Plasma switch as a temporal overlap tool for pump-probe experiments at FEL facilities

This content has been downloaded from IOPscience. Please scroll down to see the full text.

Download details:

IP Address: 183.63.81.66

This content was downloaded on 10/10/2013 at 23:09

Please note that terms and conditions apply.

Plasma switch as a temporal overlap tool for pump-probe experiments at FEL facilities

View the table of contents for this issue, or go to the journal homepage for more

2012 JINST 7 P08007

(http://iopscience.iop.org/1748-0221/7/08/P08007)

Home Search Collections Journals About Contact us My IOPscience

2012 JINST 7 P08007

PUBLISHED BY IOP PUBLISHING FOR SISSA MEDIALAB

RECEIVED: June 22, 2012ACCEPTED: July 10, 2012

PUBLISHED: August 10, 2012

Plasma switch as a temporal overlap tool forpump-probe experiments at FEL facilities

M. Harmand,a,1 C.D. Murphy,b C.R.D. Brown,b,l M. Cammarata,c T. Doppner,d

S. Dusterer,a D. Fritz,c E. Forster,e, j E. Galtier,i J. Gaudin, f S.H. Glenzer,d S. Gode,g

G. Gregori,b V. Hilbert,e D. Hochhaus,h T. Laarmann,a H.J. Lee,c H. Lemke,c

K.-H. Meiwes-Broer,g A. Moinard,i P. Neumayer,h,k A. Przystawik,a H. Redlin,a

M. Schulz,a S. Skruszewicz,g F. Tavella,a T. Tschentscher, f T. White,b U. Zastraue

and S. Toleikisa

aDeutsches Elektronen-Synchrotron DESY,Notkestr. 85, 22607 Hamburg, Germany

bClarendon Laboratory, University of Oxford,Parks Road, Oxford OX1 3PU, U.K.

cLCLS, SLAC National Accelerator Laboratory,2575 Sand Hill Road, Menlo Park, CA 94025-7015, U.S.A.

dLawrence Livermore National Laboratory, University of California,P.O. Box 808, Livermore, CA 94551, U.S.A.

eIOQ, Friedrich-Schiller-Universitat,Max-Wien Platz 1, 07743 Jena, Germany

f European XFEL GmbH,Albert-Einstein-Ring 19, 22761 Hamburg, Germany

gInstitut fur Physik, Universitat Rostock,18051 Rostock, Germany

hGSI Helmholtzzentrum fur Schwerionenforschung GmbH,Planckstr. 1, 64291 Darmstadt, Germany

iLaboratoire pour l’Utilisation des Lasers Intenses, Ecole Polytechnique,Route de Saclay, 91128, Palaiseau Cedex, France

jHelmholtz-Institut Jena,Frobelstieg 3, 07743 Jena, Germany

kEMMI,Planckstr. 1, 64291 Darmstadt, Germany

lAWE Aldernaston,Reading RG7 4PR, U.K.

E-mail: [email protected]

1Corresponding author.

c© 2012 IOP Publishing Ltd and Sissa Medialab srl doi:10.1088/1748-0221/7/08/P08007

2012 JINST 7 P08007

ABSTRACT: We have developed an easy-to-use and reliable timing tool to determine the arrivaltime of an optical laser and a free electron laser (FEL) pulses within the jitter limitation. This tim-ing tool can be used from XUV to X-rays and exploits high FELs intensities. It uses a shadowgraphtechnique where we optically (at 800 nm) image a plasma created by an intense XUV or X-ray FELpulse on a transparent sample (glass slide) directly placed at the pump - probe sample position. Itis based on the physical principle that the optical properties of the material are drastically changedwhen its free electron density reaches the critical density. At this point the excited glass samplebecomes opaque to the optical laser pulse. The ultra-short and intense XUV or X-ray FEL pulse en-sures that a critical electron density can be reached via photoionization and subsequent collisionalionization within the XUV or X-ray FEL pulse duration or even faster. This technique allows todetermine the relative arrival time between the optical laser and the FEL pulses in only few singleshots with an accuracy mainly limited by the optical laser pulse duration and the jitter betweenthe FEL and the optical laser. Considering the major interest in pump-probe experiments at FELfacilities in general, such a femtosecond resolution timing tool is of utmost importance.

KEYWORDS: Timing detectors; Instrumentation for FEL; Beam-line instrumentation (beam posi-tion and profile monitors; beam-intensity monitors; bunch length monitors)

2012 JINST 7 P08007

Contents

1 Introduction 1

2 Experiment details 2

3 Results 33.1 Timing tool in the XUV range at FLASH 33.2 Timing tool in the X-ray range at LCLS 33.3 A jitter limited method 5

4 Conclusion 5

1 Introduction

As short wavelength free electron laser facilities develop and are applied to a growing range ofscientific studies, it becomes increasingly important to elaborate reliable and easy to use diagnos-tic tools. Characterization of the timing between two independent photon sources is a particularlychallenging aspect in the femtosecond time regime for pump-probe experiments. For such ex-periments, knowledge of when temporal overlap occurs (“time zero”) and the uncertainty in thearrival time between the two pulses (the jitter) directly at the interaction point of the experimentis of paramount importance. Several timing methods have already been studied and developed:e.g., reflectivity changes of FEL excited materials [1–5], generation of sidebands in photoelectronTOF spectra [6], transient molecular Nitrogen alignment of molecules [7], or phase transitions insolids [8, 9]. We report here a competitive in-situ method to determine the temporal overlap be-tween optical laser and FEL pulses within the jitter limitation (∼ 100fs), with only few single shotsand for a broad range a FEL photon energy.

The method presented here is a shadowgraph technique where the transmission change of theoptical laser pulse is imaged through a transparent sample, in our case a simple glass slide, whichis irradiated by an intense FEL pulse. The main idea is to transform rapidly the material from ahighly transparent state to an opaque one. This is achieved by quickly changing the free electrondensity in the sample. This process is commonly used for optical gates [10] and plasma mirrorsin laser physics [11]. In contrast to those applications, the change of the free electron density isaccomplished with XUV or X-ray photons from an FEL. Photoionization and subsequently colli-sional ionization increase the free electron density within femtosecond time scale [12, 13]. In thecase of high FEL intensity, one creates a plasma on the surface that exceeds the critical electrondensity Nc, where the plasma frequency equals the optical laser frequency. For an optical laser at awavelength of λOL = 800 nm, the critical free electron density is Nc = 1.7 ·1021cm−3.

In order to reach the critical electron density on the irradiated surface, the intensity of the X-ray pump pulse has to be sufficiently high to create a significant amount of free carriers. In the case

– 1 –

2012 JINST 7 P08007

OL @ 800nm (probe)

FEL (pump)

Glass Slide

Plasma (opaque @ 800nm

for N > Nc)

CCD

Plasma shadow

Lens

Figure 1. Experimental setup of the plasma timing tool.

of optical lasers, the ultrafast ionization mechanism is mainly based on multiphoton and tunnellingionization and ensure the feasibility of 10 fs gating of optical gates [10]. In the XUV to X-rayrange, the ionization process is mainly due to single photon photoionization followed by the emis-sion of less energetic Auger electron in light elements. Subsequently, collisional ionization alsooccurs via electron impact of the primary photoelectrons and secondary electrons are produced inthis dense environment. Due to the high brilliance of FEL pulses [14], this effect is very efficient,especially in the XUV range. For example at FLASH [15], typical XUV photon intensities around1013− 1014 W/cm2 are sufficient to get to the critical density within the FEL pulse duration (typ-ically ∼ 10 fs to ∼ 200 fs). However, there is an important difference between the XUV and theX-ray regime concerning the absorption: the attenuation length. For glass (SiO2), the attenuationlength at photon energy of 200 eV is only 0.12 µm, while at 8 keV it is about 130 µm. From thisfact one could expect that it might be more difficult to reach the critical density at higher photonenergies, because the photons are then absorbed in a larger volume. But at higher energies (8 keV)also the number of electron-hole pairs which are created is much higher than in the XUV rangewhich allows to get to the critical density within ∼ 100 fs or less [16].

2 Experiment details

During previous pump-probe experiments at FLASH [17], we have developed an experimentalprocedure to overlap spatially and temporally the FEL and the optical laser pulses at the point ofinterest for pump-probe experiment. A first temporal overlap with a ∼ 50 ps time resolution isperformed with an ultrafast photodiode placed at the sample position to adjust the electronic delayset on a masterclock of the optical laser beam on the FEL arrival time [18]. In a second step, theprecise arrival time of the optical laser is adjusted with a delay line and by using the describedshadowgraph based technique in the XUV regime to reach a synchronization limited by the jitterbetween the optical laser and the FEL pulses. In our case, the FEL intensity was around 1014−1015

W/cm2, ensuring a plasma effect on the surface [17]. A sketch of the experimental setup is shownin figure 1. We position a movable glass slide in the FEL focal plane, which is backlit by a lowintensity optical laser which does not damage the glass substrate. A single FEL pulse generates aplasma on the glass surface by ultrafast ionization. If the plasma is present when the 800 nm laserpulse reaches the glass, the laser is reflected at the critical density position of the plasma and less

– 2 –

2012 JINST 7 P08007

T=+0.5psT=‐0.5ps

T=‐38ps T=‐40psT=+0.5psFELfirst

T=‐0.5psOp5callaserfirst

Figure 2. Single shot background subtracted images obtained with a single lens imaging system at FLASH.The two images correspond to two different delay steps adjusted with an optical delay line (negative delay:optical laser before the FEL), indicating that the temporal overlap position is between -0.5 ps (left) and +0.5ps (right). The shadow is due to a plasma created by the intense FEL radiation focused on the glass slidesurface.

light is transmitted through the area where the free electron density is above the critical density.This “shadow” is then observed in transmission by imaging the glass surface onto a CCD camerausing a single lens. A synchronized camera is acquiring the magnified images on a single shot basis.

If the laser pulse arrives at the same time or later than the intense FEL pulse, either the shadowfrom the critical surface, or the permanent damage at very long delays, will be visible in trans-mission. If the laser pulse arrives before the intense FEL pulse, no changes will be observed onthe transmission during the shot, while a permanent damage is observed afterwards showing theinteraction of the intense FEL pulse with the sample occurs after the optical laser arrival time. Byscanning iteratively the delay from one situation to another while observing the transmitted signalduring a single shot, one can easily find the temporal overlap between the two fs photon sources.In this way, a delay scan of few single pulses has allowed us to synchronize the optical laser andthe FEL to few hundreds of fs.

3 Results

3.1 Timing tool in the XUV range at FLASH

An example of this shadowgraph technique used at FLASH [15] is shown in figure 2. The FELbeam with a wavelength of λ = 13.5 nm, an averaged energy of 90 µJ and a pulse duration of∼ 100fs is focused to a 30 µm focal spot and then illuminated by an optical laser pulse at 800 nm with apulse duration of ∼ 60 fs [18]. The imaging system was composed of a single lens ( f = 50 mm)reaching a magnification of 10. This simple system allows the determination of temporal overlapwithin the given jitter limitation [3] and also to finalize the spatial overlap. An upper value for thejitter can be estimated by recording enough statistics for a few delay steps around the time overlapposition. At FLASH, this diagnostic has been used for various experimental conditions (energyvarying from 15 µJ to 200 µJ, pulse durations from 30 fs to 130 fs).

3.2 Timing tool in the X-ray range at LCLS

Such a diagnostic was also tested at LCLS [19] during a commissioning shift at the XPP station withFEL pulses of 8 keV photon energy and 80 fs pulse duration. The FEL beam was focused down to10 µm focal spot and the following results were obtained with a full FEL beam (0.4 mJ on target).

– 3 –

2012 JINST 7 P08007

Op#callaseronlybeforetheFEL

pulse

Duringtheinterac#on(at

2.2ps)

Op#callaseronlya=ertheinterac#on

Damage(DICmicroscopy)

a) b) c) d)

Figure 3. Images with a maximum FEL fluence, before, during and after the FEL shot. The delay betweenthe optical laser and the FEL is fixed at 2.2 ps. The last picture shows the damage spot taken with amicroscope.

Object = 20 x 15 µm Object = 1 x 1 µm

0.627 ps Damage

Object = 5 x 5 µm Object = 10 x 8 µm Object = 2 x 2 µm

1.221 ps 1.419 ps 1.617 ps

Expe

rimen

tX‐ray‐Tracing

Figure 4. The first row is showing images in transmission of the optical laser through the FEL plasma atdifferent delays. The second row is showing ray tracing Zemax calculations for an increasing object size.The first image called “Damage” is corresponding to an acquisition a long time after the FEL shot and hasbeen used to optimize initial parameters of the Zemax calculation.

Figure 3 shows some images taken at LCLS with an imaging system composed of a single lensonly and a CCD camera. The sample is a commercial microscope glass slide (Borosilicate glass) of1.2 mm thickness. The observed diffraction pattern is due to a slight defocused configuration of theimaging system. The four pictures are corresponding to the following situations (from left to right):a) an average of four reference pictures (optical laser without FEL); b) an image taken during a sin-gle shot at a fixed time delay (the optical laser pulse arrives 2.2 ps after the FEL pulse); c) an averageof five pictures taken without the FEL, few ms after the shot; and d) a differential interference con-trast (DIC) microscopy image of the damage. The optical laser and the FEL are arriving at a normalincidence to the target. As we have seen at FLASH, by taking a series of images for different delays,we can find the time overlap of the two beams with only a few shots and within the jitter limitation.

Figure 4 shows a delay scan between the optical laser and the FEL beam. The positive delayscorresponds to the optical laser arriving after the FEL. As described before, we can see the shadowof the high electron density volume evolving in time. We compared those images with a Zemax

– 4 –

2012 JINST 7 P08007

ray tracing calculation [20] including an 2-D object (a disc) placed on the glass slide. The Zemaxparameters were optimized by reproducing the image of the damage impact (images at the extremeleft) that has been precisely measured with a microscope afterwards (see figure 3). For the com-parison with the scan delay images, we only changed the object size to reproduce the diffractionpattern evolution. We can understand the image evolution by the increase of the electron densityvolume as function of time. The FEL focus has been measured with a knife-edge to ∼10 x 10 µmwhile the evolution of the diffraction pattern would suggest that at the beginning the opaque areais smaller. This would suggest that an intensity effect occurs and that only an intense central partof the FEL gaussian beam has to be taken into account. The time-dependance is related to the riseof the excited volume.

3.3 A jitter limited method

The accuracy of this method is mainly limited by the jitter between the optical laser and the FELarrival time. This has been measured up to 120fs RMS at LCLS [7] and up to 250 fs RMS atFLASH [21]. After a recent upgrade at FLASH, the jitter has been significantly improved to yieldless than 100 fs. The temporal jitter will result in fluctuations at time zero if sometimes one observesa shadow on the image and sometimes not. In this way we are able to evaluate the maximum valueof the jitter as soon as the shadow appear and disappear for a fixed delay. A statistical approachcould also allow us to extract the average value of the jitter by accumulating measurements for eachdelay step around the temporal overlap between the optical laser and the FEL pulse.

The pulse duration of the optical laser (∼ 40 fs at LCLS) is a second limitation of our tempo-ral resolution assuming that the plasma critical density is reached even faster [16]. This could beovercome by using few fs laser pulses.

By tilting the sample, it would be possible to turn this simple diagnostic into a complete hardX-rays timing tool as described in the references [2–5]. In this way, we would convert the arrivaltime information into a spatial effect by introducing a delay between one side of the FEL beamspot and the other. We would then project the time information spatially and measure directly theexact jitter shot to shot between the FEL and the optical laser. Such timing tool would then allowultrafast pump - probe experiments for a broad range of FEL wavelength. The physical phenomenabased on ultrafast photoionization is expected to be so fast for FEL [16] that it could potentiallysupport few 10’s of femtoseconds resolution.

4 Conclusion

We have developed a timing tool that has already been used successfully at FLASH and LCLS forseveral experiments. In these cases we were mainly limited by the jitter between the FEL and theoptical laser of ∼200 fs. In this paper we were able to demonstrate that one can adopt this timingtool to the X-ray regime. This diagnostic, based on a really simple set up (single lens imagingsystem) combined with a fast physical process affords a synchronization up to the jitter limitationwith only a few single shots. This study underlines the versatility of such reliable in-situ diagnosticadapted for pump-probe experiments on FEL facilities in the XUV to X-ray regime. This techniquecould easily allow a measurement of the average jitter. A major development of such diagnosticwould consist in measuring the shot-to-shot jitter online.

– 5 –

2012 JINST 7 P08007

Acknowledgments

We acknowledge all our FLASH, LCLS and Peak Brightness collaborators. We also thank B. Ziajaand N. Medvedev (CFEL, Center for Free Electron Science science, DESY) for stimulating dis-cussions. The authors are greatly indebted to the machine operators, run coordinators, scientificand technical teams of the FLASH and LCLS facilities for enabling an outstanding performance.E. Forster, U. Zastrau and V. Hilbert are grateful to the German Federal Ministry for Education andResearch (BMBF) via project FSP 301-FLASH. U. Zastrau acknowledges the VolkswagenStiftungvia a Peter-Paul-Ewald Fellowship.

References

[1] C. Gahl et al., A femtosecond X-ray/optical cross-correlator, Nature Phot. 2 (2008) 165.

[2] M. Drescher et al., Time-diagnostics for improved dynamics experiments at XUV FELs, J. Phys. B 43(2010) 194010.

[3] T. Maltezopoulos et al., Single-shot timing measurement of extreme-ultraviolet free-electron laserpulses, New J. Phys. 10 (2008) 033026.

[4] S. Schorb et al., X-ray optical cross-correlator for gas-phase experiments at the Linac Coherent LightSource free-electron laser, Appl. Phys. Lett. 100 (2012) 121107.

[5] M. Beye et al., X-ray pulse preserving single-shot optical cross-correlation method for improvedexperimental temporal resolution, Appl. Phys. Lett. 100 (2012) 121108.

[6] M. Meyer et al., Two-color photoionization in XUV free-electron and visible laser fields, Phys. Rev. A74 (2006) 011401(R).

[7] J.M. Glownia et al., Time-resolved pump-probe experiments at the LCLS, Optics Express 18 (2010)17620.

[8] K. Sokolowski-Tinten et al., Femtosecond X-ray measurement of coherent lattice vibrations near theLindemann stability limit, Nature 422 (2003) 287.

[9] D.M. Fritz et al., Ultrafast bond softening in bismuth: mapping a solid’s interatomic potential withX-rays, Science 315 (2007) 633.

[10] U. Teubner et al., Sub-10 fs gating of optical pulses, J. Phys. B 34 (2001) 2993.

[11] C. Thaury et al., Plasma mirrors for ultrahigh-intensity optics, Nature Phys. 3 (2007) 424.

[12] B. Nagler et al., Turning solid aluminium transparent by intense soft X-ray photoionization, NaturePhys. 5 (2009) 693.

[13] R. Faustlin et al., Observation of ultrafast nonequilibrium collective dynamics in warm densehydrogen, Phys. Rev. Lett. 104 (2010) 125002.

[14] V. Ayvazyan et al., Generation of GW radiation pulses from a VUV free electron laser operating inthe femtosecond regime, Phys. Rev. Lett. 88 (2002) 104802.

[15] K. Tiedtke et al., The soft X-ray free-electron laser FLASH at DESY: beamlines, diagnostics andend-stations, New J. Phys. 11 (2009) 023029.

[16] B. Ziaja et al., Unified model of secondary electron cascades in diamond, J. Appl. Phys. 97 (2005)064905.

– 6 –

2012 JINST 7 P08007

[17] S. Toleikis et al., Probing near-solid density plasmas using soft X-ray scattering, J. Phys. B 43 (2010)194017.

[18] H. Redlin et al., The FLASH pump-probe laser system: setup, characterization and optical beamlines,Nucl. Instrum. Meth. A 635 (2011) S88.

[19] S. Moeller et al., Photon beamlines and diagnostics at LCLS, Nucl. Instrum. Meth. A 635 (2011) S6.

[20] ZEMAX: software for optical system design webpage, http://www.zemax.com/, Bellevue U.S.A.(2011).

[21] P. Radcliffe et al., Single-shot characterization of independent femtosecond extreme ultraviolet freeelectron and infrared laser pulses, Appl. Phys. Lett. 90 (2007) 131108.

– 7 –