www.osa-opn.org
1047-6938/10/01/0020/8-$15.00 ©OSA
20 | OPN Optics & Photonics News
Colin Webb
History of
Gas
Lasers
Part 2: Pulsed
Gas Lasers
In this second article of a two-part series, Colin Webb explores the origins
of pulsed gas lasers, which made possible many critical applications of
laser technology, including corneal reshaping and microlithography.
Ask anyone to name the most important application
of lasers, and they will probably point to their clinical
use—for example, lasers that reshape the cornea
of the eye to correct vision defects, as explored in this month’s
cover article on “Lasers in Ophthalmology” (p. 28). The laser that
has made corneal reshaping possible is the argon fluoride (ArF)
excimer laser, which operates at 193 nm. It is a rare gas halide
(RGH) excimer laser, a type that burst on the scene in 1975.
Because of its very short wavelength and its ability to
propagate in air, the ArF laser has also become the key light
source for microlithography—which allows scientists and
engineers to produce features of size less than 100 nm on
semiconductor chips. This is the largest application for this
type of laser, and a driving force in electronic technology.
Without it—or its KrF analogue—much of today’s computer
technology would not be possible.
But the excimer laser was not created in a vacuum. There is
a considerable history of pulsed gas lasers that precedes it. Here,
following Jeff Hecht’s January feature on continuous wave gas
lasers, I explore how pulsed gas laser systems came to be.
The nitrogen laser—a DIY project
Among the very earliest of the purely pulsed gas lasers is the
molecular nitrogen system, which operates in the near ultraviolet
on the C—>B band of the N2 molecule at wavelengths
around 337 nm. H.G. Heard at Energy Systems discovered
this laser in 1963. Its performance in terms of pulse energy
(typically a few mJ), and pulse repetition frequency (typically
2-20 per second), are by no means impressive by today’s
standards. Nevertheless, it is of great significance in the
history of lasers, not only because its output was (and still is)
used to pump tunable dye lasers covering the near-ultraviolet
and visible regions, but because it was so simple to construct
and operate.
In fact, many thousands of these lasers have been built as
science projects in colleges and domestic garages all around
the world. Scientific American’s Amateur Scientist column
published do-it-yourself instructions in 1974. This has been the
inspiration for myriad designs. In 1973, Jim Piper and I built
our own nitrogen laser to be used for atomic physics experiments
at the Clarendon Laboratory at Oxford.
©UKAEA
Gas
Lasers
February 2010 | 21
Copper vapor lasers
pump dye lasers for AVLIS
experiments at UKAEA
Harwell, U.K.
www.osa-opn.org22 | OPN Optics & Photonics News
level lifetime—which, in the case of the H2 laser operating on
transitions of the VUV Lyman bands of molecular hydrogen
around 160 nm, implies rise-time requirements that are even
faster than those of the N2 laser.
In 1926, James Franck of the University of Göttingen,
Germany, suggested that, because the masses of the nuclei of
a diatomic molecule are so much larger than the mass of an
electron, the instantaneous positions of the vibrating nuclei
cannot be significantly affected when an electron changes from
one excited state to another by absorbing or emitting a photon.
Thus, electronic transitions can be represented as vertical lines
on a diagram of potential energy vs. internuclear separation.
Within a few weeks, Franck’s principle was put on a sound
quantum mechanical basis by Edward Condon of the University
of California, Berkeley. He showed that, in addition, the
most probable transitions occur at internuclear separations for
which the overlap between the vibrational wavefunctions of
the upper and lower level is maximized. This is known as the
Franck-Condon principle.
When a fast moving electron impacts a molecule, the probability
of excitation mirrors the optical transition probability.
Thus, to a first approximation, electron impact excitation
conforms to the Frank-Condon principle. As suggested by
Bazhulin and colleagues at the Lebedev Institute in Moscow,
a scheme based on the Frank-Condon loop can be used to create
transient population inversion in molecular gases. To illustrate
this scheme in the case of H2, high-energy electrons in
the discharge preferentially excite H2 molecules in the ground
vibrational level (y0 = 0) of the ground electronic level X1S+
g
to vibrational levels (y9 = 4–8) of the electronic level B1S+
u.
The selective excitation produces short-lived inversions with
respect to excited vibrational levels (y0 = 11-14) of the ground
electronic level, which are themselves far too high in energy to
be populated at thermal energies. On much longer timescales,
the population in these states eventually returns to the ground
state by thermal collisions before the start of
the next discharge, thus completing a “loop”
in the energy diagram.
In 1970, R.T. Hodgson at the IBM Watson
Research Center first demonstrated VUV
laser action in H2. He observed oscillation
on several transitions in the Lyman bands
around 160 nm. This was followed in 1971
by Waynant’s observation of oscillation on two
transitions of the Werner bands at 116 and
123 nm using a traveling wave Blumlein
excitation circuitry developed by Waynant,
Shipman and colleagues at NRL Washington.
The molecular hydrogen lasers provided
early evidence that gas lasers would be possible
in short wavelength regions of the
spectrum where most solid-state materials
have limited optical transmission. Despite
that, these systems never seem to have been
The breakdown voltage of molecular gases at the pressures
needed for laser operation tends to be much higher than in
low-pressure atomic gases. For this reason, the discharge in the
N2 laser runs transversely across the width of the active volume
rather than along its length. A capacitor bank charged to many
tens of kilovolts and switched by a triggered spark gap (or, in
more sophisticated devices, a thyratron) discharges its energy
into the nitrogen at pressures from a few Torr to one atmosphere
in a current pulse—which rises to several kiloamperes
within 50 ns or less.
The output of the laser comprises a
pulse that lasts some 10 ns for the lowest
N2 pressures or less than 1 ns in the case
of atmospheric pressure operation. Whatever
the pressure, however, the vibrational
states of the lower laser level have no way
to get rid of the population in timescales
of interest, and the laser pulse is therefore
self-terminating; longer excitation pulses
do not result in longer laser pulses.
Self-terminating VUV lasers—
The Franck-Condon Loop
This self-terminating feature is shared
by many important gas laser systems,
where the lower laser level is long-lived.
It limits pulse length to about the upper-
Nitrogen
laser at the
Clarendon
Laboratory,
Oxford.
As suggested
by Bazhulin and
colleagues at the
Lebedev Institute
in Moscow, a
scheme based on
the Frank-Condon
loop can be used
to create transient
population inversion
in molecular gases.
Piper and Webb circa 1973
February 2010 | 23
exploited commercially. This may be because their output
pulse energy and repetition rate are very limited, or it could be
because there has been—up to now—little demand for such
short wavelength lasers. Whatever the reason, these systems
remain today as scientific curiosities.
Pulsed CO2 lasers—TEA anyone?
The pulsed CO2 laser system is certainly more than a curiosity.
It provides high-power pulses at many discrete frequencies in
the mid-infrared bands around 10.4 and 9.4 μm. As Jeff Hecht
described in his January feature, at low gas pressures, the CO2
laser is basically a continuous wave system that depends for its
excitation on resonant energy transfer in collisions of N2 molecules
in the first vibrationally excited level (y = 1) with ground
state CO2 molecules:
N2(y =1) + CO2 (ground state) —>
N2(y =0) + CO2 (upper laser level) + ΔE (small).
(1)
The N2 molecules are excited by collisions of electrons in the
discharge run in a gas mixture containing He, N2 and CO2.
In 1969, R. Dumanchin and J. Rocca-Serra in France
and A.J. Beaulieu in Canada recognized that, by increasing
the gas pressure of the discharge from 5-20 Torr (the range
characteristic of the CW systems) to pressures on the order of
one atmosphere (760 Torr), the population inversion density
could be increased enormously. This is because the number of
target N2 molecules that electrons could hit, and the number
of target CO2 molecules that excited N2 molecules could hit,
would increase correspondingly. However, there is a problem
in running longitudinal discharges at such high pressures.
The voltages required to cause breakdown and to sustain a
discharge are unacceptably high, and so Beaulieu turned to
the solution of running the discharge transversely across the
width of the discharge volume rather than along its length.
Thus, the TEA (transversely excited atmospheric) CO2 laser
was born. In order to stop the discharge from collapsing into
a few isolated arcs along the length of the electrodes, Beaulieu
used distributed ballasting in the form of a closely spaced array
of 1 kΩ resistors. Each one was connected at one end to a
common rail; the free ends constituted the segmented cathode
on one side of the discharge gap, and the anode side comprised
a solid rod of conductor. If, in the early stages of the collapse
from a spatially extended glow to a constricted arc, the current
through one resistor should show an increase above the ambient
value, the increased voltage drop across that resistor would
stifle the current growth and quench the incipient arc. This is
how the ballasting action comes about.
Although Beaulieu’s device established the principle of TEA
laser operation, the discharge volume was composed of a thin
array of spatially separate discharges, which did not easily lend
itself to scaling the discharge volume. The next step demanded
the creation of a discharge volume of approximately square
cross-section; this was done by using uniform field electrodes
that were carefully shaped to Rogowski or Bruce profiles.
Several groups, including H.M. Lamberton and P.R. Pearson
working at the U.K.’s Services Electronics Research Laboratory
in Baldock, Hertfordshire, experimented with such designs
and found that pre-ionizing the discharge in the He:N2:CO2
mixture would allow uniform excitation of a large volume for a
few microseconds.
These early TEA lasers gave output pulse energies on the
order of 5J from a discharge volume of 2.5-cm width, 3.2-cm
height and 100-cm length. Pre-ionization of the main discharge
volume was provided by the photo-ionizing effect of
deep UV emitted by a corona discharge around wires situated
The energy level scheme of Hodgson’s VUV hydrogen laser.
[ Lamberton and Pearson’s TEA CO2 laser ]
[ The Franck-Condon Loop ]
Fast discharge capacitor
1 2 3
10
5
0
5
5
v0=0
H2
X9S+
g
C9u
10
Stimulatedemission
Electronexcitation
E(eV)
CO2 TEA laser discharge unit
HV DC
power
supply
Triggered
spark gap
Rogowski
electrode Preionising
discharge
wire
Reprinted with permission from R.T. Hodgson. Phys. Rev. Lett. 25(8), 494-7 (1970).
©1970 American Physical Society
From Wikimedia Commons, TEA Laser
R (Å)
B9S+
u
v9=0
5
u
B9S+
u
v9=0
5
v0=0
10
www.osa-opn.org24 | OPN Optics & Photonics News
By 1992, Livermore
had established a
copper laser MOPA
(master oscillatorpower
amplifier) array
that provided an average
power of 7 kW
to a dye laser MOPA
array with 2.5-kW
average power for
24 hours per day,
7 days per week.
to one side of the discharge volume or
behind a mesh electrode. Devices based
on this technology have found widespread
applications, including the datemarking
systems for consumer products
that have been used since the 1970s.
Throughout the 1960s, 1970s and
1980s, defense contractors worked on
scaling CO2 lasers to still-higher pulse
energies. In 1969, W.L. Nighan and J.H.
Bennett at Avco made calculations that
showed that, in order to excite N2 molecules
to the first few vibrational levels
some 0.3 to 1.2 eV above their ground
state in a gas discharge, the optimum
value of E/N—the electric field to gas
number density ratio—must be far
smaller than that needed for a discharge
in the same CO2:N2:He gas mixture in
order to be self-sustaining.
However, by using an external electron beam source to
ionize the gas mixture, researchers could run a discharge at
an E/N ratio chosen to be optimum for exciting the vibrational
levels of N2, which then go on to excite the CO2 by
the reaction of equation (1). Typical of this approach is the
device reported in 1976 by J.D. Daugherty of the Avco Everett
Research Laboratory; it employed a cold cathode electron gun
to provide a beam of 300 kV electrons 10 cm high 3 25 cm
wide. Square pulses of discharge current lasting 80 μs could
be sustained by the e-beam, leading to output pulses of some
55 μs duration, with energies of several kJ. By far the largest
part of the input energy is supplied by the discharge.
Iodine—a cure for all ills?
The photodissociation iodine laser discovered
by J.V.V. Kasper and G.G. Pimentel
at the University of California, Berkeley,
in 1964 represents a very different type of
gas laser. It operates on the magnetic dipole
transition (1.315 μm) between the fine
structure levels 2P½—>2P3/2 of the ground
state of atomic iodine. The upper laser level
is populated selectively when iodine-bearing
compounds such as CF3I or CH3I are dissociated
by intense radiation from a flashtube.
In this form, iodine lasers are selfterminating
pulsed devices, since there is
no way for the population to exit from the
2P3/2 lower level (the ground state) in relevant
time scales. Lasers of this type were developed
progressively via a series of devices
known as Asterix for laser-target studies at
the Max Planck Institute in Garching, Germany.
The technology was later transferred to Prague in the
Czech Republic, where today the Prague Asterix Laser System
can deliver pulses of up to 1 kJ at 1.315 μm.
In recent years, the iodine laser system has morphed into a
purely chemically excited device (the chemical oxygen–iodine
laser, or COIL), which does not require any electrical input to
produce long pulses of high power radiation; rather, it uses collisions
of excited oxygen molecules in the metastable 1Δ state
to excite the ground-state iodine atoms from the lower 2P3/2
lower laser level to the 2P½ upper laser level. The O2(1Δ) molecules
are produced as the result of chemical reactions of Cl2
molecules with basic hydrogen peroxide solution. In this form,
the iodine laser is effectively a continuous wave device capable
of very long pulses. Such a device is currently in development
as Northrop Grumman’s Airborne Laser.
Copper—an atomic laser
In 1966, W.T. Walter et al. at TRG Inc. in Melville, N.Y.,
U.S.A., reported oscillation on the 2P3/2 —> 2D5/2 and 2P½ —>
2D3/2 transitions of the neutral copper atom at 511 and
578 nm, respectively. Metallic copper that was vaporized
at high temperatures and excited in a longitudinal pulsed
discharge formed the basis of their copper vapor laser. A.A.
Petrash and colleagues at the Lebedev Institute in Moscow
showed that, by repetitively pulsing the discharge at a high
repetition rate, the device becomes self-heating, obviating the
need for an external oven. I. Smilanski at the Nuclear Research
Centre Negev in Israel showed that it was possible to scale the
output power with the tube diameter.
These lasers definitely belong to the self-terminating class of
pulsed gas lasers, since the 2D5/2 and 2D3/2 lower laser levels are
fully metastable, and collisional deactivation on the time scale
of the laser pulse is not possible. However, they are relatively
Courtesy of Avco
Electron gun for large
e-beam–sustained
CO2 laser.
February 2010 | 25
efficient since the electron collision
cross-section for exciting the 2P3/2
and 2P½ upper laser levels from the
2S½ ground state dominates electron
inelastic loss processes so that the
copper system is the laser analogue
of the sodium lamp. Typical commercial
copper lasers (Oxford Lasers
Ltd.) offer average powers of up to
45 W at 511 and 578 nm combined
with pulse lengths of 25 ns, and pulse
repetition frequencies adjustable in
the 10-30 kHz range.
Throughout the 1970s and 1980s,
many laboratories were developing
copper laser systems for dye laser
pumping for atomic vapor laser
isotope separation applications. The
Lawrence Livermore Laboratory in
the United States, for example, was
heavily involved in this area. By 1992,
Livermore had established a copper
laser MOPA (master oscillator-power amplifier) array that
provided an average power of 7 kW to a dye laser MOPA array
with 2.5-kW average power for 24 hours per day, 7 days per
week. However, with the collapse of the former Soviet Union
and the consequent availability of diluted weapons-grade uranium
for reactor fuel on the world market, the momentum to
develop new methods of enrichment itself evaporated.
The pure rare gas excimers—promise unbounded
The word “excimer” is a contraction of the words “excited
dimer.” Technically, rare gas halide molecules should not
belong to this class because the word “dimer” applies to a
molecule comprising two nuclei of the same atomic species.
Still, the word excimer has come to apply to any type of laser
involving a transition from the bound excited state of a molecule
to a lower level that is unbound.
The idea that transitions from the bound excited state of an
excimer to the unbound ground state would make good candidates
for laser molecules dates back to 1960. F.B. Houtermans
proposed that such transitions from excited states of mercury
dimers to the ground states would easily produce inversions,
since the ground state is unstable and easily breaks apart into
two separate Hg atoms in timescales of a few picoseconds. The
ground state is therefore virtually unpopulated, so lower-level
bottlenecking could not limit the pulse energy or duration.
Although Hg2 does not appear to have produced a viable
high-power laser system to date, it stimulated the idea that
other excimers—notably those of the rare gas molecules (Xe2,
Kr2 and Ar2)—could produce lasers. Lasers based on rare gas
dimers were reported in 1970, when N.G. Basov et. al. demonstrated
gain at 175 nm in electron-beam-pumped liquid xenon.
And the next several years brought the first operation of a pure
rare gas excimer laser with xenon (1972), krypton (1973) and
argon (1974).
From that time on, many large labs around the world began
scaling the output of Xe2, Kr2 and Ar2 lasers to high-pulse
energies. However, as C.W. Werner and E.V. George pointed
out in the 1976 book Principles of Laser Plasmas (G. Bekefi, ed.),
a barrier appeared to be set by the fact that the laser photon in
these systems has enough energy to cause the photo-ionization
of states of the excimer populated by collisional transfer from
the upper laser level. According to this explanation, photoionization
would provide an increasing source of internal loss
as the intensity of pumping increased. In contrast, the rare gas
halide’s laser photon energy is not enough to photo-ionize the
upper laser level (or states nearby); thus, one major source of
internal loss is absent.
The rare gas halide excimers—promise fulfilled
The first member of the class of rare gas halide (RGH) lasers
to be successfully demonstrated was the xenon bromide (XeBr)
laser. In May 1975, S.K. Searles and G.A. Hart of the Naval
Research Laboratory (NRL) in Washington D.C. first reported
operating this laser at 282 mm. The idea that RGH molecules
would make good candidates for powerful short-wavelength
lasers was put forward by D.W. Setser and J.E. Velazco of
Kansas State University. The previous year, Setser had been
working in the chemistry laboratory at Cambridge University
in the United Kingdom, in the group headed by B.A. Thrush.
That team had observed UV emission from the bands identified
as belonging to the argon chloride molecule excited when
Cl2 was titrated into a flowing afterglow in argon.
Sandia Labs /Courtesy of Jeff Hecht
Gary Tisone (left) and A.K.
Hays (right) with the first ArF
Laser. This e-beam–pumped
system produced pulses of
92 J at 193 nm.
www.osa-opn.org26 | OPN Optics & Photonics News
The reason why Searles could so quickly test the suggestion
that XeBr would lase in the UV is that he was already working
on pure rare gas excimer (Xe2) laser research at NRL, and he
had an electron-beam-excited device available for this work.
Thus, he only needed to change the mirrors and introduce a
small amount of bromine vapor into the laser chamber along
with the xenon to see if XeBr would lase.
At an internal seminar held at NRL in March 1975, Searles’
work on XeBr was communicated to representatives of Avco,
Northrop and Livermore and other U.S. laboratories working
on excimer laser research. J.J. Ewing and C.A. Brau at Avco
Everett lost no time in acting on the information, and, at the
June 1975 Conference on Lasers and Electro-Optics, they
announced lasing in KrF (249 nm), XeF (354 nm) and XeCl
(308 nm). Then, in 1976, Hoffman, Hays and Tisone at Sandia
Labs announced laser action on ArF (193 nm), also in an
e-beam–pumped cell.
These advances stimulated intense research. An important
strand of this work was aimed at emulating the combination of
electron-beam and discharge pumping, which had so successfully
been applied to the CO2 laser system. In the case of the
RGH lasers, the same approach did not yield such rewards in
the fraction of power that could be usefully applied as discharge
excitation. Not much more than half the energy input
to an electron-beam-sustained RGH laser could be applied as
discharge pumping.
The gas mixture of an e-beam-pumped RGH laser such as
KrF is made up of three components:
(1) the rare gas Kr at some tens of Torr pressure, (2) a few
Torr of the halogen donor (F2), and (3) two or three atmospheres
of argon or neon. The argon or neon is needed to scatter
the incoming electron beam so that it deposits its energy
in the gas cell instead of sailing straight on through. A second
important function of this background gas is the provision of
an abundant supply of “third bodies” for the three body reactions
involved in the formation kinetics of the excimer KrF*
via the ion recombination reaction of positively charged Kr+
ions and negatively charged fluorine ions F– :
Kr+ + F– + Ne —> KrF* + Ne . (2)
The main reason why the RGH lasers and CO2 differ so
much in their division of labor between sustaining the ionization
rate and exciting the upper laser level precursor species is
that the energy required to create the Kr+ ions needed for reaction
(2) is the same or very close to that required for creating
electron-ion pairs to sustain the discharge.
Recent research efforts to develop high-pulse-energy excimers
for potential applications as drivers for thermonuclear
fusion reactors have been focused on purely e-beam-excited
systems. The ELECTRA KrF laser at NRL currently produces
pulses of 250 to 700 J at repetition rates of 1-5 pulses
per second. One advantage that gas laser systems have over
their solid-state counterparts is their ability to shed waste heat
Courtesy of John Sethian and F. Hegeler
ELECTRA KrF
laser at the Naval
Research Laboratories
in Washington
D.C., U.S.A.
[ Formation of excited KrF molecules ]
Energy
Separation between Kr and F nuclei
Laser transition
249 nm
(KrF)*
Kr F
KrF+ + F–
February 2010 | 27
rapidly from the active medium by circulating the gas mixture
in a closed loop containing heat exchangers.
Throughout the 1980s and 1990s, researchers looking to
develop commercial RGH excimer lasers focused on purely discharge-excited
systems without the e-beam generators—which
tended to restrict their repetition rate capability and increase
their complexity and cost. Lumonics and Lambda Physik were
prominent developers of systems for scientific applications;
sometimes they used part of the excimer laser’s output to pump
dye lasers for pump-probe spectroscopy. Another company,
Cymer, supplied the market for UV lasers for photolithography,
where the short wavelength of the ArF laser also gives it
an advantage. Companies that supplied the ophthalmologic
markets, including Summit and Visex, set out to supply the
market for ArF systems for corneal-reshaping applications.
Another system that can be operated in the same type of
discharge-excited device as ArF or KrF is the molecular fluorine
laser. It provides pulses of a few nanoseconds’ duration at
the even deeper ultraviolet wavelength of 157 nm. Since radiation
at this wavelength is strongly absorbed in air, the transport
of an F2 laser beam necessarily involves the use of vacuum
(or helium-filled) light paths, which may make it less attractive
for many potential applications.
Discharge-excited excimer lasers typically produce pulses in
the range 0.5-2.5 J at repetition rates of 0.1-1 kHz. R.S. Taylor
and K.E. Leopold at NRC Ottawa have reported pulse lengths
as long as 1.5 μs at 308 nm for specially optimized XeCl lasers,
but, for KrF discharge lasers, the longest pulses are around
800 ns, with typical values of 10 to 50 ns.
This raises at once an interesting question. The lower laser
level of an excimer laser transition is unstable against dissociation
into neutral atoms (Kr and F in the case of KrF)—so why
are the pulses so short? These lasers cannot be self-terminating
like the N2 and H2. However, like the CO2 system, they have
an energy-level scheme that one might expect would lend itself
to CW operation. A bright young Scottish graduate student,
Julian Coutts, who was working in my group at Oxford in
1986, provided insight into this. He noted, for example, that,
in the KrF laser, the loss rate of electrons is controlled by the
dissociative attachment reaction:
	 F2 + e– —> F– + F . (3)
If now the halogen donor (F2) density in some small region
should decrease from its ambient value, the electron density
will grow more rapidly in that region. Hence, the concentration
of a neutral halogen donor would be further diminished
locally. This is clearly a positive feedback mechanism, which,
if unchecked, will lead to the formation of a column high in
electron density, ultimately forming an arc that disrupts the
uniformity of the gain region. Such an arc or array of arcs can
cause the volume of uniform gain medium to decrease rapidly
and, in extreme conditions, can result in the voltage across the
transverse electrodes collapsing.
Julian Coutts’ theory predicts the length of time after which
an initially uniform laser discharge becomes unstable due to
halogen donor depletion. The theory has been tested successfully
by many groups around the world. Coutts died last year aged
only 45, after losing his battle with cancer. Perhaps this article
can serve as a small memorial to him. t
Colin Webb (colinwebb9@sky.com) is a professor emeritus at the
University of Oxford, United Kingdom, and chairman of Oxford Lasers.Member
Andrew Kearsley
in February
1978 with the first
commercial-dis-
charge-pumped
KrF laser in the
United Kingdom.
Courtesy of Colin Webb
>>	F.G. Houtermans. Helv. Phys. Acta 33, 933 (1960).
>>	H.G. Heard. Nature 200, 667 (1963).
>>	J.V.V. Kasper and G.C. Pimentel. Appl. Phys. Lett. 5, 231-3 (1964).
>>	W.T. Walter et al. Bull. Am. Phys. Soc. 11, 113 (1966).
>>	A.J. Beaulieu. App. Phys. Lett. 16, 504-5 (1969).
>>	W.L. Nighan and J.H. Bennett. Appl. Phys. Lett. 14, 240-3 (1969).
>>	R.T. Hodgson. Phys. Rev. Lett. 25, 494-7 (1970).
>>	H.M. Lamberton and P.R. Pearson. Electron. Lett. 7, 141-2 (1971).
>>	R.W. Waynant. Phys. Rev. Lett. 28, 533-5 (1971).
>>	A.A. Isaev et al. JETP Lett. 16, 27-9 (1972).
>>	The Amateur Scientist, Scientific American, June 1974.
>>	M.F. Golde and B.A. Thrush. Chem. Phys. Lett. 29, 486 (1974).
>>	W.M. Hughes et al. Appl. Phys. Lett. 24, 488 (1974).
>>	J.J. Ewing and C.A. Brau. Appl. Phys. Lett. 27, 350-1 and 435-6 (1975).
>>	S.K. Searles and G.A. Hart. Appl. Phys. Lett. 27, 243-5 (1975).
>>	J.E. Velazco and D.W. Setser. J. Chem. Phys. 62, 1990 (1975).
>>	Ed G.Bekefi. Principles of Laser Plasmas, Wiley (1976).
>>	J.M. Hoffman et al. Appl. Phys. Lett. 28, 538-9 (1976).
>>	I. Smilanski et al. Opt. Comm. 25, 79-82 (1978).
>>	K.J. Witte et al. High Power Lasers and Applications, K.-L. Kompa and
H. Walther, eds., Springer Series in Optical Sciences, Springer-Verlag,
Berlin (1978).
>>	R.S. Taylor and K.E. Leopold. Appl. Phys. Lett. 47, 81-3 (1985).
>>	J. Coutts and C.E. Webb. J. Appl. Phys. 59, 704 (1986).
>>	I.L. Bass et al. Appl. Opt. 31, 6993-7006 (1992).
>>	J. Jones and C.E. Webb, eds. The Handbook of Laser Technology and
Applications, Institute of Physics Publishing, Bristol (2004).
>>	D. Mathew et al. J. Appl. Phys. 102, 033305 (2007).
>>	F. Hegeler et al. Journal of Physics Conference Series, 112, 032035,
IOP Publishing (2008).
[ References and Resources ]