diagnostics pour accélérateurs à champ de sillage dans un plasma

Diagnostics pour accélérateurs à champ de sillage dans un plasma

Nicolas DelerueLAL (CNRS et Université de Paris-Sud)

Accélérateurs à plasma• Une onde de sillage dans un plasma peut-être utilisé

pour accélérer des particules.• Cette onde de sillage peut-être crée par un laser

=> Cf présentations de Brigitte et François• Elle peut aussi être crée par un faisceau de particules

=> Cf présentation de Patric• Ces accélérateurs à plasma sont très prometteurs:

- Très fort gradients d’accélérations- Impulsions très courtes (few fs)

• Potentiel complet en cours d’exploration• Possibilité d’améliorer le pilote (meilleurs lasers,

meilleurs faisceaux,…)

Nicolas Delerue, LAL (CNRS) 2Journées accélérateurs, Octobre 2011Leemans et al, doi:10.1038/nphys418

I Blumenfeld et al. 2007 Nature 445 741See also CERN Courier, May 2007 and CERN Courier, June 2007

Mesure des faisceaux d’accélérateurs à Plasma

• Les faisceaux de ces accélérateurs ne sont (pour l’instant) pas aussi stable que ceux des accélérateurs classiques.

• Les caractéristiques du faisceau changent de tir à tir.=> il faut utiliser des outils de mesure d’émittance à haute énergie et à un tir.

• Les impulsions sont ultra-courtes=> mesure de profil longitudinal ultra-court à un tir.

• A Oxford, travaux effectués dans le cadre d’une collaboration avec le groupe de Simon Hooker.

• Une partie de ces travaux se sont fait en collaboration avec DIAMOND Light Source Ltd (R. Bartolini et C. Thomas).

Nicolas Delerue, LAL (CNRS) Journées accélérateurs, Octobre 2011 3J.Osterhoff et al., PRL 101,085002(2008)

Mesure d’émittance transverse

Trois méthodes ont été explorées pour mesurer l’émittance transverse à haute énergie en un seul tir:

• La reconstruction de la propagation du faisceaux à travers plusieurs écrans OTR.

• Des poivrières « longues ».• L’utilisation de « photographies » des

particules du faisceau en deux points.

Nicolas Delerue, LAL (CNRS) Journées accélérateurs, Octobre 2011 4

Instrumentation de la ligne de transfert BTS de DLS

• Un faisceau était nécessaire pour tester ces méthodes.

• Nos collaborateurs à DLS ont instrumenté la ligne de transfert entre leur anneau d’accélération et leur synchrotron pour permettre ces tests.

• Possibilité de faire des mesures « parasites » lors des journées de développement machine.


Mesure en un seul tir de l’enveloppe du faisceau (1)

• L’enveloppe d’un faisceau peut être reconstruite en mesurant sa taille à plusieurs endroits.

• Pour faire cela en un seul tir il faut que chaque mesure ne perturbe pas les suivantes.

• Cela peut se faire en utilisant des écrans à transition de radiation visible (OTR) très fins (5um).


Delerue et al., arXiv:1005.2417

Mesure en un seul tir de l’enveloppe du faisceau (2)

• Méthode validée expérimentalement sur la ligne BTS de DLS.

• Correction de la dispersion nécessaire.

• Comparaison (et accord) avec le balayage d’un quadripole.


Thomas, Bartolini and Delerue, Jinst 6 p07004 (2011)

Application à un accélérateurà champ de sillage

• L’application à un accélérateur à champ de sillage s’est heurté au problème de cohérence partielle de la lumière émise par les écrans.

• Effets similaires observés sur les LELs X (LCLS,…)

• Cet effet ne peut pas être reproduit à DLS (faisceaux trop longs).

• Ebauche d’une méthode pour faire les même mesures dans l’UV/X proche.


Poivrière étendue• La méthode de la poivrière consiste en la

décomposition d’un faisceau en plusieurs sous-faisceaux dont on mesure la position et la divergence pour obtenir l’émittance transverse du faisceau initial.

• À 100 MeV ou au delà les électrons pénètrent profondément même dans les métaux les plus dense.

• Il faut donc utiliser un réseau de trous « profonds »=> Poivrière étendue=> Effet sur l’espace des phases mesuré?


Propagation dans une poivrière étendue


N. DelerueNIM A 644 (2011) 1-10

Tests expérimentaux d’une poivrière étendue

Mesures à DAFNE (BTF)

Emittance=4.0 mm.mrad (RMS, géometrique)

Inclinaison: 0.32mrad/mm

Distance au point focal: 3.1m


Delerue et al.PAC’09 TH5RFP065

« Image » des particules• À l’aide d’émulsions

nucléaires il est possible de « voir » l’ensemble d’un faisceau de particules à condition que la densité soit inférieure à 1e-/um^2.

• Deux plaques successives peuvent être utilisées pour mesurer la position et la direction de toutes les particules du faisceau.

• Le code d’analyse est assez complexe=> travail en cours…


Comparaison des méthodes

• Les faisceaux produits par les accélérateurs à onde de sillage sont très divers.

• Ces trois méthodes permettent de couvrir une large gamme de paramètres.


Mesure de profil longitudinal (1)

• Lorsqu’un paquet d’électron passe près d’un réseau il émet un rayonnement.• Pour les longueurs d’onde plus longues que le paquet ce rayonnement a une

composante cohérente d’intensité N2.• La distribution spectrale encode le profile longitudinal du paquet. • Le réseau disperse spectralement le signal: une mesure dans une direction donnée

corresponds à une longueur d’onde donnée.=> La distribution angulaire du signal dépends du profil du paquet.

• Démontré ps (Doucas et al., PRSTAB, 9, 092801, (2006)) , capacité de descendre jusqu’à quelques fs


]|)(F|)1N(NN[),(dd

dI),(dd

dI 2eee

spNe

Expérience E-203 à FACET

• Accepté pour tests expérimentaux à FACET (20 GeV, ~200fs).

• Mesures préliminaires en Août 2011• Campagne de mesures prévue en

Mars 2012.


Winston cones

beam direction

Pyroelectric detectors

Filter mechanism

DAQ

FWHM310 fs = 93 µm

FWHM 420 fs = 126 µm Bartolini et al.

IPAC’11 WEOBB03

Analyse préliminaire

Autres applications

• Bien que développées pour les faisceaux d’électrons produits dans un accélérateur à onde de sillage, ces outils peuvent être utilisées ailleurs:

• Tests d’une poivrière étendue au CERN sur le linac 2 (50 MeV, p+) pour une éventuelle application au LINAC 4 (100MeV H-)

• Mesure de longueur de paquets de laser à électrons libre…


Conclusion• Développement de plusieurs outils pour mesurer des

faisceaux ultra-courts de haute énergie.• Important dans le cadre des recherches françaises sur

les accélérateurs à champ de sillage (CILEX,…)• Plusieurs tests sur accélérateurs classiques, à la

recherche d’opportunités pour des tests sur accélérateurs à champs de sillage.

• Ce genre de développement peut donner lieu à de nombreux petits projets attractifs pour des stagiaires.

• Travaux à adapter au contexte de financement français.


Merci pour votre attention

Collaborateurs:Riccardo Bartolini, Cyrille Thomas (DLS)

George Doucas, David Urner, Armin Reichold, Colin Perry, Simon Hooker, Andreas Walker,… University of Oxford

Anthony Ashmore, Joe Hewlett, Ewen McLean, Kate Pattle, Clémentine Santamaria, Nick Shipman, Richard Tovey, Bas-Jan Zandt … Stagiaires

Opérateurs de DLS, SLAC/FACET, DAFNE/BTF


Limitations of current acceleration techniques

• In an accelerator particles are accelerated by an electromagnetic wave creating an accelerating gradient when the particles pass.

• If the gradient is too high an imperfection in the cavity may concentrate too much EM power and create a spark (breakdown) leading to the loss of the beam and possible damages to the cavity.

• So far accelerating cavities are limited to less than 100MV/m.

Plasma AcceleratorsNicolas Delerue, LAL Orsay 19

Plasma acceleration• If the cavity is replaced by something that

can not be damaged, much higher accelerating gradients can be reached.

• This can be done by creating a wakefield in a plasma.

• Different tools are used to create such plasma accelerator:- Lasers (Tajima and Dawson, PRL, 1979)- Electron beams (Hogan et al., PRL, 2005)- Proton beams (Caldwell et al., Nature Phys., 2009)

Nicolas Delerue, LAL Orsay Plasma Accelerators 20

Wakefield in a plasmaSource: CERN Courier

Example of wakefieldSource: http://www.arwenmarine.com

Laser-driven plasma acceleration

• First proposed in 1979 (Tajima and Dawson)• In the 90s LULI demonstrated that injected

electrons can be accelerated from 3 MeV to 4.5 MeV (Amarinoff et Al., PRL, 1998).

• Significant progress in 2004: Nature: “dream beam”- RAL/IC/UK: Mangles et al.- LOA/France: Faure et al.- LBNL/USA: C.G.R. Geddes et al.


Laser-driven plasma acceleration:“dream beam”

• In the “dream beam” experiments the plasma was created in a gas jet.

• In this case the electrons are taken from the ions inside the plasma.

• In 2008 a beam of 800 MeV was produced using this technique (Kneip,…, ND, et al., PRL, 2009)


Mangles et al, doi:10.1038/nature02939

Faure et al., doi:10.1038/nature05393

GeV LPA beam• Another breakthrough occurred in 2005 when a Berkeley +

Oxford collaboration reached an energy of 1 GeV.• To do so they used a 33mm long capillary to extend the

length over which the plasma has the right properties to accelerate electrons.

• The field created during this experiment was of the order of 100 GV/m!


Leemans et al, doi:10.1038/nphys418

Typical features of LPA beams achieved so far

• Energy 50 MeV – 1 GeV• Energy spread: from very large to a few percent• Low repetition rate: 1Hz and less

(eg: 1 shot every 20s at GEMINI)• Low charge: 10-50pC • Ultra-short pulses: 5-30fs (measure very difficult)

=> High peak current• Shot to shot reproducibility is poor.• Very small footprint for the “accelerator”, laser included (few rooms).• Simulations are difficult (particle in cell on large computers)• Not all plasma acceleration experiments produce the beam

predicted/expected…


J.Osterhoff et al., PRL 101,085002(2008)

Electron driven plasma acceleration

• Electrons can also be used to generate the wakefield required to accelerate other electrons.

• This has been demonstrated using the SLAC LINAC when the energy of some electrons of a bunch was doubled from 42 to 85 GeV.

• The field was about 50 GV/m.


I Blumenfeld et al. 2007 Nature 445 741See also CERN Courier, May 2007 and CERN Courier, June 2007

Proton driven plasma acceleration• More recently there has been a proposal

to use protons as drive beam to accelerate electrons.

• Simulations indicate that the CERN SPS beam could be used to accelerate electrons to 600 GeV.


Caldwell et al., Nature Physics 5, 363 - 367 (2009)

Toward a plasma-based collider

• The next electron collider will require beams with an energy of between 500 GeV and a TeV.

• The production of such beam by a plasma-based collider has not yet been demonstrated, however there are proposal to stage several “accelerating sections” to achieve this.

Nicolas Delerue, LAL Orsay Plasma Accelerators 27Source:LBL

Source:CERN Courier

Limits

• Several important steps need to be made before such collider can be built.

• Plug to beam power efficiency is very low. In the case of laser-driven accelerators, fibre lasers may improve this but it may not be enough.

• Beam stability is another issue: at the moment each shot is different and goes in a slightly different direction…

• Very little is know about the quality (emittance) of such beam.


Getting involved

• Although a significant fraction of the developments require laser physics and plasma physics skills, particle and accelerator physicists can help by addressing several questions:

• What are the beam properties? (Energy, Energy spread, transverse emittance, longitudinal profile…)

• What happens when the beam propagates in matter?

• Etc…


Diagnostics forplasma accelerators

• Although the energy/energy spread measurement techniques are well understood (magnet + screen), other diagnostics are not so trivial.

• They must be able to measure the beam properties in a single shot at high energy.

In Oxford we focussed on:- Transverse emittance measurement- Longitudinal profile measurement


Transverse emittance• The emittance of a beam is a measure of

its quality (“temperature”).• It is the volume occupied by the beam in

the phase space.• The transverse emittance tells how

strongly the beam can be focussed => Luminosity.

• Typical high energy techniques average over several bunches.

• We studied 3 different transverse emittance measurement techniques:- Multi-OTR measurement- High energy pepper-pot- Nuclear emulsions based tracking



Multiple Optical Transition Radiation profile measurements

When an electrically charged particle experience a change of medium it radiates => Transition Radiation

Optical Transition Radiation is commonly used at accelerators to image high energy beams but it scatters the beams.

Unlike phosphorescence, OTR is a surface effect, independent of the screen thickness.

The scattering induced by an ultra-thin screen may be acceptable.

Effect of the scattering• The condition for the scattering

to be negligible can be derived:

• For a small (focussed) beam the natural divergence will dominate the effect of the scattering whereas for a collimated beam the scattering will dominate.

• GEANT4 simulations were used to validate these calculations.


Delerue et al., arXiv:1005.2417(submitted to JInstr)

Experimental validation

• This was verified experimentally using the DIAMOND (UK) BTS (3 GeV electrons).

• The beam optics validates the condition on x and is close to it in y.

• Scattering with OTR screens is seen in y and not in x.

Nicolas Delerue, LAL Orsay Plasma Accelerators 34Thomas, ND et al., IPAC’10 MOPE080

Single shot transverse emittance measurement at 3 GeV

Nicolas Delerue, LAL Orsay Plasma Accelerators 35Thomas, ND et al., IPAC’10 MOPE080

Measured value close from the DIAMOND predictions.

Shot to shot variations observed…

OTR summary

• We have shown that thin OTR screens can be used to measure the emittance of a high energy beam in a single shot.

• An attempt was made to make a similar measurements at a plasma accelerator last June but was hampered by coherent effects.=> more R&D needed to address those.


Pepper-pots• At low energy the usual transverse emittance

measurement methods uses an array of slits or holes to split the incoming beam into several beamlets.

• For each hole the position of the beamlet is known.• A screen located downstream measures the

divergence of each beamlet.• At HE e- penetrate in matter…


Extended pepper-pots

• Instead of a thin foil deep channels can be used to form the beamlets.


• Two challenges:- Mechanical assembly- Phase-space preservation


26 July 2010 Nicolas Delerue – University of Oxford 39

Creating thin channels is quite challenging from a mechanical point of view.

We form them by stacking absorbers and shims.

Require ultra-flat Tantalum (or Tungsten) => industrial collaboration.

High energy pepper-pots(extended pepper-pots)

Phase space preservation


N. DeleruearXiv:1005.3023v1(Submitted to NIM)

Extended PP: Beam tests at DAFNE

• Beam tests at DAFNE BTF: 508 MeV electrons Single shot mode possible



High energy pepper-pots(extended pepper-pots)

RMS Emittance=4.0 mm.mrad

(geometric)

Shearing: 0.32mrad/mm

Calculated distance to waist: 3.1m

Delerue et al.PAC’09 TH5RFP065

Pepper-pot summary

• We have demonstrated that extended pepper-pot can work when positioned correctly.

• We have shown that it is possible to measure the emittance and reconstruct the phase-space of a 508 MeV beam in a single shot.

• Tests are on-going at DIAMOND to extend this method to 3 GeV!


Nuclear emulsions based tracking• Perfect emittance

measurement => position + direction of each particle.

• Nuclear emulsions can resolve particles with a resolution of about 1um.

• Stacks of thin emulsions can resolve the direction and the position of a beam.

• Damaging an emulsion plate with a high power laser is less of a problem than damaging an expensive camera!


Particle tracking• Electrons above 100 MeV are

not significantly scattered by a thin layer of Nuclear Emulsions.

• We used a technique called “image registration” to match (rotate, scale,…) the simulated images from two consecutive emulsion plates.

• A motorised microscope allows the scanning of a large area.

• A low particle density is necessary. At a LPA this is usually achieved because the beam has a strong divergence.


• We demonstrated that all the bits work but did not have time to put them together within the duration of the grant.

• There may be applications with ion beams.


Why 3 techniques

• The 3 techniques we studied are meant to be complementary:- Multi-OTR requires an intense beam to have enough OTR signal.- At low energy scattering is too intense.


Why 3 techniques

- Pepper-pots won’t work with electron beams of several GeV - but they can work with very high density beams.



Why 3 techniques

- Emulsions only work with low density beams.- They work over a large range of energy.

The information provided by the 3 methods is also slightly different.



Longitudinal profile measurement


There are several techniques to measure longitudinal profiles of electron bunches.

The most straightforward is to use a RF deflecting cavity (or a streak camera) => but not suitable for ultra short beams.

Electro-optic sampling uses the wakefield induced by the beam in a crystal to modulate the field of a laser.=> but unable to reach ultra short bunches (below 30fs)

Several techniques (CTR, CDR, …) use the radiation emitted by a relativistic bunch when passing trough/near an interface.

Longitudinal profile measurement


Smith-Purcell radiation

Smith-Purcell radiation is emitted when relativistic charged particles pass near a grating.

Pioneering work on its use as longitudinal profile monitor done in Oxford by George Doucas (in the picosecond regime).


In a Smith-Purcell detector, radiation of a certain wavelength is emitted in a fixed direction.=> Allows to measure these different wavelength “easily” (the grating acts as spectrometer).=> Give access to the Fourier transform of the bunch profile.=> An inversion technique can then be used to reconstruct the bunch length but also the profile of the electron bunch.

Smith-Purcell radiation


The existing detector uses an array of 11 pyroelectric detectors able to detect far-infrared radiation.

To increase the range of wavelengths that can be measured 3 sets of filters can be inserted in front of the detectors.

Smith-Purcell detector

Bunch profile reconstruction

Different bunch profiles will give different radiation spectrums which can be reconstructed.



Smith-Purcell radiation at a LPA• To extend the range of the

detector from ps to fs the wavelength sensitivity needs to be modified.

• The previous experiment used mm radiation whereas fs electron will emit micrometric radiations.

• Simulations show that we are also sensitive to small satellites.

• We have applied to do tests at FACET (SLAC) next summer.

Delerue et al.arXiv:0912.1863

What to do with a plasma accelerator?

• Staging hundreds of plasma accelerators to build a “plasma based” linear collider is an interesting idea but there are still many challenges to solve before it can be done.

• However there are other applications for sources of electrons between 200 MeV and 1 GeV.

• The bunch energy and duration (10s of fs) are comparable to those required by free electron lasers (FELs).

Nicolas Delerue, LAL Orsay Plasma Accelerators 56Source:Lightsources.org

Source:LBL

Toward a 5th generation light source


A 500 MeV beam can be used to shoot in an undulator and create X-rays.

Very attractive source: low footprint (Fits in a university), high brightness,....

Reasonably priced 5-10MEuros (much cheaper than a conventional X-rays FEL).

Bajlekov et al.FEL’08 MOPPH081


Toward a 5th generation light source


BUT the output of a FEL is strongly correlated with the beam properties.

To use plasma accelerators as 5th generation light sources one needs to be able to measure the beam’s longitudinal profile and its transverse emittance!

Outlook• Plasma accelerators have made tremendous progress

since 1979.• They have now reached a stage where applications can

be considered.• It is likely that in the coming few years there use a X-

rays FEL will be demonstrated.• Particle physics applications will have to wait longer…• However they are many opportunities for particle

physicists to help this date come sooner:- How to measure the properties of a multi-GeV beam?- In a multi stage accelerator, what happens when the incoming electrons pass through the laser injection mirror?- …


Thank you for your attention


Positrons acceleration• The leading laser-driven plasma acceleration results

(“dream beam”, GeV beam,…) accelerated the electrons from the plasma.=> Not possible for positron acceleration

• Other results (LULI,…) used externally injected electrons. In that case the acceleration gradient is usually smaller (“linear regime”).=> Suitable for positron acceleration

• Particle driven plasma accelerators use externally injected particles.

• In the case of proton driven acceleration, it may even be possible to accelerate electrons and protons in the same train (but at different phases).


diagnostics pour accélérateurs à champ de sillage dans un plasma

Documents