OPTICAL
`ENGINEERING
`
`December 2003
`Volume 42 Number 12
`ISSN 0091-3286
`
`SPIE-
`The International
`Society for
`Optical Engineering
`
`I
`
`OE letters
`Acousto-Opticol Tunable Filters
`Diffractive Optics
`Optical Clack Generation
`Optical logic Generation
`Optical Switches
`Waveguide Gratings
`Wavelength Monitors
`
`I Featuring Papers on
`Automated Visual Inspection
`Crock Measurement
`Digital Holography
`Fiber Amplifiers
`Fiber Bragg Gratings
`Fiber Dispersion
`Image Reconstruction
`Image Restoration
`Infrared Beomsplitters
`Infrared Detectors
`Infrared-to-Visible Converters
`Interferometry
`loser Heat T reoting
`liquid Crystals on Silicon
`Motion Blur
`Optical Encryption
`Optical Sensors
`PLZT Thin Films
`Polarizers
`Position Estimation
`Profilometry
`Radiometric Calibration
`Time-of-Flight Range Finders
`Ultraviolet losers
`X-Ray Optics
`
`This resource is also available
`ontheWWW.
`Use MadCat to launch.
`
`Energetiq Ex. 2021, page 1 - IPR2015-01300, IPR2015-01303
`
`

`
`OPTICAL
`ENGINEERING
`
`Donald C. 0' Shea
`Editor
`
`Editor
`Donald C. O'Shea
`Georgia Institute of Technology
`School of Physics
`Atlanta, Georgia 30332-0430
`4041894-3992 • Fax 404/894-9958
`E-mail doshea@prism.gatech.edu
`
`Editorial Board
`James Bilbro
`NASA Marshall Space Flight Ctr.
`Optical fabrication &: testing, coherent lidar
`
`Luc R. Bissonnette
`Defence Research Establishment Valcartier
`Lidar, curosol scattering
`
`Casimer DeCusatis
`IBM Corporation
`Fiber optics
`
`Ronald G. Driggers
`Army Research Lab.
`Infrared systems and radiometry
`
`Touradj Ebrahimi
`Swiss Federal Institute of Technology EPFL
`Image/video processing and coding
`
`G. Groot Gregory
`Lambda Research Corp.
`Interferometry
`
`Jiirgen Jahns
`University of Hagen, Germany
`Optical interconnects, micro-optics
`
`F~sKajzar
`CEA Saclay
`Nonlinear optics
`
`Ali Khounsary
`Argonne National Lab.
`X-ray optics, high-heat-load optics
`
`Raymond K. Kostuk
`University of Arizona
`Holographic materials, processes, and systems
`
`H. Angus Macleod
`Thin Film tenter
`Thin film technology
`
`Dennis W. Prather
`University of Delaware
`Physical optics
`
`Gregory J. Quarles
`VLOC
`Lasers
`
`Jennifer C. Ricklin
`Army Research Laboratory
`Atnwspheric optics
`
`Giancarlo C. Righini
`IROE. National Research Council
`Integrated optics
`
`Jannick P. Rolland
`School of Optics and CREOL at UCF
`Optical system design
`
`Rajpal S. Sirohi
`Indian Inst. of Technology Madras
`Interferometry
`
`Bryan D. Stone
`Optical Research Associates
`Lens design
`
`Andrew G. Tescher
`Compression Science Corporation, Inc.
`Image compression, image &: signal processing,
`video technologies
`
`Totnaliz R. Wolinski
`Warsaw University of Technology
`Optical fiber sensors and liquid crystals
`
`Jiangying Zhou
`Summus Ltd.
`Image analysis, pattern recognition and
`computer vision
`
`Managing Editor
`Karolyn Labes
`P. 0. Bo!t.lO
`BellinaJ!am, Washington 98227-0010
`3601676-3290 • Fu 3601647-1445
`E-mail joumals@spie.org
`
`Editerlal Staff
`Monica Crabtree
`Rita J. Davis
`Michelle Luff
`Aune Munger
`
`Publisher
`Eugene G. Arthurs, Executive Dir"ilctor
`Eric Pepper, Director of Publications
`
`Mailing:
`8PIE
`P.O. Box 10
`Bellingham, Washington 98227-0010
`
`Shipping:
`SPIE
`1000 20th Street
`Bellingham, Washington 98225
`360/676-3290 • Fax 360/647-1445
`E-mail joumals@spie.org
`
`Copyright e 2003, The Society ot Phofo.Opdeal
`Instrumentation EJIIIDeers. Copying of material
`in this journal for internal or personal use, or the
`internal or personal use of specific clients, beyond
`the fair use provisions granted by the U.S. Copy(cid:173)
`right Law is authorized by SPIE subject tu payment
`of copying fees. The Transactional Reporting Ser(cid:173)
`vice base fee for this journal is $15.00 per article
`(or portion thereof), which should be paid directly
`to the Copyright Clearance Center (CCC), 222
`Rosewood Drive, Danvers, MA 01923. Other copy(cid:173)
`ing for republication, resale, advertiting or promo(cid:173)
`tion, or any form uf systematic or multiple repro(cid:173)
`duenon of any material in this journal is pnlbibited
`~ wi1b. permission in wriling from the pub(cid:173)
`lisller. The CCC fee eode for this journal is 0091·
`3286Al3J$l5.00.
`Edif,Qrial. Send manuscripts and technical corre(cid:173)
`spontfence to SPIE. Papers and reviews published
`in ()ptical Engineering reftect tbe opiniotl. of the
`llllthor(s); ittc!Usion of andes and advertise·
`ments does oot necessarily ~titute endotse·
`ment by the editor5 or by SPIE.
`SPIE MealbeniUp. Aanual membetsbip dues in(cid:173)
`clude a choice¢' ~ Sooiety's four peer~ed
`joutJliJ.s. 2003 ~ @ues: $95; $45 letirecl;
`$40 Student. For I'QORI infmDation. C0U18Ct SPIE
`M~ 5eniGes at 3601676-3290, Fa: 3CiOI
`647-1445, e-mail: memllenhip@spie.org. CIJ8IIp
`or Address. Send old and new address to SPIE
`~-
`· . ·
`C1f11i«11 ~ (IBSN 0091·1216} (USPS
`4S0.970) is publisbeci ~ by . . ~ of
`~· lallfNMt'lli® ·~ 1000
`~ ·. w ~..,. ···-~-.;.....
`lOdlSt.. - - ~---~
`&.nuat ~· pritos: ·,u. .... ~·$550
`N. America. $61S OlllsUJcl K ~ ~
`'postate paid lit ~'W'l ,..,.,... at ad(cid:173)
`di\ionai Illlli1iJig o.rtc. • .., nr t tb Sea4 - - .
`dlal1les to ~ ~,. -
`-
`St.,
`Bellingham. WA 98225.
`Primed in the Vm.t'Sflt1es oa·~:,.,_.
`CCIIIId;alia.J~ . .lliili-:aE a-.
`prodt Pill*~.~.~ aeptS
`ta&-deduetible ~-~-support
`of its efforts .~ and to advance
`knowledae in
`ailCS oproe~ 8llPlied
`.
`science and ~· With special emphasis on
`educalion, SPIE amwally awards scholarships and
`publishes a gUide to opl:ics progrtlliiJ in colleges and
`universities an:IU8d the world. The Society has es(cid:173)
`tablished a fuad to support student travel to SPIE
`symposia and is facilitating donations of technical
`publications to educational institutions and libraries
`without access to this literature. For information on
`how you can contribute or participate in a deferred..
`giving program. please contact Sharon Kirkpalrick,
`SPIE Headquarters, P.O. Box 10, Bellingham, WA
`98227..()()10, 3601676-3290, Fax 360/647-1445,
`E-mail shsronk@spie.org.
`
`Energetiq Ex. 2021, page 2 - IPR2015-01300, IPR2015-01303
`
`

`
`Optical pumping of the XeF(C-+A) and iodine
`1.315-pm lasers by a compact surface
`discharge system
`
`B. A. Knecht*
`R. D. Frasert
`D. J. Wheeler*
`C. J. Zietklewlczl
`A. A. Senln, MEMBER SPIE
`L. D. Mlkheevll
`V.S.Zuevll
`J. G. Eden
`University of Illinois
`Department of Electrical and Computer
`Engineering
`Laboratory for Optical Physics and
`Engineering
`Urbana, Illinois 61801
`
`Abstract. Details are provided regarding the design, construction, and
`performance of a compact ( ""0.6-m2 footprint), single-channel surface
`discharge system and its application to optically pumping the XeF( C
`-+A) and iodine atomic lasers in the blue-green ( =480 nm) and near
`infrared (1.315 }LITI), respectively. The system has a gain (active) length
`of ""50 em, and triggering the discharge requires no high-voltage or
`high-current switches. Measurements of the velocity of the photodisso(cid:173)
`ciation bleaching wave and the small-signal gain of the XeF(C-.A) sys(cid:173)
`tem are described. At 488 nm, the gain coefficient r was found to be
`... 0.3% cm- 1 , a value comparable to those reported previously for sys(cid:173)
`tems dissipating considerably higher power per unit length. Single-pulse
`energies >50 mJ from the XeF(C-.A) laser (-485 nm) and >0.7 Jon
`the 5p 2P112-+5p 2P312 transition of atomic iodine at 1.315 ,urn have been
`obtained with nonoptimized resonator output couplings (5% and 10%,
`respectively). The rate of erosion of the dielectric surface has been mea(cid:173)
`sured to be ... 0.1 to 0.3 ¢Tl/shot for a glass ceramic dielectric, and the
`performance of two electrical configurations for the ballasting pins
`(feedthrough and V) is compared. c 2003 Society of Photo-Optic81 tnstrtJmenta(cid:173)
`tion Engineers.
`[001: 10.111711.1624849]
`
`Subject terms: lasers; visible; optical pumping: surface discharge.
`
`Paper 030139 received Mar. 24, 2003; revised manuscript received Jun. 9, 2003;
`accepted for publication Jun. 10, 2003.
`
`Introduction
`1
`The surface discharge, an electric discharge at the interface
`of a gas and a solid dielectric, was first reported by G. C.
`Lichtenberg in 1777. Over the past three decades, renewed
`interest in surface discharges has been driven by potential
`applications in materials processing, 1 surface cleaning, 2•3
`waste remediation, and optical pumping of lasers by
`photodissociation,4- 6 all of which require intense sources
`of ultraviolet (UV) or vacuum ultraviolet (VUV) radiation.
`Excitation of lasers by a surface discharge is particularly
`attractive because of the ability of the optical source to
`produce quasi-blackbody radiation having a characteristic
`temperature above 3 X l<f K and to do so with a relatively
`simple system amenable to repetitively pulsed operation. In
`1975, Beverly5 proposed pumping the iodine photodisso(cid:173)
`ciation laser (5p 2P 112-+5p 2P312 ;A= 1.315 ~m) with a sur(cid:173)
`face discharge and subsequently measured efficiencies ap(cid:173)
`proaching 10% and 3% for the conversion of electrical
`
`*Present address: 662N 300W, Lebanon, IN 46052.
`tPresent address: Realized Technologies Inc., 1530 Barclay Blvd., Buffalo
`Grove, IL 60089.
`*Present address: National Center for Sueercomputing Applications
`(NCSA), University of Illinois, 605 E. Springfield, Champaign, lL 61820.
`'Present address: Melles Griot, 2605 Trade Center Avenue, Longmont,
`C080503.
`lpennanent address: P. N. Lebedev Physical Institute, Moscow, Russian
`Federation.
`
`power to optical radiation in the 250- to 290-nm and 170-
`to 210-nm spectral regions, respectively.7 Studies have also
`shown surface discharges to be effective sources of extreme
`ultraviolet (XUV; hw-10 to 70 eV) photons,3 and charac(cid:173)
`teristic radiation temperatures of ( 1 to 2) X 104 K are at(cid:173)
`tainable for specific energy loadings of the plasma of l to
`4 J cm- 2 (Ref. 5). In 1979, Belotserkovets et al. 8 demon(cid:173)
`strated lasing from atomic iodine when the molecular pre(cid:173)
`cursor, C3F7 I, was photolyzed by a surface discharge.
`Since that time, a variety of atomic and molecular lasers
`initiated or pumped entirely by a surface discharge have
`been demonstrated,4•9 but much of the effort has been di(cid:173)
`rected toward the XeF photodissociation laser, 10•11 partly
`because of the breadth of the absorption spectrum of the
`parent molecule, XeF2 • The other factors providing impe(cid:173)
`tus for the efforts on XeF are the two laser transitions that
`are available (B-+X and C-+A) as well as the large satu(cid:173)
`ration intensity of the C-+ A band in particular. 12 The latter,
`-400 kW cm- 2,
`is a direct reflection of both the C-A
`stimulated emission cross section (-9x 10- 18 cm2) and
`the e-state radiative lifetime (93±5 ns). 13
`In a series of papers published between 1984 and the
`early 1990s (Refs. 14 to 17), Kashnikov, Zuev and co(cid:173)
`workers reported the characteristics of a sequence of
`surface-discharge-pumped XeF laser systems producing as
`much as 174 J on the B-+ X transition of the molecule in
`
`3612 Opt. Eng. 42(12) 3612-3621 (December 2003) 0091-3286/20031$15.00 © 2003 Society of Photo-Optical Instrumentation Engineers
`
`Energetiq Ex. 2021, page 3 - IPR2015-01300, IPR2015-01303
`
`

`
`Knecht et al.: Optical pumping ...
`
`the UV at 351 nm and up to 117 J in the blue-green ( C
`-+A transition, ). .... 485 nm). Consisting of individual dis(cid:173)
`charge sections, each typically 8 to 12 em in length and
`having a dedicated capacitor and high-voltage switch, these
`optical sources had active lengths ranging from 70 to 190
`em. The first system, reported by Kashnikov et al., 14 com(cid:173)
`prised eight discharge sections in a single-channel design
`and yielded a brightness temperature of = 3 X 104 K. By
`increasing the gain length to 190 em ( 16 discharge sec(cid:173)
`tions), raising the number of discharge channels to three,
`and with careful attention given to the current rise time and
`the purity of the XeF2 vapor, the Russian group subse(cid:173)
`quently produced single-pulse energies that are, to this day,
`unsurpassed for a primary coherent source in the visible. 17
`Despite the impressive characteristics of these lasers, in(cid:173)
`cluding the ability to operate at a pulse repetition frequency
`of 1 Hz, the system was quite large, and the dedication of a
`capacitor and high current switch to each discharge section
`introduced significant timing jitter ( <0.25 JJ-S). Further(cid:173)
`more, in low-impedance circuits of this type, switches con(cid:173)
`sume substantial energies and adversely affect the complex(cid:173)
`ity and lifetime of the system.
`In the mid-1990s, we reported the design and prelimi(cid:173)
`nary operation of a compact surface discharge system that
`successfully pumF.: the XeF(C-+A) and iodine photodis(cid:173)
`sociation Iasers. 1 •19 A more detailed account of the design
`and operation of this system and an alternative structure for
`the surface discharge ignition electrodes are presented here.
`Having an active length of '"'"50 em, this surface discharge
`system requires no high-voltage (or high-current) switches,
`which has the beneficial result of lowering the firing jitter
`as well as the equivalent series inductance for the system.
`Careful attention given to the design and layout of the de(cid:173)
`vice and the power generator has resulted in a rugged sys(cid:173)
`tem having a compact footprint ( .... 0.6 m2) and an overall
`size comparable to that of a commercial excimer laser. The
`small-signal gain on the C-+ A transition of XeF has been
`measured to be 0.3% em- 1 at 488 nm, a value virtually
`identical to those reported for surface discharge systems of
`considerably higher power loadings.
`This paper is organized as follows. Section 2 describes
`in detail the design of the electrical system and laser head
`for the device. Two approaches to igniting the surface dis(cid:173)
`charge, surface feedthrough (in-line) and V configurations,
`are presented and compared. The experimental results on
`the XeF(C-+A), XeF(B-+X), and atomic iodine (1.315
`JJ-m) lasers are discussed in Sec. 3, and Sec. 4 summarizes
`the conclusions of this study.
`
`2 Surface Discharge System: Design and
`Performance
`
`2.1 Background
`Previous designs of surface discharge systems can be
`broadly classified into two groups. The first are those in
`which the discharge occurs on a dielectric surface such as
`polyethylene, Teflon, or a ceramic9•14- 17 and the distanced
`between consecutive electrodes ranges from several centi(cid:173)
`meters to beyond 10 em. The practical upper limit on the
`discharge gap is dictated by the breakdown voltage for a
`given surface and the spatial instability of the discharge,
`
`Fig. 1 Schematic diagram of the feedthrough (in-line) electrical con(cid:173)
`figuration for igniting and sustaining a surface discharge. The Mo
`pins are ballasted capacitively and resistively, and the charging volt(cid:173)
`age is typically 30 kV. The dielectric surfaces explored to date are
`machinable glass ceramic, BN, and Al20 3 and SiC-coated surfaces.
`
`which rises dramatically with increasing d. If no accommo(cid:173)
`dations are made for triggering the discharge or preionizing
`the gap, charging voltages of ;;::: 15-30 kV are required,
`even for small gaps (d=2-3 em), to obtain an adequate
`radiation temperature. Systems with long gain lengths, such
`as the devices of Refs. 14 -17, having active lengths of
`0. 7-1.9 m, rely on discharge sections operated in tandem.
`Each section, typically 8-12 em in length, requires lO kV
`when the entire discharge is triggered by a 50-kV pulse
`applied to a cable lying beneath the dielectric surface. The
`device of Ref. 16, for example, had a gain length of 1.9 m
`as a result of 16 individual discharge sections, each 12 em
`in length.
`The second general category of devices relies on dis(cid:173)
`charges produced on the surface of ferrite rods or slabs.
`Producing a stable, intense surface discharge with these
`materials requires forming a channel in the ferrite, either by
`exploding a wire on its surface20 or by micrornachining a
`narrow channel (""' 150 JJ-m) by laser ablation. 21 Such de(cid:173)
`vices typically involve rods tens of centimeters in length
`(76 em in Ref. 20) and driven by a capacitor charged to 40
`to 50 kV. To summarize, realizing in a surface discharge
`system the gain lengths normally required for high-power
`laser operation (tens of centimeters} requires a means for
`spatially confining the discharge. Long path lengths can be
`obtained with ferrites once a channel has been established
`on the surface, or shorter discharge sections ( ... 10 em) can
`be combined in tandem to yield active lengths beyond 1
`20-24
`m.
`
`2.2 Electrical Design: Power Generator and
`Discharge Ignition Configurations
`The approach adopted for the experiments reported here
`offers a novel and robust scheme for establishing surface
`discharges over paths of arbitrary length and to do so with(cid:173)
`out the need to switch each section independently. In fact,
`no high-current or high-voltage switches are used at all.
`Rather, the discharge functions as its own switch.
`A schematic diagram of the first electrode configuration
`investigated is shown in Fig. 1. The key elements of this
`surface discharge device are: a ceramic substrate that runs
`parallel to the discharge axis, a main electrode at each end,
`and a series of intermediate electrodes between the main
`electrodes which divide the discharge path into segments.
`The critical feature of this design is the latter-a series of
`resistively and capacitively-ballasted molybdenum pins,
`which serve to both establish and stabilize the surface dis-
`
`Optical Engineering, Vol. 42 No. 12. December 2003 3613
`
`Energetiq Ex. 2021, page 4 - IPR2015-01300, IPR2015-01303
`
`

`
`Knecht et al.: Optical pumping ...
`
`Fig. 3 Photograph of the surface discharge. The path length is
`48.5 em.
`
`waveform, recorded by a modified Rogowski coil for a
`charging voltage on the capacitor bank of 30 kV, is repre(cid:173)
`sentative of those observed throughout these experiments
`and shows that the current reaches its peak value of
`"""46 kA in =4 J.LS. The maximum current observed to date
`is >50 kA. Notice that the current waveform is essentially
`critically damped. indicating that the impedance of the ca(cid:173)
`pacitor bank nearly matches that of the surface discharge.
`This is a result of efforts from the earliest designs t<} mini(cid:173)
`mize the inductance of the power generator and laser head.
`Consequently, > 80% of the 2.5 kJ stored in the capacitor
`bank (for \-'= 30 kV) is deposited in the surface discharge
`in the first half cycle of the current waveform. For a power
`pulse having a temporal width of 5 J.LS (FWHM). this en(cid:173)
`ergy dissipation corresponds to a power deposition per unit
`length of discharge of ""'8 MW em-· 1 •
`The low overall inductance of the capacitor bank and
`electrical connections to the HV and ground electrodes
`manifested itself in an unusual way during the testing of
`this system. Early versions of the laser employed standard
`wire to connect the ballast circuitry to the intermediate
`electrodes. To facilitate the rapid assembly of the laser
`head, the wire was later replaced by short springs, but ex(cid:173)
`perience showed that it was necessary to carefully limit the
`number of turns in the coils. Otherwise, the inductive volt(cid:173)
`age drop produced across the spring resulted in arcing and
`severe damage to the coil and associated components.
`Although the primary purpose of the Mo pins is to en(cid:173)
`able the rapid formation of a surface discharge over a path
`of arbitrary length with a moderate voltage. the pin an-ay
`also serves to confine the discharge, thereby improving the
`shot-to-shot reproducibility with respect to previous surface
`discharge devices. Figure 3 is a photograph of the surface
`discharge viewed normal to the surface. Minor deviations
`from a straight path are evident, and the overall length of
`the surface discharge is 48.5 em.
`
`2.3 Laser Head: In-Line and V Configurations
`Side, frontal, and end-on views of the laser head with the
`in-line ballast pin array configuration described in the last
`section are illustrated schematically in Fig. 4. Each Mo pin
`was inserted into a cylindrical hole in the dielectric material
`and was mounted so that its tip was flush with the surface
`of the dielectric. For most of the experiments reported here.
`MACOR machinable ceramic served as the dielectric ma(cid:173)
`terial, but preliminary tests with pyrolytic boron nitride.
`alumina, and SiC-coated surfaces were also conducted. Be(cid:173)
`cause of its chemical stability in the presence of fluorine,
`Kynar® was the material from which the laser head was
`machined. The lower plate (not shown in Fig. 4) sealing the
`laser head was fabricated from a polycarbonate or alumi(cid:173)
`num. As noted earlier. both the main electrodes and the
`intermediate (pin) electrodes enter the laser head through
`the top and traverse the ceramic dielectric. The seal around
`the main electrodes consisted of a Kynar fitting having a
`
`Fig. 2 Oscillogram of the surface discharge current for a charging
`voltage of =30 kV. The horizontal (temporal) scale is 2 ,u.s/div, and
`the vertical scale is 20 kA/div.
`
`charge. All of the pins in the linear array were isolated from
`de ground by a !0-MO resistor and a 450-pF doorknob
`capacitor. For the first pin (trigger electrode), however, the
`capacitor was connected in series with the secondary of an
`air-core autotransformer, to which the trigger pulse was ap(cid:173)
`plied. The separation between the pins in the array, except
`that between the high-voltage (HV) and trigger electrodes
`was set at 12.7 mm (1/2 in.) to ensure that breakdown be(cid:173)
`tween two adjacent pins occurred at or below 30 kV when
`the gas pressure in the laser chamber was I atm. The gap
`between the HV electrode (left side, Fig. l) and the trigger
`electrode was set to 28 mm (designed to hold off 30 kV).
`Consequently. when the full supply voltage was impressed
`across the entire array (HV electrode to ground), break(cid:173)
`down does not occur in the absence of a trigger pulse.
`However, when a - 15 to - 18-k V trigger pulse is applied
`to the transformer primary, the voltage appearing on the
`firl't pin (trigger electrode) is -50 kV, and breakdown oc(cid:173)
`curs in the HV-electrode-trigger-electrode gap. As current
`begins to flow in this gap, the voltage at the trigger elec(cid:173)
`trode rises rapidly to a value slightly above 30 k V, owing to
`peaking effects. Since the second pin in the array is still at
`ground potential, the gap between the trigger electrode (pin
`I) and pin 2 now self-breaks. The remaining gaps in the
`array follow in quick succession, and within 5 J.LS of the
`application of the trigger pulse. a surface discharge is es(cid:173)
`tablished along the entire length of the array. A major ad(cid:173)
`vantage of this design is that the individual gaps are not
`triggered independently and, consequently, the firing jitter
`(with respect to the command trigger) is low. Furthermore.
`although negative trigger pulses were normally employed
`in the experiments described here, the use of an autotrans(cid:173)
`former reduces the sensitivity of the system to the polarity
`of the trigger (because of ringing in the secondary of the
`transformer), resulting i.n improved reliability and stability.
`Little of the energy stored in the capacitor bank (5.6 J-tF
`cf. Fig. l) is expended in forming the discharge, but the
`establishment of the surface discharge presents a low(cid:173)
`impedance path to ground. It is at this point that C 0 delivers
`virtually all of its stored energy to the discharge, resulting
`in rapid heating of the plasma and the generation of intense
`blackbody radiation. It should be mentioned that once the
`discharge is established, negligible current flows through
`the ballasting pins, which are etfectively decoupled from
`ground.
`Figure 2 is an oscillogram of the discharge current. ob(cid:173)
`tained with I atm of air in the laser head and a machinable
`glass ceramic (MACOR) serving as the dielectric. The
`
`3614 Optical Engineering. Vol. 42 No. 12, December 2003
`
`Energetiq Ex. 2021, page 5 - IPR2015-01300, IPR2015-01303
`
`

`
`Knecht et al.: Optical pumping ...
`
`IH
`
`~I
`
`Fig. 4 Drawings (to scale) of the laser head, showing side (top),
`front (lower left), and end-on views of the Mo pin array, the HV and
`ground feedthroughs (at the left and right sides of the side and fron(cid:173)
`tal drawings, respectively), and the dielectric itself. For clarity, the
`lower plate of the laser head has been removed.
`
`tube boss seal on one end and a Swagelok®-compatible
`seal on the other. Both of the main electrodes themselves
`were machined from molybdenum (chosen for the purpose
`of minimizing material ablation per shot) and attached to an
`aluminum shaft that provided direct electrical contact to the
`capacitor bank. A detailed diagram of the main electrode
`feedthrough is provided in Fig. 5. Denoted C0 in Fig. 1, the
`capacitor bank comprises 44low-inductance (equivalent se(cid:173)
`ries inductance ""'50 nH) capacitors having a rated voltage
`of 40 kV and connected in paralleL The entire bank is
`mounted in an AI and polycarbonate enclosure that is quite
`compact. When fully assembled, the capacitor bank and
`laser head together have a footprint of only 0.6 m2 ( 150
`X -40 cm2), comparable to that for a commercially avail(cid:173)
`able excimer laser.
`The major drawback of the in-line intermediate elec(cid:173)
`trode design is the failure mode. Particulates formed by the
`discharge tend to become lodged around the electrodes and
`gradually migrate into the cylindrical cavities in the ce(cid:173)
`ramic substrate. Consisting of ablated ceramic material and
`ash produced when the plasma comes in contact with Ky(cid:173)
`nar, the particulates are weakly conductive. The result is an
`alternate discharge path between the main electrodes and
`the ground plane on the exterior of the laser head, a path
`that is no less favorable (from an electrical perspective)
`than the one intended. The diversion of current into this
`auxiliary channel produces strong heating of the intermedi(cid:173)
`ate electrode insulating sleeve and the surrounding ceramic,
`often resulting in damage to both the ceramic and the laser
`head. Periodic cleaning of the laser-head surfaces amelio(cid:173)
`rates this problem, and installing a closed-loop gas recircu(cid:173)
`lation system to trap particulates should largely eliminate it.
`Nevertheless, in an effort to bypass this difficulty
`through electrode design, a second electrode configuration,
`known as the V, was designed and demonstrated. End-on
`and frontal views of the V configuration are illustrated in
`Fig. 6. One asset of this design is that, by bringing the
`intermediate electrodes to the dielectric surface at an ob-
`
`Aluminum oonductor
`
`Kynar 1i:ecldlroucll
`
`TctlonslecM:
`
`Fig. 5 Diagram (to scale) of the main electrode feedthrough, show(cid:173)
`ing end-on (left) and side views of the assembly.
`
`Fig. 6 Front (upper panel) and end-on (lower panel) views of the V
`configuration for the intermediate electrodes (molybdenum pins). A
`portion of the main electrode assembly and the bottom plate for the
`laser head have, for clarity, been removed. Note that the upper and
`lower panels are not drawn to the same scale.
`
`lique angle and from alternating sides of the laser chamber,
`the fabrication and installation complexities associated with
`introducing electrodes through the dielectric are eliminated,
`and replacing the dielectric is greatly simplified. Further(cid:173)
`more, the distance between the electrode shanks on a given
`side of the laser chamber is maximized. Although this ap(cid:173)
`proach eliminated the damaging misfires experienced with
`the in-line configuration, particulates deposited on the laser
`chamber wall caused occasional flashover between the in(cid:173)
`termediate electrodes. Our experience is that this drawback
`can be rectified by periodic cleaning and, presumably, a
`particulate trapping system.
`Most of the experimental results presented in Sec. 3 for
`the XeF(C-*A) and iodine 1.3-pm lasers were obtained
`with the in-line and V configurations, respectively.
`
`2.4 Erosion of the Dielectric Surface
`A critical parameter determining the ultimate utility of op(cid:173)
`tically pumping atomic or molecular lasers with surface
`discharges is the lifetime of the dielectric surface itself. For
`this reason, several measurements of the surface erosion
`rate were made with the system operating at full power (2.5
`kJ stored, V= 30 kV). Figure 7 is a profile of the surface of
`a machinable glass ceramic (MACOR) dielectric, recorded
`with a microstylus (Dektak) after more than 500 shots of
`the system on the same surface. Representative of other
`profiles obtained during these tests, the data of Fig. 7 show
`the nearly symmetrical channel to have a peak depth of
`=50 pm and a FWHM of -0.8 em. Thus, the average
`erosion rate is :50.1 p.m/shot. Other tests confirm the re(cid:173)
`sults of Fig. 7. Specifically, for a MACOR dielectric, which
`has a nominal composition25 of -46% Si02 , 16% Al20 3 ,
`and 17% MgO, the measured erosion rates ranged from
`0.06 to 0.3 p.m shot- 1, and varied by < 60% along the dis(cid:173)
`charge path (i.e., axially). The data suggest that the erosion
`rate is largest for the first few tens of shots on a new di(cid:173)
`electric surface but quickly falls to :50.1 p.mshot- 1 there(cid:173)
`after. Preliminary tests with alumina (Al20 3) and SiC(cid:173)
`coated surfaces indicate that they will exhibit erosion rates
`
`Optical Engineering, Vol. 42 No. 12, December 2003 3615
`
`Energetiq Ex. 2021, page 6 - IPR2015-01300, IPR2015-01303
`
`

`
`Knecht et al.: Optical pumping ...
`
`2cml
`~l
`
`J
`
`Fig. 7 Microstylus (Dektak) profile of a machinable glass ceramic
`dielectric surface after more than 500 shots of the laser system at
`full power. The FWHM of the eroded trench is -o.s em.
`
`approximately an order of magnitude smaller than these
`values, which will be required for an acceptable lifetime
`(>lOS shots for a single dielectric).
`
`2.5 Repetitively Pulsed Operation
`Although this system has been operated almost exclusively
`in the single-shot mode, it was designed and constructed so
`as to be capable of operation at pulse repetition frequencies
`(PRFs) of 1 to 5Hz. The issues that affect repetitive opera(cid:173)
`tion include rapid recharging of the capacitor bank, reliable
`holdoff and triggering characteristics, the continuous intro(cid:173)
`duction or recovery of laser fuel, clearing debris and par(cid:173)
`ticulates from the optical path and (as discussed earlier)
`electrode and dielectric surfaces, and maximizing the life of
`high-stress components. With respect to the former, the HV
`power supplies first used with this system · to charge the
`capacitor bank were a 30-kV, 10-mA system and, subse(cid:173)
`quently, an ALE model 152 L supply, rated at 40 kV and
`1.5 kJ s-t. The latter supply permitted recharging of the
`bank in less than 10 s, but introduced electrical transients
`that compromised the holdoff characteristics of the power
`generator and triggering circuitry. More recently, replacing
`the earlier power supplies with an ALE model 830 L unit
`( 40 k V, 8 kJ s- 1) has allowed for the capacitor bank, stor(cid:173)
`ing up to 4.5 kJ (for V=40 kV), to be charged in < 1 s. It
`should also be mentioned that the capacitors in the bank
`were rated conservatively. They were designed to operate
`up to 100 Hz and, largely because of the small degree of
`current reversal to which the capacitors are subjected dur(cid:173)
`ing each shot, have an expected survival of 90% after 108
`shots. Furthermore, tests indicate that prefiring of the sys(cid:173)
`tem is minimal for PRFs .eo; 5 Hz.
`A longitudinal flow system was also designed that al(cid:173)
`lows for replacing the gas mixture between shots. Although
`this subsystem has not as yet been tested, it provides for the
`removal of the fragments of the laser fuel molecular pre-
`
`3616 Optical Engineering, Vol. 42 No. 12, December 2003
`
`cursors and reconstitution of the gas mixture. For PRFs
`above ..., 5 Hz, transverse flow of the gas mixture will be
`required. In this regard, the laser head was designed to
`accommodate an array of gas flow ports, dispersed along
`the length of the surface discharge on both sides of the
`chamber, thereby allowing for gas flow along the coordi(cid:17