Laser Plasma Accelerators

V. Malka
Laboratoire d’Optique Appliquée, ENSTA-ParisTech,
CNRS, Ecole Polytechnique,
UMR 7639, 91761 Palaiseau, France

(Dated: October 30, 2018)

Research activities on laser plasma accelerators are paved by many significant breakthroughs.
This review article provides an opportunity to show the incredible evolution of this field of research
which has, in record time, allowed physicists to produce high quality electron beams at the GeV
arXiv:1112.5054v1 [physics.plasm-ph] 21 Dec 2011

level using compact laser systems. I will show the scientific path that led us to explore different
injection schemes and to produce stable, high peak current and high quality electron beams with
control of the charge, of the relative energy spread, and of the electron energy.

I. INTRODUCTION electrons were injected externally and were accelerated

through the very high electric field sustained by relativis-
The discovery of supra conductivity at a 4 Kelvin tem- tic plasma waves driven by lasers. In this former article3 ,
perature in mercury cooled with liquid helium, made by the authors have proposed two schemes: the laser beat
K. Onnes in 1911, has opened a very active field of re- wave and the laser wakefield. Several experiments have
search. After decades of effort, researchers have been able been performed in the beginning of the nineties follow-
to produce new superconductors working at “high” tem- ing their idea, and injected electrons at the few MeV
perature. Even if there remains open fundamental ques- level have indeed been accelerated by electric fields in the
tions, mastering superconductivity has allowed the dis- GV/m range in a plasma medium using either the beat
covery of tremendous applications in magnetic resonance wave or the laser wakefield scheme. Before the advent of
imaging (MRI) for medicine, in transportation (magnetic short and intense laser pulses, physicists have used the
suspension trains), and in modern accelerators (super- beat wave of two long laser pulses (i. e. with duration
conducting cavities). In parallel, since the first 1.26 MeV much greater than the plasma period) of a few tens of
Hg ion beam made by E. O. Lawrence, accelerators have ps with two frequencies ω1 and ω2 to drive the relativis-
gained in efficiency and in performance. With a market tic plasma waves in a perfect homogenous plasma at a
of more than 3 Billions dollars per year, accelerators are density for which the plasma frequency (ωp ) satisfies ex-
used today in many fields such as cancer therapy, ion im- actly the matching condition, ωp = ω1 − ω2 . The first
plantation, electron cutting and melting, non destructive observation of relativistic plasma waves was performed
inspection, etc...For fundamental research, the most en- using Thomson scattering technique by the group of C.
ergetic machines (those that deliver particle beams with Joshi at UCLA4 . Acceleration of 2 MeV injected elec-
energies greater than 1 GeV represent only 1% of the trons up to 9 MeV5 and later on, up to 30 MeV6 has
total number of accelerators) have been developed, for been demonstrated by the same group using a CO2 laser
example for producing intense X rays beams in the free of about 10 µm wavelength. At LULI, 3 MeV electrons
electron laser scheme for the study of ultra fast phenom- have been accelerated up to 3.7 MeV in beat wave experi-
ena of interest for example in biology to follow the DNA ments with Nd:Glass lasers of about 1µm wavelengths by
structure evolution, or in material science to follow evolu- a longitudinal electric field of 0.6 GV/m7 . Similar work
tion of molecules or of crystal structures. Higher energies was also performed in Japan at University of Osaka8 , in
accelerators are crucial to answer to important questions U.K. at Rutherford Appleton Laboratory9 , or in Canada
regarding the origin of the universe, of the dark energy, at Chalk River Laboratory10 . Plasma waves driven in the
of the number of space dimension, etc...The larger one, laser wakefield regime at LULI by a few J, 300 fs laser
the Large Hadron Collider, is expected, for example, to pulse have been used to accelerate 3 MeV injected elec-
reveal very soon properties of the Higgs boson. trons up to 4.6MeV11 . In all these experiments, because
Since the accelerating field in superconducting Radio- of the duration of the injected electron which was much
Frequencies cavities is limited to about 100MV/m, the longer than the plasma period and even longer than the
length of accelerators has to increase in order to achieve life time of the plasma, only a very small fraction of in-
higher energy gain. To overcome this gigantism issue, jected electrons were accelerated and the output beam
J. Dawson1,2 ) proposed to use a plasma, which, as a had a very poor quality with a maxwellian-like energy
ionized medium, can support and sustain extreme elec- distribution. Optical observation of radial plasma oscil-
tric fields. The pioneer theoretical work performed in lation has been observed at LOA12,13 and in Austin14
1979 by Tajima and Dawson3 has shown how an in- with a time resolution of less than the pulse duration by
tense laser pulse can excite a wake of plasma oscillation using spectroscopy in the time-frequency domain. More
through the non linear ponderomotive force associated recently, at CUOS, using the same technique but with
to the laser pulse. In their proposed scheme, relativistic chirped probe laser pulses, a complete visualization of
 2

relativistic plasma wave has been performed in a sin- the quality of the electron beam was improved by reduc-
gle shot. These results have revealed very interesting ing noticeably the interaction between the laser beam
features such as the relativistic lengthening of the laser and the electron beam. To improve fairly the electron
plasma wavelength15 visible on figure 1. beam quality, one has to reduce electrons injection to
a very small volume of the phase space. In general,
this means that injected electrons must have a duration
much shorter that the plasma period, i.e. much less than
ten femtoseconds which is difficult to achieve easily to-
day with current accelerator technology. I will show in
this review article how physicists have solved this cru-
cial problem by exploring different schemes like the bub-
ble regime, the density gradient injection technique, the
space limited ionization approach and the colliding laser
pulses scheme. The fil rouge of this article is the physic
of electron injection. Following this fil rouge, I will show
that the control of electron injection in a limited space
and time region led us to producing stable and very high
quality electron beam with a some level of control over
the charge, the energy spread, and the electron beam
energy.
FIG. 1. The laser pulse, that propagates from left to right,
drives a strong wake with relativistic curved front, Courtesy
of M. Downer15 .
II. CONTROLLING THE INJECTION
These first experiments have shown acceleration of ex-
ternally injected electrons. With the development of Controlled injection in laser plasma acceleration that
more powerful lasers, much higher electric fields were lead to high electron beam quality is particularly chal-
achieved, giving the possibility to accelerate efficiently lenging due to the very small value of the length of the
electrons from the plasma itself to higher energies. A injected bunch that has to be a fraction of plasma wave
major breakthrough, was obtained in 1994 at Rutherford wavelength , with typical values in the [10-100µm] range.
Appleton Laboratory, where relativistic wave breaking Doing so, electrons witness the same accelerating field,
limit was reached16 . In this limit, the amplitude of the leading to the acceleration of a monoenergetic and high
plasma wave was so large, that copious number of elec- quality bunch. Electrons can be injected if they are lo-
trons were trapped and accelerated in the laser direction, cated at the appropriate phase of the wake and/or if they
producing an energetic electron beam. A few hundreds have sufficient initial kinetic energy. Different schemes
of GV/m electric field was measured. The correspond- have been demonstrated today and allow to control the
ing mechanism is called the Self Modulated Laser Wake phase of injected electrons.
Field (SMLWF)17–19 , an extension of the forward Ra-
man instability20,21 at relativistic intensities. In those
experiments, the electron beam had a maxwellian-like A. Bubble regime
distribution as it is expected from random injection pro-
cesses in relativistic plasma waves. This regime has also In 2002, using 3D PIC simulations, A. Pukhov
been reached for instance in the United States at CUOS22 and J. Meyer-Ter-Vehn have shown the existence of
and at NRL23 . However, because of the heating of the a very promising acceleration regime, called the bub-
plasma by these relatively “long” pulses, the wave break- ble regime27 , that leads to the production of a quasi-
ing occurred well before reaching the cold wave break- monoenergetic electron beam. At lower laser intensity,
ing limit. The maximum amplitude of the plasma wave the blow-out regime28 , also allows to obtain such an
in the range 20-60 %24 was measured using a Thomson electron distribution. In those two regimes, the focused
scattering diagnostic. Energetic electron beams were pro- laser energy is concentrated in a very small sphere, of
duced with a compact laser working at 10Hz at MPQ25 radius shorter than the plasma wavelength. The associ-
in the direct laser acceleration scheme(DLA) and at LOA ated ponderomotive force expels radially electrons from
in the SMLWF26 . At LOA, the increase of the electron the plasma, forming a positively charged cavity behind
peak energy when decreasing the electron plasma density the laser, and surrounded by a dense region of electrons.
have nicely demonstrated that the dominant acceleration As the radially expelled electrons flow along the cav-
mechanism was due to relativistic plasma waves. In 2002, ity boundary and collide at the bubble base, transverse
another breakthrough was obtained in the forced laser breaking occurs29 providing a well localized region of in-
wakefield where a low divergent electron beams with en- jection in the cavity.
ergies up to 200 MeV was obtained with the 1J “salle Since the injection is well localized, at the back of the
jaune” laser at LOA. In this highly non linear regime, cavity, it gives similar initial properties in the phase space
 3

mogeneous helium gas44,45 jet is fully ionized early in

the interaction. Since the electron density value is not
high, ionization induced refraction did not play an im-
portant role. The focal position and its value with re-
spect to the sharp gas jet gradient have been measured
and varied in order to optimize the electron beam pa-
rameter. This optimum position is found when focusing
the laser beam on the edge of the plateau of the gas jet.
At high electron plasma densities, for densities greater
than 2 × 1019 cm−3 , the electron beam has a maxwellian
FIG. 2. The laser pulse that propagates from left to right, ex- like distribution with a corresponding temperature that
pels electrons on his path, forming a positively charged cavity. increases at lower plasma density. In this range of den-
As the radially expelled electrons flow along the cavity bound- sity the self modulated regime is the dominant accelera-
ary and collide at the bubble base, before being accelerated
tion mechanism. At lower density, for densities comprise
behind the laser pulse.
between 7.5 × 1018 cm−3 and 2 × 1019 cm−3 , the forced
laser wakefield regime dominates and a plateau appears
in the electron beam distribution. The bubble/blowout
to injected electrons. The trapping stops automatically regime appears only in the very low density range, be-
when the charge contained in the cavity compensates low 6 × 1018 cm−3 with a quasi-monoenergetic electron
the ionic charge, leading to the generation of a quasi- distribution.
monoenergetic electron beam that was experimentally
demonstrated in 200430–32 . Finally, the rotation in the
phase-space also leads to a decrease of the spectral width
of the electron beam33 . Electron beam quality is also
improved because electrons that are trapped behind the
laser do not interacted anymore with the electric field of
the laser. The scheme of principle of the bubble/blowout
regime is illustrated on figure 2.
Several laboratories have obtained quasi monoener-
getic electron beams in the bubble/blow-out regime :
in France32 with a laser pulse shorter than the plasma
period, but also with pulses slightly longer than the
plasma period in England30 , in the United States31 , then
in Japan34 , in Tawain35 and in Germany36,37 , and in
Sweden38 . Electrons at the GeV level were observed in
this regime using in a uniform plasma39,40 or in a plasma
discharge, i.e, a plasma with a parabolic density profile41
that allows the intense laser beam to propagate over a
longer distance, of a few centimeters. In all of those
experiments the laser beam parameters did not satisfy
fully all the bubble/blow out regime criteria. Neverthe-
less thanks to the self focusing and self shortening42 ef-
fects, the non linear evolution of the laser puls allowed FIG. 3. Electron beam distribution for different plasma den-
such transverse injection, and some of those experiments sity showing the transition between the Self Modulated Laser
are in a regime between the forced laser wakefield and Wakefield, the Forced Laser Wakefield and the Bubble/Blow-
the bubble/blowout regime. The figure 3 illustrates in a out regime. From top to button the plasma density values are
single experiment the transition between the self modu- 6×1018 cm−3 , 1×1019 cm−3 , 2×1019 cm−3 , and 5×1019 cm−3 .
lated laser wakefield, the forced laser wakefield and the
bubble/blowout regime. The experiment was performed
at Laboratoire d’Optique Appliquée with the 10 Hz, 30 It has been shown that with current laser plasma pa-
TW Ti:sapphire laser system operating at wavelength rameters, the bubble/blowout regime was not yet com-
λ0 = 0.82 µm43 . The laser pulse that had a duration pletely established. With the increase of laser power
at Full Width Half Maximum (FWHM) of τ = 30 fs systems, this regime will be reached, and significant im-
was focused using a 1 m focal length off axis parabolic provement of the reproducibility of the electron beam
mirror, onto a focal spot with Full Width Half Maxi- is expected. Nevertheless, since self-injection occurs
mum (FWHM) of 18 µm producing a laser intensity of through transverse wave breaking, it is hardly appropri-
2
3.2 × 1018 W/cm . The corresponding normalized vec- ate for a fine tuning and control of the injected electron
tor potential was a0 = 1.3. At this intensity, the ho- bunch.
 4

B. Injection in a density gradient

One solution to control electron injection with current

laser technology has been proposed by S. Bulanov et al.46
using a downward density ramp with a density gradient
scale length Lgrad greater than the plasma wavelength
λp . Injection in a downward density ramp relies on the
slowing down of the plasma wave velocity at the density
ramp. This decrease of the plasma wave phase veloc-
ity lowers the threshold for trapping plasma background
electrons and causes wave breaking of the wakefield in
the density ramp. This method can therefore trigger
wave breaking in a localized spatial region of the plasma.
Geddes et al.47 have shown the injection and accelera-
tion of high charge (> 300 pC) and stable quality beams
of ' 0.4 MeV in the downward density ramp at the exit
of a gas jet (Lgrad ' 100 µm  λp ). These results, al-
though very promising, have the disadvantage that the
low energy beam after the plasma blows up very quickly
due to space charge effect. To circumvent this issue, one
should use a density gradient located early enough along
the laser pulse propagation so that electrons can be ac-
celerated to relativistic energies48 . This can be achieved
by using a secondary laser pulse to generate a plasma
channel transversely to the main pulse propagation axis FIG. 4. Top: The target schematic representation with
for instance49 . Doing so, the electron beam energy has embedded supersonic gas jet into a capillary that is filled with
been tuned by changing the position of the density gra- hydrogen gas, Buttom : the charge (squares), energy (circles),
dient. Nevertheless, because of the small value of the and energy spread (triangles) as a function of the peak jet
laser energy, the electron beam had a large divergence density. From A. J. Gonsalves et al.53
and a Maxwellian energy distribution. 2D PIC simula-
tions have shown that this method can result in high
quality quasi monoenergetic electron beams50 . Trapping
across a laser generated plasma channel, which has the
advantage of ease of production and control has also been the gas jet with a plasma discharge53 , more energetic
considered in simulations48 . electrons at 30 MeV has been produced in the density
At LOA a density gradient across a laser created plasma ramp, and have been accelerated up to 400 MeV in a 4 cm
channel has been used to stabilize the injection51 . The parabolic plasma channel. Here also, the density gradi-
experiment was performed at an electron density close ent injection has allowed an improvement of the stability
to the resonant density for the laser wakefield (cτ ∼ λp ) and of the electron beam quality. The electron energy,
to guaranty a post acceleration that deliver high quality divergence, charge and relative energy spread were found
electron beams with narrow divergences (4 mrad) and to be respectively 400 MeV, 2 mrad, 10 pC and 11%.
quasi monoenergetic electron distributions with 50 to 100 It has been shown that steeper density transitions, with
pC charge and 10% relative energy spread. The use of Lgrad  λp , can also cause trapping54 . Using a shock-
density gradients at the edges of a plasma channel has front created at the knife-edge of a gas jet such injection
shown an improvement of the beam quality and of the re- has been successfully demonstrated experimentally55,56 .
producibility with respect to those produced in the bub- The figure 5 illustrates the improvement of injection in
ble/blowout regime with the same laser system and with a sharp density gradient, with a characteristic length on
similar laser parameters. However, the electron energy the order of the plasma wavelength and a peak electron
distribution was still found to fluctuate from shot to shot. density of about 5 × 1019 cm−3 . The experiment was
This injection scheme is promising because it permits to performed at Max-Planck-Institut fur Quantenoptik us-
relax the experimental requirements in terms of synchro- ing a multi-TW sub-10-fs laser system that delivers for
nization (ns) and spatial alignment (100 µm). The per- this experiment pulses with 65 mJ energy on target and
formances of the experiment could be further improved a duration of 8 fs FWHM. The laser pulse was focused
and could potentially lead to more stable and control- down a spot diameter of 12 µm FWHM onto the gas tar-
lable high quality electron beams. In particular, sharper 2
get yielding at a peak intensity of 2.5×1018 W/cm . The
gradients with Lgrad ' λp coupled with a long plasma comparison between the self injection and density tran-
might lead to better beam quality52 . sition injection shows a reduction of the relative energy
For example, at LBNL as shown on figure 4, coupling spread and of the charge of a about a factor of 2.
 5

FIG. 5. A few shots representative for those 10% of all the

shots with lowest energy spread for self-injection (top) and
injection at a density transition (buttom). The horizontal
axis in each image corresponds to the transversal electron
beam size; the vertical axis shows electron energy. From K.
Schmid et al.56 .

C. Injection with colliding laser pulses

In 2006, stable and tunable quasimonoenergetic elec-

tron beams were measured by using two laser beams in
the colliding scheme with a counterpropagating geome-
try. The use of two laser beams instead of one offers
more flexibility and enables one to separate the injection FIG. 6. Scheme of principle of the colliding laser pulses: (a)
the two laser pulses propagate in opposite direction, (b) dur-
from the acceleration process57 . The first laser pulse,
ing the collision, some electrons get enough longitudinal mo-
the pump pulse, is used to excite the wakefield while the mentum to be trapped by the relativistic plasma wave driven
second pulse, the injection pulse, is used to heat elec- by the pump beam, c) trapped electrons are then accelerated
trons during its collision with the pump pulse. After the following the pump laser pulse.
collision has occurred, electrons are trapped and further
accelerated in the wakefield, as shown on figure 6.
To trap electrons in a regime where self-trapping does
not occur, one has either to inject electrons with ener- try based on the use of three laser pulses. This idea
gies greater that the trapping energy or dephase elec- was further developed by considering the use of two laser
trons with respect to the plasma wave. As mentioned pulses59 . In this scheme, a main pulse (pump pulse) cre-
earlier, electrons need to be injected in a very short time ates a high amplitude plasma wave and collides with a
(< λp /c) in order to produce a monoenergetic beam. secondary pulse of lower intensity. The interference of the
This can be achieved using additional ultrashort laser two beams creates a beatwave pattern with a zero phase
pulses whose purpose is only restricted to triggering elec- velocity, that heats some electrons from the plasma back-
tron injection. Umstadter et al.22 first proposed to use ground. The force associated with this ponderomotive
a second laser pulse propagating perpendicular to the beatwave is inversely proportional to the laser frequency
pump laser pulse. The idea was to use the radial pon- is therefore many times greater than the ponderomotive
deromotive kick of the second pulse to inject electrons. force associated with the pump laser that is inversely pro-
Esarey et al.58 proposed a counter-propagating geome- portional to the pulse duration at resonance. Therefore,
 6

consequences on the dynamics of the plasma wave dur-

ing the beating of the two laser pulses, (ii) the laser pulse
evolution which governs the dynamics of the relativistic
plasma waves61 . New unexpected feature have shown
that heating mechanism can be achieved when the two
laser pulses are crossed polarized. The stochastic heating
can be explained by the fact that for high laser intensities,
the electron motion becomes relativistic which introduces
a longitudinal component through the v × B force. This
relativistic coupling makes it possible to heat electrons
even in the case of crossed polarized laser pulses62 . Thus,
the two perpendicular laser fields couple through the rel-
ativistic longitudinal motion of electrons. The heating
level is modified by tuning the intensity of the injection
laser beam or by changing the relative polarization of the
two laser pulses63 . This consequently changes the volume
in the phase space and therefore the charge and the en-
ergy spread of the electron beam. Figure 8 shows at a
given times (42 fs) the longitudinal electric field, during
FIG. 7. In red, normalized longitudinal electric field. Case and after collision for parallel and crossed polarization.
a) the laser (in pink) wakefield. Case b) the electron bunch The solid line corresponds to the PIC simulation results
(in blue) wakefield. Case c) resulting fields when laser and whereas the dotted line corresponds to the fluid calcu-
electron beams are presented together. Parameters are a0 = lation. The laser fields are also represented by the thin
1, the laser pulse duration 30 fs, ne = 7 × 1018 cm−3 , nbeam = dotted line. When the pulses have the same polarization,
0.11 × ne , bunch duration of 10 fs and diameter 4 microns. electrons are trapped spatially in the beatwave and can
From C. Rechatin PhD Thesis. not sustain the collective plasma oscillation inducing a
strong inhibition of the plasma wave which persists af-
ter the collision. When the polarizations are crossed, the
the mechanism is still efficient even for modest values of
motion of electrons is only slightly disturbed compared to
laser intensities. Upon interacting with this field pattern,
their fluid motion, and the plasma wave is almost unaf-
some background electrons gain enough momentum to be
fected during the collision, which tends to facilitate trap-
trapped in the main plasma wave and then accelerated to
ping.
high energies. As the overlapping of the lasers is short in
time, the electrons are injected in a very short distance Importantly it has been shown that the colliding pulse
and can be accelerated to an almost monoenergetic beam. approach allows a control of the electron beam energy
This concept has been validated in an experiment57 , us- which is done simply by changing the delay between the
ing two counter-propagating pulses. Each pulse had a two laser pulses57 . The robustness of this scheme has also
duration of 30 fs at full width half maximum (FWHM), allowed to carry out very accurate studies of the dynamic
with a0 = 1.3, a1 = 0.4. They were propagated in a of electric field in presence of high current electron beam.
plasma with electron density ne = 7 × 1018 cm−3 corre- Indeed, in addition of the wakefield produced by the laser
sponding to γp = k0 /kp = 15. It was shown that the pulse, a high current electron beam can also drive its own
collision of the two lasers could lead to the generation of wakefield as shown on figure 7.
stable quasi-monoenergetic electron beams. The beam This beam loading effect has been used to reduce the
energy could be tuned by changing the collision position relative energy spread of the electron beam. It has been
in the plasma as shown on figure 9. demonstrated that there is an optimal load which flat-
1D Particle in cell (PIC) simulations have been used tened the electric field, accelerating all the electrons with
to model electron injection in the plasma wave at the the same value of the field, and producing consequently
collision of the two lasers, and their subsequent accelera- an electron beam with a very small, 1%, relative energy
tion. The PIC simulations have been compared to exist- spread65 . Thanks to the beam loading effect, the more
ing fluid models58 with prescribed electric field and they energetic electrons are slightly slowed down and acceler-
show major differences, such as the plasma fields behavior ated at the same energy that the slower one. In case of
and the amount of injected charge. The fluid approach lower charge, this effect doesn’t play any role and the en-
fails to describe qualitatively and quantitatively many ergy spread depends mainly of the heating volume. For
of the physical mechanisms that occur during and after higher current, the load is too high and the most en-
the laser beams collision60 . In this approach, the electron ergetic electrons slow down too much and get energies
beam charge has been found to be one order of magnitude even smaller that the slower one65 , increasing the rela-
greater than the one obtained in PIC simulations. For a tive energy spread. The optimal load was observed ex-
correct description of injection, one has to describe prop- perimentally and supported by full 3D PIC simulations,
erly (i) the heating process, e.g. kinetic effects and their its corresponds to a peak current in the 20-40 kA range.
 7

FIG. 8. Longitudinal electric field computed at t = 43f s in

1D PIC simulation (solid red line), and in fluid simulations
(dotted blue line). The transverse electric field is also repre-
sented (thin dotted line). Parameters are a0 = 2, a1 = 0.4,
30 fs is the laser pulse duration at FWHM with a wavelength
of 0.8 micron. The laser pulses propagate in a plasma with
electron density ne = 7 × 1018 cm−3 . In a) the case of parallel
polarization and in b) the case of crossed polarization.

The decelerating electric field due to the electron beam

was found to be in the GV/m/pC range.

D. Injection triggered by ionization

Another scheme has been proposed recently to con-

trol the injection by using a high Z gas and/or a high
Z-low Z gas mixture. Thanks to the large difference in
ionization potentials between successive ionization states
of trace atoms, with a single laser pulse, one can drive
relativistic plasma waves in the leading edge of the laser
pulse that ionizes easily the low energy level atoms, and
when the laser intensity is close to its maximum ionizes
the more internal level. Such related ionization trapping
mechanism has been first demonstrated in electron beam
driven plasma wave experiments on the Stanford Linear FIG. 9. a) Evolution of the electron beam peak energy
Collider (SLAC)66 . Trapped electrons from ionization of and its energy spread as function of collision position zinj
high Z wall ions from capillary walls was also inferred in for two parallel polarized laser beams. The electron beam
experiments on laser wakefield acceleration67 . In the case peak energy is shown in red, and the energy spread in blue.
of self guided laser driven wakefield, a mixture of helium Each point is an average of 35 shots and the error bars cor-
and trace amounts of different gases was used68,69 . In one respond to the standard deviation. The position zinj = 0
of these experiments, electrons from the K shell of nitro- corresponds to injection at the middle of the gas jet, whereas
gen were tunnel ionized near the peak of the laser pulse zinj = 500 µm corresponds to early injection close to the en-
trance of the gas jet, from J. Faure et al.57 . b) Evolution
and were injected into and trapped by the wake created
of charge (red solid line with squares), ∆E at FWHM (blue
by electrons from majority helium atoms and the L shell dotted line with circles) with the angle between the polar-
of nitrogen. Because of the relativistic self focusing effect, izations of injection and pump lasers ( 0◦ , parallel polariza-
the laser is propagating over a long distance with peak tions; 90◦ , crossed polarizations). a0 = 1.5, a1 = 0.4, 3 mm
intensity variations that can produce this injection over gas jet, ne = 5.7 × 1018 cm−3 , zcoll = −450 µm. c) decon-
a too long distance and in a non homogeneous way that volved spectra with a high resolution spectrometer measure-
does not allow to produce a low relative energy spread ment. Physical parameters: a0 = 1.2, a1 = 0.35, 3 mm gas
electron beam. Importantly, it has been shown that the jet, ne = 7.1 × 1018 cm−3 , zcoll = −300 µm, from C. Rechatin
required laser energy to trap electrons is reduced, rend- et al.64 .
ing this approach of great interest to produce electron
beams with a large charge at moderate laser energy. To
overcome the problem of delocalized injection over a long electron injection to a distinct short region, in the first
distance, as shown on figure 10, experiments using two short cell filled with gas mixture (the injector stage), en-
gas cells have been performed at LLNL70 . By restricting ergetic electron beams (of the order of 100 MeV) with a
 8

parameters61 . The exploration of new numerical schemes

is important to perform accurate simulations nearly free
from numerical noises to describe higher quality electron
beams with lower emittance values and with lower rel-
ative energy spread values. These improvements should
allow for example to validate theoretical work on two
stages laser plasma design that should allow the develop-
ment of few GeV electron beams71 with a small relative
energy spread and a good emittance and to prepare rele-
vant diagnostics. In parallel, theoretical and experimen-
tal research should of course be pursued to explore new
regimes and to validate theories and numerical codes.

FIG. 10. a) Schematic of the experimental setup showing

the laser beam, the two-stage gas cell, on left the injector part
and on right the accelerator part. b) Magnetically dispersed
electron beam images from a 4 mm injector-only gas cell (top)
and the 8 mm two-stage cell (bottom). From B. B. Pollock et
al.70 .

relatively large energy spread have been generated. Some

of these electrons are then further accelerated by a sec-
ond, longer accelerator stage composed by a longer cell
filled with low-Z gas, which increases their energy to 0.5
GeV while reducing the relative energy spread to < 5%
FWHM.

III. FUTURE OF THE LASER PLASMA

ACCELERATORS

The work achieved this last decade on the development

of laser plasma accelerator has been paved by many suc-
cessful (and unsuccessful) experiments. Thanks to these FIG. 11. a) injection principle in the colliding laser pulses:
pioneering works and judging from the incredible results in the “hot” injection scheme, injection is achieved thanks to
achieved recently, the time has come where technological momentum gain (black arrow) that electrons from the plasma
issues have to be addressed if one wants to build a reliable wave (green curve) reach during the collision that allow them
to cross the separatrix (blue curve). Injection principle in
machine that will operate in a robust way. The recent
the cold injection scheme: electrons are injected by being de-
diagnostics that have been developed and/or adapted to phased from the front of the main pulse to its back without
measure the laser, plasma and electron beam parameters momentum gain (purple arrow). After dephasing, electrons
have played a major role in understanding relativistic are effectively injected over the separatrix. b) Principle of
laser plasma interaction. These accurate measurements electrons dephasing in a standing wave generated by the col-
are crucial for performing numerical simulations with in- lision with a counter-propagating circular laser pulses. From
put parameters as close as possible to the experimental X. Davoine et al.72 .
ones. In some case, one to one simulations allows to
reproduce the measured electron beam parameters us- New ideas regarding injection control, such as the cold
ing 3D PIC simulations without having any adjustable injection scheme or assisted magnetic field scheme, are
 9

always welcome for future possible electron beam quality emittance values have been found to increase for lower
improvements. In the cold injection scheme72 , two laser electron beam energy values85 . This approach for mea-
pulses with circular polarization collide in the plasma and suring the emittance is particularly pertinent when the
produce a standing wave that freeze electrons at the colli- pepper pot technique stops to be relevant, for example for
sion point in a standing wave as shown on figure 11. After higher electron energies. Measuring in the betatron angu-
the collision, these electrons are accelerate in the plasma lar distribution the correlation between the ellipticity of
wave that is restored thank to the wakefield generate by the electron beam and the laser polarization, interaction
the pump beam. In this case, no heating is needed and of the accelerated electron bunch with the laser field88
electrons cross the separatrix because they keep a con- has been demonstrated, and has also probably been con-
stant longitudinal momentum. The use of an external firmed recently89 . For the longitudinal beam emittance,
magnetic field was recently proposed73 to control off-axis Coherent Transition Diagnostics have been used to mea-
injection bursts in laser driven plasma waves. It has been sure the shortest electron bunch of 1.5 fs RMS. Time re-
shown theoretically that this magnetic field can relax the solved magnetic field measurements have confirmed the
injection threshold and can be used to control the main shortness of the electron bunch produced in laser plasma
output beam features such as charge, energy, and trans- accelerators90 and have revealed interesting features of
verse dynamics in the ion channel associated with the laser plasma accelerating and focusing fields90,91 . Opti-
plasma blowout. cal transition radiation diagnostics has been very use-
ful to identify the existence of one or two92 , or even
In the near future, the development of compact free more electron bunches produced at different arches of the
electron lasers that could deliver an intense X ray beam plasma wakefield, and the possible interaction between
in a compact way by coupling the electron beam with un- the electron bunch and the laser field92 . The improve-
dulators. Thanks to the very high peak current of a few ment of the laser plasma interaction with the evolution
kA74 comparable to the current used at LCLS, the use of short-pulse laser technology, a field in rapid progress,
of laser plasma accelerators for free electron laser, the will still improve this new and very promising approach
so-called fifth generation light source, is clearly identi- which potential societal applications in material science
fied by the scientific community as a major development. for example for high resolution gamma radiography93,94 ,
For these applications, one has to reduce the relative en- for medicine for cancer treatment95,96 , chemistry97,98 and
ergy spread and to solve the problem of beam transport radiobiology99–101 . For longer term future, the ultimate
between the electron beam and the undulator by pre- goal which is of major interest for high energy physics
serving the parameters of the electron beam. Due to will require very high luminosity electron and positron
the large divergence of the electron beam that is proper beams having TeV energies. To reach these parameters
of laser plasma accelerators, one has to use for example with laser plasma accelerators will take at least 5 decades
ultra high magnetic gradient quadrupoles to reduce the
and significant further works to develop this technology is
temporal stretching of electrons that have a longer path
required. The incredible improvement in the energy and
that the on axis electrons. Undulators radiation75 and
beam quality of laser based plasma accelerators seems
synchrotron76 radiation have been recently obtained by
promising for high energy physics purposes. But electron
coupling an undulator with an electron beam from a laser
energy is not the only important parameter and it is also
plasma accelerator. In the near term future, alternative
necessary to consider the extremely high luminosity value
schemes to produce ultra short X ray beams, are also con-
required for this objective that has to be, for TeV beams,
sidered, such as Compton, betatron or Bremsstralhung
must be greater than 1034 cm2 s−1 . Reaching this value
X rays sources. Incredible progress have been made on will require at least to produce electron bunches at kHz
betatron radiation in a laser plasma accelerators, from repetition rates with 1 TeV in energy and with 1nC per
its first observation in 200477 and its first electron be- bunch. The corresponding average power of the electron
tatronic motion observation78 , number of articles have beam will be of about 1 MW. Assuming in the best case
reported in more details on this new source, such as its a coupling of 10 % from the laser to the electron beam
sub ps duration79 and its transverse size in the microm- (today in the best case this value rise 10 % (respectively
eter range80 . The betatron radiation has been used re- 1%) for a 10% (respectively 1%) relative energy spread
cently to perform high spatial resolution, of about 10 electron beam) one has to produce at least 10 MW (re-
microns, X ray contrast phase images in a single shot spectively 100MW) of photons. Since the laser wall-plug
mode operation81,82 . This radiation has allowed physi- efficiency is below 1%, one needs at least in the most fa-
cists to determine very subtile informations of electrons vorable case 10GW of electrical power to reach this goal.
injection in capillaries83 and electrons dynamic that can- The laser efficiency conversion could be increased up to
not be obtained with other diagnostic84 . Measuring the 50% by using diode pumped systems, thus reducing the
angular and the energy spectra of this radiation gives an needed power to 0.2 GW. These considerations were done
alternative method for estimating the transverse electron neglecting several other issues such as the propagation of
beam emittance and for confirming previous measure- electron beams into a plasma medium, laser plasma cou-
ments using the pepper pot technique. The typical value pling problems, laser depletion, emittance requirements
of the normalized transverse emittance is found to be in and others102 .
the few π.mm.mrad for tens MeV electrons85–87 . The
 10

Nevertheless, before reaching an objective and more ac- and former post-docs J. Lim, A. Lifschitz, O. Lundh,
curate conclusion on the relevance of the laser plasma and B. Prithviraj, my co-worker I. Ben-Ismail, S. Corde,
approach for high energy physics, it will be necessary to E. Lefebvre, A. Rousse, A. Specka, K. Ta Phuoc, and
design a prototype machine (including several modules) C. Thaury, who have largely contributed during this last
in coordination with accelerator physicists. An estima- decade to the progress done in laser plasma accelerators
tion of the cost and an identification of all the technical research at LOA. I acknowledge the different teams from
problems that are to be solved will permit an estimate of APRI, CUOS, IC, JAEA, LLNL, LLC, LBNL, MPQ,
the risk with respect to other approaches (particle beam RAL, UCLA and others, that have in a competitive and
interaction in plasma medium, hot or cold technology, fair atmosphere made significant progresses sharing with
or others). In conclusion while a significant amount of passion this wonderful adventure. I also acknowledge the
work remains to be done to deliver beams of interest for support of the European Research Council for funding
high energy physics, the control of the electron beam the PARIS ERC project (contract number 226424).
parameters is now achieved and many of the promised
applications are become a reality.

ACKNOWLEDGEMENTS

I acknowledge warmly my former PhD students X.

Davoine, J. Faure, S. Fritzler, Y. Glinec, C. Rechatin,

[1] C. Joshi, Phys. Plasmas 14, 055501 (2007). Rev. Lett. 78, 3463 (1997).
[2] E. Esarey, C. B. Schroeder, and W. P. Leemans, Rev. [14] C. W. Siders, S. P. Le Blanc, D. Fisher, T. Tajima,
Mod. Physics 81, 1229 (2009). M. C. Downer, A. Babine, A. Stepanov, and A. Sergeev,
[3] T. Tajima and J. M. Dawson, Phys. Rev. Lett. 43, 267 Phys. Rev. Lett. 76, 3570 (1996).
(1979). [15] N. H. Matlis, S. Reed, S. S. Bulanov, V. Chvykov,
[4] C. E. Clayton, C. Joshi, C. Darrow, and D. Umstadter, G. Kalintchenko, T. Matsuoka, P. Rousseau,
Phys. Rev. Lett. 54, 2343 (1985). V. Yanovsky, A. Maksimchuk, S. Kalmykov, G. Shvets,
[5] C. E. Clayton, K. A. Marsh, A. Dyson, M. Everett, and M. Downer, Nature Physics 2, 749 (2006).
A. Lal, W. P. Leemans, R. Williams, and C. Joshi, [16] A. Modena, A. Dangor, Z. Najmudin, C. Clayton,
Phys. Rev. Lett. 70, 37 (1993). K. Marsh, C. Joshi, V. Malka, C. Darrow, D. Neely,
[6] M. Everett, A. Lal, D. Gordon, C. Clayton, K. Marsh, and F. Walsh, Nature 377, 606 (1995).
and C. Joshi, Nature 368, 527 (1994). [17] N. E. Andreev, L. M. Gorbunov, V. I. Kirsanov, A. A.
[7] F. Amiranoff, D. Bernard, B. Cros, F. Jacquet, Pogosova, and R. R. Ramazashvili, JETP Lett 55, 571
G. Matthieussent, P. Miné, P. Mora, J. Morillo, (1992).
F. Moulin, A. E. Specka, and C. Stenz, Phys. Rev. [18] P. Mora, Phys. Fluids B 4, 1630 (1992).
Lett. 74, 5220 (1995). [19] P. Sprangle and E. Esarey, Phys. Fluids B 4, 2241
[8] Y. Kitagawa, T. Matsumoto, T. Minamihata, K. Sawai, (1992).
K. Matsuo, K. Mima, K. Nishihara, H. Azechi, K. A. [20] C. Joshi, T. Tajima, J. M. Dawson, H. A. Baldis, and
Tanaka, H. Takabe, and S. Nakai, Phys. Rev. Lett. 68, N. A. Ebrahim, Phys. Rev. Lett. 47, 1285 (1981).
48 (1992). [21] W. B. Mori, C. D. Decker, D. E. Hinkel, and T. Kat-
[9] A. Dyson, A. Dangor, A. K. L. Dymoke-Bradshaw, souleas, Phys. Rev. Lett. 72, 1482 (1994).
T. Ashfar-Rad, P. Gibbon, A. R. Bell, C. N. Danson, [22] D. Umstadter, S.-Y. Chen, A. Maksimchuk, G. Mourou,
C. B. Edwards, F. Amiranoff, G. Matthieusent, S. J. and R. Wagner, Science 273, 472 (1996).
Karttunen, and R. R. E. Salomaa, Plasma Physics and [23] C. I. Moore, A. Ting, K. Krushelnick, E. Esarey, R. F.
Controlled Fusion 38, 505 (1996). Hubbard, B. Hafizi, H. R. Burris, C. Manka, and
[10] N. A. Ebrahim, Journal of Applied Physics 76, 7645 P. Sprangle, Phys. Rev. Lett. 79, 3909 (1997).
(1994). [24] C. E. Clayton, K.-C. Tzeng, D. Gordon, P. Muggli,
[11] F. Amiranoff, S. Baton, D. Bernard, B. Cros, W. B. Mori, C. Joshi, V. Malka, Z. Najmudin, A. Mod-
D. Descamps, F. Dorchies, F. Jacquet, V. Malka, ena, D. Neely, and A. E. Dangor, Phys. Rev. Lett. 81,
G. Matthieussent, J. R. Marquès, P. Miné, A. Modena, 100 (1998).
P. Mora, J. Morillo, and Z. Najmudin, Phys. Rev. Lett. [25] C. Gahn, G. D. Tsakiris, A. Pukhov, J. Meyer-ter-Vehn,
81, 995 (1998). G. Pretzler, P. Thirolf, D. Habs, and K. J. Witte, Phys.
[12] J. R. Marquès, J. P. Geindre, F. Amiranoff, P. Aude- Rev. Lett. 83, 4772 (1999).
bert, J. C. Gauthier, A. Antonetti, and G. Grillon, [26] V. Malka, J. Faure, J.-R. Marquès, F. Amiranoff, J.-
Phys. Rev. Lett. 76, 3566 (1996). P. Rousseau, S. Ranc, J.-P. Chambaret, Z. Najmudin,
[13] J. R. Marquès, F. Dorchies, J. P. Geindre, F. Amira- B. Walton, P. Mora, and A. Solodov, Phys. Plasmas 8
noff, J. C. Gauthier, G. Hammoniaux, A. Antonetti, (2001).
P. Chessa, P. Mora, and T. M. Antonsen, Jr., Phys.
 11

[27] A. Pukhov and J. Meyer-ter-Vehn, Appl. Phys. B 74, 71, 6 (2000).

355 (2002). [45] S. Semushin and V. Malka, Rev. Sci. Instrum. 72, 2961
[28] W. Lu, C. Huang, M. Zhou, W. B. Mori, and T. Kat- (2001).
souleas, Phys. Rev. Lett. 96, 165002 (2006). [46] S. Bulanov, N. Naumova, F. Pegoraro, and J. Sakai,
[29] S. V. Bulanov, F. Pegoraro, A. M. Pukhov, and A. S. Phys. Rev. E 58, R5257 (1998).
Sakharov, Phys. Rev. Lett. 78, 4205 (1997). [47] C. G. R. Geddes, K. Nakamura, G. R. Plateau, C. Tóth,
[30] S. Mangles, C. D. Murphy, Z. Najmudin, A. G. R. E. Cormier-Michel, E. Esarey, C. B. Schroeder, J. R.
Thomas, J. L. Collier, A. E. Dangor, A. J. Divall, P. S. Cary, and W. P. Leemans, Phys. Rev. Lett. 100, 215004
Foster, J. G. Gallacher, C. J. Hooker, D. A. Jaroszynski, (2008).
A. J. Langley, W. B. Mori, P. A. Nooreys, R. Viskup, [48] J. U. Kim, N. Hafz, and H. Suk, Phys. Rev. E 69,
B. R. Walton, and K. Krushelnick, Nature 431, 535 026409 (2004).
(2004). [49] T.-Y. Chien, C.-L. Chang, C.-H. Lee, J.-Y. Lin,
[31] C. G. R. Geddes, C. Tóth, J. van Tilborg, E. Esarey, J. Wang, and S.-Y. Chen, Phys. Rev. Lett. 94, 115003
C. B. Schroeder, D. Bruhwiler, C. Nieter, J. Cary, and (2005).
W. P. Leemans, Nature 431, 538 (2004). [50] P. Tomassini, M. Galimberti, A. Giulietti, D. Giulietti,
[32] J. Faure, Y. Glinec, A. Pukhov, S. Kiselev, S. Gor- L. A. Gizzi, L. Labate, and F. Pegoraro, Phys. Rev. ST
dienko, E. Lefebvre, J.-P. Rousseau, F. Burgy, and Accel. Beams 6, 121301 (2003).
V. Malka, Nature 431, 541 (2004). [51] J. Faure, C. Rechatin, O. Lundh, L. Ammoura, and
[33] F. S. Tsung, R. Narang, W. B. Mori, C. Joshi, R. A. V. Malka, Physics of Plasmas 17, 083107 (2010).
Fonseca, and L. Silva, Phys. Rev. Lett. 93, 185002 [52] A. V. Brantov, T. Z. Esirkepov, M. Kando, H. Kotaki,
(2004). V. Y. Bychenkov, and S. V. Bulanov, Physics of Plas-
[34] E. Miura, K. Koyama, S. Kato, S. Saito, M. Adachi, mas 15, 073111 (2008).
Y. Kawada, T. Nakamura, and M. Tanimoto, Appl. [53] A. J. Gonsalves, K. Nakamura, C. Lin, D. Panasenko,
Phys. Lett. 86, 251501 (2005). S. Shiraishi, T. Sokollik, C. Benedetti, C. B. Schroeder,
[35] C.-T. Hsieh, C.-M. Huang, C.-L. Chang, Y.-C. Ho, Y.- C. G. R. Geddes, J. van Tilborg, J. Osterhoff, E. Esarey,
S. Chen, J.-Y. Lin, J. Wang, and S.-Y. Chen, Phys. C. Toth, and W. P. Leemans, Nature Physics 7, 862
Rev. Lett. 96, 095001 (2006). (2011).
[36] B. Hidding, K.-U. Amthor, B. Liesfeld, H. Schwoerer, [54] H. Suk, N. Barov, J. B. Rosenzweig, and E. Esarey,
S. Karsch, M. Geissler, L. Veisz, K. Schmid, J. G. Gal- Phys. Rev. Lett. 86, 1011 (2001).
lacher, S. P. Jamison, D. Jaroszynski, G. Pretzler, and [55] K. Koyama, A. Yamazaki, A. Maekawa, M. Uesaka,
R. Sauerbrey, Phys. Rev. Lett. 96, 105004 (2006). T. Hosokai, M. Miyashita, S. Masuda, and E. Miura,
[37] J. Osterhoff, A. Popp, Z. Major, B. Marx, T. P. Nuclear Instruments and Methods in Physics Research
Rowlands-Rees, M. Fuchs, M. Geissler, R. Hörlein, Section A: Accelerators, Spectrometers, Detectors and
B. Hidding, S. Becker, E. A. Peralta, U. Schramm, Associated Equipment 608, S51 (2009).
F. Grüner, D. Habs, F. Krausz, S. M. Hooker, and [56] K. Schmid, A. Buck, C. M. S. Sears, J. M. Mikhailova,
S. Karsch, Phys. Rev. Lett. 101, 085002 (2008). R. Tautz, D. Herrmann, M. Geissler, F. Krausz, and
[38] S. P. D. Mangles, A. G. R. Thomas, O. Lundh, F. Lin- L. Veisz, Phys. Rev. ST Accel. Beams 13, 091301 (2010).
dau, M. C. Kaluza, A. Persson, C.-G. Wahlström, [57] J. Faure, C. Rechatin, A. Norlin, A. Lifschitz, Y. Glinec,
K. Krushelnick, and Z. Najmudin, Phys. Plasmas 14, and V. Malka, Nature 444, 737 (2006).
056702 (2007). [58] E. Esarey, P. Sprangle, J. Krall, and A. Ting, IEEE
[39] N. Hafz, T. M. Jeong, I. W. Choi, S. K. Lee, K. H. Trans. Plasma Sci. 24, 252 (1996).
Pae, V. V. Kulagin, J. H. Sung, T. J. Yu, K.-H. Hong, [59] G. Fubiani, E. Esarey, C. Schroeder, and W. Leemans,
T. Hosokai, J. R. Cary, D.-K. Ko, and J. Lee, Nature Phys. Rev. E 70, 016402 (2004).
Photonics 2, 571 (2008). [60] C. Rechatin, J. Faure, A. Lifschitz, V. Malka, and
[40] S. Kneip, S. R. Nagel, S. F. Martins, S. P. D. Man- E. Lefebvre, Phys. Plasmas 14, 060702 (2007).
gles, C. Bellei, O. Chekhlov, R. J. Clarke, N. Delerue, [61] X. Davoine, E. Lefebvre, J. Faure, C. Rechatin, A. Lif-
E. J. Divall, G. Doucas, K. Ertel, F. Fiuza, R. Fonseca, schitz, and V. Malka, Physics of Plasmas 15, 113102
P. Foster, S. J. Hawkes, C. J. Hooker, K. Krushelnick, (2008).
W. B. Mori, C. A. J. Palmer, K. T. Phuoc, P. P. Ra- [62] V. Malka, J. Faure, C. Rechatin, A. Ben-Ismail, J. K.
jeev, J. Schreiber, M. J. V. Streeter, D. Urner, J. Vieira, Lim, X. Davoine, and E. Lefebvre, Phys. Plasmas 16,
L. O. Silva, and Z. Najmudin, Phys. Rev. Lett. 103, 056703 (2009).
035002 (2009). [63] C. Rechatin, J. Faure, A. Lifschitz, X. Davoine,
[41] W. P. Leemans, B. Nagler, A. J. Gonsalves, C. Tòth, E. Lefebvre, and V. Malka, New Journal of Physics
K. Nakamura, C. G. R. Geddes, E. Esarey, C. B. 11, 013011 (2009).
Schroeder, and S. M. Hooker, Nature Physics 2, 696 [64] C. Rechatin, J. Faure, A. Ben-Ismail, J. Lim, R. Fitour,
(2006). A. Specka, H. Videau, A. Tafzi, F. Burgy, and V. Malka,
[42] J. Faure, Y. Glinec, J. J. Santos, F. Ewald, J.-P. Physical review letters 102, 164801 (2009).
Rousseau, S. Kiselev, A. Pukhov, T. Hosokai, and [65] C. Rechatin, X. Davoine, A. Lifschitz, A. Ismail, J. Lim,
V. Malka, Phys. Rev. Lett. 95, 205003 (2005). E. Lefebvre, J. Faure, and V. Malka, Physical review
[43] M. Pittman, S. Ferre, J.-P. Rousseau, L. Notebaert, J.- letters 103, 194804 (2009).
P. Chambaret, and G. Cheriaux, Appl. Phys. B 74, 529 [66] E. Oz, S. Deng, T. Katsouleas, P. Muggli, C. D. Barnes,
(2002). I. Blumenfeld, F. J. Decker, P. Emma, M. J. Hogan,
[44] V. Malka, C. Coulaud, J. P. Geindre, V. Lopez, Z. Naj- R. Ischebeck, R. H. Iverson, N. Kirby, P. Krejcik,
mudin, D. Neely, and F. Amiranoff, Rev. Sci. Instrum. C. O’Connell, R. H. Siemann, D. Walz, D. Auerbach,
 12

C. E. Clayton, C. Huang, D. K. Johnson, C. Joshi, C. A. J. Palmer, J. Schreiber, W. Schumaker, A. G. R.

W. Lu, K. A. Marsh, W. B. Mori, and M. Zhou, Phys. Thomas, and V. Yanovsky, Applied Physics Letters 99,
Rev. Lett. 98, 084801 (2007). 093701 (2011).
[67] T. P. Rowlands-Rees, C. Kamperidis, S. Kneip, [83] G. Genoud, K. Cassou, F. Wojda, H. Ferrari, C. Kam-
A. J. Gonsalves, S. P. D. Mangles, J. G. Gallacher, peridis, M. Burza, A. Persson, J. Uhlig, S. Kneip,
E. Brunetti, T. Ibbotson, C. D. Murphy, P. S. Fos- S. Mangles, A. Lifschitz, B. Cros, and C.-G. Wahlstrm,
ter, M. J. V. Streeter, F. Budde, P. A. Norreys, D. A. Applied Physics B: Lasers and Optics 105, 309 (2011).
Jaroszynski, K. Krushelnick, Z. Najmudin, and S. M. [84] S. Corde, C. Thaury, K. T. Phuoc, A. Lifschitz, G. Lam-
Hooker, Phys. Rev. Lett. 100, 105005 (2008). bert, J. Faure, O. Lundh, E. Benveniste, A. Ben-Ismail,
[68] A. Pak, K. A. Marsh, S. F. Martins, W. Lu, W. B. Mori, L. Arantchuk, A. Marciniak, A. Stordeur, P. Brijesh,
and C. Joshi, Phys. Rev. Lett. 104, 025003 (2010). A. Rousse, A. Specka, and V. Malka, Phys. Rev. Lett.
[69] C. McGuffey, A. G. R. Thomas, W. Schumaker, T. Mat- 107, 215004 (2011).
suoka, V. Chvykov, F. J. Dollar, G. Kalintchenko, [85] S. Fritzler, E. Lefebvre, V. Malka, F. Burgy, A. E. Dan-
V. Yanovsky, A. Maksimchuk, K. Krushelnick, V. Y. gor, K. Krushelnick, S. P. D. Mangles, Z. Najmudin,
Bychenkov, I. V. Glazyrin, and A. V. Karpeev, Phys. J.-P. Rousseau, and B. Walton, Phys. Rev. Lett. 92,
Rev. Lett. 104, 025004 (2010). 165006 (2004).
[70] B. B. Pollock, C. E. Clayton, J. E. Ralph, F. Albert, [86] C. M. S. Sears, A. Buck, K. Schmid, J. Mikhailova,
A. Davidson, L. Divol, C. Filip, S. H. Glenzer, K. Her- F. Krausz, and L. Veisz, Phys. Rev. ST Accel. Beams
poldt, W. Lu, K. A. Marsh, J. Meinecke, W. B. Mori, 13, 092803 (2010).
A. Pak, T. C. Rensink, J. S. Ross, J. Shaw, G. R. Ty- [87] E. Brunetti, R. P. Shanks, G. G. Manahan, M. R. Is-
nan, C. Joshi, and D. H. Froula, Phys. Rev. Lett. 107, lam, B. Ersfeld, M. P. Anania, S. Cipiccia, R. C. Issac,
045001 (2011). G. Raj, G. Vieux, G. H. Welsh, S. M. Wiggins, and
[71] V. Malka, A. Lifschitz, J. Faure, and Y. Glinec, Phys- D. A. Jaroszynski, Phys. Rev. Lett. 105, 215007 (2010).
ical Review Special Topics-Accelerators and Beams 9, [88] S. P. D. Mangles, A. G. R. Thomas, M. C. Kaluza,
091301 (2006). O. Lundh, F. Lindau, A. Persson, F. S. Tsung, Z. Naj-
[72] X. Davoine, A. Beck, A. Lifschitz, V. Malka, and mudin, W. B. Mori, C.-G. Wahlström, and K. Krushel-
E. Lefebvre, New Journal of Physics 12, 095010 (2010). nick, Phys. Rev. Lett. 96, 215001 (2006).
[73] J. Vieira, S. F. Martins, V. B. Pathak, R. A. Fonseca, [89] S. Cipiccia, M. R. Islam, B. Ersfeld, R. P. Shanks,
W. B. Mori, and L. O. Silva, Phys. Rev. Lett. 106, E. Brunetti, G. Vieux, X. Yang, R. C. Issac, S. M. Wig-
225001 (2011). gins, G. H. Welsh, M.-P. Anania, D. Maneuski, R. Mont-
[74] O. Lundh, J. Lim, C. Rechatin, L. Ammoura, A. Ben- gomery, G. Smith, M. Hoek, D. J. Hamilton, N. R. C.
Ismaı̈l, X. Davoine, G. Gallot, J. Goddet, E. Lefebvre, Lemos, D. Symes, P. P. Rajeev, V. O. Shea, J. M. Dias,
V. Malka, and J. Faure, Nature Physics 7, 219 (2011). and D. A. Jaroszynski, Nature Physics 7, 867 (2011).
[75] M. Fuchs, R. Weingartner, A. Popp, Z. Major, [90] A. Buck, M. Nicolai, K. Schmid, C. M. S. Sears,
S. Becker, J. Osterhoff, I. Cortrie, B. Z. R. Horlein, A. Svert, J. M. Mikhailova, F. Krausz, M. C. Kaluza,
G. D. Tsakiris, U. Schramm, T. P. Rowlands-Rees, S. M. and L. Veisz, Nautre Physics 7, 543 (2011).
Hooker, D. Habs, F. Krausz, and F. Gruner, Nature [91] M. C. Kaluza, H.-P. Schlenvoigt, S. P. D. Mangles,
Physics 5, 826 (2009). A. G. R. Thomas, A. E. Dangor, H. Schwoerer, W. B.
[76] H.-P. Schlenvoigt, K. Haupt, A. Debus, F. Budde, Mori, Z. Najmudin, and K. M. Krushelnick, Phys. Rev.
O. Jäckel, S. Pfotenhauer, H. Schwoerer, E. Rohwer, Lett. 105, 115002 (2010).
J. G. Gallacher, E. Brunetti, R. P. Shanks, S. M. Wig- [92] Y. Glinec, J. Faure, A. Norlin, A. Pukhov, and
gins, and D. A. Jaroszynski, Nature Physics 4, 130 V. Malka, Phys. Rev. Lett. 98, 194801 (2007).
(2008). [93] Y. Glinec, J. Faure, L. L. Dain, S. Darbon, T. Hosokai,
[77] A. Rousse, K. Ta Phuoc, R. Shah, A. Pukhov, E. Lefeb- J. J. Santos, E. Lefebvre, J. P. Rousseau, F. Burgy,
vre, V. Malka, S. Kiselev, F. Burgy, J.-P. Rousseau, B. Mercier, and V. Malka, Phys. Rev. Lett. 94, 025003
D. Umstadter, and D. Hulin, Phys. Rev. Lett. 93, (2005).
135005 (2004). [94] A. Ben-Ismail, O. Lundh, C. Rechatin, J. K. Lim,
[78] Y. Glinec, J. Faure, A. Lifschitz, J. M. Vieira, R. A. J. Faure, S. Corde, and V. Malka, App. Phys. Lett.
Fonseca, L. O. Silva, and V. Malka, European physics 98, 264101 (2011).
letters 81, 64001 (2008). [95] Y. Glinec, J. Faure, V. Malka, T. Fuchs, H. Szy-
[79] K. T. Phuoc, R. Fitour, A. Tafzi, T. Garl, N. Artemiev, manowski, and U. Oelfke, Med. Phys. 33, 155 (2006).
R. Shah, F. Albert, D. Boschetto, A. Rousse, D.-E. Kim, [96] T. Fuchs, H. Szymanowski, U. Oelfke, Y. Glinec,
A. Pukhov, V. Seredov, and I. Kostyukov, Phys. Plas- C. Rechatin, J. Faure, and V. Malka, Phys. Med. Biol.
mas 14, 080701 (2007). 54, 3315 (2009).
[80] K. Ta Phuoc, S. Corde, R. Shah, F. Albert, R. Fitour, [97] B. Brozek-Pluska, D. Gliger, A. Hallou, V. Malka,
J.-P. Rousseau, F. Burgy, B. Mercier, and A. Rousse, and Y. A. Gauduel, Radiation and Chemistery 72, 149
Phys. Rev. Lett. 97, 225002 (2006). (2005).
[81] S. Fourmaux, S. Corde, K. T. Phuoc, P. Lassonde, [98] Y. Gauduel, Y. Glinec, J.-P. Rousseau, F. Burgy, and
G. Lebrun, S. Payeur, F. Martin, S. Sebban, V. Malka, V. Malka, Eur. Phys. J. D 60, 121 (2010).
A. Rousse, and J. C. Kieffer, Optics letters 36, 2426 [99] V. Malka, J. Faure, and Y. A. Gauduel, Mutation Re-
(2011). search 704, 142 (2010).
[82] S. Kneip, C. McGuffey, F. Dollar, M. S. Bloom, [100] O. Rigaud, N. O. Fortunel, P. Vaigot, E. Cadio, M. T.
V. Chvykov, G. Kalintchenko, K. Krushelnick, A. Mak- Martin, O. Lundh, J. Faure, C. Rechatin, V. Malka,
simchuk, S. P. D. Mangles, T. Matsuoka, Z. Najmudin, and Y. A. Gauduel, Cell Death and Disease 1 (2010).
 13

[101] V. Malka, J. Faure, Y. Gauduel, E. Lefebvre, A. Rousse,

and K. Phuoc, Nature Physics 4, 447 (2008).
[102] K. Krushelnick and V. Malka, Laser and Photonics Re-
view 1, 1 (2009).