Mourou Lecture

96
Extreme Light Physics

and Application
Nobel Lecture, December 8, 2018 by
Gérard Mourou
University of Michigan, Ann Arbor, MI, USA & École Polytechnique,
Palaiseau, France.
I. INTRODUCTION
The advent of ultraintense laser pulses generated by the technique of
chirped pulse amplification (CPA), along with the development of
high-fluence laser materials has opened up entirely new fields of optics. A
CPA laser exhibits stunning capabilities. It can generate the largest field,
the largest pressure, the highest temperature and accelerating field, mak-
ing it a universal source of high energy particles and radiation.
CPA technology produces a wide range of intensities extending from
10 to 1025 W/cm2. In the lower part of this range, the intensity regimes of
14
1014 to 1017 have applications that include micromachining, which can be

performed on material regardless of its nature, i.e. ceramic, metal, biologi-
cal tissue, cornea, etc. Extremely clean cuts of minimal roughness even at
the atomic scale are produced. This attractive property led us to applica-
tions in ophthalmic procedures like refractive surgery, cataract surgery,
corneal transplants and glaucoma treatment. Today, a million patients a
year are benefiting from femtosecond interventions. In science, in the
same intensity level, CPA makes possible to reach the attosecond frontier,
offering a formidable tool to time resolved fundamental electronic pro-
cesses.
Gérard Mourou Lecture 97
For intensities >1018 W/cm2 laser-matter interaction becomes strongly

dominated by the relativistic character of the electron. In contrast to the
nonrelativistic regime, the laser field moves matter more effectively,
including motion in the direction of laser propagation, nonlinear modula-
tion and harmonic generation, leading to high energy particle and radia-
tion production. One of the hallmarks of this regime is Laser Wakefield
Acceleration (LWA), where the electromagnetic energy from a laser pulse
is transformed into kinetic energy producing accelerating gradients a
thousand times higher than those applied in conventional accelerators.
The electron beam can, in turn, produce a copious amount of keV radia-
tion by betatron or Compton scattering.
For intensities at 1025 W/cm2 the laser field becomes so large that pro-
tons and ions become relativist with GeV energies. The acceleration is
directly produced by the light pressure. The source size is very small, and
the large acceleration gradients combine to make this source brightness
better than any existing.
The coupling of an intense laser field to matter also has implications
for the study of the highest energies in astrophysics, such as ultrahigh-en-
ergy cosmic rays with energies in excess of 1020 eV. Intense laser fields
can also produce an accelerating field sufficient to simulate general rela-
tivistic effects in the laboratory via the equivalence principle like the loss
of information in Black Holes.
Many CPA applications offer great benefit to humankind. We discuss a
few examples capitalising on the compactness of the CPA-based source.
For instance, the generation of high energy protons and neutrons applied
to the treatment of cancer, i.e proton therapy, or in nuclear pharmacology
where short-lived radionuclides could now be created for therapy or diag-
nostics near the patient’s bed.
In the environmental arena, owing to the efficient generation of high
energy neutrons, it is one of our goals to use them to shorten the degree
of radiotoxicity of the most dangerous elements, the minor actinides in
nuclear waste.
Finally, looking onwards, particle production in “empty” space will
remain one of the main objectives of the field. It is the historical path which
guided the field in order to acquire an understanding of fundamental ques-
tions on the structure of vacuums, to give us a glimpse of the propagation
of light in vacuums and how it defines the mass of all elementary particles.
A novel laser architecture to reach intensities at the Schwinger level is
explored. Its paradigm is based on the transition of a single cycle visible
light pulse into a high energy single cycle X-ray pulse. Successful transpo-
sition would give the field a formidable boost, equivalent to the one
received when maser transitioned into laser, moving from GHz to PHz
(Light) frequencies.
98 THE NOBEL PRIZES
II. THE CPA TECHNIQUE

The key to high and ultrahigh peak power and intensity is the amplifica-
tion of ultrashort pulses in the picosecond and femtosecond time scales.
Over the past 40 years laser-pulse durations have continuously decreased
from the microsecond domain with free running to the nanosecond
regime with Q-switching, and finally to the picosecond and few-femtosec-
ond regime with mode locking (Brabec and Krausz, 2000). With the
advent of mode locking, the laser-pulse duration became so short that
pulses could not be amplified without producing unwanted nonlinear
effects. This caused a power and intensity plateau seen in Fig. 1. Before
1985, for reasonable sized systems, i.e., with a beam diameter of the order
of 1 cm, the maximum obtainable power remained around 1GW, with
focused intensities of about 1014 W/cm2. Higher power could be obtained
through the use of amplifying media with gain bandwidths that can
accommodate the short pulse spectrum and high-energy storage media
that have a small transition cross section σa. However, to be efficient, this
approach requires the use of input pulses with a high laser fluence of a
few J/cm2. As we shall see later, good energy extraction from an amplifier
calls for input pulses close to the saturation fluence, see Eq. 1. This level
of fluence delivered over a short time leads to prohibitively large intensi-
ties, in excess of TW/cm2. This is far above the limit of the GW/cm2
imposed by the need to prevent nonlinear effects and optical damage in
the amplifiers and optical components. As a result, the only alternative
seemed to use low-energy storage materials such as dyes or excimers and
increase the laser beam cross section. The latter is unattractive as it leads
to large, low-repetition rate and costly laser systems. Therefore, high-in-
tensity physics research was limited to a few large facilities such as the
CO2 laser at Los Alamos National Laboratory (Carman et al., 1981), the
Nd:glass laser at the Laboratory for Laser Energetics (Bunkenburg et al.,
1981), and excimer lasers at the University of Illinois at Chicago and Uni-
versity of Tokyo (Luk et al., 1989; Endoh et al., 1989).
In 1985 laser physicists at the University of Rochester (Strickland and
Mourou, 1985) and later (Maine and Mourou, 1988; Maine et al., 1988),
demonstrated a way to simultaneously accommodate the very large beam
fluence necessary for energy extraction in superior storage materials
while keeping the intensity and nonlinear effects to an acceptable level.
This technique was called Chirped Pulsed Amplification (CPA). CPA revo-
lutionised the field in three ways. First, a TW system could fit on a table-
top, thus delivering intensities 105–106 times higher utilising conventional
technology. Second, the CPA architecture could be easily retrofitted to
existing large laser fusion systems at relatively low cost. Today CPA is
incorporated in all the major laser fusion systems around the world such
as: Japan (Yamakawa et al., 1991), France (Rouyer et al., 1993), United
Figure 1. Laser intensity vs years.
Kingdom Vulcan Laser, United States Perry et al., 1999, etc. Third,
because of their reduced size CPA lasers could be combined with large
particle accelerators. In the case of synchrotrons (Wulff et al., 1997; Lars-
son et al., 1998; Schoenlein et al., 2000), they could be used to study
time-resolved x-ray diffraction. With a linear collider such as SLAC one
could produce fields higher than the critical field (Bula et al., 1996) and
observe nonlinear QED effects such as pair generation from a vacuum. At
present all the colliders are considering the incorporation of CPA technol-
ogy to produce γ rays for photon-photon collisions to produce a γ-γ col-
lider, (Telnov, 1990, 2000, 2001; Yokoya, 2000).
As we shall describe later, the availability of ultra-high intensity lasers
has extended the horizon of laser physics from atomic and con-
densed-matter studies to plasma, nuclear, and high-energy physics, gen-
eral relativity, cosmology, and physics beyond the standard model. It has
also had a major effect in bringing back to university laboratories science
that formerly could only be studied with large-scale facilities. Also at a
relatively low peak power level, it made precision machining and ophthal-
mic procedures in the medical arena possible.
A. Amplification: The Energy Extraction Condition

Before 1985 all amplifier systems were based on direct amplification. As
mentioned in the introduction, for extraction efficiency, a simple rule is
that the energy per unit area be of the order of Fsat, the saturation fluence
of the materials. This value is given by

100 THE NOBEL PRIZES
(1)
where ћ is Planck’s constant, ω is the angular laser frequency and σa is

the amplifying transition cross section. Fsat is 0.9 J/cm2 for Ti:sapphire
and 4 J/cm2 for Nd:glass and of the order of a mJ/cm2 for dyes and exci-
mers. It can be shown (Siegman, 1986) the output fluence Fout is given by
(2)
where G0 is the low signal gain and

(3)

(4)
The gain Gf at the end of the impulsion is given by

(5)
From Eqs. 4 and 5 we see that, to reach an efficiency close to unity, the
laser input fluence Fpulse must correspond to few times Fsat. Fig. 2 illus-
trates this point for two different initial gains G0 of 10 and 103.
Figure 2. Ampli-
fier efficiency.
This illustrates
the importance
for the input
pulse fluence
Fpulse to be few
times the sa-
turation fluence
Fsat to obtain a
good extraction
efficiency.
B. Amplification: The Propagation Condition

Prior to CPA the amplifying media were exclusively dyes, (Migus et al.,
1982) and excimers (Endoh et al., 1989; Luk et al., 1989). Typical cross
sections for these media are very large, in the range of 10−16 cm2, implying
a Fsat of only a few mJ/cm2, or a power density of 1 GW/cm2 for subpico-
second pulses. At this power density level, the index of refraction
becomes intensity dependent according to the well-known expression

(6)
Due to the spatial variation of the laser beam intensity, this will modify
the beam wavefront according to the B integral given by
(7)
Here, B represents, in units of λ, the amount of wavefront distortion due

to the intensity-dependent index of refraction, accumulated by the beam
over a length L. For a perfectly Gaussian beam, B will cause the whole
beam to self-focus above a critical power given by
(8)
For a nonlinear index n2=5×10−16cm2/W for Ti:sapphire. When the laser

beam exhibits spatial intensity modulations, n2 will cause the beam to
break up in filaments. In practice the small scale self-focusing represents
the most severe problem in an amplifier system. The maximum growth
rate gm (Bespalov and Talanov, 1966) will occur for spatial frequencies Km
given by
(9)

(10)
For intensities of the order of I = 1 GW/cm2 in Ti:sapphire, Km=200 cm−1,

corresponding to 50 μm spatial wavelength. As seen in Fig. 3, these wave-
front “irregularities” will grow at a rate of gm= 3 cm−1 with an exponential
growth rate Gm over the gain length L exactly equal to B,
Gm = B
For laser fusion, the beam is “cleaned” with spatial filters every time B
reaches 3. For high-field experiments in which the spatial and temporal
beam quality requirements are more stringent, B must be kept below 0.3
corresponding to a wavefront distortion of λ/20.
Figure 3. The intensity dependence of the Index of refraction, will create spatial variation
of the laser beam intensity, producing undesirable filaments and hot spot.
C. The CPA Concept

We have seen that amplifying media with low cross sections offer the
benefit of compactness. For instance, Nd:glass has a cross section of 10–21
cm2, which means that we can store 103 to 104 times more atoms per unit
volume and consequently, get 103 to 104 more energy before it self-oscil-
lates, than with a dye or excimer with cross section 1016 cm2. However, to
extract all the energy in a picosecond pulse would require a beam with a
fluence Fs of the order of 1 J/cm2 or an intensity of 1012 W/cm2 correspond-
ing to a B of 103, or 103 times the acceptable value.
Therefore, in order to utilise superior energy storage materials, the
laser scientist is confronted with the seemingly insoluble dilemma of
increasing the input energy needed for energy extraction, while keeping
the input intensity at an acceptable level. This problem was solved by the
CPA concept. First the pulse is stretched by a factor of 103 to 104. This
step does not change the input pulse energy and therefore does not affect
the energy extraction capability, but it does lower the input intensity by
the stretching ratio and hence keeps B to a reasonable level. Second, the
pulse is amplified by 6 to 12 orders of magnitude, i.e. from the nJ to the
millijoule or kilojoule level, before it is finally recompressed by the
stretching ratio back to a duration close to its initial value, see Fig. 4.
Figure 4. Chirped pulse amplification concept. To minimise nonlinear effects the pulse is first
stretched several thousand times lowering the intensity accordingly without changing the
input fluence J/cm2. The pulse is next amplified by a factor of 106–1012 and is then recom-
pressed by a factor of several thousand times closer to its initial value.
D. The Key Element: The Matched Stretcher-Compressor

In the first CPA set up (Strickland and Mourou, 1985) the laser pulse was
stretched using an optical fibre with a positive group delay dispersion and
was recompressed by a pair of parallel gratings (Treacy, 1969), with a neg-
ative group delay dispersion. Although this first embodiment led to a
spectacular 100-fold improvement in peak power, the problem was that
the stretcher and compressor were not matched over all orders. As a
result, after recompression the pulse exhibited unacceptable pre-pulses
and post-pulses.
This led the Rochester group to look for the ideal “matched stretch-
er-compressor.” It came in 1987, when Martinez (Martinez, 1987) pro-
posed a grating compressor with positive group delay dispersion for com-
munication applications as shown in Fig. 5. In communication systems
the wavelength of choice is 1.5 μm, a spectral region where the fibre
exhibits negative group velocity dispersion. After propagation in a fibre
the bits of information exhibit a negative chirp. It is therefore necessary
to use a dispersive delay line after propagation with a positive group
velocity dispersion to recompress the pulses. After examining this
arrangement, the Rochester group came to the conclusion that the Mar-
tinez “compressor” was in fact the matched stretcher of the Treacy com-
pressor that they were intently seeking. This can be easily shown by con-
sidering the arrangement shown in Fig. 5. When one uses a telescope of
magnification 1, the input grating located at a distance f from the first lens
will be imaged at the same distance.
Figure 5. Treacy and Martinez grating arrangements. The Martinez grating pair used as
a stretcher and Treacy grating pair used as a compressor. It was discovered and demon-
strated by Pessot et al., (1987) that these two grating arrangements are in fact matched
over all orders. The pulse can be stretched and recompressed arbitrarily keeping the initial
pulse unchanged. This grating arrangement is used in most CPA systems.
The phase conjugation properties of the ‘’stretcher/compressor’’ were

proposed and demonstrated in (Pessot et al., 1987) by stretching a pulse
of 80 fs by a factor of 1000 using the Martinez arrangement and then
compressing it back to exactly the same value using the Treacy compres-
sor.
This demonstration represented a major step in chirped pulse amplifica-
tion. This matched stretcher-compressor integrated into a CPA system was
to produce the first terawatt pulse on a table top system, the so-called T3 by
the Rochester group. It was subsequently used for subpicosecond pulse
amplification (Maine and Mourou, 1988; Maine et al., 1988) and for a pulse
duration of 100 fs by Pessot et al. (1989). This arrangement stands today as
the standard architecture used in CPA systems.
E. New Materials for CPA, OPCPA and Gain Narrowing

CPA was demonstrated initially with the two broadband-amplifying
media that were available at the time, Nd:glass and alexandrite (Pessot et
al., 1989). Shortly after, this initial work the concept was extended to
Ti:sapphire (Vaillancourt et al., 1990; Kmetec et al., 1991; Squier et al.,
1991; Sullivan et al., 1991) as well as Cr:LiSrAlF6 (Beaud et al., 1993; Dit-
mire and Perry, 1993) and Yb:glass (Nees et al., 1998). Among these mate-
rials Ti:sapphire has the advantage of the broadest bandwidth, with a
high damage threshold and excellent thermal conductivity greatly
enhanced at cryogenic temperatures (Backus et al., 1997).
Parametric amplifiers have also been proposed and demonstrated by
Dubietis et al. (1992). It was used for large scale laser applications at the
Rutherford Appleton Laboratory (Ross et al., 1997, 2000). This elegant
technique, called OPCPA for Optical Parametric Chirped Pulse Amplifica-
tion, is able, if the nonlinear propagation effects are kept under control, to
provide an extremely large gain bandwidth. It offers the great advantage
to be pumped by large-scale laser systems dedicated to laser fusion.
OPCPA could be the natural companion of any large laser fusion system.
A more detailed discussion of the OPCPA method is provided below.
In a conventional CPA system, one of the limitations in pulse duration
comes from gain narrowing. Because of their wide spectrum, short pulses
can be amplified only by materials with a gain bandwidth greater than the
pulse spectrum. We note that materials with superior energy storage typi-
cally have a low transition cross section and broad gain bandwidth. How-
ever, large gain will lead to a reduction of the laser spectrum as it is ampli-
fied and therefore a longer output pulse. In the unsaturated regime – the
linear regime – the laser spectrum will be subjected to a narrowing given by

(11)
where Δωa is the gain bandwidth and G(ωa) the exponential gain. A gain
of ten orders of magnitude will narrow the gain bandwidth by a factor of 3
to 4. A fraction of this gain, however, can be recovered in the saturated
section of the amplifier.
F. “CAN” a Novel Architecture

Many applications involving high peak power, like high energy particle
acceleration or transmutation of nuclear waste crucially need average
power. They typically need peak power in the PW range with MW average
power and excellent efficiency near 50%. Right now, a typical laser pro-
vides PW peak power but only with around tens watts average power and
0.1% efficiency. A far cry from what we would need for real applications.
To mitigate this problem, we proposed to build a novel amplifier based
on a multitude (network) of phased fibres, called Coherent Amplifying
Network (CAN). In a CAN fibre amplifier (Fig. 6) the active gain medium
is confined within the core of the optical fibre. Fibres benefit from effi-
cient heat removal due to the advantageous volume to lateral surface area.
It can also be pumped efficiently by a laser diode that is well-tuned to the
fibre. Hence the CAN laser could provide simultaneously, peak power,
average power and efficiency. Continuous-wave fibre lasers with average
powers in the kilowatt regime were first demonstrated in 2004.
Figure 6. CAN system is a fibre-laser based system built out of thousands of fibre lasers
coherently added to increase the average power and pulse energy linearly. The fibre
laser offers excellent efficiency (>30%) thanks to laser-diode pumping and provides a
much larger surface cooling area, therefore making operation at high peak power and
high average power possible with good efficiency and pristine beam quality.
Note also that CAN uses the CPA concept. An initial pulse from an
oscillator laser is stretched and split into many fibre channels as many as
104. Each channel is amplified in several stages, with the final stages pro-
ducing pulses of ~1 mJ at a high repetition rate. All the channels are com-
bined coherently, compressed and focused to produce a pulse with a pos-
sible energy of >10 J at a repetition rate of 10 kHz. Of course, the cost of
the system will be expensive but at this level any system will have a signif-
icant cost.
G. The Quest for PW Peak Power

As soon as the CPA concept was demonstrated at the millijoule and
joule levels, it became clear that it could be extended to much higher
energies using already built laser fusion systems to amplify nanosecond
pulses in the 100–1000 J range. This means that with remarkably few
alterations, that is, by chirping the pulse at the input and compressing it
at the output, a laser chain built to produce TW with ns duration pulses
could now produce petawatt PW with ps duration pulses (Maine et al.,
1987).
The first experiment started at the CEA-Limeil on the oldest laser
fusion system P102 in the early 1990s. The laser was capable of produc-
ing on the order of 100 joules in the nanosecond regime. A seed pulse
provided by an oscillator was stretched amplified to several tens of
joules and compressed to its initial value with the grating pair compres-
sor. Of course, the main difficulty was the fabrication of the gratings
with a decent damage threshold and efficiency that could handle the
large pulse energy. The largest gratings at the time, 30 cm x 40 cm, were
fabricated and integrated to the system. These first gratings were pro-
duced by Jobin-Yvon, now Horiba, with reflectivity of 90% and a dam-
age threshold of 0.1 J/cm2 and proved sufficient for a convincing
demonstration of high energy CPA.
With P102 retrofitted in CPA, we produced successfully 20 TW (Sau-
teret et al., 1991); and later 50 TW (Rouyer et al., 1993). Similar result few
months later where obtained independently at ILE in Japan by Yamakawa
et al. (1991). These results showed clearly that the CPA could work at any
arbitrary energy level and would be capable to produce PW pulse. Indeed,
as we predicted, the first Petawatt pulse was demonstrated by Perry et al.
(1999), ten years after the first terawatt. One of the impressive hurdles
overcome by Perry’s group was the fabrication of meter-size diffraction
gratings. They revolutionised the field.
Figure 7. World map of lasers with peak power >100TW.
In parallel to the Nd:glass based petawatt systems, we have today a num-

ber of high-power Ti:sapphire-based systems. They exhibit much shorter
pulses in the 20–30 fs range, and energies in the 5–10 J range, thus deliv-
ering peak power of 100 TW. A 100 TW class Ti:sapphire laser was first
demonstrated at the University of California at San Diego (Barty et al.,
1994). The leading laboratories at the present time in this area are
CoReLS Research Center, in Korea, APRC in Japan with around 500 TW
(Aoyama et al., 2002), Janus System at Lawrence Livermore, 200 TW, the
Laboratoire d’Optique Appliquée (LOA) in France with 100 TW, the Max
Born Institute in Germany with 100 TW, the Lund University in Sweden
with 30 TW, and the Center for Ultrafast Optical Science University of
Michigan with 200 TW. Today nearly one hundred 100 TW systems are
operating (Fig. 7) with about 20 systems at the PW level (Fig. 8) existing
or under construction. Groups in China (SIOM), Russia (APC), and the
USA (LLE) are now aiming at the 100 PW level.
Figure 8. World map of lasers with peak power 1PW.

H. Optical Parametric Chirped Pulse Amplification (OPCPA)

In this section we discuss the differences between the CPA and OPCPA
methods. Fig. 9 shows the conceptual layout of an Optical Parametric
CPA, OPCPA system (Dubietis et al., 1992; Ross et al., 1997). Because we
are reviewing only relativistic intensity laser we will not mention the large
number of projects related to the sub-relativistic regime using OPCPA. As
in CPA the objective in OPCPA is to stretch the pulse to a nanosecond
duration and then amplify it to the joule or higher level by optical para-
metric amplification and recompress it back to close to its initial value.
Note that the stretching is essential not only to keep the B integral low but
also to pump and extract the energy efficiently. It is only during the
stretched pulse that light can be transferred from the pump beam to the
signal beam. Consequently, the pump pulse duration and that of the
stretched pulse must be as equal as possible. The advantages of this tech-
nique are as follows:
1. Large bandwidth that could accommodate few-cycle pulses.

2. Ability to benefit from very large KDP crystals (100 × 100 cm2)
developed for laser fusion.
3. Adaptability to existing laser fusion chains, which benefit from
low-bandwidth well collimated nanosecond laser pulses at 532 nm.
4. No heat dissipation in the OPA crystal itself, this is important for
high average power.
5. No transverse amplified stimulated emission, which is a major
source of loss for large-aperture Ti:sapphire systems.
6. Ability to use an iodine laser as a pumping source.
7. Very simple amplification system.
Figure 9. OPCPA concept. In the OPCPA the pulse is amplified by optical parametric
amplification instead of regular optical amplification. Note that for efficiency the pump
pulse and the stretched pulse must have approximately the same duration and the same
spatial extend.
The disadvantages are as follows:
1. Lower efficiency than standard CPA. For a standard Ti:sapphire

CPA, the efficiency can be as high as 50% from a long green pulse,
say of 50 ns. The energy storage time of Ti:sapphire is 2 μs. So CPA
is overall a more efficient system.
2. Very large stretching ratio, in the range of 106 to 1, i.e. 10 fs to 5 ns,
necessary for energy extraction. This will make pulse compression
down to the 10 fs regime difficult.
3. Gain is a significant function of the intensity. This means the pump-
beam profile may affect the beam quality and needs a high level of
control.
In both systems, the pulse duration will ultimately be limited by the grat-
ing bandwidth. At present no large gratings have the efficiency and the
bandwidth necessary for efficient pulse compression much below 30 fs.
Beam quality from CPA has been demonstrated to be excellent. The latest
studies have shown that OPCPA can also provide good spatial beam qual-
ity (Collier et al., 1999). The potential of this technique has been demon-
strated with the production of a 35 J, 85 fs pulse, equivalent to 0.4 PW,
using a 10 cm-diameter beam (Collier et al., 2004). The possibility of
reaching high energies seems to be more straight-forward with the
OPCPA because it can benefit from kJ, ns fusion lasers that are already up
and running. The pulse duration, however, will be limited by the grating
bandwidth. CPA implementations must wait for large Ti:Sapphire crystals
grown to 20×20 cm2 dimensions. These larger-scale crystals should
become available as the demand for higher peak power increases.
III. BEYOND THE HORIZON: THE ZEPTOSECOND

AND EXAWATT DOMAIN
In their pioneering experiment Grischkowsky and Ballant (1982) used a
single mode optical fibre and were able to compress a picosecond pulse
with nJ energy to the femtosecond level. This work triggered an enor-
mous interest. To go higher in energy Orazio Svelto and his group (Nisoli
et al., 1996) introduced a compression technique based on a fused silica
hollow-core capillary, filled with noble gases to broaden the pulse spec-
trum before efficiently compressing to the 100 μJ level with chirped mir-
rors that introduce negative dispersion to the pulse phase. Refining this
technique Orazio Svelto, Ferenc Krausz and collaborators (Nisoli et al.,
1997) could compress a 20 fs into 5 fs or 2 cycles of light at 800 nm.
For higher energy, bulk compression was attempted by Rolland and
Corkum (1988). In their implementation, the pulse is free-propagating in
a solid and not guided anymore. The pulse was relatively long, around 50
fs, with an input energy of 500 μJ leading to an output pulse of 100 μJ in
20 fs. However this scheme is impaired by the beam bell shape intensity
distribution. This leads to variation in the nonlinear response across the
beam profile and a corresponding variation of the compression factor.
Figure 10. This element uses a

uniform thickness plastic element
of around .5mm. An incident flat
top pulse will induce a uniform
self-phase modulation across the
element. After spectral broade-
ning the pulse is compressed by
2 chirped mirrors to a near single
cycle pulse into a 2.1 fs after two
compression stages.
A. Large Energy Pulse Compression: Thin Film Compression (TFC)

Recently we have proposed a technique (Mourou et al., 2014) to compress
25 fs large energy pulses as high as 1 kJ to the 1–2 fs level. We call this
approach the Thin Film Compressor, or TFC, and its principle is shown in
Fig. 10. As shown in our simulation, this method is very efficient >50%
and preserves the beam quality. Here the technique relies on the flat top
nature of the high energy pulse and uses a thin “plastic” film of ~500 μm
with a diameter of 20 cm. The element, that we call plastic for simplicity,
will be transparent and exhibit a uniform thickness. The modelling of the
pulse propagation leading to spectral broadening due to the nonlinear
processes that occur within the film is described in Akhamov et al. (1992).
After spectral broadening the pulse is compressed by chirped mirrors to a
near single cycle pulse showing a pulse of 2.1 fs after two compression
stages. This represents the shortest pulse duration that can be achieved
based on the limitation defined by the laser wavelength, λ. Once the pulse
is focused on λ2 the pulse is in the desired λ3 regime. The λ3 regime is the
focus volume that achieves the highest intensity for a given pulse energy.
In order to increase the intensity, it is necessary to shift to shorter wave-
lengths that can support shorter pulse durations and tighter focusing.
Figure 11. The pulse is focused on λ2 and create a mirror with an indentation deforms due
to (a). As it moves relativistically, it provides an isolated an individual pulse (b), (c).
The proposed method for up-converting to shorter wavelengths is

through a relativistic plasma mirror, and especially in the laser conditions
of the λ3 regime (Mourou et al., 2006). The relativistic mirror is not pla-
nar and rather deforms due to the indentation created Fig 11c by the
focused Gaussian beam. As it moves relativistically in and out and side-
ways, the reflected beam is broadcast in specific directions and provide
an isolated an individual pulse Fig. 11 a, b. In the relativistic regime (Nau-
mova et al., 2004), predicts a pulse duration T compressed by the relativ-
istic mirror-scaling like T = 600 [attosecond]/a0. (Fig. 12) Here a0 is the
normalised vector potential, which is unity at 1018 W/cm2 and scales as the
square root of the intensity. For intensities on the order of 1022–1024 W/
cm2 where a0 is on the order of 100 to 1 000, the compressed pulse could
be only a few attoseconds, even zeptoseconds. Naumova et al. (2004)
have simulated the generation of thin sheets of electrons of few nm thick-
ness, much shorter than the laser period. It opens the prospect for X and
gamma coherent scattering with good efficiency.
Figure 12. Naumova et al., (2004)

predicts a pulse duration scaling
like T=600 (attosecond)/a0.
IV. LASER PLASMA ELECTRON ACCELERATION

The concept of laser-plasma wake-field electron accelerator was pro-
posed in 1979 by Toshiki Tajima and John M. Dawson (1979). In a laser-
plasma accelerator (LPA), a plasma medium (e.g. fully stripped helium or
hydrogen ions surrounded by free moving electrons) is used to transform
electromagnetic energy from a laser pulse into kinetic energy of acceler-
ated electrons by exciting high amplitude plasma density waves. See Fig.
13. An intense laser pulse (Esarey, 1996; Everett et. al., 1994; Clayton et.
al., 1994; Umstadter et. al., 1996) causes the plasma electrons to move out
of its path through the “photon pressure”. The much heavier ions barely
move and as a consequence are left unshielded. Some distance behind the
laser pulse, the electrostatic force exerted by the ions on the electrons
pulls them back to the axis, creating an electron density peak. The pattern
of alternating positive and negative charges is referred to as a plasma
wave or laser wake and supports an electric field. The wave oscillates at
the plasma frequency which scales as the square root of the plasma den-
sity and has a wavelength typically around 10 to 100 μm. This is several
orders of magnitude shorter than the typical RF period used in conven-
tional accelerators. The amplitude of the plasma wave or strength of the
electric field is proportional to the square root of the plasma density
(number of free electrons per unit volume) and proportional to the laser
intensity (for intensities ≥1018 W/cm2). For typical densities (1018–1019 elec-
trons/cm3) used in experiments, fields ranging from 10–100 GV/cm are
produced, three orders of magnitude greater than with conventional tech-
nology. The wave’s phase velocity is near the speed of light and electrons
injected at the proper phase can be accelerated to high energies. To reach
the same particle energy, plasma accelerators can then, in principle, be
three orders of magnitude shorter than their conventional counterparts.
Figure 13. The Laser Wake

Field Acceleration concept
relies on the very large
radiation pressure inducing a
plasma wave. Electrons will be
trapped in the wave and acce-
lerated to energy of GeV/cm,
a thousand time greater than
conventional technology.
The wavelength of the plasma waves is also around three orders of mag-
nitude smaller than the wavelength of the radiofrequency used in conven-
tional accelerators. The generation of low energy dispersion electron
bunches requires that the length of the bunch to be a small fraction of
these wavelengths and/or the use of complex techniques only compatible
with a large facility. In the case of plasma accelerators this condition
implies the use of electron bunches shorter than 10 fs to get an energy
dispersion below 10% typically. So far, these short pulses were produced
by controlling the wavebreaking of the plasma waves or by using laser
beam collision in the acceleration zone to produce a strongly localised
injection. In a relatively short time the major hurdles, one after the other,
have been surmounted.
Considering the beam monochromaticity, it has been simultaneously
demonstrated (Mangles et. al., 2004; Geddes et. al., 2004; Faure et. al.,
2004) to be of the order of 1%. In addition, high energy electrons have
also been shown to the GeV level by several groups in the USA at the
Lawrence Berkeley Laboratory, the Texas Petawatt Facility, University of
Texas; in Korea at CoReLS; and in China at SIOM and Peking University.
That considered, the major stumbling blocks to laser-based accelera-
tion (the repetition rate, average power and poor efficiency) due to the
inherent low repetition rate of the driving laser are being addressed. Rev-
olutionary laser infrastructures based on phased-array-optical fibre, like
the Coherent Amplification Network (CAN) described in section II. F. are
being actively studied. (Mourou et al., 2013)
Accelerator technology has immense promise for innovation for vari-
ous applications ranging from science with High Energy Particle Physics,
to applications which include betatron and free-electron light sources for
diagnostics or radiation therapy and proton source for hadron therapy as
well as homeland security.
V. HIGH ENERGY PROTON BEAMS, PROTON THERAPY

It has been shown that laser thin-target interactions can produce plentiful
MeV protons in a beam with superior transverse emittance. The proton
generation is a direct consequence of electron acceleration. Electrons that
are violently accelerated in the laser field can draw behind them protons
that are on either the front or back surface of the target.
Highly energetic proton beams have been demonstrated at Livermore,
LULI, CUOS, and Rutherford with intensities of 1018–1020 W/cm2. They
could lead to important applications such as fast ignition for inertial con-
finement fusion as was pointed out by Roth et al. (2001) and proton ther-
apy (Bulanov and Khoroshkov, 2002; Fourkal et al., 2002).
The proton used in radiotherapy and oncology provides several advan-
tages. First, proton beam scattering on atomic electrons is weak and

results in low irradiation of healthy tissues surrounding the tumour. Sec-
ond, the stopping length for the proton with a given energy is fixed and
avoids irradiation of the healthy tissues at the rear side of the tumour.
Third, the Bragg peak of the energy losses provides substantial energy
deposition in the vicinity of the proton stopping point.
Currently, proton beams with the required parameters are produced
with conventional charged particle accelerators: synchrotron, cyclotron,
and linear accelerators. The use of the laser accelerator is very attractive
because its compactness is associated with additional possibilities for
controlling proton beam parameters. The typical energy spectrum of
laser-accelerated particles observed both in experiments and in computer
simulations can be approximated by a quasi-thermal distribution with a
cut-off at a maximum energy.
The effective temperature attributed to fast ion beams is within only a
factor of a few from the maximum value of the particle energy. On the
other hand, the above-mentioned applications require high-quality proton
beams, i.e. beams with sufficiently small energy spread ΔE / E. For exam-
ple, for hadron therapy it is highly desirable to have a proton beam with
ΔE / E=2% in order to provide the conditions for a high irradiation dose
being delivered to the tumour while sparing neighbouring tissues.
In the case of the ion injector, a high-quality beam is needed in order to
inject the charged particles into the optimal accelerating phase. Bulanov
and Khoroshkov (2002) have shown that such a beam of laser-acceler-
ated ions can be obtained by using a double-layer target. Multilayer tar-
gets have been used for a long time in order to increase the efficiency of
the laser energy conversion into plasma and fast particle kinetic energy;
see, for example, Badziak et al. (2001, 2003). In contrast to the previously
discussed configurations, the use of a double-layer target was proposed in
order to produce fast proton beams with controlled quality. In this
scheme the target is made of two layers with ions of different electric
charge and mass.
A regime of ion acceleration that exhibits very favourable properties
has been identified by Esirkepov et al. (2004). In this regime the radiation
pressure of the electromagnetic wave plays a dominant role in the interac-
tion of an ultra-intense laser pulse with a foil. In this radiation pressure
dominant regime, ion acceleration appears due to the radiation pressure
of the laser light on the electron component with momentum transferred
to ions through the electric field arising from charge separation. In this
regime, the proton component moves forward with almost the same
velocity as the average longitudinal velocity of the electron component.
Thus, the proton kinetic energy is well above that of the electron compo-
nent. In addition, in the radiation pressure dominant regime the ion accel-
Figure 14. In this regime for intensity in the 1023 W/cm2, the radiation pressure is domi-
nant and scales with the laser-pulse energy.
eration mechanism is found to be highly efficient, and, as we shall explic-

itly show, the ion energy per nucleon is proportional to the laser-pulse
energy. The main results of three-dimensional PIC simulations are shown
in Fig.14.
A. Proton acceleration in the single-cycle regime

Single-Cycled Laser Acceleration (SCLA) of ions (Zhou et al., 2016) relies
on laser pulses which give access to a new ion acceleration regime. His-
torically, attempts in laser acceleration of ions relied on the multi-cycle,
high-energy pulses available. The first experimentally realised laser ion
acceleration was TNSA (Cousens et al., 2006). In this mechanism the tar-
get was thick, electrons penetrated through the thick target and ions were
not adiabatically trapped and accelerated. Rather ions were accelerated
on the surface of the fixed target over the sheath. For the foil thickness l
we define the normalised electron areal density σ = ne l/nc λ as the target
parameter. One way to increase the adiabaticity and prolong the interac-
tion for ion acceleration was to reduce the mass of the target. This
reduced σ, as in the regime of Coherent Acceleration of Ions by Lasers
(CAIL) (Yan et al., 2010) is far different from the TNSA regime. Radiation
Pressure Acceleration (RPA) (Esirkepov, et al., 2004) increases a0 and
slightly decreases σ as compared with TNSA. SCLA, by the virtue of the
decreased pulse length of the laser, also reduces σ and increases a0. Thus,
due to the reduction in σ and increase in a0 for SCLA (and RPA) relative
to TNSA, the coherence of the ion acceleration is enhanced. Two clear
advantages then arise with SCLA over the longer pulse driven RPA: first,
Figure 15. In the single-cycle regime, the acceleration efficiency of ions varies sharply.
The fewer number of cycles translates into a larger laser vector potential and yields
higher proton cut-off energy. The proton energy is increased by reducing the number of
cycles from τ = 16T (black curve) where T is the period, to τ = 4T (blue curve). In particu-
lar, with the single-cycle pulse (red curve), the cut-off energy of the ions is increased by a
significant amount and becomes quickly relativistic.
the pulse intensity is enhanced for a given laser energy as the pulse dura-
tion is reduced; second, the elimination of multi-cycle averaging over the
oscillations of a longer pulse enhances the efficiency, coherence, and sta-
bility of the ponderomotive acceleration. See Fig. 15. With this combina-
tion, the ion acceleration in the limit of single-cycle laser pulses becomes
far more robust, stable and intense due to the simple fact that the electron
ponderomotive acceleration term ‹v×B› no longer requires cycle-averag-
ing as in the case of longer, multi-cycle pulses. The application of the sin-
gle-cycle regime introduces a more coherent electron acceleration and
sharper electron layer formation that creates the fields that will ultimately
accelerate the trailing ions.
We know that in an ideal RPA light sail regime, the resultant maximum
ion energy is inversely proportional to the total mass of the accelerated
target. In a simple picture, the optimum thickness is achieved by decreas-
ing it, namely, the lower the total mass, the higher the final maximum
energy. However, other physical processes, such as transverse instabili-
ties, will strongly affect the actual acceleration process and prevent it
from reaching the optimum acceleration, particularly with current state-
of-the art multi-cycle ultra-intense laser pulses. While for shorter pulse
durations, especially for single-cycle pulses, the duration is too short for
those instabilities to develop and the constraints caused by instabilities

are strongly suppressed, which gives us more opportunity to approach to
the ideal case. So compared to the traditional RPA, the optimal target
thickness becomes smaller. Specifically, SCLA is located within the trans-
parent area (σ << a0) corresponding to smaller values for σopt.
VI. NEUTRON SOURCES

Neutron beamlines produced by CPA will enable a broad range of both
fundamental and applied science. Thermal neutron sources (~meV–eV)
are widely used for diffraction and spectroscopy experiments (Blomgren
et al., 2006). and fast neutrons (~MeV) can be applied to radiography,
medicine and material damage studies (Rynes et al., 1999). Laser gener-
ated sources of neutrons hold the advantages of a small source size and
pulse duration and synchronicity with other secondary sources. Imaging
using combined sources can yield information about the composition of
materials. The intense source of γ-radiation planned at ELI-NP, for
instance, also opens up the possibility of a revolutionary thermal neutron
source generated through neutron halo isomers. This scheme avoids the
need for moderation of the neutron energy and is predicted to generate
neutron fluxes orders of magnitude higher than at existing spallation
sources (Habs et. al., 2010).
A. Laser based sources

Laser-driven neutron sources are based on DD or DT fusion and (p,n)
and (γ,n) reactions. There are two approaches: reactions between ions
within a single laser-heated deuterium target or conversion in a second-
ary target of a laser driven ion or γ ray beam.
B. Single target scheme

Neutrons can be generated by irradiating thick solid deuterated plastic
targets with intense lasers. A series of experiments with 1–100 TW lasers
demonstrated yields up to 107–108 n sr−1 per shot (Pretzler et al., 1998;
Izumi et al., 2002). Highly energetic deuterons heated directly by the laser
at the front surface stream into the target and so initiate ‘beam-target’
fusion reactions with cold deuterons within the bulk of the solid. The
angular distribution of the neutrons is closely related to that of the accel-
erated ions and is predominantly in the forward direction for short scale-
length plasmas. Particle-in-cell code modelling of this mechanism is in
good agreement with these experimental results and identified laser and
plasma parameters, which lead to the production of an energy tuneable,
quasi-mono-energetic forward-directed fast neutron beam. Fusion reac-

tions can also be initiated in gases or a mist of sub-micron droplets pro-
duced using pulsed gas jets. The presence of solid-density clusters or
droplets dramatically increases the absorption efficiency of short pulse
intense irradiation, which can exceed 90% of a petawatt-class laser pulse
(Gumbrell et al., 2008). Because the interaction is with an extended target
several millimetres in length, ‘beam–beam’ reactions can occur between
hot deuterons within the plasma as well as ‘beam–target’ reactions with
the surrounding medium. If the emission is dominated by the thermonu-
clear process, then the duration of the neutron burst is limited to less
than a nanosecond because the ions rapidly traverse the laser-heated vol-
ume. These experiments have been performed using D2, CD4 or D2O and
in principle can use a mixture of the tritiated forms of these materials.
These sources are generally considered as isotropic, with their flux scal-
ing strongly with the laser energy.
C. Double target scheme

In ‘pitcher-catcher’ experiments a primary target (pitcher) is irradiated
with an intense laser pulse to generate an ion beam from the rear surface.
The catcher is a slab of material which acts as a neutron converter in the
same way as they are used on accelerator facility neutron sources. DD
fusion in deuterated plastic catcher targets yielding ~104 neutrons per
shot has been demonstrated using liquid droplets as the laser interaction
target (20 μm diameter D2O), enabling a high repetition rate.
The same arrangement can employ the large fluxes of γ rays produced in
intense laser plasma interactions to initiate (γ,n) reactions in the catcher
sample. These photo-neutrons are usually measured to diagnose the inter-
action rather than being optimised to investigate their potential as a source
and so yields tend to be low (~100 neutrons per shot). In the next section
we discuss a preferable option for generating neutrons using γ beams.
D. Neutron spallation
High intensity lasers can produce high energy protons in the several 100
MeV regime. These can be used to strike a target of a heavy metal, mer-
cury or tantalum. Each impact can produce 20–30 high energy neutrons
per spallation. This technique has the advantage of producing a beam of
neutrons that can be modulated. This process may be linked to conven-
tional nuclear reactor technology in an accelerator-driven system (ADS)
to transmute long-lived radioisotopes in used nuclear fuel into short-
er-lived fission products.
VII. X-RAYS
Since the dawn of the laser in 1960, scientists focused them to create
non-linear effects and very quickly, as intensity increased, plasmas. Plas-
mas are ionised media known before the existence of lasers to be good
emitters of X-rays. Naturally, laser-driven plasmas were studied for the
generation of incoherent as well as coherent X-rays. The first proposal of
a laser emitting X-rays was published in 1967 by Duguay and Rentzepis
(1967), only seven years after the demonstration of the first ruby laser by
Theodore Maiman. However, progress on the development of both inco-
herent and coherent X-rays was slow following delays in the evolution of
the laser intensity. The demonstration and then the implementation of
the CPA technique triggered a major change of paradigm for the field of
laser-driven X-ray sources. This section gives a short overview of the dif-
ferent X-ray sources generated by laser that are available today with spe-
cial attention on the impact of CPA technique.
A. Incoherent X-ray sources:

By focusing an intense laser on a gaseous, liquid or solid target, plasma is
created that emits X-rays through Bremsstrahlung or after atomic inner
shell excitation (K-α emission) of the cold solid part of the target. To be
able to excite inner-shells, the electron kinetic energy has to be higher
than the atomic level energy. Prior to CPA, acceleration of electrons
inside the plasma was just sufficient to produce K-α emission from low
atomic number elements like Aluminium (1.48 keV) and Copper (8 keV).
The jump in intensity provided by CPA induced very strong electron
acceleration in the plasma leading today to the production of K-α from
elements as heavy as Gold (68.8 keV). Furthermore, in these conditions,
Bremsstrahlung emission has been observed to extend up to several 100’s
of keV. Experiments have also used external laser-accelerated electron
beams with energy up to several 100’s of MeV leading to Bremsstrahlung
emission extending also to 100’s of MeV. All these developments are of
high interest for medicine (radiotherapy or imaging) and for non-destruc-
tive imaging.
B. Coherent X-ray sources:

The interaction of an intense picosecond or femtosecond laser with a gas
may generate the odd harmonics of the fundamental laser with a spec-
trum extending to a cut-off limit given by the law Ec ≈ 3.12 Up + Ip where
Ip is the ionisation potential and Up = I λ2 and λ being the fundamental
laser wavelength. It is apparent that by increasing the intensity, CPA
allowed an extension the cut-off to much higher photon energies,
although only up to a saturation point. Today multi-keV high harmonics

have been demonstrated using a 4 µm CPA laser. Moreover, gas ionisation
is detrimental for the phase-matching between the infrared laser and the
high harmonics, thus limiting the energy transfer between them. Using
ultrafast (the best being single-cycle) and intense laser pulses ensures
strong high harmonic emission. High harmonics having their spectral
phase locked demonstrate the emission of attosecond duration pulses.
Last but not least, high harmonics maintain the optical properties of the
fundamental laser: their wavefront might partially be controlled by the IR
laser wavefront (Gauthier et al., 2008), their angle of polarisation is fol-
lowing that of the IR laser (Vodungbo et al., 2011) and they are very
coherent. They are thus excellent X-ray sources for fundamental and soci-
etal applications.
High harmonics might be also produced by the interaction of a very
intense laser (1019 W cm–2 and above) with solid target. In that case, the
laser excites the electron from the target surface, pushing and pulling
them away every half cycle. Therefore the electrons generate a train of
attosecond X-ray bursts (Teubner and Gibbon, 2008). In some extreme
cases, the laser can be back reflected by the relativistic electrons expelled
from the surface. The laser is Doppler shifted to much higher photon
energies, reaching the keV range. The pulse is also compressed down to
100’s of attoseconds or shorter (Naumova et al., 2004; Tsakiris et al.,
2006). The X-ray source is highly coherent and controllable by adjusting
the laser and target parameters.
It is well-known that at high intensities, the interaction of a CPA laser
with gas accelerates electrons up to 100’s MeV and even at a few GeV
over few millimetres of gas. Such high acceleration gradient is achieved in
the so-called “bubble regime” where the laser pushed away the electrons
from its path, leaving on its back a bubble with a deficit of electrons and
thus an excess of ions. The expelled electrons can thus be accelerated
inside this bubble due to the high electric field while for some conditions
they also start to oscillate transversely producing betatron oscillations.
This source emits radiation with a cone aperture varying as 1/γ2, γ being
the Lorentz factor. The radiation is a “white spectrum” with a cut-off
energy depending on γ2. With a 100 TW-class laser, emission up to 20 keV
has been demonstrated within a cone of a few degrees. The emission
intrinsically has the duration of the laser (few fs) and the size of the focal
spot (few μm) making for a very promising source for medical applica-
tions as well as for imaging.
As mentioned above, in the late 60s–early 70s research proposed dif-
ferent schemes for producing a real X-ray laser i.e. by creating a popula-
tion inversion in plasma necessary for stimulated amplification of sponta-
neous emission (ASE). Although researchers were very active using non-
CPA lasers and saturated amplification was demonstrated in 1991, the

emergence of CPA created a revolution. Thanks to CPA and the possibility
of achieving ultra-high intensity, the energy of the pump laser was quickly
reduced from kJ, to 10’s of J and later joules with a subsequent jump in
repetition rate from 10–4 Hz to 10 Hz. It is worth noting that the ASE
regime without a cavity (normal X-ray mirrors are not suitable for a cav-
ity) led to weakly coherent X-ray lasers. By seeding the amplifying plasma
with fully coherent high harmonics produced by the interaction of a CPA
laser with gas, the amplified beam is demonstrated to be fully coherent,
polarised (either linear or circular depending of the incoming beam) (Zei-
toun et al., 2004) and having an excellent wavefront (Godet et al., 2009).
C. X-ray CPA
Recently, it was proposed to use the CPA technique directly in the X-ray
range for both free-electron lasers (FEL) and plasma-based X-ray lasers
(PBXRL). Although the physics behind the X-ray amplification in FEL and
PBXRL is different, the general concept is very similar: While amplifying
an ultrashort X-ray pulse, typically femtosecond, most of the energy
stored in the medium is left untouched because of a temporal mismatch
(Wang et al., 2014; Oliva et al., 2012). By stretching the incoming femto-
second X-ray pulse, then amplifying and finally compressing the pulse,
most of the stored energy may be transferred to the X-ray pulse. The con-
cept was successfully demonstrated on an FEL (Gauthier et al., 2016).
Thanks to the CPA these sources are now mature and used for applica-
tions in various domains from biology, chemistry and physics.
VIII. ATTOSECOND SCIENCE

Attosecond science represents one of the frontiers in Ultrafast Optics,
since it offers the possibility to initiate and control the motion of elec-
tronic wave packets inside atoms, molecules, nanostructures and solids. It
is now a well-established research field, which offers formidable tools for
the investigation of fundamental electronic processes (Calegari et al.,
2016).
The production of attosecond pulses is based on high-order harmonic
generation (HHG) in gases, which is a highly nonlinear process taking
place when an intense (1013–1015 W/cm2) and short laser pulse is focused
into a gas medium. The HHG process then results in the production of
coherent extreme ultraviolet (XUV) radiation with pulse duration down
to the attosecond regime. The physical processes giving rise to HHG can
be understood in the framework of a quasi-classical three-step model
(Schafer et al., 1993; Corkum et al., 1993). According to this model the
most weakly bound electron tunnels through the energy barrier formed
by the Coulomb field in the presence of the driving electric field. It then
accelerates in the oscillating field as a free particle to eventually recom-
bine with the parent ion, thus emitting a high energy photon. This pro-
cess is periodically repeated every half optical cycle of the fundamental
radiation, thus leading to a periodic emission of very short radiation
bursts, with duration in the attosecond range, as first experimentally
demonstrated by Paul et al. (2001).
For a number of important applications, such as pump–probe experi-
ments, it is essential to isolate a single pulse of the train, which can then
be synchronised with another optical pulse (Sansone et al., 2010a). The
first experimental demonstration of isolated attosecond pulses was
reported by Hentschel et al. (2001). A few developments in femtosecond
laser technology were essential to achieve the generation of isolated atto-
second pulses, in particular the stabilisation of the carrier-envelope phase
(CEP) of the driving pulses (Jones et al., 2000) and the generation of
high-peak-power sub-5-fs pulses (Baltuška et al., 2003). The most com-
mon way to generate high-peak power, few-optical-cycle pulses is to
apply suitable post-compression schemes. The most common technique
is based on spectral broadening in hollow fibres filled with noble gases, in
combination with broadband dispersive delay-line with chirped mirrors
(Nisoli et al., 1996, 1997).
Attosecond pulses were first employed for the investigation of ultrafast
electron dynamics in atomic physics. Particularly interesting is the meas-
urement of the delay in photoemission (Schultze et al., 2010; Klünder et
al., 2011), the analysis of the process of tunnel ionization (Dudovich et al.,
2006; Uiberacker et al., 2007), the investigation of electronic correlation
in helium (Ossiander et al., 2017) to name but a few examples. The first
application of attosecond pulses to molecular physics was reported in
2010, with the measurement of the electron localisation process in H2
and D2 molecules after ionisation induced by isolated attosecond pulses
(Sansone et al., 2010b). By employing high harmonic spectroscopy, atto-
second charge migration was measured and controlled in ionised iodo-
acetylene by analysing the harmonic light emitted after excitation of the
neutral molecule with a strong NIR pulse (Kraus et al., 2015). Particularly
interesting is the possibility to investigate the ultrafast electron dynamics
in complex molecules where sudden ionisation by attosecond pulses may
produce ultrafast charge migration along the molecular skeleton induce a
nuclear rearrangement (Cederbaum et al., 1999; Remacle et al., 2006).
The process of electron transfer in molecular complexes is of crucial
importance in biochemistry since it triggers the first steps in a number of
biochemical processes, such as photosynthesis and electron transport
along DNA. Experimental evidence of charge migration initiated by iso-
Figure 16. Electron density around the amine group of phenylalanine changes sign in ~2 fs.
lated attosecond pulses were first reported in 2014 in the amino acid phe-
nylalanine- (Calegari et al., 2014). Charge migration was evidenced as an
oscillatory evolution in the yield of a doubly-charged molecular fragment.
Fig. 16 shows the calculated temporal evolution of the electronic wave
packet generated by the attosecond pulses.
Attosecond technology has also been employed to study ultrafast pro-
cesses in condensed matter. The first application of attosecond pulses to
solids was reported in 2007, with the observation that photoelectrons
from the 4f band of tungsten reach the surface ~100 attoseconds later
than those from the Fermi-edge (Calegari et al., 2014). Another very inter-
esting application was reported in 2013, with the observation of an insu-
lator-to-conductor transition in fused silica (Schultze et al., 2013). It was
demonstrated that the ac conductivity of fused silica can be increased by
more than 18 orders of magnitude, within 1 fs, by using few-optical-cycle
NIR pulses (Schiffrin et al., 2013). The process of tunnel ionisation has
been investigated in solids by measuring the transmission of isolated
attosecond pulses through a single-crystalline, free-standing Si mem-
brane as a function of the delay with respect to a NIR pump pulse, which
excites electrons from the valence band to the conduction band (Schultze
et al., 2014). A photo-assisted tunnelling injection mechanism called the
dynamical Franz-Keldysh effect has been observed in thin films of poly-
crystalline diamond (Lucchini et al., 2016). This effect is based on elec-
tronic intra-band motion, which competes with inter-band transitions.
More recently, the role of inter- and intra-band dynamics in electron
transfer between valence and conduction bands in GaAs monocrystalline
samples has been investigated by employing attosecond pulses (Schlaep-
fer et al., 2018).
IX. ASTROPHYSICS
Over the last decade, ion acceleration driven by ultra-intense laser pulses
has been emerging as a very exciting potential alternative to conventional
acceleration technology. By offering radically new characteristics (Fuchs
et al., 2006; Macchi et al., 2013), such as extremely compact acceleration
(over less than 1 mm), ultra-short (less than 1 ps) and ultra-dense
bunches (over 1013 particles/bunch in a single shot), laser ion acceleration
offers very promising applications as in hadron therapy and for the pro-
duction of medical radioisotopes. But already, laser-accelerated ions are
used in ground-breaking applications in proton radiography (Chen et al.,
2012), to study the concept of rapid ignition for Inertial Confinement
Fusion (Roth et al., 2001), and in the so-called field of “laboratory astro-
physics” (Remington et al., 1999; Albertazzi et al., 2014; Gregori et al.,
2012).
A new emerging prospect in this laboratory astrophysics domain (Chen
et al., 2018) is to investigate the question of the nucleosynthesis of heavy
elements. At present, there is no complete physical scenario capable of
explaining and reproducing the observed abundances of super-heavy ele-
ments in our Solar System as well as in other systems (Arnould et al.,
2007; Thielemann et al., 2011; Reifarth et al., 2014). These super-heavy
elements (i.e. roughly half of the elements heavier than Iron, and almost
all of those beyond Bismuth) are postulated to be generated through the
r-process (Arnould et al., 2007). This process proceeds through multiple
neutron capture in a nucleus, rapidly increasing its mass number (A) until
a β--decay (emission of an electron and a anti-neutrino) takes place,
which leads to an increase in atomic number (Z). It is generally accepted
that this process can occur only under extremely high neutron flux of
more than 1020 n/[cm2.s] (Cowan et al., 1985) in order for multiple neu-
tron captures to take place despite the small cross-sections involved. The
main problem is that almost all of our knowledge relative to the dynamics
of the r-process rests on theory and simulations, but their accuracy is
hindered by large uncertainties (Mumpower et al., 2016; Panov, 2016) in
the nuclear data for the involved heavy nuclei (e.g. neutron capture rates,
β-decay and α-decay half-lives, masses). This is due to the fact that no
facility based on conventional technology- can even come close to such
an extreme neutron flux.
To reverse this, ultra-intense lasers offer a radically new prospect as
they could be used to generate ultra-bright neutron beams in order to
make direct measurements of neutron capture, i.e. (n,γ) nuclear reactions,
as well as subsequent β-decay rates of the radioactive isotopes that are
created by the neutron capture (Chen et al., 2018). For this, the required
short duration and high neutron flux could be generated, from laser-ac-
celerated protons, through spallation. Another advantage offered by
ultra-intense laser facilities for such investigations is that they are

equipped with auxiliary high-energy laser pulses, i.e. offering the possibil-
ity to perform nuclear measurements in a hot plasma environment that
emulates astrophysical conditions.
To generate the required ultra-bright neutron beams through spalla-
tion, imminent multi-PW lasers offer the prospect of pushing the focused
intensity of the laser beam on target by at least one order of magnitude
compared to current lasers, i.e. to 1022 W.cm–2, and even further, to 1023
W.cm–2, with the help of refocusing plasma optics, allowing a reduction in
the laser focal spot and a boost to the focused intensity (Nakatsutsumi et
al. 2010, 2018). The immediate consequence should be to increase the
maximum energy of the produced protons. Most notably, the domain
above 200 MeV of maximum energy should be attainable (Chen et al.,
2018), thus making it possible to significantly increase the throughput of
neutron production by spallation. Spallation is a process that occurs
when a light projectile (proton, neutron, or light ions) with a kinetic
energy from several hundreds of MeV to several GeV, interacts with a
heavy nucleus (e.g., lead) and causes the emission of a large number of
hadrons (mostly neutrons) (Van der Meer et al., 2004).
Overall, multi-PW facilities like Apollon or ELI-NP can expect to pro-
duce >1012 neutrons from the protons in the output of a Pb converter.
Using a conservative estimate of 50% for the proton beam bandwidth, it
will debunch over ~0.7 ns after 50 cm, i.e. by the time it reaches the Pb
spallation converter. Taking this as the duration of the neutron bunch (as
the individual spallation process takes place over a ~10–22 s time-scale),
and a source size imposed by the protons scattering in the Pb target (over
3 mm radius for protons at 250 MeV and 0.5 mm at 1 GeV), the conserva-
tively estimated resulting peak flux will be ~1022–5x1023 neutrons/[cm2.s].
Also, because the repetition rate of the multi-PW lasers will be
improved compared to present-day lasers, the time-averaged neutron flux
should also become quite high. Using a repetition rate of 1 shot/min (for
the largest facilities), the time-averaged flux neutron should reach ~1011–
5x1012 neutrons/[cm2.s]. This will greatly exceed those available on exist-
ing facilities and should permit multiple neutron capture (Couture and
Reifarth, 2007).
We also note that such extreme brightness neutron sources will have a
broad collateral effect aside from the nucleosynthesis application dis-
cussed here. Indeed, they could help satisfy the increasing demand for
neutron sources (Hamm, 2010), e.g. for radioisotope production and even
for more futuristic transmutation applications.
X. EXTREME LIGHT PROPAGATION IN QUANTUM VACUUM

Particle production in “empty” space is a historical path which has guided
the field of Extreme Light with the ultimate goal of investigating
laser-matter interactions in the new regime of ultra-relativistic optics
reaching into the fundamental QED and possibly QCD regimes. The laser
must necessarily be between the Petawatt and Exawatt regime to produce
synchronised, high energy radiation and particle beams with extremely
short time structures in the attosecond and zeptosecond time domain.
These unique characteristics, unattainable by any other means, could be
combined to offer a new paradigm to the exploration of the structure of a
vacuum to respond to one of the most fundamental questions: how can
light propagate in a vacuum; how can a vacuum define the speed of light;
and how can it define the mass of all elementary particles?
A. Light propagation in a vacuum

The vacuum structure defines, for example, how light propagates, as
first noted by Werner Heisenberg and his collaborators. The quantum
fluctuations in the vacuum allow light-light scattering and the conver-
sion of electromagnetic field energy into particle and antiparticle pairs.
In particular, a photon traveling in the vacuum can fluctuate into a par-
ticle-antiparticle pair that is virtual because the energy of a single light
photon is much smaller than that of a material particle pair. This trans-
mutation of a photon into a virtual pair and back, is called vacuum
polarisation since this process also alters the nature of the Coulomb
law at short distance. If a second photon arrives just when the first
photon exists in its electron-positron pair state, it can scatter from this
virtual charged particle pair. In this way, light scatters from light. In
direct extension of this argument, a strong electromagnetic field
applied in the vacuum can deflect the virtual electrons and positrons.
Therefore, there is an index of refraction of empty space filled with
fields. In principle, light can be bent by applied electro-magnetic fields,
just as gravity bends light through the deformation of space-time
geometry. In this sense, empty space has a structure not all that differ-
ent from that of a dilute gas. Because the Compton wavelength of an
electron, h/mc = 386 fm (fermi = femtometer = one in 1015 part of a
metre, the radius of a proton) is 3 million times shorter than a typical
optical wavelength, vacuum structure does not massively obstruct the
propagation of light. However, light propagating in the Universe over
cosmological distances, in the presence of external magnetic fields
experiences nonlinear vacuum effects such as photon splitting. It is
important to recall that more than 50 years ago Julian Schwinger
showed that a coherent ideal ‘plane light wave’ cannot scatter from
itself, or be influenced by itself, no matter what the field intensity is.

This is the only form of light known to which the vacuum is exactly
transparent.
B. Electromagnetic field in a vacuum

The gap between the valence and conduction band of the best insulator,
the vacuum, is twice the energy equivalent of the electron mass, V0 =
2mc2/e = 106 volts. Such high potential differences are commonly achieved
in specialised nuclear accelerators (Tandems, Van der Graaff), however
over a rather large distance. The vacuum does not begin to spark since
the electron-positron pair must materialise on two ends of the potential
well, and this is for laboratory devices a macroscopic distance apart. The
electric field strength controls the speed of vacuum sparking. The field
strength for which this vacuum decay occurs at the zeptosecond scale
(light travels a distance of the Compton wavelength λc = 10–12 m in 10 zep-
toseconds) is the so called ‘Schwinger’ critical strength E0 (named after
Schwinger, though Heisenberg was well aware of the result) for which the
potential step V0 occurs over the electron’s Compton wavelength, that is
E0 = 1.3×1018 V/m. Approach to this field strength is the intermediate term
goal of ELI. This corresponds to an intensity I0 = 4.65×1029 W.cm–2. One
speaks also of laser power when the intensity within the typical focal
domain of 1 μm2 is considered; this value is P0 = 4.65×1021 W, that is 4650
EW (E = exa = 1018) or 4.6 zettawatt.
For fields near the Schwinger value E0 we can observe massive materi-
alisation of pairs for fields existing at time scales of fs to as. The material-
isation of electrical fields into electron-positron pairs is a diagnostic tool
allowing for understanding how well the energy is focused. The abundant
formation of electron-positron pairs is the first of many vacuum effects to
study while striving to focus the laser energy into smaller and smaller vol-
umes.
The stability of the vacuum described above has been studied and eval-
uated at absolute zero temperature. At finite temperature the
valence-conduction particle-hole excitation, here pair production, can be
induced. The temperature has a significant effect on vacuum stability,
which for the attosecond time scale corresponds to the mass of an elec-
tron, k Te = mc2. Still, already at a fraction of this value massive thermally
assisted vacuum decay will be encountered. There are many other ways
which are currently discussed to bridge the 1–2 orders of magnitude in
the field strength associated with the ELI or IZEST project, such as the
Lorentz boost experienced by the electromagnetic field colliding with an
electron beam.
XI. EXTREME LIGHT AS A BLACK HOLE SIMULATOR

Extreme light offers the promise to create a “tabletop” black hole. It could
help to prove whether information is truly lost when black holes evapo-
rate. The idea that information could be lost this way has created a para-
dox in our current understanding of basic physics.
The debate over whether information is really lost during what is called
Hawking evaporation (see Fig. 17) has persisted in the 40 years since Ste-
phen Hawking combined quantum field theory with Einstein’s theory of
general relativity and discovered black hole evaporation. Almost all con-
temporary leading theoretical physicists have participated in this “black
hole war”. In quantum mechanics, the probability, or information, must
be preserved before and after a physical process. The seeming loss of
information as a result of the evaporation of a black hole therefore
implies that general relativity and quantum mechanics, the two pillars of
modern physics, may be in conflict.
So far investigations of this paradox have been mostly theoretical
because of the difficulty of observing black holes in their later stages,
when this potential contradiction is most acute. According to theory, a
solar-size black hole would take 1067 to evaporate entirely, yet our uni-
verse is only about 1010 years old. Therefore, essentially all astrophysical
black holes are too young to provide useful clues on the information loss
paradox even if they are observed, such as that responsible for the gravi-
tational waves observed by LIGO in 2016.
Recently a laboratory black hole to simulate this evaporation has been
conceived (Chen and Mourou, 2017). Using state-of-the-art laser and
nanofabrication technologies, it is projected to mimic black hole evolu-
Figure 17. Accelerating mirror as an analog black hole. Left: Black hole Hawking evapora-
tion and the trapping of the partner modes near the horizon. Right: An accelerating mir-
ror also has a horizon and can also emit Hawking particles and trap their partner modes.
The analogy between these two systems may be appreciated via Einstein’s equivalence
principle.
tions at their later stage and to reveal crucial details on how information
may be preserved during black hole evaporation.
According to Einstein’s equivalence principle, an accelerating mirror
moving near the speed of light shares some common features with a true
black hole (Einstein, 2002). In both cases, there exists an event horizon.
Interacting with quantum fluctuations in vacuum near the horizon, both
will emit Hawking particles and trap their partner modes (Fig. 17) until
the black hole evaporates entirely or the accelerating mirror suddenly
stops. By then the partner modes will be released. The purpose of this
proposed experiment is to see whether and how the Hawking particles
and their partners are entangled and therefore how the information would
be preserved.
It is known that an intense laser traversing a plasma would push the
intercepting plasma electrons to its back, the “plasma wakefields”. Under
extremely intense lasers, such density perturbations can be so concen-
trated that it can serve as a flying reflecting mirror. By properly tailoring
the increase of the density of a thin-film target using nanofabrication
technology, a relativistic plasma mirror accelerates as the driving laser
continues to enter higher density regions. At the time when the laser
leaves the thin-film target, the plasma mirror would abruptly stop its
motion, which mimics the ending of the Hawking evaporation.
XII. TRANSMUTATION
Nuclear power wrestles with the problem of its spent fuel waste that still
needs to be technically addressed (and a societal will needs to be forged).
Among the three scenarios of nuclear waste disposal depicted in Fig. 18,
i.e. (1) The no reprocessing approach (requiring 130,000 years for stor-
age); (2) The spent fuel reprocessing (10,000 year storage requirement);
and (3) The transmutation (which shortens the need for storage to 300
years and reduces the volume requirement by 100 times), one can find a
laser-assisted path that allows for approach (3), which may be the most
environmentally forward-looking.
The vision and its technology based on the combination of a CPA high
power laser amplified within an OPCPA configuration by a pump created
from efficient, high fluence fibre laser technology permits a new high rep-
etition rate laser-driven neutron source for transmutation shown in Fig.
19. The OPCPA laser irradiates a nanometric foil ejecting deuterons onto
a thin target to generate energy efficient neutrons by the fusion of D-T.
The nascent investigation of laser ion acceleration through the physics of
the Coherent Acceleration of Ions by Laser (CAIL) method (Tajima et al.,
2009; Steinke et al., 2010) opens an efficient, compact, and economic
path to the required neutron generation. Here CAIL introduces direct
Figure 18. The radiotoxicity of spent nuclear fuel may be reduced from the level of no re-
processed case by about 1000 times if we transmute the spent fuel. This is why a substan-
tial benefit may be gained by the transmutation. (Gales, 2018; Nifenecker et al., 2003).
ponderomotive acceleration of a thin sheet of electrons by the short-pulse

laser, which in turn directly pulls ions behind the electrons. The condition
of the resonant thinness of the target (and its limited mass) with the
matched laser pulse intensity is the distinguishing feature of CAIL from
the convention approach of Target Normal Sheath Acceleration, TNSA,
(Snavely et al., 2000) in which electrons are heated by the laser absorp-
tion and no specific target mass relation exists to the laser intensity.
Figure 19. Neutrons are generated by an OPCPA laser irradiation of a nanometric deuteron
foil, deuteron acceleration and interaction with tritiated solid target (Yan et al., 2009).
The simulation of the CAIL process has been performed with the EPOCH
PIC code (Brady et al., 2011). A linearly polarised laser of intensity 7.7x1017
W/cm2 and wavelength 1 micron h been focused into a 3 micron beam
footprint; this corresponds to a0 = 0.8 where a0 = eE/(me c ω0) =
0.85λL(I/1018)1/2 where λL is the laser wavelength in μm and I is the laser
intensity in W/cm2. The foil is composed of deuterium with a target den-
sity of 1.0x1023 cm–3. The efficiency of the energy conversion from the
laser to ions of 20% has been shown numerically and it sensitively
depends on the pulse length and target thickness. It has been shown pre-
viously (Steinke et al., 2010) that a0 = σ results in the highest efficiency
where σ = ne/ncr · d/ λL where ne is the electron density, ncr the critical
density (= me ωL2/4πe2 = 1.1x1021 /λ2 μm where ωL is the laser frequency
and λ μm is the laser wavelength in μm), d is foil thickness and λL is laser
wavelength. The theoretically maximum energy is given under the opti-
mal condition of σ = a0 as

(12)
Such research opens the way to aspire for the future application of emerg-
ing laser technology toward the practical transmutation technology of
nuclear waste.
XIII. FEMTOSECOND OPHTHALMOLOGY

CPA-based lasers have attracted significant interest as their potential for
high-precision micromachining due to the low damage threshold and
deterministic character of the interaction. Surgical applications, particu-
larly in the transparent delicate cornea or lens in the case of a cataract,
can take maximal advantage of these attributes. By improving existing
procedures and enabling entirely new ones, femtosecond laser technology
has the potential to become the preferred corneal laser scalpel in the 21st
century.
Corneal laser surgery to correct myopia (short-sightedness), hyperopia
(far-sightedness) and astigmatism is becoming the most commonly per-
formed medical laser procedure. Most techniques rely on altering the cur-
vature of the cornea by removing corneal tissue using direct photoabla-
tion with ultraviolet light from the excimer laser. Two techniques have
demonstrated a high degree of clinical efficacy, photorefractive keratec-
tomy (PRK) and laser-assisted in situ keratomileusis (LASIK).
LASIK has gained recent popularity due to its minimal effect on the
corneal surface, which reduces pain and recovery time. However, LASIK
requires the use of a mechanical blade (microkeratome) to give the exci-

mer laser access to deeper corneal layers. In contrast to photoablative
lasers, photodisruptive lasers operate in the near-infrared spectrum and
are not absorbed (at least to the first order) in ocular media. Near-IR
pulses can pass through transparent and limited thickness translucent
material, affecting tissue only at the focus of the laser beam. In photodis-
ruption, tissue effects are initiated by laser-induced optical breakdown
(LIOB), which requires a small focal spot size to achieve a threshold flu-
ence (energy/area) for plasma formation. Generation of a microplasma
allows the target to absorb additional laser energy.
A. Optical Breakdown Energy Thresholds

An approximately square root dependence of the fluence threshold on the
pulse duration is observed for pulses longer than 10 ps, below this value
the dependence weakens significantly. Recent measurements down to
20-fs pulse durations have again confirmed these observations in corneal
tissue (Loesel et al., 1999). These results suggest an optimal pulse dura-
tion for a corneal photo-disruptive laser in the few hundred femtosecond
pulse duration range, where energy deposited in the tissue is significantly
reduced. See Fig. 20. Further reduction of the laser pulse duration to a
sub 100-fs level adds significant technical complexity and does not pro-
duce any further significant decrease of the threshold.
Figure 20. Intrastomal cut showing

the importance of the pulse shortness:
(a) 50 ps, (b) 100 fs. The picosecond
laser cut is of poorer quality, manual
dissection required to produce corneal
cuts. The femtosecond laser produces
a contiguous cut with surface quality
similar to mechanical blades.
Figure 21. Femtosecond intrastromal scalpel.
B. Corneal Flap Cutting Procedure

(Squier et al., 1995; Ratkay-Traub et al., 2001)
Conventionally, a corneal flap is created with a mechanical microkeratome,
a small motorised knife, to give the excimer laser access to deeper layers of
the stroma during LASIK. For femtosecond laser-flap cutting (Fig. 21), the
laser focus first is scanned along a spiral pattern inside the corneal stroma
at a predetermined depth. The intrastromal cut is followed by several
semi-circular shaped cuts at decreasing depth in the cornea to connect this
intrastromal cut to the corneal surface, with a hinge present to maintain
connection to the cornea. The depth is changed by movement of the focus-
ing objective mediated by a computer-controlled galvo with accuracy of a
few micrometers and speed that can accommodate a 10-kHz repetition
rate. After the completion of the laser procedure, the contact glass is
removed and the flap lifted, similar to what is done in the mechanical pro-
cedure. The technique offers some important advantage.
1. A decrease in LASIK flap complications.
2. Higher precision thickness of flap and residual tissue bed safe thin
flaps.
3. Flexible hinge and bed parameters.
4. Planar flap, wavefront indicates less flap induced aberrations.
Immediate completion of interrupted procedures.
C. Extension of Corneal Flap Cutting to Corneal Transplants

The technique of flap cutting can readily be extended to corneal trans-
plants procedures affecting 45,000 patients in the US. As shown in the
Fig. 22, below, it can advantageously replace the trephines. It is not lim-
ited to full thickness transplants and can make easily partial thickness
ones, 24a. Moreover, it can also create complex shapes, 24b, permitting
self-locking that requires extensive suturing with a reduced healing time.
Figure 22. Today corneal trans-

plant with trephine can only
do a full thickness transplant.
However, a femtosecond scal-
pel makes possible an inter-
locking graft. It shows also that
a femtosecond laser can make
a lamellar slice possible.
XIV. CONCLUSION
CPA has revolutionised the field of optics by providing the largest fields,
the largest pressures, the largest temperatures and accelerations. The
range of its applications has been considerable starting from the eV, the
visible photon energy, and extending to the GeV, possibly the TeV. It is
enlarging the field of optics from atomic to subatomic physics to include
nuclear physics, high energy particle physics, astrophysics and cosmol-
ogy. The applications are remarkably ubiquitous, going from the simple
micromachining and eye surgery to the fundamental with TeV energy, like
astrophysics, cosmology, gravitation or nonlinear QED. At intensities
slightly above the solid or tissue damage threshold, it can be the perfect
scalpel, cutting without any collateral damage. Through its interaction
with gas, plasma, and solids, CPA can become a universal source of high
energy radiations and particles. High harmonic generation in gases can
yield attosecond X-UV pulses. The strength of its wake field in plasmas is
such that it can accelerate particle with formidable gradients of GeV /cm,
thousand times larger than conventional accelerating technique. The
large energy acceleration provided by the laser through the Einstein
equivalence principle, can be used to simulate a black hole and study, for
instance, the loss of information paradox in a black hole. In addition, the
electron being accelerated by a CPA laser in a gas can become by its lat-
eral motion a powerful betatron source that emits radiation in the keV to
100 keV regime. Interacting with a solid target, high energy protons can
be also efficiently produced. At very high intensity, the protons can
become relativistic at an energy of 1 GeV. The protons in turn can pro-
duce neutrons. The latter can find a number of applications in the medical
field like proton therapy or the production of nuclides for nuclear phar-
macology. High energy protons can also be utilised for transmuting radio-
toxic elements such as minor actinides which compose 60,000 metric
tons of nuclear waste worldwide.
The CPA System provides ultra-high peak power in the PW level. How-
ever due to its low repetition rate in the Hz regime, its average power is in
the 10–100 W range; which is far from enough for the applications such
as particle collider or nuclear waste treatment. In addition, it is still
plagued by a major flaw i.e, a very poor wall-plug efficiency of <1%. In this
document we describe the project CAN which has the capability to pro-
vide both peak power, average power and efficiency.
Going to the Schwinger limit and beyond finally, current technology
seems to limit the focused peak intensity to 1025 W/cm2. To go higher, a
novel compression technique is needed and presented. The intensity is
obtained not by an augmentation of the energy but by the reduction of the
laser pulse duration to the single cycle. By using relativistic mirrors, a
gigantic ponderomotive force in the X-ray regime can be created with
pulses as short as a zepto–attosecond. Once tightly focused it could eas-
ily produce intensity at or above the Schwinger limit. The ultra-high
intensity could engender in a solid density a wake field with an accelera-
tion gradient of TeV/cm, opening a fundamentally new vista beyond the
Schwinger regime.
references
1. Akhmanov, S. A., V. A. Vysloukh, and A. S. Chirkin, 1992, Optics of Femtosec-
ond Laser Pulses (American Institute of Physics).
2. Albertazzi, B. et al., Science 346, 325 (2014)
3. Arnould, M. et al., Phys. Rept. 450, 97 (2007)
4. Aoyama, A., J. Ma, Y. Akahane, N. Inoue, H. Ueda, H. Kiriyama, K. Yamakawa,
2002, in Technical Digest of Conference on Lasers and Electro-Optics (CLEO)
2002 “Optical Society of America”, p. 109.
5. Backus, S., C. Durfee, G. Mourou, H. C. Kapteyn, and M. M. Murnane, 1997,
Opt. Lett. 22, 1256.
6. Badziak, J., E. Woryna, P. Parys, K. Y. Platanov, S. Jablonski, L. Ryc, A. B.
Vankov, and J. Wolowski, 2001, Phys. Rev. Lett. 87, 215001.
7. Baetsle, L. H. et al., “Partitioning and transmutation in a strategic perspec-
tive,” Euradwaste, 99, 138 (2000).
8. Baltuška, A., T. Udem, M. Uiberacker, M. Hentschel et al., Nature 421, 611
(2003).
9. Barty, C. P. J., C. L. Gordon III, and B. E. Lemoff, 1994, Opt. Lett. 19, 1442.
10. Beaud, P., M. Richardson, E. Miesak, and B. T. Chai, 1993, Opt. Lett. 18, 1550.
11. Bespalov, V. I., and V. I. Talanov, 1966, JETP Lett. 3, 307.
12. Brabec, T., and F. Krausz, 2000, Rev. Mod. Phys. 72, 545.
13. Brady, C. S. and T. D. Arber, “An ion acceleration mechanism in laser illumi-
nated targets with internal electron density structure,” Plasma Phys. Control.
Fusion, 53, 074004 (2011).
14. Bula, C., K. T. McDonald, E. I. Prebys, et al., 1996, Phys. Rev.Lett. 76, 3116.
15. Bulanov, S. V., and V. S. Khoroshkov, 2002, Plasma Phys. Rep. 28, 453.
16. Bunkenburg, J., et al., 1985, IEEE J. Quantum Electron. QE-17, 1620.
17. Calegari, F., D. Ayuso, A. Trabattoni, L. Belshaw et al., Science 346, 336–339
(2014).
18. Calegari, F., G. Sansone, S. Stagira, C. Vozzi, M. Nisoli, J. Phys. B 49, 062001
(2016).
19. Carman, R. L., R. F. Benjamin, and C. K. Rhodes, 1981, Phys. Rev. A 24, 2649.
20. Cavalieri, A.L., N. Müller, Th. Uphues, V. S. Yakovlev et al., Nature 449, 1029
(2007).
21. Cederbaum, L. S., J. Zobeley, Chem. Phys. Lett. 307, 205 (1999).
22. Chen, P. and G. Mourou, 2017, “Accelerating Plasma Mirrors to Investigate the
Black Hole Information Loss Paradox,” Phys. Rev. Lett. 118, 045001
23. Chen, S.N. et al., Phys. Rev. Lett. 108, 055001 (2012)
24. Chen, S.N. et al., submitted to Matter and Radiation at Extremes (2018).
25. Clayton et al Phys. Plasmas. 1, p1753 (1994).
26. Collier, J., C. Hernandez-Gomez, I. N. Ross, P. Matousek, C. N. Danson, and J.
Walczak, 1999, Appl. Opt. 38, 7486.
27. Collier, J. L., et al., 2004, unpublished.
28. Corkum, P.B., Phys. Rev. Lett. 71, 1994 (1993).
29. Cousens, S., B. Reville, B. Dromey, and M. Zepf, “Temporal Structure of Atto-
second Pulses from Laser-Driven Coherent Synchrotron Emission,” Phys. Rev.
Lett. 116, 1–5 (2016).
30. Couture, A. and R. Reifarth, At. Data Nucl. Data Tables 93, 807 (2007)
31. Cowan, J. J., et al. Astrophys. J. 294, 656 (1985)
32. Ditmire, T., and M. D. Perry, 1993, Opt. Lett. 18, 426.
33. Dubietis, A., G. Jonusauskas, and A. Piskarskas, 1992, Opt. Commun. 88, 437.
34. Dudovich, N., O. Smirnova, J. Levesque, Y. Mairesse et al., Nature Phys. 2, 781
(2006).
35. Duguay, M. A. and G. P. Rentzepis, Appl. Phys. Lett. 10, 350–352 (1967).
36. Einstein, A. and A. Engel. The Collected Papers of Albert Einstein Vol 7: The
Berlin Years: Writings, 1918–1921. Princeton University Press, 2002.
37. Endoh, A., M. Watanabe, N. Sarukura, and S. Watanabe, 1989, Opt. Lett. 14,
353.
38. Esarey, IEEE Trans. Plasma Sci. 24 p252 (1996).
39. Esirkepov, T., M. Borghesi, S. V. Bulanov, G. Mourou, and T. Tajima, 2004,
Phys. Rev. Lett. 92, 175003.
40. Everett et al Nature 368 p527 (1994).
41. Faure et al, Nature 431 p541 (2004).
42. Fourkal, E., B. Shahine, M. Ding, J. S. Li, T. Tajima, and C.-M. Ma, 2002, Med.
Phys. 29, 2788.
43. Fuchs, J. et al., Nature Phys. 2, 48 (2006).
44. Gales, S., “Nuclear Energy and Waste Transmutation with High Power Accel-
erator and Laser Systems.” [Online]. Available: https://indico.cern.ch/
event/617648/contributions/2517094/attachments/1442136/2220662/18_
GALES_IZEST-Talk-Nuclear-Transmutation-040417.pdf. [Accessed: 22-Apr-
2018].
45. Gauthier, J. et al, Eur. Phys. Jour. D, 48, 3, 459-463 (2008).
46. Gauthier, D. et al, Nature Communication, 7, 13688 (2016).
47. Geddes et al, Nature 431 p538 (2004).
48. Godet, J.P. et al, Opt. Lett., 34, 16, 2438-2440, 2009.
49. Gregori, G. et al., Nature 481, 480–483 (2012).
50. Grischkowsky, D., A, C. Ballant, 1982, “Optical pulse compression based on
enhanced frequency chirping,” Applied Physics Letters, 41.
51. Gumbrell, E. T. et al, New. J. Phys. 10, 123011 (2008).
52. Hamm, R. (2010). Paper AP/IA-12, IAEA Proceedings Series, STI/PUB/1433,
ISBN 978-92-0-150410-4.
53. Hentschel, M., R. Kienberger, Ch. Spielmann, G.A. Reider et al., Nature 414,
509 (2001).
54. INTERNATIONAL ATOMIC ENERGY AGENCY, Implications of Partitioning
and Transmutation in Radioactive Waste Management, no. 435. Vienna:
INTERNATIONAL ATOMIC ENERGY AGENCY, 2004.
55. Jones, D.J., S.A. Diddams, J.K. Ranka, A. Stentz et al., Science 288, 635 (2000).
56. Klünder, K., J. M. Dahlström, M. Gisselbrecht, T. Fordell et al., Phys. Rev. Lett.
106, 143002 (2011).
57. Kmetec, J. D., J. J. Macklin, and J. F. Young, 1991, Opt. Lett. 16, 1001.
58. Kraus, P. M., B. Mignolet, D. Baykusheva, A. Rupenyan et al., Science 350, 790
(2015).
59. Larsson, J., P. A. Heimann, A. M. Lindenberg, P. J. Schuk, P. H. Bucksbaum, R.
W. Lee, H. A. Padmore, J. S. Wark, and R. W. Falcone, 1998, Appl. Phys. A:
Mater. Sci. Process. 66, 587.
60. Loesel, F.H., Tien, A.-C., Backus, S., et al., 1999, Proc. SPIE, 3565.
61. Lucchini, M., S. A. Sato, A. Ludwig, J. Herrmann et al., Science 353, 916 (2016).
62. Luk, T. S., A. McPherson, G. Gibson, K. Boyer, and C. Rhodes, 1989, Opt. Lett.
14, 1113.
63. Macchi, A., M. Borghesi, and M. Passoni, ”Ion acceleration by superintense
laser-plasma interaction,” Rev. Mod. Phys. 85, 751–793 (2013).
64. Maine, P., and G. Mourou, 1988, Opt. Lett. 13, 467.
65. Maine, P., D. Strickland, P. Bado, M. Pessot, and G. Mourou, 1987, Rev. Phys.
Appl. 22, 1657.
66. Maine, P., D. Strickland, P. Bado, M. Pessot, and G. Mourou, 1988, IEEE J.
Quantum Electron. 24, 398.
67. Mangles et al, Nature 431 p535 (2004).
68. Martinez, O. E., 1987, IEEE J. Quantum Electron. 23, 1385.
69. Migus, A., C. V. Shank, E. P. Ippen, and R. L. Fork, 1982, IEEE J. Quantum
Electron. QE-18, 101.
70. Mourou, G., T. Tajima, and S. Bulanov, 2006, “Optics in the relativistic
regime,” Rev. Mod. Phys., 78, 2, p 309.
71. Mourou, G., B. Brocklesby, T. Tajima, and J. Limpert, 2013, “The future is fibre
accelerators,” Nature Photonics 7, p 258.
72. Mourou, G., S. Mironov, E. Khazanov, and A. Sergeev, 2014, “Single cycle thin
film compressor opening the door to Zeptosecond-Exawatt physics,” Eur.
Phys. J. Spec. Top., 223, 6, p1181.
73. Mumpower, M.R. et al., Progr. in Part. And Nucl. Phys. 86, 86 (2016).
74. Nakatsutsumi, M. et al., Optics Lett. 35, 2314–2316 (2010).
75. Nakatsutsumi, M. et al., Nat. Comm. 9, 280 (2018).

76. Naumova, N. M., J. A. Nees, I. V. Sokolov, B. Hou, and G. A. Mourou, 2004,
“Relativistic generation of isolated attosecond pulses in a lambda-cubed focal
volume,” Phys. Rev. Lett. 92, 063902.
77. Naumova, N., I. Sokolov, J. Nees, A. Maksimchuk, V. Yanovsky, and G.
Mourou, 2004, Phys. Rev. Lett. 93, 195003.
78. Naumova, N. M., J. A. Nees, B. Hou, G. A. Mourou, and I. V. Sokolov, 2004,
“Isolated attosecond pulses generated by relativistic effects in a wave-
length-cubed focal volume,” Opt. Lett. 29, 778.
79. Nees, J., S. Biswal, F. Druon, J. Faure, M. Nantel, and G. Mourou, 1998, IEEE J.
Sel. Top. Quantum Electron. 4, 376.
80. Nifenecker, H., O. Meplan, and S. David, Accelerator driven subcritical reactors.
CRC Press, 2003.
81. Nisoli, M., S. De Silvestri, O. Svelto, Appl. Phys. Lett. 68, 2793 (1996).
82. Nisoli, M., De Silvestri, S., Svelto, O., Szipöcs, R., Ferencz, K., Spielmann, Ch.,
Sartania, S. & Krausz, F., 1997, “Compression of High-Energy Laser Pulses
below 5fs,” Opt. Letters 22, 522.
83. Oliva, E. et al Nature Photonics 6, 764–767 (2012).
84. Ossiander, M., F. Siegrist, V. Shirvanyan, R. Pazourek et al., Nature Phys. 13,
280 (2017).
85. Panov, I.V., Phys. of Atomic Nuclei 79, 159–198 (2016).
86. Paul, P.M., E.S. Toma, P. Breger, G. Mullot et al., Science 292, 1689 (2001).
87. Perry, M., P. Pennington, B. C. Sutuart, et al., 1999, Opt. Lett. 24, 160.
88. Pessot, M., J. Squier, and G. Mourou, 1989, Opt. Lett. 14, 797.
89. Pessot, M., et al., 1987, Opt. Commun. 62, 419.
90. Potemkin, A. K., M. A. Martyanov, M. S. Kochetkova, and E. A. Khazanov,
2009, “Compact 300 J/ 300 GW frequency doubled neodimium glass laser.
Part I: Limiting power by self-focusing,” IEEE Journal of Quantum Electronics
45, 336.
91. Pretzler, G. et al., Phys. Rev. E 58, 1165 (1998).
92. Ratkay-Traub I1, Juhasz T, Horvath C, Suarez C, Kiss K, Ferincz I, Kurtz R.,
Ultra-short pulse (femtosecond) laser surgery: initial use in LASIK flap crea-
tion. Ophthalmol Clin North Am. 2001 Jun;14 (2):347–55, viii–ix.
93. Reifarth, R. et al., J. Phys. G: Nucl. Part. Phys. 41, 053101 (2014).
94. Remacle, F., R. D. Levine, Proc. Natl. Acad. Sci. U.S.A. 103, 6793 (2006).
95. Remington, B. A., et al. Science 284, 1488 (1999)
96. Ross, I. N., J. L. Collier, P. Matousek, et al., 2000, Appl. Opt. 39, 2422.
97. Ross, I. N., et al., 1997, Opt. Commun. 144, 125.
98. Roth, M., et al., 2001, Phys. Rev. Lett. 86, 436.
99. Rouyer, C., et al., 1993, Opt. Lett. 18, 214.
100. Rynes, J. et al., Nucl. Instrum. Meth. A 422, 895 (1999).
101. Sansone, G., L. Poletto, M. Nisoli, Nature Photon. 5, 655 (2010).
102. Sansone G., F. Kelkensberg, J. F. Pérez-Torres, F. Morales et al., Nature 465,
763-766 (2010).
103. Schafer, K.J., B. Yang, L.F. DiMauro, K.C. Kulander, Phys. Rev. Lett. 70, 1599
(1993).
104. Schiffrin, A., T. Paasch-Colberg, N. Karpowicz, V. Apalkov et al., Nature 493,
70 (2013).
105. Schlaepfer, F., M. Lucchini, S. A. Sato, M. Volkov et al., Nature Phys. 14, 560
(2018).
106. Schoenlein, R. W., S. Chattopadhyay, H. H. W. Chong, et al., 2000, Science 287,

2237.
107. Schultze, M., M. Fieß, N. Karpowicz, J. Gagnon et al., Science 328, 1658 (2010).
108. Schultze, M., E.M. Bothschafter, A. Sommer, S. Holzner et al., Nature 493, 75
(2013).
109. Schultze, M., K. Ramasesha, C.D. Pemmaraju, S.A. Sato et al., Science 346,
1348 (2014).
110. Siegman, A. E., 1986, Lasers University Science Books, Mill Valley, CA”, p. 362.
111. Snavely, R. et al., “Intense high energy proton beams from petawatt-laser irra-
diation of lasers”, Phys. Rev. Lett. 85, 2945 (2000).
112. Squier, J., F. Salin, G. Mourou, and D. Harter, 1991, Opt. Lett. 16, 324.
113. Squier, Du, J., Kurtz, R.M., Elner, V., et al., 1995, in Ultrafast Phenomena IX,
Barbara, P.F. et al., Ed. (New York: Springer), 254.
114. Steinke, S. et al., “Efficient ion acceleration by collective laser-driven electron
dynamics with ultra-thin foil targets,” Laser Part. Beams, 28, 215 (2010).
115. Strickland, A. D., and G. Mourou, 1985, Opt. Commun. 56, 212.
116. Sullivan, A., et al., 1991, Opt. Lett. 16, 1406.
117. Tajima T and Dawson J. M., Phys. Rev. Lett. 43, 267 (1979)
118. Tajima, T., D. Habs, and X. Yan, “Laser Acceleration of Ions for Radiation
Therapy,” Rev. Accel. Sci. Technol., 2, 201 (2009).
119. Telnov, V. I., 1990, Nucl. Instrum. Methods Phys. Res. A 294, 72.
120. Telnov, V. I., 2000, Int. J. Mod. Phys. A 15, 2577.
121. Telnov, V. I., 2001, Nucl. Instrum. Methods Phys. Res. A 472, 43.
122. Teubner, U. and P. Gibbon, Reviews of Modern Physics (2008).
123. Thielemann, F.-K. et al., Progr. in Part. And Nucl. Phys. 66, 346–353 (2011)
124. Treacy, E. B., 1969, IEEE J. Quantum Electron. 5, 454.
125. Tsakiris, G. D. et al, New Journal of Physics, 8, 19 (2006).
126. Uiberacker, M., Th. Uphues, M. Schultze, A.J. Verhoef et al., Nature 446, 627
(2007).
127. Vaillancourt, T. G., B. Norris, J. S. Coe, and G. A. Mourou, 1990, Opt. Lett. 15,
317.
128. Van der Meer, K., et al., Nucl. Instrum. Methods Phys. Res. Sect. B 217 202–220
(2004).
129. Vodungbo, B. et al, Optics Express, 19, 5, 4346 (2011).
130. Wang, Y. et al, Nature Photonics, 8, 381 (2014).
131. Wulff, M., D. Bourgeois, T. Ursby, L. Goir, and G. Mourou, 1997, in Time
Resolved Diffraction, edited by J. R. Helliwell and P. M. Rentzepis (Clarendon
Press, Oxford), pp. 195–228.
132. Yamakawa, K., H. Shiraga, Y. Kato, and C. P. J. Barty, 1991, “Prepulse-free
30-TW, 1-ps Nd:glass laser,” Optics Letters 16, 1593.
133. Yan, X. Q., T. Tajima, M. Hegelich, L. Yin, and D. Habs, “Theory of laser ion
acceleration from a foil target of nanometer thickness,” Appl. Phys. B Lasers
Opt. 98, 711–721 (2010).
134. Yokoya, K., 2000, Nucl. Instrum. Methods Phys. Res. A 455, 25.
135. Zeitoun, P. et al, Nature, 431, 426 (2004).
136. Zhou, M. L., X. Q. Yan, G. Mourou, J. A. Wheeler, J. H. Bin, J. Schreiber, and T.
Tajima, “Proton acceleration by single-cycle laser pulses offers a novel
monoenergetic and stable operating regime,” Phys. Plasmas 23, 43112 (2016).

Mourou Lecture

Uploaded by

Copyright:

Available Formats

Mourou Lecture

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mourou Lecture

Uploaded by

Copyright:

Available Formats

96

Extreme Light Physics

1014 to 1017 have applications that include micromachining, which can be

For intensities >1018 W/cm2 laser-matter interaction becomes strongly

II. THE CPA TECHNIQUE

Figure 1. Laser intensity vs years.

A. Amplification: The Energy Extraction Condition

where ћ is Planck’s constant, ω is the angular laser frequency and σa is

where G0 is the low signal gain and

The gain Gf at the end of the impulsion is given by

B. Amplification: The Propagation Condition

Here, B represents, in units of λ, the amount of wavefront distortion due

For a nonlinear index n2=5×10−16cm2/W for Ti:sapphire. When the laser

For intensities of the order of I = 1 GW/cm2 in Ti:sapphire, Km=200 cm−1,

C. The CPA Concept

D. The Key Element: The Matched Stretcher-Compressor

The phase conjugation properties of the ‘’stretcher/compressor’’ were

E. New Materials for CPA, OPCPA and Gain Narrowing

F. “CAN” a Novel Architecture

G. The Quest for PW Peak Power

Figure 7. World map of lasers with peak power >100TW.

In parallel to the Nd:glass based petawatt systems, we have today a num-

Figure 8. World map of lasers with peak power 1PW.

H. Optical Parametric Chirped Pulse Amplification (OPCPA)

1. Large bandwidth that could accommodate few-cycle pulses.

The disadvantages are as follows:

1. Lower efficiency than standard CPA. For a standard Ti:sapphire

III. BEYOND THE HORIZON: THE ZEPTOSECOND

Figure 10. This element uses a

A. Large Energy Pulse Compression: Thin Film Compression (TFC)

The proposed method for up-converting to shorter wavelengths is

Figure 12. Naumova et al., (2004)

IV. LASER PLASMA ELECTRON ACCELERATION

Figure 13. The Laser Wake

V. HIGH ENERGY PROTON BEAMS, PROTON THERAPY

tages. First, proton beam scattering on atomic electrons is weak and

eration mechanism is found to be highly efficient, and, as we shall explic-

A. Proton acceleration in the single-cycle regime

those instabilities to develop and the constraints caused by instabilities

VI. NEUTRON SOURCES

A. Laser based sources

B. Single target scheme

quasi-mono-energetic forward-directed fast neutron beam. Fusion reac-

C. Double target scheme

A. Incoherent X-ray sources:

B. Coherent X-ray sources:

although only up to a saturation point. Today multi-keV high harmonics

CPA lasers and saturated amplification was demonstrated in 1991, the

VIII. ATTOSECOND SCIENCE

ultra-intense laser facilities for such investigations is that they are

X. EXTREME LIGHT PROPAGATION IN QUANTUM VACUUM

A. Light propagation in a vacuum

itself, or be influenced by itself, no matter what the field intensity is.

B. Electromagnetic field in a vacuum

XI. EXTREME LIGHT AS A BLACK HOLE SIMULATOR

ponderomotive acceleration of a thin sheet of electrons by the short-pulse

XIII. FEMTOSECOND OPHTHALMOLOGY

requires the use of a mechanical blade (microkeratome) to give the exci-

A. Optical Breakdown Energy Thresholds