WL-TR-92-2087

PLASMA PHYSICS ISSUES IN GAS

DISCHARGE LASER DEVELOPMENT

AD-A257 735

ALAN GARSCADDENMARK J. KUSNERJ. GARY EDENWL/POOC-3WRIGHT-PATTERSON ARB, OH 45433-6563

DEC 1991

FINAL REPORT FOR 10/01/84 - 12/30/91

APPROVED FOR PUBLIC RELEASE; DISTRIBUTION IS UNLIMITED.

DTICSELECTE

3EGQ 11992

AERO PROPULSION AND POWER DIRECTORATEWRIGHT LABORATORYAIR FORCE SYSTEMS COMMAND 92-30469WRIGHT PATTERSON AFB OH 45433-6563 ItII(oIIIIIIIl11111!liiiiiI !

NOTICE

WHEN GOVERNMENT DRAWINGS, SPECIFICATIONS, OR OTHER DATA ARE USED FOR ANYPURPOSE OTHER THAN IN CONNECTION WITH A DEFINITELY GOVERNMENT-RELATEDPROCUREMENT, THE UNITED STATES GOVERNMENT INCURS NO RESPONSIBILITY OR ANYOBLIGATION WHATSOEVER. THE FACT THAT THE GOVERNMENT MAY HAVE FORMULATED OR INANY WAY SUPPUED THE SAID DRAWINGS, SPECIFICATIONS, OR OTHER DATA, IS NOT TOBE REGARDED BY IMPLICATION, OR OTHERWISE IN ANY MANNER CONSTRUED, AS LICENSINGTHE HOLDER, OR ANY OTHER PERSON OR CORPORATION; OR AS CONVEYING ANY RIGHTS ORPERMISSION TO MANUFACTURE, USE, OR SELL ANY PATENTED INVENTION THAT MAY IN ANYWAY BE RELATED THERETO.

THIS REPORT HAS BEEN REVIEWED BY THE OFFICE OF PUBLIC AFFAIRS (ASDIPA)AND IS RELEASABLE TO THE NATIONAL TECHNICAL INFORMATION SERVICE (NTIS). ATNTIS IT WILL BE AVAILABLE TO THE GENERAL PUBLIC INCLUDING FOREIGN NATIONS.

THIS TECHNICAL REPORT HAS BEEN REVIEWED AND IS APPROVED FOR PUBLICATION.

/

ALAN GARSCADDEN LOWELL D. MASSIEPower Components Branch Chief, Power Components BranchAerospace Power Division Aerospace Power Division C]Aero Propulsion Laboratory . _ _

FOR THE COMMANDER j _

IF YOUR ADDRESS HAS CHANGED, IF YOU WISH TO BE REMOVED FROM OUR MAIUNGLIST, OR IF THE ADDRESSEE IS NO LONGER EMPLOYED BY YOUR ORGANIZATION PLEASENOTIFY WL/POc ,, WRIGHT-PATTERSON AFB, OH 45433-.69 TO HELP MAINTAINA CURRENT MAILING LIST.

COPIES OF THIS REPORT SHOULD NOT BE RETURNED UNLESS RETURN IS REQUIRED BYSECURITY CONSIDERATIONS, CONTRACTUAL OBLIGATIONS, OR NOTICE ON A SPECIFICDOCUMENT.

7;- IEPOiT ,T E P ) ,__DEC 1991 FINAL 10/01/84--12/30/91

. PrASMA-"PHYSICS--ISSU-9S'-IN-'-S. ..-DISCHARGE LASER DEVELOPMENT

PE 61102ALAN-GARSCADDEN PR 2301

HARK J. KUSNER TA S2J. GARY EDEN WU 03

WL/POOC-3WRIGHT-PATTERSON ARB, OH 45433-6563 WL-TR-92-2087

AERO PROPULSION AND POWER DIRECTORATE. ..

WRIGHT LABORATORYAIR FORCE SYSTEMS COMMANDWRIGHT PATTERSON AFB OH 45433-6563WL/POOC, Attn: GARSCADDEN 513-2552923

"- PUBLISHEDEIN-IEEE-TRANSACTIONS ON PLASMA SCIENCE, VOL. 19,MO. 6, DECEMBER 1991

APPROVED-FbR-PUBLIC-RELEASE; DISTRIBUTION IS - --

UNLIMITED.

An account is given of the interplay between partially ionized plasma physics andthe development of gas discharge lasers. Gas discharge excitation has provided awide array of laser devices extending from the soft X-ray region to the far infrared.The scaling of gas discharge lasers in power and energy also covers many orders ofmagnitude. The particular features of three regimes are discussed: short wavelengthlasers(deep UV to soft X-ray); visible and near UV lasers; and infrared moleculargas lasers. The current status (Fall 1990) of these areas is reviewed, and an assess-ment is made of future research topics that are perceived to be important.

PARTIALLY IONIZED PLASMAS GAS DISCHARGE LASERS 22VISIBLE AND NEAR UV ,ASERS SHORT WAVELENGTH LASERSINFRARED MOLECULAR jAS LASERS

UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED UL

Plasma Physics Issues in Gas DischargeLaser Development

Alan Garscadden, Senior Member, IEEE, Mark J. Kushner, Fellow, IEEE, and J. Gary Eden, Fellow. IEEE

(Invited Review Paper)

Abstract--An account is given of the interplay between partially longer wavelengths, since the threshold energy density of theionized plasma physics and the development of gas discharge plasma for infrared lasers is much less than that for shortlasers. Gas discharge excitation has provided a wide array of laser wavelength lasers. Shorter wavelengths also require plasmadevices extending from the soft X-ray region to the far infrared. media of high uniformity to avoid beam spreading by re-The scaling of gas discharge lasers in power and energy alsocovers many orders of magnitude. The particular features of three fractive inhomogeneities. Consequently, the short wavelengthregimes are discussed: short wavelength lasers (deep UV to soft lasers tend to. scale with energy density, while the longerX-ray); visible and near UV lasers: and infrared molecular gas wavelength molecular gas lasers can scale both volumetricallylasers. The current status (Fall 1990) of these areas is reviewed, and with energy density to the extent that the enhanced plasmaand an assessment is made of future research topics that areperceived to be important. chemistry does not deplete or quench the lasing species. Also.

the transition probabilities are generally much higher at shorter1. INTRODUCTION wavelengths so that it becomes much more difficult to achieve

CW operation unless the lower laser level population is rapidly

A LTHOUGH R. C. Tolman [1] appears to have originated depleted. Therefore as one moves into the UV regime, pulsedin 1923 the concept of population inversion leading to laser technology is generally invoked to meet these constraints.

stimulated emission and amplification, his comments suggest Further complications in developing discharge pumped shortthat he considered the situation so improbable as to be of wavelength lasers are the plasma instabilities associated withlittle value. However, at about the same time, R. Ladenburg intense pumping rates.[2] found it necessary when analyzing experiments on anoma- Much of the exploratory work on identifying and developinglous dispersion to include stimulated emission terms in his new lasers occurred in the 1960's and 70's, and now the studiesanalyses of radiation from the positive columns of narrow- are usually confined to a much more limited range of speciesbore neon discharges. The early discharge investigations of and gas mixtures. The emphasis has also changed to improvingLadenburg and of V. A. Fabrikant [3] and their colleagues the reliability and lifetime of the laser.came close to achieving the necessary conditions for pop- Given the above constraints, one concludes that the elec-ulation inversion and laser action. At that time. there was trical discharge laser is superior to chemical lasers underemphasis on understanding the approach to equilibrium at those conditions where energy density scaling is required.higher discharge currents and almost no research on enhancing Compared to solid-state lasers, it wins when volumetric scalingthe nonequilibrium features. or operation at short wavelengths ( A < 200 nm) is required.

The threshold condition for a gas laser [41 is Discharge lasers also have the advantages that gases can be

I self-healing and are much more tolerant than solids of thermal,V, \1 _ 87r (21rRT 9, o overload.9u 91 .4,1gA 3 _,N• T

where .VN, y,,, N1, and gt are the populations and statistical II. SHORT-WAVELENGTH LASERS IN DISCHARGES AND

weights of the upper and lower laser levels, respectively, Aul LASER-PRODUCED PLASMAS: UV TO SOFT X-RAY

is the transition probability, R is the gas constant, T_, is the In their classic review paper in 1976, Waynant and Eltongas temperature, .11 is the molecular mass, a is the total [5 discussed the demanding requirements inherent to theloss coefficient, and L is the length of the gain medium. demonstration of new lasers in the spectral region belowOther things being equal, it is mu-h easier to obtain gain at 200 nm. Written on the eve of the discovery of the ArF.

Manuscript received April 24. 1990: revised September 18. 1991. The ArCl, and KrCI rare gas-halide lasers, their article summarizedresearch of A. G. was supported by the U.S. Air Force Office of Scientific the properties of the handful of vacuum ultraviolet (VUV)Research (AFOSR) under Project No. 2301. that of J. G. Eden through AFOSR molecular lasers existing at the time (H,, HD. D2 , CO, Xe,,Grant 11-4X)38 (H. Schlossberg), and that of M. 1. Kushner by the NationalScience Foundation under Contract Nos. ECS 88-15781 and CBT 88-4)3170. Kr,, Ar 2), but also described with remarkable accuracy the

A. Garscadden is with the Wright Laboratory. Wright-Patterson AFB. OH plasma parameters necessary for the successful operation of-45433, atomic oscillators at still shorter wavelengths (A < 100 nm).

M. 1. Kushner and 1. G. Eden are with the Devartment of Electrical andComputer Engineering. Universitv of Illinois. Urbana. IL 61801. The past 15 years have witnessed not only the development of

IEEE Log Number 9104M18. the rare gas-halide excimer and F2 lasers, but since the 1985

report by Matthews er al. [6], a succession of XUV and soft X- 1024 , 1

ray amplifiers and oscillators based on electronic transitions in e-Collisionalmultiply charged ions. This section will briefly review recent %., Recombnoationprogress in this area (through Fall 1990) and venture a few ; - ,,i-predictions as to future directions. E .. ., Excitation and

Dielectronic,,,,,Capture

"-Z", Beam

A. Atomic Lasers z .0 ,.N ,, /

Short wavelength (A < 200 nm) atomic lasers and amplifiers Plsm

developed in the past decade have relied exclusively on laser-produced plasmas as the excitation source. In most cases.population inversion is achieved by electron-impact excitation C 1018 Photoionization

of the upper laser level while, in other instances, incoherent Plasmas

soft X-rays produced by the hot. dense plasma photoionizethe species of interest. The success of this tool is perhaps bestillustrated by noting that, at the time of this writing, more than 1016 Rs a ncfe Bam35 ionic transitions in the soft X-ray region alone have lased Charge~ronsfer'Beamsor produced gain. 10 102 103

MacGowan et al. [81 have summarized the majority of Wavelength (A)the ions that have. to date, exhibited amplification in theXUV or soft X-ray regions. By irradiating thin selenium foils Fig. 1. Wavelength depende:lce of the mean particle density i.,.VI-

mounted on Formar substrates with 7 . 1013 W-cm- 2 of calculated in 15] to obtain, by sciected collisional and optical processes. gainsof ot = )c'u -; (,, = 0. 1cm- for ion beamsi. In this graph..N, denote% the

0.53 iim (frequency doubled Nd:YAG) photons. Matthews et number density of the particle (or photon) responsible for driving the pumping

al. [6], [9] at the Lawrence Livermore National Laboratory process of interest. For electron impact excitation of a lO-nm laser in a plasma.

observed amplification on 3p - 3s transitions of the Ne-like for example. t*v,\,tI/2 = i.N_VN )112 2 10,',cm-. In the region inwhich plasma Stark-effect line-broade-iing dominates, the Stark width scales

ion Se-÷ (Se XXV) at 20.64 and 20.98 nm. Vaporization of approximately linearly with the electron ( .V. or ion i.V, particle number

the target at such intensities produces plasmas characterized densities. This figure is reprinted from [5] by permission.

by electron densities of 3 - 5- 102°m 3-1 and a - 1-keVelectron temperature. Pumping of the 2p-l5 2 3p.,,2 (J = 0)level of the multiply charged Se ion is expected to occur 1015by electron-impact excitation of the ion's 2Ip' ground state.Spontaneous emission (on the 3s - 2p transition) rapidlydeactivates the 3s (1 1) lower laser level population of the "

ion. allowing gain coefficients of -,• 4.0cm'- to be realized :- :on the 20.6-nm transition. Several other Ne-like ions have S 0

exhibited gain at wavelengths ranging from 8.2 to 28.5 nm. , so:and Table I (adapted from [8]. [10]. and [11]) summarizes AS

the results obtained with nine atomic ions. The Zn. Ga. andAs oscillator data were obtained in experiments conductedat the Naval Research Laboratory. The spectral region over zwhich gain has been obtained (at discrete wavelengths) is 10 I0:,indicated, as is the measured gain coefficient. Also. the rapid an C c. Go

rise in peak laser-pump intensity that is required as the gainwavelength decreases is clear from Table I. This trend issupported by Waynant and Elton's [5] calculations in which.in the 1-100 nm region, the volumetric pumping rate was 102predicted to increase by four orders of magnitude for each 10 100

factor of ten decrease in wavelength (i.e.. 107W-cm- 3 at 100 L.ASER WAVELENGTH ( nm)

nm to 1011W-cm- at 10 nm). Rosen et al.'s calculations [12] Fig. 2. Peak la,,cr intensities required for pumping various neon-like atomic

underscore this behavior as they show that the pump laser ion lasers (data after 18l)ý

power required to reach saturation of a soft X-ray laser scaleswith wavelength as A-4 .

Fig. 1 (reprinted from [51) also illustrates the strong func- plasma length along the coordinate orthogonal to the optical

tional dependence of the mean plasma density on A. The axis of the laser is - lOOt.n then the pump intensities of

variation of peak laser intensity with laser wavelength (shown rable I and Fig. 2 correspond to instantaneous specific power

in Fie. 2 for several Ne-likc systems demonstrated to date) loadings of 6 . 1()14 to I lOl6W-cm-3. which are several

qualitatively reflects the increased demands placed on pump orders of magnitude higher than the estimate from [5] ofpower 1y movine to shorter wavelengths [81. [131. If the 101'W-cm-t for \ 0ll nm.

TABLE I

XUV AND SoFr X-RAY ATOMIC AMPUFIERS OBTAINED To DATE (ADAPTED FROM REFS. [8], [101, AND [111)

Ion Transition ,. nm Laser Intensity, W-cm-1 Gain(A,,..,, ) Coefficient (cm,'

Neon-Like Ions

Cu 3p-.3s 22.11. 27.93, 28.47 6-1012 (1.06, m) 1.7-2.0

Zn:0÷ 3 p-.3s 21.22. 26.23, 26.72 - 2.0-2.3

Ga'] 3p-3s 24.67. 25.11 -

Ge-+ 3 p-.3s 19.61. 23.22, 23.63, 24.73. 6,1012 (1.06 pm) 3.1-4.128.65

As'3- 3 p- 3 s 21.88. 22.26 - 5.4

Se-'4. 3 p- 3 s 18.24, 20.64, 20.98. 22.03. 7.1013 (532 nm) 2.2-4.926.29

Sr28÷ 3p-3

s 16.41, 16.65 1.4. 1014 (532 am) 4.0-4.4

y 29- 3p-.3s 15.50. 15.71 1.4.1014 (532 nm) -4

Me3o÷ 3p.-3s 10.64, 13.10. 13.27. 13.94 4.1014 (532 rim) 2.2-4.2

Nickel-Like Ions

Eu 35 4d-4p 6.58, 7.10, 10.04, 10.46 7-1013 (532 rim) 0.1-1.1

yb42÷ 4d-4p 5.03 1.4-1014 (532 nm) 1.2

Lithium-Like Ions

AIll(' 4f, 5f-3d 10.57. 15.47 4.1012- 1.7"1014 (532 rim, 1.0-3.5

1.06 pm. 10.6 pm)sill+ 4f-3d 12.9 2-1013 (10.6 pm) 1.5

Hydrogen-Like Ions

C- 3-2 18.2 5.1012 (10.6 prm), 4.1-8.01.5 1014 (532 nm)

Fs' 3-2 8.1 5.7-1014 (532 nm) 5.5

Similar schemes have been developed for inverting XUV N

and soft X-ray transitions in nickel-like, lithium-like, and LIM

hydrogen-like ions. One of the earliest reported soft X-ray N-l6o eS,.oce: 3-2trtn

amplifiers, based on the hydrogen-like C VI transition at 18.2 oN*V L1-NA.S*e *: 4?-31

nm, was discovered by Suckewer et al. at Princeton University O ecu

[14], [15]. Amplification was observed in a magnetically .o " o NMconfined plasma produced by irradiating a carbon disk with ,o

75-ns FWHM, 300-J pulses from a CO, laser. As indicated sxM 0rMt0 Main Table I. single-pass gains exceeding 6.5 were produced on Z aNr IM

the 3 -l transition of the C5" ion upon populating the upper SC=t 0 0

laser level by recombination. Vc, 0l o scc= 0

As shown in Table I. the Ni- and Li-like ion systems involve k% Im

4d - 4p and nf - 3d (n = 4. 5) transitions, respectively.Hagelstein (161 has recently simulated a prospective laser at 00 0

25 nm in nickel-like Moe13. Electron-impact excitation of IONIATION ENERGY (ev)

the ions's 3d' 0 ground state in a plasma having an electron Fig. 3. Variation of the lasing wavelength of hydrogen-like and lithium-like

temperature and electron density of 150 eV and (2 - 4) • ions with the ionization energy of the ion (repnnted from 1171 by permission)

101"rm-'. respectively, is expected to result in stimulatedemission on the 4d - 4p transition of the ion with predictedpeak gain coefficients of --- t).0fciinV- (at 25.4 nm). Only a It should be noted that although the fundamental physical

few of the possible Li- and H-like ion lasers have yet been processes occurring in these laser systems are presently not

demonstrated. aid Fig. 3 depicts the range in wavelengths that thoroughly understood, it is evident that the plasma parameters

is available with multiply charged ions having these electronic and n, and T, in particular, are critical to the attainmentconfigurations. of gain in this spectral region. Both are. in turn. related

3

to the target material and peak laser intensity. To a firstapproximation. the plasma blackbody temperature T is related C I 9 sxe4d9ss25p6

to the laser intensity I by C lcS i,

SZ 0.6(j,\2)2/3 (2) wo °ZZ

where T. I, and Aare expressed in electronvolts. W-cm-2 . N.Pand centimeters, respectively. Also, the large optical pump Ar3l p4 ,p(3 p)

2 p

intensities that are necessary to realize gain underscore the 2 0-3)

validity of Hagelstein's comment that [161 "No one has yet S (developed an X-ray laser at home in their garage." U 405,6, I

At longer wavelengths. Kapteyn et al. [18], Silfvast and 20 40 60 so 1oo 120Wood [19], and Sher et al. 120] have demonstrated lasing ENERGY (*Vat 108.9 nm from doubly ionized xenon by photoionizingthe neutral ground stateoub ly with ,.,o by sot X-ray Fig. 4. Dependence of the inner-shell photoionization cross sections on

edirectly with - -eV soft X-ray photon energy for several of the atomic lasers reported in 1191 (reprinted

photons. This photoabsorption process rermoves an inner- by permission).

shell 4d elec:ron, producing the 2D33 2.5/2 5St; c of Xe", whichsubsequently decays by an Auger process i the 5so 5p 6 'Sostate of Xe Ill Both tantalum and gold-eleci plated stainless- velocity distribution is roughly Maxwellian. Consequently, thesteel targets ere used to produce the soft '-rays, and peak excitation rate (av) to the upper laser level-the metastablelaser intensit s (A = 1.06 /m) were 2- 101 - 1012 W-cm- 2 . Cs 5p 5 5d 6s 4 D, . level-is 3.5.10-9cma-s-' which yields aThe efficiency for converting 1.06-/pm infraied pump power peak upper state population of 1.2.10 14 cm-3 . Lasing occurredinto soft X-r:.v radiation was estimated to be '-7% [181, [201, in the extreme ultraviolet at 96.9 nm, and the small signal gainand the blac0.body temperature of the plasr i was calculated coefficient was measured to be 4.9 cm-'.to be 30 eV. Kapteyn et al. [181 also four ! that preventing Efforts by Rocca et al. to produce UV lasers in CW electron-photons of energy < 50 eV from reaching tf: Xe gas enhanced beam-pumped plasmas (221 have yielded 14 mW on the 318.1the laser output by almost an order of rr. enitude. Photons nm [5s 'G 4 - 5p 3 F3 1 transition of the silver ion [231.below this energy threshold photoionize '-c atom, but eject Discharge currents exceeding 1 A at 2.5 kV from a twoa 5s, rather than 4d, electron-thus unne, -,sarily increasing electron-gun arrangement have made it possible to achievethe electron density without contributing ) pumping of the continuous lasing on this and several visible lines of Ag, Zn,upper laser level. By exciting the Xe III aser in traveling- Cd, and F by direct electron-impact excitation. In order to im-wave geometry, Sher et al. [20] measured a mall signal gain of prove the specific power loading of the medium and extend thee40 (-Yo = 4.4 cm- 1 ) and single pulse outp, . cnergies of 20 pJ range of available wavelengths into the VUV, Rocca's groupin the forward direction and 0.4 tLJ in the ,:;ckward direction, is also exploring capillary discharges 1241. Pulsed capillary-The key to making this system practical ua is the development confined discharges in He produce electron densities in excessof a novel. oblique-incidence pumping geometry. Presently, of 10 7 cm-3 , and have the added advantage of rapid plasma

the Xe laser can be effectively pumped with < 500 mJ of cooling by conduction to the discharge cell walls. Strong1064 nm (Nd:YAG) radiation and operates at a pulse repetition emission was observed from He II in an effort to obtain gainfrequency of 2 Hz. It is anticipated that further development on its hydrogenic 3 - 2 line. Similar experiments with lithiumof these techniques will lead to practical sources of even produced XUV radiation at 72.9 nm from doubly charged Lishorter wavelength radiation. The reduced pump energies will ions formed in discharges in capillaries fabricated from lithiumpermit higher repetition rates (>10 Hz) and result in accessi- hydride. Peak current densities up to 2.5- 106A-cm- 2 in a 40-bility of the sources to nonlaser scientists for applications in kV discharge yielded emission on the 3 - 2 line of Li", whichspectroscopy, imaging, high-resolution holography, nonlinear noticeably intensified in the plasma afterglow, presumably dueoptics. and materials studies. to pumping of the Li Ill, n = 3 state by three-body electron-ion

In addition to the Xe Auger laser, Silfvast and Wood [191 recombination in the cooling plasma.have reported soft X-ray photoionization-pumped lasers in In' Experiments conducted by Petersen [25] have recentlyat 185 nm and He at 164 nm. Both involve the removal of an yielded more than 80 new CW lasing lines from ionicinner shell electron, and Fig. 4 shows the energy dependence transitions of the rare gases. With a - 2.6-mm bore. butof the photoionization cross sections of interest for several otherwise conventional ion-laser plasma tube, he reportedelements that have lased when pumped by soft X-rays [19]. continuous lasing on 46 lines in the 200-300 nm region, and 43

A photoelectron-pumped laser in neutral Cs vapor was in the 300-400 nm interval. Most exhibited threshold currentsrecently demonstrated by Barty et al. 121). Although initiated of 40-70 A ý730 - 1280A-cm- 2 ), which is a reflection ofby the production of soft X-rays at a stainless-steel target. this the fact that the upper levels for these laser transitions lielaser is actually collisionalily pumped by hot photoelectrons. as much as 125 eV above the ground state for the neutralThe conversion efficiency of the 1.3,) 1.01'2W-cm-'. 1.064- species. Single-line output powers up to - 0.65 W have been/im pump intensity into soft X-rays was - 3%. giving rise obtained, and. for Kr. the output power for all the lines in theto a plasma radiating a - 25-eV blackbody spectrum. At this 242-266 nm region was 2.5 W. In Ar. the lasing transitions intemperature. the photoelectron density is - l(l16ct - and the the 275-306 nm range yielded as much as h W of CW power.

4

B. Molecular Lasers s

The introduction of new UV and VUV molecular lasers over 0(

the past several years has slowed considerably as emphasis has 12eV Blackbodyshifted to the development of existing systems. and to the F: 3_ c

laser, in particular. Recent experiments conducted by Yamada aet al. [26] increased the output pulse energy of this laser 2

by almost an order of magnitude over previous devices (to 2beyond 100 mJ/pulse). Improvements in the laser's extractionefficiency were attributed to a reduced anode-cathode gap(and, hence, increased E/N) and the use of diluent pressures (b)as high as -- 6000 torr (-. 8 atm). Modeling of this system 10by Kim et at. [271 and Ohwa and Obara (28] bear out these •results. Efficient operation of the laser requires specific powerloadings of the active medium in the -- 4 - 23 MW-cm- 3 6

range [26], which is roughly an order of magnitude larger ,.O 4

than that required for the rare gas-halide systems. This. of ZU 2

course, is a reflection of the large ionization potentials of 2

both of the constituents of the F, laser gas mixture (He. F2 ). 0 (c)For a He/F: (99.85/0.15%) gas mixture at a total pressureof 2.6 atm. for example. Ohwa and Obara [28] calculatethe electron temperature to be 4.6 eV for E/N = 7.7 ., ,,- 210_1TV cm- 2 . 0 r

c0-.The higher power deposition requirements for the F: laserstem not only from the higher power deposition inherentwith high-pressure operation. but also from several kineticsfactors. Simulations [28] suggest that the acceleration in theremoval rate for the laser's lower level [F_(A')] population (d)and temporal broadening of the laser pulse are critical issues zfavoring increased background pressures. Consequently, to 9iobtain intrinsic laser efficiencies exceeding 1% requires total e

,X Werner Bandgas pressures of at least 4 atm [281. Similar conclusions were Z 30

reached by Kim et al. [27]. '20The F2 laser has been developed to such an extent U,

that commercially available systems now offer several watts 10

of output power at pulse repetition frequencies exceeding O 50.0 50 100 :so 2CD50 Hz. While the gas lifetime is below 106 shots, ENERGY (eV)

continuing improvements in laser-head construction materials Fig. 5. Data relevant to the operation of the H:- 16-nm laser pumped

and design promise to extend the gas-fill lifetime and by a photoionization electron source. Comparison of: (a) the photon energy

improve single-pulse output energies by at least a factor of distribution emitted by a 12-eV blackbodv. (b) the H: phototonization crosssection, (c) the calculated electron density, and (d) the electron-impact

two. excitation cross section for the Weiner band (C-X) of H: (reprinted from

A striking example of the utility of laser-produced plasmas [291 by permission).

in pumping molecular lasers by electron impact is the H, VUVlaser experiments reported by Benerofe et. al. [291. Molecular C. Future Prospectshydrogen was photoionized with soft X-rays produced by atraveling-wave laser plasma. When the plasma is produced Since the single-pulse energies and repetition rates availableby - 580 mJ. 200-ps pulses from an Nd:YAG laser, free with a transverse discharge-pumped F2 laser exceed those fromelectrons generated by the photoionization process have an other sources in the VUV, near-term efforts will undoubtedlyaverage energy of 10 eV. and the electron density is estimated focus on further improving the efficiency of this laser. How-to be :3 • 101'cm-l in the active volume. Figs. 5(a) and (b) ever, since the D' - A' lasing transition in F, occurs betweencompare the photon energy distribution for a 12-eV black- two bound states, the tunability of this system is limited and thebody with the H, photoionization cross section. The electron development of primary sources in the VUV (or XUV for thatenergy distribution calculated in (29] is illustrated in Fig, 5(c) matter) that are tunable would be extremely valuable for sev-and. Fig. 5(d) shows the H, ClY, - Xl'v1 " (Werner band) eral applications. including spectroscopy, optical imaging. andelectron-impact. excitation cross section. With this pumping materials analysis. Toward that end. Sauerbrey and Langhofftechnique a gain coefficient of 0.9 cm-1 was measured at [30] have proposed and explored the excimer (bound - free)room temperature for a gain length of 27 cm. Lowering the transitions of the singly charged alkali-halide ions as possiblegas temperature to 11)0) K results in the gaia coefficient rising VUV laser-active media. Following the trend toward higherto -• 1.5 ,'tt - stages of ionization in atomic systems, increased attention to

ionized small molecules seems inevitable in order to satisfy the are He-Ne lasers which operate on a CW basis at gas pressurescompeting requirements of tunability and short wavelength, of a few torr and power depositions of a few to 10's W-cm- 3 .

For the laser plasma-pumped atomic lasers, increased efforts Due to the specialized nature of gas discharges for VNUVin the optical properties of materials in the VUV. XUV. and lasers, it is difficult to present a general discussion ofsoft X-ray spectral regions and the application of MBE and their properties. This uniqueness is illustrated by discussingMOCVD film-deposition techniques to multilayer fabrication how gas discharges may be designed to enhance certainwill lead to steady improvements in the performance of short- plasma-chemical attributes which, in this case. result in laserwavelength optical components. A major impediment in the oscillation. In this section we will discuss excimer lasers aspast to progress has been the quality of available optics. but an exemplary system of how the gas discharge may set limitsthe plenary paper given at CLEO'91 by R. Freeman of Bell on the performance of the laser.Laboratories highlighted the advances that have been realizedrecently in the fabrication of reflective optics in the soft X-rayregion. Improvements in the interfacial quality of multilayer A. General Properties of Excimer Lasersstructures have yielded reflectivities exceeding 60-70% at 130 Electric discharge excimer lasers provide highly efficientnm. for example. Over the next five years, one can expect even sources of pulsed coherent ultraviolet (UV) radiation [31].greater strides to be taken in this area. which will translate The term "excimer'" has now come to refer to a class ofinto soft X-ray oscillators that have higher peak powers excited molecules which are bound in their excited states.and yet are more compact. Coupled with more widespread and repulsive or weakly bound (Do < kT) in their grounduse of the traveling-wave excitation technique developed at states. Laser transitions typically occur between the lowestStanford. such advances will lower threshold pumping powers. excited states of these molecules and their ground states.bringing these atomic lasers within the financial reach of more These molecules therebv constitute nearly ideal laser media.laboratories. Improvements in target design so as to more Since their ground states are unstable, inversions are readilyefficiently convert near-IR or visible laser power into soft obtained by pumping the upper level. Transitions to the groundX-ravs of the desired energies will encourage this trend. state then result in dissociation of the molecule. The fact

that their ground states are not stable, though, means that

1Il. VISIBLE AND NEAR-UV LASERS excimer species themselves cannot be feedstock gases andmust be generated "real time" in the plasma. Excimer lasers areInfrared lasers which operate on transitions between vibra-thrfeacutlydsibds"psm-eia"lsr.

tional levels of the lowest electronic state have upper levels A a rity o ex cimer m s havea demosate laser

which are at most a few tenths of an electronvolt above the Avreyo xie oeue aedmntae aewhich arel. Atmst a fewultenths ofschangesectonvt aboveay t- action. The most successful in terms of efficiency, laser energy.ground level. As a result, gas discharges can be easily con- adtcnlgcldvlpetaeterr a aoe RHand technological development are the rare gas halogen (RgH)structed which have characteristic energies commensurate with excimers. The RgH lasers that have been most intensivelythese values. Visible and near-UV lasers, however, operate on developed are ArF (193 nm). KrF (248 nm). XeCI (308 nm).electronic transitions whose upper laser level is a few to 10's of and XeF (351 rm). The excited RgH excimer results from theelectronvolts above the ground state. It is difficult, particularly ion pair consisting of the rare gas (Rg) positive ion and theon a CW basis, to construct gas discharges which have halogen (H) negative ion. (See Fig. 6.) These excimer lasersthese characteristic energies. As a result, visible and near-UV typically operate between the bound B state of the ion pair and(VNUV) lasers usually rely on a fortuitous sequence of kinetic the repulsive (or weakly bound) X state. The B-X transitionsprocesses to excite the upper laser level. For example, many have wavelengths ranging from 192 nm for ArF(B-X) to 351metal ion lasers are excited by excitation transfer from an

nm for XeF(B-'X). Visible wavelengths can be obtained byelectronically excited noble gas atom whose levels fortuitously operating on the lower gain (C-A) transitions of the XeCIalign themselves with the appropriate levels in the metal atom. and XeF molecules.For example. in the Cd÷ laser, the upper levels are produced A schematic of a typical discharge laser appears in Fig. 7by excitation transfer from helium metastable atoms. [311[1331. The electrode separation is a few to 10 cm. with

Ilvt 2 'S. 2"S) + C Cd~~- (5,..D-, 2) _; 12,) lengths of 10's of cm to > I m. The gas pressure is 1-5 atm

(3a) (see below). The discharge circuit consists of a pulse-forming

Laser oscillation then follows on the 441.6-nm transition line (PFL) with electrical lengths of 10's to 100's of ns. Thecharging voltage of the PFL is 1-4 kV/cm-atm. resulting in

('(I-(.Qs' 2 D-,/ 2 ) - Cd+(5t) 2 P 1/2 ) ÷ h-,(2.N'V) (3b) a power deposition of 100 kW-cm-i to many MW-cm-.

The voltage is switched onto the laser head using a thyratron(Gas discharges used for VNUV lasers therefore tend to or low-inductance rail-gap switch, causing breakdown and

be somewhat specialized, each being designed to persuade sustaining a discharge current of 100's of A-cm--. Thethe energy flow to follow a certain kinetic pathway which impedance of the discharge is typically 0.1-0.5 Q-m basedenables threshold to be reached. For example. at one end of on a square aperture. Variants of the discharge circuit usethc extreme are excimer lasers ( 193-351 nm) which are excited saturable magnetic inductors (351 to aid in switching, or ah% pulsed ieas discharges having pulse lengths ot 10-100(1 ns. separate "'spiker'" circuit to breakdown the gas [36). thereby!as prcssurcs o• 1-5 iam. and a power deposition ot 101) kW- allowing better impedance matching between the sustainer'M I,:tc%% %IW-cr n At the opposite end oh the catceor, PFL and laser head. It is mandatory that the discharge volume

12 Kr-" + F- -- %I -- KrF(B. v) - Kr + %I l5)

10 where M is a stabilizing third body. The neutral channelS.consists dominantly of harpooning reactions. exemplified by

8- 1 3/2-Kr x (z P3,'21÷F" [461

6. C ltI, Kr' - Fý -- KrF(B.v0 - F (6)02

Z4 where Kr* represents Kr in an excited state. As argon is// often used in the gas mixture for a KrF. additional excitation

2 reactions such as

0-A2nv.a KF(2.,z ArF(B) + Kr - KrF(B. vi I- Ar (7)

x It/,,z are possible where the ArF exciplex was previously formed by2 3 4 5 6 7 8 reactions analogous to those shown above. These formation

R (1) reactions are exothermic bv a few to many electronvolts.

Fig. n. Energy levels of the K'F exciplex. Laser oscillation at 248 nm takes resulting in KrF(B. v) being formed high in the vibrationalplace on the B--X transition 131]. manifold. Lasing, however, occurs out of KrF(B. v=0). There-

fore collisional relaxation of the vibrational manifold must

SPARK-GAPOR occur before lasing [47].THYRATRONTYPICAL) HIGH IMPEDANCE

SPIKER PFL (OPTIONAL) B. Scaling Considerations

OISCHARGE ( "'TO CHARGING The formation efficiency (power channeled to the upperELECTRODE -..- CIRCUITS) laser level compared to power deposited in the plasma) of

"- KrF excimer lasers may be as high as 40%. This highLOW IMPEDANCE efficiency results from the fact that with the exception of

S.I. rSUSTAINER PFLLASER _[ \ ". ,, the repulsive A states of the F, molecule (which has a smallM)RRlOR" TO PULSE POWERI��R� FOR PREIONZAVrO• mole fraction anyway), the upper laser level is energetically

PREIONIZATION SOURCE the lowest excited electronic state in the system. Therefore4_ power deposited in the plasma will naturally channel by

excitation transfer down to the upper laser level [48]. ForFig. 7 Schematic of a discharge-excited excimer laser. The discharge ex a tleo suppose a n to th p er lan ele [48).im or

electrodes are typically separated by a few to 10 cm and have lengths of examp an Ar* is produced by an electronic-impact10's of cm to I m. The optical axis is parallel to the long dimension of the collision. Dimerization rapidly follows, after which chargeelectrodes transfer produces Kr*. This reaction is followed by an ion-ion

neutralization which produces the upper laser level:

be uniformly preionized to 107 - 10'(cn- 3 prior to applying Ar Kr Kr F-high voltage. Preionization is accomplished using UV spark e + Ar Ar- - Ar' - Kr - Kr- KrF(B). (8)arrays [371. X-rays [381, [39]. and corona bars [401.

Commercial excimer lasers are currently available whichdeliver a few joules of pulse energy at repetition rates of Gas mixtures for excimer lasers are usually tertiary, con-100's of Hz. yielding as much as 500 W [41]. The discharge taining a lower atomic weight noble gas in addition to the rareapertures on these devices are a few centimeters. wallplug gas and halogen for the excimer molecule. For example. inefficienc, , are a few percent, and the laser pulse lengths are discharge-excited KrF lasers a typical gas mixture is He/Kr. F-usually < .30 ns. Higher pulse energies (<_ 66 J) [42], [43], = 99.5/0.5/0.05. Although binary gas mixtures (e.g.. KriF:)larger apertures (:_ 20 cm) [431, higher efficiencies of (!_ 4%) have been investigated, optimum performance is usually ob-[44], and longer pulse lengths of (< 1.5 /Ls) [371 have been tained with a tertiary mixture. The tertiary mixture is necessar,obtained with XeCI laboratory lasers, although not all on a due to the large momentum-transfer cross section of the Rgsingle device. which forms the exciplex. The cross sections for electron-

Excited states of the RgH exciplex are typically generated impact momentum transfer am for Ar. Kr. and Xe exceedvia an ion-ion neutralization process constituting the ion 10 A-' for energies in the range of 5-10 eV [49].channel, or a harpooning reaction constituting the neutral These values of am would require excessively high oper-channel. For purposes of discussion we will use the KrF laser ating and breakdown voltages for large aperture dischargesin Ar/Kr;F2 and He/KriF2 mixtures as a model system. The (greater than a few centimeters). Therefore the average ,:processes discussed for KrF are directly analogous to other mixture must be reduced by diluting with either He or Ne.RgH systems. For example. the KrF(B) molecule is formed both of which have lower momentum transfer cross sectionsthrough the ion channel by [44]. 145] compared to the heavier rare gases. In practice. avalanche-

discharge excimer lasers (pulse lengths < 20-30 ns) tend to useKr- - F- - N KrF(B. v) - \1 (4) He as a buffer gas. Long-pulse I> many 10's ns) discharge

excimer lasers tend to use Ne as a buffer gas. since its lower where (T. is the optical stimulated-emission cross section.momentum transfer cross section at energies < 10 eV is Since the excimer transition is essentially bound-free, theadvantageous. optical emission cross section tends to bc small compared

The requirement that excimer lasers must operate at pres- to that for atomic bound-bound transitions. For example. thesurcs exceeding an atmosphere results from the fact that the optical stimulated-emission cross section for KrF is 2.6 xformation of the upper laser level through the ion channel 1l0- ci" ' [31), which combined with a formation efficiencyrequires a three-body reaction in order for its rate to be of 25%- and effective lifetime of 5 ns. the power deposition incompetitive with other excitaion or quenching processes. For a KrF laser must be 250 kW-cm 3 to obtain a small signal painexample, the rate constant for ioa-ion neutralization resulting of 0.1 cti-. With this power deposition. the system must bein atomic products, operated on a pulsed basis.

The fact that high-power deposition is required sets furtherKr- - F- - Kr + F. (9) limits on the gas pressure. Power deposition in an electric

discharge is given bv P = jaE_, where j,j is the discharge( 1 -i(rruui~s- (501. The ion-ion neutralization current density and E. is the self-sustaining electric field. The

reaction resulting in formation of the exciplex dimer is a self-sustaining electric field is determined almost entirely bythree-body reaction [--4. (45): the gas mixture. It is that value where

Kr- - F- - % - KrF(B) 4- MI (10) = Etfsk,, (12)

,.where %I is any third body and is required to conserve momen- where f, is the mole fraction of component .). ki,tum. In order for the exciplex-torming reaction to effectively is its ionization rate coefficient, and k,, is its attach-

compete with the neutralization reaction, the density of th.- ment rate coefficient. For typical lasers E,'.V%: 2 -

third body must be .NI] >, k•/k 1 . Although there is a 5Td ( m= I x 1O-tV-cm-) [331. [35) or E/P -0.4 -

complicated dependence of kl 0 on [MI. over the range of 0-3 IA) kVicm-atm. The current density is j = Ezn-/ 771 , I

atm. ki,,.\I = .\r! A 3 x I0-'cm3s-<atm-1 [39]. Therefore and in typical gas mixtures. i,,, = *2 X 1i012 p(atm) s-1. These

the total gas pressure must exceed an atmosphere in order for values result in Ji z 3.5 x 10'3.-, (kV/cm-atm) p(atm) f kA-

the reaction (10), which forms the exciplex. to dominate. cm-'. where f is the fractional ionization. Power deposition

Due to the efficiency with which deposited power is chan- in an excimer laser discharge is therefore:neled to the upper laser level, the characteristics of laser P(kW-cm 3.5 x I0•E 2 (kV/cm-atmtp-2(atnif. (13)operation are largely a function of local power deposition.

Therefore the manner and uniformity of power deposition Since discharge lasers typically operate with fractional ion-is often the greatest concern in constructing the device and ization of 10-U to 10-'. to obtain power depositions of 100'soptimizing laser performance [341. [51)-(541. The uniformity kW-cm the gas pressure must be at least a few atmospheres.Of povwcr deposition is largely determined by the distributionOf the preionization density. Since near its optimum value. C. Unijormitty and Instabilitieslaser efficicncy is relatively insensitive to small chances inother parameters such as gas mixtures, power deposition is The purpose of the discussion above is to motivate the

the only remaining parameter which may be optimized and reader to view discharge-excited excimer lasers as systems

which characterizes the laser. For example. in discharge- whose pertormance depends primarily on local power deposi-

excimer lasers, discharge instabilities limit power deposition tion. The details of the kinetic processes leading to excitation

and constrain laser performance. as opposed to being a limiting of the upper laser level bracket the specific gas pressure

aspect of the kinetics. and power deposition which are required. To the extent that

From past experience and simple scaling laws one could excimer lasers are power-driven devices, their performancerecommend that excimer lasers he operated at low-power is typically lin;ted by issues relating to power deposition asdeposition to enhance their discharge stability. The short opposed to any particular kinetic mechanism. For example. theradiative lifetime of the upper laser level, however, sets laser pulses in electric discharge-excited excimers are usuallystringent requirements on pump rate and gas pressure. For terminated bv discharge instabilities (e.g.. arcing. constriction.example. the radiative lifetime of the rare gas-halide excimers streamers) which prevent the gas from being uniformly excited13 state ranges from 0 ns for KrF to 15 ns for XeF [31]. rather than bottle-necking in the kinetics. These dischargeWhen also considering quenching processes. the total lifetime instabilities, which develop on time scales of 10's to 100's

ot the B state is 50-70'% of the radiative value. The short of nanoseconds. originate from a ,,nuniformitv in gas mix-lifetimes ot these molecules sets a requirement on the rate of ture. preionization. density, or electric field fresulting from

pumping (t the medium. For example. for power deposition electrode contouring).I' , W-cm ý and formation efficiency if. the unsaturated To examine the limiting aspects of discharge stability ondensity of the upper laser level is .V,, = rfl½/, where hi, 1;'- performance of excimer lasers it may be useful to viewis the laser photon energy. To obtain a specific small signal the plasma as simply a network of nonlinear resistors. Since

gain. b., :1 . the power deposition must he cxcimer lasers arc typically transversely excited, the plasmamay, be logically divided into segments 'w'.hich are perpendicu-

' ;,, , I I l) lar it both the electrodes and the optical axis. thereh, making

a parallel array. The conductivity of each segment a, and the 20, - -, , 3I0

total conductivity of the discharge or are -- rxeony ,-• 5C! - Voltoge

,, (4/),(nq/.,,. a = E,or, (14) C o - -e20C

where I is the length of the segment. .4 is its cross-sectional U. Zarea, rn is the electron density in the segment, and v,, is the /o\'\\ oelectron momentum-transfer collision frequency. The power 0> 0 Ddeposition in a particular segment is therefore: '

P =jE, = T, E2 X rz,. (15) 00 25 075 00o 125

If we limit ourselves to plane parallel electrodes, the E/N TIME (ns)

across each plasma segment is the same. The relative power Fig. 8. Experimental and theoretical waveforms for the discharge-exctted

deposition between segments therefore scales to first order as laser constructed by Watanabe and Endoh [331.the ratio of their electron densities. Instabilities resulting inlargely disparate power deposition must then have originatedfrom processes which cause or reinforce small changes in electron collisions, and also includes models for the externalelectron density, electrical circuitry and laser extraction. The model is a lime-

In excimer lasers, where electron loss is dominated by dependent one-dimensional representation and is based onattachment to the halogen donor (e.g.. F- in a KrF laser) models described in [351 and [55]-[57]. The experimentaland in which the halogen donor is consumed during the conditions foi this demonstration are those of Watanabe anddischarge pulse. small changes in the initial preionization Endoh [331. The device uses a He/Kr/F2 = 99.47/0.44,0.09electron density can lead directly to constriction and arcing. gas mixture at 4.5 atm. The discharge active volume is 7 xUnder quasi-steady-state conditions the local self-sustaining 7 x 80 cm 3. An intrinsic laser-energy efficievcy of 1.2% waselectric field is determined by a balance between ionization obtained with an output energy of 10 J. The experimental andand electron loss. in this case attachment. The operating point theoretical current and voltage waveforms are shown in Fig.of the discharge is then the value of E/N which satisfies (12). 8.The mole-fraction-weighted rate coefficients for ionization An instability can result from a nonuniformity in electricand attachment are functions of halogen density, excited-state field. preionization electron density, or gas mixture whichdensity, and electron density, in addition to being functions of results in nonuniform halogen burnup. A common source isE/N. Therefore the operating point in a discharge may vary nonuniformities in preionization electron ctknsitv. To demon-as a function of position as these parameters also are spatially strate this effect we have varied the preionization densityvarying, profile in the discharge KrF laser. Power deposition at the

An instability will occur, because when the operating end of the current pulse is shown in Fig. 9 for fractional edge-points of various segments of the discharge are different, to-center nonuniformities in a preionization density of 0-4).1the discharge cannot operate at all of those points (highest on axis). The spatial dimension in the figure is parallelsimultaneously. Since the operating point E/N will be to the electrodes and perpendicular to the optical axis. In allhighest where the halogen density is largest, then there cases, power deposition at the leading edge of the dischargewill be an ionization excess at locations where the pulse is fairly uniform as a function of position and no worsehalogen density is lower. The electron density thereby than in the ratio of the preionization density. At these early

increases in those plasma segments. If the halogen donor times the F2 concentration has not been significantly depletedis consumed by the attachment process (e.g., e + F: - so that all spatial locations have the same E/.V operatingF + F) and is not replenished during the discharge pulse. point. As the F2 is gradually depleted more rapidly near thethen the rate of burnup of the halogen also increases, axis where the preionization density is the highest. the stable

leading to a higher electron density and further burnup. operating point at those locations decreases to a value smallerThis process is unstable and will eventually lead to than the average E/N, thereby creating an ionization excess.arcing. An electron avalanche is then initiated, resulting in more rapid

The rate-limiting process of the performance of a discharge- depletion of F2 . This results in constriction of the dischargeexcited excimer laser is therefore the uniformity of power and termination of laser oscillation.deposition. The uniformity of power deposition is determined Another effect which can cause an instability is Ei.V notbv the uniformity of the preionization density, electric field, being uniform, which may occur as a result of nonalignedand reactants (e.g., halogen density) which may cause the electrodes. To demonstrate this effect we varied the alignmentoperating point of a given plasma segment (that is. its lo- of the electrodes to cause an edge-to-center fractional changecal self-sustaining E/N) to be different from that provided in E/.V of 0-4).01 (highest on the axis). while keeping thebv the discharge. The degree of nonuniformity which may preionization density uniform. These results are shown in Fig.

be tolerated is quite small. To demonstrate this effect an 10. Only small variation I<< 0.l7,) -hanges in E,'.N can beelectric discharge KrF(B-.X) laser has been modeled. This tolerated before constriction of the discharge and terminationsimulation includes rate equations for the heavy particle and of laser oscillation occurs.

MAXIS TABLE If5 TIME CoNsTANTsis) OF EsERGY REL4.AXntON

t 12 nsPROCESSES IN MOLECU~LAR GAS DISCHARGES

EDGE TO CENTER Space-charee relaxation lIY-I1-loE NONUNIFORMITY IN Electron energy relaxation 1`-0-

PREiONIZATION =0 1 Ionization and attachment o-oM Vibrational relaxationZ 3 003 Thermal energy iliffusion o41-

0.0 Gas residence lime Io0`-l0l

0.01 Plasma chemistrv 10s11Io0I

the gas mixture too rapidly, or alternatively the system should0 1 allow for these changes. In practical situations this usually

means that dissociation processes should not severely deplete0 1 1 1the lasing molecule density and the products should not quench

0)0 0 20 30 40 the lasing species. Therefore E/N should not be too highPOSITION (cml and the discharge should not provide many energetic electrons

capable of causing dissociation, If the new species formedFig,. 9. Power deposition at the end oftIhe current pulse when the preioniza- have large attachment rates. this can lead to a macroscopiclion densits has ai tractional edve to the center nonunitormilv of0 01 (highest discharge instability. The third feature is that the temperatureon the axis).

of the gas mixture must not be allowed to become comparableto the energy of the lower level. otherwise thermal pumping

AXIS of this level will prevent the population inversion. Heating of

I the gases occurs due to vibrational-rotational to translationalt 725 rsenergy transfer from excited vibrational states (including those

= 125 ISwhose populations are increased by lasing), and to a lesser

extent by electron molecule elastic collisions.EDGE TO CENTER Depending on the boundary conditions. several of these

MNONUNIFORMITY IN effects can be occurring simultaneously. A summary of theE 00 estimated time constants for the principal phenomena in these

discharges is given in Table 11. Some of the important lasers0 and the range of lasing transitions are listed in Table Ill.

0003 The molecules CO,, N.O. OCS. CS,. and HCN are linear4 00000 triatomics. and all have-two stretchine vibrational modes

and a bending mode. The CO,. N.O. and CS, systems canbe excited directly in discharges and also have operated by

000 -.,j7~---.---Z2~,energy transfer from vibrationallv excited nitrogen. The OCS00 0T 20 3,0 40 transitions around 3.25 pmi appear in high-current pulsed

POSITION (cml discharges at low pressures. Various I-CN transitions occur

Fig. 10. Power deposition at the end of the current pulse when the applied in pulsed and/or CW discharges from I-CN or a variety ofelectric field has a tractional edge-to-cenler nontuniformitt of 0.01 (highest precursors such as CH., and NH-,,N, or CH 4CN.on the axis). The fortuitous near-coincidence of the vibrational energy

levels of N, (I' = 1) and CO,. (001) and the consequent high

~ INRARE MOECULR-GS LAERSefficiency of this laser have led to a dominance of researchon the C02-N2 -- He laser among infrared lasers 1581. Thehighest efficiency laser (- 30'/u), summing over all lasing

A. Introduction lines. is the carbon monoxide laser, which also depends on

Infrared lasers generated by electrical discharges in molecu- vibrational excitation. This high efficiency is achieved in CO.lar gases must satisfy three principal requirements if they are to because an absolute population inversion is not required and.be efficient and useful. First, the discharge must excite certain under appropriate conditions, there is fast redistribution ofvibrational modes efficiently. These can be the vibrational the vibrational energy by anharmonic pumping (591. Also.modes of the lasing species or of donor molecules. In terms multiple wavelength cascade lasing from levels as high asof the gas-discharge parameters. this means that selection r= .17 to i,= :16 down to the r=I to r 0 transitionof the mean electron energy is accomplished by control of can occur.the E/X parameter. One attempts to optimize the overlap The characteristic energies involved in the infrared lasersbetween the EEDF (electron energy distribution function) and determine the operational parameters. The 5- and 10-pmthe excitation cross sections. Secondly, the plasma chemistry wavelengths correspond to approximatelly 0.2- and 0.1-eViflduced !-% -he discharce should not change the composition (t'f transitions, respectively. The ionization energies of the lasing

TABLE III and so had little or no control on the gas chemistry: and

PRINCIPAL TRANSmONS IN MOLECULAR GAS DiSCtARGE IR LASERS finally, considerably more research was concentrated on the

Molecule Wavelength Range Important Lines unstable forms of the discharge. rather than attempting to(Pm) extend the regimes of uniform plasma operation. Assistance

Co 4.9--6.66 5.263. 5.113 to the development of laser physics came primarily from the

DBr 5.8--6.33 6.19. 6.224 knowledge of cross sections [60], the collisional processes

DCI 5.04-5.6 5.15. 5.18 involved, and the theoretical infrastructure for the calculation

DF 3.83-4.02 2.71. 2.76. 2.83 of electron energy distribution functions and energy depositionin gases.

H, 0.835-1.3 1.12. 1.30

HBr 4.0-4.65: 19.4-23.4 4.2. 4.46. 4.57

HCI 3.7-4: 13.9-19.8 3.84 B. Specific Molecular Laser Processes

HF 2.64-3.05:10.2-21.8 2.83 Research in gaseous electronics and gas discharge physics

N, 0.75--8.2 0.87. 0.9 actually had indicated the possibility of laser action in molec-

CO, 4.32-4.38; 9.12-17.4 10.6 ular gases very early, but the clues were not pulled togetheruntil Patel's experiments. In his book Electrons in Gases, Sir

H•0 26.4-218.5 33.9. 35.1 John Townsend [641 noted that there is a large increase in

H2O 2.28--67.2 79. 90 electron-energy losses in collisions with N2 molecules above

HS 33.48-225 162. 225 a mean energy of 0.8 eV. Reviewing data obtained in 1929, heHCN 12.85-773 337 wrote: "In N, it must be assumed that in the experiments on

the determination of the mean energy and drift velocities, thereN2,O I0.34-i 1.04 10.5. 10.9are losses of energy in large amounts in some of the collisions

SO2 10.85-215.3 140.85 of the electrons with atoms of nitrogen where the value of

OCS 8.24--8.42: 123-132 - E/P is 20, and the mean energy of the electrons is 2.2 eV."

CH. 3.04 - In this quote the word "atoms" is used in the generic sense tomean atoms or molecules. It must be remembered that it wasnot readily apparent at that time how an electron collision in

species are approximately 15 eV, and the electron-impact nitrogen could cause vibrational excitation efficiently or whydissociation energies are in the range of 6 to 10 eV. The direct it would require 2.3 rather than 0.3 eV, the vibrational modevibrational excitation energies are 0.3 eV, and the much more energy. Thus the understanding of the efficient vibrationalimportant resonant vibrational-excitation energies typically are excitation of nitrogen lay dormant until the temporary negative2 to 4 eV. Thus. one aims to maintain an EEDF that provides ion mechanism was identified (cf. [65]).vibrational excitation and sufficient ionization to sustain the A large number of molecular gas-laser transitions in various

power loading into the discharge, but not too high an average gases have been shown to be possible using discharge or

electron energy, otherwise dissociation is rapid. There are discharge-assisted excitation [66]. Only the higher efficiency

various methods by which these constraints can be satisfied. and noncorrosive systems have received concentrated attentionUnder low-pressure nonflowing conditions, one can add a gas and full investigation. There are many reports of different lineswith a low ionization potential and a large cross section for lasing under rather specific excitation conditions. However.nfot for many gases these systems are specialized, offer lowionization. Xenon satisfies these requirements and does notparticipate in the chemistry of the discharge. Alternatively, atof volumetric or

p c e the imizatiyon cne dsu ied beternately, a power scaling. The results were of interest primarily forhigher pressures tthe spectroscopic information that they provided. There issources such as high-energy electron beams [60].

Gasan approximate correlation between the energy of the laserG transition and the electron-impact dissociation energy of the

cooling, precooling, or gas flow and expansion methods. Thesegas. This means that at longer wavelengths the gas is verymethods will be discussed further below. In the case of the easily dissociated. Eventually, for rotational transitions it is

carbon monoxide laser, cryogenic cooling is required for more appropriate to use optical pumping.sealed-off, low-vibrational transition lasing. The low gas tem- The first proposal for molecular gas-infrared lasers appearsperature reduces the competition of vibrational-translational to be that of Polanyi [67], and the discovery of laser action in acollisions with vibrational-vibrational energy exchange. pulsed gas discharge (in nitrogen) was reported by Mathias and

Discharges in complex molecular gas mixtures had been Parker in 1963 [68]. Fast energy-transfer processes betweenstudied for many years and empirical observations on the phe- N,, CO, and CO 2 had been appreciated for many years by thenomenology had been made on electronegative gases [611 for shock tube community. There is an interesting discussion in theexample. the ionization forms in oxygen-rich discharges and Proceedings of the 12th Solvav Institute Chemistry Conference.the onset of constrictions [621 in gas mixtures. However. this where Hornig et al. [69] discuss the close coupling betweenbody of work was not generally useful in laser development. vibrationally excited nitrogen and CO or CO2 . In fact. theFirst of all. the investigators did not work with gas mixtures very fast relaxation of CO, by water vapor is also mentioned.of particular interest to lasers: they rarely used flowing gases Continuous wave operation (at I mW) in CO, was first

II

announced by Patel (70]. with excellent descriptions of the TABLE IV

lasing mechanism and transitions involved. Improvements in SIMILARtY REL.tAONS FOR MOLECULAR LASER A.,O ATOMICpower output came with the addition of nitrogen (initiall as a LASER DISCHARGES (AFTER KONYU'KHOV. 1741). THE TABLE

INDICAtES THE DISCHARGE RESPONSE To CHANGES i1 TUBE

mixing laser). The nitrogen N_, ( = 1) level is easily excited RADIUS AND GAS PRESSURE THAT KEEP PR CONSTAT,via a temporary negative ion resonance, and it rapidly transfers Parameters Molecular Atomic Laser

energy to the CO, (001) asymmetric stretch vibration. The Laser Discharge

addition of helium to the mixture gave further improvements Discharge low-currentdensitv I

(now at levels up to 100-kW CW). The helium assists in (1) Dimensions R,,R-=a Assume

maintaining low gas temperature and also increases the re- L, =L:a.laxation rate of the lower level. Scaling of the laser power I: Gas temperature Tg2 =Tg, RA =R2 a

has occurred in gas density, volume, and energy loading. The 2G e aoAssumean

related problems in discharge physics were primarily to avoid (3) Electron temperature Te_,=TeI Te2=Te!

the glow-to-arc transition and to provide sufficient gas cooling. (4) Gas pressure P-=P a P.=P.a'(5) Gas number densitv N:=Nla N-=Nja

(6) Electron densttv Ne-=Ne, a Ne-=Ne a-

(7) Current density J'=Jta J,=j=a-"C. Low-Energy Lasers (8) Current 1=-11 a [,=I!

The scaling laws for positive-column-dominated gas dis- 9) Electric field E-E at0) Power Input leneth P-=P P-=P a

charges at low pressures and at low input power were de- (lla) Gain coetficient X =x, a IHgher current

veloped in the 1920's and 1930's. They were assembled in (I lb) - x 2 =X laser dischareesSI '2a) Saturation flux densttN S'=S1 a in'ol'e

the landmark book by von Engel and Steenbeck in 1934 [711 ,-2b) - P=P..a forbidden

and later reviewed by Francis [72]. While these similarity (13) Power out .olume PL-=PI_ ',milarII.

laws appeared to be useful in application to helium-neon procsse,, ,uchlaser discharges. the higher power densities of molecular as %tep

excitation and

gas lasers necessitated revisions. Abrams and Bridges [731 Ionization, andupdated these relationships for laser discharges (cf. Table IV), recomrtration

using the revisions provided by Konyukhov [741. These allowfor the high-power densities of the laser discharge and the The Schottkv diffusion theorv (t thc psittxe column. at-consequent gas heating. Konyukhov noted that if molecular gas though often quoted with respect to molecular gas lasers, doesdischarges of different diameter have the same gas temperature not apply. The Schottky model - otageand electron temperature. they obey siiart rules (not no pl.TeShtk. oe predicts a discharge voltg

esimilarity that is independent of current. While nonlinear terms in ion-quite the same as atomic gases) with consequent relations ization or loss will give a falling characteristic D. Schuocker ctthat apply to laser gain and efficiency. At the higher gas a. [75] showed that for narrow-bore helium-neon dischargespressures appropriate to many waveguide lasers. the transition

these effects are insufficient to describe the experimentallinewidth becomes proportional to pressure rather than being observations. The inclusion of a wall sheath of reasonable

Doppler broadened. Then the gain is independent of density extensions shown o model the o r resuleextension was shown to model the observed results. The

and tube diameter. These discharge and laser similarity laws sheath thickness increases with electron temperature and withare summarized in Table IV, where relations ( IIa) and (12a) decreasing charge density. Therefore the effects of the finiterefer to a constant linewidth that is Doppler broadened, wall sheath are more important at lower discharge currentsrelations (1 lb) and (12b) refer to linewidths that are pressure when the discharge plasma volume is actually constricted

broadened. The results show that the waveguide or narrow- by the shatee These olusis whle orictedbv the sheath extension. These conclusions. while originall

diameter lasers scale as power per unit length. and not aspower per unit volume. This result occurs, because it is the derived for rare gas lasers, also apply to molecular gases.

former th:. ntrols gas temperature in waveguide lasers. The For low-pressure larger radius discharges. the principalformr t: nrol gastemeraurein avegidelasrs.The cooling is by thermal diffusion to the boundaries; that is.

result alsk, . csages the importance of gas cooling in larger

devices. div(k grad T) + Q = 0. (16)In many waveguide lasers (atomic and molecular) it is

extremely difficult to obtain a stable dc discharge without The profile of the heating term Q is controlled by the electronadditional RF excitation. As demonstrated long ago. two density profile. To first order, the Bessel function may beconditions must be satisfied to obtain a stable discharge: the approximated by a parabolic function (Q = Q,,j1 - 5r' 'R`t2 .ballast resistor (or its diode impedance equivalent) must be where (2,) is the on-axis value. The temperature differencegreater than the absolute value of the differential resistance between the discharge axis and tube wall is then given byR of the discharge, and the ballast also must be lower 3EI1 - (1)than ( '1,C!?), where L is the inductance of the discharge AT- (17)and C is the capacitance between anode and cathode. Many SPkmixdischarges in capillaries or with limited cathode area have where A ..,, is the species-averaged thermal conductivity. Erapidly changing V - I curves at low currents, and it is difficult is the discharge electric field. I is the discharge current. andto run a stable discharge. Such circumstances led to adoption '; is the laser efficiency. The difference in gas temperature isof RF excitation, independent of tube radius. For a CO_:N,:He. 1:2:3 mixture.

1I

i I that the impedance of the load as seen by the RF source isthe complex conjugate of the impedance of the generator asseen by the load. An analysis relevant to waveguide lasers

100has been provided by Moghbeli et al. [77]. and while it was

\ I derived with CO, waveguide lasers in mind. it has reasonably

E so •_o generic applicability. There are two similar principal RF exci-SFAST FLOW- tation approaches. The first uses metal electrodes sandwiched

L LASERS between ceramic pieces to provide a hollow waveguide. It

"" Lis found that more efficient operation is obtained when theoi- \0-.J, excitation frequency is high where the sheath fields are lower.

LUSHowever. at higher frequencies and Ionzer waveguide lengthsLU I ._\ _ \the RF wavelength is comparable to the discharge length, and

10.h ~nonuniform fields cause nonuniform discharges. This effect is

A ,overcome by using distributed shunt inductors. The inductors5 provide a negative admittance to compensate for the variation

in phase angle of the transmission-line reflection coefficientalong the length of the laser. The inductors are of equalvalue except the end inductors, which have twice the value

2 of the others. The second approach is the parallel-resonantS2 t to 20 30 distributed inductance developed bv Newman et al. [78]. The

-EECTRODE SPACINGiTUBE DIAMETER miniceramic pieces are sandwiched between the electrodes to

Fig. 11. Electrical input power density in diffusion-cooled and tast-flow form a relatively high capacitance structure. Many discretelasers. The power densttv for alpha-to-gamma discharge-mode transition is inductances are used to approximate uniformly distributedshown for diffusion-cooled lasers at the optimum excitation frequency (after1761). One wishes to operate at high frequencies in the alpha-mode at power

densities below the transition line to maintain efficient and uniform excitation, provide an LC resonance at the excitation frequency. Thissecond method is becoming more important, because it permitspower scaling by increasing the area of the electrode or by use

k , 10 3 W/cm/K. The maximum temperature difference of coupled discharges to give phased-array structures.tolerable (-- 300 K) translates to a maximum laser output

of 50 W/m at an efficiency of 20%. Several methods have D. High-Energy Lasersbeen used to overcome this temperature limitation. Hall et al.[76] have readdressed the waveguide laser and pointed out Scaling to power levels of more than a few hundred watts re-

the importance of area scaling. For higher power lasers this quires convective cooling of the laser. The mass flow required

limitation is overcome by high velocity flow to provide heat ii. to limit the temperature rise to -\T K. is given by

rejection (see below). lt P(I - HIf one considers a uniform discharge between a cooled metal CAT

plate of width ic. length L. and spacing ,4. then the total The CO, 1:2:3 mixture corresponds to an equivalent 18.7laser power extractable from the volume is P, = fP,(wi/d)L. gm/mol and a specific heat. ip = 26 J, mol:K. The gas densitywhere f is a geometric factor allowing for two-wall cooling, is •.4 x l0-4rnict 3 . These figures translate into -b 1 )-43

and PL is the specific power for a square cross-section gain Ibis/kW. For powers greater than a kilowatt. this flow is

medium (defined by Hall et at. as approximately 80 W/m). difficult to achieve at low pressures and so gives an incentiveConsequently, the power of a waveguide laser should scale to consider pressure scaling. While the CO, laser again will beasP = f(PL/d).4. Subject to the requirement of no decrease considered, the concepts apply to other molecular gas systems.in energy deposition efficiency, this suggests that for a chosen The input power is given by P = jE = q,.iiE., whered the lasers will scale with electrode area .1. Experiments it. is the electron drift velocity. To achieve high efficiency.validating these ideas have recently been reported [761. A an (E/.V) is selected and the fractional ionization is kept376-mm-long CO, discharge excited at 125 MHz gave power constant. Therefore P = q(nE/.\N.tI.. V or P , N 2 ifextraction scaling as 20 kW/m 2 in the one-dimensional cooling the conditions are met. At low pressures. E/N must he highregime. The scaling of the input power density versus the enough to provide an ionization frequency that balances theelectrode spacing (for waveguides) or versus the tube diameter losses due to diffusion. recombination. and attachment. At highfor a free-space cavity mode laser is illustrated in Fig. 11. Fast pressures the ionization can be provided by electron beams.flow lasers are really only economical for higher power lasers which then permit tuning of the E/N of the pumping dischargeand so have been given a power densit` threshold. to obtain maximum efficiency. This approach benefited from

The importance of efficient coupling of RF energy to the the energy-deposit ion studies performed many years earlierlaser discharges must be emphasized. RF matching networks by Berger and Seltzer [791. Pumping efficiency curves haveare used in RF plasma reactors. and a similar but distributed been provided bv various authors 180].matching is required for the RF-excited lasers. It is necessary Eventually. the CO2 electrical laser becomes a victim of itsthat the matching condition be at least approximated. such high efficiency. This occurs when the relaxation of the lower

13

laser level is fast enough and provides sufficient energy to electron attachment and dissociation were invaluable. Theremodify the flowing-gas properties. The effect has been named are several methods to overcome the adverse effccts of dis-"'mode-medium interaction." sociation of CO2 on laser operation. In sealed systems it is

The RF discharge was investigated by Levitskii [81], who possible to provide the required ionization level with reducedclassified two principal modes: a so-called alpha discharge dissociation levels by adding a small percentage of xenon toat low-current discharges with distinct electrode sheaths and the traditional gas mixture. The use of xenon is attractive, sincevolume (Townsend) ionization: and a second, gamma-mode. it does not react chemically with the other species and has aat higher current densities where the discharge is maintained large ionization cross section and a low ionization potential.by secondary electron emission caused by ion bombardment In principle, one could use metal additives, but these lead toof the electrodes. (A contemporary approach [82] would difficulties with control of the vapor pressure and condensationintroduce the influences of excitation frequency, collisionless on optical components. Alternatively. CO or H, may be addedenergy transfer, and wave-riding.) These excitation mecha- to the mixture to enhance detachment of negative ions. Anisms can cause spatial discharge inhomogeneities. Optimum third method to reduce the dissociation of CO- is to includeexcitation frequencies which vary inversely with the electrode heterogeneous catalysts in the system. Materials that have beenspacing are found to be typically 120 MHz at d = 2.75 used include hopcalite. platinum, and gold [91]. Browne andmm. Too low a frequency gives the -,-mode, whose fast Smith [92] demonstrated that while a platinum catalyst in aelectrons cause excessive dissociation for CO2 -based lasers low-pressure CW CO_-N,-He-Xe laser resulted in a longand probably most other molecular infrared lasers as well. lifetime, the further addition of hydrogen or deuterium wasThe transfer of the models and analyses developed for RF deleterious. In other words, the benefits of reassociation hadprocessing discharges to RF-excited lasers has not been made. already been obtained and the hydrogen then only decreasedHowever, empirical advances in coupling several waveguide the population of the upper laser level.lasers by means of ridge coupled areas, driven by one RF In flowing lasers and at high pressures it is not economicalsource. point the way to new discharge configurations suitable to use xenon. and then one does use additives such as H-0. 02,for phased-array operation. and H2 . Shields et al. [93] constructed a comprehensive kinetic

Scaling to high pressures was also achieved by pulsed high- model of negative ion formation in such discharges and alsopressure discharges. These relate to the early work on impulse compared the low-pressure results of Prince and Garscaddenbreakdown studies by the school of Meek and Craggs [831. To [94]. At high pressures, CO,. is the dominant negative ion.avoid extremely high voltages, Dumanchin and Rocca-Serra while at low pressure the negative ions of the oxides of[84] first excited high-pressure (420 torr) CO2 in transverse nitrogen and the oxygen negative ions are more important.geometry. Beaulieu [85] used 17-kV short-10's of ns pulses to An analysis by Hokazono and Fujimoto [95] used 175 kineticexcite an atmospheric pressure discharge where the distributed equations for the plasma chemistry in a TEA laser and obtainedexcitation was achieved by using a row of sharp pins as a a useful engineering rule that the fractional decomposition wascathode. Many configurations and preionization schemes have almost proportional to the energy deposited/pulse. Calculationsbeen investigated, including X-rays. double pulses, photoion- of the effects of additive CO. H,, and H20 showed the lastization. and auxiliary discharges. Lin and Levalter [86] defined to be most effective in reforming CO-: however, this criterionthe preionization conditions necessary to achieve uniformity alone is not sufficient, since no account is made of the effectsof the laser discharge. The technology has been reviewed by of the additives on the plasma stability.Wood [87], and more recently by Brown [88].

F. Discharge InstabilitiesE. Discharge Chemistry The problem of arcing at high pressures is an old one.

The detailed analysis of infrared lasers has invoked many as- Several authors [96]-[98] have suggested that the presence ofpects of discharge physics. Calculations of the electron-energy a sufficient density of negative ions enhances arcing in TEAdistribution functions by solving the Boltzmann transport lasers based on CO 2 and other gases or gas mixtures. At high-equation were assisted by the methods established much energy loading the actual arcing often appears to be a combi-earlier by discharge research and electron transport research. nation of an ionization and a thermal instability. The ionizationThese calculations were made for given input cross sections. instability in an infrared molecular gas mixture is. strictlyE/N values, and an assumed gas mixture. Rapid progress speaking, an attachment-ionization-recombination instability.was possible since the codes were essentially established by It occurs in both self-contained and externally sustainedearly work. especially that of Frost and Phelps [89]. These charges. These instabilities have properties similar to the clas-calculations were not self-consistent, in that they did not sical moving striations [97]. Temporal-spatial inhomogeneities"consider dissociation or the resultant chemistry. Furthermore. adversely affect the quality of the laser medium and sometimesthe traditional drift experiments studiously had avoided such provide a trigger mechanism for the onset of constrictionreal-life problems. Nighan et al. [90] were among the first and arcing. Thus the uncontrolled plasma chemistry can leadto show that the influences of plasma-induced chemistry and to degraded long-term performance of low-power lasers andnegative ions were very important for long-life, sealed-off catastrophic failure of high-power lasers. The degradation canoperation and for higher energy-density discharge stability, occur also due to vibrational quenching by the impurities, and.To analyze these phenomena. the quantitative databases on more seriously, to the dissociation of the lasing species.

The attachment instability requires an electron-attachment local heating and expansion of the cathode gas layer againstrate which increases rapidly with electron temperature. The the more dense background gas. A possible explanation isnecessary condition was expressed by Nighan and Wiegand that there is a coupling of the fluid-dynamic Rayleigh-Taylor[971 as instability with the ionization and thermal instabilities of the

ka ka plasma.kk> 1 (19) The carbon monoxide laser is one in which all ofk,k, the degrees of freedom of the molecule play important

where k,, and k, are the attachment and ionization rates, roles in the laser mechanism and efficiency. The laserand the caret notation denotes their logarithmic derivatives depends on the vibrational excitation by anharmonic pumpingwith respect to T,_. The mechanism of the instability can through vibrational-vibrational (V-V) energy transfer. Carbonbe described heuristically as being similar to a striation, monoxide is slightly anharmonic and the energy spacingNeglecting the phase relations, an increase in electron tem- between vibrational levels becomes smaller with an increaseperature will cause an increase in negative ion density, and in the vibrational level. The energy is redistributed by thea decrease in the local electron density. There is therefore process:a decrease in the local plasma conductivity. If the circuitconditions require a constant current, the local electric field CO( I,)- CO(e') -. CO( + 1) + CO( t" - 1) (20)will increase, which further increases the attachment rate andthus leads to instability. The sufficient conditions for instability low translational-rotational temperature. The CO(r) is excitedinvolve the electron-ion and ion-ion recombination rates and

by electron collisions up to approximately v=7. The highermost importantly, the detachment rate k,t. As the currentdensity in the discharge increases. kj must be made larger vibrational levels are then populated in a non-Boltzmann distri-doensityuinrthe dschargity.The increasdes. ye mt be msadela bution bv anharmonic pumping. In carbon monoxide at lowerto ensure stability. The amplitude cycle of the insta bility s, the V-V energy transfer rates are orders of magnitudecan be established by numerically solving the continuity faster than the vibrational-translational (V-T) energy-transferequations for the three charged species. Examples of such rates. Trainor et at. [1021 showed that when there is strongcalculations are give by Long et al. [991 for a C02 :N2 :He excitation of the lower vibrational levels and decoupling cf thelaser mix and the effects Of 02 and CO. These results were vibrational manifold from the translational mode. the upperconfirmed by Beverly [1001. In principle, the influences of thedischarge modulations on the laser output can be calculated. levels o the vibratioltmanifd re ov rp te withThe neutral plasma chemistrv occurs on a longer time scale respect to the equivalent Boltzmann distribution for the same

vibrational energy content. Due to their higher small signal(Table I1) and hence can be calculated separately to establish i on a er conent Due o their h rasignalthe neutral densities. The negative ion-chemistry calculations gains. almost all CO lasers operate on the P transitions fromvibrational rotational states (V, J)-(V-1, J+l), which permitby Shields et al. [931 showed that small amounts of oxygen i on partial

were especially troublesome to the uniformity of atmospheric tasing s, pauinversions. The lasing occurs in cascadingTEA CO, lasers. The preionization electrons are lost rapidly transitions, because when one transition reaches threshold.to mleclaroxyen v treebodyattchmnt.Thi efectit depletes the lower level of the transition above it andto molecular oxygen by three-body attachment. This effectgreatly reduces the permissible time window between the populates the upper level of the transition below it. Thesegreianiyatidncpulsehandemainsdischarge.Thusothebcollision effects lead to operation on a series of transitions and apreionization pulse and main discharge. Thus the collisio naliher thanprocesses and plasma chemistry actually determine the require- 6gh qhave een o verall effigh oy h ig nments and sophistication of the circuit hardware needed for p0%shave been rorted v13] Thvigh opt couplinreproducible and efficient lasing. possible with CO produces very obvious optogalvanic effectswhich illustrate the need for inclusion of collisions of theThe interest in high-pressure lasers led gas discharge re-search into parameter regimes that were entirely new. The seodkninm elgthsaerPirtoheC lsrthere had been few studies on cryogenically cooled discharges.e-beam-sustained discharge may satisfy the quoted conditions Terqieeto o rnltoa eprtrsldt

The requirement of low translational temperatures led to afor stability: however, the discharges will often still form revival of interest in discharges with cryogenic wall coolingarcs. The theory previously referenced is for a homogeneousunbounded plasma: there is no account of the phenomena or in supersonic expanding flows. Reilly [104] identifiesat the boundaries. Arc formation in such e-beam-sustained the latter as a class where the discharge has volumetricplasmas was studied by Douglas- Hamilton [1011. A major temperature control. In this case, there is a one-dimensional

plasas as tuded b Dogla-Hailto [11].A mjorpower balance of flow heating versus the discharge powerfinding was the identification and measurement of discharge pus cstreamers (usually but not always initiated at the cathode) input:

which propagate into the discharge and often lead to shorting dT Eof the gap. The streamer growth rate was determined to be rE-p-(- r? i . 1 (2exponential. Based on the framing camera measurements andshadowgraphs. a model was devised wherein the streamer is where r and cp are the gas density and specific heat. re-represented as a quasi-thermal arc concentrating the current spectively. U is the gas velocity. ,Tq/dz is the rate offlow and enhancing the electric field at its leading tip. The temperature increase with distance, and v, is the electroninitiation of the streamer at the cathode is caused by a local thermal speed. The important point is that as the gas densityinhomogeneity in cathode emission, which in turn results in the increases while the power input increases at a fixed i E/N'.

15

the capability of the gas to absorb the power also increases not come easily, since a large database is required to modelwith g-as density, and design such lasers. The modeling requires information on

Shapiro [105] has analyzed the situation of heat input to a large number of processes: (a) electron-impact excitationgas flow in a channel of constant-area cross section and and ionization: (b) ion impact. especially dissociative chargeshowed that both subsonic and supersonic flows will tend transfer in polyatomic species: (c) V-T relaxation rates. (d)to Mach number = 1. Ravleigh line calculations for energv V-V-T near resonance exchange collisions: (e) spontaneousinput approximately equal to the translational flow energy decay rates, including A. = 1 and Au = 2; (0) stimulatedshow that inlet Mach numbers must be less than 0.15 to emission cross sections and line-broadening parameters. andavoid density changes greater than a factor of two. Chok- (g) miscellaneous processes involved in the conditions relateding will occur for inlet Mach numbers of -. 0.4 [1041. to discharge uniformity and stability such as UV ionizationThus in CW and repetitively pulsed devices the average processes, attachment, and detachment rates. Research intopower is limited by the maximum flow velocity: in short- collisional processes has benefited the CO laser-some ofpulse (less than the flow transit time) devices the pulse the recent impressive gains in the CO-laser performance andenergy is limited bv the maximum gas temperature permis- applications are reviewed by Maisenhalder [1101. Similarsible. improvements are considered possible for other laboratory

Gas purity is very important in the carbon monoxide infrared lasers.systems, since it is necessary to avoid any species that It is anticipated that more practical designs for high-pressurequench the vibrational manifold. The anharmonicity of CO operation of gas discharges will evolve due to the interest inmeans that there is usually a resonance energy transfer to tunable lasers and in very short pulse lasers. The necessarythe impurity from some levels of the CO. Some species operating pressure for continuous frequency tuning is about 5such as the pentacarbonvls cause extremely fast V-T atm for a mixture of CO, isotopes, and 10 a"tr for a singlerates. isotope 11111. There are many applications in spectroscopy.

As one moves into the submillimeter region the discharges short-pulse phenomena. ranging. optical pumping. and relax-required become longer and larger in cross section. These ation processes to encourage this development. It appears thatoscillators have been used as reference frequencies and for distributed (meander-line or slow-wave applicator) microwaveplasma interferometrv. The difficulties, such as toxicity, of excitation may play a role in discharge excitation, in additionworking with HCN. CS,. etc.. have restricted their uses thus to the waveguide devices utilized so far.far. (There are also considerable problems with sealed-off Gas-phase approaches to the near-millimeter wave-sourceoperation that have received little attention.) As an interesting problem have been reviewed by De Lucia ei al. [1121.example of such discharges. Pollack et al. [1061 ran a I-A The gas-phase approach is still attractive for higher powerwater-cooled discharge. 3 m in length with a 7.5-cm ID. The devices. It should be possible to revisit these devices withwater vapor laser displayed at least 65 pulsed lines, ranging a better understanding of the plasma conditions requiredfrom 47 to 79 jim and 9 CW lines, for laser operation. Also. new excitation techniques and

The extension of discharge excitation techniques to far- configurations offer enhanced performance. Several of theinfrared transitions that involve pure rotational transitions ideas used in plasma processing. such as extended multipolebecomes one of limiting returns. The quantum efficiency plasma confinement and electron-cyclotron resonance plasmas.becomes the ratio of the rotational quantum to at least one appear attractive and may serve to change such lasers fromvibrational quantum. The discharge excitation also introduces laboratory devices to established sources. Electrical-chemicalthe problems of plasma chemistry and gas cooling. There- lasers operating in the infrared, similar to those discussedfore it is much easier, if appropriate laser pumping sources in Section III. are anticipated. Materials and catalystexist, to use optical excitation of the far-infrared transition. developments may not only regenerate the lasing species.The overall efficiencies are about the same since, in either but also deliver vibrationally excited molecules [1131.case, one is still using a discharge to provide the vibrational [114].laser. Other configurations of great interest will be those that

Discharge lasers that are electrical-chemical in nature can are modular [78], [1151 and compatible with phased-arrayaccommodate more intense and higher E/N discharges. Such operation. Hart et al. [1151 have demonstrated mode lockingmethods include microwave excitation [107], [1081 or e-beam and efficiency in CO2 using a staggered waveguide approachstabilized discharges [109]. In these systems there are proba- (68 W in a 6-element array). The machining techniques andbly contributions to population inversion from recombination optical gratings for longer wavelength operation should bereactions into excited products. easier and less expensive than those required below 12 jim.

Some remarkable claims of a very large number (61) ofcoupled discharge lasers have been made 11161 at 10.6 jim.

G. Future Prospects

Initially. much of the research tended to concentrate onthe scaling of infrared lasers: however, more recently there ACKNOWLEDGMENThas been a new emphasis on long lifetime and reliability. The authors thank their colleagues for comments and refer-This trend is expected to continue and expand from the CO, ences. especially E. A. McLean for suggestions regarding theand CO systems to other infrared laseis. Improvements do short--wavelength laser portion of the text.

I 0

REFERENCES 597-599. Feb 1989.[271 Y.-P. Kim. M. Ohara. and T. Suzuki. "Theoretical evaluation ot electron

II] R. C. Tolman. "Duration ot molecules in upper quantum states." Phvv. beam excited vacuum ultraviolet F- lasers. J. Appl. Phi's,. sal. 59. pp.Rev.. sol 23. pp. 693-709. 1924. 1815-1818. Mar. 1986.

[2] R. Ladenburg. "'Dispersion in electrically excited vases." Rev. Mod. 1281 M. Ohwa and M. Obara. 'Theoretical evaluation of high efficiencyPhvs.. vol. 5. pp. 243-256. 1933 oand references theremi. operation ot discharge-pumped vacuum ultraviolet F. lasers." Appl

13] V, A. Fabrikant. "The emission mechanism of gas discharges." Trans. Phvs. Lett.. vol. 51. pp. 958-960. Sept. 1987.Electrotekhn. Inst., vol. 41. p. 254. 1941) (quoted in C. G. B. Garrett. [291 S. J. Beneroje. G.-Y. Yin. C. P. J. Party. J. F. Young. and S. E. Harris.Gas Lasers. New York: McGraw-Hill. 1967. chap. 1). " 116 nm H-." Phys. Rev. Lett.. vol. h6. pp. 3136-3139. June 1991

[4] B. A. Lenrgyel. Introduction to Laser Phsvics. New York: Wiley. 1966. [301 R. Sauerbrey and H. Langhoff. "Excimer ions as possible candidatesp. 184, for VUV and XUV lasers." IEEEJ. Quant. Electron.. vol. QE-21, pp

[51 R. W Waynant and R. C. Elton. "Review of short wavelength laser 179-181. Mar. 1985,research." Proc. IEEE. vol. h4. pp. 1059-1092. July. 1976. [311 C. A. Brau. in Excimer Lasers." C. K. Rhodes. Eds. Berlin: Springer.

[6] D. L. Matthews et al.. "Demonstration of a soft-X-ray amplifier." Phvs. 1979. pp. ,7-134.

Rev. Lett.. vol. 54. p. 110. 1985. 132] R. S. Taylor and K. E. Leopold. "Ultralong optical-pulse corona preton-[7] Proc. Opt. Soc. Amer. on Shortwavelength Coherent Radiation: Gener- ized XeCI laser." J. Appl. Phvs., vol. 65. pp. 22-27. 1989

ation and Applications. R. W. Falcone and 1. Kirz. Eds. Washington. [331 S. Watanabe and A. Endoh. "'Wide aperture self-sustained discharge KrFDC: Opt. Soc. Amer., 19)88. and XeCI lasers." Appl. Phvs. Lett.. vol. 41. pp. 79Q-801. 1982.

[8] B. J. MacGowan et al.. "Progress toward a 44 A X-ray laser." in Proc. [341 G J. Hirst. V Rivano. and C. E. Webb. "Spatially resolved gain

Opt. Soc. Amer. on Shorrvaveleneth Coherent Radiation: Generation measurements in a discharge-pumped KrF laser amplifier." I Appl.and Applications. R. W Falcone and J. Kirz. Eds. Washington. DC: Phs's.. vol. 61. pp. 2438-2444. 1987.

Opt. Soc. Amer.. 1988. pp. 2-11) [35] C. H. Fisher et al.. -High efficiency XeCI laser excitation %% ith magnetic[9] M. D. Rosen etal., "Exploding-foil technique for achieving a soft X-ray switching." Appl. Ph's. Lett.. vol. 48. pp. 15"4-1576. 1986.

laser." Phvs. Rev. Leet.. vol. 54. pp. 1106-109, Jan. 1985. [36] W H. Long. M. J. Plummer. and E. A. Stappaerts. "Efficient discharge

[10] T. N Lee. E A. McLean. and R. C. Elton. "Soft X-ray lasing in neonlike pumping o1 an XeCI laser using a high-voltage prepulse." Appl Pht s

germanium and copper." Phvs. Rev. Lett, vol. 59. pp, 1185-1188. Sept Lett.. vol. 413. pp. 735-737. 1983.

1987. [37] R. S. Taylor and K. Leopold. "Microsecond duration optical pules from

[11] T. N Lee. E. A. McLean. J. A. Stamper. H. R. Griem, and C. K. Manka. a UV-preionized XeCI laser." Appl. Phvs. Lett.. vol. 47, pp. 81-X3. IQS.5.

"ULaser-driven soft X-ray experiments at NRL." Bull. Amer. Phvs. Soc.. [381 K. Midorikassa. M. Obara. and T. Fujioka. "X-ray pretonization ot rare

vol. 33. p. 1920. 1988. gas-halide lasers." IEEEJ. Quant. Electron.. vol. QE-20. pp 198-2105.

112] M D. Rosen. R. A. London. and P. L. Hagelstemn. "The scaling of Ne- 1984.

like X-ray laser schemes to short wavelength." Phss. Fluids. vol. 31, [39] M. R. Osborne. "Preionized electron density and ion deca, measure-

pp. (66--.70, 1988. ments in an X-ray pretonized rare gas fluoride laser." J. Appl. Physs

[13] R. C. Elton. X-Ray Lasers. New York: Academic. 1990. vol. 63. pp. 32-37, 1988.

[141 S. Suckevser. C. H. Skinner, H. Milchberg. C. Keane. and D. Voorhees. [40] T. S. Fahlen. "Efficient quarter-joule KrF laser with corona pretoniza-

"Amplification of stimulated soft X-ray emission in a confined plasma tion." IEEE J Quant. Electron.. vol. OE-15. pp 311-313. 1979

column." Phvs. Rev. Lett.. vol. 55. pp. 1753-1756. Oct. 1985. [41] P. Oesterltn and D. Bastling. "Eureka's excimers." Phvi. World. %ol. 1.

[15] S. Suckeweretal., "Divergence measurements of soft X-ray laser beam,- pp. 43-46. 1990.Phys. Re'. Lett.. vol. 57. pp. 1004-1007. Aug. 1986. [421 T. Hasama. K. Miyazaki. K. Yamada. and T Sato. "50 J discharge

[16] P. L. Hagelstein. "Short wavelength lasers: something new, something pumped XeCI laser." IEEE J. Quant. Electron.. vol. 25. pp 113-120).

old." in Proc. Opt. Soc. Amer. on Shortwavelength Coherent Radia- 1989.

tion: Generation and Apphcations. R. W. Falcone and J. Kirz. Eds. [43] L. F. Champagne. A. J. Dudas. and N. W. Harris. "Current rise-time

Washington. DC: Opt. Soc. Amer.. 1988. pp. 28-35ý limitations of the large volume X-ray preionized discharge-pumped

[17] S. Suckeser. "X-ray related expenments and theory at Princeton." in XeCi laser." J. Appi. Phvs.. vol. 62. pp. 1576-1584. 1987.

Proc. Opt. Soc. Amer. on Shorrwavelength Coherent Radiation: Gener- [44] iL organ. r. N. a lasmas. " Kn nd . L. WhitAen. 'horv of ion-ion recombination in plasmas." Phvs*. Rev. sal A26i. ppation and Applications. R. W. Falcone and J. Kirz. Eds. Washington. l%96.-1703. 1982.DC: Opt. Soc. Amer.. 1988. pp. 36-4h. [451 NI. R. Flannerv and T. P. Yang. "Three-body ion-ion recombination in

[18] H. C. Kapteyn. R. W. Lee. and R. W Falcone. "Observation of a short mercurv-hahde lasers.* Che.- Phys. Lett.. vol. 5h. pp 143-1-.7. 19-Swavelength laser pumped by Auger decay." Phys. Rev. Lett.. vol. 57. [461 J. E. Velazco. . H. Koltz. and D. W. Setser. 'Quenching rate constantspp. 2939-2942. Dec. 1986. for metastable argon, krypton and xenon atoms by fluoride containing

[19] W T. Silf' ast and 0. R. Wood Ii. "Phototonization lasers pumped by molecules, and branching ratios for XeF and KrF" formation." J ('hem.broadband soft X-ray flux from laser-produced plasmas." J. Opt. Soc. Ph's.. - ol. t)5. pp. 3468-34801. 1976.Amer. B. hrol. 4. p. 609--618. Apr. 1987. [47] A. Kvaran. M. J. Shaw. and J. P. Simons. Vibrational relaxation at

1201 M. H. Sher. J. J. Macklin. J. F. Young, and S. E. Harris, "Saturation KrFF and XeCI by rare gases." AppL. Phvs B. vol. 46. pp 95-1012.of the Xe 111 109 nm laser using traveling-wave laser-produced plasma 1988.excitation.' Opt. Lett.. vol. 12. pp. 891-893. Nov. 1987. 1481 W J. Witteman and B. M. H. H. Kleikamp. "On the electron-beam

[211 C. P. J. Barty et aL. "12.8 eV laser in neutral cesium." Ph/s. Rev. Lett.. pumped KrF laser." J. Appl. Phs.. vol. 55. pp. 1299-1307. 1984vol. 61. pp. 2201-2204. Nov. 1988J. C [491 M. Havasht. "Recommended values of transport cross sections." Nagoya

[22] 3. 3. Rocca. 3. 0. Meyer. and G. 3. Collins. "1 W CW Zn ion laser." Instit. Tech.. Rep. No. IPPJ-AM-19. 1981.Appl. Ph/s. Lett.. vol. 43. pp. 37-39, 1983; see also. B. Wemsman et [501 R. E. Olsen. J. R. Peterson. and J. Moseley. "Ion-ion recombination totalal.. "CW silver ion laser with electron beam excitation," IEEEJ. Quant. cross-sections atomic species." J. Chem. Phvs.. vol. 53. pp. 3391-_1'307.Electron.. .ol. OE-24. pp, 1554-1556. 1988. 1970.

[211] 3. Wern.sman. J. J. Rocca. and H. L. Mancini. "CW ultraviolet and 151] R. S. Taylor. 'Pretonization and discharge stability study of lone opticalvisible laser action from ionized silver in an electron beam generated pulse duration UV-pretonized XeCI lasers." Appl. Phys. B. vol. 41. ppplasma." IEEE Photon. Tech. Lett.. vol. 2. pp. 12-14. Jan. 1990. 1-44., 1986.

1241 J J. Rocca. M, C. Marconi. M. Villagran Muniz. and D. C. Beethe. [521 S. Suminda et al.. -"Effect of pretonization uniformttv on a KrF laser.'"Capillary discharge plasmas as extreme ultraviolet laser sources." in J. Appl. Phvs.. vol. 52. pp. 2682-2686. 1981.Proc. Opt. Soc. Amer. on Shortwavelength Coherent Radiation: Gener- [53] S. Watanabe. A. 1. Alcock. K. E. Leopold. and R. S. Taylor. 'Spattall,anon and Applications. R. W. Falcone and J. Kirz, Eds. Washington. resolved gain measurements in UV pretomnized homogeneous dischargeDC: Opt. Soc, Amer.. 1988. pp. 99-105. XeCI and KrF lasers." Appl. Ph/'s. Lett.. vol 38. pp. 3-h. 1981

[25] A B. Petersen. 'CW ion laser operation in argon. krypton and xenon 154] A. B. Treshchalov and V E. Peet. "Spatial-time dynamics of theat wavelengths down to 232 nm." in Proe. Ann. Meet. IEEE Lasers discharge pumping and lasing in a XeCI excimer laser." IEEEJ Quantand Electro-Optic.s Soc (LEOS '89) (Orlando FL). Oct. 1989: see also. Electron.. vol. 24. pp. 169-176. 1988."Short wasvelength CW rare gas ion lasers." in Proc. Eni. Found. Conf [55] M. J. Kushner and A. L. Pindroh. -Discharge constriction. photodetach-on Future Prospects and Appl. for UV and 1,'V Laserv iSanta Barbara. ment. and ionization instabilities in E-beam sustained discharge excimerCA). Feb 0(1) lasers." .1 Appl. Phys.s.. vol. 60. pp. (X04-914. 1986.

[26] K Yamada. K. Mivazaki. T Hasama. and T Sato. "High power [561 NI. I Kushner. A. L. Pindroh. C. H. Fisher. T A. Znotins. and Idischarge-r-umped F, molecular laser." Appl. Ph's. Lett. vol. 54. pp. 1 Ewmg. "Afultidimensional modeling of transverse avalanche laser

17

discharges," J. Appl. Ph/s., vol. 57, pp. 2 , 6-2423. 1985. slow electrons in He. At. Kr. and Xe using transport coefficients." Phis.[57] M. 1. Kushner. "Discharge instabilities initiated by nonuniform laser Revt. vol. 127. pp. 1621-1626. 1962: see also. Phvs. Rev.. %Al. 136. pp,

extraction in E-Beam sustained discharge KrF lasers." J. Appl. Phys., 1538-1546. 1964.vol. 62. pp. 101-107. 1987. [901 W 1J. Wiegand and W. L. Nighan. 'Plasma chemistrv of CO,-N_:-He

[58] Handbook of Mlolecular Lasers, P. K. Cheo, Ed. New York: Marcel discharges. "Appi. Phvs. Lett.. vol. 11. pp. 583-586. 1973.Dekker. 1987. [91] J. A. Macken. S. K. Yagnik. and M. A. Samis. "CO_ laser pertormance

[591 R. E. Center. "High-power. efficient electrically-excited CO lasers." in with a distributed gold catalyst." IEEE J. Quant. Electron.. vol. 25. pp.Laser Handbook. vol. 3. M. L. Stitch. Ed. Amsterdam: North-Holland. 1695-1703. 1989.1979. [92] P. G. Browne and A. L. S. Smith. "Long-h.led CO- lasers Aith

[601 J. D. Daugherty. "'Electron-beam ionized lasers." in Principles of Laser distnbuted heterogeneous catalysis." J. Ph's. D: Appl. Phv . vol.Plasmas. G. Beketi. Ed. New York: Wiley. 1976. pp. 2464-2470. 1974.

[61] K. G. Emeleus and G. A. Woolsev. Discharges in Electronegaive Gases. [93] H. Shields. A. L. S. Smith. and B. Norris. "Negative ion effects in TEANew York: Barnes and Noble, 1970. CO, Lasers," J. Phys. D: Appl. Phvs.. vol. 9. pp. 1587-1603. 1976.

[62] C. Kenty. "'Dissociative recombination and the constriction of the [94] J. F. Prince and A. G-irscadden. 'Negative ion species in CO,-N:-Hepositive column." in Ionization Phenomena in Gases. vol. 1. H. Maecker. discharges." Appl. Phys. Lett.. vol. 27. pp. 13-15. 1975.Ed. Amsterdam: North-Holland. 1962, pp. 356-366. [95] H. Hokazono and H. Fujimoto. "Theoretical analysis of the CO:

[63] L. J. Kieffer. 'A compilation of electron collision cross section data for molecule decomposition and contaminants yield in transversely ex-modeling gas discharge lasers." Joint Instit. Lab. Astrophys., Boulder. cited atmospheric CO: laser discharge." J. Appl. Phwvs.. %ol. 62. pp.CO. Rep. No. 13, 1973 (revised 1978). 1585-1594, 1987.

[64] 1. S. E. Townsend. Electrons in Gases. London: Hutchinsons. 1947, [96] D. H. Douglas-Hamilton and S. A. Mani. "'An electron-attachment[65] G. J. Schulz, "Vibrational excitation of N2, CO and H 2 by electron plasma instability." Appl. Phys. Lett.. vol. 23. pp. 508-510. 1973

impact." Phws. Rev. A. vol. 135A. pp. 988-995. 1964. [97] W. L. Nighan. in Principles of Laser Plasmas. G. Bekefi. Ed. New[661 M.A. Pollack. "Molecular gas lasers," in Handbook of Lasers. R. J York: Wiley. 1976: see also, W. L. Nighan and W. J. Wiegand.

Pressley. Ed. Cleveland. OH: CRC Press. 1971. pp. 298-349. "Influence of negative-ion processes on steady-state properties and[67] J. C. Polanyi. "Proposal for an IR maser dependent on vibrational striations in molecular gas discharges.' Phvs. Rev. .4, vol. AlO. pp

excitation." J. Chem. Phys.. vol. 34. pp. 347-349. 1961. Q22-945. 1974[68] L. E. S. Mathias and J. T. Parker. "Stimulated emission in the band [98] A. S. Kovalev. I. G. Peersiantsev. V. D. Pis'mennvi. and A. T Raakhi-

spectrum of nitrogen.' Appl. Phys. Lett., vol. 3, pp. 16-18. 1963. mov, 'Instabilities of high-pressure glow discharges." in Molecular ;as[69] "Energy transfer in gases." in Proc. 12th Chem. Conf.. R. Stoops, Ed. Lasers, E. P. Velikov. Ed. Moscow: MIR. 1981. chap. 3.

New York: Interscience. 1963. [99] W. H. Long. W. F. Bailey. D. R. Pond. and A. Garscadden. presented at[70] C. K. N. Patel. "Continuous-wave laser action on vibrational-rotational the IEEE st Int. ConFi. on Plasma Sci.. Knoxville. TN. 1974 published

transitions of CO,," Phys. Rev. Lett.. vol. 336A. pp. 1187-1193. 1964. ts in W. J.Con. an Plasma .hsheoj711 A. von Engel and M. Steenbeck. Elektrrsche Gasetladungen (Ihre Phvsik excerpts i W. 1. Wiegand and W. L. Nighan. 'Plasma chemistr ot

und Technik. I. 11). Berlin: Springer-Verlag. 1934. CO 2-N 2-He discharges." Appl. Phvs. Left.. vol. II. pp. 583-586. 114'3.

[72] G. Francis. "The glow discharge at low pressure." in Encyclopedia of and A. Garscadden. Gaseous Electro.cs. vol. 1 New York: Academic.

Physics. vol. 22. Berlin: Springer-Verlag, 1956. 1978, chap. 2.2.[73] R. L. Abrams. "Waveguide gas lasers." in Laser Handbook. vol. 3. M. [100] R. E. Beverly. "Ion aging effects on the dtssociative-attachment in-

L. Stitch. Ed. Amsterdam: North-Holland. 1979. stability in CO2 lasers." Opt. Quant,. Electron., vol, 14, pp. 51)1-513,

1741 V K. Konyukhov. "Similar gas discharges for CO, lasers." Soviet 1982.

PhYs.-Tech. Ph's.. vol. 15. pp. 1283-1287. 1971. [101] D. H. Douglas-Hamilton. "Investigation of the production of high[751 D. Schuocker. W. Reif, and R. Erlacher. "Properties and current-voltage density uniform plasmas." U.S. Air Force Wright Aeronautical Labs.

characteristics of discharges in waveguide gas lasers," Appl. Phvs.. vol. Rep. 80-2087. AD No. A095508, 1980.14. pp. 277-282. 1977. [102] C. E. Treanor. J. W. Rich. and R. G. Rehm. "Vibrational relaxation of

[761 D. R. Hall and H. J. Baker. "RF excitation of diffusion cooled and fast anharmonic oscillators with exchange-dominated collisions." J. Chem.axial flow lasers." in Proc. 7th Int. Syvmp. on Gas Flow and Chemical Phvs., vol. 48, p. 1728, 1968: see also, G. E. Caledonia and R. E. Center.Lasers. SPIE. vol. 1031. pp. N"-67. 1988. "Vibrational distribution functions in anharmonic oscillators." J. Chem.

[77] F. Moghbeli. D. He. G. Allcock. and D. R. Hall. "Impedance matching Ph/s.. vol. 55, pp. 552-528. 1971.in radio-frequency discharge excited waveguide lasers."J. Phys. E. vol. 11031 M. M. Mann. D. K. Rice. and R. G. Egucht. "An experimental17. pp. 1159-1164. 1984. investigation of high energy CO lasers." IEEEJ. Quant. Electron . vol

[78] L. A. Newman and R. A. Hart. "Recent advances in sealed-off CO, OE-10, pp. 082--683. 1974.lasers." Laser Focus. vol. 23. pp. 80-86. 1987. [104] J. P. Reilly. "High-power electric discharge lasers." Astronaut. Aero-

[79] J. J. Berger and S. M. Seltzer. Nat. Acad. Sci.. Washington. DC. NAS- naut., pp. 52-63. Mar. 1975.NRC 113 (Nucl. Sci. Series Rep. No. 10). [105] A. H. Shapiro. The Dynamics and Thermodynamics of Compressible

[801 D. H. Douglas Hamilton and R. S, Lowder, "Carbon dioxide electric Flow. New York: Ronald Press. 1953.discharge laser kinetics handbook." Avco Everett Res. Lab.. Apr. 1975: [106] M. A. Pollack. T. J. Bridges, and W. J. Tomlinson. "Competitive andsee also. C. J. Elliott. 0. P. Judd. A. M. Lockett. and S. D. Rockwood. cascade coupling between transitions in the CW water vapor laser.""-Electron transport coefficients and vibrational excitation rates for Appl. Phss. Lett.. vol. 10. pp. 253-255. 1967.electrically excited CO, gas lasers." Los Alamos Nat. Lab.. Los Alamos. [107] D. S. Vasyutinskii. V. A. Kruzhalov. T. M. Perchanok. D. K. Terekin.NM. Rep.-5562-MS. 1974. and S. A. Fridrikhov. "Pulsed microwave discharge as a pump for the

[81] S. M. Levitskii, "An investigation of the breakdown potential of a high- CO, laser." Soy. Ph/s.-Tech. Phys., vol. 23. pp. 189-194. 1978frequency plasma in the frequency and pressure transition regions." Sos'. [1081 K. G. Handy and J. E. Brandehik. "Laser generation by pulsed 2.45 GHzPhvs.-Tech. Ph/s.. vol. 2. pp. 887-891. 1958. microwave excitation." J. Appl. Ph/s.. vol. 49. pp. 3753-3756. 1978

[82] J. P Bouet and P. Belenguer. in Non-Equilibrium Processes in Partially [1109] S. R. Byron. L. Y. Nelson. and G. 1. Mullanev. 'HF and DF lasersIonized Gases, M. Capitelli and J. N. Bardsley. Eds. New York: by direct electrical discharge excitation." AppL. Phvs. Lett., vol. 23. pp.Plenum. 1990. 565-567. 1973.

[83] 1. M. Meek and J. D Craggs. Electrical Breakdown of Gases. New [110] F Maisenhalder. "High power CO lasers and their application potential."York: Wiley. 1978. in Proc. 7th Int. Svmp. on Gas Flow and Chem. Lasers. SPIE. vol. 1031.

[84] R. Dumanchin and J. Rocca-Serra. "Augmentation de I'energie et de Ia pp. 98-111. 1988.puissance tournie par unite de volume dans un laser a CO2 en regime 111I] S. Lovold and G. Wang, "Ten atmospheres high repetition rate RF-pulse." (.'. R, Acad. Sci.. vol. 269. pp. 916-919. 1969. excited CO, waveguide laser." Appl. Phys. Lett,. vol. 40. pp. 13-!

[85] A J. Beaulieu. "Transversely excited atmospheric pressure CO, laser:" 1982.Appl. Ph/s. Lett., vol. 16. pp. 504-506. 1970. [112] F C. DeLucia. E Herbst. M. S. Field. and W Happer. 'Gas phase

1861 S. C. Lin and J. I. Levaiter. "X-ray preionization for electric discharge approaches to the near-millimeter wave source problem." IEEEJ. Quantlasers." Appl. Phvs. Lett.. vol. 34. pp. 505-508. 197', Electron.. vol QE-17. pp. 2171-2186. 1981.

[871 0 R. Wood. "High-pressure molecular lasers." Proc. /rEE. vol. 62. pp. [1131 R. I Hall. I. Cadez. M Landau. F Pichou. and C Scherman, V''bra-355-397. 1974. tional excitation ot hydrogen via recombmnative desorption ol atomic

1881 R r Brown. "CO., TEA Lasers" in Handbook of.Molecular la•ser,. P hydrogen on a metal surface." Ph/s. Ret I.et'r. vol 00. pp 337-141).K. Cheo. Ed New York: Marcel Dekker. 1997 1988.

1891 L S. Frost and A V Phelps. "Momentum transfer cross sections lot 1114] M Kon and B L. Ilalpern. "Vibranonall, excited ('O' trom the reaction

1 8

of 0 atoms and adsorbed CO on platinum." Chem. Phys. Lett., vol. 110. J. Gary Eden (S'75-M'TI--SM'92-F'88) receivedpp. 223-229. 1984. the B.S. degree in electrical engineering from the

[1151 R. A. Hart. L. A. Newman. A. 1. Cantor, and J. T. Kennedy. "Laser University of Marvland. College Park. and the MS.havezuide array." Appl. Phvs. Lett.. vol. 51. pp. 1057-1059. 1987. and Ph.D. degrees from the University of Illinois

[I I] A, F. Glova. Y. A. Dretzin. 0. R. Kachurin. F. V. Lebedev. and in 1973 and 1976. respectively. He was awardedV D. Pismennyt. "Phase locking of a two-dimensional array of a National Research Council Postdoctoral ResearchCO waveguide lasers.' Soy. Tech. Phy-s. Lett., vol. 11, pp. 102-103, Associateship from the Naval Research Laboratory1()85 (NRL). Washington. DC. in 1975.

In November 1976 he joined the staff of the

"Laser Physics Branch of the NRL. During his tenurethere he made several contributions to the areas of

visible and ultraviolet lasers and gas laser spectroscopy. Since joining theAland andrecaivdte MB69SMc 71 and boD.derne inrSot- faculty of the University of Illinots in 1979. he has conducted experiments

ueends University. Belfast. N. Ireland. focusing on electronic transition gas lasers and their applications. His researchueemUigeratito thel.s. Nanddis nowaResear group has demonstrated several new molecular lasers and amplifiers in the

He emigrated to the U.S. and is now a Research visible and near IR (including CdBr and Znl and has developed laserPhysicist at Wright-Patterson AFB (Ohio). where he spectroscopic techniques for studying the Rvdberp states of diatomic andstudies collisional plasmas. electron transport, and cntdlaser diagnostics. triatomic excimer molecules. Recently. his efforts have concentrated on laser

Dr. Garscadden is a Fellow of the American photochemical processes for growing semiconductor and metal films and on

r.Phasical Societs. nonlinear techniques for generating coherent VUV radiation. He has publishedextensively, and has been granted II patents.

Dr. Eden is a Fellow of the Optical Society of America and is an AssociateEditor of IEEE PHo'ro.ics TLcHNOLOGN LET'Rs. He received a ResearchPublication Award in 1979 for his work at the NRL. in which he codiscoveredthe first ton-beam-pumped laser.

Mark J. Kushner (S'74-M'79-SM'89-F'Qfl wasborn in Los Angeles. CA. in 1952. He receivedthe B.A. degree in astronomy and the B.S. degreein engineering from the University of California.Los Angeles in 1976. He received the M.S. andPh.D. degrees in applied physics from the Cali-fornia Institute of Technology, Pasadena. in 1977and 1979. respectively, where he also held theposition of Chaim Weizmann Postdoctoral ResearchFellow.

He served on the technical staffs of the SandiaNational Laboratory and Lawrence Livermore National Laboratory beforejoining Spectra Technology, Inc. (formerly Mathematical Sciences Northwest).where he was a Principal Research Scientist and Director of electron, atomic.and molecular physics. In August 1986 he joined the Department of Electricaland Computer Engineering at the University of Illinois (Urbana-Champaign).where he currently holds the rank of Associate Professor. He has published70 refereed papers and presented more than 120 conference papers on topicsrelated to gas and solid-state lasers, pulse power plasmas. plasma chemistry.chemical lasers, plasma processing of semiconductors, plasma treatment oftoxic gases, and laser spectroscopy.

Dr. Kushner is a member of Phi Beta Kappa. Tau Beta Pi, Sigma Xi. theOptical Society of America. the Materials Research Society, and is a Fellowof the American Physical Society.

'9