

3) Pulsed sourcef ~ 5 THz

OXC

Mode locked

Laser

t ft f=


9



Optical synchronization system

Narrow

Band.

Direct/

Interferometer

EOMs/

SeedingTwo color bal.

Opt. cross-corr.

End-station

Laser

MLOMO-RF

Laser pulse Arrival beam/laser DWC/Kly

A & cavity

Desired point-to-point stability ~ 10 fs

<5fs

Direct

Distribution

LO-RF

Optical link Optical link

<5fs <5fs

Optical link

FB

EDFL, t~100-200 fs,

f=50-250MHz, P > 100mW,

phase noise < 10fs ( 1kHz)

Free space + EDFA /

HP-EDFA + fiber dist

OXC / RF

Polarization contr.

Dispersion comp.

Other lasers

Main issue: robustness, stability and maintainability Prototype at FLASH


10



FLASH synchronization system (Sep. 2009)

FLASH accelerator and optical end-stations for synchronization:

Master laser oscillator (EDFA) with 4 stabilized optical links to

• 3 Beam arrival monitors (BAM)

• 1 cross-correlator for locking diagnostic TiSa laser system (OXC)

• and large aperture BPM for energy measurements (EBPM)

Courtesy: S. Schulz


11



Master laser oscillator (MLO)

Specification/Requirements

- Mode-locked erbium-doped fiber laser

- Repetition rate of 216.66MHz

- Average power > 100mW

- Pulse duration < 100 fs (FWHM)

- Integrated timing jitter < 10 fs [1kHz,40MHz]

- Amplitude noise < 2 10-4 [10Hz,40MHz]

- Mechanical robust/easy to maintain

1st generation MLO 2nd generation MLO 3st generation MLO

Original design:

J. Chen et. al., Opt. Lett. 32,

1566-1568 (2007)


12



RF Lock


Courtesy: S. Schulz

Carrier = 1.3 GHz

Performance:

Integrated timing jitter 11.5 fs @ [1kHz, 10MHz] / 4fs @ [10kHz,10MHz]

Very short pulses < 90 fs

Power stability: 2% pkpk and 0.4% rms over 240 hours! (no FB stabilization)

Problems

Output power < 70mW otherwise double pulses

Reproducibility and maintance


13




Pulse duration (sech2(t/ p)-fit)

Courtesy: S. Schulz

Output stability

Performance:

Integrated timing jitter 11.5 fs @ [1kHz, 10MHz] / 4fs @ [10kHz,10MHz]

Very short pulses < 90 fs

Power stability: 2% pkpk and 0.4% rms over 240 hours! (no FB stabilization)

Problems

Output power < 70mW otherwise double pulses

Reproducibility (spectrum/pulse duration/phase noise) & maintenance

Requires special diagnostics and complex exception handling


14




0 5 10 15 20 25 30 35 40 4599

99.5

100

100.5

Long-Term Amplitude (Menlo M-Comb 2 x ET3010)

Time [h]

Rel. P

ow

er [%

]

0 5 10 15 20 25 30 35 40 4560

70

80

90

100

Time [h]

Pie

zo [V]

0 5 10 15 20 25 30 35 40 4521

21.5

22

22.5

23

Time [h]

Tem

pera

ture

[°C

]

Commercial lasers

Based on same technologyAfter some trouble shooting ok

High output power > 100 mW

Similar problems in terms of reproducibility

Long term measurements: observed instabilities

New laser based on different mode lockingType: saturable absorber laser (SESAM)

Output power > 100mW

Pulse duration p < 150 fs

Robust mechanics, high reproducibility

Integrated timing jitter 4.3 fs [1kHz,10MHz]

Inspected at tested at PSI

Changed specification to our purpose

Delivery April 2010

Problem: long term live time

Instabilities


15



Distribution system

Courtesy: S. Schulz

120 mW input

5 mW out

90% incoupling eff.


16



Distribution system


17



Fiber Link Stabilization

216 MHz

Er-doped

fiber laser

J. Kim et al., Opt. Lett. 32, 1044-1046 (2007)


18




Single crystal background free optical cross-correlator

Detector 1

Detector 2


19





Detector 2

Detector 1


20





Detector 1

Detector 2

Link is locked at zero crossing

Amplitude fluctuations are converted to gain fluctuations


21



• Three layer design:-Free space optics with OXC (top)

-Fiber layer (middle)

-Electronic layer (bottom)

• 5 links completely assembled

• mechanically very robust

• several problems identified-Yawing/pitching of delay line

-Some optics holders

-Easy to assembly

• 2nd prototype design ready for prod.

Drawback link stabilization:- requires dispersion compensation

- precise overlap of laser pulses

- to cost intensive to supply many end-stations

Room for further improvements

(packaging/costs)

DESY design

Fiber Link Stabilization: engineered version


22





23





24



Calibration of detector signal Still varies from link to link (learning curve …)

Courtesy: S. Schulz

Fiber Link Stabilization: cross-correlator

Varies likely due to dispersion compensation & optical power achieved in link

Calibration is used to determine remaining in-loop error


25



Calibration of detector signal Still varies from link to link (learning curve …)

Courtesy: S. Schulz

Fiber Link Stabilization: cross-correlator

Out-of-loop jitter cannot be determined easily

LINK15 uses different link end configuration


26



link 1

link 2

8 days

link 1

link 2

2 month

Installation of two fiber links at FLASH Much smoother correction

For 6 month in operation

Day/night periods & spring -> summer observable

Smooth operation (interruptions understood)

Courtesy: F. Löhl

Fiber Link Stabilization: long term behavior


27



Principle of the bunch arrival time monitor (BAM)sampling times of ADCs

4.7 ns

(216 MHz)The timing information of the electron bunch is

transferred into a laser amplitude modulation.

This modulation is measured with a photo

detector and sampled by a fast ADC.

ADC1

ADC2

See: EPAC’06, Loehl et al., p.2781

Courtesy: F. Loehl

17mm14.5mm 6.2mm

1.2mm thick

Alumina disk

New pickup design &

Improved readout

Courtesy: K. Hacker

resolution < 10 fs

Bunch arrival time monitor


28




Bunch arrival time monitor (BAM)

• Coarse and fine measurements

• 10GHz Mach-Zehnder interferometer

• Precision encoder + high duty cycle delay stages

• Controls: Bias, Temp. readout & regulation

• New engineered design (2nd iteration)

Performance & findings:

• Typical resolution: 6-10 fs, (down to 4 fs) @ 1nC

• Stable operation after establishing FB on coarse

• Assembly + slicing ~ 4 weeks

• Problematic fiber management (too long fibers)

• hardware + firmware + automation software


29




Optical front-end in accelerator tunnel

Temperature stabilized (40mK pkpk)

High duty cycle optical delay line

10GHz Mach-Zehnder modulators

Coarse/fine channel

Remote controlled (Bias/Temp/Motors)

Heidnhain encoder


30




Scan of laser phase

Fine channel, dynamic range ~ 4 ps,

intrinsic resolution ~ 4-10fs

Coarse channel, dynamic range ~ 60 ps,

intrinsic resolution ~ 60-100fs

Fluctuations within pulse train

Finding the signal


31



uncorrelated jitter

over 4300 shots:

8.4 fs (rms)

<6fs rms resolution

BC3BC2 ACC 23

Diag.

ACC 456ACC 1

Diag.

CDR

LOLAISR

BPM

BAMBAM 1

ORSUndulators

PP-laser

ISR

BAM 2

60 m drift

• Complete test: includes timing jitter of MLO & 2*LINK & 2*BAM

• Remark: old configuration (no distribution & bread board type LINK)

Courtesy: F. Löhl

Bunch arrival-time monitor (BAM)

Some measurements last user run

Arrival time drift & jitter from photo-injector

0.8ps

t = 86fs

Gun rf

30%

Laser rf

70%

Tuning

pulse train

arrival

& exit of linac

t = 82fs

1.1ps

4.0ps

drifts

jitter

intra-train

300 s

t = 64fs

residual

Without FB!Without FB! Without AFF!Without AFF!

For FB missing: robust exception handlingFor FB missing: robust exception handling


33



Infrastructure for FLASH optical synchronization 300 cables to optical table (including 120 signal cables)

58 motors

20. . . 24 laser diodes for MLOs and EDFAs

16. . . 20 piezo stretchers

4 VME crates (change to uTCA in planning)

optical fibers to all critical end-stations

42 temperature sensors, 2 active temperature

regulations

air conditioning and option for water cooling

. . . and almost everything assembled in-house!

Status: ~ 95% done

In house electronics developments:

VME laser diode driver (FPGA)

VME ADCs 130MHz, 16bit

LN Piezo driver 300V

DSP regulation (old)

ADC (9MHz, 14bit)

DAC (9MHz, 14bit)

Vector modulator

Status: exist or in productions


34



Thanks for your attention

Documents

Holger Schlarb - indico.desy.de