In this issue NATURE PHOTONICS FOCUS: X-rays
DECEMBER 2010 VOL 4 NO 12
www.nature.com/naturephotonics
A new era for X-rays
QUANTUM OPTICS
Robust entanglement
NANOSCALE CIRCUITS
Rewritable photodetectors
PHOTONIC CRYSTALS
Soliton compression
nphoton.2010.CoverDec.indd 1
16/11/10 17:19:49
REVIEW ARTICLE | FOCUS
published online: 30 noveMber 2010 | doi: 10.1038/nphoton.2010.256
the attosecond nonlinear optics of bright
coherent X-ray generation
tenio popmintchev, Ming-chang chen, paul arpin, Margaret M. Murnane and henry c. Kapteyn*
The frequency doubling of laser light was one of the first new phenomena observed following the invention of the laser over
50 years ago. Since then, the quest to extend nonlinear optical upconversion to ever-shorter wavelengths has been a grand
challenge in laser science. Two decades of research into high-order harmonic generation has recently uncovered several feasible routes for generating bright coherent X-ray beams using small-scale femtosecond lasers. The physics of this technique
combines the microscopic attosecond science of atoms driven by intense laser fields with the macroscopic extreme nonlinear
optics of phase matching, thus essentially realizing a coherent, tabletop version of the Roentgen X-ray tube.
T
he promise of nonlinear optics became clear soon ater the invention of the laser 1, when Peter Franken focused light from a ruby
laser into a quartz crystal and observed how a small amount of
laser light was converted to the second harmonic2. However, the full
impact and realization of nonlinear optics came only with an understanding of how a signal wave generated through a nonlinear optical polarization grows as it propagates through a nonlinear medium.
he beautiful and elegant perturbation treatment of nonlinear optics
developed by Nicolaas Bloembergen and co-workers made it possible to develop both birefringent-phase-matching and quasi-phasematching (QPM) techniques for eiciently converting laser light from
one wavelength to another 3. In the early days, upconverting laser light
to very short wavelengths was considered in the context of perturbative nonlinear optics. In this regime, the nonlinear medium is not
damaged, and the nonlinearity of the medium is a weak perturbation
such that higher orders of nonlinearity become successively weaker.
As a result, the most practical approach for generating vacuum ultraviolet light of around 100 nm was to use excimer lasers or conventional harmonic generation to make near-ultraviolet light, followed
by phase matching using excited-state atomic resonances4.
his situation changed completely in 1987, when Mcpherson et al.
discovered high-harmonic generation (HHG)5. In experiments seeking to characterize perturbative low-order harmonic generation, they
unexpectedly observed a large number of odd harmonics of the fundamental driving laser, corresponding to the coherent combination
of an odd number of photons (3rd, 5th, 7th, 9th, 11th, 13th, 15th
and 17th harmonics). Moreover, the intensity of successively higher
harmonic orders did not decrease signiicantly, as was expected in
perturbative nonlinear optics. hese exciting observations were rapidly reproduced using infrared driving lasers6 that produced an even
more remarkable comb of harmonic frequencies. Basic symmetry
considerations can easily explain why only odd harmonics are generated — a gas medium has an inversion symmetry that precludes
the generation of even-order harmonics. However, a completely new
non-perturbative understanding of nonlinear optics was necessary to
explain why many harmonics — all of comparable intensity — were
generated simultaneously. he unique physics of the HHG process
also necessitated a new perspective, and a new formalism, of phase
matching in dynamic media that could not have been foreseen from
the perspective of perturbative nonlinear optics.
Microscopic attosecond science of hhG
he key to uncovering the attosecond physics underlying HHG is to
understand how atoms respond to strong laser ields. his was a topic
of great interest in the 1980s, when above-threshold multiphoton ionization was observed and studied7. he time-dependent Schrödinger
equation simulation of how a single electron is ionized from an atom
by a strong laser ield, which was successfully used to explain abovethreshold multiphoton ionization, was used by Kulander and coworkers to reproduce the characteristic multiple harmonic spectra
of HHG8. Other work discussed the basic HHG process in as early
as 19829,10, considering it to be a Raman-like process via continuum
states and describing the recollision process in classical terms in the
context of above-threshold multiphoton ionization11.
he next few years saw an intuitive picture of HHG emerge that
is now known as the three-step model. In this picture, an electron is
irst liberated from an atom through strong ield ionization, is then
accelerated by the laser ield, and inally recombines with the parent
ion, emitting any excess energy as a high-energy photon (Fig. 1a). he
explicit relation between the energy of the emitted HHG photon and
the ponderomotive energy (the oscillation energy) of the free electron
in the laser ield can be used to write a simple cut-of rule for the
maximum HHG photon energy that can be generated from a single
atom or ion:
hνSA cut-of = Ip + 3.17 Up ≈ ILλL2
(1)
where Ip is the ionization potential of the gas atom and Up is the
ponderomotive energy of a free electron in an intense ield (that is,
the average energy of an electron driven by the oscillating electric
ield of a laser of intensity IL and wavelength λL). his expression
was irst identiied through quantum simulations12,13, and can also
be derived from Newton’s second law (F = mea = −eEL) applied to a
free electron of charge e, mass me and acceleration a in an electric
ield of magnitude EL (ref. 14). hese models were developed to
be consistent with a variety of observations at the time, and later
rigorous systematic tests15 bore them out. he linear single-atom
scaling of harmonic cut-of against laser intensity means that harmonics can in-principle be generated at very high photon energies, simply by increasing either the intensity or wavelength of the
driving laser 16.
In 1994, the development of a full semi-classical theory 17 established the microscopic quantum picture of HHG. his theory applied
a path-integral approximation to the time-dependent Schrödinger
solution for the electron motion, resulting in an expression for the
time-dependent harmonic dipole that identiied the electron pathway,
JILA, the Department of Physics and the NSF Engineering Research Center in Extreme Ultraviolet Science and Technology, University of Colorado at
Boulder, Boulder, Colorado 80309, USA. *e-mail: henry.kapteyn@colorado.edu
822
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
FOCUS | REVIEW ARTICLE
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
a
Laser field
X-ray
a Macroscopic phase matching
Laser field
Coulomb potential
e–
Constructive addition
of X-ray fields
b Tight focus
c Plane wave
b
Gas-filled waveguide
Laser
Gas jet
Electron
wavefunction
Ion
X-ray beam
Lmedium independent of confocal parameter
Gas-filled cell
~50 Å
Lmedium < confocal parameter
~1–1
,00
0
X-ray field
Å
Lmedium < confocal parameter
X-ray
wavefront
Wavefronts
Laser
wavefront
X-ray
Laser
wavefront wavefront
c
Tunnel ionization
Laser field
Laser beam
X-ray beam
Atomic gas
Atomic
gas
Quantum ionization
Geometrical Guoy phase and CEP
ΔϕGuoy
ΔϕGuoy = 0
Cos
Sin
Cos
–Cos
~Cos
Wavevector picture
Katomic
kX-ray
qkL
kX-ray
Time (cycles)
Figure 1 | the microscopic single-atom physics of hhG. a, Classical
schematic of HHG. The electric field of an intense laser extracts an electron
from an atom through tunnel ionization. The laser field then accelerates the
electron, with a small fraction of the electron returning back to the ground
state of the same atom, liberating its excess energy as a high-energy
photon. b, The quantum nature of HHG. The electron’s wavefunction is
driven by the laser field, giving rise to quantum interference between the
bound and free portions, as well as to transverse spreading. The rapidly
varying time-dependent dipole modulation gives rise to short-wavelength
radiation (courtesy of I. Christov). c, The total fractional ionization of the
medium grows stepwise with each laser cycle. The more realistic quantum
model includes the modification in ionization due to the recolliding electron.
including ionization, propagation and recombination, while preserving the de Broglie-wave quantum nature of the free electron driven
by the laser ield:
t
–
–
–
– * [p– – A
(t)]e–iS(p, t, t')EL(t')dx [p– – AL(t')] + c.c.
x(t) = i∫ dt'∫d3pd
x
L
0
(2)
where x(t) is the electron displacement, p is the momentum, AL is the
vector potential of the driving laser ield, EL(t') is the applied electric
ield and c.c. denotes the complex conjugate. In equation (2), the irst
Katomic
qkL
Figure 2 | Macroscopic phase-matching of hhG in the spatial domain.
a, Phase-matched signal growth ensures that harmonic emission from
many atoms over an extended medium (of length Lmedium) adds together
coherently. b, Tight focusing geometry with associated curved laser and
X-ray wavefronts. The geometrical Gouy phase shift (ΔφGuoy) results in a
dynamical phase change between the carrier wave and envelope of the
laser pulse. In a wave vector picture, the strong intensity gradient leads
to a radial atomic wave vector Katomic that favours non-collinear phase
matching before the focus and collinear phase matching after the focus.
c, Plane-wave propagation in a waveguide or cell minimizes the curvature
of the laser and HHG wavefront. The result is a well-directed, fully phasematched HHG output. The carrier–envelope phase change is absent in
vacuum. The wave vector picture is also significantly simplified in this
geometry, in which only near-collinear phase matching is possible.
factor corresponds to ionization of the atom; that is, a transition from
a bound state into the continuum. he second factor, e−iS(p-,t,t'), where
S is the action integral, takes into account how the quantum phase
of the electron wavefunction evolves over the ~1 fs timescale during
which the electron resides in the continuum. he inal factor represents the transition from the continuum back to the ground state.
Physically, the origin of high harmonics is the rapidly spatially and
temporally modulated electron wavefunction that evolves as an atom
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
823
REVIEW ARTICLE | FOCUS
a
b
t’r
e–
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
X-ray
amplitude
tr
Snapshots in frame
moving at speed c
vL > c
ELinitial
2Lcoh
Lcoh
+
–
3Lcoh
+
4Lcoh
–
Lmedium
Ion
Lmedium
ELfinal
t’b
tb
EX-ray
initial
EX-ray
final
EX-ray
EL
vX-ray = c
Out of phase
c
d
Δk(t) = 0
Lcoh→ ∞ (Lcoh >> Lmedium)
+
vL = c
+
+
+
vX-ray = c
In phase
Figure 3 | Macroscopic phase-matching of hhG in the temporal domain. a,b, Non-phase-matched build-up. Without phase matching, the evolution of the
laser electric field due to geometrical and dynamically changing dispersion leads to a variation in the moment of birth and recombination of the
rescattering electron (for any given energy) (a). The superimposed laser and HHG fields are plotted at the input (ELinitial and E ) and output (ELfinal and E
) of the nonlinear medium, shot in a frame moving with the speed of the X-rays (that is, the speed of light, c). The X-ray amplitude oscillates over a
characteristic length called coherence length (Lcoh), which prevents strong HHG signal build-up (b). c,d, Phase-matched build-up. In dynamically changing
conditions, phase-matching Δk(t) = 0 manifests itself through a minimal phase variation of the driving laser field for a fraction of a laser cycle, within
which the electron can return (c). Under such conditions, the HHG amplitude (that is, the quadratic HHG intensity) grows linearly with distance (d).
ionizes in a strong laser ield (Fig. 1b). his leads to very high order
harmonics in the radiated dipole ield, even when driven by visiblewavelength lasers.
he expression for x(t) in equation (2) can immediately be translated into a polarization response P(t) of the medium with atomic
density n, P(t) = −nex(t), which can be inserted into Maxwell’s wave
equation3. However, the non-instantaneous nature of HHG as a nonlinear process, in which the portion of the electron wavefunction that
is free acquires a quantum phase, means that there is a phase diference
between the driving laser and the HHG dipole that directly appears in
the high-frequency polarization response driving the harmonic emission. his quantum phase diference can be quite large (>>π), because
the electron, when free, can oscillate with a large excursion at a short
de Broglie wavelength. As we discuss below, this quantum or intrinsic
phase can be exploited to manipulate the HHG phase, spectrum and
pulse duration, and also to implement powerful and general QPM
schemes using additional light ields.
Macroscopic phase-matching of hhG
For a nonlinear optical frequency conversion process to be eicient,
the phase velocities of the fundamental driving laser ield, vL, and
the harmonic light, vX-ray, must ideally be matched so that the harmonic emissions from many atoms in the medium add coherently
(Figs 2,3). If the HHG conversion process is not phase-matched, the
high-harmonic signal builds up only over a propagation distance
824
in which the relative phase of the driving laser and harmonic ields
slip by π radians (Fig. 3b). his distance is the coherence length
Lcoh(q) = π/Δkq, where Δkq is the phase mismatch between the driving
laser and harmonic wave vectors for harmonic order q, denoted kω
(or kL) and kqω (or kX-ray), respectively. Moreover, in HHG the nonlinear medium is usually far from transparent and is also being ionized
as the harmonics are emitted (Fig. 1c). hese efects, combined with
the quantum phase shits between the driving laser and the harmonics, result in a dynamically and spatially varying index of refraction
(and dispersion). Finally, HHG uses a gas medium with no birefringence (precluding most phase-matching techniques). As a result, the
prospects of achieving full phase-matching look challenging. he
story of how these challenges were overcome — by developing a new
ield of extreme nonlinear optics to generate a useful lux of coherent
X-rays — is one of the most exciting chapters in the recent progress of
nonlinear optics and quantum physics.
he importance of phase matching to HHG was recognized very
early on. Initial experimental work required a tight focus to reach an
intensity capable of ionizing a noble gas atom. his in turn caused
a large Gouy phase shit through the focus, which is particularly
problematic for HHG because of the large diference in wavelength
and divergence of the harmonics compared with those of the driving laser (Fig. 2b)18. Furthermore, the Gouy phase shit interacts
with the intrinsic quantum phase of the electron19–21. However,
pioneering work by Balcou et al. treated this problem elegantly by
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
FOCUS | REVIEW ARTICLE
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
a
of phase matching emerged from experiments using short-duration
(~25 fs, or <10 optical cycle) driving laser pulses with suicient energy
(~1 mJ) such that a tight focusing geometry was not required24–26.
By propagating the driving laser in a gas-illed hollow waveguide,
a non-diverging plane-wave phase-matching condition could be
maintained over an extended distance, making it possible to buildup a fully spatially coherent and fully phase-matched high-harmonic
beam (Figs 2–4). he waveguide also makes it straightforward to create a region of controlled, conined and well-characterized gas pressure, with optimal diferential pumping at the entrance and exit of the
waveguide. Plane-wave propagation also means that, neglecting loss,
the laser peak intensity is constant along the direction of propagation,
thus avoiding any variation in the intrinsic quantum phase (Fig. 2c).
In such a one-dimensional geometry, the major sources of phase
mismatch are due to both a pressure-dependent (neutral and free-electron plasma dispersion) term, and a constant, pressure-independent
waveguide propagation term. Full phase-matching can be achieved
within some fraction of the optical cycles of the laser pulse by eliminating the phase mismatch, which is given by:
HHG beam profile
b
∆k(t) = kqω ‒ qkω
30 nm
13 nm
3 nm
≈q
4
4
3
3
Intensity (a.u.)
y (mm)
c
2
1
2
1
0
0
0
1
2
x (mm)
3
4
0
1
2
3
4
x (mm)
Figure 4 | spatially coherent hhG emission for applications in imaging. a,
Under phase-matched conditions, a near-perfect Gaussian HHG beam can
be generated. b, Young’s double-pinhole interferograms show near-100%
fringe visibility and full spatial coherence of HHG beams at 30 nm, 13 nm
and 3 nm (refs 32,41,93). c, Also shown is a point-difraction interferometry
measurement to characterize the HHG wavefront, where a pinhole in a
thin-film filter leads to interferences between the residual beam transmitted
through the filter and the diverging wavefront from the pinhole94. Figure c
reproduced with permission from ref. 94, © 2003 OSA.
considering the intrinsic phase as an efective additional wave vector added to the overall phase matching of the HHG process22. his
interplay between geometrical and atomic phase means, for example, that the maximum HHG yield with a laser focused into a gas
jet favours a geometry in which the laser is slowly diverging from
a focus as it propagates through the gas (Fig. 2b)23. In this case, the
Gouy-induced phase mismatch is compensated for by a decreasing
laser intensity, which results in a quantum wave vector opposite to
that of the Gouy phase. he phase mismatch can therefore be minimized over distances corresponding to a small fraction of the confocal parameter.
his method is far from ideal, however, because the intensity of
the driving laser will diverge and eventually fall below the threshold
for generating harmonics. Even more importantly, early work did
not consider any phase mismatch contributions due to the index of
refraction of the gas and the generated plasma, which are typically the
dominant factors for eicient phase matching. he complete picture
{
u211λL
2π
‒P [1 ‒ η(t)] δn ‒ η(t)NatmreλL
λL
4πa2
}
+ ∆kquantum(t)
(3)
In equation (3), q is the harmonic order, u11 is the lowest-order
waveguide mode factor, λL is the centre wavelength of the driving
laser, a is the inner radius of the hollow waveguide, P is the pressure, η(t) is the ionization fraction, re is the classical electron radius,
Natm is the number density of atoms at 1 atm, δn is the diference
between the indices of refraction at the fundamental and harmonic
wavelengths, and Δkquantum is the quantum phase mismatch associated with the intrinsic phase of the rescattering electron. Physically,
the phase velocity of the high harmonics is close to the speed of
light in vacuum, c, because the refractive index for high photon frequencies (far above most atomic resonances) is close to unity. he
phase velocity of the driving laser, however, is altered through both
the static and dynamic terms in equation (3). Although waveguide
propagation increases the phase velocity of the driving laser, the
medium dispersion is dynamic; that is, the neutral gas index slows
the phase velocity, but the phase velocity rapidly increases as free
electrons are created through ionization. Phase matching occurs
during ionization of the pulse (typically on the leading edge or ideally at the peak), for a fractional ionization η, when Δk = 0. In a
waveguide, this occurs at a level of ionization that can be controlled by the pressure and laser intensity in the waveguide. Tuning
the pressure in this geometry therefore provides a clear signature
of phase matching 24–27. A near-plane-wave geometry can also be
realized using a large focus and high laser energy when the confocal parameter of the beam is much longer than the gas medium.
Such a high-energy HHG geometry has been efectively used in
experiments that beneit from high peak power rather than high
repetition rate28–30.
In conventional nonlinear optics, the medium is usually transparent, and relatively long crystal lengths can be used to saturate
the conversion. For HHG, particularly in the extreme-ultraviolet
region of the spectrum, the medium is absorbing, dispersive and
refractive. Re-absorption limits the useful medium length to ~5–10
absorption lengths31. Under these conditions, conversion eiciencies of laser light to individual harmonic orders are ~10–5 per harmonic at photon energies of around 45 eV (using argon gas as a
nonlinear medium; ref. 26) and 100 eV (in helium). hus, a useful
lux of harmonics (~microwatts) can be produced by few-wattlevel laser systems. When phase matched, the harmonics naturally
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
825
REVIEW ARTICLE | FOCUS
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
a
λ L = 0.8 µm
λ L = 1.3 µm
EX-ray(t)
EX-ray(t)
e–
e–
20
20
0
oh
rr
40 adiu
s)
0
60
6
0
0
0
20
L
20
(B
L(
)
ius
rad
4
0
B
r
oh
EL(t)
40
80
4
0
8
0
EL(t)
Phase-matching cut-ofs
5
He
hν PM cut-of ∝ λ 1.7
L
Extreme-UV
Soft X-ray
Ne
1
0.5
Ar
Water
window
0.05
0.8
1
2
3
4
5
6
Water
window
0.4
0.3
He
15
as
Ne
30
as
0.2
Region I
Perfect phase
matching
0.1
λ L = 1.3 µm
0.5
Photon energy (kev)
Region II
QPM
0.1
7 8 9 10
Ar
0
Laser wavelength λL (µm)
Transform-limited pulse duration
0.6
10
Attosecond
pulses
Photon energy (keV)
c Phase-matched HHG spectra
Zeptosecond
pulses
Hard X-ray
b
λ L = 2.0 µm
10
as
20
as
65
as
80
as
0
1
Normalized HHG intensity
1
Figure 5 | phase matching of hhG using mid-infrared lasers. a, As the driving laser wavelength (period) is increased, the electron’s wavefunction spreads
transversely due to quantum difusion, which rapidly reduces the single-atom HHG yield. (courtesy of I. Christov). 1 Bohr radius = 0.53 Å. b, Theoretical
HHG phase matching cut-of as a function of driving laser wavelength for three- (dashed lines) and eight-cycle (solid lines) pulses. This global phasematching picture has been validated experimentally (solid circles) using several laser wavelengths and nonlinear media demonstrating that bright, phasematched HHG emission is possible over broad regions of the X-ray spectrum38–41,43. c, Experimental phase-matched soft-X-ray supercontinua that support
very short isolated attosecond pulses38,39,41,43.
emerge as a fully coherent beam (Fig. 4)32, with pulses in the femtosecond-to-attosecond regime that are perfectly synchronized to
the driving laser. he phase-matched bandwidth can also be broad,
simultaneously spanning many harmonic orders (Fig. 5). Finally,
for driving laser pulses that are 15 fs or shorter, phase matching can
occur only over a single half-cycle of the laser, providing a straightforward way to generate sub-optical-cycle attosecond pulses using
gated phase matching (Fig. 6)33,34. Other approaches for attosecond pulse generation shown in Fig. 6 require very short driving
laser pulses.
he cut-of rule of equation (1) shows that very high harmonic
orders are generated at higher laser intensities and, as a result, at
826
higher ionization levels (Fig. 1c). Moreover, from equation (3) it
is also clear that for ionization levels η higher than some critical
ionization fraction ηcr, there is no way to satisfy Δk = 0. his critical ionization level is given by:
ηcr =
λL2 reNatm
1
1 – 2 +1
q
2πδn
–1
and corresponds to the ionization level at which phase matching is
achieved for large focal diameters (that is, plane-wave propagation).
Above this critical ionization level, the phase mismatch due to the
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
FOCUS | REVIEW ARTICLE
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
b
100
Circularly
polarized
laser field
80
60
Linearly
polarized
laser field
1.0
40
20
–4
–2
0
2
4
Time delay (fs)
Ne
0.4
3
80 as
2
1
0.2
0
–200
0
5
130 as
0
0.2
0
λL = 0.8 µm
10
0.4
4
0.8
0.6
0.8
Contributing
subcycle
Contributing
subcycle
EUV intensity
1.0
Phase (rad)
λL = 0.8 µm
Ar
0.6
–300 –150
0
150
Phase (rad)
Ionization
Laser field
Single HHG burst
EUV intensity
Electron energy (eV)
a
–5
300
Time (as)
200
Time (as)
0
163 as
–4
EUV intensity
ELinitial
20
10
0
initial
EX-ray
final
EX-ray
vX-ray = c
30
20
10
0
2
4
1
λL = 2.0 µm
1
He
210 as
20
10
30
2
–1
0
1
2
Time (fs)
Theory,
multiple pulses
20
10
–1
Time delay (fs)
e
Ar
30
0
4
Time delay (fs)
Theory,
isolated pulse
ELfinal
0 200 400
Time (as)
2
Photoelectron
energy (eV)
Bi-colour field
with shaped
polarization Single HHG
burst
4
Photoelectron
energy (eV)
1.0 Ar
0.8
0.6
0.4
0.2
0
–200
Phase (rad)
EUV intensity
vL = c
λL = 0.8 µm
1
Experiment
30
Phase (rad)
λL = 0.8 µm
HHG bursts
Laser field
Photoelectron
energy (eV)
d
c
0
1
Time delay (fs)
λ L = 1.3 µm
He
10 as
15 as
1.0
0.5
1
0
X-ray intensity
X-ray intensit
ntensity
teensi (a.u.)
intensity
X-ray intensity (a.u.)
0
Ne
20 as
0
1
1
30 as
0
1
Ar
Ne
Ar
65 as
80 as
0.0
0.2
0.3
0.4
0.5
0.6
0
Photon energy (keV)
0
–100
0
Time (as)
100
–100
0
Time (as)
100
Figure 6 | attosecond pulse generation. a, Isolated attosecond pulse generation driven by a few-cycle laser field. The cut-of photon energy varies
significantly on a cycle-by-cycle basis, so that simple spectral filtering can isolate HHG emission from a single electron recollision at the peak of the
laser field95. This approach has yielded isolated extreme-ultraviolet (EUV) pulses as short as 80 as (ref. 96). b, Isolated 130 as EUV pulse generation
using polarization gating97,98. HHG emission is suppressed except during the most intense central cycle of the field, where the laser polarization
is linear. c, If a second laser field of a diferent colour is added to this polarization gating method, even shorter EUV pulses can be generated99. d,
Isolated attosecond pulse generation using gated phase-matching confined to a narrow temporal window near the critical ionization level. This
approach results in isolated 210 as EUV pulses at photon energies around 50 eV, using multicycle driving lasers33. e, Isolated attosecond pulses
through gated infrared phase-matching in the soft-X-ray region of the spectrum. Bandwidths suicient to support 10 ± 1 as pulses are observed at
around 400 eV. Figure reproduced with permission from: a (right), ref. 96, © 2008 AAAS; b (right), ref. 97, © 2006 AAAS; c (left), ref. 100, © 2010
NPG; c (right), ref. 99, © 2010 APS.
plasma cannot be eliminated, even by using very short driving laser
pulses. As higher harmonics are generated at higher levels of ionization, the phase-matching cut-of is always less than the single-atom
cut-of given by equation (1), and is <150 eV even for helium driven
by Ti:Sapphire lasers35. For harmonics generated at ionization levels of >ηcr, the inite phase mismatch due to uncompensated plasma
dispersion reduces the coherent build-up length to the micrometre range, which reduces the yield by several orders of magnitude.
Overcoming this phase-matching limit to generate bright harmonics in the sot- and hard-X-ray regions (required for many applications in spectroscopy and imaging) has therefore been a grand
challenge in extreme nonlinear optics.
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
827
REVIEW ARTICLE | FOCUS
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
c
a
Periodically modulated hollow waveguide
Successive gas jets
2Lcoh
Lcoh
+
Harmonic beam
Laser beam
Elaser
150 μm
0.25 mm
3Lcoh
4Lcoh
+
0
0
ECP1
ECP2
Lmedium
EX-ray
2Lcoh
Lcoh
+
E laser
0
3Lcoh
+
4Lcoh
0
+
Lmedium
λCP << Lcoh
cτ CP ~ Lcoh
EX-ray
b
d
X-ray-generating
laser pulse
+
Argon-filled
waveguide
2Lcoh
Lcoh
3Lcoh
+
—
4Lcoh
—
+
EQCW laser
CP laser pulses
Lmedium
Pulse collision point in plasma
41
45 0
2
cτQCW laser ~ Lmedium
lse
1
pu
29
33
37
Harmonic order
CP
25
λ QCW laser ~ Lcoh
s
+
4
Figure 7 | QpM techniques for extreme nonlinear optics. QPM relies on switching of or phase shifting the HHG emission in zones that contribute
destructively to the signal. a, Geometrical QPM schemes. A modulated waveguide can periodically modulate the peak intensity of the driving laser beam
and hence the HHG phase. A sequence of gas jets can also be used, in which the spacing and density can be varied. b,c, All-optical QPM approaches are
superior to their geometrical predecessors because a versatile periodic light structure can adaptively compensate for any laser propagation dynamics and
can be implemented over many coherence zones. A sequence of weak counter-propagating (CP) pulses (of wavelength λCP, pulse duration τCP and electric
field strength ECP) can create a light structure inside a waveguide that is matched to the coherence length of a particular harmonic, in order to selectively
enhance a quasi-monochromatic HHG bandwidth59,60. d, All-optical grating-assisted phase matching. At higher photon energies, the coherence length
reduces significantly, and the laser electric field itself can be used to modulate the recolliding electron and HHG phase. QCW, quasi-continuous-wave.
Figure a (right) reproduced with permission from ref. 56, © 2007 NPG.
phase-matching of hhG in the kev region
A variety of approaches for phase matching the HHG process at
higher photon energies have been explored, including the QPM
schemes discussed in the next section. However, over the past two
years, a simple method of extending phase matching to the keV
region and beyond has been developed, based on the same phase
matching physics described above but using longer-wavelength
driving lasers. Somewhat paradoxically, by increasing the wavelength of the driving laser, the phase-matching cut-of shits to
shorter wavelengths (Fig. 5).
It had long been understood from equation (1) that because the
HHG cut-of photon energy is proportional to Up ≈ ILλL2, using a
longer-wavelength driving laser should result in a higher singleatom cut-of photon energy. Experimentally, however, most highharmonic experiments used Ti:Sapphire ultrafast lasers operating
at a wavelength of ~0.8 μm. Direct veriication of the wavelength
scaling of the single-atom cut-of was therefore irst tested in 2002,
ater Ti:Sapphire laser and optical parametric ampliier technology had advanced to the point at which mid-infrared (mid-IR)
pulses of suicient energy were practical36. Furthermore, it was
also known that the longer excursion of the oscillating electron
means that the probability of rescattering — and thus the efective
828
nonlinear susceptibility of the process — would decrease rapidly
with increasing driver wavelength (Fig. 5a). Recent theoretical
and experimental studies have veriied a detrimental scaling in
single-atom HHG eiciency of between λL–5 and λL–9, depending
on how the scaling is deined14,37,39,43. his severe eiciency penalty,
in addition to an already painful 10–20% conversion eiciency of
the Ti:Sapphire driver energy to mid-IR optical parametric ampliier light, has limited the prospects for mid-IR laser drivers for
practical HHG sources.
However, this picture changed dramatically when phase
matching was taken into account 38,39. As explained previously,
phase matching is possible only up to a critical ionization level
that depends on the gas species and driving laser wavelength39.
However, the maximum cut-of photon energy of HHG scales
as ILλL2, meaning that for a longer wavelength driving laser, the
laser intensity required to generate a given high-harmonic energy
rapidly decreases. Because the tunnel ionization rate (the irst
step of HHG) is a quasi-d.c. process, it depends exponentially on
the laser intensity but is reasonably independent of laser wavelength. herefore, even for very high harmonic photon energies,
the medium can be less ionized when longer-wavelength lasers
are used, making full phase-matching possible. Taking all the
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
contributions to the phase mismatch into account (equation (3))
a favourable and strong scaling of the phase matching cut-of —
the maximum photon energy that can be generated under ideal
phase matching conditions at a given λL — is predicted to vary as
hνPM cut-of ~ λL1.6–1.7 throughout the X-ray region39,43, which is almost
as favourable as the single-atom cut-of hνSA cut-of ~ λL2. his prediction was recently veriied by several experiments in diferent
laboratories, as shown in Fig. 5b, including the recent demonstration of a fully phase-matched, spatially coherent source with
a bandwidth spanning the entire water window 38–43. Finally, this
generalized picture of phase matching as a function of the laser
wavelength is valid not only for infrared lasers but also for λL in
the visible and ultraviolet regions.
Fortunately, several factors work to countervail the low singleatom yield for HHG driven by mid-IR lasers. he required phase
matching pressures are very high (multi-atmosphere) for pressuretuned phase matching in waveguides. Moreover, gases generally
become more transparent at higher photon energies, except just
above their inner shell electron ionization thresholds. his means
that although each atom is generating less HHG light, there is also
less re-absorption of the generated harmonics by the nonlinear
medium, resulting in a rapidly increasing optimum density length
product (around several absorption lengths). When all the scaling
parameters are included, in the case of helium the predicted conversion eiciency actually increases with increasing HHG photon
energy 39,43. hus, there is a very real prospect of generating bright
harmonic beams in the hard-X-ray region of the spectrum, with
averages powers of tens of microwatts, from a tabletop setup.
Another interesting aspect of mid-IR phase matching is the
large phase-matched bandwidths available39,41,43, spanning an
energy range of more than 300 eV for HHG driven by 2.0 μm
light 41,43. he bright X-ray supercontinuum shown in Fig. 5c is
the broadest coherent bandwidth generated so far using any
light source. Moreover, past work has shown that because phase
matching is conined to only a few half-cycles of the laser — even
when using relatively long six-cycle driving laser pulses — the
harmonics emerge as a single attosecond burst 33,34. Such macroscopic phase-matching gating functions as a femtosecond Pockels
cell. hus, when combined with the reduced attochirp (1/λL) of
longer-wavelength driving lasers, and together with previously
demonstrated chirp-compensation techniques44, it is likely that
pulses as short as 11 ± 1 as can be supported (Fig. 6e). hese
phase-matched bandwidths, ΔEX-ray, are related to the transformlimited X-ray pulse duration by an approximate rule of thumb:
τ
[as] ~ 1.8/ΔEX-ray [keV] (ref. 43). hus, for harmonics driven
by λL > 4 μm light, the phase-matching bandwidth for HHG in
helium is predicted to span over >1.8 keV, providing a coherent
X-ray supercontinuum that is useful for elemental-speciic spectroscopy and imaging, capable of supporting zeptosecond-duration
pulses (Fig. 5b,c). Experiments employing this technique will
probably use the phase coherence and synchronization of diferent
parts of the X-ray continuum emission to excite and probe complex materials and molecular systems.
Finally, we note that the phase matching of HHG at high
photon energies using mid-IR driving lasers generally requires
a waveguide geometry, for several reasons. First, the waveguide
conines the multi-atmosphere-pressure gas through the ends of
the waveguide, with optimal diferential pumping and extended
plane-wave propagation of the driving laser. Second, a longer driving laser wavelength results in a shorter Rayleigh length for the
focused beam. Although in principle one can always choose the
laser focal spot size such that a long enough confocal parameter
is maintained and optimal conversion eiciency is reached, this
results in a pulse energy requirement that quickly becomes prohibitive without a guided geometry.
FOCUS | REVIEW ARTICLE
controlling the quantum phase on attosecond timescales
A remarkable property of HHG is the existence of the large intrinsic quantum phase that arises from the phase advance of the electron wavefunction in the continuum. Although phase matching of
the HHG process can occur regardless of this phase, the fact that it
depends on the laser ield provides a beautiful means of manipulating the HHG process in several useful ways.
Understanding how to take full advantage of the quantum phase
for optimizing high-harmonic emission emerged with a deeper
understanding of the HHG process itself. he irst proposal for quantum manipulation was suggested in 199519, using a geometry that
intersected laser pulses. A few years later, Chang et al.45 took advantage of the interaction between the time-varying chirp of the laser
and the intrinsic chirp in the HHG output due to the quantum phase
to manipulate the spectral bandwidth of the HHG emission. When
harmonics are driven by a positively chirped laser pulse, the laser and
quantum phase subtract, leading to narrow-bandwidth harmonic
peaks. However, for negatively chirped laser driving pulses, the quantum and laser chirps add, leading to near-continuum HHG emission.
hese simple experiments represented some of the irst experiments
in attosecond science, in which time-domain manipulation of the
laser electromagnetic ield incident on a single atom enabled timedomain control of the radiating electron wave function.
he next step in complexity for exploiting the quantum phase
in HHG was to shape the driving laser pulse itself, thus providing
better control of the HHG spectrum. By controlling the nonlinear
chirp of the driving laser pulse, it became possible to sculpt the
nonlinear response of the atom (that is, the electron wavefunction)
to selectively emit a preferred frequency harmonic46. he mechanism for this selective control can be understood in the context
of the recollision picture of HHG47. he intensity of the driving
laser couples quite strongly to the intrinsic quantum phase, as the
de Broglie phase advance of the electron during its excursion is
quite large, increasing with the energy of the emitted harmonics
(an approximate rule-of-thumb is ~1 radian per harmonic order).
herefore, an optimal nonlinear chirp on the driving laser can
interact with the intrinsic quantum phase to allow emission from
a speciic harmonic order from diferent cycles in the laser ield
to add constructively for that harmonic order, while suppressing
adjacent orders. his single-atom efect is known as intra-atom
phase matching, to distinguish it from macroscopic phase matching 47, and clearly shows that the HHG spectrum is in fact a single,
broad-bandwidth structured spectrum — not a series of discrete
peaks representing independent conversion processes48.
QpM in the hhG process
In visible nonlinear optics, QPM is a powerful and widely
employed scheme for regions in which phase matching is not
possible. Instead of matching the phase velocities of the driving
laser and nonlinear polarization throughout a medium, QPM
periodically corrects the phase mismatch every coherence length
(Fig. 7), thus preventing back-conversion of the harmonic light 3.
QPM schemes in the visible region of the spectrum periodically
vary the crystalline structure of the nonlinear material (by periodic polling, for example). he original proposals for QPM also
considered alternating linear and nonlinear materials, so that the
harmonic signal is always generated in phase. Finally, other QPM
schemes considered using graded-index materials to modulate the
harmonic phase. All of these visible-wavelength schemes can be
adapted to the extreme nonlinear optics of HHG. Moreover, the
dependence of the quantum phase of HHG emission on the total
light ield present in the medium makes all-optical QPM schemes
practical and very promising.
Many diferent ways of implementing QPM for HHG have been
proposed. As early as 199449, calculations by Shkolnikov et al.
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
829
REVIEW ARTICLE | FOCUS
proposed that periodically modulating the medium density could
achieve QPM. By 2003, experiments by Paul et al. and Gibson et al.
demonstrated the proposal of Christov, which suggested that by
periodically modulating the diameter of a hollow waveguide, the
resulting laser intensity modulations could impose the necessary
phase modulation on the HHG ield50–54. Other experiments by
Zepf et al.55 used HHG in a very tight-focus geometry in a simple waveguide, thereby exciting very high order modes in the
waveguide that interfere to create an on-axis modulation of the
laser intensity. he idea of achieving QPM through a periodic
modulation of density was also demonstrated by Seres et al.56.
However, laser-induced defocusing, laser mode scrambling and
other geometric limitations together make such QPM schemes
diicult to scale to the very long interaction lengths needed for
good eiciency.
So far, the most powerful approach for achieving QPM in the
HHG process is to superimpose additional light ields on the nonlinear medium (Fig. 7b,c). As explained above, the quantum phase
shit acquired by the electron during its trajectory in the continuum is linear in laser intensity, and can reach tens to hundreds of
radians. hus, by generating harmonics in the presence of a pair of
interfering laser pulses of wavelength λL, spatiotemporal interferences over distances of ~λL/2 (≪Lcoh) then lead to laser intensity
and phase variations that can easily scramble the quantum phase
of the HHG emission, even to the extent that HHG is essentially
turned of, as demonstrated by Voronov et al.57 (based on theoretical suggestions19,58).
When HHG in a certain region can be successfully turned of,
the next step is to implement QPM by sending the correct light
pattern or sequence of counter-propagating pulses through the
medium in the direction opposite to the driving laser (see Fig. 7b,c
and work of Zhang et al.59–61). A gas-illed waveguide is an ideal
geometry for these experiments because it creates an approximately
plane-wave interaction with explicit control over the phase mismatch, and because it is straightforward to align the forwards- and
backwards-travelling beams. Lytle and co-workers demonstrated
that the correct light pattern can be experimentally measured by
sending a single counter-propagating pulse through the medium,
as the harmonic output oscillates when the single pulse travels through the diferent coherent zones in the medium61,62. A
sequence of only four pulses (Fig. 7b) will signiicantly and selectively enhance a particular narrow range of harmonics (700-fold),
as each successive harmonic is generated at a slightly higher ionization level corresponding to a slightly shorter coherence length
and pulse sequence. Using a longer pulse sequence should allow
for even greater spectral selectivity.
he idea that the phase of the high-harmonic ield can be engineered with an additional light ield opens up many new possibilities beyond using a simple counter-propagating pulse sequence to
establish discrete zones in the medium. For example, the HHG
phase can be modulated by the oscillating ield of a single long-duration counter-propagating pulse, as proposed by Cohen et al.63. In
this scheme (Fig. 7d), the electric ield of the backwards-travelling
pulse, when added to the driving pulse, can shit the phase of the
HHG emission, ideally with a periodic phase shit of π between
the peak and trough of the backwards-travelling wave — which is
then matched to the coherence length of the forwards-travelling
pulse. his grating-assisted phase-matching concept will ideally be
implemented with a mid-IR backwards-travelling pulse to match
coherence lengths in the range of micrometres, ideally suited for
multi-keV generation (with very short coherence lengths in the
micrometre range).
Development of these all-optical phase-matching concepts also
led to the realization that they represent special cases of spatiotemporal QPM64. In standard QPM, energy conservation is assumed,
830
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
whereby the energies of the input and the output photons sum
to zero for the nonlinear process, and the spatial structure of the
medium represents a virtual k-vector that can be included in the
conservation of momentum. In-fact, QPM can also be implemented by matching k-vectors and compensating for any energy
mismatch by a temporal modulation of the medium. In the case of
QPM of HHG, the backwards-travelling wave corrects for both the
energy and momentum mismatch. his means that the period of
the backwards-travelling wave does not strictly match the coherence length of the forwards-travelling wave. hese new concepts
of spatiotemporal QPM are quite general, and can also in principle be applied to low-order nonlinear optics. his makes it clear
that the topic of high-order, non-perturbative nonlinear optics is a
non-trivial extension of nonlinear optics that promises to contribute broadly to nonlinear optics and nonlinear science.
QPM schemes are very attractive because they can use any
wavelength for the driving laser, and therefore do not sufer from
the low λL–5–λL–9 dependence of the single-atom yield for mid-IR
driving laser wavelengths. hey remain more complex to implement experimentally than the perfect phase-matching schemes
described above, but promise to play a major role as the ield of
extreme nonlinear optics develops further.
applications of ultrafast coherent X-rays
When considering attosecond timescales, it is important to start by
bearing in mind the Heisenberg uncertainty principle. he timeenergy uncertainty — one form of this relation — gives ΔEΔt ≥ ħ/2,
which in physical units is ΔE [eV] Δt [fs] > 1. In other words,
dynamics on a subfemtosecond timescale will always involve
large energy bandwidths of >1 eV. In this context, the HHG process itself is the premier example of new physics on an attosecond
timescale. A typical spectrum of high-harmonic emission spans
tens-to-hundreds of electron volts in the extreme-ultraviolet/sotX-ray regions — in a single coherent spectrum — with a structure
that indicates complex dynamics and that can be manipulated to
understand the HHG process45–47.
In terms of its application, one can argue that HHG is following
a trajectory similar to that of NMR. NMR started as a basic physical
observation, progressed to spectroscopic applications, then to very
basic imaging, and is now a staple technique for both basic science
and medical imaging. Advances in X-ray science and technology
have resulted in breakthrough discoveries ranging from unravelling the structure of DNA and proteins to visualizing molecular
and material structure. Ultrashort coherent pulses of X-rays are
powerful probes of the nanoworld; X-rays can penetrate and (by
virtue of their short wavelength) image very small objects. By using
elemental absorption edges, X-rays can provide information that
is speciic to each element or chemical species. Ultrashort pulses
of X-rays can also be used to capture even the fastest dynamics of
electrons, giving a tabletop source of coherent X-rays the potential
for a broad range of applications in medicine and nanotechnology.
he many emerging applications of HHG sources are bearing-out this promise, making it possible to capture dynamics in
atoms65–68, molecules69–77, surfaces78–80 and materials81,82, even at
the fastest timescales, and making coherent X-ray imaging 83 with
nanometre resolution possible on a tabletop84–86. Moreover, as the
useful photon energy range for high-harmonic sources expands
considerably over the next few years, a wealth of new applications
will open up in the ields of nano-, bio-, magnetic-, molecular- and
materials science and technology. Finally, the ultrafast time durations of high harmonics — in the femtosecond-to-attosecond (and
soon possibly zeptosecond) regime — makes this source the ultimate strobe light, perfectly synchronized to the driving laser and
capable of capturing the fastest motion in our natural world, even
at the electron level.
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
summary and perspective
he past two decades have seen a revolution in optical science as a new
region of the spectrum has become available to coherent light. he
physics of HHG involves previously inaccessible regimes of energy
and timescales, and despite the fact that attosecond timescales might
be considered to be fundamentally incompatible with dynamics in
atoms and molecules, the high-order harmonic process shows that we
can observe — and usefully manipulate — dynamic processes in these
systems on such timescales. hese developments open up a wide range
of studies that will make use of ultrafast, laser-like X-ray beams for a
multitude of applications across all areas of science and technology. he
useful wavelength range of tabletop HHG sources is poised to expand
signiicantly in the near future. Furthermore, large-scale free-electron
laser X-ray sources87 such as the Linac Coherent Light Source88 are
reaching new regimes of high-intensity X-ray interactions89,90 that will
provide capabilities complimentary to tabletop HHG sources. Other
concepts such as the generation of high-order harmonics by relecting
ultra-intense lasers from a solid surface91,92 also promise to add to the
mix of new sources. he future for coherent X-ray sources over the
next two decades is indeed very bright.
references
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Maiman, T. H. Stimulated optical radiation in ruby. Nature 187,
493–494 (1960).
Franken, P. A., Hill, A. E., Peters, C. W. & Weinreich, G. Generation of optical
harmonics. Phys. Rev. Lett. 7, 118–119 (1961).
Armstrong, J. A., Bloembergen, N., Ducuing, J. & Pershan, P. S. Interactions
between light waves in a nonlinear dielectric. Phys. Rev. A 127,
1918–1939 (1962).
Reintjes, J. F. Nonlinear optical parametric processes in liquids and gases
(Academic Press, 1984).
McPherson, A. et al. Studies of multiphoton production of vacuum-ultraviolet
radiation in the rare gasses. J. Opt. Soc. Am. B 4, 595–601 (1987).
Ferray, M. et al. Multiple-harmonic conversion of 1064-nm radiation in raregases. J. Phys. B 21, L31–L35 (1988).
Agostini, P., Fabre, F., Mainfray, G., Petite, G. & Rahman, N. K. Free-free
transitions following 6-photon ionization of xenon atoms. Phys. Rev. Lett. 42,
1127–1130 (1979).
Kulander, K. C. & Shore, B. W. Calculations of multiple-harmonic conversion
of 1064-nm radiation in Xe. Phys. Rev. Lett 62, 524–526 (1989).
Shore, B. W. & Knight, P. L. Enhancement of high optical harmonics by
excess-photon ionization. J. Phys. B 20, 413–423 (1987).
Gontier, Y. & Trahin, M. Frequency-conversion involving Raman-like
processes above the ionization threshold. IEEE J. Quant. Electron. 18,
1137–1145 (1982).
Kuchiev, M. Y. Atomic antenna. JETP Lett. 45, 404–406 (1987).
Krause, J. L., Schafer, K. J. & Kulander, K. C. High-order harmonic generation
from atoms and ions in the high intensity regime. Phys. Rev. Lett. 68,
3535–3538 (1992).
Kulander, K. C., Schafer, K. J. & Krause, J. L. in Super-intense laser–atom
physics Vol. 316 (eds Piraux, B., L’Huillier, A. & Rzazewski, K.) 95–110
(Plenum, 1993).
Corkum, P. B. Plasma perspective on strong ield multiphoton ionization.
Phys. Rev. Lett. 71, 1994–1997 (1993).
Chang, Z. H., Rundquist, A., Wang, H. W., Murnane, M. M. & Kapteyn, H. C.
Generation of coherent sot X-rays at 2.7 nm using high harmonics.
Phys. Rev. Lett. 79, 2967–2970 (1997).
Spielmann, C. et al. Generation of coherent X-rays in the water window using
5-femtosecond laser pulses. Science 278, 661–664 (1997).
Lewenstein, M., Balcou, P., Ivanov, M. Y., L’Huillier, A. & Corkum, P. B.
heory of high-harmonic generation by low-frequency laser ields.
Phys. Rev. A 49, 2117–2132 (1994).
L’Huillier, A., Schafer, K. J. & Kulander, K. C. High-order harmonic
generation in xenon at 1064 nm: the role of phase matching. Phys. Rev. Lett.
66, 2200–2203 (1991).
Peatross, J., Fedorov, M. V. & Kulander, K. C. Intensity-dependent phasematching efects in harmonic-generation. J. Opt. Soc. Am. B 12,
863–870 (1995).
Salieres, P., L’Huillier, A. & Lewenstein, M. Coherence control of high-order
harmonics. Phys. Rev. Lett. 74, 3776–3779 (1995).
Lewenstein, M., Salieres, P. & L’Huillier, A. Phase of the atomic polarization
in high-order harmonic-generation. Phys. Rev. A 52, 4747–4754 (1995).
FOCUS | REVIEW ARTICLE
22. Balcou, P., Salieres, P., L’Huillier, A. & Lewenstein, M. Generalized phasematching conditions for high harmonics: the role of ield-gradient forces.
Phys. Rev. A 55, 3204–3210 (1997).
23. Balcou, P. & L’Huillier, A. Phase-matching efects in strong-ield harmonic
generation. Phys. Rev. A 47, 1447–1459 (1993).
24. Zhou, J., Peatross, J., Murnane, M. M., Kapteyn, H. C. & Christov, I. P.
Enhanced high-harmonic generation using 25 fs laser pulses. Phys. Rev. Lett.
76, 752–755 (1996).
25. Schafer, K. J. & Kulander, K. C. High harmonic generation from ultrafast
pump lasers. Phys. Rev. Lett. 78, 638–641 (1997).
26. Rundquist, A. et al. Phase-matched generation of coherent sot X-rays. Science
280, 1412–1415 (1998).
27. Durfee, C. G. et al. Phase matching of high-order harmonics in hollow
waveguides. Phys. Rev. Lett. 83, 2187–2190 (1999).
28. Takahashi, E., Nabekawa, Y., Otsuka, T., Obara, M. & Midorikawa, K.
Generation of highly coherent submicrojoule sot X-rays by high-order
harmonics. Phys. Rev. A 66, 021802 (2002).
29. Mashiko, H., Suda, A. & Midorikawa, K. Focusing coherent sot-X-ray
radiation to a micrometer spot size with an intensity of 1014 W/cm2. Opt. Lett.
29, 1927–1929 (2004).
30. Kazamias, S. et al. Global optimization of high harmonic generation.
Phys. Rev. Lett. 90, 193901 (2003).
31. Constant, E. et al. Optimizing high harmonic generation in absorbing gases:
Model and experiment. Phys. Rev. Lett. 82, 1668–1671 (1999).
32. Bartels, R. A. et al. Generation of spatially coherent light at extreme
ultraviolet wavelengths. Science 297, 376–378 (2002).
33. homann, I. et al. Characterizing isolated attosecond pulses from hollow-core
waveguides using multi-cycle driving pulses. Opt. Express 17,
4611–4633 (2009).
34. Sandhu, A. S. et al. Generation of sub-optical-cycle, carrier–envelope–phaseinsensitive, extreme-UV pulses via nonlinear stabilization in a waveguide.
Phys. Rev. A 74, 061803 (2006).
35. Rundquist, A. Phase-matched generation of coherent, ultrafast X-rays using
high harmonics (Washington State Univ., 1998).
36. Shan, B. & Chang, Z. H. Dramatic extension of the high-order harmonic cutof
by using a long-wavelength driving ield. Phys. Rev. A 65, 011804 (2002).
37. Tate, J. et al. Scaling of wave-packet dynamics in an intense midinfrared ield.
Phys. Rev. Lett. 98, 013901 (2007).
38. Popmintchev, T. et al. Extended phase matching of high harmonics driven by
mid-infrared light. Opt. Lett. 33, 2128–2130 (2008).
39. Popmintchev, T. et al. Phase matching of high harmonic generation in the
sot and hard X-ray regions of the spectrum. Proc. Natl Acad. Sci. USA 106,
10516–10521 (2009).
40. Takahashi, E. J., Kanai, T., Ishikawa, K. L., Nabekawa, Y. & Midorikawa,
K. Coherent water window X-ray by phase-matched high-order harmonic
generation in neutral media. Phys. Rev. Lett. 101, 253901 (2008).
41. Chen, M.-C. et al. Bright, coherent, ultrafast sot X-ray harmonics spanning
the water window from a tabletop light source. Phys. Rev. Lett. 105,
173901 (2010).
42. Xiong, H. et al. Generation of a coherent X-ray in the water window region
at 1 kHz repetition rate using a mid-infrared pump source. Opt. Lett. 34,
1747–1749 (2009).
43. Popmintchev, T. Tunable coherent ultrafast light in the sot and hard X-ray
regions of the spectrum. PhD thesis, Univ. Colorado (2009).
44. Lopez-Martens, R. et al. Amplitude and phase control of attosecond light
pulses. Phys. Rev. Lett. 94, 033001 (2005).
45. Chang, Z. et al. Temporal phase control of sot-X-ray harmonic emission.
Phys. Rev. A 58, R30–R33 (1998).
46. Bartels, R. et al. Shaped-pulse optimization of coherent emission of highharmonic sot X-rays. Nature 406, 164–166 (2000).
47. Christov, I. P., Bartels, R., Kapteyn, H. C. & Murnane, M. M. Attosecond
time-scale intra-atomic phase matching of high harmonic generation.
Phys. Rev. Lett. 86, 5458–5461 (2001).
48. Bartels, R. A., Murnane, M. M., Kapteyn, H. C., Christov, I. & Rabitz, H.
Learning from learning algorithms: Application to attosecond dynamics of
high-harmonic generation. Phys. Rev. A 70, 043404 (2004).
49. Shkolnikov, P. L., Lago, A. & Kaplan, A. E. Optimal quasi-phase-matching
for high-order harmonic generation in gases and plasmas. Phys. Rev. A 50,
R4461–R4464 (1994).
50. Christov, I. P., Kapteyn, H. C. & Murnane, M. M. Quasi-phase matching of
high-harmonics and attosecond pulses in modulated waveguides.
Opt. Express 7, 362–367 (2000).
51. Paul, A. et al. Quasi-phase-matched generation of coherent extremeultraviolet light. Nature 421, 51–54 (2003).
52. Gibson, E. A. et al. Coherent sot X-ray generation in the water window with
quasi-phase matching. Science 302, 95–98 (2003).
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.
831
REVIEW ARTICLE | FOCUS
53. Paul, A. et al. Phase-matching techniques for coherent sot X-ray generation.
IEEE J. Quant. Electron. 42, 14–26 (2006).
54. Christov, I. P., Kapteyn, H. C. & Murnane, M. M. Dispersion-controlled
hollow core iber for phase matched harmonic generation. Opt. Express 3,
360–365 (1998).
55. Zepf, M., Dromey, B., Landreman, M., Foster, P. & Hooker, S. M. Bright
quasi-phase-matched sot-X-ray harmonic radiation from argon ions.
Phys. Rev. Lett. 99, 143901 (2007).
56. Seres, J. et al. Coherent superposition of laser-driven sot-X-ray harmonics
from successive sources. Nature Phys. 3, 878–883 (2007).
57. Voronov, S. L. et al. Control of laser high-harmonic generation with
counterpropagating light. Phys. Rev. Lett. 87, 133902 (2001).
58. Birulin, A. V., Platonenko, V. T. & Strelkov, V. V. Generation of high-order
harmonics in interfering waves. Zh. Eksp. Teor. Fiz. 110,
63–73 (1996).
59. Zhang, X. H. et al. Quasi-phase-matching and quantum-path control of
high-harmonic generation using counterpropagating light. Nature Phys. 3,
270–275 (2007).
60. Lytle, A. L. et al. Quasi-phase matching and characterization of high-order
harmonic generation in hollow waveguides using counterpropagating light.
Opt. Express 16, 6544–6566 (2008).
61. Lytle, A. L. et al. Probe of high-order harmonic generation in a hollow
waveguide geometry using counterpropagating light. Phys. Rev. Lett. 98,
123904 (2007).
62. Zhang, X., Lytle, A. L., Cohen, O., Murnane, M. M. & Kapteyn, H. C.
Quantum-path control in high-order harmonic generation at high photon
energies. New J. Phys. 10, 025021 (2008).
63. Cohen, O. et al. Grating-assisted phase matching in extreme nonlinear optics
Phys. Rev. Lett. 99, 053902 (2007).
64. Bahabad, A., Murnane, M. M. & Kapteyn, H. C. Quasi-phase-matching of
momentum and energy in nonlinear optical processes. Nature Photon. 4,
570–575 (2010).
65. Schins, J. M. et al. Observation of laser-assisted anger decay in argon.
Phys. Rev. Lett. 73, 2180–2183 (1994).
66. Drescher, M. et al. Time-resolved atomic inner-shell spectroscopy. Nature
419, 803–807 (2002).
67. Baltuska, A. et al. Attosecond control of electronic processes by intense light
ields. Nature 421, 611–615 (2003).
68. Loh, Z. H. et al. Quantum state-resolved probing of strong-ield-ionized
xenon atoms using femtosecond high-order harmonic transient absorption
spectroscopy. Phys. Rev. Lett. 98, 143601 (2007).
69. Velotta, R., Hay, N., Mason, M. B., Castillejo, M. & Marangos, J. P. High-order
harmonic generation in aligned molecules. Phys. Rev. Lett. 87, 183901 (2001).
70. Itatani, J. et al. Tomographic imaging of molecular orbitals. Nature 432,
867–871 (2004).
71. Zhou, X. B. et al. Elliptically polarized high-order harmonic emission from
molecules in linearly polarized laser ields. Phys. Rev. Lett. 102,
073902 (2009).
72. Le, A. T., Lucchese, R. R., Tonzani, S., Morishita, T. & Lin, C. D. Quantitative
rescattering theory for high-order harmonic generation from molecules.
Phys. Rev. A 80, 013401 (2009).
73. Baker, S. et al. Probing proton dynamics in molecules on an attosecond time
scale. Science 312, 424–427 (2006).
74. Gagnon, E. et al. Sot X-ray driven femtosecond molecular dynamics. Science
317, 1374–1378 (2007).
75. Sandhu, A. S. et al. Observing the creation of electronic Feshbach resonances
in sot X-ray-induced O2 dissociation. Science 322, 1081–1085 (2008).
76. Wagner, N. L. et al. Monitoring molecular dynamics using coherent
electrons from high harmonic generation. Proc. Natl Acad. Sci. USA 103,
13279–13285 (2006).
77. Li, W. et al. Time-resolved dynamics in N2O4 probed using high harmonic
generation. Science 322, 1207–1211 (2008).
78. Miaja-Avila, L. et al. Laser-assisted photoelectric efect from surfaces.
Phys. Rev. Lett. 97, 113604 (2006).
832
NATURE PHOTONICS doi: 10.1038/nphoton.2010.256
79. Saathof, G., Miaja-Avila, L., Aeschlimann, M., Murnane, M. M. & Kapteyn, H. C.
Laser-assisted photoemission from surfaces. Phys. Rev. A 77, 022903 (2008).
80. Cavalieri, A. L. et al. Attosecond spectroscopy in condensed matter. Nature
449, 1029–1032 (2007).
81. La-O-Vorakiat, C. et al. Ultrafast demagnetization dynamics at the M edges
of magnetic elements observed using a tabletop high-harmonic sot X-ray
source. Phys. Rev. Lett. 103, 257402 (2009).
82. Siemens, M. E. et al. Quasi-ballistic thermal transport from nanoscale
interfaces observed using ultrafast coherent sot X-ray beams. Nature Mater.
9, 26–30 (2010).
83. Miao, J. W., Charalambous, P., Kirz, J. & Sayre, D. Extending the methodology
of X-ray crystallography to allow imaging of micrometre-sized noncrystalline specimens. Nature 400, 342–344 (1999).
84. Sandberg, R. L. et al. Lensless difractive imaging using tabletop coherent
high-harmonic sot-X-ray beams. Phys. Rev. Lett. 99, 098103 (2007).
85. Sandberg, R. L. et al. High numerical aperture tabletop sot X-ray difraction
microscopy with 70-nm resolution. Proc. Natl Acad. Sci. USA 105,
24–27 (2008).
86. Ravasio, A. et al. Single-shot difractive imaging with a table-top femtosecond
sot X-ray laser-harmonics source. Phys. Rev. Lett. 103, 028104 (2009).
87. Ayvazyan, V. et al. First operation of a free-electron laser generating GW
power radiation at 32 nm wavelength. Eur. Phys. J. D 37,
297–303 (2006).
88. Emma, P. et al. First lasing and operation of an angstrom-wavelength freeelectron laser. Nature Photon. 4, 641–647 (2010).
89. Hoener, M. et al. Ultraintense X-ray induced ionization, dissociation, and
frustrated absorption in molecular nitrogen. Phys. Rev. Lett. 104,
253002 (2010).
90. Young, L. et al. Femtosecond electronic response of atoms to ultra-intense
X-rays. Nature 466, 56–61 (2010).
91. Dromey, B. et al. Bright multi-keV harmonic generation from relativistically
oscillating plasma surfaces. Phys. Rev. Lett. 99, 085001 (2007).
92. Teubner, U. & Gibbon, P. High-order harmonics from laser-irradiated plasma
surfaces. Rev. Mod. Phys. 81, 445–479 (2009).
93. Zhang, X. et al. Highly coherent light at 13 nm generated by use of
quasi-phase-matched high-harmonic generation. Opt. Lett. 29,
1357–1359 (2004).
94. Lee, D. G., Park, J. J., Sung, J. H. & Nam, C. H. Wave-front phase
measurements of high-order harmonic beams by use of point-difraction
interferometry. Opt. Lett. 28, 480–482 (2003).
95. Christov, I. P., Murnane, M. M. & Kapteyn, H. C. High-harmonic generation
of attosecond pulses in the ‘single- cycle’ regime. Phys. Rev. Lett. 78,
1251–1254 (1997).
96. Goulielmakis, E. et al. Single-cycle nonlinear optics. Science 320,
1614–1617 (2008).
97. Sansone, G. et al. Isolated single-cycle attosecond pulses. Science 314,
443–446 (2006).
98. Corkum, P. B., Burnett, N. H. & Ivanov, M. Y. Subfemtosecond pulses.
Opt. Lett. 19, 1870–1872 (1994).
99. Gilbertson, S. et al. Isolated attosecond pulse generation using multicycle
pulses directly from a laser ampliier. Phys. Rev. A 81, 043810 (2010).
100. Pfeifer, T. Easing attosecond generation. Nature Photon. 4, 417–418 (2010).
acknowledgements
he authors thank S. Backus of KMlabs and I. Christov of Soia University. Several past
students and postdocs also made critical contributions to the work discussed here,
including A. Bahabad, R. Bartels, Z. Chang, O. Cohen, C. Durfee, E. Gibson, A. Lytle,
J. Peatross, A. Rundquist, X. Zhang and J. Zhou. his work was funded by the National
Science Foundation, a National Security Science and Engineering Faculty Fellowship, the
US Department of Energy (DOE) and the DOE National Nuclear Security Agency.
additional information
he authors declare competing inancial interests: details accompany the paper at
www.nature.com/naturephotonics.
nature photonics | VOL 4 | DECEMBER 2010 | www.nature.com/naturephotonics
© 2010 Macmillan Publishers Limited. All rights reserved.