Standoff spectroscopy via remote generation
of a backward-propagating laser beam
Philip R. Hemmera, Richard B. Milesb, Pavel Polynkinc, Torsten Sieberta,d, Alexei V. Sokolova,
Phillip Spranglee, and Marlan O. Scullya,b,c,1
aTexas A&M University, College Station, TX 77843;
dFreie Univeristät, 14195 Berlin, Germany; and
bPrinceton University, Princeton, NJ 08544;
eNaval Research Lab, Washington, DC 20375
cUniversity of Arizona, Tucson, AZ 85721;
Contributed by Marlan O. Scully, October 5, 2010 (sent for review August 26, 2010)
In an earlier publication we demonstrated that by using pairs of
pulses of different colors (e.g., red and blue) it is possible to excite
a dilute ensemble of molecules such that lasing and/or gain-swept
superradiance is realized in a direction toward the observer. This
approach is a conceptual step toward spectroscopic probing at a
distance, also known as standoff spectroscopy. In the present
paper, we propose a related but simpler approach on the basis
of the backward-directed lasing in optically excited dominant con-
stituents of plain air, N2and O2. This technique relies on the remote
generation of a weakly ionized plasma channel through filamenta-
tion of an ultraintense femtosecond laser pulse. Subsequent appli-
cation of an energetic nanosecond pulse or series of pulses boosts
the plasma density in the seed channel via avalanche ionization.
Depending on the spectral and temporal content of the driving
pulses, a transient population inversion is established in either
nitrogen- or oxygen-ionized molecules, thus enabling a transient
gain for an optical field propagating toward the observer. This
technique results in the generation of a strong, coherent, counter-
propagating optical probe pulse. Such a probe, combined with a
wavelength-tunable laser signal(s) propagating in the forward
direction, provides a tool for various remote-sensing applications.
The proposed technique can be enhanced by combining it with the
gain-swept excitation approach as well as with beam shaping and
adaptive optics techniques.
air laser ∣ atmospheric surveillance ∣ threat detection ∣ Raman
tant and challenging problem, with applications in environmental
science and national security. Measurements using light detection
and ranging (LIDAR) techniques coupled with differential ab-
sorption LIDAR have been reported (1, 2). These techniques
provide valuable tools for the measurement of trace impurities
in the atmosphere.
However, to enhance sensitivity and information content re-
trieved by the return signal, a different technology is needed.
In ref. 3, we presented standoff spectroscopy (SOS) technique
for detection of trace impurities in the atmosphere via gain-swept
superradiance (4). In that scheme, we first pump the molecules
of interest, e.g., nitric oxide or phosgene, at some prearranged
distance, say, 300 m. These molecules decay spontaneously to a
lower state via a specific radiation frequency. Then the impurity
molecules at, say, 290 m, are excited at a later time that is delayed
from the first pulse by Δτ ¼10m
realized in the second region. Subsequent regions of inversion
are generated by later pulse pairs as in Fig. 1. In ref. 3, which will
be briefly reviewed in Gain-Swept Lasing of Impurities in Air, the
impurity molecules themselves constitute the lasing medium. In
the present paper, we propose a simpler alternative approach that
is based on transient backward-directed lasing in atmospheric
gases such as N2and O2on the far side of the region to be probed.
The basic idea behind the remote generation of a back-propa-
gating laser probe beam is illustrated in Fig. 2. This technique
he continuous monitoring of the atmosphere for traces of
gases and pathogens at kilometer-scale distances is an impor-
c≅3 × 10−8s, so that some gain is
nels via controlled prechirping of intense femtosecond laser
pulses at the launch site. These pulses self-compress through
natural dispersion of air as they propagate. At a predetermined
remote location, the peak power of these pulses exceeds the
threshold for self-focusing in the air. The pulses collapse transver-
sely and form laser filaments, leaving weakly ionized plasma
channels in their wake. A high-energy, nanosecond laser pulse
that immediately follows the seed pulse deposits energy into the
plasma channel and creates conditions for optical gain in the
The above approach is particularly plausible if applied tolasing
ing has been previously demonstrated at atmospheric pressure by
using electrical discharge and microwave pumping and detailed
analysis shows that such a filamentation N2air laser is feasible.
Another possibility is to utilize various visible and near-infrared
transitions in the ionized oxygen, although this route is more
at gas pressures substantially lower than atmospheric pressure.
However, oxygen lasing in realistic open-air conditions has now
been demonstrated (5). An additional degree of control over
optical excitation of atmospheric gases can be offered by spatial
beam shaping and adaptive optics.
The utilization of the strong backward-propagating optical
improvement of the detection sensitivity compared to the stan-
dard LIDAR techniques. LIDAR is based on the detection of
the nearly isotropic scattering of a forward-propagating probe
between pulses in each pair is decreasing. The second pulse in each pair
has a higher velocity because of atmospheric dispersion. The first pair of
pulses overlaps near the back of the cloud, creating a small region of gain.
Subsequent pairs overlap at closer and closer regions of the cloud, producing
a swept-gain amplifier that lases back toward the observer.
SOS. Multiple pairs of pulses are generated such that the spacing
Author contributions: P.R.H., R.B.M., P.P., T.S., A.V.S., P.S., and M.O.S. designed research;
P.R.H., R.B.M., P.P., T.S., A.V.S., P.S., and M.O.S. performed research; and P.P., T.S., A.V.S.,
and M.O.S. wrote the paper.
The authors declare no conflict of interest.
Freely available online through the PNAS open access option.
1To whom correspondence should be addressed. E-mail: email@example.com.
3130–3134 ∣ PNAS ∣ February 22, 2011 ∣ vol. 108 ∣ no. 8www.pnas.org/cgi/doi/10.1073/pnas.1014401107
beam. The component of this scattering that propagates toward
the observer is a priori weak. On the other hand, the sensing
approach proposed here employs a bright and highly directional
backward-propagating probe, thus increasing the signal-to-noise
ratio of the detected signal dramatically. Potential schemes of
remote spectroscopic analysis using a back-propagating coherent
optical probe are described in Experimental Plan.
Gain-Swept Lasing of Impurities in Air
Let us briefly summarize the ideas of ref. 3, which provides the
conceptual background for the present work. As depicted in
Fig. 1, the key idea is the generation of a gain-swept excitation
via pulse “catch up” due to atmospheric dispersion. Consider
the initial pulse pair that overlaps at the far side of the cloud.
We arrange such conditions that the first pulse in the pair is at
a higher frequency (e.g., λ1¼ 400 nm), and the second pulse is
at a lower frequency (e.g., λ2¼ 580 nm). The two pulses have
different group velocities because of the dispersion of air. The
initial delay time Δt12between the first two pulses determines
the distance z at which the faster-moving lower-frequency pulse
catches up with the slower-moving higher-frequency pulse.
The group velocity in air is given by
1 þ λdn
where n is the refractive index of air whose dependence on the
wavelength is given by the following formula:
λ2½nm?þ3.777 × 104
For example, the difference in group velocities of the two pulses
for the case when λ1¼ 400 nm and λ2¼ 580 nm is Δvg≈ 4×
105cm∕s. Thus the pulses will overlap at a time τ such that the
initial pulse separation ΔL ¼ Δvgτ. Hence, for ΔL ¼ 1 mm,
τ≅2 × 10−5s; i.e., the pulses will overlap at a distance of about
5 km from the transmitter.
The following pulse pairs are to be delayed so as to generate a
population inversion swept back toward the sender. The relaxa-
tion time of the population inversion caused by collisions in the
air is in the nanosecond range. Therefore, each pair of visible
pump pulses should follow the previous pair by the time Trep
on the order of 1 ns.
This swept-gain configuration yields a beam of exponentially
amplified spontaneous radiation that is gain-guided in the back-
ward direction. Because, in principle, the gain path can be made
arbitrarily long, it is possible to generate a strong signal from only
partsper million concentration ofthe impurity in the atmosphere.
Lasing of N2and O2in Air
The gain-swept approach described above turns the analyte to be
sensed into a wavelength-selective active optical medium. This
approach is very powerful because it provides for the ultimate
control over temporal and spatial distributions of optical gain
throughout the sensing region. However, the practical implemen-
tation of this technique will be extremely challenging because of
the overall complexity of the approach. Each particular analyte
gas will have its own unique lasing parameters. Therefore, the
pulse-shaping apparatus involved in the generation of adequate
pulse sequences will either have to be dedicated to sensing a par-
ticular gaseous substance at a particular range or be excessively
complex. A simpler SOS scheme is needed to make the overall
approach more practical. This simpler alternative that we pro-
pose here relies on using major constituents of air, such as nitro-
gen and oxygen, as gain media for the generation of a backward-
propagating coherent optical probe. This probe will be used in
combination with a tunable forward-propagating laser beam(s)
for a remote Raman-type spectroscopic analysis of trace amounts
of various gases in the atmosphere.
In the gain-swept scenario, the extremely low concentration
of the analyte that at the same time was acting as the laser gain
medium was translated into the very low per-unit length value of
gain. Consequently, a long gain path was required. In the case of
the dominant constituents of the ambient air, the gas concentra-
tion is orders of magnitude higher, and a much shorter gain path
(of the order of 1 m or even less) would suffice. In fact, there are a
variety of commercial gas lasers, including a molecular nitrogen
laser (6), that are compact and efficient devices. These lasers can
serve as models, which we will attempt to emulate in ambient
air. The main challenge in such emulation will be to replace the
discharge excitation mechanism used in the overwhelming major-
ity of practical gas lasers with a remote optical excitation using
laser beams. In addition, contrary to commercial gas lasers, there
will be no freedom in optimizing the gas pressure and composi-
tion. The conditions for the creation of population inversion will
have to be realized in plain air under atmospheric pressure,
whereas most commercial gas lasers operate at much lower pres-
sures and utilize particular selected mixtures of gases.
Central to the alternative SOS strategy, on the basis of lasing
in the major constituents of the ambient air, is the process of fem-
tosecond laser filamentation (7–9). When a light beam with peak
power exceeding the self-focusing threshold (∼3 GW ¼ cm2)
propagates in the air, its transverse intensity profile shrinks
because of self-focusing. In the absence of any defocusing
mechanism, the self-focusing would result in a collapse of the
beam profile into a singularity. What stops the collapse is the
defocusing action of plasma generated on the beam axis via multi-
photon ionization. The peculiar object that is composed of the
hot and intense core and the generated plasma on the beam axis
is termed the laser filament. Laser filaments in air are typically
about 100 μm in diameter. They have been shown to propagate
over distances orders of magnitude longer than the Raleigh range
corresponding to their transverse dimensions.
The light intensity in the filament core is clamped to the level
approximately equal to the multiphoton ionization threshold,
about 1013–1014W∕cm2in the air. The combination of this high
level of intensity and the extended propagation distance facili-
tates efficient nonlinear-conversion processes in the filament
way that they become compressed by the air dispersion at a prearranged
distance behind the cloud. Self-focusing collapse of these pulses results in
the generation of weakly ionized seed plasma channels. The plasma density
in these seed filaments is increased by several orders of magnitude by the
application of a longer drive pulse. The properties of the pulses are tailored
to produce population inversion in the ionized N2and O2. The air-laser
radiation at frequency ν1is combined with an interrogation pulse at ν2to
identify trace amounts of gas, such as NO2, in the cloud.
An ultrashort laser pulse or sequence of pulses is prechirped in such a
Hemmer et al.PNAS
February 22, 2011
core, which result in the bright emission of broadband supercon-
tinuum light (10). This emission is a signature phenomenon of
filamentation; it makes femtosecond laser filaments potentially
useful as light sources in LIDAR, but it cannot be used in our
SOS scheme because the supercontinuum emission by filaments
Because oxygen has the lowest ionization potential of all con-
stituents of air, the overwhelming majority of ionized species
inside laser filaments constitutes singly ionized oxygen molecules
and freed electrons. This plasma is very dilute. Although reported
experimental data on the plasma density inside filaments vary by
about 2 orders of magnitude, it is certain that less than 0.1% of all
oxygen molecules are ionized. Thus, ultrafast laser filaments
themselves cannot serve as practical gain media for efficient oxy-
gen or nitrogen lasers. The ionization fraction of oxygen is too
low, and that of nitrogen is even lower. Simple scaling of the pulse
energy is not going to help because filamentation is a self-consis-
tent process, and increasing the pulse energy typically results in
the creation of multiple filaments inside the transverse profile of
the laser beam, not in a higher plasma density in a single filament.
However, high plasma density is needed for the lasing action in
the ambient air to occur.
The solution that we propose to this problem is the use of a
dual-pulse excitation scheme. In this approach, the femtosecond
pulse with peak power above the self-focusing threshold but
limited pulse energy of the order of tens of millijoules (the igni-
ter) creates the filament, thus providing seed electrons that are
liberated from oxygen molecules by multiphoton ionization.
A second, much longer pulse with joule-level pulse energy (the
heater) accelerates the seed electrons and initiates an electron
avalanche that ionizes a substantial fraction of nitrogen and oxy-
gen in the air. A sequence of multiwavelength heater pulses may
have to be applied in order to select a particular excitation path-
way in the air molecules and thus affect the wavelength of the
backward-generated probe beam.
In order to initiate an electron avalanche, the duration of the
heater pulse needs to exceed the electron collision time in air,
∼350 fs (11). In addition, the complete ionization of air inside
a 1-m-long filament with a diameter of 100 μm will require pulse
energy of the order of 1 J. A standard Q-switched lamp-pumped
Nd∶YAG laser and its harmonics easily satisfy both of these
The igniter–heater scheme described above has been used
previously for the creation of centimeter-long dense plasma chan-
nels in laser wake-field tabletop accelerators and Raman lasers
(12, 13). We extend this approach to meter-scale plasma channels
in ambient air.
Nitrogen, the most abundant constituent of air, offers the most
straightforward choice for the application of this approach. Nitro-
gen is one of the few gases that have been shown to produce effi-
cient lasing in the UV range at atmospheric pressure. In fact,
stimulated emission in the backward direction from atmospheric
nitrogen excited by an ultrafast laser filament has been reported
(14). The femtosecond pulse was the only source of excitation
in that work, and no heater pulse was used. Consequently, the
density of the ionized nitrogen molecules was extremely low, and
the emitted radiation was too weak to be practical or even to be
properly characterized in terms of power, pulse energy, and pre-
cise spectral content. Furthermore, remote excitation of atmo-
spheric nitrogen by microwave pulses has been shown to result
in quite efficient transient lasing (15). The backward-propagating
laser pulse in that case had three spectral lines at 316, 337, and
358 nm, and the total generated pulse energy was in tens of
microjoules. Commercial nitrogen lasers that operate at about
one-tenth of atmospheric pressure and utilize electric discharge
excitation produce UV pulses at the 337-nm wavelength, in Fig. 3,
with energy in the hundreds of microjoules (6). This level of
performance can be achieved by using our dual-pulse excitation
approach. A potential complication may be related to the pre-
sence of oxygen, which provides for an alternative, preferential
pathway to ionization with a lower ionization potential than that
of the nitrogen. For that reason, oxygen is typically viewed as
being toxic in nitrogen lasers with discharge excitation. At the
same time, it is possible that smart application of coherent-
control techniques may turn this disadvantage into a benefit
and utilize nitrogen-oxygen collisions as a mechanism for empty-
ing the lower lasing level in the nitrogen laser.
The second major constituent of air is oxygen. Because the
ionization potential of oxygen is the lowest among air constitu-
ents, oxygen is the only air species that becomes appreciably
ionized through ultrafast filamentation. Under electric discharge
excitation, oxygen has been shown to lase in various spectral lines
throughout the UV, visible, and near-IR parts of the spectrum,
but in the past lasing was demonstrated only at pressures much
lower than atmospheric pressure (16, 17). Moreover, we have
now made oxygen lase under optical filament excitation at atmo-
spheric pressure (5). Technical realization of neon-oxygen and
argon-oxygen lasers offers insights into possible routes to inver-
sion (18, 19). These lasers belong to the general class of dissocia-
tive excitation transfer lasers. Specifically, inversion is produced
by the collision of the excited-state neon and argon atoms with
molecular oxygen. This process leads to the production of two
oxygen atoms—one in the ground and the other in the excited
state. Two inversion mechanisms are possible depending on
the energy of the excitation transfer from the respective noble
gas. For the case of neon in the3P2excited metastable state,
the resonant energy transfer of approximately 16.6 eV leads to
oxygen dissociation and production of oxygen atoms in the
ground and 33P states. In this case, lasing occurs in the fine struc-
ture on the 33P-33S transition at 844.6 nm as well as on the
35P-35S transition at 777.3 nm. For the case of the inelastic colli-
sions with excited-state argon, atomic oxygen in 21S and 21D can
be further excited by electron impact in an electric discharge to
yield emission from the same transitions discussed above. The
intensities of the femtosecond laser pulses in a laser filament
are on the order of 1013–1014W∕cm2. This intensity level sug-
gests a possibility of the nonlinearity that is sufficiently strong
to mimic inelastic collisional excitation through multiphoton ex-
citation by the laser pulses at the 800-nm wavelength commonly
used for filament generation. Multiphoton excitation of molecu-
lar oxygen to the states accessible by the collisional energy trans-
fer from neon and argon atoms, primarily in the Schuman–Runge
continuum above the3Σustate, can yield excited-state atomic
oxygen and create conditions for lasing.
In summary, ultrafast laser filamentation can be applied to
mimic physical processes involved in the creation of population
inversion in the constituents of ambient air by electrical dis-
charge. The possibilities of the interplay between constituents of
the medium, in this case molecular nitrogen, oxygen, and their
atomic dissociation products, as well as the electron impact in
the plasma provide numerous possibilities for pumping inversion
in any of these constituents. Whereas filament configuration is
predominantly defined by the properties of the initial ultrashort
pumping laser pulse, the excitation of the medium to states
enabling stimulated emission at the specified wavelengths can
be achieved and optimized via intelligent or adaptive pulse shap-
ing of the filament pulse itself or via the application of auxiliary
Our immediate experimental plan is as follows. We will first study
cavityless air lasers in a controlled laboratory setting. We will use
a single millijoule-level femtosecond igniter pulse to produce a
relatively short (meter-scale) plasma filament, approximately
0.1 mm in cross-section. This low carrier-density plasma will then
be heated by a joule-level nanosecond laser pulse timed with
www.pnas.org/cgi/doi/10.1073/pnas.1014401107Hemmer et al.
respect to the igniter. We will determine optimal conditions
(pulse shapes, frequencies, and relative timing) that result in
denser plasma and investigate various nitrogen and oxygen lasing
schemes discussed above. At that step we expect to demonstrate a
single efficient “open-air” bidirectional laser (lasing simulta-
neously forward and backward). The main difference of our air
laser from other easily constructed air lasers will be the plasma
preparation mechanism (using a pump laser beam instead of, for
example, an electric discharge). We expect the probability of
success at this first step to be high.
by anegativeprechirpoftheigniter pulse.Thereissufficient prior
improved control over the distance at which filamentation occurs
can be achieved by adjusting the igniter beam focusing geometry,
as was also demonstrated in the past (22).
After an efficient open-air bidirectional laser is attained, and
after we demonstrate an ability to control its position (with re-
spect to the source laser), we will move to the next step, which
will be to demonstrate production of multiple laser gain regions
at controlled distances with controlled timing, so as to achieve the
swept-gain scheme and demonstrate the preferentially backward
lasing. We will work toward this more difficult task in a laboratory
setting first. An extension to a backward swept-gain laser many
kilometers away will pose additional challenges (due to, for ex-
ample, beam distortion by air turbulence) that could be addressed
through an adaptive optical technique, such as was recently
applied in order to enhance filaments at large distances (23).
SOS via Raman Spectroscopy
We next briefly discuss remote-sensing scenarios using our coher-
signal at the detector. This back-propagating signal is in contrast
to the conventional (1, 2) (e.g., LIDAR) weak signals that are
based on incoherent scattering with no preferred direction.
We consider spectroscopic configurations that use the remote
air laser to interrogate the suspect “cloud” and detect trace
scopy will clearly be of little use, because the air laser will typically
operate at a “single” wavelength, producing a relatively narrow
spectral band without the possibility for wavelength tuning. An
obvious route is to do nonlinear spectroscopy and supplement
the backward-generated light by additional laser pulses (multiwa-
velength, tunable, or broadband) sent in the forward direction.
The end goal is to induce a coherent, directional, signal beam
that carries the molecular fingerprint straight back toward the
Molecular excitations produced by pairs of photons—one from
the backward air-laser beam and the other from the forward
interrogation laser—provide a sensitive molecular sensing me-
chanism. Examples include two-photon absorption (TPA) and
stimulated Raman scattering (SRS). Ordinary TPA and SRS
(Fig. 4 A and B) can be directly used for remote sensing utilizing
the backward laser, because these mechanisms do not require
phase matching (or, alternatively, one may say that they are
always automatically phase-matched). Then the concentration
of trace gasses can be quantified by measuring absorption of
the backward-propagating beam as a function of wavelength of
the interrogation laser. For gas molecules, both TPA and SRS will
show species-specific (fingerprint-like) structure of vibrational
and rotational energy levels. For gas-molecule fingerprinting,
each (TPA and SRS) will have its pros and cons. SRS will benefit
from smaller collisional dephasing of the ground-electronic-state
molecular oscillations, whereas TPA will have the advantage of
a nearly Doppler-free configuration (when nearly the same wave-
lengths are used for the two counterpropagating laser beams).
For remote spectroscopy of small particles and aerosols, SRS will
be preferred because TPA will typically be relatively featureless
because of large electronic broadening in the condensed phase.
Ordinarily, TPA and SRS will suffer from shot-to-shot fluctua-
tions of the air laser, which is because of the fact that measure-
ments will have to be made by comparing signals obtained
from different laser shots at different interrogation laser wave-
lengths. In principle, this difficulty can be circumvented if the
air laser can be made to operate at multiple wavelengths. We also
note that traditional coherent anti-Stokes Raman scattering
(CARS) cannot be used in scenarios involving both forward
and backward laser beams because of phase-matching constraints
forward direction (and the air laser is used to provide a probe
beam), k-vector mismatch for “backward” CARS will be reduced
but will still remain substantial and will prevent backward anti-
Finally, we note that higher-order nonlinear processes can be
considered as alternatives to TPA and SRS. Those higher-order
processes will require larger laser intensities and will generally
produce weaker signals, while possibly allowing higher molecular
selectivity by addressing several molecular levels at once.
Summary and Conclusion
The present paper focuses on generating laser action in O2and
N2in air, as a backward-propagating spectral probe. The basic
idea involves production of plasma filaments by chirped high-
power laser pulses that are further heated by longer high-energy
pulses in order to produce population inversion. By varying the
chirp, individual gain regions can be sequentially prepared in a
with counterpropagating beams at frequencies υ1and υ2as shown in Fig. 2.
Whereas TPA and SRS are automatically phase-matched, CARS is not.
Energy level and k-vector diagrams for TPA (A), SRS (B), and CARS (C)
tion R, that are most relevant for the nitrogen laser.
Sketch of the energies of N2, as a function of the internuclear separa-
Hemmer et al.PNAS
February 22, 2011
gain-swept superradiant mode. Such gain-swept superradiance is
discussed at length in ref. 3.
Weak backward lasing from excited nitrogen molecules inside
femtosecond laser filaments has been observed by Chin and co-
workers at the University of Laval in Canada (14). Recent
theoretical work also supports the concept of N2lasing through
filamentation. Furthermore, recent experiments by Miles and
coworkers at Princeton show that remotely initiated lasing in
atmospheric O2is feasible (5).
ACKNOWLEDGMENTS. We gratefully acknowledge support for this work by
National Science Foundation Grant EEC-0540832 (MIRTHE ERC), the Office
of Naval Research, and the Robert A. Welch Foundation (A-1261 and
A-1547). P.P. acknowledges support from US Air Force Office of Scientific
Research Grants FA9550-10-1-0237 and FA9550-10-1-0561.
1. Steinbrecht V, Rothe KW, Walther H (1989) Lidar setup for daytime and nighttime
probing of stratospheric ozone and measurements in polar and equatorial regions.
Appl Opt 28:3616–3624.
2. Bisson SE, Goldsmith JEM, Mitchell MG (1999) Narrow-band, narrow-field-of-view
Raman Lidar with combined day and night capability for tropospheric water-vapor
profile measurements. Appl Opt 38:1841–1849.
3. Kocharovsky V, et al. (2005) Gain-swept superradiance applied to the stand-off detec-
tion of trace impurities in the atmosphere. Proc Natl Acad Sci USA 102:7806–7811.
4. Bonifacio R, Hopf F, Merystre P, Scully M (1975) Steady-state pulses and superradiance
in short-wavelength, swept-gain amplifiers. Phys Rev A 12:2568–2573.
5. Dogariu A, Michael JB, Scully MO, Miles RB (2011) High-gain backward lasing in air.
6. See, for example, Stanford Research Systems, Inc., NL100: 337 nm nitrogen laser
7. Couairon A, Mysyrowicz A (2007) Femtosecond filamentation in transparent media.
Phys Rep 441:47–189.
8. Sprangle P, Peñano JR, Hafizi B (2002) Phys Rev E 66:046418.
9. Bergé L, Skupin S, Nuter R, Kasparian J, Wolf J-P (2007) Ultrashort filaments of light in
weakly ionized, optically transparent media. Rep Prog Phys 70:1633–1713.
10. Nibbering ETJ, et al. (1996) Conical emission from self-guided femtosecond pulses in
air. Opt Lett 21:62–64.
11. YablonovitchE, BloembergenN (1972) Avalancheionization and the limiting diameter
of filaments induced by light pulses in transparent media. Phys Rev Lett 29:907–910.
12. Gaul E, et al. (2000) Production and characterization of a fully ionized He plasma
channel. Appl Phys Lett 77:4112–4115.
13. Pai C-H, et al. (2008) Backward Raman amplification in a plasma waveguide. Phys Rev
14. Luo Q, Liu W, Chin SL (2003) Lasing action in air induced by ultrafast laser filamenta-
tion. Appl Phys B 76:337–340.
15. Vaulin VA, Sinko VN, Sulakshin SS (1988) Air ultraviolet laser excited by high-power
microwave pulses. Sov J Quantum Electron 18:1457–1458.
16. Bridges W, Chester A (1965) Visible and UV laser oscillation at 118 wavelengths in
ionized neon, argon, krypton, xenon, oxygen, and other gases. Appl Opt 4:573–580.
17. Tunitskii LN, Cherkasov EM (1968) Investigation of a pure-oxygen laser. J Appl Spec-
18. Bennett WR, Jr (1965) Inversion mechanisms in gas lasers. Appl Opt 4(Suppl 1):3–21.
19. Bennett WR, Jr, Faust WL, McFarlane RA, Patel CKN (1962) Dissociative excitation and
optical maser oscillation in Ne-O2and Ar-O2rf discharges. Phys Rev Lett 8:470–473.
20. Méchain G, et al. (2005) Range of plasma filaments created in air by a multi-terawatt
femtosecond laser. Opt Commun 247:171–180.
21. Shverdin MY, Goda SN, Yin GY, Harris SE (2006) Coherent control of laser-induced
breakdown. Opt Lett 31:1331–1333.
22. Liu W, et al. (2006) An efficient control of ultrashort laser filament location in air for
the purpose of remote sensing. Appl Phys B 85:55–58.
23. Daigle J-F, et al. (2009) Remote sensing with intense filaments enhanced by adaptive
optics. Appl Phys B 97:701–713.
www.pnas.org/cgi/doi/10.1073/pnas.1014401107Hemmer et al.