Papers Electron Beam Triatomic Excimer

2268 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. QE-17, NO. 12, DECEMBER 1981 Papers

Electron Beam iatomic and Triatomic Excimer

Abstruct-The spontaneousand stimulated emission characteristics ENERGY, eV 12'10 9 8 7 4 .~~--~~~~.-~~~~-~~~...~~~~~.~~6 5 ~.2 of three recentlyreported broad-band blue-green rare gas-halide ex- Ar? Ktl Xe? cimers, XeF (C+A) at 486 nm, Xe2CI at 518 nm, and KrzF centered RAPE GCSE5 6 6 IrFn.n5 KrCl KIF XeBrUb" YeCl XeFIB-XIn at 436 nm, are reviewed. The influence of different halogen donors RG HCuDES and buffer gases as well as optimization of both the gas mixture and y 4'0 K,O RG 3LIDES E1 8 opticalresonator configuration for wavelength tuning were studied. E CI. CIF Br I. BrF IF ,bq.., Rd. The kinetic mechanisms which describe the formation and quenching hALOGENS o^ 'G of Xe2Cl* are discussed. Problems with achieving optimum gain as well HgCl Z"l MlL HALICES 0 0 d1' as understanding both molecular andatomic absorptions in electron XaF ic -4, beam excited raregas-halide mixtures aredetailed. XeF {C-AI ezE ill F Kr,F Xe,Cl [WF] L',6 "~ , , . .x "7 ." . ," Rg>X TRIMERS f5l CZ I. INTRODUCTION

~ ~-~~~ ~-~~r~. ~1~~ , ~ ~., ~ ~~ -~ ---- ~~~~~~~ INCE 1975, numerous excimers have been reported that 103 2c3 300 ioc 530 600 650 Sare capable of generating high power laser reduction in the WAVELENGTH. nrn __ visible and ultraviolet spectrum [l] [5]. The relatively sim- Fig. 1. Wavelengths and tuning ranges of the various rare gas, rare gas- - halide, rare gas-oxide, halogen, mercury halide, and broad-band RgX ple techniques required to pump such lasers, aswell as their and Rg2X excimer lasers. For Ar2F and Xe2F, thespectral ranges of demonstrated high efficiency, have made them useful lasers in fluorescence are indicated. many interesting applications. Recently,there has been in- terestin exploring the broadbandwidth emission,which is the wavelength selecting element, a tuning range of 4 nm for observed from several diatomic and triatomic excimers as an Ar2 excimer laser at 127 nm [7 J and -2 nm for ArF, KrF, wavelength tunable lasersources (see Fig. 1). Alternatively, XeC1, and XeF (B +X) excimers [8], [9] has been achieved. conversionof narrow-band UV excimer laser emission into In this paper, the general characteristics of three new electron longer wavelength tunable dye laser radiation has been dem- beam pumped lasers, XeF (C--A) [lo], Xe2Cl [ll], and onstrated [4] .l Wavelength tuning ofan excimer was first Kr2F [12], capable of broad-band tunability will be discussed. achieved over a 6 nm tuning range with the Xe, laser centered Fig. 2 summarizes the complexes associated with high pressure at 172 nm [6]. With either a prism or diffraction grating as mixtures of the rare gases argon, krypton, and xenon with the halogens fluorine and chlorine. The pure rare gas excimers, Manuscript received July 20, 1981. This work was supported by the Arf, Krf, andXef with UV emissions at 126, 146, and 172 Office of Naval Research, the National Science Foundation, and the Robert A. Welch Foundation. nm, respectively, are shown in the first column.The other F. K. Tittel and W. L. Wason, Jr. are with the Department of Electri- columns list their respective exciplex transitions:the short cal Engineering and Rice Quantum Institute, Rice University, Houston, wavelength B+X transition,the longerwavelength broad- TX. -+ G. Marowsky is with the Max-Planck Institute, Gottingen, Germany. band C A transition, and the emission of the triatomic ex- M. C. Smayling was with the Department of Electrical Engineering ciplex. For Ar2F* and Xe,F*, the wavelengths listed are the and Rice Quantum Institute, Rice University, Houston, TX. He is now center of the fluorescenceemission, whereas thedata given with Texas Instruments, Houston,TX. 'Commercial excimer laser pumped dye laser systems are offered by for Kr2F* and Xe2Cl* correspond to the center of the laser several manufacturers. emission, as obtained in the experimentsdescribed below.

0018-9197/81/1200-2268$00.75 b 1981 IEEE TITTEL et al.: BROAD-BAND DIATOMIC AND TRIATOMIC EXCIMER LASERS 2269

I Rg X.

Rg+X

Fig. 4. Chemical structure and potential energy diagrams for a typical (a) diatomic RgX* and (b) Rg2X* triatomic molecule. Fig. 2. Raregas halide complexes and their associated emission wavelengths. The C -+ A transition of ArF has not yet been identified.

Ar LOOtorr Botm Ar Xe 1.2 torr CCIL

F CI Br 1 J n I

300 350 400 450 500 550 lArl WAVELENGTH, nm - I -- 36 54 Fig. 5. Xenon pressure dependenceof XeCl (B -t X)and Xe2Cl fluorescence spectra. ATOMIC NUMBER - Fig. 3. Relative dissociation energy of rare gas-halide excimers versus the atomic number of the rare gas; represents the pure rare gas Excimer lasers operate on electronic transitions involving a excimers such as Ne;,AI;, Kr;, and Xe;; represents a B -+X ex- bound ionic excited state and a repulsive, or at most weakly ciplex; o represents a C-t A exciplex; * stands for triatomic exciplexes. Thedouble lines indicatethose trimers whichhave shown bound, lower state that rapidly dissociates. The lowest elec- laser action. tronic level is designated the X state, with higher levels labeled A, B, C, etc. [Fig. 4(a)]. Due to the weakly bound nature of 11. CHEMICALPHYSICS the X state, the allowed B +X transition is inherently narrow Homogeneously broadened excimer emission bands have banded. The C+A transition, however, terminateson the been observed on the long wavelength side of most B +X elec- highly repulsive A state, which gives rise to the broad band- tronic laser transitions as incidental companion fluorescence width of this transition. For the triatomic exciplex [Fig. 4(b)] [l], [13]. These emission spectra are attributed to diatomic Rg2X*, the ground state is repulsive, similar to the A state of C -+ A transitions and triatomic homonuclear and heteronu- the dimer, and thus, the emission is characteristically broad clear rare gas-halide excitiers. banded. : In Fig. 3, the dissociation energies of the rare gas-halide ex- The formation of the trimer depends on several factors. It is cimers listed in Fig. 2 are plotted versus atomic number of the strongly enhanced by increasing both the buffer gas pressure rare gas. Each energy is normalized to the dissociation energy and the concentration of the rare gas. High pressure favors ex- of its respeetiw lare gas excimer Art, Krt, and Xet . The un- cimer formation by, increasing therecombination rate of stable Ne; and NeF* exciplexes have been adde{ for complete- atomic excited states into the molecular upper level.Fig. 5 ness. From this figure it is apparent that: 1) the dissociation depicts the effect of xenon pressure by showing the relative energy of the RgX exciplex decreases with increasing atomic Xe2C1* fluorescence emission for two different Xe' pressures number, and 2) the dissociation energy increases regularly for inelectron beam excited Ar/Xe/CC14 mixtures.The XeCl" a particular rare gas with F, Cl, Br, and I. This diagram is use- fluorescence intensity is not only reduced byopening the for- ful in the search for wavelength positions of new unknown mation channel towards Xe, C1*, but also by increased nonradi- candidates for broad-band C+A transitions (marked by the ative quenching of XeCI" due to the higher Xe partial pressure, rectangular symbols) or new trimers (marked by the triangular as discussed in Section VI. symbols) such as Ar2CI* or the bromides of Kr and Xe. Because of the broad bandwidth characteristics of these ex- Schematic potential energy curves for the various RgX" and cimers, the stimulated emission cross section of the bound-free RgzX" excimer lasers under consideration are shown in Fig. 4. transitions is small. Therefore, pumping rates larger than those *' 2270 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. QE-17, NO. 12, DECEMBER 1981

TO OMA I* PHOTODIODE ACCELER- tt FARADAY CAGE , PDP 11/23 I INTRA CELL OPTICAL RESONATOR

FARADAY CUP PROBE

TRANSIEWT pdqPLOTTER GRAPHICS ;I PRESSURE GAUGE

TO BUFFER GASES +VACUUM SYSTEM

Fig. 7. Schematic diagram of the transverse electron beam pumped laser cell. Fig. 6. Blockdiagram of the experimental arrangement and the data acquisition system. and Faraday probe, which can both be mounted inside the laser cell, monitor energy deposition and current density along of atomic level lasers are necessary for reasonable gain. This the optical axis. With an 8 cm X 1 cm cathode and 6 atm of requires intense electron beam or other powerful pumping ex- argon, the current density is -100 A/cm2 and the energy den- citation (fast discharge or optical) of the excimer species in sity is -2 J/cm2 at a distance of 2.5 cm from the anode foil. order to achievelarge excited state populations. The forma- The laser cell is connected to an all stainless steel gas-vacuum tion of the excimer relies primarily on ion recombination and manifold. The quality of the gases used in the experiments chemical reactive collisional processes involving excited rare and the filling techniques employed are important factors in gas species and an appropriate halogen donor. obtaining consistent experimental results, since quenching by ., even small amounts of impurities can be severe. The rare gases used-neon, argon, krypton, and xenon-were of research 111. EXPERIMENTAL DETAILS grade, 99.995 percent pure. Halogen donors such as CCI4 and The basic experimental arrangement, illustrated in Fig. 6, NF3 were of spectrophotometric grade, with a purity of 99t consists of an electron beam excitation source, a high pressure percent. The halogen donors were allowed to flow into the re- laser cell, and various electrical and optical instrumentation. action celi slowly to permit precise pressure measurement of Direct electron beam excitation is one ofthe most general less than 0.1 torr. The pressure was monitored with a capaci- techniques for pumping excimer lasers [ 11 - [3] on account of tive MKS baritron gauge. The gas manifold was evacuated to the high peak powers available (in excess of IO3 MW/cm3), less than 10 mtorr between fillings of the gas constituents. broad spectral applicability, and scalable and proven technol- Buffer gases were introduced into the cell through high-flow ogy, Efficiencies for conversion of electrical energy into ex- regulators to allow turbulent mixing of the halogen donor and cited states of the main buffers is close to 50 percent. Further- the rare gas.Because of deterioration of the halogen donor mqre,electron beam excited reactions can be conveniently molecules, refding of the laser cell was necessary after a few transferred to dischargelaser systems. Different techniques shots. are available for coupling the electron beam into the laser me- The cell arrangement shown in Fig. 7 has several features dium: These include transverse, radial, or longitudinal pumping which are importantfor achievinglaser output from low- of the laser cell. In this paper, we employ transverse excita- gain media. The intracell resonator, comprised of two high- tiola'of the active media by means of an electron beam pro- reflectivity mirrors separated by 10 cm, eliminates reflection duced by the field emission diode of a Physics International losses which would occur at window surfaces in the case of Pulserad 110 electron beam generator. This machine is capable external optics. Internal optics also allow the use of a short of producing 15 kA pulses of 1 MeV electrons with an 8 ns cavity length to decrease the roundtrip transit time. The reso- pulse duration. A device such as the Pulserad 110 is especially nator may be conveniently aligned from outside the cell, since useful as a pump source for excimer lasers, since the short out- the mirrors are held by flexible bellows mounts. Alignment is put pulse may not only be used to pump potentially low-gain necessary after everylaser shot to insure consistent perfor- laser media, butthe short excitation pulse permits detailed mance. The bellows mounts are made versatile so that they studies ofpertinent reaction kinetics. Theelectron beam is can be adapted to hold a variety of cavity optics, including transyersly injected into the laser cell through a 50pm thick mirrors, prisms, and gratings. The optimum cavity configura- titanium foil which serves as the anode of the field emission tion consisted of a high reflectivity (>99 percent) curved mir- diode, as well as a pressure barrier to separate the laser media ror (r = 2 m) and a flat output coupler (2-5 percent) with its from the high vacuum diode region. The laser cell and field optical axis 2.5 cm from the anode. Wide-band (+15 nm) di- emission diade are shown in Fig. 7, and are described in detail electric coatings were used with a protective SiOz outer layer in [14]. Several electrical monitors are useful in determining to prevent attack by fluorine and chlorine on the sensitive op- how well the e-beam generator is performing. A Marx bank tical coating. For such a resonator, the volume of the active voltage probe and a Rogowski coil (for measuring diode cur- region was only about 0.04 cm3, but this'geometry insured rent) monitor the basic machine performance. A calorimeter alignment stability against the e-beam induced shock wave. TITTEL et al.: BROAD-BAND DIATOMIC AND TRIATOMIC EXCIMER LASERS 2271

I AE-O.I~V I L- 1

>13- T

2- LL2 W z W 1- TIME, ns - hrra - Xe + F t XeFc,n LASER TUNING

Fig. 8. Energy level scheme of the diatomic excimer XeF (B -f X)and XeF (C -+ A) transitions.

Unstable resonators would yield a much larger interaction volume, but laser threshold conditions would be difficult to ob- tain with the low-gain excimers. The optical emission of the electron beam excited rare gas- halide mixture in the laser cell could be monitored in several ways. Timeintegrated spectral data were recordedwith a Princeton AppliedResearch optical multichannelanalyzer (OMA1) using a model 1254D intensified vidicon detector. WAVELENGTH, nm - Temporal data were obtained simultaneously with a fast vac- Fig. 9. Spectra of XeF (C-t A) laser showingtuning effects using reflectorswith maximum reflectivity centeredat 465 nm and 485 uum photodiode (ITT F4000 S5) or a photomultiplier (RCA nm, respectively. C31000B). Narrow-band interference filters were used to ob- tain temporal characteristics in a limited spectral range. The signal fromthe photodiode was recordedwith a Tektronix transition at 351 and 353 nm. Fig. 9 demonstrates the broad- R7912 transient digitizer. Careful electrical and X-ray shield- band tunability of the C-+A transition using a wide-band 100 ing of the detection apparatus minimized noise that resulted percent reflector, and two 2 percent output couplers centered from firing the e-beam machine. Both the transient digitizer at 465 and 485 nm. The figure shows a potential tuning range and the OMA 1 were interfaced to a DEC PDP 11/23 minicom- of 65 nm, from 450 to 515 nm, for an optimized mixture puter system, controlled by means of an HP 2648A graphics of 16 torr Xe, 8 torr NF3, and 6 atm Ar. An intracell Littrow terminal, as shown in Fig. 6. Temporal and spectral data re- prism or a high efficiency grating could also be used to tune duction and feature extraction were accomplished with soft- the laser output, but only in a longitudinally pumped cell with ware routines developed for this purpose. increased gain length [30] . The strong modulationof the laser output shown in Fig. 9 is due to transient absorption processes, IV. THE XeF (C+A) LASER and will be discussed in Section VIII. The laser pulses occur in the afterglow ofthe electron beam excitation because ofinitial As a first example of a broad-band tunable excimerlaser, the molecular absorptions arising from the argon buffer gas and XeF (C + A)laser centered in the blue-green at 486? 20 nm is the long ring-up time due to the limited gain [31] , which de- briefly discussed. This laser has been pumped photolytically lay the laser action by -35 ns [ IO]. An interesting feature of [ 151 , [ 161 and by both electron beam [lo], [ 171 and dis- the XeF C + A transition occurs when verya high-Q resonator charge [ 181 , [ 191 excitation. Continuous broad-band tunabil- cavity is used. A sudden increase in the output power by as ity and an output energy in excess of 6 J have been reported much as two or three orders of magnitude, accompanied by in [20]. Amplification by injecting tuned dye laser pulses is enhanced spectral and temporal narrowing, is observed. This described in [21] . The relatively high gain of XeF (C + A) may be due to enormously increased photon flux inside the [22]-[24] accountsfor its successful operationunder such high-Q cavity. In this case, a few of the absorption features diverse conditions.The relevantenergy levels of the C+A (indicated by 1-3 in Fig. 10) have been bleached out and ap- and B +X transitions of XeF are given in Fig. 8. Contrary to pear in emission. A possible explanation for this high output the usual case, the Cstate lies about 0.1 eV (700 cm-l) below could be that the extremely high-Q cavity for the C + A tran- the B state in this system 1251 - [27]. The ratio of C and B sition might lead to the channeling of all of the inversion en- state quantum yield is about 0.1 [28]. The small energy dif- ergy into the C-t A bound-free emission. Assuming a Boltz- ference between the B and C states may be used to enhance mann temperature distribution between the C and B states, the population density of the C state by cooling the gas mix- and a level spacing of -0.1 eV, one can only account for a fac- ture to below room temperature [29]. Heating the mixture tor of 50 increase, which is about 10 times too small for the ob- increases the B state population, and thus favors the B 3 X served increase in output intensity. An alternative explanation 2272 IEEE JOURNAL OF QUANTUMELECTRONICS, VOL. QE-17, NO. 12, DECEMBER 1981

TABLE I SPECTROSCOPIC PROPERTIES OF THE TRIATOMICRARE GAS HALIDES

X AX *rad Rg2X (nm) (nm) (ns) References

NezF 145a NezF - - [371 Ar2F 290 45 220 b, [49], ArzCl 246 30 - Kr2F 420 70 >I30 b, [33],[ll],[48], [55] KrzCl 325 30 >500 [13] KrzBr 318 - - l201 Xe2F 630 100 -200 b XezCl 490 80 >135 b, [13], [43], [45], [70] XezBr 430 80 250 1131, 1201 XezI 375 - - [I31 aCalculated, possible spectrumin [37 1. bThis work. R. Sauerbrey and H. Langhoff, private communication,T&[& F] = 230 f 100 ns.

!L5O 460 L70 LBO 490 500 510 520 WAVELENGTH (nml Fig. 10. Influenceof cavity-& on XeF (C-t A) laser intensity.For high-Q, cavity (upper spectrum) laser intensity increases by a factorof lo2. Atspectral positions @, @, and @, absorptionfeatures of low intensity spectrum are reversed into emission lines. Mercury and Ar' laser lines are included for calibration. would be that the high intercavity photon flux may lead to nonlinear bleaching [32] of an unusually broad-band absorp- WCVELENGTH nm - tion feature of unknown origin. In any event, the large output -1200- powers observedsuggest that for efficient utilization of the -9 aotrn Ar BatmAr x 20 torr Kr LOO torr Kr C+A transition, either a cavity dumping technique or a fast 10 torr NF~ 10 torr NF3 - injection procedure might prove very useful. z,,,> 6-x jKrFi Lot z 12~3nm) W +loo. / V. THE TRIATOMICRARE GASHALIDE EXCIMERS: b IL20nml XezCl AND Kr2F I Inaddition to emission from the raregas-halide excimer RgX", broad-band emission is observed at longer wavelengths from triatomic exciplexes RgzX" [I], [14], [33] - [36], as described in Section 11. Such trimers are primarily formed by three-body collisions between the excimer RgX" and a rare (b) gas. The ionic potentials of the excited states of these trimers Fig. 11. Fluorescence spectra of (a) ArzF and ArF (B+ X) and (b) of Kr2F and KrF (B -+ X). In the case of (b), the fluorescence peaks of have been calculated using ab initio computations [36] - [39] ArzF and KrF (C+ A)are superimposed at 290 nm. and a "diatomics in molecules'' (DIM) approach [40]. Table I lists the homonuclear raregas-halide triatomic molecules, as a rare gas, and several atmospheres of a buffer gas. Composite well as such important parameters as the emission wavelengths fluorescence spectra of the excimers studied in this work are for transitions to the repulsive ground state [Fig. 4(b)J, the shown in Figs. 11 and 12. Such fluorescence measurements bandwidth Ah, and the radiative lifetime (reciprocal Einstein served to determine various spectroscopic and laser parameters. A coefficient) 7,ad. It is immediately obvious from Table I In addition, kinetics andoptimum gas composition of the that the bandwidths for these bound-free transitions, typically broad-band exciplex media were studied under fluorescence 50-80 nm, are much larger than the 1-2 nm for the mainline conditions. B +X transition. The radiative lifetimes are also much longer, A laser spectrum for XezCl is shown in Fig. 13. The laser typically 100-200 ns for the trimers, as compared to 4-40 ns spectrum is characterized by spectral narrowing with the for the diatomic excimers. Of the 10 trimers listed in Table I, FWHM decreasing from 80 to 13 nm. Furthermore,the only XezC1 [ll] and Kr2F [12], [41] have been observed to intensity of the stimulated emission increases compared to lase so far. The spectral and temporal fluorescence character- fluorescence intensity. In addition, the peak laser intensity is istics of high pressure Ar/Xe/NF, , Ar/NF,, and ArIXeICCl, red shifted by 30 nm from the fluorescence peak. This is partly mixtures were evaluated. These mixtures were typically com- caused by the h4 gain dependence [ 13 and partly by the cavity posed of a few torr of ahalogen donor, several hundred torr of reflectors centeredat 520 nm.Cavity enhanced atomic ab- TITTEL et al.: BROAD-BAND DIATOMIC AND TRIATOMIC EXCIMER LASERS 2273

-t 200 "o Batm Ar E otm Ar Y 20torrXe 200 torr Xe 1 torrCCli 1.2 torr CCI, 150 i

WAVELENGTH .nm- Fig. 14. Temporalcomposite of XezCl laserand fluorescence,XeCl (B -+X)fluorescence, and the e-beam pump pulse.

pulse and the laser output reaches its maximum peak intensity later than the XezCl fluorescence peak. As discussed pre- viously, the delay of the laser pulse is a consequence of the long ring-up time of the cavity oscillation as discussed in [3 1). A series of experiments was conducted to find optimal conditions for maximum laser output. The gas mixture was studied in detail over the following ranges: CC14, 0.4-3 torr; Xe, 100- WAVELENGTH .nm- 800 torr; Ar, 2-10 atm. The strong influence of CCI4 and Xe (b) partial pressure on the laser output is evident in Fig. 15(a) and Fig. 12. Fluorescence spectra of (a) Xe2C1,XeCl (B-t X), andXeCl (b). At 200 torr of Xe, the laser intensity had a sharp peak at (C -t A)and (b) XezF, XeF (B --f X), and XeF (C + A). 1.2 torr CCl, with the peak shifting for higher Xe pressures. As can be seen, the effective pressure range for CC14 and Xe I is quite limited. The argon buffer gas had a surprising effect on the intensity, asFig. 15(c) shows. The saturation at high 6atm Ar Ar pressures reflects botha limitation in theformation of 400 torr Xe Xe2Cl* and absorptionsby excited molecular argon [42]. Neon and nitrogen werealso used as buffer gases to isolate some of the effects of argon absorption. However, the trimer did not achieve threshold laser conditions for up to 10 atm of N2 , and showed only weak laser action with up to 12 atm of neon. Other chlorine donors species such as C12, HCl, CHC13, CH2C1,, and CH3Clwere also tried, but CC4 proved to be better in terms of optimizing laser performance.

Hg-5L6 Besides spectral and temporal narrowing, the trimer output showed spatial narrowing as well. The laser beam divergence Ar-51L.5 was measured to be about 8 mrad. As in the case of XeF (C-t A), tuning was accomplished with cavity reflectors of different center wavelengths. Theother triatomic laser Kr2F[12], although somewhat I I I I I 350 400 450 500 550 weaker thanthe Xe,C1 laser, showed all the characteristics WAVELENGTH (nml expected for stimulated emission behavior, which weredis- Fig. 13. Laser spectrum of the trimer Xe2Cl. Various Hg and the 514.5 cussed in detail for the XezCl laser. The peak laser output oc- nmargon ion laser line are indicated for calibration of the OMA I curred with 10 torr NF3 as the halogen donor, while the output spectrum. showed no real maximum for both Kr and Ar but, rather, increased with pressure. Atomic absorptions by krypton meta- sorption lines are apparent as in the case of the XeF (C+A) stables were identified in the laser spectrum. laser. Fig. 14 shows typical normalized temporal characteristics of the W and visible fluorescence and laser output from VI. KINETICSOF TRIATOMICEXCIMERS a high pressure Ar/Xe/CC14 mixture. The laser pulse exhibits The energy transfer and excited state kinetics for trimers is temporal narrowing compared to the fluorescence pulse, and is complex, especially for electron beam pumping [ 11 - [4] . The considerably delayed (by -35 ns) from the e-beam pulse due detailed kinetic route by which the excimer is formed depends to buffer gas transient absorptions. Furthermore, the Xe, C1" upon the excimer, the halogen donor molecule, the excitation emission occurs delayed with respect to the W fluorescence conditions, and the partial pressures of the constituent gases. 2274 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. QE-17, NO. 12, DECEMBER 1981

Xe,CI 8 arm Ar 200 Torr Xe

I of Xe,CI Fig. 16.Schematic diagram the Xe2C1 kinetic process. Radiative 8 aim Ar lifetimes and quenchingrate constants arealso depicted. "O' n 1.2 Torr CC14 t 04 I \

01 I 0 200 400 600 800 1000 1200 XENON PRESSURE, Torr ---) (b) 10 20 30 LO 50 TIME, ns - Fig.17. Comparison of experimentally observedXezCl fluorescence pulses andcalculated pulsesbased on the kineticsmodel of Fig. 400 Torr Xe 16 1361.

A reaction scheme forthe formation of Xe2C1* from XeCl* via a termolecular reaction is illustrated in Fig. 16. For XeCl the production of Xe2C1 corresponds to an alternative decay channel, together with radiative decay and collisional quenching due to Ar, Xe, and CCL, with quenching rate co- efficients as indicated in Fig.16. The XeCl* emission data 260 460 660 SO0 lob0 1'200 appear to indicatea further possible quenching mechanism, ARGON PRESSURE,KPA - with kXeCI(Ar,Xe) 4 X IOw3' cm' s-l byanother three- (4 body collisional process that involves a highly unstable inter- Fig. 15. Dependence of Xe2Cl laser intensity upon (a) CCl4 concentra- mediate species such as ArXeCl* [5 11 , [52] . For the pressure tion, (b) xenon concentration, and (c) argon concentration. range required for optimum trimer fluorescence, XeCI* suffers The main excitation mechanism takes place via three-body severe quenching by the high xenon pressures. The termolecu- reactions of excited and ionized excimers. In this paper, we lar formation mechanism may be tested in several ways, as dis- limit the discussion to Xe2CI* kinetics based on temporal fluo- cussed in [44]. For example, Fig. 17 shows the good agree- rescence measurements of electron beam excited Ar/Xe/CC14 ment between the experimentally measured and calculated mixtures [43]. It is apparent from Fig. 14 that the Xe2Cl* Xe2Cl pulses for a mixture of 2 atm Ar, 400 torr Xe, and 2 fluorescence peak occurs after the W fluorescence had effec- torr CC14. tively decayed to zero. From this observation and previous Rateconstants for the formation reaction and for several work [44] [45] itmay be concluded thatthe diatomic ex- termolecular quenching reactions of Xe2C1* have been mea- cimer XeC1* is a precursor in the reaction chain leading to the sured [43] and are given in Fig. 16. The value of the formation formation of Xe2CI*, subsequent to channeling of initial elec- constant kl (Ar, Xe)was found to be 1.5 k 0.5 X cm6 - tron energy into the excited dimer state. Such a conversion s-l. The dominant loss process for both XeCl* and Xe2C1 is of RgX* into Rg,X* has been considered as the predominant quenching by the halogen donor CC4. Quenching by argon and formation mechanism for other ionic and substitutive reactions by xenon is less severe. The quenching conditions play an im- involving two or three bodies for Ar2F, Kr2F7and Kr2Cl [I31 portant role in determining the laser operating conditions, as E461 - [501 . was shown in Fig. 15. The rate of collisional quenching of the TITTEL et al.: BROAD-BAND DIATOMIC AND TRIATOMIC EXCIMER LASERS 2215 200 torr Xe ;i/r 10 I I 0.5 0- 0- 0 10 20 0 4 8 12 0 2 4 6'8 10 12 lL CCI, PRESSURE, torr - NF, PRESSURE, torr- c -z NF3 PRESSURE,torr (a) @) - Fig. 18. Stern-Volmer plots of (a) Xe2CI* against CC14 pressure, and (b) Kr2F against NF3 pressure. trimers was established by measuring thedecay constant 7 for a number of mixtures containing constant amounts of two of the gases but different amounts of the third 'component. Some of the experimental data are shown in Figs. 18 and 19. In Fig. 18, the XezCl and KrzF decay frequency7-l is plotted as afunction of the CC14 and NF, pressures, respectively. Fromthe slopes, and k&-, were caldated as (6 +- 1) X 0 0 1.0 20 0 cm3 s-' and 5 X lo-" cm3 s-l, respectively. Thus, CCl' PRESSURE,torr NF3 is a much less severe quencher than CC4. This observa- - Fig. 19. Pressure dependence of Xe&l* and KrzF*fluorescence on tion is substantiated by the halogen donor pressure dependence the halogen donorpartial pressure. of the fluorescence of Xe2C1 and Kr2F as shown in Fig. 19. Hence, it ispossible to employ much higher NF3 pressures are included for comparison. It is immediately evident froq than CC4 to optimize fluorescence intensity. However, as dis- this table that the broad-band excber lasers possess far lesq cussed in the previous section, a search for another chlorine gain than their narrow-band rare gas-halide counterparts. This donor has not been successful thus far. The radiative lifetime is due to thesmaller stimulated emission cross section, and the of the donors can be estimated from the zero interceptin Figs. fact that the trimers and the C+ A dimers represent decay: 18 and 19'. This intercept contains contributions due to col- channelsfor the primarydiatomic exciplexes. Except for lisional quenching of Xe2C1* by Xe and Ar, and of Kr2F* by Kr2F, the general agreement is good. In the case of Kr2F, the Kr and Ar. However, when these were subtracted, the respec- gain measurement was made at a wavelength corresponding t@ tive radiative, lifetimes were computed as 135 and 130 ns. a strong absorption in the laser spectrum. Hence, little or no The postulated trimer formation mechanism predicts only a gain would be expected, and in fact, onli absorption was ob- limited efficiency for the production of Xe2C1 from XeCl in served [53], [54]. The optic4 gains reported for XeF (C+ A) the absence of an intermediate heteronuclear excimer species. [22], [23] and Xe2Cl [55] are in, agreement with those mea- The laser output powers for electron beam pumped triatomic sured in this work even though 'the experimental conditions excimers so far are limited by the relatively small formation were quite different. Thelower gain for discharge pumped I. rate, and by the presence of transient molecular and atomic XeF (C + A) [ 171 is a result of the lower pumping powerused. absorptions. In the absence of absorbers, the unsaturated gain coefficient g of an excimer system may be calculated using fluorescence VII. OPTICALGAIN CONSIDERATIONS data in the expression [ 1J , [2] A critical issue of the performance of low-gain lasers such as Xe2C1, Kr2F, and XeF (C + A) is the relationship between the gain of the excimer transition and absorptions from other molecular or atomic species. The gain nqy be measured experi- where u is the stimulated emission cross section, h is the central mentally by 'a variety of methods. A comparison of three of wavelength of the transifion, Ah is the bandwidth (FWHM) of these methods is presented in [22], This reference describes the fluorescence spectrum, N* is the excited state population gain measurements made on XeF by a shutter technique, in density in the upper laser level, and c is the velocity of light. which different volumes of the 'gas are excited by the e-beam, The gain data pf the broad-band transition lasers in Table I1 by a mirror method, in which amplified spontaneous emission are based on the probe-beammeasurement 'of go for XeF is used, and adirect measurement using a CW laser probe beam, (C-+A) [23]. For this excimer, N* = 7 X lo1' cm-,.has Gain may &so be determined by measuring the ring-up time of been calculated, and this figure has then beenused to compute the laser pulse [28], [31]. Table I1 provides a comparison be- the other excited state densities based on comparable OMA tween the calculated and measured gains for each of thelaser and diode neasurepents, taking into account the spectral re- mixtures studied. The data for the B + X transitions af XeF sponse ofthe detectors. 2216 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. QE-17, NO. 12, DECEMBER 1981

TABLE I1 COMPARISONOF CALCULATED AND EXPERIMENTAL GAIN DATA

Excimer u, cm2 X 10-l~ gcalc, cm-l gexp, cm-l

KrzF 0.7 0.013a - Xe2Cl 0.8 0.024 0.027b, 0,025 1551 A12 F 0.2 0.002 - XeF (C + A) 1.1 0.07 0.05 [22],[23], 0.01 [17], 0.3 [22] XeF (B+ X) 50 -

aOnly absorption was observed in [53] and [54]. bThis work.

1.5 I Xe&I 8 otm Ar 400 Torr Xe 1.4 Torr CCI, \ --..

2.0 I I 0 50 100 150 200 NANOSECONDS ----I (a)

3 1 .-.. I 10 20 30 40 50 60 70 Xe&I 8 aim Ar TRANSMISSION T % --- 400 Torr Xe 2- 1.4 Torr CC14 Fig. 21. Calculated extraction efficiency of e-beam pumped low-gain excimers as a function of output coupling for the indicated value of E v go. Internal losses are characterized by ~i =O,Ol/cm (detzils in [47]). $1- z z \\ multiple transits (3-5 passes) through the excimer medium, \ --. "0- -c^ the direct probe method is particularly useful. This method makes it, possible to measure conveniently both the magnitude and temporal behavior of the effective optical gain. Fig. 20

L I I I shows that the net gain varies strongly as a function of dis- 0 50 100 150 200 NANOSECONDS tance from the anode foil. Near the anode foil, a strong initial (b) molecular absorption in the immediate afterglow of the pump pulse is seen, followed by a weak gain of - 1 percentlcm: The intermediate case, at the optic axis, shows gain and absorption Xe2CI canceling initially, followed by a stronggain of 2.7 percent/cm. 8 atm Ar 2- 400 Torr Xe The lack of initial absorption was unexpected, although the 1.4 Toir CCll E delay in the gain peak corresponds well to the delay in the laser output. At the furthest distance from the anode [Fig.'20(c)] no initial absorption was present, and a gain of 1.7 percent/cm z --- occurred soon after the e-beam pump pulse. The gain measure- 0- ments on Xe,Cl at 5 14.5 nm are useful, since they confirm the gain calculations as well as provide insight into the competi- tion between e-beam induced molecular absorption and population inversion. 0 50 100 I50 200 NANOSECONDS So far all the output power from, the broad-bandexcimer la-

(C) sers is in the llowatt range, primarily due to the small active Fig. 20. Temporal behavior of XezCl gain with optical axis (a) 1.5 cm mode volume available from the stable resonator. However, it from foil, (b) 2.5cm from foil, and (c) 3.5 cm from foil (vertical should be pointed out that even in low-gain systems, the ex- pass). traction efficiency in terms of photons coupled out divided by the number of excited states produced by theparticular transi- Because of the availability of several Ar+ laser lines in the tion could be as hgh as 50 percent with an appropriate choice Xe2Cl excimer band, and the possibility of increasing the sensi- of output couplers. Fig. 21 depicts the dependence of extrac- tivity of the gain measurement by many horizontal' or vertical tion efficiency versus transmission of the output coupler, TITTEL et al.: BROAD-BANDDIATOMIC AND LASERSTRIATOMICEXCIMER 2211

i- Both Ar: and Ar"" are capable of interacting with Xe to form Xe" [61] , [62] by the following reactions:

Ar"" + Xe -+ Xe" + Ar

Ar; + Xe 4 Xe" + 2Ar. Xenon has two excited metastable states, 'P,, {5p5(zP312) 6s[3/2],}, and 3P0,{5p5(2P,~2)6s'[1/2]o}. Inaddition, there are two other excited states with 5p56s configuration, 6s[3/2] (3P1)and 6s'[1/2] 1' ('PI), but they have allowed dipole transitions to the ground state, and are thus short-lived in situations where the xenon concentration is low. The 3P2 metastable is located at 8.3 eV, whereas the 3P0 metastable Fig. 22. Kinetic scheme leading to atomic absorption from 3P~xenon state is at 9.5 eV. These states may absorb radiation and un- metastables to high-lying Rydberglevels. dergo resonant transitions to higher excited Rydberg levels near theionization limit, as depicted in Fig. 22. For large tlnder the assumption of steady-state conditions for various principal quantum numbers n, the Rydbetg atom is hydrogenic gain figures ranging from go = 0.02 cm-' to go = 0.08 cm-I in the sense that it consists of a single electron moving in an and L = 7.5 cm, as discussed in detail in [56]. Fig. 21 shows orbitaround a unit charge. Some possible transitions from that the output powers of the broad-band lasers could be en- xenon metastables are to the nf(3/2), , fif(5/2), , n~(1/2)~, hanced by nearly one order of magnitude by selection of an and np(3/2),excited states. The selection nile is thatthe optimum output coupler. parity must change during the transition, and that AJ= +1,0 [55] . For principal quantum number n up to about 12, the VIII. MOLECULARAND ATOMICABSORPTION location ofthe observed excited states has been tabulated BY ELECTRONBEAM PUMPEDBROAD-BAND [63]. The energy levels of Rydberg states with higher prin- RARE GASHALIDES ciple quantum numbers can be calculated using a quantum de- The existence of transient absorptionin electron beam fect model. For such a model, one calculates the ionization pumped rare gas halides limits the performance af excimer la- energy of the higher states using the expression sers by lowering extraction efficiency and limiting scalability. Two of the most prominent characteristics of low-gain XeF (C-+A),Xe2C1, and KrzF electron beam excited lasers is that stimulated emission is delayed by broad-band ionic absorbing where Ry is the Rydberg constant (13.6 eV) and 6 is the quan- species and the presence of a large number of intense absorp- tumdefect, which is calculated from known lines withthe tion lines present in the emission spectrum [2] , [ 101 - [ 121 , expression [28], [57]. The broad-band absorbers have been identified / 13.6 as molecular in origin [42] , and come from the argon buffer gas in the form of dimers such as Ar;. The absorbing excited molecules are created at the onset of electron beam excitation, IP is the ionization potential of the metastable (1 2.1 eV for and have a lifetime of10-20 ns. This absorption seriously Xe) and E, is the energy of a known level with principle quan- limits the laser efficiency even though the excited state pro- tum number n. duction efficiency is very high. Thus, laser action occurs in In order to identify the absorptionlines in the laser spectrum the afterglow regime of the electron beam pulse. The transient of XeCl (C-+A)Xe,C1 and Kr,F, carefully calibrated time- nature of themolecular absorption is ciearly shown in Fig. 20. integrated spectra were recorded with the optical multichannel The absorption at 515 nm is on the order of 30 percent/cm analyzer. Calibration spectra were taken before and after each at 8 atm of Ar [42]. The other absorption has been identified shot, using mercury emission or argon ion laser lines as refer- as involving longer living excited atomic species, notably the ence. A computer program was used to plot the spectra with a calibrated wavelength scale. rare gas metastables, which surVive forhundreds of nano- t' seconds. These atomic metastables can give rise to discrete Fig. 23 shows typical high-resolution laser spectra for both absorption involving transitions to excited Rydberg states XeF (C + A) and Xe2C1. Absorptions by Xe metastables have [58] - [60]. been identified at many wavelengths throughout the excimer Metastable states are most likely created by excitation from laser band. Assignments of the transitions indicated in Fig. 23 excited argon species, following the kinetic chain listed below were based on data from [57], [58] , [63]. Absorption lines (also shown in Fig. 22): identified in XezCl ind XeF (C -+ A)laser spectra arising from the 3P0 metastable state are listed in Table 111. There was one e- + Ar -+ Ar+ + 2e- transition from the 3Pz metastable state to the 7~(3/2)~Ryd-

Ar+ + 2Ar -+ Arl + Ar berg state at 461.3 nm. However, no absorptions originating from the Xe 3Pt level were observed as in [28]. It should be Ari + e- -+ Ar2 -+ Ar"" + Ar noted that there are other absorption lines visible on the spec- Ar"" + 2Ar -+ Arjl' + Ar. tra which have not yet been identified. 2278 IEEE JOURNAL OF QUANTUMELECTRONICS, VOL. QE-17, NO. 12, DECEMBER 1981

I Ill Ill /I I I1 I I I I ‘€8 L7L L79 La2 187 L93 502 IIII I I I I I I I I L82 LE7 L93 502 516 539 I 3e- nflM), 1; ,b 8 7 6 =n

I I I I lIIIlilil L50 LEO 470 LBO L90 500 510 480 5m 520 540 560 WAVELENGTH, nm - WAVELENGTH, nm - (a) (b) ’

TABLE I11 435 nm, whereas the high-lying Rydberg levels have transitions IDENTIFIED TRANSACTIOh’S BETWEENXENON 6s” Po METASTABLESAND corresponding to around 350 Thus, the overlap of the HIGH-LYING RYDBERGSTATES nm. emission spectrum and the rich region of the absorption spec- n Transition eV in nm trum is not as great as in the case of XeF (C+A) or Xe2C1. Transitions from 3Pz to 6~(3/2)~and 6p(l/2), can be identi- 11.84 516.4 fied, as wellas atransition from the 3P0 metastable to the 11.91 502.5 5f(3/2), Rydberg level. The unidentified lines may be due to 9 11.96 9 493.4 absorption by NF3 orreaction byproducts. 1oa 11.99 487.1 11 12.01 482.5 Additional confirmation of the metastable nature of the ab- 12 12.03 12 479.1 sorbing species was obtained by comparing the temporal be- 13 12.05 476.5 havior of the XeF (C-t A) laser at 486 nm, where the OMA 14 12.06 14 474.5 15 12.07 472.8 spectrum indicates a maximum in output intensity with a cor- 16 12.07 471.5 responding minimum in the spectrum at 487 nm. In both hP(1/2)1 cases, the laser output had the same temporal characteristics. 11 11.88 509.3 Since the XeF (C -+ A) pulse length was greater than 60 ns, the 12 11.94 12 497.5 absorption must have remained relatively constant throughout 13 11.98 491.1b 14’ 12.00 484.2 the duration of the pulse. Hence, the initial level of the ab- 16 12.04 477.5 sorbing species must be metastable on ths time scale. 18 12.06 18 473.4 Several attempts were made to reduce the effects of the ab- nP(3/2)1 sorptions. For example, the KrzFlaser was operated at - 55°C 11.80 619.7 10 11.80 to shift the absorption to lower wavelengths [53], [64]. No 11.88 508.3 11 11.88 12 11.94 497.3 11.94 12 major change was observed. The XeF (C-+ A) and XezCl lasers were also operated at lower temperatures with different aValues for n = 10-16 calculated using6 = 0.043. buffer gases such as Ne, Xe, and N2 but most of the absorp- bThisvalue differs from a transition of 489.3 which is calculated , from the data in [63]. tion features remained unaffected. CValues forn = 14-18 calculated using 6 = 3.45. Absorption features are a significant characteristic 6f e- beam excited broad-band excimer lasers. Not only do these The laser spectra of Kr,F also shows a number of character- absorption lines reduce the total output power from the laser, istic absorption lines. Some of these can be identified with but they play a significant role in limiting the tunability of the transitions from krypton metastables to low-level Rydberg laser. In principle, it should be possible to eliminate a large states. The emission spectrum of KrzF is centered at around number of these lines by introducing aquencher for the xenon TITTEL ef at.: BROAD-BAND DIATOMIC AND TRIATOMIC EXCIMER LASERS 2279 metastableground state. Efforts in that direction have had ELECTRON BEAM limited success thus far, although recent workreported in *CCELERAToRY [28] suggests that the atomic metastable absorptions may be quenched by rapid transfer to Nz when using photolytic ex- (a) citation at 172nm with anXe: radiation source. IX. SUMMARY

Laser characteristics were studied extensively for the broad- LASER MIRROR band rare gas-halide excimers Xe2Cl, KrzF, and XeF (C-tA). / HELMHOLTZ COIL Since all three excimers have unbound ground states, a spec- trally wide emission bandwidth is available for wavelength tun- t I ing. Other excimers with wide fluorescence bandwidths-ArzF TO VACUUM PUMP I AND GAS HANDLING I and KIF (C+ A) at 290 nm, XeCl (C + A) at 345 nm, and XezF at 630nm-were studied but exhibited gains which were too small to achieve laser threshold in the 10 cm transversely pumped laser cavity used in these experiments, Spectral and DlOdE/ temporalfluorescence studies were conducted as an inter- ADAPTER COIL7 mediate step in establishing optimum conditions for obtaining L Fig. 24. Longitudinally electron beampumped laser cell in (a) the h laser action, and to derive kinetic constants. A detailed spec- geometry and (b) the head-on configuration. troscopic study was performed for Xe2C1*, which yielded formation and quenching rate constants as indicated in Fig. 16. Tuning was achieved for all three excimers by using optics cen- cavity reflectors. Preliminary experiments using the h configu- tered at different wavelengths. Efficiencies forthese lasers ration have yielded three orders of magnitude increase in power were small because of the small active volume defined by the for an Ar-Nz laser mixture as compared to the output when stable optical resonatorgeometry. The optical gain forthe using the transverse cell shown in Fig. 7. trimer Xe2C1was measured up to 2.7 percentlcm. The gain The longer available gain length of the longitudinal arrange- was delayed by transient absorptions involving the argon buf- ment wiU allow a variety oftuning experiments. Insertion fer gas. In addition, atomic absorptions from metastable lower losses encountered by using prisms, gratings, or etalons should states to Rydberg upper states were also identified in all the not exceed the per pass gain of the medium. This condition is laser spectra. Gain data are listed in Table 11, and the xenon not satisfied for low-gain broad-band excimers pumped trans- 3P0 metastable absorption lines observed in the XeF (C+ A) versly.Besides more efficient e-beam energy deposition, the and Xe2C1 laser spectra are compiled in Table 111. longer cell will also permit using reflectors which will greatly Several futureapproaches to the development of tunable increase the active volume, and hence yield much greater laser broad-band excimer lasers appear to be feasible. For example, output power. Such improvements will make broad-band ex- other RgzX molecules as indicated in Table I, such as Xe2Br, cimer lasers more useful for potential applications on account Kr2C1, or Ar2F, may be studied as potential new laser media. of their wide wavelength tunability. Second, new methods of coupling.the e-beam energy into the buffer gas should be tried to increase the laser output power. ACKNOWLEDGMENT For example, the approach of using optical pumping may be The authors would like to acknowledge the experimental as- evaluated since it avoids the problem of e-beam induced tran- sistance of G. Zhenhua, and helpful discussions with Dr. G. sient absorption effects, as reported in [15], [55] . Eventually, Glass and Dr. F. B. Dunning. Much of the early work reported discharge excitation of these laser media may be possible by here was performed with the assistance of Dr. W. E. Ernst and using optical or X-ray preionization [65] of high pressure rare Dr. R. E. Stickel, Jr. gas-halide mixtures in view of the difficulty of operating stable discharges in high pressure gases. Recently, supersonic flow- REFERENCES electron b‘eam stabilized discharge excitation of diatomic and C. K. Rhodes, Ed., Excimer Lasers. Berlin: Springer, 1979. J. J. Ewing, “Excimer lasers,” in Lasers Handbook, voL 3. Am- triatomic excimers has been proposed [66] to lower the ex- sterdam, The Netherlands: North-Holland. 1979. cited species quenching and absorptionlosses. M. J. Shaw, “Excimer lasers,” Prop. Qu&m Electron., vol. 6, Improved energy couplingcan be accomplishedby longi- PP. 3-54, 1979. M.H.R.Hutchinson, “Excimers and excimer lasers,” AppZ. Phys., tudinal pumping of the laser cell. In the “X” arrangement, il- VO~.21, pp. 95-1 14, 1980. lustrated in Fig. 24(a), the e-beam is coaxial with the laser axis I. S. Lakobaand I. S. Yakovlenko,“Active media of exciplex instead of being perpendicular to it. This method of e-beam lasers,” Sou. J. Quantum Electron., vol. 10, pp. 389-410,1980. D. J. Bradley,M.H.R. Hutchinson, and C. C. Ling,“Tunable pumping was developed bya group at SandiaLaboratories VUV excimer laser systems,” in Tunable Lasers and Amlications. [67]. A “head-on” version of this design, shown in Fig. 24(b), A. Mooradian, T. Jaeger, and P. Stockseth, Eds. Berlk springer; has beendescribed in [68]and [69]. Although this second 1976, pp. 40-48. W. G. Wrobel, H. Rohr, and K. H. Steuer, “Tunable vacuum W configuration is simpler than the “X” geometry, since no turn- laser actionby argon excimers,” Appi. Phys. Lett., vol. 36, ing coil is needed, it does not allow the use of high quality pp. 113-115,1980. cavity optics, since the e-beam must pass through one of the T. R. Loree, K. B. Butterfield, and D. L. Baker, “Spectral tuning 2280 IEEE JOURNAL OF QUANTUM ELECTRONICS, VOL. QE-17, NO. 12, DECEMBER 1981

of ArFand KrF discharge lasers,” Appl.Phys. Lett., ~01.33, body quenching of KIF* by Ar and broad-band emission at 415 pp. 171-173,1978. nm,”Appl. Phys. Lett., vol. 31, pp. 26-28, 1977. [9] R. T. Hawkins, H. Egger, J. Bokor, and C. K. Rhodes, “A tunable [34] N. G. Basov,V. A. Danilychev, V. A. Dolgikh, 0. M. Kerimov, ultrahighspectral brightness KrF excimerlaser source,” Appl. V. S. Lebedev, and A. G. Molchanov,“New excimer emission phys. Lett., vol. 36, pp. 391-392, 1980; and J. C. White, J.Bokor, bands of noble-gas halides,”JETPLett., vol. 26, pp.16-19,1977. R. R. Freeman, and D. Henderson, “Tunable ArF excimer laser [35] V. S. Zuev, A. V. Kanaev, L.D. Mikheev, and D.B. Stavrov, source,” Opt. Lett.,vol. 6, pp. 293-294, 1981. “Blue luminescence mechanism of ArKrF* and KrNzFexcimers,” [lo] W. E. Ernst and F. K. Tittel, “A new electron-beam pumped XeF Sop. J. Quantum Electron., vol. 7, pp. 1562-1563, 1980. laser at 486 nm,” Appl. Phys. Lett., vol. 35, pp. 36-37, 1979. [36] W. R. Wadt and P. J. Hay, “Electronic states of ArzF and Kr2F,” [11] F. K. Tittel, W. L.Wilson, R. E. Stickel, G. Marowsky, and J. Chem. Phys., vol. 68, pp. 3850-3863,1978. W. E. Emst, “A triatomic XezCl excimer laser in the visible,” [37] C. F. Bender and H. F.Schaefer, 111, “Ionicexcited states of Appl. Phys. Lett., vol. 36, pp. 405-407, 1980. Ne2F,” Chem. Phys. Lett., vol. 53, pp. 27-30, 1978. [12] F. K. Tittel, G. Marowsky, M. C. Smayling, md W. L.Wilson, [38] H. H. Michels, R. H. Hobbs, and L. A. Wright, “The electronic “Bluelaser action by the rare gas halide trimer KrZF,” Appl. structure of ArFand Ar2F,” Chem.Phys. Lett., vol. 48, pp. Phys. Lett., vol. 37, pp. 862-864, 1980. 158-161,1977. [13] D. C. Lorents, D. L. Huestis, M. V. McCusker, H. H. Nakano, and [39] W. R.Wadt and P. J. Hay, “The low-lying electronicstates of R. M. Hill, “Optical emissions of triatomicrare gashalides,” Ar2F,” Appl. Phys. Lett,, vol. 30, pp. 573-575, 1977. J. Chem Phys,,vol. 68, pp. 4657-4661, 1978. [40] D. L. Huestis and N. E. Schlotter, “Diatomics-in-molecules [14] G. Marowsky, R. Cordray, F. K. Tittel, and W. L.Wilson, potentialsurfaces for the triatomic rare gas halides:RgzX,” “Versatilehigh-temperature high-pressure vapor celldesign for J. Chem. Phys.,vol. 69, pp. 3100-3107,1978. electron beam excited laser studies,” Appl.Opt., vol. 17, I411 N. G. Basov, V. S. Zuev, A. V. Kanaev, L.D. Mikheev, and pp. 3491-3495, 1978. D. B. Stavrovskii, “Laser actionin an opticallypumped Kr2F 1151 W. K, Bischel, H. H. Nakano, D. J. Eckstrom, R. M. Hill, D. L. triatomic excimer,” Sov. J. Quantum Elecrron., vol. 7, pp. 2660- Huestis, and D. C. Lorents, “A new bluegreen excimer laser in 2661,1980. XeF,” Appl. Phys. Lett., vol. 34, pp. 565-567, 1979. [42] E. Zamir, D.L. Huestis, H. H. Nakano, R. M. Hill, and D. C. I161 N. G. Basov, V. S. Zuev, A. V. Kanaev, L. D. Mikkeev, and D. B. Lorents, “Visible absorption by electron-beampumped rare Stavrovskii, “Laser action due to the bound-free C(3/2)-A(3/2) gases,” IEEE J. QuantumElectron., vol. QE-15,pp. 281-288, transitionin the XeFmolecule formed by photodissociation,” 1979. Sov. J. Quantum Electron., vol. 9, p. 629, 1979. [43] G. Marowsky, G. P. Glass, M. C. Smayling, F. K. Tittel, and [17] J. D. Campbell, C. H. Fischer, and R. E. Center, “Observation of W. L. Wilson, “Dominant formation and quenching kinetics of gain and laser oscillations inthe blue-green during direct pumping electron beam pumped Xe2C1,” J. Chem.Phys., vol,75, pp. of XeF by microsecond electronbeam pulses,” Appl. Phys. Lett., 1153-1158,1981. VO~.37, pp. 348-350, 1980. [44] D.L. Huestis, “Studies of the triatomic rare gashalides,” pre- [ 181 R. Burnham, “A discharge pumped laser on the C->A transition sented at the 1979 Top. Meeting Excimer Lasers, Charleston, SC, of XeF,” Appl. Phys. Lett.,vol. 35, pp. 48-49, 1979. Sept. 11-13, 1979. [19] C.H. Fisher, R.E. Center, G. J. Mullaney, and J. P. McDaniel, [45] K. Y. Tang and D. C. Lorents, “Studies of excimer laser excited “A 490-nmXeF discharge laser,” Appl. Phys. Lett., vol. 35, kinetics using synchrotronradiation,” presented atthe 1980 pp. 26-28,1979. Lasers Conf., Orleans,LA, Dec. 15-17, 1980. [20] D. L. Huestis, “Tunable visible rare gas halide lasers,” presented [46] V. H. Shuiand C. Duzy,“Theoretical study of theformation at the 1980 Lasers Conf., New Orleans, LA,Dec. 15-19, 1980. rates of rare-gas halide trimers,” Appl. Phys. Lett., vol. 36, pp. [21] C. H. Fisher, R.E. Center, G. J. Mullaney, and J. P. McDaniel, 135-1 36,1980. “Multipass amplification and tuningof the blue-green XeF C -+ A [47] D. L. Huestis, R. M. Hill, D. J. Eckstrom, M. V. McCusker, D. C. laser,” Appl. Phys. Lett., vol. 35, pp. 901-903,1979. Lorents, H. H. Nakano, B. E. Perry, J. A. Margevicius, and N. E. [22] W. E.Ernst and F. K. Tittel, “Gain studies of electronbeam Schlotter, “New electronictransition lasersystems,” SRI Int., excited XeF laser mixtures,” IEEE J. Quantum Electron., vol. Menlo Park, CA, SRI Project PYU-6158,Tech. Rep. 1, 1978. QE-16, pp. 945-948, Sept. 1980. [48] G. P. Quigley and W. M. Hughes, “Lifetime and quenching rate [23] R. M. Hill, P. L. Trevor, D. L. Huestis, and D. C. Lorents, “Mea- constants for KR2F* and KR2*,” Appl. Phys. Lett., vol. 32, pp. surement of gain on the XeF (C-A) blue-greenband,” Appl. Phys. 649-651,1978. Lett., vol. 34, pp. 137-139, 1979. [49] C.H. Chen and M. G. Payne, “Ar2F radiative liietime measure- [24] G. Marowsky, F. K. Tittel, W. L. Wilson, and E. Frenkel, “Laser ment,” Appl. Phys. Lett., vol. 32, pp. 358-360, 1978; andC. H. gain measurements bymeans of amplified stimulated spontaneous Chen, M. C. Payne, and J. P. Judish, “Kinetic studiesof ArF* and emission,”AppZ. Opt., vol. 19, pp. 138-143, 1980. ArzF*in proton excited Ar-F2 mixtures,” J. Chem.Phys., [25] D. Kligler, H. H. Nakano, D. L. Huestis, W. K. Bischel, R. M. Hill, VO~.69, pp. 1626-1635, 1978. and C. K. Rhodes, “Energy ordering of the excited statesof XeF,” [50] K. L. Hohla, T.R. Loree, C.A. Brau, and W. E. Stein, “A linac Appl. Phys. Lett., vol. 33, pp. 39-41,1978. pumped excimer laser: Fluorescence and scaling studies,” Appl. [26] T. G. Finn, L. J. Palumbo, and L. F. Champagne, “The role of Php., VOL 25, pp. 329-336, 1981. the Cstate in the XeF laser,” Appl. Phys. Lett., vol. 34,pp. [51] G. P. Glass, F. K. Tittel, W. L.Wilson, M. C. Smaying, and G. 52-55,1979. Marowsky, “Quenching kineticsof electron beam pumped XeC1,” [27] M. Rokni, J. H. Jacob, J. C.Hsia, and D. W. Trainor,“The Chem. Phys. Lett., to be published. origin of thebroadband emissionof XeF,” Appl.Phys. Lett., [52] H. C. Brashears, Jr., D. W. Setser,and Y. C. Yu, “Emission VO~.35, pp. 729-731, 1979. spectra of KrXeCl*, KrXeBr*, KrXeI*, ArKrF*, and ArKrCl*,” [28] W. K. Bischel, D. J. Eckstrom, D.L. Huestis, and D. C. Lorents, J. Chem. Phys.,vol. 74,pp. 10-17,1981. “Photolytically pumped XeF (C->A ) blue-green laser,” presented [53] R. 0. Hunter, J. Oldenettel, C. Howton,and M. V.McCusker, at the 1979 Lasers Conf., Orlando, FL, Dec. 17-21, 1979; also “Gain measurements at 4416A on ArXeF* and KrzF*,” J. Appl. inJ. Appl. Phys., vol. 52, pp. 4429-4434, 1981. Phys., VO~.49, pp. 549-552, 1978. [29] W. E. Ernst and F. K. Tittel, “XeF laser characteristics studied at [54] J. G. Eden, R.S.F. Chang, and L. J. Palumbo, “Absorption in the elevated temperatures,” J. Appl.Phys. vol. 51, pp. 2432-2435, near ultraviolet wingof the KrzF* (410 nm) band,” IEEE J. 1980. Quantum Electron.,vol. QE-15, pp. 1146-1156, 1979. [30] J. Leigel, F. K. Tittel, W. L. Wilson, and G. Marowsky,“Broad [55] K. Y. Tang, D.C. Lorents, and D. L. Huestis, “Gain measure spectral tuning of an ebeam pumped XeF (C->A) laser,” Appl. ments on the triatomic excimer Xe2Cl,” Appl. Phys. Lett., vol. Phys. Lett., vol. 39, pp. 369-371, 1981. 36, pp. 347-349,1980. [31] G. Marowsky, M. Mum, and F. K. Tittel, “Excimer laser gain by [56] J. K. Rice, G. C. Tisone, and E. L. Patterson, “Oscillator perform- pulse-shape analysis,” IEEE J. Quantum Electron., vol. QE-17, ance and energy extraction from a KrF laser pumped by a high pp. 1281-1285, July 1981. intensity relativistic electron beam,” IEEE J. Qugntum Electron., [32] G. Haag and G. Marowsky, “Transmission of organic laser dyes vol. QE-16, pp. 1315-1326, Dec. 1980. under inclusion of stimulated emission,” IEEE J. Quantum [57]L. F. Champagneand R.S.F. Chang, “Transientabsorption Electron., vol. QE16, pp. 890-897, Aug. 1980. studies in pure rare gases from 2500 A to 4000A,”J. Phys., voL [ 331 J. A. Mangano, J. H. Jacob, M. Rokni, and A. Hawryluk, “Three- 41, C9, pp. 445-447,1980, TITTEL et al.:LASERSTRIATOMICEXCIMER BROAD-BAND AND DIATOMIC 2281

R. F. Stebbings, C. J. Latimer, W. P. West,F. B. Dunning, and Gerd Marowsky, for a photograph and biography, see p. 356 of the T. B. Cook, “Studies of xenon atoms in high Rydberg states,” March 1981 issue of this JOURNAL. Phys. Rev. A, vol. 12, pp. 1453-1458, 1975. R. D. Rundel, F. B. Dunning, H.C. Goldwire, Jr., and R.F. Stebbings, “Near-threshold photoionization of xenon metastable atoms,” J. Opt. Soc. Amer.,vol. 65, pp. 628-633,1975. R. F. Stebbings and F. B. Dunning, “Autoionization from high- lying 3p5 (*PllZ)np‘ levels in argon*,” Phya Rev. A, vol. 8, p. 665,1973. Wiim L. Wilson, Jr. (S’68-M’71) was born in Bay Shore, NY, on Feb- M. Bourene, and J. LeCalve, “De-excitation cross-sectionsof ruary 6, 1943. He received the B.S. degree in 1965, the M.E.E. degree metastableargon by various atoms and molecules,” J. @em. in 1966, and the Ph.D. degree in 1972, all in electrical engineering, Phys., vol. 58, p. 1452, 1973. from Cornell University, Ithaca, NY. R.E. Gleason, T. D. Bonifield, J. W. Keto, and G. K. Walters, From 1971 to 1972 he was an Instructor-Research Associate withthe “Electronic energy transfer in argon-xenon mixtures excited by Department of Electrical Engineering at Cornell. In 1972 he became electron bombardment,”J. Chem. Phys., vol. 66, pp. 1589-1593, anAssistant Professor of Electrical Engineering at Rice University, 1977; and C. A. Brau, Excimer Lasers,C. K. Rhodes, Ed. Berk. Houston, TX, where he is currentlyan Associate Professor. His re Springer, 1979, ch. 4. searchinterests include far-infrared devices andthin-fim magnetic C. E. Moore, Atomic Energy Levels, vol. 111, Nat. Bur. Standards, materials Circ. 467, U.S. Gov. Printing Office, Washington, DC, 1958. Dr. Wilson is a member of Tau Beta Pi, Eta Kappa Nu, Sigma Xi, the M. Rokni, J. H. Jacob, and J. A. Mangano, “Dominant formation IEEE Magnetics Society, the IEEE Microwave Theory and Techniques andquenching processes in ebeam pumpedArF* and KIF* Society, and the IEEE ElectronDevices Society. lasers,” Phys. Rev. A, vol. 16, pp. 2216-2224, 1977. S. C. Lin and J. I. Levatter, “X-ray preionization for electric di5 charge lasers,” Appl. Phys. Lett., vol. 34, pp. 505-508, 1979. B. Forestier, B. Fontaine, and P. Gross, “Supersonic flow low temperature electronic transition excimerlaser,” J. Phys., vol. 41, C9, pp. 455-462,1980. R. A. Klein, “A-3, Sandii’s lOOJ HF laser system,” Sandia Nat. Lab., Alberquerque, NM, Rep. SAND 79-1659, Sept. 1979. R. Sauerbreyand H. Langhoff, “Lasing inan e-beam pumped Michael C. Smayling was born on August 31,1952, in Kansas City, KS. Ar-Nz-mixture at 406 nm,” Appl. Phys., vol. 22, pp. 399-402, He received the B.E.E. degree from the University of Minnesota, Min- 1980. neapolis, and the M.S. and PbD. degrees in electrical engineering from R. W. Dreyfusand R. T. Hodgson, “Molecular-hydrogen laser; Rice University, Houston, TX, in 1972, 1976, and 1981, respectively. 1098-1613 A,”Phys. Rev. A, vol. 9, pp. 2535-2648,1974. He waswith the du Pont Company, Victoria,TX, in 1973, transferring H.P. Grieneisen, H. Xue-Ting, and K. L. Kompa Gem. Phys. to the du Pont Experimental Station in 1976. His work included elec- Lett., to be published. trooptical instrument design and design of a clinical patient identifica- tion system to be used with the du Pont “aca” (patent pending). He is currentlya Staff Member of TexasInstruments, Houston, TX.His present interests includeVLSI process development, submicrontransis- Frank K. Tittel (SM’72), for aphotograph and biography, see p. 1285 tor characterization, plasma kinetics,and laser chemistry. Heis a the Julyof the 1981 issuethis of JOURNAL. LicensedProfessional Engineer.