LA-7755-PRProgress Report

UC-21Issued: November 1980

Inertial Fusion Program

July 1—December 31, 1978

Roger B. Perkins and theLaser Fusion Program Staff

Compiled byFrederick Skoberne

- DISCLAIMER .

This book was prepared as an account of work sponsored by an agency o ( l h e United States Government.Neither the United Stales Government nor any agency thereof, nor any of iheir employees, makesanywarranty, express or implied, or assumes any legal liability or responsibility lor the accuracy,completeness, or usefulness of any information, apparatus, product, or process disclosed, orrepresents that its use would not ir' i ingc privately owned rights. Reference herein to any specificcommercial product, process, or service by trade name, trademark, manufacturer, or otherwise, docsnol necessarily constitute or imply its endorsement, recommendation, of favoring by ihe UnitedStates Government or any agencv thereof. The views and opinions of authors expressed herein do notnecessarily stale or reflect those of the United States Govern men t or any agency thereof.

OF TtiiS ESSliiSEHT IS MIL

CONTENTS

ABSTRACT 1SUMMARY 2

CO2 Laser Program 2CO2 Laser Technology 3Experiments, Diagnostics, and Military Applications 3Theoretical Support and Direction 4Laser Fusion Target Fabrication 5Applications of Laser Fusion Systems Studies 6

I. CO2 LASER PROGRAM 7Gemini System 7Helios System 8References 13

II. ANTARES—HIGH-ENERGY GAS LASER FACILITY 14Introduction 14Optical System 14Front-End System 16Power-Amplifier System 20Energy Storage System 23Target System 25Controls System 26HEGLF Site and Structures 30

III. CO2 LASER TECHNLOGY 32Propagation Studies 32Saturable-Absorbcr Recovery 35High-Efficiency Phase-Conjugate Reflection in Germanium and in Inverted CO, 37Calculation of Small-Signal Gain Coefficients in CO2 40Direct Measurement of Nonlinear Dielectric Properties in Germanium 41Subnanosecond Extinction of CO2 Laser Signals via Bulk

Photoionization in Germanium 43High-Pressure Enrichment of Light Isotopes 44References 45

IV. TARGET EXPERIMENTS, DIAGNOSTICS, AND MILITARY APPLICATIONS . . 47Experiments 47Diagnostic Development 57Military Applications 73References 80

V. LASER FUSION THEORY AND TARGET DESIGN 81Theoretical Support 81Target Design 89Code Development 89References 91

VI. LASER FUSION TARGET FABRICATION 93Introduction 93Target Fabrication 94Inorganic Coatings Development 98Organic Coatings Development 101Organometallic Coatings Development 105Polymer Foam Development 105Cryogenic Target Development !06References 108

VII. APPLICATIONS OF LASER FUSION—FEASIBILITY AND SYSTEMSSTUDIES 109

Reactor Design Studies 109Integrated Plant Design Studies 118Engineering Development Studies 121References 127

VIII. RESOURCES, FACILITIES, AND OPERATIONAL SAFETY 128Manpower Distribution 128Operational Safety 128

IX. PATENTS, PUBLICATIONS, AND PRESENTATIONS 129Patents 129Publications 129Presentations 130

VI

INERTIAL FUSION PROGRAMJuly 1—December 31, 1978

by

Roger B. Perkins and theLaser Fusion Program Staff

ABSTRACT

Progress at Los Alamos Scientific Laboratory (LASL) in the development of high-energy short-pulse CO2 laser systems for fusion research is reported. Improvements toLASL's two-beam system, Gemini, are outlined and experimental results are discussed.Our eight-beam system, Helios, was fired successfully on target for the first time, andbecame the world's most powerful gas laser for laser fusion studies. Work on Antares,our 100- to 200-TW target irradiation system, is summarized, indicating that designwork and building construction are 70 and 48% complete, respectively. A baselinedesign for automatic centering of laser beams onto the various relay mirrors and the op-tical design of the Antares front end are discussed.

Optical phase conjugation as a means of automatic target alignment is summarized,and we report on work with SF6-based gas isolators containing H2, which may lead toalmost total recovery of isolator absorption in a few tens of nanoseconds. In experi-ments with exploding-pusher targets, neutron yields exceeding 10" were obtained for thefirst time with CO2 radiation. Progress in the development of x-ray diagnostics is out-lined, including the fast high-voltage triggering of our x-ray streak camera. Ultravioletspectroscopy of highly ionized beams yielded results that had not been observedpreviously. Muitiburst simulation and opacity experiments, part of our modest militaryapplications effort, are described. Improvements in calculations! techniques, and resultsof studies on optical loading of targets and their energy absorption by shifted laser fociare presented. Significant problems in the fabrication of laser fusion targets culminatingin the successful manufacture of the 20-times-liquid-density target are outlined, and thedevelopment of a heat-transfer code that calculates the nonuniformity of DT ice layerson opaque inner target shells is discussed.

The results of various fusion reactor studies are summarized, as well as investigationsof synthetic-fuel production through application of fusion energy to hydrogen produc-tion by thermochemical water splitting. Studies on increased efficiency of energy extrac-tion in CO2 lasers and on lifetimes of cryogenic pellets in a reactor environment aresummarized, as well as the results of studies on pellet injection, tracking, and beamsynchronization.

SUMMARY(R. B. Perkins and Laser Fusion Staff)

The Laser Fusion Program at the Los Alamos Scien-tific Laboratory (LASL) is pursuing the dual goal ofdeveloping inertia! confinement fusion for commercialand military applications. It is essential for both goals toachieve scientific breakeven; that is, a fusion energy out-put that equals the laser energy incident on the target.For this purpose, we invented, and are developing, high-power short-pulse carbon-dioxide gas lasers that willprovide the efficiency and repetition-rate capability re-quired for use in commercial power plants. As our gaslaser systems become available, they are used intensivelyin a vigorous experimental program aimed at achievingthermonuclear burn of fuel pellets. This goal requiresthat we perform basic physics experiments to provide anunderstanding of the processes involved, and integralpellet burning experiments, which we expect to achievebreakeven by the mid-1980s. In our experimentalprogram, we are expending significant effort in targetdesign, target fabrication, laser facility support, anddiagnostics development. In addition, a modest ex-perimental effort directed toward military applications isunder way. Last, a systems group is exploring designconcepts for future commercial fusion and fusion-fissionhybrid reactor systems and subsystems to identify poten-tial problems.

CO2 LASER PROGRAM

Gemini

The Two-Beam Laser System, Gemini, was devotedprimarily to target experimentation and, to a smalldegree, to system characterization. During the past sixmonths, 170 target shots were fired, 73% of whichproduced the desired energy on target with nomeasurable prelasing. Some experiments were delayedbecause of problems with the pumping-chamber pulsers.The following salient features characterized the Geminioperations during the past six months.

• Encircled energy of the focal spot was measured.• The internal saturable - absorber gas cell was

received.• A single-target insertion mechanism was designed.• A new oscillator-preamplifier system was designed

to replace the existing front end.

Helios

In late July, we entered a new era in the laser fusionprogram when the Helios laser system was fired suc-cessfully on target for the first time. Although the initialenergy on target was limited to only 1.6 kJ because of atarget-amplifier parasitic mode, investigations into thenature of this mode enabled us to develop a means to in-crease the on-target energy to the 6-kJ level. At theselevels the Helios laser facility now ranks among theworld's most powerful laser systems, and is the world'smost powerful gas laser. Because the carbon dioxidelaser is considered by most people the most promisingcandidate for a laser-fusion-driven reactor, our recentHelios results will play a significant, if not critical, role indetermining the direction of this country's overall laserfusion program. The results are preliminary, but ex-perimental evidence of the first high-density core com-pressions has generated a good deal of excitementthroughout the laser fusion community.

Antares

Substantial progress was made in all areas of the An-tares project. Raising the prestressed concrete walls ofthe larger buildings that will house Antares and thetarget systems was the most obvious undertaking. Alllong-lead-time Antares laser components were releasedto vendors for fabrication in preparation for the start ofinstallation in August 1979.

A baseline design for automatic centering of the laserbeams on the power-amplifier and target-system relaymirrors was completed. We subsequently preparedvariations on the baseline system and an alternative con-cept to allow an assessment of the impact on cost andperformance.

The suggestion to sample the full system of the outputwith wire transmission gratings was accepted after testsof the gratings on Gemini. The gratings were fabricatedin-house on a winding machine designed to produce wirespark and drift chambers.

The optical design of the Antares front end was com-pleted. Preamplifiers were ordered and delivered, andone of the required six driver amplifiers was ordered forevaluation. A prototype of the Antares front end was

assembled in the old Single-Beam System area. Evalua-tion and development of multiline oscillators and mul-tipass amplifier geometries were planned for theprototype.

We reactivated the power-amplifier prototype toevaluate modifications in the Antares power-amplifierdesign that were a result of the prototype program.Reliability and control of the gridded cold-cathode gunwere improved. Stands, pressure vessels, and gunparts —all long-lead-time items—for the power-amplifierproper were promised for delivery in mid-1979.

A firm, fixed-price contract for the engineering designand fabrication of the Antares energy storage systemwas awarded. All prospective vendors were willing toquote on a fixed-price basis because a prototype unit hadbeen designed and tested by LASL. This prototype Marxunit met the specified inductance requirements as well asthe basic voltage and energy requirements. In addition,we developed a spark gap that transfers the charge andcurrent anticipated under fault conditions withoutfailure.

A design-construct contract for the Antares target-chamber and vacuum system was also awarded. A uni-que feature of this 13OO-m3 system was the requirementthat the pumping units be cryogenic pods. The contrac-tor was to proceed with the procurement of long-lead-time items before final design review.

A tree-structured hierarchical architecture wasdefined for the controls system. Detailed specificationsfor the control computers and the communicationsnetworks were released for competitive quotations.

Because construction was three months behindschedule, the Department of Energy negotiated with theContractor to obtain beneficial occupancy of the LaserHall by mid-July 1979.

CO2 LASER TECHNOLOGY

Optical phase conjugation continues to promise im-proved beam quality and, possibly, automatic targetalignment in laser fusion systems. The use of longer ger-manium samples permitted us to increase the observedefficiency of phase-conjugate wave generation from 2 to20%. Also, conjugation in inverted CO2 was seen for thefirst time. In a hybrid arrangement where the CO2

medium helped to amplify the conjugated wave, efficien-cies as high as 250% were measured without loss ofwave quality; the conjugated signal was stronger thanthe original aberrated signal

In related work designed to identify upper limits to im-proving phase-conjugation efficiency in germanium byusing higher 10-um intensities, we found that dense, bulkplasma forms in intrinsic and in optical-grade ger-manium above 200 MW/cm2 with associated free-carrierdensities as high as 2 X 10'Vcm3 without permanentmaterial damage. However, p-type germanium did notshow these effects, renewing our interest in doped ger-manium for conjugation.

In the gas-isolator development area, absorptionrecovery was studied to better understand the potentialimpact of these systems on the retropulse problem. Wefound that the recovery rate of the SF6-based mixturesdepends only weakly on the saturating wavelength, andslows with increasing optical fiuence. However, byadding H, to the isolator mixture, the absorption at somewavelengths recovered up to four times faster thanpreviously. The goal is to obtain almost total recovery ofisolator absorption in a few tens of nanoseconds.

Work on the reinjection laser prototype culminated indemonstration of reliable operation at pressures up to7600 torr. Results of a full characterization are reported,including an operation reliability level of 99.9S during a4000-shot sequence. This laser uses multiple passes togenerate energetic, subnanosecond pulses, and will beused in the upgraded Gigawatt Test Facility (GWTF)system.

A prototype of the GWTF was also characterized,revealing some discrepancies between predicted and ob-served propagation of subnanosecond multiline pulses.Although predicted and observed extracted energiesagreed well in these tests for 1.0- and 0.5-ns final-amplifier-input-pulse durations, the amplified pulse wassubstantially broader than expected in the latter case.

Investigations begun this period include the first directmeasurement of the nonlinear 10-um refractive index(n2) in germanium via ellipse rotation and determinationof an upper limit to n2 in NaCI.

EXPERIMENTS, DIAGNOSTICS, AND MILITARYAPPLICATIONS

Experiments

The Helios laser has delivered previously unattainableCO2 laser power and energy to various targets.Exploding-pusher targets served as a test of the newfacility. Neutron yields exceeding 10* were obtained forthe first time using CO2 radiation. Plastic-coated glass

microballoon (GMB) targets designed to achieve highfuel density were used in experiments aimed at targetsthat are direct predecessors to fusion reactor targets.These targets yielded sufficient neutrons for meaningfulparametric studies. High-density targets are difficult todiagnose, but initial experiments are providing informa-tion consistent with our code calculations, which inten-sifies our diagnostic development efforts. The measuredspeed of the fastest plasma protons at Helios gavefurther support to the expressions for hot-electron tem-perature scaling, (IA.2)1'3, where I is the ratio of focalpower vs unit area and X is the laser wavelength.

Diagnostics Development

The laser fusion program depends on our ability toprovide the instruments required. This is increasinglytrue with large complex laser facilities and complextargets.

During this reporting period, researchers tried to im-prove the time resolution of neutron detectors so as tomore clearly separate the effects of ion temperature,burn time, and run-in time on neutron spectra.

We have continued our vigorous program of x-raydiagnostics development. Advances during the past 6months clarified the triggering problems that haveplagued our x-ray streak camera work. We continuedfast high-voltage switching and triggering studies, whichshould be useful for various diagnostics. Opticaltelephotography with high resolution is evolving towardthe use of precisely timed optical probe beams.

Military Applications

Many feasibility studies were performed and a con-siderable amount of data was collected during the pasttwo years using the two-beam glass laser.

The work on vacuum uv spectroscopy of highlyionized metals has been concluded successfully, and weobtained good, usable spectra of 22 metals. Many of theresults had never before been observed.

By contrast, the GEAR x-ray streak camera con-tinued to have triggering problems in its internal cir-cuitry, and no satisfactory data have been obtained.

Our equation-of-state (EOS) studies continue alongtwo distinct lines: impedance matching and Shockwavestructure. We are using a fast, visible-light streak camerato observe the shock front emerging from the rear of

multistep targets. Because the EOS is known for the sub-strate, it can be deduced for the other layers. Work todate has confirmed the feasibility of this technique.

The work on multiburst simulation and blast wavesused two-wavelength, visible-light holographic inter-ferometry to observe one- and two-dimensional shocksin air at various pressures. Experimental results to datedo not agree completely with calculations; i.e., gas den-sities are in approximate agreement but electron densitiesare not.

In the opacity experiments, we made progress in ob-serving optical self-emission from the back of aluminumfoil targets simultaneously along a 30° and a normal lineof light. Analysis indicates a time-dependent anisotropyin this emission.

THEORETICAL SUPPORT AND DIRECTION

The final report of the DOE Ad Hoc Experts' Groupon Fusion (chaired by J. S. Foster) was issued in June1978. One recommendation was that DOE establish afull-spectrum design and theory effort at LASL to coverthe range of possible ICF (Inertial Confinement Fusion)drivers, rather than CO2 lasers only. We have presenteddetailed plans to the DOE Office of Laser Fusion, andanticipate that our effort will increase in that direction inboth depth and scope over the next year.

Significant work was performed in the four sections ofour theoretical group. Our unclassified target-design ef-fort concentrated on the two Helios experiments: the 20-Times-Liquid-Density Milestone experiment, also dis-cussed in preceding reports, and the Exploding-Pusherdesigns reported herein. An important achievement inour code development effort was the preparation of anoff-axis ray-trace procedure, which modifies previoustreatment that caused us to overemphasize the non-spherical optical loading and to overcorrect our two-dimensional designs. Our laser physics work shows im-proved calculations of the small-signal gain coefficient inCO2 lasers.

Our theoretical target design studies produced resultson the optical loading and absorption by shifted laserfoci and on an extension of the suprathermal electronscaling law; progress was also made in understandingthe coherent acceleration of hot electrons and their sub-sequent transport in the plasma. Finally, we presentcorrected calculations on the loss of fast ions from thetails of the fuel-ion distribution in low-pR pellets.

LASER FUSION TARGET FABRICATION Metal Coatings Development

General

Our ability to fabricate targets Tor laser fusion experi-ments has increased in several areas, culminating in thesuccessful manufacture of the 20-times-liquid-density(20XLD) target. In meeting this goal, we improved ourplastic-coating techniques and can now apply plastic(CH) coatings up to 350 um thick with a peak-to-valleysurface smoothness better than 1 urn to stalk-mountedGMBs. In addition, we were able to remove the coatedstalk, leaving an imperfection of only ~ I um insmoothness and uniformity at the stalk location. Also,we can coat about 100 unmounted, quality-selectedGMBs with up to 15 nm of very-smooth-surfaced plasticand recover essentially all the coated shells. Metalcoatings of gold and molybdenum up to 5 um thick wereapplied to levitated GMBs by sputtering. Both unifor-mity and surface smoothness are adequate for our20XLD target.

In cryogenics, we developed a heat-transfer computercode that calculates the nonuniformity of the DT icelayer in a target frozen by the fast-isothermal-freezing(FIF) technique. We are applying this code to targetswith opaque inner shells in which the DT layer cannot bemeasured directly.

Target Assembly Progress

We supplied more than 1000 targets to our threeoperating laser systems and made more than 100 partsfor diagnostic devices. About 320 targets werefabricated for Main-Sequence experiments. Another 340were supplied for a wide variety of Target-Essentials andSupport Physics experiments, and over 400 were madefor military applications experiments. We completed thedevelopment of fabrication techniques for 20XLDtargets to be tested on Helios.

During development of the 20XLD target, wedeveloped a micromachining technique Co remove thecoated stalk from the plastic-coated GMB and toremount the resulting spherical target. To fill these high-aspect-ratio glass shells with 30 atm of DT fuel gas, wemeasured the GMB crush strength and optimized ourpressure-staging procedures.

In preparation for fabricating the Rigel-B target(20XLD with a metal pusher layer) we developed techni-ques to coat small numbers of quality-selected GMBswith metal layers. We applied coalings up to 5 um thickof sputtered gold and molybdenum to GMBs levitatedby our gas-jet levitator. The surface smoothness of thesecoatings is better than I um peak-to-vallcy.

To improve the quality of all-metal pushers, we aretrying to make free-standing, high-quality metal shells bymetal-coating a removable spherical mandrel. Spheres ofpolymethyl methacrylatc (PMMA) made by our dropletgenerator technique were overcoated with ~2 urn ofCVD nickel and leached in solvent. Although some ofthe resulting hollow shells were of high quality, somecrumpled without the mandrel support.

Plastic-Coatings Development

We finally discovered how to plastic-coat a fewpreselected GMBs by low-pressure plasma polymeriza-tion without losing them or fusing them to the coater inthe process. Up to 15 um of poly p-xylenc was applied toas few as 8 GMBs. with 100% recovery and extremelysmooth surfaces. For the thicker coatings needed for the20XLD target, we applied poly p-xylylene by the vapor-phase pyrolysis technique in coatings up to 350 urnthick. Surface smoothness was improved by lowering thedeposition rate, applying a 20-nm-thick passivating layerof aluminum to the GMB before coating, or by adding acomonomer during deposition. Surfaces with < ±0.2-umpeak-to-valley variation are achieveable.

All experiments to produce small-cell-size, low-densityplastic foam failed to achieve one of these requirements.We are, therefore, reassessing foam fabrication methodsand investigating alternative buffer materials.

Organometallic Coatings Development

Because we cannot measure the uniformity of a DTice layer inside a metal pusher shell, we developed aheat-transfer computer code to calculate the freezingrate and DT shell nonuniformity for complex, multilayer,multishell targets. Calculated nonuniformities for glass

shell targets were in good agreement with experimentaldata, and we are trying to compare calculated and ex-perimental data for different configurations of glassshells.

Work on the cryogenic-target freezing apparatus forHelios is under way. The present retraction speed of theheat shield is 3 cm in 6.8 ms. which is probably fastenough, but more work is needed on the shock absorbersystem that stops the retracting shroud.

APPLICATIONS OF LASER FUSION SYSTEMSSTUDIES

In our reactor design studies we programmed themathematical description of the new plasma modelrepresenting the magnetically protected reactor cavityfor use on the CRAY computer. We also investigatedmethods of removing the significant amount of fuel-pelletmicroexplosion energy by means of magnetic fields in amagnetohydrodynainic (MHD) energy converter thatdecelerates the ions without suffering excessive wall ero-sion and even generates electric energy in the process.

The concept of absorbing the neutron energy at hightemperature in a blanket of boiling lithium-containingmetal was studied further and led to conceptual solutionsto the production of high-temperature process heat formore efficient energy conversion.

Our studies of proliferation-resistant fuel cycles, basedon laser-driven ICF reactor technology, focused onquasisymbioiic concepts that maximize fissile-fuelproduction and minimize heat production.

Hydrogen production studies continued. They arecoordinated with a parallel study, supported by the Of-fice of Fusion Energy, of thermochemical hydrogenproduction. As part of this effort, we are also engaged inprocess design of a bismuth-oxide electrother-mochemical cycle.

Finally, in engineering development studies that attestto the impact of advanced technologies on long-life,high-repetition-rate ICF systems, we investigated meansof increasing the efficiency of energy extraction in CO,lasers and calculated the lifetime of cryogenic pellets in areactor environment. Results of a study performed for usby United Technologies Research Center on pellet injec-tion, tracking, and laser beam synchronization are alsoreported.

I. CO, LASER PROGRAM(G. Schappert)

Research and development programs on high-energy short-pulse CO : lasers began atLASL in 1969. The first system, the Single-Beam System, was designed in 1971. beganoperation in 1973. and was phased out in November 1977. Two large systems nowoperating are the Two-Beam System called Gemini, and the Eight-Beam System calledHelios. Target experimentation continued on Gemini, which will ultimately generatepulses of 2 to 4 TW for target-irradiation experiments. Helios became operational inApril 1978. and surpassed the design goal on June 21, 1978. with an output of 10 kJ ata power level of ~20 TW. The third system. Antares. is in the design and prototypestage. This system, described separately in Sec. II. will generate laser pulses of 100 to200 TW, with the objective of demonstrating scientific breakeven.

GEMINI SYSTEM

Introduction (J. P. Carpenter)

The Gemini System was devoted primarily lo targetexperimentation and. to a small degree, to systemcharacterization. During the past 6 months. 170 targetshots were fired and 73% of those produced the desiredenergy on target with no measurable prelasing. Some ex-periments were delayed by problems with the pumpingchamber pulscrs.

Laser Performance and Diagnostics (J. P. Carpenter, J.J. Hayden. J. McLeod)

Measurements of encircled energy at the focal spotshowed that 70% of the total beam energy passedthrough a 100-um-diam pinhole. The Strehl ratio of~0.25 derived from these data should improve substan-tially with the addition of adaptive optics to the triple-pass amplifiers. The adaptive-optics contract withHughes Research Laboratories entered the finalhardware phase.

The internal saturable-absorber gas-cell hardware wasreceived and inspected. This cell will house the defor-maUe recollimating mirror and should allow the deliveryof ~1 TW to a target.

A target insertion mechanism that holds one target ata time was designed and should be installed during thesecond quarter of FY 79. This insertion system will per-

mit an unlimited number of target shots before breakingthe vacuum in the larget chamber. The present targetwheel holds seven targets, bat the vacuum must bebroken each time a new wheel is inserted.

The optical diagnostics facility can now measure preand postpulsing power to 10 ' of the main pulse peak.

Oscillator Preamplifier System (P. Goldstone. V.Romero)

A new oscillator preamplifier system, to be installedearly in FY 80. was designed to replace the existing frontend. A smoothing tube in addition to a switched out os-cillator, similar to that in the Helios system, will be used.The new front end will be installed directly east of thedual-beam module. Downtime will be minimal becausethe present front end will remain operational until thenew system is installed. The new oscillator-preamplifierroom is shown in Fig. 1-1.

Computer and Control System (S. Hackenberry, P.Castine)

A computer simulation of our pulse-power capacitorbanks was started. This computer model, which shouldsimplify troubleshooting the pulse-power systems, usesthe NET 2 electronics code in the LASL Central Com-puting Facility.

14 ft (cl««f«nc« 14 ft 4 in.)

GEMINI FRONT END

Fig. 1-1.Gemini front end.

HELIOS SYSTEM

Introduction (J. S. Ladish)

The second half of 1978 marked the transition of theHelios Laser Fusion Facility from construction to thetarget-irradiation phase. After successfully firing thelaser in excess of the design output of 10 kJ intocalorimeters in June 1978 (FIST II), we decided to per-form the Earliest Laser On-Target (ELOT) experimentas soon as possible to gain experience and to becomefamiliar with the problems to be encountered in target-irradiation tests. On July 27, 1978, ELOT was per-formed at an on-target energy of 1.6 kJ and peak inten-sity of just over 2 TW. The energy on target in that ex-periment was limited because of a parasitic oscillation

mode caused by target feedback coupling to the am-N plifier gain medium.

The discovery of the parasitic mode triggered a con-siderable effort to determine the feedback source at theamplifier end, which would lead to its elimination. Amore suitable gas mix (Mix 907) was subsequentlydeveloped for the amplifier saturable-absorber cell, andthe entrance beam tube used in the power amplifier wasremoved. Also, we discovered a small, cone-shapeddepression, made by micromachining, at the center ofthe large micromachined recollimating mirror. We usedNextel paint to cover the depression because the coneshape could act as a direct retroreflector, which wouldproduce parasitic paths involving the power-amplifiergain medium and the target and/or diagnostics locatednear the target. These modifications raised the on-targetenergy to ~6 kJ, equal to a peak power of ~13 TW.

Although Helios is still in the shakedown phase, ourtarget experimentation has already provided valuablenew information regarding laser plasma interaction, andhas produced the first clear evidence of high-densitytarget implosions.

Target experimentation provided the principal excite-ment during the second half of 1978, but most of our ef-fort on Helios centered on many less exciting but equallyessential tasks, which are described below.

Front End (R. Carlson, R. Quicksilver, M. Weber)

The eventual requirement on beam simultaneity inHelios has been set at ± 1 cm (±33.3 ps). In preparationfor ELOT. the individual beam paths of the eight Heliosbeams were adjusted for a beam simultaneity of ±50 ps,using a method similar to that used in the Geminisystem. The time of flight (TOF) of a front-end pulse tothe target was measured by the time difference betweenthe arrival of an optical pulse on a fast pyroelectricdetector (Molectron P5-00) referenced against the laser-triggered spark-gap pulse (LTSG) of the three-stagePockels-cell system. The simultaneity measurementswere made with the LASL-built 3-GHz oscilloscope torecord the optical and the LTSG pulses; a sweep calibra-tion trace was added to each record to permit future tem-poral deconvolution. About 1300 photographs weretaken yielding 200 records that met the pulse amplituderequirements. The final data after several iterations ofpath-length adjustment are presented in Table II .

TABLE II

BEAM SIMULTANEITYDATA FOR HELIOS

Path DifferencePrimary Beams from Mean (ps)

1A.IB.3B.3A.

2B2A4A4B

+51-24

^

-23

±±±+

40334361

This method of beam simultaneity measurement istedious and can cause several systematic errors.Therefore, we developed, prototyped, and tested an im-proved technique that enables us to perform routinemeasurements of beam simultaneity with a resolution ofless than ±5 ps. Briefly, the cw (continuous wave) CO,alignment laser in the front end will be modulatedelectro-optically at 40 MHz and aligned onto a surrogatereflecting sphere located in the target chamber. Thereflected return signal will be detected by a HgCdTedetector located in the front end. The phase of this returnsignal will be compared continuously to the phase of theelectrical signal that drives the modulator; this phase dif-ference is proportional to the path difference. The systemis shown schematically in Fig. 1-2. All equipment for thissystem is on hand, and installation is ~30% complete.

As part of our continuing improvements to the Heliosfacility, we installed the following components:

• An improved oil-filled TEA oscillator and itsassociated Invar-stabilized optical support struc-ture;

• Four motorized mirror slides at Spatial Filter 2 topermit repeatable injection of the krypton-ion visi-ble alignment laser; and

• A 16-port voltage divider that provides an LTSGpulse to Gallery West (our diagnostics setup) and tothe target diagnostics room for timing purposes.

Laser Physics (G. T. Schappert. J. S. Ladish, D. Casper-son, R. F. Haglund)

The first eight-beam target shots on Helios were per-formed during the second half of 1978. On July 27. theentire system was fired for the ELOT test at reducedPFN voltage (42 kV/stage) with 6 torr of Mix 804 ineach saturable-absorber cell. The total energy output onthese shots was 1.6 kJ and was limited because ofparasitic instabilities brought about when the entiresystem was carefully aligned onto a GMB target. An in-side view of the Helios target chamber is shown in Fig. I-3.

After these first successful low-energy eight-beamshots, we performed a systematic study of the parasiticthresholds with targets in place. We used several techni-ques to identify the transitions involved in the self-lasingfrom a target. An Optical Engineering spectrometer, and

Cd T* ModulatorStructure

HPB640BRF

Source

BcamSplitter

^-Optical PolSurrogate

Sphere

Analyzer

Foeutmg Ltnt

'HgCdTe Detector

DC Bios)

ACPrt-Aimpl

HP8447F

Reference Signal

Vector VoltmeterHP8405A

Fig. 1-2.Beam simultaneity schematic.

Fig. 1-3.Inside view of Helios target chamber.

later, 9- and 10-u.m bandpass filters were placed atGallery West to detect any energy due to self-lasing.However, these efforts were unsuccessful because ofalignment problems and because of the low energyavailable at Gallery West. The most useful techniquewas to use various absorber gases in the amplifier sidearm and in the saturable-absorber cell to differentiallyquench 9- and 10-um light. Mix 804 did not perform wellat 9 urn and was replaced with Mix 907.

With this new mix, the stable pump voltage attainableon the power modules was raised from 46 to 50 kV, witha corresponding increase in energy from ~200 to ~500 Jper beam.

To determine the sources of parasitic instabilities in amodule, we disassembled the saturable-absorber cell ofAmplifier 1A and removed the reentrant beam tube. Thistube guides the oscillator pulse through the absorber cellwithout attenuation, but may offer a partially reflectingsurface that, in combination with the target, can produceself-lasing. In addition to having the input-beam tuberemoved, Amplifier 1A was also corrected for a machin-ing fault in its 40.5-cm (16-in.) recollimating mirror. Thesingle-point diamond turning process leaves a smallcone-shaped depression at the very center of the mirror,which can reflect light at angles other than those inten-ded. The fault was corrected by painting the center withNextel paint.

Subsequent threshold measurements with Amplifier1A aligned onto a target indicated that the amplifier wasstable at 54 k V, with only 11 torr of Mix 907 in the ab-sorber cell. Energy extraction measurements later

showed that the energy output of this amplifier underthese conditions is ~800 J. The same correctiveprocedure was then carried out on the remaining poweramplifiers. With these modifications and with Mix 907 inthe saturable-absorber cells, Helios can stably producemore than 6 kJ on target.

After retropulse damage was observed on the 10-cm(4-in.)-diam NaC! windows that separate the laser gasmix in each amplifier from the 3 torr of air in thesidearm, a Mylar transporter system was installed inModule 3A, which placed a "sacrificial" 1/4-mil film ofMylar over this window. The forward pulse passedthrough the Mylar after the first pass of the triple-passamplifier at an energy density of ~0.1 J/cm2 with onlyslight attenuation. The retropulse, however, struck thewindow at >5 J/cm2 and created a plasma that blockedmost of the retropulse from reaching the 10-cm (4-in.)NaCl window. The window has remained undamaged af-ter ~15 shots since installation of the transporter.Similar systems have been installed in all modules.

Other modifications include the replacement of thestraight amplifier anodes with curved anodes, whichshow improved gain uniformity.

Controls (E. L. Jolly, M. D. Thomason, J. Sutton, D.Remington, F. D. Wells, W. Hanna, K. M. Spencer, L.Sanders)

The Helios master control program evolved into athree-task multiple overlay (three overlay areas) withalmost 28 k of core storage on the Data General EclipseS/200 control computer, using the latest revision of DataGeneral's operating system, QDOS, Rev. 6.32. Recentchanges to the control program were the addition ofabsorber-gas pressure-change routines and the replace-ment of the monitor task with a more efficient version.'

Work on supporting stand-alone programs is sum-marized as follows.

• The saturable-absorber control program was writ-ten.

• Studies continued of the new Asynchronous LineMultiplexor (ALM) and of its interface with theremote microprocessor substation, to be used forbeam diagnostics. Some subroutines to support theALM were written and tested. The ALM will beused to communicate with several microcomputerswithin the Helios facility and with the PDP 11/70computer of the Target Diagnostics Group.

10

• The front-end checking program was modified toaccommodate changes in the front-end hardware.

• The support programs for Helios, CRASHSV andTAPESEV. are being updated. CRASHSV waswritten to allow data retrieval from the digitizers inthe event of a system crash at shot time, whereasTAPESV allows data retrieval from the on-linetarget if data are lost from the disk. Both supportprograms were modified to update the operationslog file.

• The front-end spectral-line-content program LINESwas used in conjunction with laser system firings byrunning the background terminal. This program willsoon be incorporated into Helios.

• Testing of the prototype microprocessor-basedtarget-chamber mirror-position encoding systemrevealed problems in both software and hardware.The system works well with a smoothly rotatingmotor driving the mirror mount, but does notalways count correctly when a stepping-motordriver is used. Several test programs were written todiagnose the problem; some progress has beenmade.1

• The automatic alignment program was completedwith the addition of a subroutine to operate the cwCO2 alignment laser. This program is usedroutinely.

To carry out the necessary microprocessor program-ming, we developed the capability to assemble programsand to encode them into programmable read-onlymemories (PROM) automatically. A cross assembler(which allows us to assemble Intel 8085 assemblylanguage programs) was obtained from Boston Systems,Inc., and installed on the Eclipse S/200. A program toencode automatically from the cross-assembler outputwas developed. Object files from the disk are sent acrossour Asynchronous Line Multiplexer (ALM) to aborrowed portable PROM programmer. AnotherPROM programmer was ordered for exclusive use onHelios.

The second computer processor for Helios arrived inearly December 1978. Installation is about half completeand proceeding well. No computer downtime is an-ticipated as a result of the installation.

In addition to these tasks, the following hardware con-trol systems were completed, installed, and madeoperational.

• The target-chamber vacuum control• The target three-axis manipulator control•» The building safety and warning control

• The saturable-absorber-gas control• The amplifier side-arm pressure control• The motorized mirror-mount beam-positioner con-

trolWe designed and fabricated an optical isolation

scheme to eliminate all wired inputs to the shielded con-trol computer room. A total of 124 digital lines and 8analog lines will be transmitted optically through thescreen-room/control-room interface.

Mechanical Assembly and Engineering (E. L. Zimmer-mann, B. Maestas, D. Martinez, J. Valencia, L.Rodriguez)

Although mechanical assembly was completed sometime ago, Helios upgrading and maintenance work con-tinued.

Several titanium foils (windows) separating theevacuated and pressurized sections in the power modulesdeveloped pin-hole leaks and were replaced. Data ac-cumulated indicate an average window life of ~75 shotswith a typical foil-replacement time of less than a day.Although the phenomenon of pin-hole leak formationwas investigated in some detail, no satisfactory explana-tion has been found. Helios operations are not affectedtoo severely.

The following steps were taken to ensure a constantavailable supply of foil windows.

• A contract was awarded to produce acceptable win-dow assemblies routinely.

• A metal-fabrication group at LASL was funded toproduce foil-window assemblies (at present as theprimary source).

• A mill run of titanium foil was purchased to ensurean adequate supply.

• An inventory of at least ten complete assemblies ismaintained.

The new curved anodes are showing signs of fatigue-cracking in the weld joints, and an investigation of possi-ble solutions to the problem is under way.

The electron-beam vacuum metal bellows are begin-ning to crack from fatigue and will be replaced soon withstainless steel bellows.

Miscellaneous tasks that were completed include• installation of safety chains, work platforms, and

railings on the power modules;• installation of rupture disks on two of the amplifier

modules;

11

• design of safety covers for the amplifier output saltwindows; and

• design of oil troughs for the amplifier modules toeliminate oil overflow.

Concrete pads were poured to support the temporarymodular building, which will be used until the permanentHelios addition is completed. The vestibules for thesouth doors were completed. Plans for the permanentaddition are nearly complete, with groundbreakingscheduled for April 1979.

Two l/24th-scale Helios models were completed bythe Model Shop and displayed at several meetings. Onemodel will be housed in LASL's Science museum.

Helios Laser Beam Diagnostics (I. Bigio, S. Jackson, R.Ainsworth, C. Smith, A. Laird)

A 1-W krypton-ion laser was installed in the front endas an alignment aid for Gallery West. Alignment time re-quired by Gallery West for preparing a shot sequencewas thereby reduced to less than one hour.

Initial optical problems were resolved by the im-plementation of CaF, beamsplitters and of a redesignedoptical layout to collimate the sample beams to theshielded screen room. Studies have shown that carefulplacement of lead shielding is required to protect theliquid-helium-cooled prepulse detectors from x rays. Atthe end of 1978, the electrical problems were beingsolved and the last of the beam lines were madeoperational. Preliminary plans for an automatic align-ment system were initiated.

The following laser beam parameters can now bemonitored.

• Total-energy—all eight beams on target• "3Q" ratio (ratio of prepulse: main-pulse: post-

pulse energy)—one beam only• Temporal pulse shape—one beam only• Front-end energies—all four beams• Front-end spectral contentIn addition, both 8085 microcomputer substations

became operational, and the optical fiber links from thesubstations to the main computer were installed and arealso operational. The final link to the ALM remained tobe made, but the remote substations can be operatedfrom the main control room.

Optical Systems (J. Hanlon, V. K. Viswanathan, M. D.Bausman, J. J. Hayden, J. Murphy, P. Bolen, R. Parnell,I. Liberman)

The optical systems tasks fall into four majorcategories: day-to-day alignment and maintenance of theHelios laser alignment system; evaluation of adaptiveoptics; design and manufacture of the ir microscope; andsalt-window control, recoating, and refurbishing.

Day-to-day alignment and maintenance tasks have in-cluded upkeep of the automatic alignment system; align-ing the laser for firing; installing the remote-controlmotorized mounts for the orthogonal alignmenttelescopes inside the target chamber; changing optics; in-itiating the design of new holders for the Hartmannmirrors; initiating the design of covers for the power-amplifier windows; transferring maintenance of theHartmann cameras from EG&G to L Division; remov-ing all beam tubes; reestablishing the beam lines for allamplifiers; cleaning mirrors; and making and installingnew masks for the 7.1-cm (3-in.) mirrors inside thepower amplifiers. Attempts to automate Spatial Filter 4were not successful; this task should be completed earlyin 1979.

Primary efforts in the Rocketdyne and Hughes adap-tive optics contract centered on liaison and on themanufacture and use of a CO2 Smartt interferometer toevaluate the adaptive optics. Preparations for testingthese optics in our optics laboratory were made.

The overall mechanical design of the ir microscope iscomplete, but some detailing of parts remains to be done.The lenses for the system were designed and ordered. Amajor future task will focus on establishing and design-ing an evaluation system for processing the relayed im-age.

Five coated salt windows for the Helios system werereceived. These windows were coated on one side onlyand will be used on the target chamber. One new windowwas installed; recordings of the pulse shape indicate thatthe second surface reflection has been eliminated bycoating, as expected. The worst of the existing windowswill be replaced as AR-coated salts arrive. The presentsalt windows will be gradually replaced with coatedones. Windows that can be refurbished will be returnedto Harshaw for repolishing.

12

Monitoring, documenting, and coordinating the ship-ping, refurbishing, and coating of salt windows remainedan important task.

Experimental Plan 13 (Target Alignment Accuracy inHelios) was carried out successfully. As the principalresult, we established that the average pointing error forthe GMBs tested is 34 um. By improving the tests, weexpect to reduce this error to less than 25 nm, as re-quired.

An important modification made to the opticaltransport code LOTS enables us to routinely calculatethe interferogram produced by an arbitrary opticalsystem in conjuction with a Twyman-Green or Smarttinterferometer. This important accomplishment permitsus to compare the observed and calculated inter-ferograms and allows us to make specific statementsabout system alignment and about the sensitivity of thewavefront error affected by individual optical compo-nents. Applications of this powerful code to the Helios

system are being considered. Experimental verificationof its prediction capability and a general description ofthe work was presented at the Optical Society Con-ference in San Francisco.2

An ir interferometer was constructed in-house and isbeing used routinely in conjunction with LOTS.

REFERENCES

1. F. D. Wells and D. Remington, "Mirror Position Dis-play Equipment for the Target Chamber MirrorMounts of the Helios Laser Fusion ResearchFacility." Proc. CUBE Symposium, Los Alamos,1978.

2. B. D. Seery, "Design and Assembly of Carbon Diox-ide Laser Systems Using IR Interferometry," presen-ted at the Optical Society Conference, San Francisco,1978.

13

II. ANTARES—HIGH-ENERGY GAS LASER FACILITY(T. F. Stratton)

Antares and associated facilities are being constructed to provide a CO2 laser fusionsystem with an optical output of 100 to 200 TW with a maximum energy of 100 kJ. Webelieve that scientific breakeven is within reach of this machine. All laser-system sectionwork proceeded on schedule toward the 1983 completion date.

INTRODUCTION (T. F. Stratton, J. Jansen)

Substantial progress was made in all areas of the pro-ject. Raising the mostly prestressed concrete walls of thelarger buildings that will house Antares and the targetsystems was the most eye-catching undertaking; all long-lead-time Antares laser components were released tovendors for fabrication in preparation for the start of in-stallation in August 1979.

A baseline design for automatic centering of the laserbeams on the power amplifier and target-system relaymirrors was completed. We prepared variations of thebaseline system and an alternate concept to allow assess-ments of the effect on cost and performance.

The suggestion to sample the full system output withwire transmission gratings was accepted after tests of thegratings on Gemini. The gratings were fabricated in-house on a winding machine designed to produce wirespark and drift chambers.

The optical design of the Antares front end was com-pleted. Preamplifiers were ordered and delivered, andone of the required six driver amplifiers was ordered forevaluation. A prototype of the Antares front end was setup in the old Single-Eeam System area. Evaluation anddevelopment of mul'.iline oscillators and multipass am-plifier geometries v/ere planned for the prototype.

We reactivated the power-amplifier prototype toevaluate modifications incorporated in the Antarespower-amplifier design as a result of the prototypeprogram. Reliability and control of the gridded cold-cathode gun were improved. Stands, pressure vessels,and gun parts —all long-lead-time items—for the power-amplifier proper were to be delivered in mid-1979.

A firm, fixed-price contract for the engineering designand fabrication of the Antares energy storage systemwas placed with Maxwell Laboratories. All prospectivevendors were willing to quote on a fixed-price basisbecause a prototype unit had been designed and testedby LASL. This prototype Marx unit met the specified in-ductance requirements, which was a primary uncer-

tainty, as well as the basic voltage and energy require-ments. In addition, we developed a spark gap thattransfers the charge and current anticipated under faultconditions without failure.

A design-construct contract for the Antares target-chamber and vacuum system was awarded to thePittsburgh-Des Moines Steel Company (PDM). A uni-que feature of this 13OO-m3 system was the requirementthat the pumping units be cryogenic pods. We advisedPDM to proceed with the procurement of long-lead-timeitems before the final design review.

A tree-structured hierarchical architecture wasdefined for the controls system. Based on this definition,we released detailed specifications for the control com-puters and the communications networks for competitivequotations. Digital Equipment Corp. was the successfulbidder.

Finally, because facility construction was threemonths behind schedule, the Department of Energynegotiated with the Contractor to obtain beneficial oc-cupancy of the Laser Hall, the most critical part of thefacility, by mid-July 1979.

OPTICAL SYSTEM (A. Saxman, D. Blevins, W. Miller,J. L. Munroe, W. H. Reichelt, C. Silvernail, J. Sollid, T.Swann, W. Sweatt, P. Wolfe)

Introduction

The Antares beam-alignment and diagnostic concep-tual designs and analyses are partially completed. Theoptical-mechanical end-to-end system performance re-quirements have been specified. The initial analysis ofsubsystems indicated that the required hardware isavailable commercially. The bigger part of the initialanalysis has involved choosing cost-effective hardwareto modify and integrate into the Antares design.

Prototype beam alignment, beam diagnostics, andmirror positioners are in the initial design phase, with

14

some of the follow-on initial prototypes being tested,such as centering detectors, power detectors, signal-conditioning electronics, and large-mirror positioners.Final designs of the support structure for the beam-alignment and diagnostics package are being incor-porated into the power-amplifier and target-system de-signs.

Beam Alignment

A number of beam-alignment tasks were completedsuccessfully and new efforts were started. Descriptionsof some of these tasks follow.

Hughes Research Laboratories Study Contract.Phase II of the Hughes Research Laboratories (HRL)study contract for a conceptual baseline design of a com-plete beam-alignment system for the Antares power-amplifier system was completed. The task had thefollowing engineering objectives: to ensure that an align-ment beam entering the power amplifier is colinear withthe output beam of the front-end driver amplifier; that itis properly centered and pointed into the power am-plifier; that it remains aligned as it passes throughvarious elements, including all mirrors, windows, andspatial filters; and to ensure that the beam exiting thepower amplifier reaches the target area withoutvignetting. The equipment proposed in the baselinedesign would allow automatic alignment of each of the72 optical paths through the 6 power amplifiers betweenthe front-end output and the back reflector of the poweramplifier. The details of the proposed flip-in detectorsystem were outlined in the last semiannual report, LA-7587-PR.

Baseline Design Effort and Analysis. Using the basicpower-amplifier beam-alignment conceptual designgenerated by HRL and LASL as a baseline design, ourbeam-alignment group initiated a design and analysis ef-fort to simplify the mechanical, electrical, and electro-optical designs of the alignment laser source and itsbeam sector insertion device, the beam-centering andpositioning sensor package, beam sensor positioningdevices, and the general approach of maintaining theprecise placement of components and devices. Thesedesign considerations greatly reduced the projected costof the system hardware and its integration as comparedto an estimate generated for the baseline concept.

Alternative Beam-Alignment Technique. Wegenerated another beam-alignment technique for the en-tire Antares system. The conceptual design is referred toas a "See-Through Imaging System," in which each ofthe 72 optical beam paths from the front end to thetarget is manually or automatically aligned. The basicsystem uses a visible imaging system positioned at theoutput of each driver amplifier. The system colinearizesits optical axis with the driver-amplifier output beam andthen properly centers each component of the entire op-tical train through to and including the target. Studiesbegan on a number of technical areas for this system,such as solid-angle constraints, reflection losses from thediamond-turned mirrors, various combinations of visibleimaging camera systems, optics, and image ccntroid-determining hardware and software packages.

Conceptual Beam Alignment. A conceptual beam-alignment design effort for the target system withHughes Aircraft Corporation (HAC), was initiated andwill continue through March 1979. The basic tasks in-volve upgrading an analytical model to describe thetarget system optics, performing a sensitivity analysis ofthe Antares optical train, verifying our analysis that in-dicated noninterference of the target-system beams andoptical elements, developing conceptual alignmentschemes for the target-system optics, selecting an overallbaseline alignment scheme, and identifying long-lead-time items required for a prototype single-beam align-ment subsystem for the entire optical train.

Beam Diagnostics

Design efforts were started on the diagnostic packagefor the Antares system as referenced in the last semian-nual report (LA-7587-PR).

Wire Diffraction Grating. The final design of the wirediffraction grating and support structures was com-pleted, and in-house fabrication was initiated. A develop-ment prototype wire grating was tested successfully onGemini. The grating sustained fluxes exceeding 3 J/cm2

without damage and sampled the total beam energy aspredicted.

Spatiil-Fater/Thernul-Detector Testing. Initial testingof the spatial-filter/thermal-detector assembly was com-pleted. The net repositional error product (REP) of beam

15

displacement and energy was 5 mmmJ; that is, with 1mJ of total energy on the detector, it resolved the beamposition accurately, to within 5 mm. Typically, we an-ticipate a total energy deposition of 150 mJ, which im-plies a resolution of 0.04 mm under Antares conditions.

Input Diagnostics Package Designs. Design startedon the retropulse and power-amplifier input diagnosticpackages, the outpui-beam sampling calorimeters for theturning chambers that collect the 12-sectored, high-power, wire-grating-samplfd beams from the power am-plifier: and the diagnostic spool calorimeter that collectsthe special-purpose full output of a power amplifier.

A beam diagnostic and alignment test facility was par-tially constructed to evaluate the detector packages andthe CAMAC/computer data collection and analysissystem.

Window and Mirror Fabrication

Harshaw Chemical Company (HCC) produced theirfirst three successful forgings of Antares windows. Theproduction-process checkout remained promising. If thewhole production can be judged by these results, nodelay in meeting schedule requirements is anticipated.Recent HCC in-house efforts in Kyropoulos crystalgrowth were encouraging. Boules grown by this techni-que had minimum haze and sparkle. In fact, HCCdelayed the procurement of the last four Antares growthfurnaces until the results of this new growth techniqueare evaluated. The salt window and cell pressure testswere concluded successfully. Mechanical and opticalparameters were well within tolerances up to theoperating pressure, and no plastic flow of the salt wasdetected after a 3 5-day pressure soak.

Orders for 82 periscope substrates and 48 rear reflec-tor substrates were placed. The polyhedron beam reflec-tor design was reviewed and the first attempt to machineit was scheduled for April 1979. A preliminary analyticaldesign for the parabolic-mirror fixture was completed.

The Y-12 Plant of Union Carbide Corporationdemonstrated repeatable surface figures of less than1000 nm rms at 10.6 urn and surface finishes of less than500 nm peak-to-valley. These values are adequate for theAntares system.

Mirror Positioners and Mirror Cells

All the parts for the mirror positioners were received.Tests on the motorized positioner continued. The fine-screw actuator responded and repeated reliably; thecoarse-screw actuator needed minor rework. The Uni-versity of Tennessee made a detailed deflection analysisof the periscope mirrors supported by the three-postkinematic mount. Initial results indicated less than 125-nm peak-to-valley distortion of a mirror in the worstpossible attitude. An alternative kinematic mountingpost was designed in which flexures replace sphericalball bearings. Trial parts were ordered.

FRONT-END SYSTEM (W. Leland, M. Kircher, C.Knapp, D. Swanson, G. York)

Introduction

A detailed design of a one-beam prototype front endwas completed, and procurement of hardware began asdesign drawings and specifications became available. Abaseline design of the Antares front end was alsoprepared. The oscillator systems for Antares and theprototype are different, but all other features of theprototype front-end design are identical to those of onebeam line in Antares. However, the Antares designduplicates the one-beam design six times for all majorcomponents after a six-way beam splitter. The Antaresoscillator complex will produce a multiline pulse outputwith the capability of shaping the pulse, whereas theprototype beam line will, at least initially, use a multilineoscillator complex without beam-shaping capabilities.The one-beam prototype preserves important distancesbetween major components but has a different con-figuration from the Antares front end to permit installa-tion in an existing building.

General Description of Beam Line and Arrangment ofComponents

A schematic layout for the one-beam prototype isshown in Fig. II-1. The Antares layout is shown in Fig.II-2. The following description of the prototype beam

16

Fig. III.Layout for one-beam Antares prototype front end.

LI

J.

45.

I

».

Ml

! t

STAME OSCILIATOHS

ownsATTMMTOK A * MAM COAWNEt

M M IMOM MWSUIS

PWMP OPTICS TAKE

4-PASS MKTBI PKAMP IUM0MCS «

SPATIAL FHH* » l .SWIKKOKI.SMI IHFRHtPf .A U O M M - K W HUECTMM STATION

MPMP W W OPTICS TAKE.MWMirmiscopi.M M SMMNt STATM).PMMMV CASSECMM OPTICS

PAHASniC SIPK«SSMN CHAMH

PM*«P|U«NMCSia

m w mnw unici UMJ.McmMn CASSWHM OPTICS.IIAMSTBMWSUTla)

u.M.

15.

M.II.

M.

H.

X.

SATIMAHE AISOMEt CHAMKR

K A n s n i n aSPATIW F i l m 1] ,AIICNMM'KAM HtNCIHM iTATHH

PAW IHKW MWSKIKWMVtl WPtfl OPTICS TAME.

AtlBWENT miSCOH.•cm sianwe SIAIION,PIHMAItV CASSCCIIAIN OPTICS

MIVCIT ODIPU1 OPTICS.SECONMtV CASSECMW OPTICS.KAM SKEIHNC STATION,CAIC4MK1O

M M AtlCNMEW STATION

ML11.

3Z

S.

H

».

*ST.

%

VACUUM PUMP

Eucraw-ciM P i u a

PKAMP POHEt SUPPtV

mSTMMMATICN HACKS

OSCKIATOt MAOHSTICS

SCMMDOOM

A. npnw• K C M I I E *C. HK« OPTICS IMCirACE

E. CMWCMCtF. SPAKMCKC. FUPPVOISKH. COMfUTtll niMINAl1. HAM COPT (MITJ. TIMHK MM COMMUNICATIONS HACKK. DISPUVS

TMNSFOMEIt

CLEAN WOM CLOMINC SHELVES

SINK

Fig. II-2.Layout of six-beam Antares front-end system.

line applies in all important respects to the Antaressystem, with a sixfold duplication required in Antares af-ter the beam splitter. In the following paragraphs, num-bers in parentheses refer to numbers associated withparts shown in Fig. I I I .

The beam, which originates in a multiline oscillator(1), is sent through a spatial filter (2), a beam-steeringstation (3), and a switchout station (4). At this point it

emerges from the oscillator complex as a multiline milli-joule pulse of the desired duration (typically, 1 ns). Thespatial profile is Gaussian with a characteristic width(1/e) of about 0.6 cm. The beam proceeds from theswitchout station to a beam-steering station (S), and thenthrough another spatial filter (6). An alignment-beam in-sertion station (6a) is coupled to the spatial filter. Visibleor CO2 alignment beams are inserted at this point

17

The alignment beam is adjusted to be colinear with themain beam as defined by the spatial filter. Divergence ofthe alignment beam is also adjusted to be equal to thedivergence of the main beam. After passing through thespatial filter and another beam-steering station (7), thebeam travels through a preamplifier complex (8). Thepreamplifier is fitted with on-axis Cassegrain optics toprovide three-pass amplification. Parasitic suppressioncells are located at each end of the preamplifier. Provi-sion is made to insert an optical tooling telescope justbefore the preamplifier at station (8a). The Cassegrainoptics and preamplifier produce an annular beam of 9-cm o.d. and 3-cm i.d. with a nominal energy density of50 mJ/cm2. Next, the annular beam passes through asaturableabsorber cell (9). and is then directed into abeam-splitter station (10). The beam splittergeometrically cuts out a 2.5-cm circular beam (six beamsfor Antares). Trie unused beams are routed to thediagnostics station (10a). but the 2.5-cm beam cut out bythe beam splitter passes through a beam-steering station(II). where it is directed to another spatial filter (12). Thealignment-beam insertion station (12a) is associated withthe spatial filter (12) and has the same functions as sta-tions (6) and (6a). After the spatial filler (12) the beampasses through a path-length adjustment complex (13)and (14). through another beam-steering station (IS),and then through the driver-amplifier complex (16).Provision is made to insert an optic tooling telescope at(16a) just before the driver complex. The driver complexconsists of half an Antares dual-beam amplifier fittedwith on-axis Cassegrain optics and parasitics-suppression cells at each end. The dimensions of the an-nular output beam are 15-cm o.d. and 9-cm i.d., with al-J/cm2 energy density of the multiline output beam. Theoutput beam is directed either to a beam-diagnostic sta-tion (17) or to the retropulse simulators (18) and (19).The retropulse simulator can produce a backward-running pulse, which is delayed as it would be if it werereflected by a target in Antares, by reflecting a chosenfraction of the output pulse. Implementation of theretropulse simulator is not planned for early prototypebeam operation.

As noted earlier, the Antares beam layout shown inFig. II-2 follows the same general sequence of compo-nents. Output from each of the six driver amplifiers isdirected onto stations that are a part of the opticalsubsystem.

Description of Major Components

Driver Amplifiers. Driver-amplifier procurement con-tinued. Systems. Science and Software (S3) was selectedto build the three dual-beam units according to ourdetailed specifications. Design reviews were held with S3

in August 1978, and fabrication was started. Due to un-favorable component delivery schedules and other dif-ficulties, the driver could not be used in the prototypefront end by the requested December completion date.

The driver amplifiers are high-pressure (1200-torr)electron-beam-type amplifiers. They are packaged asdual-beam units with two high-pressure discharge cham-bers coupled to a single, two-sided, gridded, cold-cathode gun. Each high-pressure discharge has its ownenergy supply, but all three electron guns are connectedto a common energy supply. Switching relays providefor operation of any subset of the three electron guns.

The 2-m gain length will provide gain-length (gL)products up to 7. although adequate output energy ispredicted at gL products of 6. A parasitics-suppressioncell is attached at each end of the amplifier to ensurestable operation with triple-pass, on-axis Cassegrain op-tics. The 21-cm-diam, 4.5-cm-thick N a d windows aretilted 15° to the optic axis to avoid on-axis reflection.Beam diagnostics can be performed by using the reflec-ted energy from the salt windows of the parasitics sup-pi ession cells. Optics are mounted on support standsoutside the amplifier isolator cells. Margin to provide ad-ditional output cr to compensate for unforeseenproblems is provided by the capability for higher-than-nominal gL products and by the designed ability tooperate at 1500 torr rather than at the nominal pressureof 1200 torr.

Preamplifier. We selected a standard commercial unit(Lumonics 602) for the preamplifier. The amplifier andmost auxiliary equipment (e.g., stands, power supplies,parasitic isolator cells, windows) were delivered andassembled. The unit was operated routinely at specifiedgain (gL = 4) and gain uniformity. Tests with the triple-pass Cassegrain optics awaited delivery of the optics.

Oscillator and Switchout. The oscillator to be usedfirst in the prototype front end is a modified Laser

18

Development Corporation (LDC) unit. The LDC os-cillator combines a pulsed TEA laser of our design withmultiple low-pressure plasma tubes, one for each desiredline. The optical cavity for each line includes one low-pressure plasma tube, but shares the TEA laser. Agrating combines the beams to give the unidirectionalmultiline operation. The switchout is a standard LASLdesign with three stages of CdTe Pockels cells and ger-manium plate polarizers. The Pockels cells are driven bya fast (nanosecond) high-voltage (25-kV) pulse generatedby a charged line and a spark-gap switch. All equipmentfor the prototype oscillator-switchout system wasdelivered, and checkout was begun.

Laser energy extraction calculations and predictionsof the performance of some target designs indicate theneed for pulse-shaping. In the LDC design, the commonbeam volume in the TEA laser (part of the oscillatorcavity) produces coupling between lines and precludespositive relative-line-amplitude adjustment. The Antaresoscillator design provides an extremely versatile system,which allows positive amplitude control for each line aswell as pulse shaping by variable timing of pulses fromeach line. The final switchout unit is identical to theprototype unit and can. in principle, also provide pulse-shaping by control of the switchout voltage puise. TheAntares oscillator design uses six stable, 1-W, cw lasers,each operating on a selected line. Each of the six beamsis switched out individually and directed through its ownpath-length adjuster before being combined into a com-mon beam path. Adjustment of the relative timing of thepulses from different lines, either electrically or by path-length difference, will produce pulse-shaping. Controlover the individual line amplitude further adds to theline-shaping capability. Amplitudes can be adjusted byvarying the power input to the cw lasers and by insertingattenuators. The 1-W power level in the cw lasers mustbe amplified tc 1 MW. The desired amplification is ac-complished with four-pass amplification in a commer-cially available laser (Lumonics K922S), which has a gLproduct of 4. Amplification is performed after combiningthe six beams into a single beam by combinations ofzinc-selenide beam splitters. No serious problems withcoupling between lines are expected because we areoperating the Lumonics laser in the small-signal regime(output energy of 1 mJ vs available stored energy inexcess of 100 mJ). The four-pass amplification designdepends on a polarization rotator, germanium Brewster-

angle plates, and a Pockels-cell switch with germaniumplate polarizers. All four passes lie on a common paththat requires a minimum of gain volume. Good contrastratio is obtained because two passes are made throughthe Pockels-cell switch. Design of the unit was completedexcept for some detail drawings of minor parts. Partswere ordered and fabrication was initiated.

The repetition rate for the Antares oscillator complexcan be as high as 3 Hz. Longer pulses, along with thehigh repetition rate, will give output power useful forsome alignment tasks.

Beam-Line Components Not Associated Directly withAmplifier. The spatial filters consist of two poweredmirrors with a pinhole of appropriate size located be-tween them at the focal point of the input mirror. Thefilters are used in the front end for spatial beam qualitycontrol; beam-line definitions for use in alignment;alteration of beam characteristics (size and divergence);and reiiopulse interruption. The design of these units iscomplete, and parts were either delivered or have beenordered.

Beam-steering stations are used at various locations toadjust beam position and direction, and to facilitatepositioning of the beam line relative to large units suchas amplifiers. These designs are complete, and parts weredelivered or have been ordered.

The saturable-absorber cell is included to enhance thecontrast ratio as needed. It is different from theparasitics-suppression cells in that it needs to functiononly on the selected lines of operation, whereas theparasitics cells must cope with all lines that can produceparasitics. The gas filling for the saturable absorber isSF6. Design was completed and procurement started.

The beam splitter is a perforated mask whose holesseparate the number of needed beams. Appropriate op-tics pick up these physically operated beams and directthem as needed. Design was completed and parts havebeen ordered.

The path-length adjuster compensates for the differinglocations of amplifiers (and hence, for the path length tothe target) by adding length to the shorter paths. Designwas completed and parts were ordered.

The alignment-beam insertion stations are similar tothose in Helios. The design was completed and partswere ordered.

19

Controls and Diagnostics

The prototype front-end controls are identical to themanual backup, checkout, and maintenance controlsdesigned for Antares. The manual controls are compati-ble with switchover to computer control when com-puters, interface hardware, and software becomeavailable. Diagnostic or monitoring information from thevarious components in the prototype front end is ob-tained by Antares design sensors, signal generators, andtransmitters. The design incorporates fiber optics forsignal transmission wherever possible. The output of theoscillator complex, the preamplifier, and the driver am-plifier will be examined for total energy and for the tem-poral and spatial character of the pulse.

Detailed lists specifying type and characteristics (e.g.,voltage levels and impedance) of all Antares control andmonitor signals were prepared.

System Studies

Performance. Pulsa energy and temporal charac-teristics were calculated for the system. Our results in-dicated acceptable performance, but uncertainties ex-isted in the characterization of parasitics suppression bysaturablc-absorber gases under the wide range ofoperating conditions considered.

Gaussian-beam propagation calculations were madeto account for first-order diffraction effects. We ex-amined beam stability and factored it into the design ofall components. The geometric magnification producedby the Cassegrain optics used on the driver amplifier andpreamplifier desensitizes the system to directionalchanges of the beam at the input end, but it enhances thedisplacement sensitivity. All beam-line components aremounted directly and rigidly to the building floor.Whenever possible, mirror combinations are mounted insuch a way as to desensitize effects of their motion onbeam-line direction and position.

Reliability and Maintenance. Reliability received highpriority in all design work. We used conservative designand specifications to achieve an expected misfire rate ofless than 1%. Maintenance for Antares was designed forreplacement of readily accessed components or groupsof components. Diagnostics were included for rapididentification of the faulty replaceable part or group ofparts. Routine maintenance of energy storage units wasplanned.

Safety. All components presenting electrical hazardswere designed with regard to personnel safety considera-tions. Radiation hazards were addressed by controlled-access procedures. Both the prototype front end and theAntares front end were designed for a single controlledaccess whenever radiation hazards are present. Specialhandling equipment was designed or specified toeliminate mechanical hazards to personnel.

Equipment is protected by hard-wired interlocks andthrough computer check routines.

POWER-AMPLIFIER SYSTEM (R. Stine, R. Scariett,G. Ross, W. Turner, W. Miller, E. Yavornik, N. Wilton,G. Allen, W. Gaskill)

The Antares power amplifier and a cross section areshown in Figs. H-3 and -4. The unit is 16 m long, 3.7 min diameter, and the centerline is 3.2 m above the floor.The basic design incorporates a centrally located 12-sided grid-controlled cold-cathode electron gun, whichdischarges radially into 4 annular pumping volumes.

Fig. II-3.Antares power amplifier.

a«rr

Fig. 11-4.Cross section of Antares power amplifier.

20

All large long-lead-time components for th: pressure-vessel and electron-gun portions of the power amplifierwere ordered. The first electron gun. as well as the sup-port stand for the first power amplifier, will be deliveredby July 1979. The pressure-vessel components will bedelivered soon thereafter for installation on the supportstand.

The first electron gun will be tested in Building 85. in-dependent of the Laser Hall.

Pumping Chamber

Electrical Design and Testing. We designed and testeda solid anode bushing using a compressed-rubber dielec-tric. Initial voltage testing at 600 kV showed adequateholdofT. but lifetime data were still required. A metal-powder-loaded epoxy bushing was fabricated andawaited testing. The backup position depended on thewater-graded bushing tested in the prototype power am-plifier. The bushing holes in the pumping chamber wereenlarged to a diameter of 35 cm. to accept any of thebushings under consideration. We completed a dielectricdivider design that is consistent with the revised electron-gun support system. The dielectric divider is not an in-tegral part of the support system, but it prevents ring-mode optical parasitics and improves gain uniformityand efficiency by containing the electron flux.

Mechanical. The pumping-chamber and spacer-spoolsections are being fabricated: delivery is expected in May1979. The salt-window domes and back-reflector domeswere redesigned to eliminate costly forgings. These unitswere ordered, and delivery is expected in September1979. The design of the electron-gun support system wasnearly completed.

Electron Gun

Electrical Tests. Modifications were made to theprototype power amplifier to allow in-depth testing ofelectron-beam characteristics for the final Antaresdesign. The modifications included a new control-griddesign, improvements to the vacuum system, some ad-ditional corona ri.igs on the cathode assembly, viewingports in the grid, and a careful cleaning of all parts of theelectron gun. The major problem to be solved by thesemodifications was grid emission at high voltages and atgrid resistance, which led to a loss of control of the

electron-beam current. In addition to the physicalmodifications to the electron gun. we also developed agrid-conditioning procedure, which consisted of shortingthe grid to the cathode and then operating the gun with aseries of pulses of increasing voltage. We tested thesechanges in a series of runs at voltages between 300 and500 kV and grid resistances between 250 and 1200 ft.The improvements allowed operation at higher voltagesand grid resistances than previously possible, and therewas decreased frequency of grid emission at all voltages.The results of these tests, which were incorporated intothe Anlares electron-gun design, showed that the An-tares electron gun should perform well within the re-quired specifications.

We also conducted tests to determine the electricalperformance of the prototype dielectric dividers at lowvoltage. When the electron-gun voltage was reduced to300 kV. flashover was observed between the ribs of thedividers. At 400 kV the flashover stopped. The new An-tares designs for anodes and dielectric dividers were in-stalled and will be tested.

We selected a test facility for the electron guns. Draw-ings of building modifications required for the testingprogram were submitted for approval. The modificationsincluded a pit for the vacuum system, lead shielding, anda large screen room. We considered these provisionsnecessary because of the late completion date for theHEGLF laser building.

Mechenical. All four electron-gun wcldments weremachined successfully. A typical weldment with themachined hibachi slots is shown in Fig. 115. The weld

Fig. IIS.Typical electron-gun weldment with the machinedhibachi slots.

21

test program proceeded well, with all the flat samplescompleted. Once the final ring sample is welded andtested, the electron-gun section will be machined inpreparation for final welding and will be shipped to thewelding shop. We designed a revised sealing concept forthe junction of electron gun and pressure vessel. Thisconcept uses bellows and eliminates tight tolerances inthese large parts.

Three vacuum spools were being fabricated, withdelivery starting May 1979. Both the coaxial feedbushing and the electron-gun support bushings were be-ing fabricated, with delivery expected in April 1979.Brazed-foil and glued-foil electron-gun windows were or-dered for the first: electron gun.

Gas and Vacuum Subsystem

The electron-gun high-vacuum lurbomolecular pumpwith its associated mechanical backing pumps wasreceived. We will install the system in an existingbuilding for electron-gun testing. Detail drawings of thecomplete gas and vacuum system were started. TheVacuum System Evaluation and Test Laboratory wascompleted.

Handling Equipment

We emphasized the completion of the electron-gunhandling equipment because it is needed first. A low-boyfixture, which will serve both as a handling fixture and atest fixture for the electron-gun testing, was designed.The fixture ran be rotated to allow easy access to all 48foil windows. Drawings for the grid/cathode assemblyfixture and for the 3.2-m-high gun-insertion fixture to beused in the Laser Hall were completed.

Optical Support Structures

Detailed structural analyses of the back reflector,relay optics, and in/out oprics support structures werecompleted. The drawing for the optical support structureof the back reflector was completed; the relay and in/outoptical support-structure drawings will follow. The long-lead-time aluminum plate for these elements was or-dered.

The optical components and their locations on the op-tical support structures were defined. The first turningmirror will have the same kinematic mount as the backreflector and the periscope mirrors. The second turningmirror and the focusing mirror will use a modified com-mercial mirror mount. Prototypes of the trombone andspatial filter were fabricated. A small mirror clampingdesign was detailed, and mirrors were fabricated andtested. The transmission gratings for large beam sampl-ing were integrated into the optical support structures,and three prototypes are being fabricated.

Support Stand

Weldment fabrication and final machining of thepower-amplifier support stand were ordered. The steelbox beams were ordered earlier and will be delivered tothe stand fabricator in mid-February 1979. The first sup-port stand was scheduled to arrive by July 1979.

High-Voltage Cable

Both General Cable Company and Sieverts Kabelverkdry-cured high-voltage cables were certified for use at550 kV in the power amplifier. Specifications for bothcables were completed and sent out for bid.

Diagnostics

The design of the small-signal gain measurementsystem was completed. We ordered the CO3 laser as wellas all mirrors, mirror mounts, and other accessories.This system will require a screen room in the Laser Hall,which will be shared with the optical diagnostics.

A completely computer-integrated diagnostics systemwas designed for the electron-gun testing.

Air Bearings

A new set of electron-gun air bearings was installedand tested on the test vehicle during December 1978(Fig. II-6). The new air bearings were a second-generation version of the prototype bearings tested inJuly 1978. The major difference between the prototypeand the new air bearings was the size of the internal air-supply line, which was increased to 1.3-cm o.d. This

22

Energy. kJVoltage. MVInductance, uHJitter, rms. nsPrefire Rate

300.1.2

<3.0<20<0.0l

Fig. II-6.Ant ares electron-gun air-bearing assembly (showninverted).

change was made to improve the performance of the air-bearing assemblies. The smaller copper tubing on theprototypes appeared to restrict the air flow to the bear-ings, resulting in nonuniform lift and high frictional drag.Our test results showed that the new air bearings hadmore uniform lift, operated at a lower pressure, exhibiteda higher flow rate through the assembly, and required alower tow load than the prototype. These improvementsin performance we; • the direct result of increasing thesize of the air-supply line.

ENERGY STORAGE SYSTEM (K. Riepe, G. Allen, J.Bickford)

The Energy Storage System supplies the high-powerhigh-voltage pulses to the gas discharges and electronguns in the power amplifiers. The system consists of 24gas pulsers storing 300 kJ each, and of 6 gun pulsersthat store 70 kJ each.

The gas pulsers will be built by a contractor. We builta prototype gas pulser to provide information for evalua-tion of proposals, to help us review the contractor'sdesigns as they progress, and to provide a test bed forcontrols and diagnostics.

Because requirements for the gun pulser have not beenfully defined, these pulsers will be designed andfabricated in-house.

Gas Pulsers

Requests for quotation for 25 Marx generators (in-cluding 1 spare) were sent to 4 vendors. Basic specifica-tions were

The contractor will be required to build one Marx in hisfacility and to run it for 2000 shots to allow us toevaluate its reliability. The first trigger system will betested for 50 000 shots. Bids were received from threevendors. After some changes in specifications, best andfinal ofl'ers were received in November 1978. Letter-contract approval was predicted for late January 1979.

Testing of the prototype gas pulscr was slowed byfailure of the charging-power supply. Because a powersupply of comparable size was not available, testing con-tinued with a much smaller power supply.

The Marx generator was reconfigured to measure theeffect of spark-gap position on inductance. In theoriginal configuration, the spark gaps were placed awayfrom the centerline of the stage, toward the centerline ofthe Marx. It was thought that this would minimize theloop and. thus, the total inductance. However, we foundthat moving the gaps back to the centerline of the stage,and thus decreasing the stage inductance while increas-ing the loop inductance, decreased the Marx inductancefrom 2.7 to 2.4 uH.

New spark gaps, based on the spark-gap testprograms, were installed in the Marx. Before installation,these gaps were tested to determine the effect of dimen-sions and electrode condition on the self-breakdownvoltage (SBV) measured at atmospheric pressure with aslight flow of air through the gap. The gap was isolatedfrom the power supply with a large (1-MfJ) resistor, anda small capacitor (2.4 nF) was charged and dischargedthrough the gap. SBV was measured after several tens ofconditioning shots. The standard deviation of the 17gaps tested was 1.8%, and the largest difference was3.4%. indicating an acceptable tolerance stack-up.

The coated Carborundum resistors to be installed inthe gas-pulser prototype were received and tested beforeinstallation. The test consisted of soaking the resistors inoil for one week. The resistance increased if there werelarge holes in the coating. A 10-kJ, 10-kV capacitor wasthen discharged through each resistor. When the resistorwas heated by the dissipated energy (30 kJ produced atemperature rise of 50 K), a stream of small bubbles wasemitted from any pinhole leaks. About 20% of theresistors failed this test. We continued to work on this

23

problem with Carborundum Corporation. They recom-mended a double costing of polyurethane, rather thanthe usual jjolyurethane-over-epoxy coating.

Electron-Gun Falser

The electron gun pulsers provide the energy to drivethe electron j,ins. The beams from the electron gunscontrol ihc ion;/xHion density, hence the impedance, ofthe gas discharges. The gun current density controls theoverall impedance, whereas the voltage controls the elec-tron range, and thereby controls the gain distribution.iVe -J .inducted a Jieoretical and experimental programto determine me optimum voltage and waveform.

Jf 'he or.'Mium waveform is square or trapezoidal, aGuiiUrnin network can be used. We built and tested tofull voltage a megavolt, two-mesh, type-A Guillemin-Marx generator. In addition, we used the Marx portionof the network to life-test the high-voltage cable samples.This procedure provided a good life-test of the Marx it-self. The disadvantage of the Guillemin network was thatit limited the risetime. Experiments indicated that avoltage rise of several kilovolts per nanosecond was re-quired for uniform ignition of the cathode. Thus, theGuillemin network required a peaking capacitor andswitch to provide this voltage rise.

Theory indicated that the optimum gun waveform wasa high-voltage spike (500-600 kV), followed by a lowervoltage plateau (300-400 kV). The high-voltage spike oc-curred before voltage was applied to the gas, fully ioniz-ing the entire volume. After voltage was applied to thegas, the applied electric field allowed the use of lowerenergy electrons, while maintaining ionization across thefull discharge gap. We found experimentally that,without the initial high-voltage spike, gas breakdown oc-curred. This waveform can be generated by a Guilleminnetwork. However, a good approximation may also begenerated by a straight Marx generator, which has theadvantage that it may be built with low inductance, andthus does not require a peaking circuit to provide goodignition of the electron gun. These considerations led tothe selection of a simple inexpensive Marx generator asthe Antares gun pulser.

Component Development

After completion of the spark-gap life-testing under

fault conditions, we performed some experiments to

characterize the gap under normal operating conditions.A dummy load was inserted in series with the spark gapto limit the peak current to ~25O kA with 25% reversal.The parameters of interest were trigger jitter and prefirerate.

It is difficult to relate spark-gap jitter to Marx jitterbecause the Marx jitter depends on the internal couplingof the Marx. However, if the spark-gap jitter is high, theMarx jitter will obviously also be high. Switch jitter wasless than 10 ns with a 200-kV trigger pulse at 120 kVand at a safety factor of 2. This implied that a triggerpulse of more than 200 kV, and probably at least 300kV, will be needed for the Marx to achieve a Marx jitterof <20 ns.

To achieve the specified Marx prefire rate of about 1in 100, the individual spark-gap prefire rate should beless than I in 2500. Verification of this value would takemore than 5000 shots. This was not consideredreasonable, and a less direct approach was taken. Thedistribution of SBV at a pressure corresponding to amean SBV near 100 kV was measured, and theprobability of self-breakdown at or below one-half themean SBV was calculated by assuming that the distribu-tion was Gaussian, and that the standard deviation wasindependent of pressure. The standard deviation of SBVwas between 2 and 3% of SBV over the range of 50 to120 kV. If the distribution is Gaussian, then the prefireprobability at one-half SBV is negligible (< 10"100). Fromexperience, we know that the prefire rate will be higher,so either or both assumptions were incorrect. However,the standard deviation of gaps operating at much lowercurrent is also on the order of 2 to 3%, so we may expectthat the prefire rate of the Antares gaps will be no worsethan that of conventional gaps.

Power Supplies

Specifications for the Antares power supplies werewritten. We require a gas-pulser power supply that willcharge four gas pulsers in one minute. Six will be re-quired. The gun-pulser power supplies will charge onegun pulser in ~20 s, depending on the gun-pulser energy.Six will be required. One spare of each will be purchased.

The first unit of each power supply will be required toundergo a proof-of-design performance test at themanufacturer's facility, as follows.

a. Operate high-voltage transformers without rec-tifiers at 200% of rated voltage for 7200 cycles.

24

b. Short the output through a 100-ft-long cable at 60-kV charge with a vacuum relay.

c. Operate the full power supply into an open circuitat 130% of maximum charge voltage (78 kV) or at max-imum open-circuit voltage, whichever is higher, for 6hours.

d. Operate with both polarities short-circuited for 1hour;

e. Operate at full power into a resistive load for 1hour.We expect these tests to guarantee the reliability of thepower supplies. Each power supply will be required topass an acceptance test consisting of the transformer testat 200% of rated voltage, and of 2-hour open- and short-circuit tests.

TARGET SYSTEM (P. Wolfe, J. Allen, W. Sweatt, N.Wilson, V. Zeigner)

Introduction

The scope of the target system is illustrated in Fig. II-7, an artist's sketch of the target vacuum system. Sixbeam tubes, each 1.7 m in diameter and ~35 m long, ex-tend underground from the north wall of the Laser Hallto the Target Building, where they connect to six 3-m-diam turning-mirror chambers. Six smaller beam tubeslead to a centrally located target chamber, a 7-m-diamvessel that houses a space frame, which, in turn, carriesarrays of folding and focusing mirrors for each beam, aswell as a precision target-insertion mechanism. Figure II-7 shows the space frame in its rolled-out position, whichrequires the removal of one target-chamber end dome

Fig. 11-7.Artist's sketch ofAntares target vacuum system.

and of three short beam tubes. This position will be usedprimarily for initial installation and major overhaul ofthe 12 mirror arrays. For normal operation the targetchamber will remain closed; targets and diagnosticequipment will be changed by means of air locks withoutbreaking vacuum; and minor maintenance will be carriedout by backfilling the target vacuum system with air andentering through a man-access hatch.

Activity related to the target system was markedly in-creased, which made it appropriate to organize a moreformal target system team, made up of the regular con-tributors from the various design disciplines, primarilyfrom L-Division's Antares Group, and of representativesfrom the Target Diagnostic and Target FabricationGroups. Brief weekly coordination meetings werevaluable for spotting action items and for keeping the ef-fort on schedule.

Optical Train

As described in some detail in the previous semian-nual progress report (LA-7587-PR), the portion of theoptical system within the target chamber was radicallyredesigned to decrease the focal length of the focusmirrors that illuminate the target. Although this changehad the desired effect of decreasing the spot size ontarget, it made the layout of components in the spaceframe and the use of space near the target for experi-ments much more critical. In addition, it increased thediversity of required sector mirror shapes over that ofthe earlier, more symmetric design. Therefore, we con-ducted several detailed studies of outline ray trajectoriesbetween mirrors and of beam footprints on mirrors andmirror arrays before completing the design. An indepen-dent check of the representative results of our in-housecalculations was performed on the Antares optical trainas a part of the Hughes Aircraft study contract.

Target Positioner

This component of the target system might bedescribed more properly as a target insertion and preci-sion positioning mechanism. As a key element in the op-tical train, its design must satisfy not only obviousfunctional requirements, but also must provide forreproducibly locating targets within a few tens of

25

micrometers of a fixed point in space, and for accom-modating optical alignment tooling to ensure that theremainder of the optical train is aligned and focused onthat point. In addition, for cryogenic targets, thepositioner must incorporate retractable means for cool-ing the target up to a few milliseconds before shot time.

On the basis of the successful design developed for theHelios system we commissioned the same LASL groupto develop a successor for use on Antares. A design anddevelopment schedule consistent with the overall targetsystem schedule was worked out, and design workbegan, with first tasks consisting of firming the require-ments for precise target positioning and for interfaces ofthe alignment system.

Space Frame

Upon completion of interference checks for therevised optical layout within the target chamber, wereinitiated the design effort on the space frame. Thegeneral frame envelope of this structure was alreadyfairly well defined, because it had to be established topermit the design of the target chamber to proceed (seeTarget Vacuum System, below). The remaining structureconsists of 12 supplementary frames to hold the 6folding-mirror arrays and the 6 focusing-mirror arrays,plus members to provide rigidity. We could not use acommon design for the supplementary support frames;this added complexity to the task of designing andanalyzing the overall space frame. Overlying the integra-tion of these members into an overall structure were therequirements that no optical-beam interferences be in-troduced, and that adequate target access be providedfor fusion experiments.

Two design layouts were generated and supplemen-tary scale models were built. These models permitted acritique of structural-member placement from thestandpoint of the frame's probable rigidity; providedclear insight on interferences and accesses; and served asindispensable aids in setting up the necessarymathematical models of the overall structure for rigidityand vibration analyses. At the end of 1978, we selectedone design and initiated mechanical analyses.

Target Vacuum System

This part of the target system composes the entirevacuum envelope, as shown in Fig. II-7 and described in

the introduction to Sec. II.F, together with the pumpsand controls needed to rough and maintain an operatingpressure of 2 x 10"' torr in the target chamber. InAugust 1978, we awarded a contract to design and buildthis system to the Pittsburgh-Des Moines Steel Com-pany (PDM), and a meeting of PDM and LASL designgroups was held at LASL.

At the end of a 10-week conceptual design effort byPDM, a 3-day preliminary design review was held inPittsburgh. Structural concepts presented by PDM wereaccepted with minor revisions, such as providing accessstairs instead of ladders, and aluminum platforms ratherthan steel. However, acceptance of the high-vacuum-system design was delayed, pending further calculationsand verification of pumping capabilities and com-patibilities of the proposed combination of Freon-chilledbaffles to handle water vapor, and helium-refrigeratedsurfaces to pump other gases.

A review of a revised pumping scheme was held inDecember 1978, but some apprehensions remained, andan alternative design based on helium-refrigeratedpumps alone was requested by LASL. Our experiencewith this approach on Helios led us to believe it to bereliable and adequate; to our knowledge, the combina-tion system has never been tried on applications likeours. PDM accepted our recommendation.

In December, authorization was given PDM to initiateprocurement of certain long-iead-time items andmaterials. A Final design review was scheduled forMarch 1979. before beginning fabrication.

CONTROLS SYSTEM (M. Thuot, D. Call, D. Carsten-sen, D, Gutscher, A. Kozubal, R. Lindberg, F. Maestas,F. McGirt, W. Seifert, B. Strait)

General

Final requirements for the Antares Controls Systemwere established. These requirements are listed below.Absolute Requirements

• Accept and execute operator commands accordingto predetermined algorithms.

• Prepare and fire the laser in a specified sequence.• Acquire, store, and display laser and control system

information.• Meet the schedule determined by the installation

and checkout of other Antares subsystems.Design constraints that tend to limit the manner in

which absolute requirements are met are listed below,

26

followed by additional design constraints ordered ac-cording to priority.Absolute Design Constraints

• Personnel must be protected from severe safetyhazards.

• The control system must operate in an adverse EMIenvironment over a large area.

• The control system must have sufficient generalityand flexibility to accommodate change.

• A comfortable man-machine interface must beprovided.

• System control from a central location must beprovided.

• Distributed subsystem control must be provided.Additional Design Constraints. Priority-Ordered

• The control system will be hierarchical and tree-structured.

• The control elements will be designed for highreliability and low mean repair time.

• The control system will be implemented accordingto the best techniques available for specification,design, and documentation.

• Testing and integration will proceed at many levelssimultaneously, with tools and controls that preventhazards to personnel and equipment.

• A common control language, common hardware,and a common operating environment will be sup-ported at each level of control.

• Decisions to make or buy hardware and softwarewill be based on requirements, interfacing, support,schedule, and budget.

• Command acknowledgment and command execu-tion times will be minimized.

• A programmable central control sequencer will beprovided to fire the laser.

• All laser data and status information will beavailable at the central control station.

• The control system will provide archival storage oflaser and control data.

• Computer-generated hard copy will be available atthe central and subsystem levels.

The configuration of the Antares control system asspecified above was established as shown in Fig. II-8.This configuration is a four-level hierarchical tree-structured network of 11 minicomputers and ~100microcomputers.

The top-level processor, the Integrated Control Center(ICC), provides an integrated central control point for allof Antares. In addition, auxiliary support functions suchas data-base management will also be provided at thislevel.

The second-level processors, the Subsystem ControlCenters (SCC), provide individual control of each majorAntares subsystem.

The third-level processors. Beam Line Controllers(BLC), exist only under the Laser Hall SCC and serve ascontrollers for each Antares beam line. The fourth-levelprocessors. Machine Interface (MI) processors, executethe desired control and data-acquisition functions.

The experimental data-system node of the controlnetwork refers to the experimental data system, which isunspecified at present.

M - NMHIWINTERFACE PROCESSOR

Fig. II-8.Antares control-system configuration.

27

Communication facilities between all network nodeswill be supplied by a commercially available package.

Hardware

After competitive bidding, computer hardware wasselected as shown in Fig. II-8 with Digital EquipmentCorporation (DEC) supplying all the computers andcommunication facilities. The network hardware will beordered in stages. The first stage, obligated from FY-79funds, includes the ICC processor; the three SCCprocessors for the front end, laser hall, and opticssystems; and two third-level processors. The remainingfive third-level processors for beam-line control will beprocured from FY-80 and FY-81 funds.

The configuration of the machine interface processor(MI) has not been finally determined. Work is inprogress to establish the control and data-acquisition in-terfaces to be supported.

Software

Schedule and budget required that as much of the An-tares software be purchased as possible. This par-ticularly applied to operating systems, communicationssupport, and even to the use of device drivers, where ap-propriate. Only in the case of specialized applicationstasks or when vendor-supplied software is not availablewill we write our own software.

Design and implementation of these application tasksbegan with the definition of software modules. The firstor top-level modules for the Antares software are shownas a hierarchy chart in Fig. II-9, which only describesthe major software functions and the interrelationshipsbetween these functions. The data paths are notdescribed. Lower level modules are shown in Figs. 11-10to II-12.

ANU»tS SOfTWAM HIHAHCHV CHAM - C(WTHOU 1YSTE*

I . orCIATE SVSTIM

Fig. 11-10.Antares software hierarchy chart of the lower levelmodules (Operate System).

ANTAHES SOFTWARE HIERARCHY C H O I - CONTROLS SYSTEM t

2. MANAGE DATA

Fig. 11-11.Antares software hierarchy chart of the lower levelmodules (Manage Data).

AWTAIItS SOFTWARE HIEKAKCHY CHART • CONTROLS SYSTEM

3. MANAGE COMMANDS

Fig. 11-9.Antares software hierarchy chart.

Fig. 11-12.Antares software hierarchy chart of the lower levelmodules (Manage Commands).

28

Front-End Controls Antares Timing System

Detailed control requirements, complete with cost andmanpower estimates and schedules, were integrated intothe total control-system effort, which contributed to finaldefinition of the overall control structure describedabove. Tests of hardware in several areas such asC A M A C crate control lers for the LSI-11microprocessors were completed. A specification waswritten for the standard Antares crate controller.

Laser Hall Controls

Detailed control requirements specified by SystemManagers for the power amplifiers, electron-gun test,laser diagnostics, and energy storage were integratedinto the overall control-system effort. Interactions withother systems were also specified. This work contributedto the establishment of the controls-system architectureand general approach.

Beam Alignment

Hardware designs for stepping-motor controls, hand-held controllers, and sensor electronics were completed.

Network Management

Most of our effort was expended for site preparationand procurement of computer hardware. In addition,plans for computer relocation in the Antares central-control screen room and installation of DECNET II be-tween a PDP 11/60 processor and an LSI-11 processorwere completed.

Requirements for the Antares timing system werespecified. These requirements will be used to write aspecification before design, implementation, andprocurement of the timing system.

ICC and Central Control

Requirements for the operator-machine interfaces forAntares are being established, including specifyingoperator control and status displays, video and audiosignals, control panels, and terminals. In addition, alarge amount of time was spent in specifying the dataarchival system for Antares.

MI Configuration

A team was formed to establish the MI configurationrequired for Antares control. Work was started on anevaluation of the input/output interfaces, the fiber-opticpackaging, and an EMI enclosure specification.

Antares Software

A team was formed to implement the Antares produc-tion software using structured programming methods.Work on the top-level modules was begun.

Early Support

A team was formed to provide early temporary sup-port for data-acquisition and control activities in thefollowing areas.

29

• Electron-gun test facility• Small-signal gain measurement development• Optical systems development• Laser diagnostics• Front-end systems development

vVork has begun to establish a general package for theparallel I/O interface driver for PDP 11/03 computers.

HEGLF SITE AND STRUCTURES (J. Allen)

Construction Package I

As of December 31. 1978. the job was 48% completewith 74% of the time consumed. This meant a reschedul-ing of the completion date from mid-July to late Septem-ber 1979. The delay was caused by abnormally badweather, change orders, bad concrete samples, and im-proper planning by the Contractor. Figure 11-13 is anaerial view of the site, taken in late summer 1978. Thestatus of the elements that make up Package I is outlinedbelow.

Laser Building. The high-bay portion of the laserbuilding, which houses the power amplifiers and energy-storage equipment, is shown in Fig. 11-14. According tothe construction sequence chosen by the Contractor, thispart of the package will be completed later than most ofthe other elements. Power-amplifier installation in thisarea is to start by early August 1979. A joint-occupancy

MNUTMIUM fMOMMIFIH

Fig. 11-13.Aerial view of HEGLF site.

Fig. 11-14.High-bay portion of laser building housing thepower amplifiers and energy-storage equipment.

arrangment must be worked out between LASL and theContractor if this schedule is to be maintained.

Most of the wall panels were erected and a portion ofthe support slab for the power amplifiers was poured. Afortunate selection of a low-profile crane allowed us togain 2 ft of hook height with only a minor change to thecolumn design.

The office, laboratory, and shop portions of the laserbuilding have progressed farther than the high bay, withgrade beams poured and the T-beam roofing sections in-stalled over two-thirds of this area. All underground con-duit and piping were installed.

The front-end room received one coat of paint and theair-conditioning ducting was installed.

Mechanical Building. The floors, equipment pads,walls, and roof concrete work were completed. Themechanical building and the laser-building shops andlaboratory areas were completed first to accommodatethe mechanical and electrical subcontractors' work.Water chillers, boilers, an air conditioner, and some pip-ing were installed. Part of the tritium stack was erected.

Office Building. The walls and T-beam roof sectionswere installed. The underground conduit and piping wereinstalled. Some concrete test specimens taken from theT-beam pours had low 28-day break values, but ad-ditional cores taken from the T beams later, showed ade-quate concrete strength. Office and laser-building con-struction was delayed until the discrepancy was re-solved.

30

Warehouse. A prefabricated design that satisfiedeverybody was selected. Construction will start when theground thaws—probably in April 1979.

Package II

The north and west walls of the target building werecompleted. The south wail was two-thirds complete, withthe beam-tube penetration inserts in place. The scaf-folding to support the forms for the 5-ft-thick roof was inplace. This building should be finished in July 1979,which is compatible with the scheduled start date ofAugust 1. 1979 for target vacuum system installation.This package was 42% complete with 76% of the timeconsumed.

The 500 000-gal. water tank and the backup fireprotection system were nearly complete. The tank wasnot painted, pending resolution of safety procedures tobe used by the Contractor.

The rebuilding of the power line leading into the site toaccommodate the extra HEGLF load was completed.

High-Voltage and Optical Evaluation Laboratories

Both laboratories were essentially complete. Thelaboratories were scheduled for occupancy in mid-January 1979. The original contract completion date ofOctober 23. 1978 was missed because of late delivery ofthe boiler and hot-water pumps. Several design deficien-cies became apparent during the final walk-through andcontributed to the delay.

III. CO2 LASER TECHNOLOGY(J. F. Figueira)

Each of our CO, laser systems described earlier represents a significant advance inthe state of the art of reliable CO2 laser subsystems, components, and diagnostics. Thedesign, construction, and improvement of the systems require, therefore, basic supportof CO, technology. Some important areas are the development of short-pulse multifre-quency oscillators; amplifier optimization; development of subsystems for the preven-tion of system self-oscillation and removal of prepulse energy, including the study ofrelaxation processes in bleachable gases; and the development of passive means toenhance laser beam quality and simultaneously to provide automatic target alignmentthrough phase conjugation.

PROPAGATION STUDIES (S. J. Czuchlewski, J. F.Figueira, A. V. Nowak, E. J. McLellan, C. E. Knapp)

Introduction

We evaluated the performance of a 10-GW sub-nanosecond CO3 laser system prototype capable ofreliably generating pulses of good optical quality andhigh peak irradiance at a repetition rate of 0.1 Hz. Thissystem will be used to conduct development studies insupport of our CO2 Laser Fusion program, includingtarget diagnostics development, optical damagethreshold studies, the evaluation of gaseous and solid-state saturable-absorber isolators, and the developmentof phase-conjugate reflectors. Comparisons were madebetween the energy and pulse shapes obtained with thissystem and those predicted by a theoretical model thathas been used extensively in high-power CO2 amplifierdesign. In this case, the model is examined in a regimewhere the input pulse width is less than half the effectiverotational relaxation time, a domain in which it has notpreviously been tested.

Apparatus

The prototype consisted of a reinjection CO2 os-cillator, an electro-optic switch system for producingsubnanosecond pulses, and a power amplifier (Fig. III-1). A commercial TEA laser was used as the power am-plifier. Both the oscillator and the power amplifier weremultipassed to optimize system performance.

The oscillator and electro-optic switch were describedpreviously.1 The Lumonics-600A TEA power amplifier

E-0 SWITCH, " -70 -ns pulse

Fig. HI-1.Two-stage TEA laser system, showing the reinjec-tion oscillator, the electro-optic switch, and thetriple-passed Lumonics power amplifier.

is a double-discharge, uv-preionized device with an ac-tive volume of 8.9 by 10 by 50 cm,2 capable of operatingat a repetition rate of 0.1 Hz. The gas mix used was61:15:24::He:N2:CO2 at 580 torr. Small-signal gain was4%/cm, measured with an attenuated multiline, gain-switched pulse that sampled a 25-cm2 cross section of

32

the active region on a path inclined 2.5° relative to theaxis of the device.

Experimental Procedure

Two types of measurements were made. First, wedetermined the response of the full system to a l-ns inputpulse, and second, we took a set of measurements to in-vestigate pulse shaping with a 570-ps input pulse.

Case 1. One-Nanosecond Pulses. For the first experi-ment. 2-mJ, l-ns FWHM pulses with 200-ps risetimewere produced by adjusting the Pockels cells in theelectro-optic switch for in-phase operation. Reinjectionto the oscillator medium for four stages of amplificationyielded a 230-mJ, 1.02-ns FWHM pulse with a 190-psrisetime for input to the Lumonics amplifier. The pulseenergy was distributed among the 10 um P(16), P(18),and P(20) lines. The beam diverged slightly as it passedthrough the amplifier, occupying a cross-sectional areaof ~25,40, and 46 cm2 on the three successive passes.

Energy and pulse shape measured and calculated aftereach pass are shown schematically in Fig. Ill 2. Thepulse width decreased appreciably on the second andthird passes, giving an 8.7-J output pulse with 778-psFWHM duration. Because the amplifier was underfilled,these values can be increased by 30% to produce an out-put power of 12 GW.

Beam quality was evaluated by measuring the Strehlratio for the system via a diffraction-grating technique.The measured ratio of 0.25 would allow a focused inten-sity of IOU W/cm2 with an f/2 optical system.

Case 2. One-Half-Nanosecond Pulse. The theoreticalmodel we use predicts that a 1.0-atm CO2 amplifiersystem should effectively amplify pulses as short as 250ps when the optical flux is sufficient to saturate the am-plifier medium. To test this prediction, the electro-opticswitch was adjusted for out-of-phase operation, yieldinga 0.3-mJ, 270-ps FWHM pulse with 120-ps risetime forsecond measurement series. During passage through thefour-pass reinjection amplifier, the pulse energy in-creased to 19 mJ, but the pulse broadened to 570 ps.

Because we wanted to study the performance of thepower amplifier for short pulses under saturation condi-tions, the beam expander was adjusted to provide asmaller beam than that used in the first experiment. The

10\ i i

TRIPLE PASS AMPLIFIER PERFORMANCE

MEASURED

CALCULATEDENERGY

INPUT PiSS I PASS I PASS 31

Fig. 111-2.Performance of the tripled-passed TEA amplifierwith a laser mixture of 61:15:24::He:N2CO2 at580 torr total pressure. Length of active gainregion was 50 cm.

area of the diverging beam during the third pass was 6.1cm2 compared with 46 cm2 in the previous case. Input-output parameters and pulse shapes for the Lumonicsamplifier under these conditions are shown in Fig. III-3.Note that, in this case, the injected pulse was broadenedsignificantly, to 880 ps FWHM, from the predicted valueof 370 ps.

Theory

The performance of the Lumonics amplifier has beenmodeled by using a coherent density-matrix kineticscode developed by Feldman,2'3 which describes thepropagation of pulses shorter than about 5.0 ns. Thiscode accurately predicts the energies and shapes of 200-J, 1.5-ns pulses propagated through our 2000-torrelectron-beam-controlled Gemini system.3 However, themodel has not been fully tested with subnanosecond

33

, MEASURED

CALCULATED

HEUCUmJ!

19

Tp(ns)

INPUTrowatw, 1

17f!.U«l»O/c»!

IS

Eit».:>,»

— 67"

— 71;

Tpi nt)

OUTPUT

i> fWEUHr £49f lit

fVg. ///-J.parameters and shapes from the triple-

passed Lumonics power amplifier for a 570-ps in-put pulse. Solid curves are experimental; dashedcurves are theoretical.

pulses. Because the characteristic relaxation times of thelaser medium vary inversely with pressure, the presentwork explores a substantially different regime from thatstudied in Ref. 3.

Modeling a strongly saturated, multipassed amplifierrequires accounting for the energy extracted from themedium on preceding passes. In addition, the 16-ns in-terval between passes may require accounting for ad-ditional kinetic processes. The first of these processes isintramode V-V relaxation.4 In addition to the upper(001) and lower (100) CO2 laser levels, higher lyingvibrational levels are also excited by the discharge. TheseJeveJs can reJax to the (001) state, thus repumping theupper laser level. Because this relaxation time has beenestimated to be ~50 ns,4 we ignored this process in ourpresent work. Ho.vever, this relaxation time has not beenaccurately established and may be shorter.

The second process that must be considered is relaxa-tion of the lower (100) laser level. By using Stark's ex-perimental data,5 this relaxation time is estimated to be45 ns for our conditions, whereas the Jacobs et al. data6

predict a value of ~7 ns. Because the latter results wereobtained more directly, we assumed in our model thatthe lower laser level relaxes completely between passes.Nevertheless, the discrepancy in the measured time con-stants for this process warrants further investigation.Under these assumptions, the gains on the second andthird passes of the amplifier are reduced slightly from theinitial values (80% of maximum gain in the extremecase).

Discussion

Because of the difficulty of modeling a multipassedsystem, comparison between theory and experimentmast be somewhat qualitative, and agreements within25% are considered good. Figures IH-2 and -3 showthat, although the theoretical and experimental energiesagree to within 15%. the predicted pulse shapes are con-sistently 40 to 60% narrower than those observed. With1.0-ns injected pulses, the pulse did narrow significantly,although not as much as expected. However, with 570-ns input pulses, the width of the observed pulse increasedfrom 570 to 880 ps instead of decreasing to 370 ps aspredicted. Figure 1113 shows that the second peak of theobserved output pulse is higher than predicted. This rein-forces the conclusion that the amplifier is not saturatingas completely as expected.

Two parameters affect the response time of the lasermedium: the dipole dephasing time T2 and the effectiverotational relaxation time x',.

Trot (1II-D

where T, is the longitudinal relaxation time, M is thenumber of laser transitions on which the system isoperating, and K is the rotational partition function.7 Thelatter function is obtained from a "reservoir" model ofthe rotational equilibration process within a givenvibrational state. The ratios TriM/T2 and TFWHM/TJM,where trist and TFWHM are the risetime and width of theinput pulse, should then characterize the response of themedium to a given pulse. These characteristic ratios arelisted in Table III-I for our two experiments and for theexperiment reported for the high-pressure amplifier.3 Inall three cases, the input-pulse risetimes were longer than

TABLE III-I

CHARACTERISTIC LASER-MEDIA

RESPONSE TIMES FOR THREE EXPERIMENTS

FWHM/Tf

CaselCase 2Ref. 3

8383273

1.031.540.88

2.02.3

22.0

1.00.371.9

34

T,. and the risetimes should not have been degraded bythe amplifiers, which agrees with the experimental data.For Case 2. however, the pulse width equals 0.4 TJOT,

which may explain why the pulse broadened as ittraveled through the amplifier. The reservoir model usedin the computer code has not been tested in this regime(especially for multiline pulses), which could explain theapparent discrepancy with theory.

Conclusions

The performance of a highly reliable, relatively simpleand compact. 10-GW. subnanosecond CO : laser systemhas been demonstrated successfully. Results suggest thatthe 1.0-atm TEA amplifier is not effective at propagatingpulses shorter than ~78O ps. Our experimental results donot agree with calculations of short-pulse propagation.This discrepancy has not been resolved. Further ex-perimental work is under way to study the propagationof intense subnanosecond pulses during a single pass ofthe Lumonics amplifier, when the process has not beencomplicated by multipass phenomena. A new high-pressure (2000—8000 torr) double-discharge reinjectionoscillator is being installed to facilitate this investigation.

The present study has also demonstrated the need formore detailed measurements on the kinetics of CO, lasermedia for pulse durations >IO ns. to determine the ef-fects of repumping by intramode V-V processes.

SATURABLE - ABSORBER RECOVERY (A. V.Nowak)

Introduction

An earlier report* described double-resonance experi-ments designed to measure the relaxation behavior of10.6-um gaseous saturable absorbers after excitation bya l-ns pulse from a CO2 laser. We continued theserecovery measurements primarily to clarify several un-resolved questions.

Experimental

Three modifications were made to the double-resonance apparatus previously described. TheHg:Cd:Te detector was replaced by a faster, more sen-

sitive liquid-helium cooled Hg:Ge detector; the 10-cmgas cell in which the two laser beams cross was replacedby a 3.1-cm cell tilted ~45° with inspect to the beams;and a germanium polarizer stack was placed in themonitor beam between the double-resonance cell and theHg:Gc detector to reduce the intensity of unwanted scat-tered light.

Results and Discussion

The relaxation rates for vibralionally excited SF6 in 10torr of Mix 804 (1.5% total SF6 content), extrapolated tozero pump-pulse energy are shown in Table III-ll. Notethat the relaxation time varies from a maximum of 1200ns for 20M14 | i.e.. pump laser tuned to the P(20) CO,laser line, monitor laser tuned to the P(I4) line] to aminimum of 210 ns at 2OM2O. The relaxation times aresimilar for a fixed monitor and varying pumpwavelength, but are very dissimilar for a particular pumpline and changing monitor line.

We show the dependence of the relaxation rate onpump-pulse energy and mix pressure in Fig. III-4. Notethat the relaxation time slows with increasing pulseenergy. This trend was observed for all pump-monitorpairs studied, although the actual value of the slopedt/dF differed from pair to pair. The Pt values ex-trapolated to zero energy are 7.84, 7.88, and 8.43us torr at 5.06. 10.0. and 19.93 torr, respectively. Here,P is the pressure of Mix 804 and T is the correspondingrelaxation time.

We also studied the relaxation of SF6 in a system con-sisting of Mix 804 and hydrogen. We chose H, because

TABLE III II

RELAXATION TIMEFOR VARIOUS PUMP LINES

10-u.m Relaxation Times forMonitor 10(im Pump Lines (ns)Lines

P(14)P(18)P(20)P(26)

P(I4)

<450...—

P(18)

350240...

P(20)

1200350210800

35

T24C

— 2 00'

£ !

£ -20-X i

3 }UJ i

*-080

040__£ . .

40 BO 120 160INCIDENT ENERGY (mJI

20C

Fig. UI-4.Relaxation time xfor Mix 804 at 20M26 after ex-citation by a l-ns CO-, laser pulse. The dashedlines are a least-squares fit. The slopes are 4.66,2.13, and 0.77 ns/mJ.

it is among the most effective known deactivators ofvibrationally excited SF4.' If used as an additive to anySF6-based saturable-absorber mix, it could produce arelaxation rate fast enough to protect against retropulseamplification in large CO2 laser fusion systems.

Our measurements of the relaxation rate of SF6 in Mix804 buffered by H2 are summarized in Fig. HI-5. Thepressure of Mix 804 was 10 torr. The relaxation due toH, was separated from that caused by Freons in the mixby use of the relation

PT » P [T804

x8043- l (III-2)

where t ^ is the relaxation time observed in 10 torr ofunbuffered Mix 804, and T I 0 4 + , is the relaxation time ob-served in 10 torr of Mix 804 buffered by P torr of H2.Note that at monitor lines P(18) and P(26) the PT valueincreases with H2 pressure above 100 torr. This behavioris inconsistent with a simple bimolecular model for V-Trelaxation of SF6 by H2. If the relaxation in the 804 + H2

system were dominated by the bimolecular process

18 -

16-

14 -

12 -

10 -

8 -

6 -

A -

2 -

• 20 M 18

A 20 M IB [N?>

A 14 M 18

• IB M IK

O20MI4

V20 M ?r,

I! ?0 M 20

A.10

, i . . _ . x . . l . _ J _

400 r.oo 1000 2O00

BUFFER PRESSURE (torr)

no ion '" ?oo

Fig. III-5.Pi values for the relaxation of SF6 by H2

molecules in the system Mix 804 + Hv ex-trapolated to zero pump-pulse energy. Thepressure of Mix 804 was 10 torr in all cases. Thesymbol XMY refers to a /0-um P(X) pump lineand to a /0-um P(Y) monitor line.

S F ; • (Hl-3)

where SFJ denotes a vibrationally excited SF6, then xwould be inversely proportional to the H2 pressure, or PT= constant.

We also studied the relaxation of the bimolecular mixSF6 + H2. Figure III-6 illustrates two extreme cases. For20M26, the relaxation time increased drastically withpump pulse energy at 10.8 torr of H,, but onlymoderately at 37.9 torr of H2. For I4M18, we observedthe opposite effect: the relaxation time decreased with in-creasing energy, very noticeably at 9.99 torr, but onlyslightly at 20.1 torr. Only 18M20 at 21.3 torr of H2

showed a very pronounced increase in the relaxationtime with increasing energy. The other cases studied, in-cluding 20M14,18M18, and 20M40, showed only a mildincrease of the relaxation time with pulse energy.

The PT values extrapolated to zero pulse energy forSF6 buffered by H2 are summarized in Fig. III-7. Notethat our data are consistent with those of Steinfeld at al."for the one value at which our results can be compared.The agreement is striking when we consider that theirpulse was longer (O.S us vs 1 ns), and weaker (severalkW vs several MW), and that the H2 pressure was muchlower (160 to 430 mtorr vs 10 to 80 torr).

We offer these results without interpretation. The SF6

system has a long and deserved history for being difficultto model. We will defer any complete interpretation of

36

laoo

1600

1400

11

i200;

v>

lOCO'

Oi—<x

a. eoo

400

200-

• 20 M 26, P lMj)'iO8tcrr-j

A 20 M 26 ,P lM?]«379torr j

O 14 MiB .P (Hj ! = 999ijrr

A 14 M 18 , P ( H j ) = ? 0 Horr

40 *0 120 160INCIDENT ENERGY (mj)

200

Fig. III-6.Relaxation time xfor SFt in the bimolecular gasmix SFt + H2, following excitation by a 1-ns CO2

laser pulse. The SF^ pressure was 155 ± 5 mtorr.The slopes for the case 20M26 are 5.66 and 0.151ns/mJ, and for the case 14M18, are -2.39 and-0.313 nslmJ.

the data until a more comprehensive data base isobtained.

Future Experiments

More work remains to be done to completely charac-terize the relaxation of Mix 804. First, we will prepare atable similar to Table III-II, to include five lines at whichthe Helios CO2 laser amplifiers operate, P(14) throughP(22). Second, we will flrmly establish the dependence ofthe relaxation of Mix 804 on the energy of the pumppulse.

—o

IO

1

k

16

14

,2

•

6

4

2

°l

A 14 M I t

a 18 M IS

V 20 M 26<e« HESULT

. OBTAINED

o

D

BY

20 M 14

20 M 20

18 M 20

STEINFELDtloil

(P a 10)

2 t

V*

•

10PRESSURE

ir 20

•c

2 0

(10'fI

M 40

a*

50

—

-

•<

- i

D ^j

1IOC

Fig. 111-7.Ft values for the relaxation of SF6 by H2 inbimolecular gas mix SF^ + H2 extrapolated tozero pump-pulse energy. The SFt pressure was 155± 5 mtorr in all cases.

It would also be desirable to perform a detailed studyof the bimolecular system SF6 + buffer gas. This mix iseasier to model than multicomponent mixes. The resultson these types of systems, for example, the relaxationtime as a function of the buffer series He, Ne, Ar, Kr,and Xe. could yield a useful pattern of information andwould be substantially more interesting to the scientificcommunity than the results for multicomponent mixes.

HIGH-EFFICIENCY PHASE-CONJUGATEREFLECTION IN GERMANIUM AND IN INVER-TED CO2 (B. J. Feldmin, R. A. Fisher, I. J. Bigio, E. E.Bergmann)

Introduction

During this period, we generated phase-conjugate,10.6-um reflections from the phase grating establishedwith counterpropagating waves in germanium sampleslocated within an oscillating TEA double-discharge CO2

laser cavity as well as in the CO, gain medium itself. Thework reported* for germanium was the first demonstra-tion of nonlinear phase conjugation in the infrared, and,more generally, of intracavity degenerate four-wave mix-ing. The work within the CO2 gain medium is the first

37

demonstration of infrared phase-conjugation in an inver-ted medium. These intracavity techniques have generalapplicability for any laser system.

In such a conjugation process, a nonlinear interaction(such as stimulated Raman scattering, stimulatedBrill; "iin scattering, or degenerate four-wave mixing)causes a "reflected" wave, which is the amplitude com-plex conjugate of the incident wave. Characteristic of theprocess is the fact that a phase-conjugate reflectionexactly retraces the optical path of an incident wave,regardless of any phase distorters that may be in thepath of the incident wave. In contrast to a normal mirror(which inverts only the component of the k-vector nor-mal to its surface), a conjugating mirror inverts all threecomponents.

Conjugate optics can improve the double-pass opera-tion of laser amplifiers. As a high-quality optical beamtraverses an optical system, it is unavoidably distortedbecause of imperfections, improper surfaces, and in-homogeneities in the gain medium. If the distorted beamis reflected by a phase conjugator, it is redirected in sucha way that each distorted ray returns to itscorresponding aberrating spot and the beam emergeswith excellent optical quality. If so, large optical systemscould be arranged so that there are many elementswhose imperfect optical quality would not impair thepassage of a diffraction-limited beam, and this couldresult in large savings for high-power laser designs.

Among the many potential applications of such aphase-conjugator, the application to laser fusion isperhaps the most exciting. In one arrangement, the fu-sion target is diffusely illuminated- The scattered lightcould travel backward through the amplifier chain, bereflected by a conjugating four-wave mixing medium,and return through the imperfect optical elements; anamplified pulse wouM then impinge directly on thetarget, obviating the u-id for ultraprecise target align-ment. Such a phase-conjugating mirror will undoubtedlyfind other applications, such as providing stable laseroperation even in a gain medium experiencing time-varying inhomogeneities, and, possibly, diagnosing laser-plasma interactions.

How Four-Wave Mixing Produces Phase Conjugation

For the simplest example of this effect, consider amaterial in which the index of refraction depends linearlyupon intensity. Such a material would, for example,

cause strong self-focusing or strong self-defocusing(depending on whether the refractive index increased ordecreased with intensity). The nonlinear response can bedescribed as an index of the form n = n0 + n2 <E2>, oras a polarizability % = xm + X<3) E2. The coefficients arerelated through n2 = 2jix(3)/n0.

Consider waves E,, E2, and E3 (all of the same fre-quency) propagating through this nonlinear material.Waves E, and E2 are arranged to be precisely counter-propagating plane wave- ird both are assumed to be farstronger than E,. In the ,fied theory, the pair of op-positely propagating (. up) waves (at the same fre-quency co) in a material exhibiting a nonlinear index ofrefraction, provides the conditions for which a third(probe) wave E,, also at frequency to, incident on thematerial from any direction, would result in a fourthwave being emitted from the sample precisely retracingthe k-vector of the third wave. Although a nonlinearphenomenon is utilized, the effect is linear in the field tobe conjugated. Therefore this technique is far more at-tractive than, say, stimulated Brillouin backscattering.Because the effect is linear, a superposition of E,'s willgenerate a corresponding superposition of E4's, and onecan readily see that any complex field is time-reversed insuch an arrangement.

It is valuable to consider how the interference gives aconjugated E4 wave. Consider first the interaction of theweak wave E, and of one strong wave E,. The indexchange contains the term (E, + E3)

2, which includes across term corresponding to a phase grating perfectlyphase-matched so that the other strong wave E2 isBragg-scattered into the E4 wave. Concurrently, E2 andE, interfere to form a phase grating that scatters E, intothe E4 wave. This process is analogous to volumeholography in which the writing and reading are donesimultaneously.

Experimental Results in Germanium. Numerousdemonstrations of extracavity phase-conjugate reflectionvia degenerate four-wave mixing in the visible have beendescribed recently. These experiments used beams fromfrequency-doubled YAG lasers, from dye lasers, fromruby lasers, and from cw argon-ion lasers; four-wavemixing media included liquid CS2; resonant, near-resonant, and two-photon resonant absorbing sodiumvapor; and ruby crystal. Yet, before our work, no experi-ments had been performed at 10.6 um. Demonstratingthis effect in the infrared was made difficult because (1)

the effect has an <o2 dependence, and thus, is reduced bya factor of 400 from the corresponding visible effect, (2)very strong pump waves are needed, and (3) the pumpwaves must be perfectly collimated and preciselyretrodirected.

We recognized that one could reduce the constraintson the pump waves by requiring that they merely becomplex conjugates of each other. This means that if E,is diverging, E2 must be converging, and vice versa.Because the two counterpropagating waves inside an os-cillating laser cavity are already complex-conjugates ofeach other, if the nonlinear material is placed within thecavity of an oscillating stable-mode laser, the material isguaranteed to be a conjugator for light of the laser fre-quency. As an added bonus, the circulating power withina laser can be much higher than the output power.

In our first experiments,8'10 we obtained 2% phase-conjugate reflection by reversing a 95% flat germaniumoutput coupler so that the substrate was internal to theoptical cavity. The laser output was then used to probethe sample's conjugate reflectivity. In our recent experi-ments, we used thicker germanium samples. A 2.5-cm-thick. 95% reflecting end piece gave a conjugate reflec-tivity of 10%. whereas a 15-cm-long single-crystal bouleplaced in the center of the cavity gave 20% conjugatereflection efficiency. The effect did not scale simply asthe sample length squared because, as the length of thenonlinear sample was increased, the circulating powerwithin the cavity decreased. This behavior may be due tothe intensity-dependent formation of an electron-holeplasma" in the germanium, to surface damage, or tophase-front durations of the pump waves due to non-linear refraction. To verify that the gain medium doesnot impair the conjugation effect, we first passed thebeam through a different path in the amplifier on theway to the germanium conjugator. When including thegain of the amplifier, the effective conjugate reflectivityof the hybrid system was 40% (using a 0.5-cm-thick ger-manium conjugator).

Conjugate Reflection in Inverted CO,

In the second set of experiments, we chose to obtainconjugation using the nonlinear properties of theresonance associated with the CO, gain medium itself,and again, our intracavity technique used the counter-propagating waves already present in the oscillating lasercavity. As an added bonus of working with the non-

linearity of a partiaily saturated gain medium, the phase-conjugate signal is automatically amplified if the probebeam is directed through "unused" gain volume on itsway to and from the interaction region.

The nonlinear process used in the CO2 gain resonanceis not merely the four-wave mixing process; because ofsaturation, the effect is the coherent superposition offour-wave, six-wave, eight-wave, ... processes. Becauseof the ambiguity in the number of waves involved, weprefer to identify this mixing process as resonant light-by-light scattering. The theory of phase-conjugate reso-nant light-by-lighl scattering was first developed byAbrains and Lind.12 They used the well-known nonlinearsusceptibility of a homogeneously broadened two-levelabsorber (or amplifier) which, on resonance, sets up anamplitude grating that couples the standing field and theprobe field to generate the conjugate wave. As should beexpected, the maximum scattering efficiency is attainedwhen the strong counterpropagating standing wavesonly partially saturate the transition. Consequently, alimitation occurs because the depth of the amplitudegrating becomes reduced as the strong waves exceed thesaturation flux, and this, in turn, explains why the effectis not merely a third-order (four-wave) process.

In this experimental setup, our conventional hybridhigh-pressure/low-pressure TEA laser is operated with a10-m radius of curvature, a 35% reflecting mirror, and aflat 10% reflecting mirror. These unusually low reflec-tivities were chosen to reduce the circulating powerwithin the cavity. The laser operated on the P(20) line ofthe IO-|im CO2 branch; the pulse emitted from thecurved output coupler was ~150 ns (FWHM) and hadan energy of ~ 100 mJ. This corresponded to a cir-culating intensity of ~1.5 MW/cm2 inside the cavity.The output of the laser was attenuated, passed through abeam splitter, and redirected through the TEA laser gainmedium. Phase-conjugate reflectivity was diagnosed byusing an infrared vidicon or a high-speed SAT detectoron the return-signal reflection from the beam splitter.The phase-conjugate nature of the reflected light wasconfirmed by studying the spatial characteristics of thebeam in the presence of an aberrator and in the presenceof a cross-hair image placed on the beam in front of thebeam splitter. In both cases the initial incident probe-beam image was preserved in the reflected image. In asingle-pass configuration where the probe beam overlap-ped the lasing region of the gain medium only in the last10 cm of the 91-cm discharge length, effective reflec-tivity was 0.5% (reflectivity defined as the peak intensity

39

of the phase-conjugate beam divided by the peak inten-sity of the incident probe beam). In a double- and triple-pass arrangement, where only the last 10 cm of the finalpass through the gain medium overlapped, the lasing-region reflectivities were 35 and 250%, respectively. Theenhanced reflectivity in these cases is due entirely to theamplification of the outgoing probe and incoming phase-conjugate signal en route to and from the conjugationregion. The actual phase-conjugate reflectivity in the 10-cm overlap region is, in all cases, less than 10~2, in ac-cord with the Abrams and Lind theory12 for the field andgain-medium parameters involved.

Concluding Remarks

We demonstrated that a very simple intracavityarrangement is useful in obtaining phase-conjugatereflections. The technique is readily applicable to anynonlinear material placed in an oscillating laser cavity; ifthe laser operates, the material is guaranteed to be a con-jugate reflector for light of that frequency. Table III-IIIindicates the conjugate reflectivities we obtained.

By using this technique, we demonstrated for the firsttime a phase-conjugate reflection in the infrared usingboth germanium and the CO2 gain medium itself.

TABLE III-III

PHASE-CONJUGATION RESULTS IN THEINFRARED OBTAINED WITH GERMANIUM

CALCULATION OF SMALL-SIGNAL GAINCOEFFICIENTS IN CO2 (J. C. Goldstein)

The small-signal gain coefficient of a CO2 amplifier isfrequently used as a measure of the energy stored in thedevice, and therefore, as an indicator of the amplifier'sperformance. The stability of a laser amplifier systemwith respect to the growth of spurious parasitic modes ofoscillation—and the suppression of such parasitics bysaturable absorbers—is determined in large measure bythe magnitude and frequency variation of the small-signal gain coefficient. It is important, therefore, to beable to infer accurately the molecular energy-level pop-ulation densities from small-signal gain measurements,and conversely, to be able to predict the small-signalgain coefficient from calculated medium conditions(such as vibrational and rotational temperatures).

The usually straightforward relationship of the small-signal gain coefficient to molecular populations issomewhat obscured in CO2 by the presence of gain onmany closely spaced lines. The gain coefficient at anyparticular frequency is the sum of the gains, each oneevaluated at that frequency, contributed by many dif-ferent transitions. In addition to the contributions fromvarious rotational lines of the main vibrational transition(the 00° 1 -> 10°0 transition in the case of the 10.6-umband), there can also be contributions from transitionsbetween higher lying vibrational states due to near coin-cidences of some of the rotational lines of these transi-tions with those of the main transition.

We constructed a code to compute the small-signalgain coefficient at any specified frequency by summingover P- and R-branches of the following transitions inCO2.

InteractionLength (cm)

12.545.085.08

10.1615

Hybrid

Power(MW/cm2)

408.78.78.08.0

2512

Reflectivity (%)

20.82.33.4

10.32040

10 urn

00°l - .00°2-*00°3 - .O l ' l - » •

• 1 0 ° 0

- 10° 1- 10°2ll'O

mainsequencesequencehot

00000001

9 urn

° 1 —*•

°2 - •

°3 -*' 1 - > • 1

020202

D3'l

°0°1°20

40

This notation abbreviates the correct notation of thelower levels, which are coupled Fermi resonance pairs;e.g., "10°0" means |10°0, 02°0|, and "02°0" means|10°0, 02°0|,,. Only the principal isotopic variant isconsidered |CI2(016)2; natural abundance, 98.4%].

The code was used to compute the line-center absorp-tion coefficient of the 10.6-um P(20) line in pure CO2

and of mixtures of He, N2, and CO2 in thermalequilibrium at various temperatures and pressures. Theresults agree with recent published measurements13'14 towithin a few percent—better in most cases than theclaimed experimental uncertainty. In comparing thepredictions of the code with a recent gain measurement,"we varied the vibrational and rotational temperatures un-til we obtained good agreement between the code and themeasured value of the 10.6-um P(20) line-center small-signal gain coefficient. We then computed the line-centergain coefficients for other lines without changing anyparameter in the code. The results are shown in Figs. III-8 and -9 for various lines of the P- and R-branches of the10.6-um transition. The effects of including all overlapp-ing lines (as opposed to a single transition) in the calcula-tions are shown in Fig. Ill-10.

The code appears to calculate small-signal gain or ab-sorption coefficients in CO2 to an accuracy of a few per-cent. It is conceivable, however, that for higher tem-peratures and gas pressures, it may be necessary to in-clude more transitions in the calculation.

~ 4.25 -

DIRECT MEASUREMENT OF NONLINEARDIELECTRIC PROPERTIES IN GERMANIUM (D.E. Watkins, C. R. Phipps, Jr., S. J. Thomas)

Introduction

The use of germanium in high-power infrared lasersystems and in four-photon mixing16 and phase-conjugation17 experiments has renewed interest in itsnonlinear optical properties. Conversion efficiencies ob-tained in these experiments suggested very large valuesfor the infrared nonlinear susceptibility in germanium. Inaddition, more accurate knowledge of the nonlinear in-dex of refraction in NaCl and germanium are of greatimportance for our Helios and Antares laser systems, butno direct measurements of these crucial quantities in the10-urn region were available. We initiated some ap-propriate experiments and used time-resolved rotation ofthe axes of elliptically polarized light to measure the non-linear dielectric tensor in germanium at 10 umwavelength.

Owyoung" has shown that, for an ellipticallypolarized plane wave propagating along a 100 axis in atransparent cubic crystal, the angle of rotation can bedetermined when the major axis of the ellipse lies along a100- or 110-crystal axis.

u4.00 -

3.75 -

o£ 3.50

3.25 h

3.00

2.75

1

-

0

1

1

8

I

- r r

8 o

1 I

i 'i •

9

I i

• T

o

1

- r

0

1

LELAND'

•

* go

I I

CODEo

-

0_

110 12 14 16 18 20 22 24 26 28 30

LINE NUMBER, I. BRANCH

Fig. III-8.Calculated and measured16 line-center gain coefficients for several lines of the P-branch of the 10-\im transition.

41

o

1

-

o

1

1

o

1

1

o

1

1

0

1

1

o

1

1

§

1

1

0

1

1

0

1

1

LELAND

o

I

i

CODE0

-

o

1

^ 3 . 7 5

z 3.50LU

t 3.25LOoo

% 3.00CO

27510 12 14 16 18 20 22 24 26 28 30

LINE NUMBER, R BRANCHFig. III-9.

Calculated and measured16 line-center gain coefficients for several lines of the R-branch of the 10-uw transition.

4.25

Tg" 4.00uJ

3 3.75

i—

S 350u_

S 325

S 3.00CO

275

I

-

-

0

*

I I

0

0 •

•

1 1

1 !

o

0 •

•

1 1

1 ••[ 1 1 1

o

NO OVERLAPWITH OVERLAP

0

* • °• o

o

1 1 1 1 1

1

o

-

-

-

0

_

110 12 14 16 18 20 22 24 26 28 30 32

LINE NUMBER, 2 BRANCH

Fig. 111-10.Comparison of calculated gain coefficients showing the effects of neglecting overlapping spectra

from adjacent lines.

Results

An elliptically polarized CO2 laser beam was passedthrough samples of Ge, CS2, and NaCI. A polarizationsplitter was oriented at 45° to the axis of the ellipse. The

separated S- and P- polarized signals were time-resolvedvia fast pyroelectric detectors to obtain peak intensitiesI5 and Ip. The ratio of these intensities can be related tothe angle of ellipse rotation, and hence, to the compo-nents of the nonlinear susceptibility of the material.

42

1.0

0.90

080

M 0.70

0.60

a.H

0.50

ELLIPSE ROTATION IN 12cmINTRINSIC Ge, 110 AXIS.,,1221 , , ,^ll* 3 = I 3x10 esu

O.I 1.0 10 100

PEAK INTENSITY (MW/crrf

Fig. Mil.Dependence of ratio IS:IP on peak intensity for thevalue of nonlinear susceptibility %j given. The dotsare experimental values.

Our system was calibrated against CS2. In 10 cm ofCS2. we found x?n = (3.9 ± 1.0) X 10~13 esu comparedto (4.7 ± 1.6) x I0"13 esu reported by Moran et al.19

Measurements made on a 12-cm-long single-crystalintrinsic germanium sample are shown in Fig. Ill-11.The values obtained from this sample are x\u = (2.5 ±0.3) X 10"" esu and x!"1 = (1.3 ± 0.2) X 10"" esu.These values are substantially lower than previouslyreported. Wynne20 used third-order optical mixing to in-fer x\H = (1 ± 0.5) X 10-10 csu and x\21i = 0.6 x""-However, recent work in four-photon mixing16 andphase conjugation17 has produced smaller effects thanexpected on the basis of Wynne's work.

A 17.8-cm-long piece of polycrystalline NaCIproduced no measureable result. This experiment wouldhave been sensitive to x'221 = 3 X 10~14 esu in NaCI,consistent with Smith et al.,21 who found x\221 = (2.6 ±1.3) X 10~14 esu via four-photon mixing in NaCI at 1.06urn.

The results for germanium indicate a lower efficiencyfor four-photon mixing and phase conjugation thanpreviously assumed. Also, they imply a 1/4-wavelengthdistortion for intensity-length products of ~0.5 GW/cm,which has been confirmed by brightness degradationmeasurements.22 In NaCI, these measurements implythat 1/4-wavelength phase distortion will occur forintensity-length products that are greater than 45GW/cm. Further measurements with 20 times greatersensitivity are under way to correct the upper boundreported here with accurate measurements of %3 in NaCI.

SUBNANOSECOND EXTINCTION OF CO2

LASER SIGNALS VIA BULK PHOTOIONIZA-TION IN GERMANIUM (C. R. Phipps, Jr., S. J.Thomas, D. E. Watkins)

Introduction

Recent reports of efficient phase conjugation23 andfour-wave mixing2'1 of CO2 laser signals in germaniumhave renewed interest in the nonlinear optical propertiesof this material. The ultimate efficiency of theseprocesses is limited by the magnitude of the nonlinearsusceptibility and by the intensity-length productavailable in the interaction region. The former issomewhat smaller than had been supposed,25 and theuseful interaction length has practical limits set by loss,nonlinear refraction, or the signal duration itself.

We found that a definite intensity limit also exists forpropagation in intrinsic or optical-grade germanium,determined by the sudden onset of bulk photoionizationat levels only moderately higher than those required forefficient mixing or conjugation. The threshold intensitywas relatively independent of wavelength in the 9- to 10-um region, and varied only slightly for pulse durationsranging from 1 to 50 ns. lonization thresholds werelower in optical grade than in intrinsic material, butmuch higher in p-type gallium-doped germanium.Because the saturation intensity26 of the intraband holeabsorption is 3.2 MW-crrT2, and because the Abrams-Lind Model27 predicts efficient phase-conjugation insuch saturable absorbers, the increased threshold in p-type germanium may permit the attainment of four-wavemixing efficiencies considerably higher than those seenpreviously.

Experiments and Results

The effects studied were completely reversible, usingintensities well below surface damage thresholds. Bulkionization was revealed by transverse photoconductivesignals with decay times varying from nearly 1 ms inpure germanium to 50-100 us in optical-grade material.These results were consistent with absorption decayrates observed in low-power, 9-um probe measurements.Transmission was measured for intense, collimated,single- or multiline CO2 laser signals in samples ofvarious lengths. Both time-averaged and instantaneousquantities were measured; time resolution of ourdiagnostic system was 100 ps.

43

The time-resolved data for uniform illuminationclearly show that the instantaneous, high-intensitytransmission function drops to zero (irreversibly onnanosecond time scales) whenever the input intensity ex-ceeds a critical value Ib « 200 MWcm"2. The peak out-put intensity is essentially a constant, and the fall time ofthe output signai can be as small as 400 ps, indicatingvery rapid plasma formation.

Bearing in mind that absorption in germanium isdominated by the intravalence-band transitions in thisspectral range, and taking proper account of saturation,we infer peak photoplasma densities of 2 X 10" cm"3

from the experimental transmission data, as indicated inFig. 111-12.

We also used a second method to measure the non-linear transmission properties of germanium, with the in-put CO2 laser beam brought to a real internal focus. Weobserved that internal damage was prevented by plasmaformation. In addition, the transmission for a space-timeGaussian beam in this configuration was highly predic-table over four orders of magnitude in input intensity.Applications of this highly reproducible transmissionfunction include passive intensity-limiting for detectorprotection, and passive pulse-shortening. An additionalapplication of this work is passive intensity-range com-pression for more effective comparison of prepulses andmain pulses in target-illumination diagnostics, where the

50 100 200 300 400 500PEAK INPUT INTENSITY (MW cm"!)

Fig. 111-12.Free carrier densities vs input intensity inferredfrom residual transmission measurements on ger-manium samples for lengths of 5.7 cm (0), 9.7 cm(A), and 14.2 cm (*). The trend line represents n ocI6.

intensities to be compared differ by many orders ofmagnitude.

HIGH PRESSURE ENRICHMENT OF LIGHTISOTOPES (J. B. Marling, I. P. Herman, LLL; S. J.Thomas)

Introduction

A series of joint LLL-LASL experiments, designed todemonstrate enhanced deuterium enrichment in certainFreon compounds by using ns-duration CO2 lasersignals, was completed successfully at LASL. Higherthan 1000-fold enrichment was obtained, and the shortlaser pulses permitted the use of relatively high gaspressures for this type of work.

Laser separation of light isotopes by multiple-photondissociation is becoming more attractive, using pulsedCO2 lasers as an excitation source. Very high single-stepenrichment factors were observed when the proper work-ing molecule was chosen. For deuterium separation,single-step enrichment factors above 1000 with highyield have been obtained in 2,2-dichloro-1,1,1, -trifluoroethane (Freon 123) at 0.5 torr using standardCO2 TEA laser technology near 10.65 urn.28 For 13Cseparation, single-step enrichment factors up to 65 wereobserved in CF,Br at room temperature,29 but enrich-ment dropped rapidly above 1 torr.2'

Extension of these performance parameters into the100-times higher pressure regime is essential to reducegas pumping costs to competitive levels and to achieve arealistic photochemical reactor size, especially fordeuterium separation. Deuterium and 13C enrichment us-ing these molecules was studied experimentally and isreported here as a function of pressure and incidentfluence for a variety of CO2 laser pulse lengths (2 to1000 ns).

Experiment and Results

The CO2 laser used was either an oscillator-amplifierwith variable short-pulse duration or (at LLL) a stan-dard TEA laser with a 100-ns gain-switched initial spikefollowed by a weak 1-us tail. The LASL oscillator wasmodelocked with a 12-ns switchout gate for selection ofa single 2-ns FWHM pulse for amplification. With themodelocker turned off, a 12-ns FWHM pulse wasswitched out for amplification. Pulses of 80 ns without

44

tails were available by bypassing both the switchout andmodelocker, and by controlling the laser gas composi-tion.

We evaluated deuterium using mixtures of CF,CHC12

(Freon-123) containing ~0.4% CF3CDC12. We also per-formed several experiments using Freon-123 withnatural deuterium isotopic abundance (~0.015% D).The primary deuterium-containing photoproduct istrifluoroethene. (CF, = CFD).28 and the isotope contentof the product was determined by mass spectrometry.The CF,Br (bromotrifluoromethane) decomposes toyield "C-enriched CF3CFj as primary photoproduct;2'CF,Br of natural isotopic abundance, containing ~1.1%13C. was always used. For deuterium separation, weused the P(26) 10.65 urn line for photolysis, and for I3C,we used the P(24) 9.59-um CO2 laser line.

In the initial "C enrichment studies in CF,Br, repor-ted in Ref. 29. the low-pressure single-step enrichmentfactor of 65 dropped 50% by 0.8 torr and dropped 10-fold by 3 torr. By using 2-ns CO2 laser pulses in ourstudy, the single-step I3C enrichment factor at 3 torr was43. It decreased as a very slow function of pressure to 27at 100 torr. and to 22 at 225 torr.

Deuterium enrichment was measured by using 12-nsdissociating pulses at a peak focal fluence of 8 to 10J/cm2. The single-step deuterium enrichment factor wasconstant at 900 ± 50 in the trifluoroethene product overthe range 1 to 600 torr. However, the yield at 100 torrwas too low to determine isotopic enrichment. Hence,high pressure quenched the photoproduct yield withoutdecreasing the enrichment. Use of shorter (2-ns) dis-sociating pulses again resulted in near-1000-fold enrich-ment, but with apparently higher yield because we ob-tained a sufficient amount of photoproduct at 100 torrfor isotope analysis. Freon 13 of natural isotopic abun-dance was also examined, and yielded a deuteriumenrichment factor of 1900 ± 400 at 30 torr, and of 1500± 100 at 100 torr. The use of relatively long (80-ns) dis-sociating pulses did not significantly decrease theisotopic enrichment at higher pressures. However, ahigher fluence of 20 J/cm2 was used, reducing overallenrichment (see Fig. 2 of Ref. 28), but it remained ashigh as 220 ± 50 throughout the pressure range from 0.3to 30 torr.

In ir photolysis of a sample in which the minorityisotope is the absorbing molecule, the effect of collisionson yield should be analyzed as a function of laserfluence. Above the saturation fluence (for the resonantlyabsorbing molecule of the minority), an increase inpressure can lead to increased ir dissociation of the non-

resonant species, which can absorb photons only aftercollisions with the resonant molecule. The increase oc-curs without greatly decreasing the association yield ofthe resonantly absorbing minority isotope. Therefore, athigh fiuences the isotopic enrichment will fall with in-creasing pressure, perhaps with little change in yield.However, at fluences between dissociation threshold andsaturation, collisions will quench the resonant isotope,and the nonresonanl species, therefore, will not be ableto dissociate. Consequently, as pressure increases, netyield will fall, but isotopic enrichment may vary onlyslightly. However, we ignore the importance of collisionsin removing the multiple-photon absorption bottleneckthat appears at very low fluences. Our conclusions seemto be especially true for deuterium separation in Freon-123 where the single-step enrichment factor is veryfluence-sensitive.28 but showed only slight sensitivity tooperating pressure.

Conclusion

The practicality of using lasers for the separation oflight isotopes has been greatly enhanced bydemonstrating that high single-step isotope enrichmentcan be maintained even at pressures up to 1/3 atm.Isotopically selective multiple-photon dissociation wasachieved with 2-ns CO2 laser pulses, permitting 1500-fold enrichment of naturally occurring deuterium in pureFreon-123 at 100 torr and 22-fold enrichment of I3C inCF,Br at 225 torr.

The LASL part of this work was supported by the In-ertial Confinement Fusion Program because of the ex-istence of high-intensity, short-pulse CO2 lasers. Im-proved efficiency of deuterium separation will benefit thefusion program by providing an improved fuel source.

REFERENCES

1. F. Skoberne. Comp., "Laser Fusion Program, July1—December 31, 1977," Los Alamos ScientificLaboratory report LA-7328-PR (December 1978).

2. B. J. Feldman, "Multiline Short Pulse Amplificationand Compression in High-Gain CO2-Laser Am-plifiers," Opt. Commun. 14, 13 (1975).

45

3. H. C. Volkin, "Calculation of Short-Pulse Propaga-tion in a Large CO2-Laser Amplifier," J. Appl.Phys.. to be published.

4. R. J. Harrach. "Effect of Rotational and IntramodeVibrational Coupling on Short-Pulse Amplificationin CO,." IEEE J. Quantum Electron. QE-11, 349(1975).*

5. E. E. Stark. Jr., "Measurement of the 10°0-0-02°0Relaxation Rate in CO2." Appl. Phys. Lett. 23, 335(1973).

6. R. R. Jacobs, K. J. Pettipiece, and S. J. Thomas,"Rate Constants for the CO2 020-100 Relaxation,"Phys. Rev. A 11. 54 (1975).

7. G. T. Schappert, "Rotational Relaxation Effects inShort-Pulse CO2 Amplifiers," Appl. Phys. Lett. 23,319 (1973).

8. F. Skoberne, Comp.. "Inertial Fusion Program,January 1—June 30, 1978," Los Alamos ScientificLaboratory report LA-7587-PR (May 1980).

9. J. I. Steinfeld, 1. Burak, D. G. Sutton. and A. V.Nowak. J. Chem. Phys. 52, 5421 (1970).

10. E. E. Bergmann, 1. J. Bigio, B. J. Feldman, and R.A. Fisher, Opt. Lett. 3, 82 (1978).

11. C. R. Phipps, Jr., Los Alamos ScientificLaboratory, private communication (1978).

12. R. L. Abrams and R. C. Lind, Opt. Lett. 2, 94(1978), and Erratum, Opt. Lett. 3, 205 (1978).

13. A. M. Robinson and Y. -K. Hsieh, J. Appl. Phys.48. 1589 (1977).

14. A. M. Robinson and N. Sutton, Appl. Opt. 16,2632(1977).

15. W. T. Leland, M. J. Kircher, M. J. Nutter, and G. T.Schappert, J. Appl. Phys. 46, 2174 (1975).

16. N. Lee, R. Aggarwal. and B. Lax, Opt. Commun. 9,401 (1976).

17. E. Bergmann. I. Bigio, B. Feldman. and R. Fisher.Opt. Lett. 3. 82 (1978).

18. A. Owyoung. I.E.E.E. J. Quantum Electron. QE-9,1064 (1973).

19. M. Moran, C. She. and R. Carman, I.E.E.E. J.Quantum Electron. QE-11, 259 (1975).

20. J. Wynne. Phys. Rev. 178, 1295 (1969).

21. W. Smith. J. Bechtel, and N. Bloembergen, Proc.Laser Induced Boulder Damage in OpticalMaterials. 7th Annual Symp. (1976), p. 321.

22. C. Phipps, Jr., Los Alamos Scientific Laboratory,unpublished data (1978).

23. E. E. Bergmann, 1. J. Bigio, B. J. Feldman. and R.A. Fisher. Opt. Lett. 3, 82 (1978).

24. N. Lee. R. L. Aggarwal. and B. Lax, Opt. Commun.19. 401 (1976).

25. D. E. Watkins. University of Washington, un-published data (1978).

26. C. R. Phipps. Jr., and S. J. Thomas, Opt. Lett. 1, 93(1977).

27. R. L. Abrams and R. C. Lind, Opt. Lett. 2, 94(1978).

28. J. B. Marling and I. P. Herman, Appl. Phys. Lett. 34(February 1979).

29. M. Gauthier, P. A. Hackett, M. Drovin, R. Pilon,and C. Willis. Can. J. Chem. 56, 2227 (1978).

46

IV. TARGET EXPERIMENTS, DIAGNOSTICS, ANDMILITARY APPLICATIONS

(G. H. McCall)

In an integrated program of target experiments, theory, and design, we are es-tablishing a fundamental understanding of laser-target interactions, particularly of therelevant plasma physics and hydrodynamics. Experimental and theoretical efforts haveaddressed the scaling of consistent models to higher laser intensities. Emphasis has beenon the development and demonstration of experimental techniques that will determineconclusively the performance of present and future targets. A modest experimental ef-fort is directed toward military applications.

The laser fusion process calls for new diagnostic techniques having spatial and tem-poral resolutions in the submicrometer and 1- to 100-ps regime. These needs are beingmet with a vigorous diagnostics program in such areas as laser calorimetry, charged-particle and neutron detection, x-ray spectrometry, and subnanosecond streak cameradevelopment.

EXPERIMENTS (R. P. Godwin, D. B. Giovanielli)

Introduction

This reporting period was an exciting one. The Helioslaser is maturing as a complete laser-fusion researchfacility. It has delivered previously unattainable CO,laser power and energy to various targets. Exploding-pusher GMB targets served as a test of the new facilityand verified our confidence in the understanding and in-tensity scaling of those targets. Neutron yields exceeding10" were obtained for the first time using CO2 radiationpowers of 3 TW. In our first experiments, we usedplastic-coated GMB targets designed to achieve high fueldensity, but relatively low temperature. These targetswere direct predecessors of long-range fusion reactortargets. Their neutron yields (up to 10*) provided enoughdata for statistically meaningful parametric studies. Thisclass of targets is sensitive to such important aspects oflaser fusion as hot-electron preheat, which is unimpor-tant in exploding-pusher experiments. Although thesehigh-density targets are difficult to diagnose by simpletechniques that provide uniquely interpretable informa-tion, initial experiments are providing information con-sistent with our code calculations and are determiningour diagnostic development efforts, particularly in theareas of neutron diagnostics with high temporal resolu-tion and x rays. The measurement of the fastest plasmaprotons at Helios gives additional support to the hot-electron temperature scaling as (IA.2)1/3, where I is the

focal power/area and X the laser wavelength, which hasproved to be a useful laser diagnostic. Energy transportquestions continue to be dominant in laser-plasma in-teractions, particularly in the interpretation of experi-ments. Conceptually simple experiments performed atGemini showed that electron currents flowing towardtargets through support structures, and through targetpotentials of several hundred kV may be important vec-torial factors in transport—-factors previously discoun-ted in hydrodynamics calculations.

Helios Exploding-Pusher Experiments (D. Giovanielli)

Exploding-pusher experiments were the first to be un-dertaken at the Helios laser facility. Because thebehavior of such targets is well known, they were chosenfor the first full-scale integration of laser operations,beam diagnosis, alignment and focusing, target handling,and target diagnosis systems. Performance of all of thosesystem elements can be normalized against earlier ex-perience.

In the first sequence, 32 DT-filled GMB targets wereirradiated with total laser energies varying from 750 to4100 J. Simultaneity of the eight Helios laser beams waspreviously measured to be within 50 ps. Several targetswere irradiated with fewer than eight beams, but thesuprathermal electrons, which drive the explosion of thethin-walled glass shell, symmetrized the target. Therewas no loss of ability to calculate target thermonuclear

47

yield when two or more laser beams were used toirradiate the targets (single-beam experiments were notattempted).

Target diagnostics included neutron yields, fast(isothermal front) ion velocities, x-ray pinhole images,high-resolution x-ray spectra (Si lines), and multichannelx-ray spectra from 10 eV to 4 keV.1 The high-resolutionspectroscopy used instruments placed within 3 cm of thetarget position. No reduction in parasitic oscillationthreshold of the laser amplifiers could be observed insuch proximity. X-ray line spectra displayed features notpreviously observed, such as both hydrogen-like andhelium-like silicon continuum edges simultaneously,allowing triply redundant temperature determinations.The successful operation of these instruments verifiedour ability to design and produce thin filter windowassemblies that can be used near the Helios targets. In-tegration of the multichannel spectrometer and x-raydiode data yielded x-ray emission efficiencies (incidentlaser light re-emitted as x rays) of 2%. in agreement withpreviously obtained values.

X-ray pinhole camera images from seven shots areshown in Fig. IV-1. Rippling of the inner imploded glasssurface is suggestive of hydrodynamic instability:however, because laser light remains incident long afterthe time of peak compression, and these are time-integrated photographs, deduction of quantitative infor-mation from this apparent shell breakup is difficult.Radiochemical diagnostics now being developed may bemore sensitive to shell breakup at the time of peak com-pression and thermonuclear burn.

Table 1V-I lists the initial target parameters,equivalent trapezoidal laser pulse shapes, calculated andmeasured neutron yields, and velocities of the fastestions observed. The ion velocities match data obtained atthe Single Beam System and Gemini, and lead to a scal-ing of suprathermal electron temperature with incidentlaser intensity of I1", in agreement with simulationresults. Neutron yields were calculated by using ananalytical model2 and the measured target and laserparameters. The absorbed laser energy fraction was notmeasured in these experiments; however, in the calcula-tions of expected yield we assumed 25% absorbedenergy, as measured previously with spherical targets.3

We assumed 35% of the absorbed energy to be lost toreradiation, fast-ion production,4 suprathermal electronemission,5 and conduction along the supporting stalk.6

The excellent agreement between measured andcalculated neutron yields confirms both the model andour input assumptions.

Evidence of Adiabatic Implosion of GMBs (T. H. Tan)

A most significant result from our target experimentsat Helios is evidence of adiabatic implosion. Ft is essen-tial for the success of the laser fusion program thatadiabatic implosion processes be investigated as soon aspossible. Exploding-pusher experiments have contributedvaluable information, but the conditions that exist inthese targets are far from those that are believed requiredto produce breakeven of higher yields. Helios nowproduces enough energy on target so that, for the firsttime, we can design targets with predicted densitieshigher than that of solid DT. and we can produce thesame adiabatic conditions in the fuel as those of currentbreakeven designs.

The 20XLD goal (4 g/cnv') was chosen even thoughpredictions indicate that, for certain designs, muchhigher densities can be reached by targets irradiated byHelios. Also, it is important to study phenomenaassociated with the early adiabatic phases of laser-drivenimplosions, and complicated targets depend on poorlyunderstood physical processes that are extraneous to thestudy of the implosion itself. Therefore, we needed assimple a target as possible so that the implosion systemcould be diagnosed in a straightforward way that wouldrequire only modest extrapolation from existing ex-perimental results.

The achievement of adiabatic compression to high fueldensities requires, first, adequate shielding of the targetinterior against laser preheat, and second, efficient con-version of absorbed laser energy into compressional mo-tion. In our targets, shielding from hot-electron preheatis accomplished by adding a low-density paryleneablator to a bare GMB (exploding-pusher) target. Thelow-density material allows a geometrically thickablator. This results in the largest possible outer radiusfor a target of given mass. Assuming the laser beams canbe defocused to flood the target surface nearly unifor-mly, thereby minimizing the intensity, the energy spec-trum of hot electrons produced during laser absorptionshould be cooler for these than for smaller targets. Thus,less material is required for hot-electron preheatshielding because the electron range decreases rapidlywith energy.

With such targets, we have measured neutron produc-tion time-delays that are significantly longer than the 1-ns laser pulse, implying that the implosion is sustainedadiabatically after the laser pulse has been turned off.The experimental data are crucial for demonstrating the

48

881107 06 881107 08 881107 09

i l I f—500/im

881222 04 881222 06

890201 08 890202 02

Fig. IV-l.Time-integrated x-ray pinhole camera images from seven exploding-pusher shots at Helios.

4 9

TABLE IV I

INITIAL TARGET PARAMETERS AND DATA FROM 32 EXPLODING PUSHED EXPERIMENTS AT HELIOS USING DT FILLED GMB TARGETS

Initial Target Parameters Equivalent Trapezoidal Laser Pulse Neutron Yield

Shot

Number

88100517

88101215

88102411

88102602

88102603

88102604

88IO2AO5

R8102607

88110705

881IO7O6

88110708

88110709

88110916

88112803

88113002

88113004

88113005

88113007

88113008

88122107

88122204

88122206

89010405

89OI23IO

89012311

89012903

89012904

89012905

89013003

89013005

89020108

89020202

Radius

(urn)

196

162

150

168

117

159

258

275

173

217

189

198

225

195

195

200

183

144

140

101

148

145

099

168

161

175

165

158

157

142

148

146

Wall

Thick

ness

(Mm)

0.99

1.29

1.0

1.22

1.08

0.99

1.25

1.01

1.57

0.98

0.88

1.18

1.16

0.99

0.95

0.83

0.85

0.99

1.02

0.99

1.05

1.13

1.1!

0.58

1.03

1.12

0.85

0.88

1.16

I.I148

0.80

DTFiH

(mg/cm')

2.0

2.0

fi.75

2.0

2.0

5.85

2.25

1.58

2.0

2.25

1.8

2.25

2.255

6.75

2.0

2.0

2.25

6.75

6.5

1.8

6.75

6.75

2.0

2.25

2.25

2.25

2.25

2.25

6.75

6.75

1.05

6.5

IonFront

Velocity

(10'cm/s)

3.40

3.32

3.45

3.12

2.30

3.21

2.83

3.54

2.95

2.89

3.07

3.04

•2.97

3.25

3.19

3.39

3.39

3.20

2.60

3.20

2.52

2.52

2.50

2.50

2.90

2.40

2.30

2.76

2.90

2.95

6.75

2.90

Rise

time

(psl

374

471

320

360

375

374

221

374

2.90

Peak

Time

<ps)

115

068

071

110

198

115

097

115

FaNTime

(ps)

855

636

989

763

729

855

1179

1191

1179

Total

Energy

(J)

2905

2189

3990

3984

753

3217

2154

1494

1428

2476

2329

1840

2296

1770

2693

313.1

3015

2529

1537

2356

1841

M616

4100

717

1395

6X9

684

1444

2280

3040

2820

2634

Peak

Power

(TW|

4.0

3.0

5.5

5.5

1.0

4.4

3.0

2.1

2.3

3.4

3.5

2.5

3.2

2.4

3.1

3.5

3.3

2.8

1.7

2.6

2.0

1.8

5.1

0.79

1.5

0.76

0.76

1.6

2.5

3.4

3.1

2.9

No. of

Laser Beams

7

7

7

8

2

7

54

4

6

6

6

6

4

6

7

7

6

4

8

6

6

7

2

4

2

2

4

6

8

8

6

Obs.

(10')

1.6

1.42

0.42

2.0

.066

0.8

0.26

0.41

2.3

3.5

1.1

2.3

0.22

2.2

1.6

0.57

.05

0.14

.045

.055

0.13

.04

0.3

0.4

1.3

0.56

0.51

Cak.

(10")

5.4

1.7

3.4

8.1

.074

2.3

2.2

0.79

0.41

5.4

4.2

1.25

2.9

0.27

2.7

4.2

3.1

0.48

0.10

0.53

0.18

0.12

6.7

.042

0.29

.021

0.3

0.34

0.39

0.90

0.71

0.49

Obs.

Cak

0.30

0.84

0.12

0.25

0.89

0.35

0.33

1.00

0.43

0.83

0.88

0.79

0.81

0.81

0.38

0.18

0.10

1.40

0:25

0.46

0.45

1.33

0.88

1.03

1.44

0.79

1.04

achievement of high fuel compression and high burn ef-ficiency by inertial confinement. They also provide animportant check on the reliability of the code predictionthat is so necessary to the laser fusion program.

These, results were obtained with GMB targetstailored for systematic investigation of high-densitypellet compression. Each GMB is 300 urn in diameterwith a wall thickness of I urn. They are coated with dif-ferent thicknesses (0-100 urn) of parylene and filled with30 atm of DT gas.

The implosion time is measured as a delay in excess ofthe expected arrival-time difference between the x-rayfiducial and the prompt-neutron pulse from anexploding pusher target. Ultrafast TOF detectors andscopes are used.

Targets with coatings of parylene 54-62 urn thickproduced consistent data. By increasing the amount oflaser energy on target from 2.2 to 3.0 kJ, the neutronyield increased from 9 x I06 to 2.6 X IO\ while theneutron production time delay decreased from 2 ± 0.5ns to 1.5 ± 0.5 ns. in agreement with code predictions.Longer drive in time is due to smaller energy transferresulting in lower fuel temperature and smaller yield. Thelongest implosion time was 2.8 ± 0.5 ns on a target witha plastic coating 88 um thick and illuminated by 3.6 kJof laser energy. It produced 3 X 10* neutrons. X-raypinhole photography corroborated the behavior of thesetargets that we inferred from neutron measurement.

Preliminary data and code predictions both indicate aconsistent relationship among implosion time, yield, andcompressed fuel density.

Measurement of Fastest ton Velocity (T. H. Tan)

The measurement of the fastest ion velocity has been apowerful plasma diagnostic for laser-matter interactions.These fastest ions, which are protons with velocities ex-ceeding 10* cm/s. were first detected by us in measure-ments done at KMS Fusion with a scintillator-photomultiplicr detector designed for measuring chargednuclear reaction products.

We reported previously' that v, cc iji,"6. where v, is thefastest ion velocity and «(», is the laser flux. From theisothermal expansion model (Ref. 8) v, oc Cs, where Cs isthe ion sound speed, the hot electron temperature Te canbe related to <j>, as Tc oc fyui. This agrees with computerplasma simulation work.* which suggests the prominenceof resonant absorption and strong densitv-profilesteepening due to the presence of the ponderomotiveforce. The data have been extended using the Heliossystem to about 5 X 10" W/cm2. Figure IV-2 shows thedata points ranging over four decades in intensity. Therelationship established earlier at lower intensities re-mains valid.

Fluctuations of ion velocity for a given <)>, are due tothe systematic ^certainties in the laser pulse and plasmainteraction parameters rather than to any statistical errorassociated with the measurement. Possible uncertaintiesinclude laser risetimc and pulselength. focused spot ontarget, presence of hot spots in the beam, low-levelprecursors, and coherent effects in the simultaneouspresence of several laser beams. Analysis reveals that the

IOO

u

9 IO

uoS

io1 4

Vc*",ft>0 17

r.-ihoi,

I0«su m l j i i l i i i i i i . i. i i.i n i l i_

Loser Flui __„ _ i _ *•

IO15 1 0 *

(Laser Klu»)Xz,

10" 10"

10" 10" ~

Fig. IV-2.Plot of the fastest ion velocity (vj vs laser flux on target.

51

cause of variation in ion velocity can be accounted forby these kinds of uncertainties.

For any target design, two critical code inputs thatgenerate the simulated implosions are fraction of energyabsorbed and hoi-electron temperature. The agreementof the fast-ion data with the simulations implies that theassumption of Tc_ oc §['3 is still valid at an intensity of10" W/cm2. Because the analysis assumes a constantabsorption fraction, it suggests that the absorptance isthe same as that measured at lower intensity.10 Both in-puts were incorporated into our high-compression targetdesign. The agreement between the experimental resultsand calculations" bears out the diagnostic capability offast-ion detection.

We have also established direct correlation betweenfast ion spectrum and target implosion as seen by the x-ray pinhole camera. The abrupt appearance of thefastest ions essentially guarantees a bright core in the im-plosion center. Furthermore, for DT-filled GMBs, themeasured neutron yield increases as a direct function offast-ion velocity. This implies that, as long as the targetshell thickness is comparable to or less than the range ofthe hot electrons, which is true in the case of exploding-pusher targets, the yield is determined primarily by hot-electron temperature, and hence, by the laser irradiance.

Line and Continuum Spectra from Laser-IrradiatedGMB Targets (K. Mitchell, A. Hauer, W. Priedhorsky)

X-ray spectroscopy provides a variety of informationcrucial to the understanding of laser-plasma interactions.

O 35

High-energy bremsstrahlung provides the temperature ofsuprathermal electrons important to fusion target designbecause of the fuel preheat they cause. Low-energybremsstrahlung is important to the energy balance oflaser fusion targets and provides the thermal electrontemperature. Line spectra offer a multitude of diagnosticpossibilities and very selective sampling of various targetregions by using the lines of various materials.

Very strong silicon spectra were observed from 300-um-diam GMB targets irradiated by the Helios laser.Two spectra taken at laser energies of 2 and 4 kJ areshown in Figs. IV-3 and -4. These spectra are spatiallyintegrated so that the radiation is observed from both thelaser interaction region and the compressed core. Thespectra are qualitatively different from those obtained inlower energy experiments. This is the first time that free-bound continua have been seen from 10-um laser-produced plasmas. If the log of continuum intensity isplotted as a function of energy, the electron temperaturecan be estimated from the slope of the curve. The tem-perature indicated from the (Is2-continuum) spectrumshown in Fig. IV-3 is 310 eV. Figure IV-4 shows boththe "helium-like" (Is2-continuum) and "hydrogen-like"spectra, but they exhibit different slopes. This indicatesthat the two cortinua are emitted from different regionsin space or m emitted at different times.

The arrow on the x-axis in Fig. IV-3 shows the posi-tion of the helium-like (Is2-continuum) recombinationedge for an atom in vacuum. As shown, the actual posi-tion of this edge is shifted considerably toward lowerenergies. This effect is caused by high-density plasmamicrofields that reduce the ionization potential. For the

3 0 20

0 10

1.7 IS 19 2.0 ;• I V / ? .5 ? 4 } ? S 7.6

Fig. IV-3.Spectrum from 300-\im-diam GMB irradiated with laser energy of 2 kJ.

52

IB 20 2 2 2.4 X C. PB JO

Fig. IV-4.Same as Fig. IV-3 at 4 kJ.

case shown, the helium-like recombination edge isgradual, which indicates a density gradient in the source.The density of the compressed silicon determined by thismethod is at least IO23/cm3.

The temperature of 310 eV, derived from the slope ofthe free-bound continuum, appears to be inconsistentwith the broad-band temperature estimate of 580 eV. Alimited number of space-resolved spectra from 200-umGMB targets show the free-bound continuum emittedprincipally from the core region. Also intensity ratios ofhydrogen-like and helium-like lines indicate that the coreis hotter than the coronal region. As far as temperaturemeasurements are concerned, these data appear to raisemore questions than they answer. More space-resolvedspectra from GMB targets are expected in the nearfuture. Also, we are proceeding to analyze the intensitydistribution among the "satellite" lines.

Broad-Band X Ray Spectral Measurements on GMBTargets (P. Lee, K. Mitchell)

Broad-band spectrometers were used to record x-rayspectra in the range 30 eV to 5 keV for several Heliosshots >2 kJ on GMB targets. Spectral discriminationwas obtained with K- or Ledge absorption filters. Spec-tra in the region 30 to 600 eV were recorded with XRDs,whereas spectra from 780 eV to 5 keV were recordedwith NE HI fluor photomultiplier-tube detectors. Exam-

ples of spectra from two shots on 300-u.m-diam GMBtargets arc shown in Fig. 1V-5. Spectra in the range 1 to5 keV are identified principally as x-ray continua due tofree-free bremsstrahlung. There is some blend of line andfree-bound continuum that contributes to the signal, par-ticularly in the 2.6-keV channel, but this is not evidentfrom these data. Assuming x-ray continuum, the slope ofthe curve between 1 and 5 keV indicates an electron tem-perature of 580 eV.

5 icr1

GMB

GMB

TBS

SHOTNO.

• 06• oa

El.

ZS32 J2 377J

150 J

3 4Energy, E(keV)

Fig. 1V-5.Broad-band x-ray spectra—6 beams of Heliossystem on GMB targets.

53

About 90% of the x-ray emission occurs belowenergies of 0.8 keV. The character of this x-ray emissionis unknown. The total x-ray emission into 4n steradiansis about 2% of the laser energy incident on target. This isconsistent with t^e x-ray emission measured at Geminiwith about 0.25 TW on target. Hence, the bulk of the x-ray emission at energies below 0.8 keV scales with laserenergy.

Experiments using the Gemini system, with <1.0 TWon 180-jim GMB targets, shows that the electron tem-perature Te oc 500 I? " . For a laser power IL of 2.4 TW,we would expect Te to be 680 eV; data from Heliosshowed Te to be 580 eV. This difference, which is notgreat, probably exists because the GMBs used in Heliosand in Gemini experiments were different sizes.

Analysis of Pinhole Camera Images (W. Priedhorsky)

X-ray imaging on the Helios system is performed bytwo pinhole cameras (Fig. 1V-6). Each has a pinhole 13urn in diameter located 4.5 cm from the target and

Fig. IV-6.Pinhole camera for the Helios laser system.

shielded by a beryllium window 7.5 urn thick. Ascurrently operated, with a pinhole-film separation of39.87 cm, the camera magnification is 8.86. At thatmagnification, the x-ray flux from GMBs coated with upto 50 urn of plastic is enough to produce an optical den-sity > 1.5 on Kodak 2490 film. The camera resolution islimited by the geometric size of the pinhole, rather thanby diffraction effects.

The analysis of pinhole images for quantitative infor-mation about target performance (e.g., compression) iscomplicated by the fact that the x-ray emission per unitvolume from a compressed target is a function of bothdensity and temperature. It is not possible to distinguishbetween a density spike and a temperature spike fromimage data alone. Moreover, the targets currently beingshot are, in many cases, optically thick at soft x-rayenergies, making recovery of density information from x-ray images difficult. Because of these problems, we havechosen to compare the pinhole images obtained ex-perimentally with intensity profiles calculated by amodeling program in an attempt to correlate predictedand actual target behavior.

The code uses a one-dimensional calculation; thus, thetarget and implosion were assumed to be sphericallysymmetric. The code uses the laser pulse profile andtarget parameters to calculate radial profiles of density,temperature, emission, etc., as functions of time. Theemission profile is folded with the response of 2490 filmand the pinhole camera filter transmission function; it isthen integrated over time and summed over the line ofsight (LOS) to yield the two-dimensional radial intensityprofile expected for the pinhole images. The calculatedemission comes primarily from the glass shell.

Figure IV-7 shows the pinhole image obtained fromthe parylene-coated GMB that had the greatest apparentcompression of the Helios shots to date. The film wasraster-scanned by a microdensitometer to produce two-dimensional, digitized density data. The center of weightof each image was then calculated, and a mean radial-density profile was obtained by averaging the densitysamples in a series of annuli about the center. Inspectionof the images revealed that they were not entirely sym-metrical, as can be seen in Fig. IV-7. The effect of suchasymetry on the profile was estimated by comparing theprofiles from 20° wedges to the 360° average profile.We concluded that, for the images examined thus far, themean profile accurately represented the local profile: theprofile was not significantly flattened by the averagingprocess. Film density was converted to x-ray intensity

54

Fig. IV-7.Pinhole picture of Shot 88110707,145-\m-radiusGMB with 55-fim CH coating, 2232 J on target.

via the known calibration of 2490 film. Because the filmgamma is a function of photon energy, the characteristiccurve appropriate to 2.62 keV was chosen as typical forthe calculated x-ray spectrum. The characteristic curvefor 2490 film shows little change in shape over the region1.2-3.0 keV.

Figure IV-8 shows the experimental (error bars) andcalculated (solid line) radial x-ray intensity profiles fromthe plastic-coated high-density sequence target of Fig.

IV-7. The laser and target parameters assumed in thecalculations were chosen to match the experimental con-figuration as closely as possible. The target was a 145-um-radius GMB coated with 55 urn of plastic (CH den-sity, 1.0). Energy on target was 2232 J from six of theeight Helios beams. The simulation assumed sphericallysymmetric illumination and implosion. The calculatedprofile was scaled to fit peak experiment) intensitybecause we are interested in comparing the shapes of theprofiles only. Though the shapes of the calculated andobserved profiles are similar, the observed profile isslightly broader than the calculated profile at its half-width and has a longer tail. The maximum compressionachieved in the simulation was 7 times liquid density.

The minimum radii of the fuel-glass and the glass-plastic interfaces obtained in the simulation run are in-dicated by the arrows in Fig. IV-8. Neither interfacecorresponds to any feature in the calculated profile.Thus, there appears to be no way to estimate target per-formance for such targets without resorting to a detailedsimulation. The analysis of more shots using a com-parison with code simulations is in progress.

Return-Current Heating (R. F. Benjamin, G. H. McCall,A. W. Ehler)

Our interest in lateral heat transport at long distances(i.e., more than several focal radii from the irradiatedregion) is motivated by observations during exploding-pusher experiments. X-ray photographs revealed that theglass stalk that supports the gas-filled GMBs is self-luminous in x-rays. This heat flow along the supportfiber can influence implosions by modifying the heatingsymmetry or the electron spectrum or by simply drain-ing energy.

We performed two experiments to determine that thedominant mechanism for this energy transport is returncurrent. The first experiment was to irradiate a targetconsisting of a tortuous path of 10-um-diam fibers (seetarget schematic in Fig. IV-9a). A focused CO2 beam(300 J, 1.5 ns) impinged on a GMB at point A, produc-ing heat that traversed the fibers. Assuming that opticalluminosity is a signature of heating, we identified theheat-flow path using time-integrated optical photographs(Figs. IV-9b, -9c). The most luminous path, ABCJ, isbetween the irradiated region at point A and the nearestelectrical ground, the target holder, as seen in thephotographs. This indicates that the heat flow is driven

55

SHOT * 881107072232 J, 6 BEAMSGMB'300^m diam

54 ̂ m CH coating

Minimum radius,gloss-plotticinttrfacc

40 60 80Source Radius l/im)

100 120

Fig. JV-8.Mean radial intensity profile. Shot 88110707. Solid line indicates calculated profile for similar con-ditions.

Fig. IV-9.(a) Schematic of target and heat flow in target consisting of tortuous path of 10-yim-diam fibers, (b)and (c) Time-integrated photographs of heat flow along fibers.

by an electrical mechanism. We also observed that whenthe fiber section CJ was removed, making ABCDEG theshortest ground path (Fig. JV-9c), the fibers shielded byK were as luminous as those between points A and Gthat were directly exposed to the illuminated region.Thus the electrical mechanism is more important thanplasma radiation or scattered light.

The second experiment discriminated between thedirections of electrical current flow. An aluminum slabmounted on the center conductor of a 50-fi coaxial cablewas irradiated by 150 J of CO2 laser light (1.5-ns pulsewidth). The voltage pulse, shown in Fig. IV-10, was 180kV positive, indicating a net flow of electron returncurrent toward the target.

56

Fig. IV-10.Positive voltage pulse indicating new flow of current toward target.

We conclude that lateral heat transport at long dis-tances along support fibers is driven by return currentthat replenishes emitted hot electrons. The opticalluminosity is ohmic heating. A simple theoretical modelbased on a Maxwellian electron distribution with ameasured temperature agrees with the value of thevoltage pulse and the number of emitted electrons. Moresophisticated current flow treatments must be implemen-ted in computer simulations as a result of thisexperiment.

DIAGNOSTIC DEVELOPMENT (R. P. Godwin)

Introduction

A viable laser fusion experimental program depends inlarge part on being able to anticipate and develop thekinds of instruments required. This is increasingly truewith large, complex laser facilities and ever more com-plex targets.

During this reporting period, the researchers whodeveloped the invaluable fast charged-particle detectorsturned to the problem of improving the time resolution ofneutron detectors. Their objective is to improve ourability to separate the effects of ion temperature, burntime, and run-in time on neutron spectra.

We have continued our vigorous program of x-raydiagnostics development with a balance of near-termgoals (such as the development of a PIN diode x-rayspectrometer) and longer term goals (such as charge-coupled imaging devices, which eventually shouldreplace film detectors). Our x-ray streak camera workwas plagued with triggering problems that were more dif-ficult than we anticipated. Advances made during the

past six months, however, clarified the problems in thisarea. In addition, we have a general project in the area offast high voltage switching and triggering, which shouldbe useful for various diagnostics, including optical probebeams and x-ray backlighting.

Although a superficial glance at 10.6-um laser-plasmainteractions indicates that we should concentrate eitheron infrared or x-ray photons, experience has shown thatvisible-light imaging or spectroscopy provides importantinformation. Seeing is believing. Systematic developmentof optical telephotography with high spatial resolution isevolving from time-integrated photographs through tem-porally resolved photography to the use of preciselytimed optical probe beams.

Ultrafast Neutron Diagnostics (T. H. Tan, A. Williams)

Recently, we developed several fast particle detectorsthat can generate a signal pulse with a full width at halfmaximum (FWHM) of 600 ps or less. Each detector isassembled by coupling an ultrafast, quenchedscintillator12 of some optimum size to a high-gain ITTmicrochannel plate photomultiplier (MCP-PMT). Atypical assembly of the detector is shown in Fig. IV-11.Figure IV-12 shows the response of such a system. Witha fast GHz oscilloscope as recorder, it is possible to ob-tain time resolution of less than 100 ps in time shift andpulse broadening. This greatly extends our capability todiagnose the performance of the high-compression laserfusion targets.

For example, measurement of the implosion time fromthe delay of the neutron arrival provides the first ex-perimental clue of the hydrodynamic process associatedwith laser fusion. With greater yield, the time history of

57

Fost /scintiMator-'

If

MCP

'-Photo cathode

t guide

Fig. IV-ll.Typical assembly of an ultrqfast neutron detector.

neutron production can be streaked by these electrons toyield more refined details of the implosion processes. Letus consider the relationship Y ~ n2 <av> tc V = N2/V<ov> tc. where Y is yield, N is number of DT atoms(known), V is instantaneous volume, tc is burn time, and<av> is cross-section rate, which is a function of fueltemperature Ti alone; tc can be measured from pulsebroadening at short distances and T, at larger distances.The burn volume and the final density can be deduceddirectly.

We are pursuing the experiments to achieve high com-pression at high enough yield to use this fast diagnosticcapability. Such diagnostics of laser fusion performancebecome easier as the yield and confinement time in-crease. At near breakeven, when it may not be possibleto use other diagnostics in the harsh environment nearthe target, these fast neutron diagnostics may providemuch of the important data relevant to laser-induced fu-sion processes.

PIN Diode X-Ray (PDX) Spectrometer (P. Rockett)

A PIN diode x-ray spectrometer has been designed tooperate in the 4.5- to 29.2-keV range. The instrument,complete except for K-edge filters, will be mounted onthe Helios vacuum chamber to monitor the spectralregion between that covered by photomultiplier tube/NEIII scintillators on the low-energy end and CsF detectorson the high-energy end.

Calculations were completed on optimizing PIN diodethickness and sensitivity with K-edge filter transmissionand x-ray spectral form. An analytical model wasgenerated, describing the response of the filter-diode

(a)

(b)Fig. IV-12.

Typical response of the ultrafast detector to a 50-ps e-beam pulse, (a) Intrinsic response of the detec-tor as calibrated by the Cerenkov light, (b) Com-parison of detector response between 3% quenchedNE III scintillator and the standard NE III scin-tillator.

combination (in C/mJ) to any given spectral input.Variations of the diode thickness show clearly its effectas a low-pass filter. A thin diode simply cut off the high-energy response. This enhanced the amount of energy

58

R(E)

.05

04

0 3

02

.01

o

- 1-•

1/

t n i = , 0 . 4 F m

y -MPo« - band11

10 20E(keV)

30 4 0

Fig. IV-13.Spectral sensitivity of filtered diode detector forsilicon depletion layer thickness tSj = 25 urn.

R(E)

0.15

01

.05

J

1-"K-band"

= 10.4/im

= 250 f in

= 0.8/tm

—'Po«»-band"

20ElkeV)

30 40

Fig. IV-14.Same as Fig. IV-13 for t]{ = 250 um.

recorded in the K-edge region relative to that in thehigher energy pass-band. The dramatic effect of diodethickness is illustrated in Figs. IV-13 and -14. Here, onesees the spectral sensitivity of the Filtered diode detectorsfor two different silicon depletion layer thicknesses (tSi).The K-edge-detected energy constitutes the primary por-

tion of the total signal for tSi = 25 um (Fig. IV-13), butonly a minor portion of the total signal for tS| = 250 um(Fig. IV-14). These design principles were used in choos-ing both the diodes and filters for the PDX spectrometer.

X Ray Sensitive Charge-CoupledDevelopment (P. Rockett)

Device (CCD)

A contract has been awarded to Westinghouse Cor-poration for fabrication of an x-ray-sensitive (1-10 keV)CCD to replace film in x-ray pinhole cameras. Theproposed device will have a sensitive area 3 mm by 6mm with 30 um resolution at the detector and 2 X 104

imaging pixels. It will be highly quantum-efficient overthe entire range of I -10 keV. Detailed high-resolutionimages at 10 keV will enhance our insight into the sourceof high-energy electrons in laser-induced plasmas. Thereadout into our PDP 11/70 computer will be im-mediate, allowing shot analysis on a real-time basis.

X Ray Streak Camera Development (P. Rockett, J. S.McGurn)

A streak camera mount for the Helios vacuum cham-ber was designed and drawings were completed. Themount provides tracks along which the camera may betransported. An air lock allows the camera to beremoved from the chamber without breaking vacuum.This will permit film to be reloaded without interferingwith experiments. Repositioning accuracy will be ±10um with a remote station for motor control.

A pinhole translator was constructed for use with thex-ray streak camera in an imaging mode. The position-ing accuracy will be ±2 um in 2-um steps with both x-and y-control. The translator will be attached to thestreak camera mounting rail independent of the camera.The target, pinhole, and streak camera slit will be alignedusing a telescope of high magnification.

A new 36-stage, avalanche transistor trigger unit wasreceived from General Engineering and Applied Re-search (GEAR) and was installed in the GEAR Pico-Xx-ray streak camera. This replaced the original fastoptical-spark-gap trigger, which could not be triggeredwith 10.6-um light. Tests indicated a sweep rate of 100ps/mm.

We used our two-beam Nd:glass laser to determinethe transition trigger delay (by finding the x-ray streak intime) and to calibrate the streak rate and linearity.

59

The laser provided a 300-ps FWHM, 1.06-u.m pulse X-Ray Scalingto illuminate a 200-um nickel microballoon. Two beams Mitchell)were incident upon the target with 5 to 20 J per beam,one beam delayed 1.5 ns with respect to the other. Atrigger beam was picked off and directed into a fastphotodiodc. whose output then triggered the avalanchestack in the Pico-X. Triggering delay, measured by os-cilloscope, was 16 ns with <0.5 ns jitter when the unitwas fired into a 10-kfi, high-voltage probe.

When the camera was placed in the vacuum chamber,however, the temporal location of the streak could not befound. Occasional appearances of partial streak imagesindicated that the avalanche stack did not functionproperly and was randomly triggering on noise duringlaser flashlamp discharge. Direct-current shots provedthat the streak tube was functioning and that sufficient x-ray flux was being generated for film exposure. Inspec-tion of the trigger unit showed that many transistorswere being biased within a few volts of breakdown. Thecamera is being relocated to our single-beam Nd:glasslaser (5 J. 30 ps) where the lower background noise levelis more suitable for further tests of delay, jitter, andsweep rate.

Spectral calibration of the camera will continue at I4.93 keV (Ti Ka). and at or near 17.8 keV (Zr Ka).These will provide six performance data points and willgive us an accurate picture of the Pico-X spectralsensitivity.

Experiments (W. Priedhorsky, K.

An investigation was made to determine whether x-ray calorimetry could be performed using intensitymeasurements of the Si resonance line (Is2p — Is2) todetermine absorbed laser energy on target.

Spectrometer and spectrograph data were obtainedfrom 50 shots with the south beam of the Gemini systemin and out of focus (to vary intensity) on a variety ofSiOj targets. Laser energy on target was varied throughnominal values of 50. 100. 200. and 300 J.

Figure IV-15 shows the integrated signal in the SiXIII resonance line (Is2p - Is2) as a function of laserenergy. The upper line is the least-squares linear fit to thefocused shots on target (slab, lollipop, and GMB),whereas the lower line is a fit of shots -360 urn fromfocus on glass slab targets. There is a notable differencebetween the focused and unfocused shots: none of thepoints from the unfocused shots falls above the meanline of the focused shots.

The focused shots fit the relationship

1.10 ± 0.33l ine

(IV-1)

FOCUSED

UNFOCUSED .

0 = Rl«B, FOCUSED

• ' SLAB.f-36O^m" -- ISO/im CMBa . 90Mm GMBO = LOLLIPOP

IOO

LASER ENERGY (J)

1000

Fig. IV-15.

Integrated signal in the Si XIII resonance line (Is2p — 2s2) as a function of laser energy.

60

where \lim is the intensity of the Si (Is2p - Is2) line andE, is the laser energy. The unfocused shots fit therelationship

I line1.02 i 0.26 (IV-2)

We conclude that there is a small but consistent inten-sity dependence of the resonance line in this range ofshot parameters. The focused and unfocused data areconsistent with a common fit of

I line0.25 (IV-3)

where IL is the laser intensity. For a constant spot size,such a fit would require

l ine1.25 (1V-4)

This is consistent with the data.A second measurement was made with a K-edge filter

spectrometer, which includes the Si line (Is2p — Is2), inchannel 4 covering the range from 1.8 to 2.45 keV. Thisresonance line represents a sizeable fraction of the totalenergy recorded by this channel. The spectrometer datafor focused and unfocused shots are shown in Fig. IV-16. The two lines that are linear fits to the focused andunfocused points are given by

ch 4

and

1.68 ± 0.26

1.29 : 0.15Ch 4

(focused)

(unfocused),

(IV-5)

(1V-6)

where Ich 4 is x-ray intensity in channel 4 and EL is laserenergy. Hence, the difference between the focused andunfocused shots in the channel 4 signal is less thanmeasured with the spectrograph. The spectrometerprovides more detailed data than film recording in aspectrograph because the spectrometer signal is a blendof lines and some continuum. In any event, the laserenergy on target can be estimated to within a factor of 2with 90% confidence from spectrometer measurementsol" the resonance line x-ray intensity.

Narrow-Aperture Methods for Determining the Emis-sion Structure of Imploded Targets (M. M. Mueller)

A general treatment of single-aperture imaging wasgiven for slit, square, and circular apertures as sum-marized in the last progress report." It was shown that,for irradiance data of sufficient quality to allowmeaningful first derivatives, a simple technique decon-volves a wide-aperture image to give the image thatwould have been obtained using an infinitesimal slit.

10-' :

inzUlz*

10" * :

;

:

• , . , . , i

O i

« =

D «

o .

y

SLAB, FOCUSED

SLAB, f -36Opr180pm GM8

90pm GMB

LOLLIPOP

FOCUSED T

UNFOCUSED

n

:TTT

, , 1 1

100

LASER ENERGY (J)IOOO

Fig. IV-16.Measurements made with a K-edge filter spectrometer from 1.80-2.45 keV, including the Si line(Is2p - Is2).

61

Although the problem of instrumental broadening hasthus been overcome in principle, wide-aperture decon-volution is often limited in practice because the data aretoo noisy to yield useful first derivatives. Usually,because of diffraction as well as limited sourceradiances, very small apertures are also prohibited.Thus, there is a need for methods to handle noisy datataken with narrow apertures. Such methods, with par-ticular reference to microballoon targets, are the subjectof this study. The numerical simulations were carried outusing a version of the code developed during theprevious study."

The two main questions to be addressed are:• Can the base-intercept method be used to determine

the size of a small part of a radiating source? Inparticular, can the diameter of an imploded core bedetermined from its image, even though the image isperturbed by the irradiance from the unimplodedmicroballoon wall? An ancillary issue is whetherthe FWHM is a valid measure of the imploded corediameter.

• Can a method be developed to determine the emit-tance of the core relative to that of the implodedwall from images formed by pinholes of narrowaperture? This question bears directly on the issueof inner-wall (pusher-fuel) stability during implo-sion. If significant instability occurs, mixing of thewall material with the nearly nonradiant DT gaswould be expected to produce radiant cores without

much "hollowness" at the center. Conversely, acore that is nearly nonradiant or "hollow" at thecenter would indicate pusher-fuel hydrodynamicfuel stability.

Figure IV-17 is a microdensitometer trace through thecenter of the time-integrated, x-ray pinhole image of animploded microballoon target (Shot VV29 #4). It is agood example of images obtained with the Gemini lasersystem. They exhibit a dim outer ring having a diameterapproximately equal to that of the original microballoon,and a brighter core about one-fourth the originaldiameter. While some asymmetry in the core is usuallyapparent, it will be ignored in this discussion.

Such time-integrated x-ray images may be interpretedas being the superposition of two "snapshots"—one ofthe laser-heated microballoon wall before the implosionhas begun and the other of the compressed core at "turn-around" before disassembly has begun. According to thesimulations, a model based on the superposition of thesesnapshots is adequate to explain the measured image dis-tributions. Thus, we tentatively conclude that theremainder of the implosion history, when the glass shellof reduced density moves at high speed, contributes onlya background irradiance to the time-integrated image.However, a possibility remains that the core diameterdeduced from such photographs is appreciably largerthan the minimum core diameter attained.

The first question has been addressed by numericalsimulations using a variety of source emission models.

0.5 10 1.5 JO 2.5 3 0

Disptocement on Detector Plan* (mm)

Fig. IV-17.Microdensitometer trace through the center of a time-integrated x-ray pinhole image of amicroballoon target imploded by the Gemini C02 laser system.

62

The conclusion is that the base-intercept method ofdetermining the core diameter is not only valid in thenormal range of experimental conditions, but is accurateto one part in a thousand even when the emittance fromthe unimploded shell is 30 times that from the implodedcore. Hence, we infer that the base-intercept method ofdetermining the sizes of "hot-spots" within a target isamazingly robust against smooth but intense perturba-tion from other parts of the source. Of course, themethod fails if the images of two or more hot-spotsoverlap. The answer to the ancillary question is that theFWHM is not generally a valid measure of the corediameter, which is strongly dependent on core structure.For example, for a homogeneous core, use of theFWHM would result in overestimating the volume com-pression by about a factor of 2.

The answer to the second question is more involved,for we have developed two independent methods fordetermining the x-ray emission from the central region ofan imploded microballoon. Both methods assume thatthe absorption of the x-rays by the target is negligible, asis reasonable for targets that have been used until now.

The first method is straightforward and potentially ac-curate, even with data noise, but is tedious. It involvesassuming a source emission model, calculating theradiance distribution, performing a two-dimensional con-volution under the experimental configuration, and thenperforming a second convolution under the conditions ofthe microdensitometer scan. This procedure is theniterated until the calculated irradiance pattern best fitsthe data from the microdensitometer scan. Although thismethod cannot provide a unique solution, our limited ex-perience indicates that uniqueness is not a problem forrealistic source models.

An example of the modeling procedure will be givenwithout demonstrating the iterative process. Theparameter values have been chosen to give approximatecongruence with the data of Fig. IV-17. The emissionmodel used is an outer (unimploded) shell of 180-um o.d.and an inner (imploded) shell of 45-um o.d., with respec-tive wall thicknesses of 1.8 and 5.6 um. The resultingradiance pattern is then convolved with a 12-um-diampin hole with a system magnification of 9, as shown inFig. IV-18 in isometric projection. A microdensitometerscan convolution using a 6.1 -urn-square aperture(referred to the plane of the source) of this irradiance dis-tribution is shown in Fig. IV-19. This second convolu-tion partially fills in the central dip and broadens thebase intercepts, but leaves the FWHM of the core prac-tically unchanged.

Fig. 1V-18.Irradiance distribution in the pinhole image of theemission model.

- : • : ; -ss -sc - « : -20 0 2a «c

Souret Radius (>im)

s ; 100

Fig. IV-19.Simulated microdensitometer scan through thecenter of the irradiance distribution shown in Fig.IV-18.

The main features of the simulated microdensitometerscan in Fig. IV-19 are similar to those of Fig. IV-17. Themeasured base intercept widths in Fig. IV-17 arereduced by the standard method13 to give 185-um o.d.for the unimploded shell and 45-um o.d. for the implodedshell. The microballoon target had a shell diameter of180 urn. It seems remarkable that such a simple sourcemodel can give good congruence with a time-integratedphotograph. Many other simulations with varyingsource models have been computed, and imply that the

data of Fig. IV-17 are consistent only with a model oflow central emission, which is consistent with thehypothesis of pusher-fuel hydrodynamics stability.

When the model-adjusting method is used with arather narrow pinhole, the emission structure of the im-ploded core can be determined accurately. However, itcan be time-consuming, so it would be desirable to havea more expedient method to serve as a first diagnostic.

To this end, we have investigated the effectiveness ofthe ratio of the FWHM of the core image to the corediameter in determining the emission from the centralpart of the core.

We found that the above ratio can indicate the relativeemission of the central region for a broad class ofrealistic models. It appears that the FWHM could bedeveloped into a useful diagnostic of central core emis-sion and that the FWHM method could be used withlarger apertures than those of the model-adjustingmethod.

Reconstruction of Source Emission Structures from SlitImages (M. M. Mueller)

The methods of source reconstruction discussedabove are integral methods that are generally superiorfor noisy data, but they cannot guarantee a unique solu-tion. For cases in which useful first derivatives can beobtained, the derivative methods would be valuable as acheck on the uniqueness of integral methods, or as theprimary method, because they are suitable for automaticcalculation without human intervention. However, thederivative methods require irradiance data from in-finitesimal slits, which leaves only data from the wide-aperture deconvolution method.13 Because wide-aperturedeconvolution is itself a differential method, its use toprovide input to other differential methods compoundsthe requirement of low noise levels to impracticablestringency. Hence, we are forced to consider the use ofdifferential methods on image data taken with practicalrealizable slits.

Both available differential source-reconstructionmethods, a dual-sequential Abel inversion or the Vestand Steel method,14 require radial source symmetry andinfinitesimal-slit images. However, the simpler Vest andSteel method (dl/xdx, where I is the irradiance and x thedisplacement) is the one to be discussed here. Our ques-tion is whether we can use the Vest and Steel methodwith undeconvolved irradiance data taken with practical,narrow-slit apertures.

Although the answer depends somewhat on sourcestructure and the investigation is not completed, the Vestand Steel method does not appear to work well withpractical apertures. The matter has been investigatedanalytically for sources that give a uniform disk ofradiance and numerically for spherical-shell sources.

The conclusion is that appreciable source distortionoccurs for ratios of aperture width to overall targetdiameter greater than ~0.01. For aperture ratios as largeas 0.03, which is about the lower limit of practical,diffraction-free slit imaging for larger targets, sourceemission structure is rather grossly distorted. However,the locations of emission structures are correctly pre-served and the width of spherical shells can be roughlydetermined by the naive application of the base-interceptmethod, even though this method is strictly correct onlywhen applied to source diameters, not to shell widths.Hence, the Vest and Steel method may be of some valuein source reconstruction from practical slit images eventhough the emittance ratios between different parts of thesource may be badly distorted. Also, numerical simula-tion could help in estimating the sense and magnitude ofthe distortion, thus allowing correction.

The issue is complicated and is still under investiga-tion. There is a possibility that a modified and elaborateversion of the Vest and Steel method could be devised togive adequate source reconstruction from real data, butit is clear that the straightforward application of thismethod cannot give accurate emission ratios when prac-tical slit widths are used.

Extension and Generalization of Two-Dimensional Im-age Simulations (M. M. Mueller)

The convolution code13 developed to simulate single-aperture images of randomly asymmetrical sources wasmodified to provide more realistic simulations. Onemodification replaces the circular or square aperturewith something intermediate: the generalized circle(Oberkreis) xp + yp = 1. At p = 2, the circular apertureis recovered, while the square aperture is approached asp —»• oo. Because small "pinhole" apertures are neverreally circular, a more general aperture shape for thesimulation is usually desirable, and in many cases con-siderably improves the simulation accuracy.

Another modification improves the calculation of theradiance from a three-dimensional source, which isbasically an Abel integral for cases of radial symmetry.

64

In the interest of more realistic simulations, another in-dependent two-dimensional convolution was added tosimulate the effect of a microdensitometer scan of thefilm image. In many cases this second (microden-sitometer) convolution is quite important.

Uniformly Redundant Array Imaging of Laser-DrivenCompressions (E. E. Fenimore, P Division; T. M. Can-non, M Division)

We have been investigating the problem of artifactsfrom the image process of coded apertures. These ar-tifacts can obscure portions of the image, but imagesproduced by uniformly redundant arrays (URAs) aresupposed to be free of artifacts. An x-ray camera basedon coded-aperture imaging with URAs was tested usingthe Gemini laser system. Five single-pinhole pictureswere also taken to test the faithfulness of the URA im-age.

Figure IV-20 is a single-pinhoie x-ray picture of alaser-driven compression of a DT-filled GMB, which isscaled to correspond to the URA image of a similartarget (Fig. IV-21). Figures IV-22 and -23 are one-dimension intensity scans through Figs. IV-20 and -21,respectively. These figures show two important features.(1) The URA produced a faithful picture without the ar-tifacts common to other coded aperture imagingsystems. We believe that the small differences betweenthe URA and the single pinhole in these pictures are dueto the pictures being taken from different angles. Thelack of artifacts (and the fact that a digital analysisavoids the nonlinear problems of optical reconstruction)means that the URA has a larger dynamic range andthat intensity ratios from different parts of the image canbe determined accurately. (2) The URA consisted of9521 apertures and, therefore, collected many morephotons than did the camera with the single pinhole.Note that the peak in the URA intensity trace (Fig. IV-23) represents ~7100 times more photons than the peakin the pinhole camera trace (Fig. IV-22). This is near thetheoretical maximum for the URA, which would be9521 times as many photons, or a total of 6.60 X 104.

With its much greater photon-collecting ability, theURA can form images in a weak x-ray flux where thesingle pinhole cannot. Also, the URA can be mountedfarther from the target, thus avoiding debris and makingspace available for other diagnostics nearer the target.

Optical Telephotography (R. Benjamin, J. Riffle)

We developed an optical telephotographic system forexperiments on the Gemini target chamber. During thisreporting period, we used this instrument on lateral heatflow experiments, discussed elsewhere in this report, andon parametric studies involving GMB targets. In the for-mer experiment, the photographs produced by thiscamera were crucial in discovering the importance ofheating due to electrical return current along target sup-port fibers. The latter experiment produced a variety offeatures that we shall attempt to correlate with otherdiagnostic results. Preliminary comparisons indicate thatwe can detect prelase by an excessive amount ofluminous plasma blowoff and we can detect verticalalignment errors by the location of intensely luminousregions at the GMB. Many photographs showeddramatic plasma blowoff from the interface between theGMB and mounting stalk.

The telephotographic system produces high-resolutionimages covering a wide latitude of optical exposure. Thephotographs are time-integraled visible images asdefined by the ir cutoff of the film and the uv cutoff ofthe glass optics. The lens is a modified commerciallyavailable 90-mm-diam telescope, and the film holder is amodified tri plane camera with two internal pellicles. Thecomponents are rigidly mounted on a plate and attacheddirectly to a vacuum flange, as shown in Fig. IV-24.Because the instrument is entirely external to the cham-ber, film-handling, alignment, and focusing are easilyperformed. To cover a wide latitude of exposure, wetypically use two black and white films (ASA 75 and1000) and a color transparency film. The deployed in-strument has a spatial resolution of 80 lpflive pairs)/mmreferenced to the source and a magnification of 12.

Optical Probe (R. Benjamin, J. Riffle)

In preparation for the optical probe experiments at theGemini and Helios lasers, an optical system based on ahigh-power, pulsed argon laser has been designed andfabricated. The source assembly, shown in Fig. IV-25,consists of a commercial argon laser (modified to syn-chronize it with the CO2 laser), an autocollimating align-ment telescope, and beam transport optics. The structureis extremely stable, yet portable. Initial experiments us-ing this system will study plasma expansion, thermal

65

Fig. IV-20.Single-pinhole x-ray picture of a laser-driven com-pression of a DT-filled GMB scaled to correspondto the image shown in Fig. IV-21.

Fig. IV-21.URA x-ray of a target like that shown in Fig. IV-20.

66

•4.9llO*

100 fim

f

Fig. IV-22.Densitometer trace of image in Fig. IV-20.

Fig. IV-23.Densitometer trace of image in Fig. IV-2I.

* ' • •

Fig. IV-24.Telephotography system consisting of a 90-mm-diam telescope and a modified tri-plane cameraused as a film holder. The assembly is rigidlymounted on a plan attached to a vacuum flange.

PERFORATED MOVEABLEALIGNMENT MIRROR

SYNCHRONOUSSHUTTER

TO TARGET

NAUTOCOLLIMATINGALIGNMENT SCOPE

TARGET CHAMBERSUPPORT STRUCTURE

Fig. IV-25.Optical probe source assembly.

67

transport, and laser-driven shock propagation intransparent media.

Incoherent, uv-Triggered Spark Gap (R. Carman, N.Clabo, F. Wittman)

We are pioneering in the development of fast high-voltage switching and triggering for use with variousdiagnostics such as optical probe beams, x-ray and visi-ble streak cameras, and x-ray backlighting.

A pressurized spark gap with low-interelectrodecapacitance was used in this investigation. A double-electrode trigger was used with a uv-grade quartz tube toisolate the trigger electrons from the main gap. Sul-furhexafluoride (SF6) at 5 psi was mixed with argon at400 psi to pressurize the spark gap. A 75-MC2 chargeresistor was used in series with the spark gap and a 20-kV pulse was used for the trigger. Peak outputs andpulse widths for several different values of appliedvoltage and spark-gap pressure are shown below. Thelow-output peak with respect to applied voltage was at-tributed to corona or leakage current from anode toground and to the cathode of the gap causing a largevoltage drop across the 75-Mfi resistor. A new cavity isbeing designed to withstand 2000 psi and have lesscapacitance and a longer leakage path to ground. Thisshould allow us to use higher voltage, decrease pulsewidth, and increase efficiency.

Voltage Applied to Pulse WidthCharge Resistor Spark-Gap Pressure Peak Output Voltage (FWHM)

(kV) (psiL (V) (ns)

466

>V

67-110-125-125

7001080

-1000800

0.480.640.600.64

High-Voltage, Short-Pulse Technology (F. Wittman, R.Carman, N. Clabo)

We continued to develop technology to produce low-jitter, fast-risetime, short electrical pulses of high-voltageoutput. Several areas were explored, including the use ofovervoltaged spark gaps to sharpen the pulse risetime,the use of pulsed charging techniques for krytrons, andthe design of very low capacitance, electrically triggeredspark gaps for producing electrical pulses of <300 ps.

Also, several practical applications of this technologywere identified, one of which will be discussed in thisreport.

Because laser-triggered spark gaps require carefulalignment and a large amount of maintenance, it wouldbe desirable to develop a simple, electrically triggeredkrytron circuit that would produce a subnanosecondoutput pulse of 15-20 kV. This could replace the op-tically triggered spark gap used for pulse generators inPockels cell optical pulse selectors and pulse-clippingsystems. With this in mind, we attempted to understandthe process of pulsed charging of a krytron switch. Wefound that KN22 krytrons (EG&G) were not useful inthis application because the pulsed charging processcaused them to trigger uncontrollably. Further, by ex-ceeding the maximum anode applied voltage rating, itwas necessary to take into account the internal charac-teristic 50-fl impedance of the KN22, as well as the in-creasing inductance of this impedance above 100 A. Inpractice, a larger external load impedance was requiredas the pulse charged voltage applied was increased.

The current capacity of the KN6B is ~3000 A, andthe principal use is in ~50-ns risetime applications,therefore, we decided to try pulsed charging in this tube.We found that while the dc applied voltage limit on theKN6B anode was 8 kV, at least 20 kV could be appliedin a pulsed mode for up to ~ 100 ns. Then, we noted thatif the applied voltage is pulsed to 50 or 60 kV, internalconduction occurs. However, the resulting output pulsefrom the KN6B can be very short. A pulse charging cir-cuit was built using 10 series KN22B tubes to apply an~60-kV pulse to the anode of one KN6B. No keep-alivevoltage is used on the KN6B so that full-charge voltageis developed on the anode before the tube starts to con-duct. Figure IV-26 shows a 519 (Tektronix) oscillogramwhere the pulse is ~900-ps FWHM with an 18.75-kVpeak. Notice that the trace does not return to thebaseline, but rather should be ~5% of the peak due tothe steady-state discharge of the large anode capacitorused in the KN22 string, while the output is taken into apiece of 50-Q coaxial cable. In many applications, thisfeature is of no consequence, but when it is important, acrowbar system could be used to eliminate it. The pulserlifetime appears to be quite high, including the KN6B,even though ~375 A must flow to develop the full outputvoltage. We are packaging one such pulser to use as apulse driver for a CdTe CO2 pulse clipper. We intend topursue the answer to why 60 kV must be applied to the

68

Fig. 1V-26.Oscillogram of an electrically triggered krytron pulser output as recorded on a Tektronix 519 scope.The amplitude corresponds to 18.75-kV peak and the pulse width is 900-ps FWHM with a risetimeof < 50 ps.

KN22 krytron string in order for the KN6B to producea pulse of 18.75-kV peak.

Electrical pulses of 900-ps FWHM are interesting, butother Pockels cell applications require 15- to 20-kV elec-trical pulses with ~IOO-ps risetime and ~200-psFWHM. At this time, we do not believe that a krytronsystem can be built to satisfy this requirement. On theother hand, we wish to retain electrical initiation of thepulse output.

CO2 Laser System for Hydrodynamic Studies (R. L.Carman, F. Wittman, N. Clabo)

The study of the self-consistent evolution of electrondensity profiles near the CO2 critical density is fun-damental to understanding laser-plasma interactions.The ponderomotive force apparently ameliorates manyof the deleterious effects once assumed to be present withlong-wavelength lasers. This conclusion has beendeduced from indirect measurements but should be con-firmed directly. To do so, careful interferometricmeasurements with high spatial and temporal resolutionare required. We determined that the most effective wayto perform such measurements is with a small laserdedicated to this experiment. To this end, operation of asnii»!l CO2 laser system began during the later part of1978, but there are several problems that must beremedied before we can use the laser as a facility for in-terferometric measurements. All six cw CO2 oscillatorsoperated successfully in early summer 1978, delivering~25 W total. Upon arrival of the Ar:Ge detectors, we

demonstrated the piezoelectric translator (PZT) schemeof oscillator output frequency and amplitude stabiliza-tion. In this scheme, the front mirror is translated to ad-just the cavity length to provide automatic feedback. Thespecific CO2 line and band are controlled with thegrating rear reflector. The diameter of the intercavityapertures was determined by manually translating thePZT-controlled mirror to cause more gain in the off-axiscavity modes and then stopping down the intercavityapertures until lasing ceases. Although each cavity hasan independent feedback system and controller, no dif-ficulties were encountered in obtaining the same relativeamplitudes from the six lasers.

The beam-combining optics described in Ref. 13 wereinstalled in the summer of 1978. We demonstrated adja-cent line operation in the 10-um P-branch shortlythereafter. Alignment was carried out in two phases.First, one beam was aligned to pass through the am-plifier chain. Each of the other beams was centered andaligned to the first using the quadrant detector sensorsdeveloped for Helios. We achieved a centering accuracyof ±20 urn for several beams and determined that theangular drift of the beam-combining system was withinthe required tolerance of ±25 urad.

The first GaAs Pockels cell could not withstand thecw beam. A large movable beam stop was then insertedin series with the two gates ahead of it to accomplishreliable 50- to 200-ns adjustable pulse generation.

The low-pressure amplifier chain can be divided intotwo parts. The first consists of five amplifiers, each 50cm long, operating with a 1:1:1 mix of CO2 :N2 :He at anabsolute pressure of 15 torr. By varying the duration of

69

the input laser pulse, we established that pulse durations<75 ns were required for the 10-u.m (P20) gain to be in-dependent of other lines in the 10-|im band.

The second GaAs Pockets cells separated the low-pressure amplifier sections and was synchronized withthe first GaAs Pockels cell. The second portion of thelow-pressure amplifier chain consisted of four amplifiers,each 50 cm long, also operated with a 1:1:1 mix ofCO,:Nj:He. but at an absolute pressure of 25 torr. Wereliably obtained ~1 mJ for 75-ns pulse widths.

Firing only the 9- to 60-kV Marx banks for the low-pressure amplifier and the 10-kV GaAs Pockels cellpulse driver caused no timing problems; however, timingproblems did arise when the 60- and 120-kV Marx banksof the TEA amplifiers were included.

High-Pressure Reinjection Laser (E. S. McLellan)

Introduction. CO2 laser oscillators capable of highlyreproducible spectral content will be required in ourfuture laser fusion systems. A multiatmosphere pream-plifier is also required for efficient generation of high-power subnanosecond pulses. Such signals are neededfor diagnostics development and materials studies, aswell as for power-amplifier drivers. We developed ahighly reliable, high-pressure reinjection oscillator thatsatisfies both requirements. The reinjection concept per-mits a single gain medium to be used as both oscillatorand preamplifier." Here, we discuss numerous improve-ments made during the present reporting period, and inparticular, those leading to reliable operation atpressures up to 7600 torr. Measurements of spectral andspatial gain uniformity for various pressures, and ofoutput-energy reproducibility arc presented. Figures IV-27 and -28 show the Invar-stabilized reinjection os-cillator assembly. This system integrates the oscillatorand four-pass amplifier optical components into a single,stable structure that includes the pressure vessel for thehigh-pressure discharge.

Hardware Design Considerations. In our design, asimple capacitor and spark-gap system replaces the con-ventional Marx system for laser gas pumping. This sim-ple system was found to enhance the laser reliability. TheO.O75-nF main discharge capacitor is charged to 75 kV,regardless of gas pressure. High-pressure operation isfacilitated by adjusting the laser gas mixture to take ad-vantage of the V oc PR2/3 dependence of the self-

Fig. IV-27.High-pressure reinjection oscillator with four-passoptics integrated into the Invar-stabilized struc-ture.

Fig. IV-28.High-pressure laser components.

sustained discharge voltage on total pressure P andproportion R of nonhelium laser gas components. FigureIV-29 plots the maximum arc-free operating pressure asa function of helium fraction in the laser mixture.

Our data show that with 60- by 5-cm electrodesseparated by 1.5 cm, stable operation is obtained forPR2'3 = 1.19.

70

ao

0-

e

*!t-z

oy l O -

1,000

100

90

BO

70

60

50

40

30

20

2.000 1.000 4.000 b.OOO 6,000 7,000 8.000

PRESSURE ( T O R H )

Fig. IV-29.Maximum pressure obtained before arcing between the electrode as a function of helium fraction inthe laser mix and measured small-signal gain at that pressure.

Laser Performance. The measured pressure-dependence of P-20 (10-um) centerline gain is shown inFig. IV-29 and in Table IVII. Gain variation vsoperating wavelength is deseribed in Table IV-II1. andmeasurements of gain variation with radial positionwithin the aperture are presented in Fig. IV-30 and TableIV-IV. Greater gain homogeneity than that shown inFig. IV-30 was obtained by series rather than paralleloperation of the two prcionizer circuits located behindthe anode screen.

Specific stored optical energy was determined bymeasuring the energy extracted from a 0.5Y dischargevolume by gain-switched oscillation. When the entire1.5- by 5-cm aperture was illuminated, 12 J wasproduced with a 36% reflecting output coupler. An un-coated NaCI output coupler (4% single-surface reflec-tance) produced > 9 J.

Tests were conducted to determine whether shock andvibration caused by laser operation resulted in misalign-ment of the oscillator optics. Results of a 4000-shot test

TABLE IV II

DEPENDENCE OF P20 CENTERLINE GAIN FOR10.4 u.m BRANCH OF VARIOUS LASING MIXES

Gain at Various Pressures, torr, and He:N2:CO2 Mixes (%/cm)

Mean GainGain. Standard

Deviation

118537:25:38

4.524.594.554.564.464.540.05

142050:20:30

4.404.414.394.364.414.400.02

180068:14:19

3.553.493.493.493.573.520.04

264580:10:10

3.063.002.952.962.922.980.05

420091:4:5

1.982.012.001.992.0!2.000.02

550095:2:3

1.501.491.511.521.521.510.01

760096:2:2

1.241.171.191.251.191.210.03

71

TABLE IV-III

P AND R LINE DEPENDENCE OF GAIN AT CENTERL1NE OF CAVITY FOR67:I4:19::He:N2:CO2 AT 1800 torr

Gain at I0.4-nm P-Lines (%/cm)

12 16 18 22 24

2.88 3.162.87 3.25

3.24 3.04 3.053.18 3.11 3.06

2.93 3.24 3.28 3.122.872.93

2.973.22 3.33 2.99 2.983.22 3.35 3.14 2.95

Main Gain 2.90 3.22 3.28 3.08 3.00Gain. Standard 0.03 0.03 0.07 0.06 0.05

Deviation

Gain at 10.4-utn R-Lines (%/cm)

Mean GainGain. Standard

Deviation

20

3.183.173.193.233.14

Main Gain 3.18Gain, Standard 0.03

Deviation

14

2.852.842.872.902.932.880.04

Gain

16

3.343.313.263.303.233.290.05

16

2.952.912.942.992.972.950.03

(8

2.932.932.942.972.952.940.02

20

2.752.792.732.752.762.760.02

22

2.612.632.622.602.602.610.01

at 9.4-um R-Lines (%/cm)

14

3.143.383.373.363.393.380.02

14

2.932.922.922.983.022.950.04

15

3.103.103.113.223.073.120.06

18

3.183.123.113.213.153.150.04

20

3.2J3.203.203.193.213.200.01

conducted during a 6-day period are shown in Fig. IV-31. No drop in laser energy was observed after 4000shots within the 1% resolution of the measuring system.These results show that mechanical coupling of the dis-

charge and optical structure in this design did notdegrade system reliability.

During this test, there were 6 shot failures in a 4000-shot sequence. This 99.85% reliability was obtained

72

-I 0 I 2DISTANCE FROM CENTER (cm)

Fig. IV-30.Gain uniformity across the discharge measuredparallel to electrodes at 1800 torr with67:13:20::He:N2CO2 laser mixture on P-20 linesof 10-nm branch.

without warniup or purging of the laser. Start-up timewas determined entirely by the time required to reachoperating pressure. This laser will be used in the GWTFupgrade. In this application, previous experience at highpressure predicts pulse widths of 0.1 to 0.5 ns with peakpowers of 1.0 to 5.0 GW.

MILITARY APPLICATIONS (J. H. McNally)

Introduction

During the past 2 years (January 1977 to December1978). the two-beam glass laser facility was used for avariety of experiments most of which are related tomilitary applications. The laser can deliver a pulse to thetarget chamber from two temporally and spatially coin-cident opposing beams. The pulse width can be 70, 300,or 1000 ps. Normally, a 300-ps pulse is used to deliver15-50 J on target for a power density of ~1014 \V/cm2.

Two series of experiments not directly connected withmilitary applications involve calibrating the GEAR Pico-X streak camera and obtaining spectroscopic data in thevacuum uv of highly ionized metals. The laser is used toproduce a short burst of x rays of known duration forcalibrating the camera. The laser beam impinging on avariety of flat metallic targets produces a high-temperature plasma for spectroscopic measurements.

The following experiments were directly related toweapons applications.

1. Equation-of-state studies. By observing high-pressure laser-generated shock fronts, an improved ver-sion or the equation of state (EOS) can* be deduced.

2. Multiburst simulation and blast wave interactions.Laser fireballs created from single or multiple targetssimulate full-scale nuclear events with respect to densityand temperature.

TABLE IV-IV

DEPENDENCE OF GAIN ON RADIAL POSITION WITHIN APERTURE,

Mean GainGain, Standard

Deviation

+2.5

2.152.262.252.182.192.210.05

+2.0

3.303.253.343.243.353.300.05

Gain

+ 1.5

3.143.103.133.323.083.160.09

67:!4:l9::He:N 2:CO3

at Various Distances

+ 1.0

3.353.383.343.353.323.350.02

+0.5

3.523.483.473.453.443.470.03

0

3.523.553.553.553.603.560.03

from Centerline

-0.5

3.213.223.263.213.253.230.02

-1.0

3.153.042.853.122.913.010.13

(%/cm)

-1.5

2.812.962.812.912.962.890.07

-2.0

3.002.863.212.783.112.99

-2.5

.40

.62

.50

.53

.45

.500.18 0.08

73

5:00 pm 800 am

WED1000SHOTS

THUR.1000

SHOTS

FRI.1000

SHOTS

MON.1000

SHOTS

Fig. IV-31.Chart recorder output showing the relative energy obtained for each of 4000 shots fired during theintegrated structure stability test.

1. Opacity experiments. The laser-produced plasmafireball is observed in the visible region at various anglesthrough a series of filters to gain information onisotropy.

Equation of State Studies (L. Veeser, J. Solem. A.Lieber—P Division)

We did not have a suitable streak camera during mostof the past 6 months, therefore, we spent considerabletime upgrading our optics to collect more light from thetargets and to reduce the background light. When wereceived a GEAR Pico-V camera late in December, wedecided to begin using it immediately and to postponedetailed studies of camera sensitivity and sweep speeds.This was dictated by the laser schedule, by our desire tomove as quickly as possible toward our goal of animpedance-match equation-of-state measurement, andby the lack of a convenient calibration facility.

Our new camera is very reliable. Image magnificationlo about 25x on the streak camera slit still leaves enoughintensity and resolution to see more than one step on atarget foil. Therefore, we can measure how the shockvelocity changes with target thickness. We looked at animpedance-match foil made of a 3 urn step of gold and a4 uni step of aluminum on a 13-um-lhick aluminum sub-strate. Although these data have not been analyzed, thestreaks indicate that we can observe the quantitiesneeded to demonstrate an impedance-match experiment.We also made some risetimc measurements foraluminum foils, finding risetimes of ~50 ps measuredwith ~l5-ps resolution. Such risetimes imply a viscosityat 2 Mbar of ~ 102 P. which is three orders of magnitudeless than Mineev and Savinov15 measured in 1967. Wefeel that our numbers are more reasonable.

Future projects include upgrading the diagnostics andthe laser beam quality for high-quality impedance-matchmeasurements, improving the beam spot uniformity andmeasuring its size for each shot, and studying the timedependence of the pulse.

74

Multiburst Simulation and Blast Wave Studies (S. N.Stone, M. D. Wilke—X Division)

We continued laboratory experiments of laser-drivensimulations of the nuclear multiburst environment. Ouroptical diagnostic techniques were expanded during thisreporting period with the successful introduction of twowavelength experiments. Records were obtained ofShockwaves in 50- and 580-torr air for laser energies of10 and 30 J. delivered with pulse durations of 300 ps, attimes ranging between 0.5 and 50 us. Maximum electrondensities near the center of the shocked region, asdeduced from the observations, are about an order ofmagnitude lower than calculations based on thermalequilibrium conditions. On the other hand, maximumgas densities observed in the shock front differ fromcalculations by only —30%.

We improved our set up geometry for observingshock coupling experiments from air into transparentplexiglass blocks, with the result that shock-couplingstrengths were considerably lower than observed earlier.

For the first time in our MAL experiments, one-dimensional and two dimensional shocks were generatedin air at very low pressure (1 toi r). Previously, the lowestpressures were 10 and 20 torr. The fast-framing emissionpictures of the I D shocks showed luminous plasma torelatively large radii (> 1 cm) and late times (> I us) withsubstantial deviations from uniform spherical shape.

In the 2 D case of two colliding shocks at 1 torr, weobserved an appreciable increase in luminosity from theplasma ring formed by the intersection of the twospherical shock waves: this luminosity was ~ 3 3 %greater than the sum of the light from the separate shockwaves.

Laser Opacity Equipment (N. M. Hoffman, L. W.Miller, J. M. Mack, H. W. Kruse—X Division)

During this period, data from our August 1978 experi-ments were reduced and analyzed. These experimentsused a channel plate intensifier (CPI) camera gated at5.7 us to image the optical self-emission from the back ofaluminum foil targets simultaneously along a 30° and anormal LOS (see Fig. IV 32). We saw. for the first time,a time-dependent anisotropy in this emission, which wediscuss below.

The two-dimensional images on film were scannedwith a densitometer, and film densities were converted to

Chonn«1PloHDttecror

Fig. IV-32.Schematic of experiment showing channel-plate-intensifier camera gated to image optical self-emission from the back of aluminum foil targetssimultaneously along a 30° and a normal LOS.

Iv(30")Whilt 1/2 milWhile 1/4 milRed 1/2 mil

£ io

on0

i _ .. J i i5 1° 15 20

Gate Delay Time, A (ns)

Fig. IV-33.Anisotropy in self-emission along the 30° and nor-mal LOS vs gate delay lime.

relative intensities. The resulting two-dimensional arrayof intensity values was smoothed and then searched forthe peak value of intensity in each LOS. The vertical axisshown in Fig. IV-33 shows the anisotropy R, which isthe ratio of peak intensity in the 30° LOS to the peak in-tensity in the normal LOS. The horizontal axis is At, thetime difference between the laser pulse (300-ps FWHM)and the trailing edge of the gate; i.e.. at At = 0, the gatejuM begins to overlap the laser pulse. Increasing Atmeans the gate is delayed with respect to the laser pulse.The ratio R was corrected for progressive differential

75

obscuration of the glass plate protecting the optics; thiseffect is always <12%. Each point represents a singlelaser shot: pulse energies range between 9 and 40 J.

Most of the data in Fig. IV-33 are for an unfilteredband of white light, defined essentially by the S-20response of the CPI photocathode. However, 3 pointswere taken using a red filter which passed only X > 6000A. These points are shown as open circles.

These data reveal a trend from "limb-brightening" (R> I) to "limb darkening" (R < 1) as the gate is delayedwith respect to the laser pulse. At late enough times (At

> 10 ns). R returns roughly to unity. The l/4-mil foilsbehave much like the 1/2-mil foils, although we have lit-tle data for them. Interestingly, red light seems to be"limb brightened" at the same time that white light is"limb darkened" at At = 2 ns.

Figure IV-34 shows relative intensity in the normalLOS vs gale delay time for laser pulses with energies be-tween 26 and 31 J striking 1/2-mil foils. This profile iswhat one expects from smearing the true intensityhistory of the target backside with the ~6-ns FWHMgale: from earlier streak camera observations we knowthat the intensity history consists of a bright flash of <1-ns duration, followed by a fainter tail of >IO-ns dura-tion.

Comparing Fig. IV 33 with Fig. IV-34 shows that thelimb darkening (R =: 0.75 at 2 ns < At < 10 ns) isassociated with the smeared-out bright flash. Also, thereturn of R to roughly unity is associated with the fainttail at At > 10 ns. This is expected as the plasma

26J<E<3IJW»,lt, I / ? mil

O0L

0. i -_ A L—.

10 15 20Gite Delay Time, AMns)

Fig. IV-34.

Relative intensity in the normal LOS vs gate delay

time.

becomes nonplanar. which certainly occurs by 10 ns af-ter the laser pulse.

The limb-brightening for At < 2 ns has been amystery, though we now believe it may be an instrumen-tation elTect. We can rule out a straightforward inter-pretation using any isotropic source function in a planarplasma: it is impossible to get limb-brightening strongerthan 1/cos 9 in this way. At 6 = 30°. 1/cos 6 = 1.15,whereas we have R between 1.5 and 2.0.

Thus, we are led to consider a dynamic instrumenta-tion effect as a source of spurious limb-brightening. Sup-pose that the CPI tube does not gate simultaneouslyacross iis surface. Specifically, suppose that the portionof the lube viewing the 30° LOS gates "off" slightly laterthan the portion of the tube in the normal LOS. Then forAt ~ 0. the 30° view will integrate more of the risingedge of the bright flash than will the normal view, givinga spurious limb-brightening. If this does happen, and itprobably does (see below), it nevertheless does not in-validate our observations of limb-darkening.

If the back-side emission were not intrinsically limb-darkened, then the spurious limb-brightening caused bynonsimultancous gating would persist at least until theleading edge of the gate is moved later than the laserpulse: i.e.. until At ~ 6 ns. After At ~ 6 ns, it might bepossible for nonsimultancous gating to give spuriouslimb darkening: then the emission intensity is decreasingwith time over most of the gate duration. This argumentdoes not affect the limb-darkening for 2 ns < At < 6 ns,except possibly to make the observed limb-darkeningsomewhat weaker than the intrinsic limb-darkening.

Recent experiments show that, in fact, the CPI tubedoes not gate simultaneously across its surface. It ap-pears that the outer edge of the circular tube gates "on"about 1 ns before the center of the tube; the edge gates"oil" slightly earlier, too. so that the total gate durationis about the same across the lube. The gating pattern iscircularly symmetric about the center of the tube; thus,to remove this effect one could simply ensure that thetwo target images are equidistant from the tube center.We are reviewing the data to determine how far the twoimages were from being centered.

Vacuum uv Spectroscopy of Highly-Ionized Metals (J.Reader and G. Luther, NBS, Washington, DC)

In October 1978, researchers from the NationalBureau of Standards used LASL's high-power Nd.-glass

laser to observe spectra of highly ionized atoms belong-ing to isoelectronic sequences important to the diagnosisof fusion plasmas generated in tokamaks. These obser-vations were made by photographing the spectra oflaser-produced plasmas of various metals with a 2-mgrazing-incidence spectrograph. Spectra of the followingmetals were obtained: Fe. Sr. Y, Zr. Nb, Mo. Ru, Rh,Pd. Ag. Cd. Sn. Ba. La. Nd. Sm. Gd. Dy. Er, Yb, Ta,and W. All spectra were obtained with a laser pulsewidth of 300 ps at a total energy of ~25 J. Six laserpulses were generally required to produce a usable spec-trum. The preliminary results are catagorized by isoelec-tronic sequence.

Zinc sequence. Figure IV-35 summarizes theNBS/LASL observations for the 4s2 'So - 4s4p "P,resonance line. Zc is the net charge of the atomic coreseen by the valence electron, or stage of ionization. Thepoint for Xe is from tokamak observations by Hinnov."The points for Rb6+ to Sb21+ are from recent work at

1600

1200

5 800<£uiZill

400

I—1

LASER-PRODUCED.PLASMAINBS/LASL

fzn,

1600

1200

—i 1 1 14

/

$ 1LASER-PRODUCED/

PLASMA •'-' 1(NBS/LASL) /

W

/

'Yb

Ta

> 800s

400

f<Gd

N d ,

Ba.Xe

1/2-3/2,

M o .Flu.

/ ^ 1/2-1/2

'XeTOKAMAK

^ (THEORY)

1

Ag

Kr

Cu\

15 25Z •

35 45

15 25 35 45

Fig. IV-36.Energy transitions for the 4s-4p resonance linesfor the copper sequence.

NBS with low-inductance sparks17 and laser-producedplasmas.18

Copper sequence. Figure IV-36 summarizes theNBS/LASL observations for the 4s - 4p resonance lines.The Xe points are again tokamak observations fromRef. 16. Previous NBS work17*18 on these ions coveredRb7+ to Sb22+. The present observations for W4i+ repre-sent the highest charge states ever observed with diffrac-tion gratings.

Comparisons with theory for the copper and zincsequences from Br and Kr to W are given in Figs. IV-37ant! -38. The points for Xe (with error bars) represent in-terpolated corrections to the theory, which permitderivations of accurate semiempirical wavelength values.

Fig. IV-35.Energy transitions for the 4s2 'So - 4s4p%resonance line for the zinc sequence. Zc is netcharge of the atomic core seen by the valence elec-tron, or stage of ionization.

Sodium sequence. In a recent NRL/NBS/LASLcollaboration involving spectra for molybdenum ob-tained with a Nchglass laser at the Naval ResearchLaboratory in Washington, DC," two close lines wereobserved in the neighborhood of a strong tokamak line1*at 129 A. This was attributed to the 3s 2S,/2 - 3p 2Pr3/2

77

+2.0

0.0

-2.0

W

Fig. IV-37.Comparison of theory with observed wavelength for the zinc sequence.

W

Fig. IV-38.

Comparison of theory with observed wavelength for the copper sequence.

transition of Mo31 ' Figure IV-39 gives the newNBS/LASL results for this sequence (Sr to Pd). Theyshow that the lower energy line of the two close lines inthe laser-produced plasma is the 3s 2S, /2 — 3p 2PV 2 tran-sition of Mo31 \ and the higher energy line, the 3 p 2 P , / 2 -3d 2D, /2 transition of this ion. The Kr25+ point is atokamak observation of Hinnov.1*

Magnesium sequence. In Fig. IV-40, the points be-tween Sr and Rh represent the NBS/LASL results. They

confirm the identifications in Rcf. 19 for Mo30f for thethree transitions shown. The Kr point is due to Hinnov.16

Figure IV-41 gives a comparison between theory and ex-periment for the Mg isoelectric sequence. The point forKr represents an interpolated correction to the theory,from which an accurate semiempirical wavelength maybe derived.

78

1000

800

600

> 400ODCUl

5 200

3p 2 P 3 / 2 - 3d 2 D 5 / 2

Cu 3s 2 S 1 / 2 - 3p 2 P 3 / 2

10 15 20 25 30 35

Fi&. IV-39.Energy transitions for the sodium isoelectronicsequence showing two transitions of ' Molu.

Fig. 1V-40.Energy transitions for the magnesium isoelectronicsequence showing three transitions ofMo30+.

1000

800

o 600

O 400Hi

HI

200

3s 3p 3 P 2 — 3s3d 3 D

3s3p 1 P 1 - 3 s 3d Sr

Kr-TOKAMAK

Se

Cu

Fe 3s

_ . . . . ! . . I .. I . . . . I10 15 20 25 30 35

0.30

0.20

««? 0.10

3 o.oo-0 .10

-0 .20

-0 .30

Fig. IV-41.Comparison of theory with experiment for transition wavelengths in the magnesium isoelectronicsequence.

79

REFERENCES

I. D. B. vanHulsteyn, P. Lee, K. B. Mitchell, and P. D.Roekett. "X-Ray Continuum Spectra," in "LaserFusion Program. July I—December 31,1977," LosAlamos Scientific Laboratory report LA-7328-PR(December 1978).

2. D. Giovanielli and C. W. Cranfill, "Simple Modelfor Exploding Pusher Targets," Los Alamos Scien-tific Laboratory report LA-7218-MS (May 1978).

3. V. Cottles. D. Giovanielli, L. White, and A.Williams. "'Absorption Measurements-10.6 urn CO2

Laser Light." in "Laser Fusion Program, July1 — December 31. 1977." Los Alamos ScientificLaboratory report LA-7328-PR (December 1978).

4. F. Young. "Thomson Parabola Measurement ofFast-Ion Spectra and Population of Ion-ChargeStates." in "Laser Fusion Program, July 1—Decem-ber 31, 1977." Los Alamos Scientific Laboratoryreport LA-7328-PR (December 1978).

5. D. Giovanielli, J. Kephart, and A. H. Williams,"Spectra and Angular Distributions of ElectronsEmitted from Laser-Produced Plasmas" J. Appl.Phys. 47, 2907 (1976).

6. R. F. Benjamin. G. H. McCall, and A. WayneEhler. "Measurement of Return Current in a Laser-Produced Plasma." Phys. Rev. Lett. 42, 890 (1979).

7. T. H. Tai. "Inertial Fusion Program, January 1—June 30, 1978," Los Alamos Scientific Laboratoryreport LA-7587-PR (May 1980).

8. J. G. Crow. J. Plasma Phys. 14, Part 1, 65 (1975).

9. D. Forslund. J. M. Kindel. and K. Lee, Phys. Rev.Lett. 39, 284 (1977).

10. V. Cottles, "Inertial Fusion Program, January1—June 30, 1978," Los Alamos ScientificLaboratory report LA-7587-PR (May 1980).

11. T. H. Tan, 'inertial Fusion Program, January 1—June 30. 1978," Los Alamos Scientific Laboratoryreport LA-7587-PR (May 1980).

12. P. B. Lyons, S. E. Caldwell, L. P. Hocker, D. G.Crandall, P. A. Zaganiro, J. Cheng, G. Tirsell, andC. R. Hurlbert, IEEE Trans. Nucl. Sci. NS-24(1977) p. 177.

13. F. Skoberne, Comp., 'inertial Fusion Program,January 1—June 30, 1978," Los Alamos ScientificLaboratory report LA-7587-PR (May 1980).

14. C. M. Vest and D. G. Steel, "Reconstruction ofSpherically Symmetric Objects from Slit-ImagedEmission: Application to Spatially Resolved Spec-troscopy," Opt. Lett. 3, 54-6, (1978).

15. V. N. Mineev and E. V. Savinov, Sov. Phys. JETP52. 629 (1967).

16. E. Hinnov, "Highly Ionized Atoms in TokamakDischarges," Phys. Rev. 14A, 1533 (1976).

17. J. Reader and N. Acquista, "4s-4p Resonance Tran-sitions in Highly Ionized Cu- and Zn-like Ions,"Phys. Rev. Lett. 39, 184 (1977).

18. J. Reader, G. Luther, and N. Acquista, "Spectrumand Energy Levels of Thirteen-Times IonizedMolybdenum (Mo XIV)," J. Opt. Soc. Am. (Decem-ber 1978).

19. P. G. Burkhalter, J. Reader, and R. D. Cowan"Spectra of Mo XXX, XXXI and XXXII from aLaser-Produced Plasma," J. Opt. Soc. Am. 67, 1521(1977).

80

V. LASER FUSION THEORY AND TARGET DESIGN(D. B. Henderson)

Our theoretical support activities are closely coupled to our experimental efforts, withthe intent of gaining a fundamental understanding of laser-target interactions, par-ticularly of the relevant plasma physics and hydrodynamics. The close coupling oftheory and experiments has made it possible to eliminate theories that are not supportedby experiment. In general, basic studies have shown that the design difficulties as-sociated with long wavelength are less severe than believed earlier, and that breakeventarget designs are attainable even in the presence of a hot-electron spectrum. Theseresults have increased our confidence that scientific breakeven can be demonstratedwith our efficient, inexpensive CO2 lasers.

THEORETICAL SUPPORT

Introduction (D. W. Forslund)

The supporting physics section continued its efforts todevelop models for laser light absorption and transportof the resulting deposited energy. The effects ofdeliberate or inadvertent targeting error on the absorp-tion of light by spherical targets in the Helios systemwere calculated. The calculations indicate that absorp-tion may be enhanced by off-center focusing. Thevalidity of the empirical scaling model of hot-electrontemperature was further verified, as well as the detailedacceleration mechanism of the hot electrons. A more ac-curate Monte Carlo treatment of hot-electron transportwas developed, which includes important cross-couplingterms between the hot and cold electrons. Finally, a moreprecise calculation of the loss of fast ions from the tailsof the fuel-ion distribution in low pR pellets is presented.

Light Absorption by a Spherical Target as a Function ofLaser Focal Shift (H. Brysk, A. J. Scannapieco)

The illumination of a sphere by a system of finite laserbeams and the associated energy deposition were discus-sed previously.1 However, the complete computation ofthe amount of light absorbed by a spherical target re-quires further consideration of the time variation in thelaser pulse and of the motion of the critical surface underhydrodynamic expansion. On the basis of a two-dimensional study, the expansion velocity can be approx-imated by

V r 200 n I 1/3

where I is the largest absorbed irradiance up to that timein the problem, and where the numerical factor has beenevaluated with the energy expressed in kilojoules, time inpicoseconds, and distance in microns (thus, the units of vare 108 cm/s and those of I are I0" W/cm2).2

Laser light absorption by spherical targets in Helioswas computed with the LISP code (Laser Illumination ofSpherical Pellets). The initial target radii were 100 and200 urn (i.e., the bracketing current design values). Thetargets were assumed to expand during irradiation (totalbeam energy, 4 kJ) to ~!000 urn. The focal spots wereaimed in unison at the center of the target or at variouslateral distances from the center, both along and perpen-dicular to the polarization direction, to illustrate the im-pact of deliberate or inadvertent targeting. The focal spotwas assumed Gaussian with a width of 68 urn, and thetemporal pulse as triangular, with a risetime of 200 ps.and a fall time of 1 ns.

The fractional absorption (integrated over the pulselength) as a function of the displacement of the focal spotfrom the center of the target is plotted in Fig. V-l for thetwo initial target sizes and for P, the displacement alongthe polarization direction, and S, perpendicular to P.With the focal spot at the center of the target, there is, ofcourse, no polarization dependence; also, there is no sen-sitivity to the initial target radius because the targetquickly grows to a size much larger than the focal spot(and to about the same size for both initial radii). Notethat the absorption changes little for focal shifts (in anydirection) of the order of the focal spot, so that thequality of targeting cannot be safely inferred from targetperformance. When the beam is shifted perpendicularlyto the polarization direction, there is a monotonicdecrease in absorption with displacement. On the otherhand, for shifts along the polarization direction there is

81

100 200DISPLACEMENT^)

Fig, V-l.Light absorption by a spherical target vs beamshift P along polarization direction, and S perpen-dicular to polarization.

300

an absorption maximum (as much as 25% higher) for anoff-center position. More significant may be the fact thatthis maximum occurs with a displacement —150 urnaway from the specular point; there may. then, also bean enhancement of laser output because of suppressionof parasitics.

Target performance depends not only on total energyabsorption but also on its time dependence. Figures V-2and -3 display the temporal profile of the absorbedpower for the two target sizes with various focal spot dis-placements. With the focal spot at the center of thetarget, the absorbed power profile is quite similar to thetriangular incident power profile. For displacementssmaller than the initial target radius, the pulse shape isonly moderately altered while the amplitude changes, sothat Fig. V-l approximately presents the relative perfor-mance. As the displacement increases, the pulse rise isslowed because only the wing of the beam pattern is inci-dent on the target until the latter has expanded enough tointercept a major part of the beam; hence, the relativeperformance for displacements greater than the initialtarget radius is expected to be worse than Fig. V-lsuggests. At late times, the absorption becomes insen-sitive to the shift because the target has then expanded somuch that its size is much larger than the displacementand the beam size.

i i

Ro-200^m

I |

1

-

200 400 600 800TIME (ps)

1000 1200

200 600 800TIME (ps)

Fig. V-2.Absorbed power profile for beam shifts (pm)perpendicular to polarization; (a) for Ro = 200\xm; (b)for Ro = 100 \im.

82

2 -

1/5CO

1inn /s.

Jim

~ m

'11//

5̂0

1

i

1

R

(

1

1

0- 200 pm

i

1

-

-

200 400 600 800

TIME (ps)1000 1200

200 400 600 BOOTIME(ps)

Fig. V-3.Absorbed power profile for beam shifts (\im) alongpolarization (a) for /?„ = 200 \un; (b)for ra = 100um.

Extension of Suprathermal Electron Scaling Law (D. W.Forslund)

Of great importance in the performance of laser-driventargets is the spectrum of the suprathermal electronsgenerated in the absorption process. We previouslyreported3 a scaling law that fits the suprathermal electronspectrum produced in two-dimensional WAVE simula-tions and agrees rather well with the experimental data.We verified that this scaling law is valid over an intensityrange an order of magnitude wider than reportedpreviously, now ranging from 10" to 2 X 10" W •um2/cm2. This result was obtained on the CRAY-1 com-puter, which accepts grid sizes and performs at speedsunattainable on the CDC 7600. At the highest intensityrun, IX2 = 1.6 X 1017 W • umVcm2, still below the max-

imum attainable intensities in Helios, the measured hottemperature TH (for Tc = 0.625 keV) was 35 keV,whereas the formula3

= 30(n2)1/3T

yields 30 keV; here IX2 is in units of 10" W • umVcm2.and TH and Tc are in keV. At the lowest intensity of 1.2X 1O15 W • umVcm2, TH was 9.3 keV, in agreement withthe above formula. The new data points are shown inFig. V-4. This result increases our confidence in using theformula to predict the hot-electron temperatures.

Other noteworthy features of (he highest intensity runillustrate that the absorption process is extremely non-linear. First, the hol-electron temperature is less than theoscillating energy of 41 keV of an electron in the vacuumelectromagnetic field; it would be interesting to seewhether this behavior continues at higher intensities. Se-cond, the hot-electron density is three times higher thanthe critical density, indicating that the hot-electron pres-sure gradient at that density is insufficient to overcomethe balance between the cold-electron pressure gradientand the laser pressure gradient. This situation exists eventhough the hot pressure exceeds the cold pressure bymore than a factor of 3. The absorption remains above25% over the entire intensity range.

Coherent Acceleration Of Hot Electrons (P. Bezzerides,D. DuBois, D. Forslund)

In an effort to understand the physical mechanismsresponsible for the high-energy tails observed in simula-tions of resonance absorption, we developed a simplemodel of the acceleration process. The ultimate goal ofthis work is to identify the controlling factors in thedevelopment of the hot-electron spectrum and thereby todetermine the extent to which the spectrum can bemodified experimentally.

The model was studied numerically and analytically.Basically, it assumes that the heating is due to coherentacceleration when an electron passes through a large-amplitude, localized, traveling wave with spatially vary-ing phase velocity. Our numerical work, particularly thephase-space plots, illustrates that the particle in the fieldoscillates at large amplitudes about its guiding center,and that this coherent oscillatory motion is interruptedby resonant acceleration in the neighborhood of the

83

J--Tc-IOke«

^

0.33

& 1.06-pm X-RAY DATA

o IO. t -Mm X-RAY DATA

• lO.C-pm IDN DATA

A * WAVE SIMULATION

J I

K>

/i*. V-4.Hot-electron temperature vs laser intensity. The solid data points are WA VE simulation points andthe rest are experimental data. The new points are at 1.2 X 10" and 1.6 X 10". The dashed curvesare the empirical scaling laws for three different background electron temperatures, Tc = 0.625,2.5,and 10.0 keV. The solid squares correspond to Te - 625 eV, the solid circles to Tt = 2.5 keV, and thesolid triangles to Tc = 10 keV.

resonance where the velocity is equal to the phasevelocity, leading to the final velocity of the particle. Thefinal time-averaged distribution function in v2 is obtainedfrom the v, — v2 transform plots, where v, is the velocityat which the electron was injected into the field at earlytime to exit the field with value v2 at a given final time.

Most attempts to understand the heating analyticallyhave invoked the transit-time heating effect, which is theonly way to explain how particles that are already fastcan acquire still higher energies. We derived a generalformula relating v, to v2 by using a generalization ofguiding-center theory, which allows for large-amplitudeadiabatic motion, resonant interaction, and transit-timeeffects. We found that

J(v2,t)T

<h uH(x(t),t)!

(v-i)

relating the velocities of the velocity transform plot for agiven time t at which v2 is attained, where T is the injec-tion time. In Eq. (V-l), UH[X(T),T] is the oscillatoryvelocity of the particle at its position X(T) at time T. Equa-tion (V-l) reduces to the well-known limits of guiding-center motion, resonant heating, or transit-time heating,but more important, it shows how the acceleration scales

with the inhomogeneity of the field. Evaluating the in-tegral in Eq. (V-l) is, in general, as difficult as obtainingthe detailed trajectories. For the envelope of the vs — v2

transform corresponding to maximum acceleration for agiven v2, we find

- (V 2 -V (V-2)

with vph(xr) denoting the phase velocity at the resonantpoint given by vDh(xr) = v2 - Av, where Av =4/«^vOi(xr)Vph(xj). The most important limit occurswhen, in the frame moving with the phase velocity, theacceleration due to the field is greater than that due to thespatial phase inhomogeneity. Note that Av is equivalentto the scaling result for TH =s eE ,̂ where t~x is to be un-derstood as k, the spatial phase. The predictions of Eq.(V-2) compare favorably with the numerically deter-mined v, - v2 envelope for the velocity transform plot;we believe, therefore, that the coherent-accelerationmodel is well understood. Progress is being made in un-covering the role of coherent acceleration in oursimulations.

84

Hot-Electron Transport (R. Mason)

Elsewhere we described a double-diffusion scheme4 forhot-electron two-dimensional transport in which thesuprathermal electrons were treated as a fluid. This treat-ment, however, limited the precision of preheat predic-tions. The need for accurate preheat calculations hasrenewed our interest in Monte Carlo modeling.

The one-dimensional Monte Carlo transport code ofearlier reports (see LA-6245-PR) has been improvedsignificantly. As before, the hot electrons are treated asarea-weighted PIC particles, whereas the cold electronsare treated as a fluid. The hot electrons are scattered, inaccordance with Jackson's Gaussian distributions forboth angular deflections in Coulomb interactions anddecelerations from Coulomb drag. Their energy loss isdeposited in the cold electrons, which are also ac-celerated in the local area-weighted electric field.

As an important new feature, we have calculated Efrom Poisson's equation in the "current" form, i.e..

in which j s is the suprathermal number flux, and ncuc isthe flux of cold electrons. We accumulate j s as a particleproperty at the end of each cycle; ncuc is obtained fromthe momentum equation for the cold electrons

there are very few cold electrons present (''—>• oo, so thatuc —*• 0, or for nc —*• 0), Eq. (V-4) becomes

- E 1 " 3 * ' > 1 i t - v i t r . . , , . ' < * • • > .c J C "'(V-4)

In general, with strong collisions (when vAt » 1), Eq. (V-4) reduces to Ohm's law, so that, in steady state,

with dPjdx small, Eq. (V-3) yields

if

which could lead to enormous fields on a typical fluiddynamic calculation mesh in which the cells are manyDebye lengths wide. To avoid such large E-fields, welocally decrease the electron charge e so as to stretch thelocal Debye length up to the calculational mesh dimen-sions. This ensures that the E-field is of the order

determining E(m+1).Alternatively, for extremely high collision frequencies

so that the cold-electron velocities are nearly zero, or if

which we would derive near steady state from a momenttreatment of the suprathermal electrons.

This new Debye stretching technique is extremely im-portant because, for the first time, we can track MonteCarlo or multigroup electron transport in the low-densitycorona around a pellet, where the density of suprather-mal electrons exceeds the density of cold electrons. Inpractice, the E-fields calculated by this technique es-tablish quasi-neutrality over one or two calculationalcells. The stretched Debye length is somewhat akin to thestretched mean free path employed with von-Neuman ar-tificial viscosity.

Figure V-5 shows typical ouput from a run with theMonte Carlo code. A fixed gold profile has been il-luminated with 10'* W/cm2 of 1.06-um light with 20%absorption for ?. ps. Frame I shows the density profilesfor hot and cold electrons and the critical density Zn,.Suprathermal electrons are created at the critical density(Zn, = 1021 cm"3) and are emitted toward the laserbeyond the corona. Frame (d) shows the phase space forsuprathermal electrons. They have been reflected by theself-consistent E-field and flow with u < 0 out the leftside of the problem area. Electrons with u > 0 in x < xcri,(xcr!t ss 31-um point) have been either scattered by theions or retarded by the E from Ohm's law. Frame (b)shows the reflecting E > 0 field In x > xctlt and theretarding ohmic E-field in x < xcrlt. Frame (c) shows thehot and cold temperatures set up in the plasma at 2 ps.The hot electrons are at roughly 20 keV. They aregenerated in accordance with TH = 31.6 (IX2)113 VJ3, andthe increasing background temperature of cold electronscan, therefore, affect TH. To model this dependence witha fixed number of simulation particles emitted during

85

(cm"3) |0

•h .c(keV) ,„> U

10

. . 1 1

WITHOUT h

- M INHIBITION

To

1 1 i

V

25

U(crtvWO

150.

0.

-150.

-300.

50 0

( d ) 1

— * » *• •

I

':*"»•••]

i

*• • •_

i

25 50

Ffc. K-5.Typical results from the Monte Carlo code with Debye stretching after 2 ps. See text for descriptionof individual plots.

each cycle, we found it necessary to weight the PIC par-ticles. Finally, two temperature profiles for cold electronsare shown. The higher Tc curve is for thermal diffusionalone, whereas the lower curve plots diffusion plus ther-mal inhibition stemming from enhanced thermoelectriceffects and convective heat-flow effects, due to the strongcold return currents in response to the incidentsuprathermal current.

To treat the above flow effect of cold electrons withprecision, the model now performs complete donor-cell5

fluid dynamics calculations for the cold electrons. Thus,we avoid flux-limiters in the transport of cold electrons,rendering accurate drift velocities for the calculations ofconvective thermal transport, PDV work, and jouleheating.

High-Energy Ion Losses In Laser Fusion Pellets (A. G.Petschek,* D. B. Henderson)

The mean free path for the 90° deflection of a test ionin a hot plasma of identical ions is

X = V\ v"(er r ? V £n,',)"1 , (V-5)

where M, is the mass, v the velocity, n the backgrounddensity, Ze the ion charge, and (n\ the Coulomblogarithm. Thus, high-energy ions have long mean freepaths and will be preferentially lost from a finite plasma.

•Consultant: Permanent address. New Mexico Institute of Mining andTechnology. Socorro. NM 87801.

86

On the other hand, in the quadratures leading to thereaction rate

2 J" v?o(v)dvv v !

(V-6)

high relative velocities v are emphasized because of therapid increase of the cross section o. Other symbols inEq. (V-6) are g, and g2, the distribution functions of thereactants, and their velocities v, and v2. The emphasis onhigh velocities is most easily seen if g, and g2 are Max-wellian so that they can be replaced by a Maxwellian g ofthe relative velocity and by another of the center-of-massvelocity. Integrating, we obtain

10

1/2.'g(v)v?o(v)dv (V-7)

If a involves a simple Gamow penetration formula ap-propriate to DT, the integrand has a maximum at arelative energy of E/kT = 6.8(keV/kT)"3. Hence, loss ofthe high-energy tail of the distribution function might beexpected to have a substantial effect on <ov>. In anearlier publication,6 this effect was estimated by trun-cating the distribution g at the velocity above which theions could transit the pellet radius with less than one 90°scattering. Because the values of <ov> presented in Ref.6 are erroneous, we give, in Fig. V-6, <ov> as a functionof the temperature and cutoff, which is for distributionsthat are Maxwellian at some temperature up to a cutoffand zero beyond. Reference 6 used Eq. (V-7) rather thanEq. (V-6) and showed a much larger effect than Fig. V-6.

To better calculate the effect of ion tail loss, we used acomputer to solve the Fokker-Planck equations, as out-lined by Rosenbluth et al.,7 and have added an ion lossterm using a diffusion approximation, assuming

fr) r(x,v) (V-8)

where n is the density of fast ions as position^ and speedv. This leads to a loss per unit time independent of posi-tion

1 IEr dt

(V-9)

Our distribution function is, therefore, also independentof position.

2 5 10 20 50 I00 200 500CUT-OFF ENERGY,Ec(keV)

Fig. V-6.Fusion reactivity <ov> integrated to cutoff energyEe/or various temperatures T.

A problem arises as to what to assume for the fate ofthe lost particles. Should they be returned to the pellet orlost permanently, and if they are returned, with what dis-tribution of speeds? Clearly, the answer depends on com-plicated dynamics in the pusher surrounding the ther-monuclear fuel. We chose to return the lost ions to thefuel with the speed distribution of the fuel. That is, afterthe loss calculation, we reseated the distribution to reachthe original particle number. This assumption un-doubtedly affects the reaction rate and the mean ionkinetic energy, but probably has much less effect on theshape of the distribution function and on the deviation of<ov> from its Maxwellian value. For simplicity we alsoused a single ion of mass 5/2 instead of carrying D and Tions separately. We calculated ftiA following deWitt,8 us-ing the mean ion energy, and thus, ignoring some of theenergy dependence of A.n A. We have also neglected colli-sions with electrons because ion-ion collisions are muchmore important than ion-electron collisions for both themomentum and energy change of the ions.

87

ION ENERGY ( k e V )

. . 4 0 1 2 3 4 5 6 7 B 9 10 M 12

01

DENSITY-T IME PRODUCT, n t ( s cm )

Fig. V-7.Evolution of fusion reactivity <mT>, temperatureT, and the ratio <av>M(T)/<av>, for an initialMaxwellian distribution at T = I keV. Evolutionincludes ion losses for pR = 1O~* g cm~2.

Figure V-7 presents the reaction history of a pelletwith a density of 180 g/cm\ a density-radius product pR.of IO~4 g/enr. and a 1-keV Maxwellian initial ion dis-tribution. The <ov> average drops very rapidly, due.largely, to a decrease in the average ion kinetic energy.The ratio <ov>.M/<av>. where the numerator iscalculated with a Maxwellian distribution at the meanion kinetic energy and the denominator is calculated withthe distribution from the Fokker-Planck code, reaches amaximum of 3.9 at a density-time product nl % I012

s.crrT3 when the mean ion kinetic energy has dropped bya factor of 2 and <ov> has dropped by a factor of ~600.Thus, the dominant effect is energy loss, but the tail lossis not altogether negligible. Earlier in the history it is

>

°£i—

(£Ula.SUJ\—oz<IAt)

bV

I U

,n"510

5 io"6

2 | 0 " 7

OD

ac

^ I0 * 8

I 0" 9

lO'10

-

I

:I

-

-

_ _

V\\\ \\ 'V

\

1 1 1

ORIGINAL MAXWELLiAN :

( T - 1 k»V)

MAXWELLIAN AT PRESENT "=ENERGY ( T - 0 7 5 « k e V ) \

-CALCULATtD DI5T«IBuilON _.

^ ^ FUNCTION I

VC1'" '" J\\ \ '--\ \ \ E\ \ ^\ \ \ n\ \ \

\ \ \ \\ X '-

, , , l \ , , , \ , , , X \5 10 15ION EH£M1 (NANOfBGS)

20

Fig. V-8.Distribution function for case of Fig. V-7 at time nl- 2.01 X lOn s c»r}. Also shown are the original

1-keV Maxwellian and the Maxwellian at the pre-sent energy. The latter yields a <av> equal to theevolving distribution function if truncated at ET —4.37 keV.

relatively more important. For example, at nt = 2.01 XJO" s.cm'i when the mean ion kinetic energy is 0.76 ofthe starting value. <ov> is down by a factor of 14.8, ofwhich a factor of ~6 is due to kinetic energy loss and 2.4is due to tail depletion. Figure V-8 shows the distributionfunction for this case, together with a Maxwellian dis-tribution of the same mean energy. To reduce <ov> by afactor of 2.4 at this temperature would require a cutoff at4.37 keV. wh'di is also shown.

To summarize, we have shown that losses from the tailof the distribution function in small laser targets (thosewith an areal mass = pR IO~2 g/cm2) significantly affectthe reaction rate, partly because of the decrease in ionkinetic energy and partly because of the deviation from aMaxwellian distribution that results from preferentialloss of fast ions.

88

TARGET DESIGN

Exploding-Pusher Target Designs for Helios (W. P.Gula)

From the time Helios became operational, we haveneeded the capability to predict the performance ofvarious targets. Accordingly, we performed numerouscalculations. The two main purposes of the study are toaid in the design and interpretation of experiments onHelios and to calibrate our codes with the experimentalresults.

Single shell DT filled GMHs were chosen for thisstudy because they are relatively simple and have only afew variable parameters, the targets are easy to diagnoseexperimentally, and we understand the physics of theseexpkxling-pusher targets. In our one-dimensionalcalculations the glass shells were perfect spheres with aninitial density of 2.4 g/cnv\ filled with an equimolar mix-ture of deuterium and tritium.9

Many parameters of the laser pulse and of the targetswere varied. The results show that the GMBs shouldhave a shell thickness of ~1 \im, a radius of 75 to 300uni, depending on the laser power, and a fill pressure of10 aim or less. The neutron yield as a function of laserpeak power for a given target and pulse width is shownin Fig. V-9. The steepness of these curves emphasizes theimportance of the dependence on peak power of theexploding pusher targets in this regime. The one excep-tion is the curve for targets of 100-|um radius: its slope ismuch less steep, consistent with data from earlier workwith pulses of lower power. The steepness is less becausethe implosion time of a 100-um target is so short thatonly limited amounts of energy can be transferred to thefuel before disassembly begins, even for very high peakpowers.

Preliminary target shots on Helios were made at lowpowers (2 to 5 TW). Neutron yields ranged from 2 to 3X 10*11. in agreement with calculations for the specifictargets used.

5 K) 15 20PEAK POWER (TW)

Fig. V-9.Neutron yield as a function of laser peak power.

tions. and the rewriting of the laser ray-trace package toallow for rays that do not pass through the coordinate Z-axis. The effort to adapt some codes to the CRAY-Icomputer was redirected slightly to aim at FTN compila-tion. This effort is progressing well and will be finished inlate summer 1979.

CODE DEVELOPMENT

Introduction (K. A. Taggart)

The two major efforts in code development during thepast 6 months were the incorporation of nonlocalthermodynamie-equilibrium physics for opacity calcula-

Oflf-Axis Ray-Trace Routine (G. A. Rinker)

The laser ray-tracing routine in previous versions ofour codes required that each ray pass through the Z-axisat some point for the sake of simplicity. If we view acylindrical coordinate system in three dimensions andimagine representing a laser beam by a bundle of

89

straight-line rays, we can see that a general descriptionmay be obtained only if some rays are allowed to missthe Z-axis. The on-axis description can be complete onlyfor a beam focused to an infinitesimal spot size on theaxis.

Although a complete three-dimensional ray trace (inCartesian coordinates, for example) would increase ex-ecution lime significantly, a suggestion made earlier maybe used to exploit the cylindrical symmetry of the targetand remove analytically the angular degree of freedom.10

This treatment does not introduce additional physical ap-proximation because the noncylindrically symmetric raypath is effectively averaged over all angles when coupledto the cylindrical target.

The geometrical ray-trace equation is

r cos y L2 cot2 y + (2 sir y - (V-14)

r = J'.n2(r) (V-10)

where r describes the position of the unit-mass particlewhose path describes the ray trajectory, and

n 2 ( r ) = (VII)

is the index of refraction, where pe(r) is the electron den-sity measured in units of the critical density. The speedof thy particle has been set to I. For a cylindrically sym-metric target. pe(r*) does not depend on 6. so that theabove equation becomes

and

where

(VI2)

(V-13)

is the conserved angular momentum. Thus, the completethree-dimensional solution may be obtained from theformer restricted solution merely by the addition of anew term to the density, yielding an effective densitype(r,z) + LJVr2. In vacuum, the equations may be solvedanalytically, with the resulting ray path

where y and rB are parameters that describe the rayasymptotes. The distance of closest approach of the rayto the Z-axis is

-Lz/sin y (V-15)

The ray-trace equations are reduced to those usedpreviously in our design codes when L,, = 0.

This new capability was implemented in the codes. Agiven ray is traced in vacuum according to the hyper-bolic path given in Eq. (V-14). When a material boun-dary is intersected, the ray is traced through with linearapproximations to the term L|/r2 in each cell as is nor-mally done for the true density. This overall proceduresutlers from a numerical problem due to the singularityin L|/r2 as r approaches zero, with the result that raysarc not traced accurately in the cells adjacent to the z-axis. However, this inaccuracy has little physicalsignificance, because there is very little volume in suchcells. Further problems arose in implementing thiscapability because the previous critical-surface turning-point detector is not valid in the new context. This situa-tion was corrected, and an approximation to the ob-served angular-dependent resonance absorption was ad-ded at the same time.

The problem of physical interest concerns a laserbeam focused to some finite spot. To approximate thissituation by using geometrical ray-trace optics, it isnecessary to produce a bundle of rays distributed ac-cording to some intensity pattern in the focal plane. Thispattern is obtained in the following way. A nominal focalpoint is described as before in the (r.z) plane; then ran-dom fluctuations in Lz and the angle y are imposed toyield an intensity pattern corresponding in the limit of aninfinite number of rays to a Gaussian of specifiedFWHM. More general intensity patterns may be ob-tained with superpositions of such Gaussians, subjectalways to the cylindrical averaging process mentionedbefore, which transforms off-axis focal points into rings.These fluctuations are rerandomized at each cycle. Ad-ditional random fluctuations in Lz and y may be imposedat reflections within the target.

An example of the effects of these modifications isshown in Figs. V-10 through V-13. Figure V-10 shows aball with a laser focused to a point on each side of the

90

zoo.

-200 200

Fig. V-10.On-axis laser illumination of spherical target.

10'

1

wzUl

10'-100 •so SO 100

fig. V-Jl.Laser intensity on sphere of Fig. V-10 as a func-tion ofz.

200

zoo

Fig. V-12.Off-axis laser illumination of sphere in Fig. V-10.

-100 100

Fig. V 13.Laser intensity on sphere of Fig. V-12 as a func-tion of:.

sphere: the corresponding intensity as a function ofz isshown in Fig. V 11. Figures V-12 and -13 show the sameconfiguration except that finite focal spots of 100 urnhave been imposed. A significant change in overall il-lumination is evident from the trend of the graph in Fig.V-13. Note that the fluctuations are statistical and tendto disappear over time because of the rerandomizing car-ried out during each cycle. The increase in symmetry isthe most obvious manifestation that our new laser ray-tracing capability reflects the physics of pellet irradiationmore accurately.

REFERENCES

1. H. Brysk and A. J. Scannapieco, Bull. Am. Phys.Soc. 23, 750 (1978).

2. M. A. Stroscio, D. B. Henderson, and A. G.Petschek, Nucl. Fusion 18, 142.5 (1978).

3. D. W. Forslund, J. M. Kindel, and K. Lee, Phys.Rev. Lett. 39, 284.

4. R. J. Mason, "Double Diffusion in Hot ElectronTransport in Self-consistent E and P Fields," Phys.Rev. Lett. 42 (1979).

91

5. R. A. Gentry. R. B. Martin, and B. J. Daly. "Anl-ulerian Differencing Method for UnsteadyCompressible How Problems." J. Comp. Phys. I,

6. D. B. Henderson. Phys. Rev. Lett. 33. 1. 142(1974).

7. M. N. Rosenblmh. W. M. McDonald, and D. L.Jtidd. Phys. Rev. 107. I (1957); and W. M.McDonald. M. N. Rosenbluth. and W. Chuck. Phys.Rev. 107. 350 (1957).

8. H. deWitt. Lectures in Theoretical Physics Vol. IX.Pan C (Gordon and Breach. 1967) p. 621.

9. C. W. Cranfill. "One Dimensional ComputerSimulations of the Implosion of Single-Shell Targetswith the Los Alamos Two-Beam CO, Laser." LosAlamos Scientific Laboratory report LA 6827 MS(1977). Fig. II.5.

10. \:. Tappert. University of Miami. Florida, and II.Lindman. Los Alamos Scientific Laboratory, privatecommunication (1978).

92

VI. LASER FUSION TARGET FABRICATION(R. J. Fries!

Our target fabrication effort, supported by extensive theoretical investigations, supplies thermonuclear fuel containing pellets for laser drixen compression arul heating experiments. These targets, which range from simple, deuteraled. flat plastic films to coinplcx. mullilayercd structures containing cryogenic solid DT fuel, are optimized for usewith high power CO, lasers. After a target has been designed, we develop thetechnologies to produce the materials and properties desired. Then, we rill the targetwith thermonuclear DT fuel, assemble the necessary parts, and develop methods tomeasure and characterize all these properties. Finally, we insert the target in the targetchamber and position it at the exact laser focal spot.

INTRODUCTION (E. H. Farnum. R. J. Fries)

Our target fabrication effort has two objectives:(1) We supply targets of current design lo satisfy the

needs of the experimental program. Targets of variousdesigns are used: thermonuclear compression targets forMain Sequence experiments, partial and modified compression targets for Target-Essentials experiments. ;.wide variety of targets for Support Physics experiments,and several types of targets for Military Applications expcrimems. All are irradiated in one of our laser systems.

(2) We develop new techniques as needed for thefabrication of future thermonuclear compression targetsenvisaged in the Inertial Fusion Program plan. Becausethese target designs are not fixed, we must continue to in-vestigate all materials that have a high probability of be-ing used in the final designs. We also develop methods ofmeasuring and characterizing these materials and ofassembling them into required configurations.

In addition to these goals, we provide micromachin-ing. inieroassembly. and materials fabrication services toother groups.

A recent design for the Polaris-A multishell laser fu-sion target is shown in Fig. VI-1. Laser fusion targetshave evolved slowly to this stage. Initial targets, calledSirius, comprised only one shell filled with room-tcmpcralurc DT gas. This single shell acts as an ex-ploding pusher in which the fuel is preheated as it is com-pressed. Most of these shells arc GMBs with only a fewatmospheres of DT gas till. Although these designs can-not lead to a high yield, they do produce some ther-monuclear reactions and neutrons. The target isimproved by adding a 100-u.m-thick layer of plastic,which is vaporized by the laser pulse causing the glass

-ABSORBER

-OUTER PUSHER

DT ICE OR FOAM

INNER PUSHER

OTICE

OVERALL WAM£TER«2*m

Fig. VI-1.Polaris-A, a multilayer multishell target designed

for An tares and Helios.

shell to implode hydrodynamically. By coating the GMBwith molybdenum or other high-Z metal before plasticdeposition we reduce preheat, and thus, make the desiredadiabatic hydrodynamic compression possible. Such adesign will be used to attain our 20XLD milestone.

In our Polaris-A target, this high-Z metal coating isadded as a heat shield for the fuel, the ablator layer ischanged to low-density plastic foam to provide ahydrodynamic cushion, and a second pusher layer is ad-ded outside the foam.

Initial targets will irradiate the outer pusher shelldirectly, but in subsequent experiments other layers willbe added to provide further thermal insulation, to

93

enhance laser absorption, and to improve hydrodynamiccoupling of the laser energy to the target.

In all these targets, from the simple Sirius to Polaris(Fig. VI-1), the fuel will be frozen as a solid layer of DTice onto the inside surface of the innermost pusher shell.Calculations show that this cryogenic modification willimprove the yield substantially. We are developingtechniques to freeze such layers in place within theHelios target chamber.

These and other targets of recent design call for thedevelopment of high-Z metal shells with diameter andwall-thickness uniformity deviations of net more than1% and a surface smoothness tolerance of <1000 A.They also require low-density, small-cell-size plasticfoams, thick layers of plastic loaded with metals of lowor high Z, metal foams, and smooth metal layers ofmoderate or low Z. In addition, we are developingmethods to prepare alternative fuels that are solid atroom temperature and contain fuel atoms at high density(e.g., polyethylene, lithium hydrides, or ammonia borane,in which the hydrogen is replaced by an equimolar mix-ture of deuterium and tritium). Even though the nonfuelatoms in these compounds dilute the nuclear fuel andreduce target performance, the fact that these materialsare solid at room temperature may be an advantage, es-pecially in designs that require fuel-containing layers inthe outer parts of the target.

Because any target must be characterized completelyto understand its performance, we have devoted much ofour effort to measuring and documenting all targets wedeliver and to developing new, automated high-resolutionmethods of characterizing target parts. We are develop-ing a surface acoustic-wave resonator-driver to sortbatches of target shells for diameter and wall uniformityusing the efficiency (Q-value) and frequency at which theshells bounce as an indicator of quality. We are alsoautomating our x-ray microradiographic method forobserving defects and nonuniformities in opaque shells.

TARGET FABRICATION (R. J. Fries, E. H. Farnum)

General

Our primary assignment in target fabrication is the as-sembly, delivery, and postshot analysis of targets for ourthree operating laser systems. We therefore devotewhatever fraction of our effort is required to meetdelivery requests from experimenters on these systems.

In addition, we try to maintain a large and very flexibleinventory of materials and techniques so that we can res-pond rapidly to changes in target design or specifica-tions. As part of target assembly, we also mount andalign the targets on the appropriate target-changingmechanism to eliminate the need for any further positionadjustment in the target chamber. If desired, wephotograph and analyze any remnants to supplementtarget diagnostics.

The process of target fabrication typically includes thefollowing steps. A type of GMB is chosen that is mostsuitable for the desired diameter, wall-thickness, and gas-permeation characteristics. These GMBs are availablefrom the 3M Company, KMS Fusion, Inc., or theLawrence Livermore Laboratory. The GMBs areprecision-screened to the desired size on an air-jet sieve,separated according to density by flotation in SF6 gas,crushed by external pressurization (only for cases wherethe desired wall thickness is greater than 1.5 urn), andcleaned by flotation in ethanol. The refined batch is op-tically preselected with an interference microscope andthen filled with DT or DT:Ne gas mixtures by permea-tion through the wall at elevated temperature and pres-sure. The filled GMBs are cleaned in ethanol and ex-amined carefully with an interference microscope.Targets passing optical inspection are examined andmeasured in three planes. Coatings of metal and/orplastic are then applied. These coatings are examined inthree views by x-ray microradiography to determinecoating smoothness and uniformity, often after each suc-cessive layer has been applied. In the meantime, otherparts are being fabricated. For example, for the two-shellvacuum insulation target (see Fig. VI-2), we form theouter plastic shell by coating an appropriately sized cop-per ball bearing with plastic (by our low-pressure plasmapolymerization method described below), cutting thecoating into two halves with a laser knife, and etching thecopper mandrel away in nitric acid to leave the plastichemispheres. When the central fuel container is opaque,we use x-ray microradiography to determine shell unifor-mity.

Fabrication Activity (B. Cranfill, V. Woods, M. Calcote,J. Feuerherd)

We supplied more than 1000 targets for our threeoperating laser systems and made more than 100 partsfor diagnostic devices. About 320 targets were fabricated

94

Fig. VI-2.A two-shell target consisting of a GMB and asingle concentric plastic shell.

for main-sequence experiments. Another 340 targetswere supplied for a wide variety of Target-Essentials andSupport-Physics experiments, including preliminaryshots on the Helios laser, and over 400 were made formilitary applications experiments.

We completed the development of fabrication techni-ques for the current Vega-B target. We also improvedour low-pressure plasma polymerization technique andcan now coat mounted GMBs with over 350 urn ofparylene for the 20XLD milestone target. The surfacesmoothness of this coating is acceptable (± 1 urn). Wesolved the problem of perturbations near the mountingstalk, which are introduced because both the GMB andthe stalk are coated simultaneously in the coatingprocess. We are replacing the coated stalk with a new,uncoated one by micromachining with a single-point dia-mond tool that has a concave tip. This techniqueproduces a smooth spherical surface at the cutoff. Wesuccessfully fabricated several targets for an x-ray point-source experiment which requires a tungsten needlesuspended behind an aperture in the configuration shownin Fig. VI-3.

Gas Analysis of Laser Fusion Targets

General. To predict the performance of hydrogen-,deuterium-, and DT-filled microballoons correctly, wemust be able to measure the gas fill with a high degree ofaccuracy. Also important are the permeation rates of thehydrogen isotopes through thin layers (shells) of avariety of coating materials at various temperatures. Thisinformation will be helpful in deciding under what condi-tions a given target type can be filled in real time.Permeation data at room temperature and below willalso be important in selecting the best storagetemperature for targets filled with the desired quantity offuel gas.

Filling Large-Aspect-Ratio Shells for the 20XLDMilestone Target (Betty S. Cranfill). The 20XLDmilestone target requires 300-um-diam, 1-um-wallGMBs filled to 30 atm (3 MPa) of DT gas at roomtemperature. Two problems were encountered in fillingthese targets. (1) Commercially available batches of 300-um-diam GMBs are of very poor quality and of a glasscomposition that permits high DT loss rates even whenstored at - 80°C (193 K); and (2) the aspect ratio(diameter/wall thickness) of these shells was so large thatthey collapsed by elastic buckling during the fill (theresulting shock wave and/or fragments break other near-by shells) with the result that even with precrushing andcareful staging, only ~10% of the GMBs loaded for DTfill survived. Both problems were solved during thisreporting period; we are now filling 300-um-diam, 1-um-wall GMBs with 30 atm of DT and achieving 90%recovery.

In trying to solve the first problem, we obtained someGMBs from KMS Fusio Inc., that were not only ofvery high quality with respect. wall uniformity and sur-face finish, but also had very low permeability to DT atroom temperature due to ~\-¥o CaO content in theglass. We measured the outgassing rate of GMBs fromKMS Fusion, Inc., with and without CaO addition bymonitoring the x-ray emission of a number of shells overa period of time; the results are shown in Fig. VI-4. Thehigh quality of these shells significantly reduces the timeneeded for selection and improves the gas-filled yield byreducing elastic buckling failure.

We solved the buckling problem by isolating the GMBduring fill and measuring buckling failure so that fill-staging might be optimized. Isolation was provided by a

95

Fig. VI-3.An x-ray backlighting target. The tungsten needle pointed toward the hole emits x-rays from its tipwhen healed bv the laser.

Fig. VI-4.X-ray count rate data from KMS Fusion, Inc.GMB as a function of time. The relative time forDToutgassing is shown for GMB with and withoutCaO.

4 5TIME (months)

96

wire with holes drilled into it, each of which acts as an in-dividual container for a GMB during the DT fillingprocess. Thus, each GMB is shielded from its neighbor.The wire containing the GMBs is then placed into aclose-fitting glass capillary which retains the GMBs intheir holes. Another advantage (other than 90% recoveryfrom 30-atm DT fill) of this technique is that the identityof the GMB is maintained during fill, which permits therecovery of precharacterized shells. Because the fill gascontributes to the wall thickness measured by optical in-terferometry and must be subtracted, wall thickness canbe measured more accurately on an empty GMB. In ad-dition, wall thickness carefully measured before and afterfilling can be compared to obtain a nondestructive, quan-titative gas assay. Although this method is much moretedious than x-ray counting, we used it as an aid incalibrating the x-ray counting system for glass shells.Variations among different shells indicated an x-raycalibration error of ~±7%.

Data for interferometrically measured gas content inGMBs at several fill pressures are plotted with the ex-pected fill density in Fig. VI-5. All these GMBs, whichwere held at fill conditions for 24 h, showed a lower-than-expected gas content. In contrast, a 10-day fill to anexpected 28 amagat at the same temperature showed

5 0 -

II 2.0

100 ISO 200 250 300 350Outside Diameter!urn)

Fig. Vl-5.Interferometric measurements of the amount ofDTgas in GMB. The difference in optical path lengthbefore and after DT fill in a GMB is plottedagainst the outside diameter of the shell for severalJill pressures and is compared to the curve expectedfor the Jill conditions used.

29.8 ± 1.4 amagat in 10 shells. These data, which wouldindicate a permeation time longer than 24 h conflict withthose of still another experiment in which similar GMBswere emptied of gas by heating in air for only 4 h.Further experiments to resolve these contradictions arein progress.

Buckling Failure in Large-Aspect Ratio Shells (JohnV. Milewski). The tensile strength of shells determines thedesign limitations on the maximum achievable fuelloading. A method of testing for tensile strength was dis-cussed in the iast report (LA-7587-PR). However, thecurrent limitation in loading shells with fuel is not thetensile stress from internal gas pressure but rather com-pressive buckling failure during diffusion-fill. The lowbuckling strength of larger diameter spheres with thinnerwalls severely limits the external filling pressure that canbe applied to the shells during diffusion-filling. For exam-ple, at present, we use 100-psi fill stages for 300-um-diam, 1 -um-wall GMBs; to achieve a fuel density of 30amagat at 673 K, we must use 11 stages at 24 h each.Because relatively few preselected GMBs are filled in asingle run, optimizing the fill staging is obviously veryimportant.

We built an apparatus for measuring the crushingstrength of shells and determined the room-temperaturecrush strength for GMBs in the size range currently be-ing used for targets. These data agree very well with thetheoretical relationship for elastic buckling (Fig. Vl-6)given by

exx

where Ptxx is the externally applied pressure at whichbuckling occurs, Y is Young's modulus, v is Poisson'sratio, Ar is wall thickness, and r is shell radius. Note thatthe buckling-failure pressure is inversely proportional tothe square of the aspect ratio (diameter.wall thickness)and that an increase in diameter from 100 to 300 urn atconstant wall thickness should reduce the buckling-failure pressure by a factor of 9. For example, 100-um-diam shells with 1-um-thick walls do not fail by compres-sive buckling at room temperature until ~5000 psi of ex-ternal pressure is applied, whereas 300-um-diam, 1.0-um-thick targets fail when ~55O psi of external pressureis applied.

We do not expect this relationship to apply at elevatedtemperatures because the glass becomes plastic rather

97

20,000

100200 300 «C0

ASPECT RATIO (diomeler/wollthicliKJsl

500 COO

Fig. VI-6.Data obtained by hydrostatically crushing GMB at room temperature are plotted for shells of dif-

ferent aspect ratio and are compared to the theoretical curve for elastic buckling failure.

than remaining an elastic material. It is, therefore, impos-sible to predict the properties of thin-walled glass shellsat elevated temperatures; they must be measured direct-ly. We are adding a furnace to our apparatus, which willallow crush strength determinations up to 65O°C (923K).

INORGANIC COATINGS DEVELOPMENT (A. T.Lowe)

General

Because future targets will require layers and shells ofmetal or plastic, we continued to develop a wide varietyof coating techniques. Metal coatings were applied by thewet chemical methods of electro- and electroless plating,by chemical vapor deposition (CVD), by physical vapordeposition, and by sputtering. We focused our attentionon high-quality coatings and on the coating and recoveryof a small number (~100) of preselected GMBs.

Electroless Plating (S. Armstrong, W. Doty)

Previous electroless nickel plating studies used eitherSplacels or CVD nickel-coated GMBs as the substrates.We developed simple and reliable chemical treatmentsteps that allow us to plate thin layers of electrolessnickel onto GMBs without the need fcr CVD nickel

coating. To carry out the chemical treatments on smallnumbers of GMBs, special glassware and handlingtechniques were also developed. We applied thesechemical pretreatment steps—which consist of cleaningin a hot aqueous solution of Na2CO3, followed by sen-sitizing and activating with SnCl2 in HCI—to GMBsmade by 3M, LLL, and KMS, and achieved metalliza-tion. However, this initial electroless nickel coating isvery fragile, and adhesion to the glass surface is verylow. Suspension and movement of the GMBs duringelectroless nickel plating is essential, yet too hard an im-pact against the retaining screens, or against each other,can cause surface defects. Another cause of defects onGMB surfaces is the presence of gas in the form of air,H2, or water vapor from near-boiling solutions, whichcauses the GMBs to align themselves at the gas-liquid in-terface and results in rough, pitted deposits.

An all-fluorocarbon plastic electroless nickel-platingfixture that will fit into a \-( beaker and that contains an0.83-ntf plating chamber was constructed. GMBs wereagitated by a slow uniform up-and-down motion throughTeflon screens of 74-um opening. Two frequent types ofdefects were observed: small holes and tearing. Webelieve these defects are caused by collisions with thescreen and by poor adhesion to the glass substrate. If thedamage occurs in the very early stages of deposition, asmall piece of electroless nickel is completely removedand a hole is left. At a later time, the electroless nickeldeposit is thicker and a screen impingement produces atear, but not complete removal of a piece of nickel. This

98

flap of nickel will lie unevenly on the surface and subse-quent deposition forms a rough modular part. We are in-vestigating ways of improving the adhesion between theGMB and the initial nickel coating.

Chemical Vapor Deposition (G. J. Vogt)

We completed construction of the fluidized-bed CVDreactor designed to allow us to quantitatively analyze thefluid bed. This reactor system, designed specifically tostudy the WF6+H2 reaction, is also useful for other CVDreactions. We are investigating the system controls anddiagnostics.

Experimental work on a Solacel fluidized bed wasstarted with the WF,+H2 CVD reaction to ascertainoperating conditions for the optimum coating qualitiesand to relate these conditions to fluid-bed parameters.

We also completed the literature search for availablethermodynamic and molecular data for the WF (+H2

reaction. The subsequent computation for the ther-modynamic yield of CVD tungsten as a function of H2

and WFC feed ratio was performed by using these col-lected data. In general, for all pressures andtemperatures, a 99% tungsten metal yield can be ob-tained at equilibrium for an input ratio of H2 to WF,between 3.0 and 3.3. Note that the stoichiometric inputratio of H2:WF6 is 3.0, as given in the reaction

WF,6(g) + 3 H2(g) W(s) + 6HF(g) -

Because CVD is generally a nonequilibrium process dueto a short residence time in the reaction zone, additionalH : is necessary to ensure a high W(s) yield. Also, theyield of tungsten metal by CVD appears to be specifical-ly dependent on the kinetics of the process.

A similar thermodynamic analysis for W(C0), CVDis being considered as an alternative to WF4 CVD forGMBs because, in the earliest stage of the WF6 process,the GMBs are quite susceptible to HF attack, yieldingbadly damaged bed material. We could alleviate thisproblem by either fully coating the GMBs with theW(CO), process or by flash-coating the glass with theW(CO)6 process before further coating with theWFS process.

Sputter-Coating, a Gas-Jet Levitator Development (A.T. Lowe, C. Hosford)

We continued our development of sputter coatings onGMB and other spherical mandrels. By using the gas-jetlevitator described in the previous progress report,1 wehave successfully applied molybdenum and gold tolevitated GMBs. The configuration that allowed success-ful GMB levitation is shown in Fig. VI-7.

The GMB is observed from outside the chamberthrough a 10-cmdiam viewport with a Questartelescope, and illumination of the sphere is maintainedthroughout the run with a 4-mW He:Ne laser whosebeam passes through a second viewport and is directedonto the GMB with mirrors inside the sputteringchamber.

The gas flow (measured with a mass flowmeter) re-quired to levitate a GMB is between 0.1 and 0.2 cm'/minat sputtering pressures from 20 to 100 mtorr. At theseflow rates, additional argon must be added to the bell jarto establish sputtering pressures of 40 mtorr.

Initial liftoff of the GMB from the collimated holestructure (CHS) is difficult because of static charging ofthe GMB. Increasing the gas flow sufficiently to lift thesphere would often blow the GMB away (typical GMBweight, 2 ug). However, an electromechanical driver (asused for the shaker table) attached to the CHS vibratedthe structure sufficiently to gently dislodge the GMB and

KFIHTCMMfiNeriMMK

Fig. VI-7.Schematic of rf sputtering unit with its gas-Jetlevitator.

99

to permit liftoff at low gas flows. To keep static chargesto a minimum, the GMB must be coated with a thin(200- lo 500-A) metallic layer before levitation, stored inmetal containers, and treated with commercialelectrostatic-charge-eliminating devices.

The mctallographic cross section of a gold coating 1.4± 0.2 (am thick is shown in Fig. VI-8. This coating wasdeposited at a rate of 4.2 nm/min onto a 200-;im-diamGMB with a 2.0-um-thick wall.

To keep the sphere levitated, a delicate balance mustbe maintained between the two counteracting forces ofgravity and levitation-gas pressure. Therefore, thelevitation-gas flow must be increased as the weight of theGMB increases during coating. For the gold coating inFig. VI-8. we h~d to increase the flow from an initial rateof O.I cmVmin by 0.02 em'/min for every 100 nm ofgold film added. Preliminary coatings made with thistechnique show much cleaner, smoother surfaces thanthe vibrating-table scheme used previously.1

The disadvantage of this technique is the low deposi-tion rate achievable at the high pressures necessary forstable levitation. Finer gas-flow controls and improveddesigns of directional jets, which should allow operationat higher rf power and lower pressures, are being tested.

Microflow Valve (J. R. Miller, J. Feuerherd)

The need for precise gas flow with a controllable rateof 0.001 to 1.0 std cmVmin for the gas-jet levitatorcoupled with the lack of a suitable commercial device ledto our development of a microflow valve.

The simple device, shown in Fig. VI-9, relies on thethermal expansion difference between a precisely lappedtungsten plug and the surrounding brass body. As thetemperature of the assembly is raised by the resistanceheater coiled around the valve body, ihe brass expandsmore than the tungsten and a gap forms around the plug,allowing gas to flow through the device. Typical flowrate vs temperature at three pressure differentials for one

- J ~ \ - ^ — STAINLESS STEEL TUBING

| \ s ~ BRASS BODY

- RESISTANCE HEATER

TUNGSTEN PLUG

THERMOCOUPLE

STAINLESS STEEL TUBING

Fig. VI-8.Optical metallograph of the cross section of a 1.4-\im-thick gold coating sputtered onto a levitatedGMB. 2500X

GAS F L O *

Fig. VI-9.Schematic of temperature-controlled microflowvalve.

100

0014

0012

ioa7?" 0 0 M>—«

| oo«

0004

000?

. ' ' '

-

l i i

° 20 40

i 1 r T i l — » r ~i 1—

fiP=93KK)*l"«/

r===^*'i , , . , ,"« 0 BO 100 l?0 HO

TEMff RMURf (Cl

Fig. Vl-10.Flow rate for helium gas as a function of orificetemperature for the micro/low valve.

of these valves is shown in Fig. VI-10. The sensitivity ofthis valve can be changed by choosing plug and bodymaterials that have either greater or less difference intheir thermal expansion coefficients.

Mandrel Development (S. Butler, M. Thomas)

Work toward making mandrels for metal microbal-loons continued. This effort should producemicrospheres that can be coated with nickel by chemicalvapor deposition (CVD), and that can be removedthrough the coating, leaving a shell strong enough for ad-ditional coating. The resulting product is a free-standingmetal microballoon, with the inside diameter dependentupon the mandrel size and the wall thickness determinedby the coating process. We further investigated condi-tions for spherical droplet formation, using water in thedroplet generator described in Ref. i. The size of thedrops was measured by viewing them in free fall with atelescope, video-camera assembly, and stroboscopiclight, which have been calibrated for such measurements.Droplet size measured for a 140-um orifice agrees wellwith that predicted from mass balance, liquid-jetvelocity, and excitation frequency. However, dropsformed with a 68-um orifice are substantially larger thanpredicted (by 1.5 to 2.5 times) and the measured sizedoss not correspond to any overtone or simple divisor ofthe excitation frequency. We cannot offer an explanationfor the discrepancy yet.

Sodium nitrite spheres ~25O to 500 urn in diameterwere produced by dropping the molten material at 563 Kthrough a 300-ntn orifice. Although the drops solidifiedin the 1.5-m drop length, some spheres contained bub-bles. We have not yet metal-coated this product.Polymethyl methacrylate (PMMA) spheres were CVDnickel-coated at ~355 K. Although the PMMA does dis-solve through a thin (~2-um) coating without adverseswelling, most of the shells crumple without mandrel sup-port. Observation with an optical microscope showedthat some shells with a wall thickness of 1.1 to 1.4 urnremained intact after removing the PMMA. Becausehigh-energy radiation is known to decompose PMMA.promoting dissolution, this fact may allow easier removalof the mandrel material through the nickel wall.Therefore, we have irradiated some PMMA spheres witha linac electron beam. Evidence of decomposition of thematerial appears at doses of ~50 Mrad. but quantitativeanalysis of the change in solubility is not yet complete.

ORGANIC COATINGS DEVELOPMENT (R.Liepins)

General

Many of our multilayered laser fusion targets use anouter shell of low-density, low-Z material as an ab-sorber/ablator layer. This layer absorbs energy from theincident laser, is heated and vaporized, and streamsaway from the pusher shell causing the pusher shell toimplode via the rocket reaction forces. For target pelletsthat do not depend on the strength of this ab-sorber/ablator to contain the fuel-gas pressure, wegenerally use plastic. We also provide free-standingcylindrical and spherical shells of plastic as targets andfor special diagnostic measurements. These latterspecimens are generally fabricated by coating ap-propriate mandrels, which are then dissolved in acid toleave the free-standing plastic shells, but we are also try-ing to make plastic shells by using our droplet generatordescribed earlier. Finally, because the Polaris-A targetrequires a shell of low-density small-cell-size plasticfoam, we are developing techniques for making this foamin the form of spherical shells that can be placed aroundtarget cores.

101

Gas-Phase Coatings (R. Liepins. M. Campbell)

General. The primary problems encountered incoating GMBs with plastic are obtaining uniform coatingthickness and high surface smoothness. Of the threegeneral coating approaches (solid, liquid, and gas phase)the gas-phase coating approach gave the most promisingresults. Two gas-phase techniques were investigated: (1)a low pressure plasma (LPP) process, and (2) a vapor-phase pyrolysis (VPP) process, i.e.. the p-xylyleneprocess.

The primary requirements for coatings on laser fusiontargets are sphericity and uniformity of better than 5%and smoothness or defects less than I urn. We achievedthese requirements with a low-pressure plasma process,and are approaching success with a vapor-phasepyrolysis process. In addition, we can now apply thesecoatings to a few preselected GMBs with 100%recovery.

Low Pressure Plasma Process (LPP).

Introduction. Previous work demonstrated that manydifferent coating effects were caused by electrostaticcharge build-up on the GMB support structures. Wehave used this electrostatic charge effect to enable us tocoat as few as eight free-standing GMBs in the LPPprocess, a feat not previously possible. A study that cor-related the electrical power in the plasma with depositionrate and chemical structure of the coating revealed thatthe ionization potential of the organic species is one ofthe most important parameters affecting the depositionrate. Similarly, the ionization potential of the backgroundgas strongly affects the plasma density, and thereby af-fects the deposition rate as follows: the lower the ioniza-tion potential the higher the plasma density and the fasterthe deposition rate. Asa result, we were able to doublethe thickness of smooth, uniform coatings (up to IS urnthick) by changing the background gas in the plasmafrom argon to xenon, which has a lower ionization poten-tial.

Coating a Few Preselected GMBs. In all previous LPPcoating work, batches of many hundreds of GMBs werecoated at one time. The large number facilitated bounc-ing in the plasma. Attempts to coat fewer than 100GMBs failed because the GMB could not be made tobounce. We can now coat as few as eight GMBs byprecoating the aluminum plate electrodes with a thin

(<:o.5-um) coating of poly(p-xylene). Apparently, thisdielectric coating leads to a rapid charge build-up on theplates, which stimulates bouncing of the GMB.

Effects Caused by the Ionization Potential (IP) of theBackground Gas and Monomer. Previous LPP workused an argon background gas because the presence ofargon produced stiffer, stronger coatings and higher sur-vival rates for thin shells than coatings made withoutbackground gas. We studied the deposition rate andmaximum thickness of smooth-surface coatings forargon (IP ~ 15.8 eV). nitrogen (IP = 14.5 eV). andxenon (IP =12.1 eV). We found that substituting xenonfor argon triples the deposition rate, doubles the max-imum smooth-coating thickness, and improves theachievable surface finish. The data for nitrogen lie be-tween those of argon and xenon, as expected from itsionization potential.

To determine the effect of the ionization potential ofthe monomer on deposition rate and smoothness, wemade coatings with p-xylene, 1,3-buladiene. andethyleneat an input power of 700 mW. Increased ionizationpotential reduced the coating rate as follows.

Monomer

p-Xylenc1.3-ButadieneEthylene

lonization CoatingPotential rate

(eV) (nm/min)

8.459.1

10.5

25.811.93.6

Note that plasma polymerization differs strikinglyfrom free-radical chain-reaction polymerization in which1.3 butadiene polymerizes almost explosively and p-xylene will not polymerize at all.

Vapor-Phase Pyrolysis Process (VPP).

General, To extend our capability in the preparationof thick coatings with precise control of coatingthickness and surface smoothness, we also developed theVPP process. In principle, the process has no coating-thickness limitation and inherently leads to a conformalcoating of high surface smoothness and uniformity. Thecoating p-xylylene possesses high mechanical strength,good thermal stability, and high resistance to most sol-vents and chemicals. We increased the thickness of

102

coatings on GMBs to greater than 350 um, with surfacesmoothness approaching 100 nm. This coating can beapplied to mandrels that are etched out through a smallhole, thus leaving free-standing shells with aspect ratiosgreater than 22 and wall thicknesses at least as thin as 20um. We also installed a gas-jet levitator in a VPP ap-paratus and coated some levitated GMBs.

TUek, Smooth Coatings. To apply thick coatings, acold chamber with a wide temperature and gas-flowcapability was constructed (Fig. VI-11). Coating deposi-tion rates from 40 to 500 nm/min and depositiontemperatures from 283 K (+10°C) to 238 K (-35°C)were explored. Deposition rate has a direct effect on thecoating surface smoothness—the higher the rate, therougher the surface. By going from a deposition rate of180 nm/min to 40 nm/min, the surface roughnessdecreases from ±1 um to <±0.05 um. Thus, in prepar-ing very thick coatings a compromise must be madeamong coating thickness, surface smoothness, anddeposition time. Typically, a 100-um-thick coating witha I -uin peak to valley surface roughness can be obtainedin 8 h. A GMB with an 86 urn thick coating of p-xylylcnc applied at a rate of 177 nm/min and showing 1-um surface smoothness is illustrated in Fig. VI-12.

MONOMER

6MB

BAFFLES

- * - C O L 0 N 2

VACUUM

Fig. VI-11.Schematic of vapor-phase pyrolysis (VPP) coalingapparatus used to obtain thick (>350-\m),ultrasmooth (±100-nm) coatings on GMBs.

Fig. Vl-12.Scanning-electron micrograph of the surface of an86-^m-thkk VPP plastic coating deposited at 177nm/min. 5000X

We found two techniques that will improve surfacefinish even at high deposition rates. One uses a co-monomer that interferes with the crystallization processand the other uses a substrate coating witii a relativelylow number of nucleating sites. Addition of the co-monomer 2-chloro-p-xylylene to p-xylylene dramaticallyimproves surface finish even at high coating rates. FigureVI 13 shows an 87-um-thick coating polymerized with33 wt% of the above comonomer at a rate 1.9 timesfaster than the coating in Fig. Vl-12. To obtain anequivalently smooth surface with only a single monomer,we would have to use a deposition rate less than 40nm/min. which is eight times slower than the co-monomer rate used for the coating shown in Fig. Vl-12.A similar, but somewhat less effective smoothing can beachieved by applying a smooth aluminum coating to theGMB before plastic deposition. A GMB mounted on anAramid fiber was coated with 20 nm of sputteredaluminum and then with 41 um of p-xylylene at a rate of150 nm/min. The resulting surface (shown in Fig. VI-14)is better than that of a similar coating on a bare GMB(Fig. Vl-12) but not as smooth as a comonomer deposi-tion (Fig. VI-13).

Pimstk Coatings OH Levitated GMBs (with A. Loweand C. Hotford). To coat VPP plastic on a GMB for20XLD targets, we must mount the GMB on an Aramid

103

Fig. VI-13.Scanning-electron micrograph of the surface of an87-iim-thick VPP plastic coating deposited at 336nmlmin, but with a comonomer added to reducecrystallite growth. 4250X

Fig. VI-14.Scanning-electron micrograph of the surface of a41-\im-thick VPP plastic coating deposited at 150nmlmin on a GMB passivated with a thin sputteredcoating of aluminum. 4450X

fiber, apply the coating, and then machine the coatedfiber off to produce a smooth, spherical target. This is adifTicuit. time-consuming process which, at best, onlymarginally meets target specifications for sphericity andcoating uniformity. Therefore, we are trying to find away to apply thick plastic coatings to unmounted GMBs.Although we can apply LPP plastic up to 15 urn thick onbouncing, unmounted GMBs, we were unable to keep anumber of unmounted spheres in motion during the VPPprocess which will allow thicker coatings. To solve thisproblem, we installed a microballoon levitator in ourVPP cold chamber and applied up to 5-um-thickcoatings of VPP plastic to levitated GMBs. This levitatoris similar to that developed for sputter-coating (describedabove). We are optimizing the coating and levitationparameters to allow application of thick, smooth-surfaced coatings.

Chemical Analysis of Plastic Coatings by irSpectrometry. Qualitative and quantitative chemical

analysis of as-applied plastic coatings provides necessarydata to obtain better control of coating uniformity, agingcharacteristics, and composition, especially for the high-Z-loaded plastics. This information has been difficult toobtain because of the very small size and sphericalgeometry of the substrate. Consequently, we were forcedto analyze samples taken from the coater wall or from alarge number of shells, even though we knew that thesesamples might not be representative of the coating on asingle selected shell. We recently developed a techniquefor analyzing plastic coatings on GMBs as small as 600um in diameter by measuring the ir spectrum of a singleshell on a Fourier transform spectrometer. Because thisinstrument has the capability of subtracting the spectrumof the GMB substrate from that of the coated GMB, weare able to obtain spectra, and hence, to identify thequalitative chemical composition of the coating in situ.We are trying to calibrate the instrument so that we mayobtain quantitative data.

104

ORGANOMETALLIC COATINGSMENT (R. Liepins, M. Campbell)

General

DEVELOP- POLYMER FOAM DEVELOPMENT

General

The outer pusher of the Polaris target requires a high-Z. low-density (1.3 to 3 g/cm3) material as an x-ray andhot-electron shield for the DT fuel. We are investigatingboth metal-loaded plastic and high-Z metal foam for thislayer. By using easily vaporizable organomeiaLiemonomers, we have successfully deposited layers ofr.ietal-loaded plastic with both the VPP and LPP proces-ses.

Lead and Tantalum Organometallics

We applied LPP organometallic coatings to GMBswith tetraethyl lead (CH3CH2)4Pb and with pentaethoxytantalum (CH,CH2O)5Ta, both pure and mixed 50 wt%with p-xylene. Although very little tantalum was incor-porated in the applied film in any case, preliminarycoatings made with tetraethyl lead showed up to 30 wt%lead. In the VPP process we introduced 10 wt% iron viathe organometallic compound ferrocene into the p-xylylene vapor stream and observed 9.3% iron inclusionin the resulting film. Because some of this iron could beextracted by soaking the film in hot methanol we believethat at least some of the iron was molecularly dispersedbut did not react with the p-xylylene. These results aresummarized in Table VI-I.

Several advanced laser fusion target designs require~250-um-thick spherical shells of low-density foam(<0.05 g/cm3) with very small cells (< 1 urn). Ideally, thefoams should consist only of carbon and hydrogen.However, to increase the chances for producing thedesired foams, we also used polymers containingnitrogen and/or oxygen atoms. This broadened thechoice of polymers greatly, although heavy metals andother elements with atomic numbers higher than 20 aregenerally considered undesirable. Because the foamsshould have a uniform structure throughout, cell size anddensity should be uniform.

Summary of Small-Cell-Size Foam Development atMonsanto (J. L. Schwcndeman, Monsanto ResearchCorporation)

We have supported a development effort at MonsantoResearch Corporation to form small-cell-size, low-density plastic foam shells. The desired goal was a cellsize less than 2 urn and a density less than 0.1 g/cm3.This contract terminated in October 1978. Threematerials were investigated for meeting cell size and den-sity requirements in foam coatings for small spheres,with the following results.

TABLE VI-I

FORMATION OF HIGH-Z METAL-LOADED PLASTIC COATINGS

Appearance

Reflected light Transmitted light

CoatingThickness

(tun)

Metal(wt%)

Calculated Found

50:50 Pb:p-Xylene Dark brown, metallic Brown 7.9 — 13.350:50 Ta:p-Xylene Tan Tan 1.4 — 0.13Tetraethyl lead Black, metallic Dark brown 5,8 64.1 30.7Pentaethoxy tantalum Dark brown, metallic Brown 7.3 44.5 < 1.0Ferrocene :p-Xylylene Bright orange — 107.0 10.0 9.3

105

Evaluation of cellulose acetate asymmetricmembranes proved that these foam films have the re-quired small cell size. These films can be prepared atroom temperature and lend themselves to molding partsin matched positive-pressure molds and, possibly, tovacuum-forming of the desired foam shells. However, wewere unable to produce these foams in the desired den-sity range. At present. 250-nm-thick cast foam films,when dried, are 70 to 80 urn thick, with 2.0-um andsmaller cells and a density greater than 0.30 g/cm3.

Aropol water-extended polyester foam systems (WEP66 I P ® and D-WEP 123®) were also evaluated as apossible means of forming small-cell-size foams. We suc-ceeded in making water-extended polyester foams with adensity as low as 0.12 g/cm3, but failed to form foamshaving cell sizes less than the required maximum of 2.0um. We could mold these foams into small spheres.

Gelatin/water/toluene emulsion foams were also in-vestigated. Upon film draw-down, foams of small cellsize were produced. However, work on this material wasstopped because of concern as to the reproducibility ofthe properties of gelatin and its relatively low thermalstability.

As a result of the failure of all experiments in this con-tract, we are reconsidering alternative possibilities formaking foams that meet requirements for both densityand cell size and using pressurized gas in place of foamfor the buffer.

CRYOGENIC TARGET DEVELOPMENT (J. R. Mil-ler)

General

Laser fusion targets fueled with cryogenic liquid orsolid DT offer the advantage of high initial fuel densitywithout the disadvantage of diluent atoms being presentas they are in room-temperature solids having a highhydrogen density [e.g., lithium in Li(D,T), carbon in(-CDT)n, or boron and nitrogen in NH3BH3]. In addi-tion, calculations indicate that the yields from targetsFueled with liquid or solid DT can be considerably higherthan those from targets of the same design, but fueledwith high-pressure DT gas. Therefore, we actively pur-sued the development of cryogenic targets despite thesignificant experimental complications encountered intheir fabrication and in their use in laser-target interac-tion experiments. We also tested a cryogenic insertion

mechanism for loading and cooling these targets in theHelios target chamber.

Cryogenic targets receiving the greatest attention areuniform, hollow shells of solid DT condensed onto theinside surface of a glass or metal microballoon thatserves as the pusher shell. We previously developedtechniques to freeze the DT into a uniformly thick layeron the inside surface of the glass and to measure thethickness uniformity of the condensed DT shell. Becausehigh-yield targets require metal pusher shells and we arenot able to directly measure the uniformity of DT layersinside opaque shells, we made a calculational analysis ofthe FIF technique.

Fast Isothermal Freezing (FIF) (J. R. Miller, W. J.Press, R. J. Fries)

The investigation of any possible limitations of thefast isothermal freezing (FIF) technique for producinguniform, solid DT layers in target fuel cores continued.As reported in LA-6616-PR and LA-6834-PR, the FIFmethod forms layers of solid DT on the inner surface ofbare GMB fuel cores with a uniformity that is betterthan can be determined interferometrically (±10%variation in layer thickness). Because laser fusion targetsof advanced design have multiple shells and multiplelayers around the cryogenic fuel core, we continued theinvestigation of inherent limitations that may apply tothe FIF technique.

Because it is not possible to measure the solid DTlayer's uniformity in opaque, multishell targets in-terferometrically, and because no other diagnostictechniques are available to measure the uniformity of athin, low-Z shell (DT) inside one or more opaque high-Zshells (gold, for example), we developed a computermodel to calculate how long liquid exists in the fuel coreas it is cooled. This time can then be used to predict thegravitationally induced nonuniformity (NU) produced inthe frozen DT shell.

The NU produced in the frozen DT shell, defined asthe difference between the maximum and minimum shellthicknesses divided by the average shell thickness, de-pends on the length of time liquid is present in a givenfuel core. For example, a 100-um-diam fuel core filledwith DT to 10 MPa and frozen in 25 ms, or a 500-um-diam fuel core filled to 10 MPa and frozen in 100 ms willeach produce a NU of 0.095.2 Our heat-flow model

106

calculates the cooling rate of the fuel core, and hence, theexpected thickness uniformity of the solid DT layer.

Although laser fusion targets are radially symmetricand therefore can be modeled mathematically in onedimension, the cooling rate of a multishell target cannotbe solved analytically. This inability stems from thestrong temperature-dependence of both the thermal con-ductivity and the specific heat of most materials over thetemperature range of 4 to 40 K and from the thermal in-teraction of each layer in a multimaterial target.

The problem of calculating the length of time liquid ex-ists in a fuel core can be separated into two parts. Thefirst part consists of determining the time required tocondense all the DT gas into a liquid shell, and the se-cond part, determining the time required to freeze this li-quid shell. The sum of these two times is the quantityused to predict the NU of the frozen DT shell. Physical-ly, this partition of the problem assumes that the heatreleased by the condensing gas prevents the liquid fromcooling below the triple point until all the DT has con-densed.

An exact numerical solution to the heat equation forgas condensation in a multishell sphere withtemperature-dependent thermal properties was obtainedfrom a standard finite-difference formulation.3'4 Stable,physically realistic results for all times were obtainedwith a Crank-Nicholson scheme using a time-dependentweighting factor.3 With a constantly spaced nodal mesh,the transient target temperature distribution had minimaldependence on the time step used. Therefore, incorpora-tion of a variable time step into the finite-differencescheme had the added benefit of minimizing the com-putational effort. Less than 1 min of CDC 6600 cpu(central processing) time was needed for each multishelltarget evaluated.

The numerical solution for the DT condensation timein a simple two-shell target is shown in Fig. VI-15. Thefamily of curves, reduced temperature vs reduced radius,plots the condensation time of DT gas in a 100-um-diam,1-urn-thick glass fuel core positioned inside a second 1-um-thick glass shell. The target, initially at T, = 40 K,cools to TF = 19.7 K in 63 ms. In general, thicker shellsor larger target diameters increase the DT condensationtime.

Because the DT condensation time is strongly affectedby the helium diffusivity, and because thermal-propertydata for cryogenic, gaseous helium at low pressures arescarce,6 the temperature-dependence of the thermal dif-fusivity for helium was obtained by matching calculated

Fig. VI-I5.DT condensation-time calculation for a target withtwo concentric glass shells separated by 50 \im ofhelium exchange gas. the temperature of the DT isplotted against radial position.

and experimental results of the freezing time for a bare100-um-diam glass fuel core. The condensation time isalso a function of the density of the gaseous helium.

Although metal shells are not considered in Fig. VI-15, condensation-time results for such shells were ob-tained. The effect on condensation time of adding a metalshell is minimal: no significant time increase is seenbecause of the relatively high thermal conductivity ofmetals compared to that of other target materials.

The time required to freeze a liquid DT shell is the se-cond part of the liquid-existence-time problem. As stated,when the freezing time is added to the condensation time,we can predict the layer nonuniformity of solid DT in atarget fuel core.

An approximate solution to the heat-flow problem in-volving a liquid-solid phase change in spherical bodieshas been formulated.7 This approximate solution wasused to compute freezing times of liquid DT shells. Theresults of this calculation indicate that the DT fusiontime is short (< 1 ms) compared to typical DT condensa-tion times (>10 ms). Consequently, layer nonunifor-mities arise during the DT condensation time rather thanduring the DT solidification time.

Model calculations were performed to determine thelength of time liquid DT is present during the formationof cryogenic laser fusion targets. Our results make it pos-sible to predict the thickness uniformity of solid DT

107

shells in opaque, multishell targets. As shown, gas con-densation occurs at a much slower rate than does thesubsequent solidification of the liquid shell.

Helios Cryogenic Target-Producing Prototype (J. R.Miller, J. K. Feuerherd, H. E. Tucker, C. Cummings, R.Day)

The experimental apparatus designed to producecryogenic targets for the Helios laser system was testedextensively. Three specific areas investigated were (1) theheat load to the apparatus, (2) the cryogenic-shroudretraction system, and (3) the seals between the targetmount and the cryogenic shroud.

To cool the apparatus to the desired 6 K, we had toredesign the liquid-helium transfer system, the thermalstandoff between the target mount, the target cart, andthe target holder. In addition, a new cryogenic-shroudretraction system was designed, fabricated, and as-sembled. Evaluation of the performance of this devicewith high-speed motion pictures, discrete position sen-sors, and an accelerometer showed that it needs 6.8 msto move the required 3 cm. We expect this time to be suf-ficiently short. However, work is continuing on theshock-absorber subsystem needed to stop the rapidlymoving cryogenic-shroud retraction device with uniform,constant deceleration, which is needed to minimizemechanical fatigue in the entire system and vibration inthe Helios target chamber.

REFERENCES

1. F. Skoberne, Comp., "Inertial Fusion Program,January 1—June 30, 1978," Los Alamos ScientificLaboratory report LA-7587-PR (May 1980).

2. Extrapolated from T. M. Henderson, R. J. Simms,and R. B. Jacobs, "Cryogenic Pellets for Laser-Fusion Research—Theoretical and Practical Con-siderations," Adv. Cryog. Eng. 23, 682 (1977).

3. M. N. Ozisik, Boundary Value Problems of HeatConduction (International Textbook Co., 1968), p.413.

4. B. Carnahan, H. A. Luther, and J. O. Wilkes, AppliedNumerical Methods (John Wiley and Sons, Inc.,1969), 462.

5. S. V. Patonkar and B. R. Baliga, "A New Finite/Dif-ference Scheme for Parabolic Differential Equations,"Numerical Heat Transfer 1, 27 (1978).

6. D. C. Pickles and B. Colyer, "Heat Transfer ThroughLow Pressure Helium Gas at Temperatures Below 4.2K," Rutherford High Energy Laboratory reportRHEL/R 145 (1967).

7. Sung Hwang Cho and J. E. Sudderland, "PhaseChange of Spherical Bodies," J. Heat Mass Transfer13, 1231 (1970).

108

VII. APPLICATIONS OF LASER FUSION—FEASIBILITY AND SYSTEMSSTUDIES

(L. A. Booth and T. G. Frank)

Our feasibility and systems studies are being performed to analyze the technicalfeasibility and economic aspects of various commercial and military applications oflasers and laser fusion. Commercial applications include electric power generation,fissile-fuel production from fusion-fission hybrid reactors, and production of syntheticfuels such as hydrogen. Our studies also include assessments of advanced technologiesthat require development for the ultimate commercialization of laser fusion. The generalobjectives of these studies are the conceptualization and preliminary engineering assess-ment of laser fusion reactors and other energy production subsystems; the developmentof computer models of integrated plant systems for economic and technology tradeoffand comparison studies; and the identification of problems requiring long-term develop-ment efforts.

REACTOR DESIGN STUDIES

Introduction

Our reactor design studies determine the fundamentalengineering feasibility of reactor design concepts and es-tablish scaling laws for the sizing of reactors in in-tegrated plant designs for energy production applica-tions. In converting pellet output energy to useful forms,two primary design considerations prevail: (1) protectionof the reactor chamber structural wall from x rays andion debris generated by the pellet, and (2) conversion ofthese energy forms and of the high-energy neutrons touseful energy while producing the fuel constituent,tritium.

In general, these design considerations are essentiallyindependent of each other. Because ~80% of the pelletoutput is in the form of high-energy neutrons, this energyis most conveniently converted to thermal energy in a"blanket" surrounding the cavity and containing lithiumor lithium compounds for tritium breeding; the design ofthis blanket is essentially not dependent upon the first-wall protection scheme. Likewise, means of protectingthe first wall are generally not dependent upon theblanket design. Therefore, a variety of blanket designscan be incorporated with a particular wall-protectionscheme and vice versa.

We are studying three methods of first-wall protec-tion: (1) a lithium wetted wall,1 (2) a magnetically

protected wall with a low-Z sacrificial liner,2 and (3) amagnetically protected wall with a background fill gas.In the lithium wetted-wall concept, the x-ray output andion debris are absorbed by the ablation of a thin film oflithium that coats the inside surface of the first wall. Inthe second concept, the ion debris is ducted out the endsof a cylindrical configuration where the ions are eithercollected on large surface areas or slowed down byMHD decelerators. The x-ray energy is absorbed byablation of small amounts of the low-Z sacrificial liner.In the third concept, the ion debris is converted in thesame manner as above, but the x-ray energy is at-tenuated and converted into a blast wave, whichbeneficially increases the interaction time of this energyfraction with the first wall.

Our studies of blanket concepts include (1) a low-temperature (< 1000 K) liquid-lithium blanket for electricpower generation, (2) a high-temperature (1500-2000 K)boiling-lithium blanket for high-efficiency electric powergeneration and as a high-temperature process heatsource for hydrogen production, and (3) a lithium-cooledarray of elements containing fertile 238U and/or 232Th forthe production of fissile fuel and electricity as abyproduct.

Reported below are results from our studies on themagnetically protected wall concept, on the high-temperature boiling-lithium blanket concept, and on alow-power-density fissile-fuel production blanket(hybrid) concept.

109

Magnetically Protected Cavity Phenomena Modeling (I.O. Bohachevsky, J. J. Devaney)

We have programmed the mathematical description3

of the new plasma model representing the magneticallyprotected reactor cavity for use on the CRAY computer.The part of the code that solves the coupled ion-transport and electromagnetic field equations has beentested successfully; the part describing the behavior ofthe continuum background medium is being tested. Inaddition to completing these time-consuming tasks, ourtheoretical analysis has led to an improved under-standing of the physics of cavity phenomena and ofimplications of the postulated cylindrical symmetry.

We have derived expressions for the momentumtransfer from the high-velocity ions to the backgroundmedium. Some results are summarized in Table VIM,where we have used F, as the drag force in dynes actingon an alpha particle moving with the energy E(keV)through helium at the density of pg/cm3. Although ourexpressions are currently valid only for neutral gases,their use may be justified in cases of low backgroundionization. In addition, their availability in a formsuitable for computation enables us to test the code andprovides assurance that additional results will be derivedwhen needed.

MHD Decelerators for the Magnetically Protected WallConcept (I. O. Bohachevsky)

The use of structured pellets in inertial confinement fu-sion (ICF) reactors and of a buffer gas for cavity wallprotection4 implies that a significant fraction of fuel-pellet microexplosion energy will be deposited in thecavity environment. Therefore, there is a need to in-vestigate methods of removing and recovering thisenergy. Calculations indicate that the sputtering erosionof energy sinks,5'6 proposed for the removal of pellet-debris ion energy in the magnetically protected reactorcavity design, may be unacceptably high. Consequently,we had to seek alternative methods. We found that theenergy may be removed and recovered from the pelletdebris and the cavity medium by means of magneticfields in a magnetohydrodynamic (MHD) energy con-verter that decelerates the ions without suffering exces-sive wall erosion and that even generates electric energyin the process.

Previous studies7 assessing the applicability of MHDgenerators for energy removal from the fusion fuel-pelletdebris led to a more extensive analysis* of these devices,as summarized below.

There are conventional and induction-type MHDgenerators. In a conventional generator, the electrodes

TABLE VIM

DRAG FORCE ON AN ALPHA PARTICLE INHELIUM AT DENSITY pg/cm3

r , = p X 2.4113 X 10"4

S H = 0 . 0 4 4 0 7 E

SL = O.58EO.59

S = SH E > 10 MeV

50 keV < E < 10 MeV

+ AS 5 keV < E < 50 keV

1 keV < E < 5 keV

E (keV) 50 40 30 20 10 8 5AS 0.10 0.15 0.20 0.33 0.50 0.60 0.70

110

are exposed to the high-energy plasma flow and thereforetend to be eroded rapidly. In an induction-typegenerator, the power is extracted with induction coilslocated behind insulating refractory walls, which,therefore, are not exposed to the destructively hot plasmaenvironment.

The plasma exiting from the reactor differs from thatin a conventional MHD generator in two respects. (1) Itsconductivity is very high because it contains a largeproportion of metallic ions in high ionization states, and(2) it can be collimated with the magnetic field into ahigh-velocity beam.5'9 Consequently, the magneticReynolds number in the MHD channel will be high (>I)and the energy density will not be less than in a conven-tional (electrode) generator. These conditions make itpossible to exploit the engineering advantages of theinduction-type generator without penalties commonly as-sociated with low-energy density and efficiency if a con-ventional combustion-generated plasma is used in thisdevice.

To investigate the compatibility of magneticallyprotected reactor cavity designs with MHD generatorswe imposed two performance requirements. (1) Thegenerators should reduce the ion exit velocity so that thedebris, possibly with the aid of diffusers, could be used in

conventional heat exchangers, and (2) they shouldgenerate sufficient amounts of electricity to energize theirown magnets.

In our analysis we solved the one-dimensionalchannel-flow equations, determined the magnetparameters, and estimated the cost of MHD decelera-tion. Representative results summarized in Tables VII-IIand -III and in Figs. VIM and -2 show that nearly 65%of the ion kinetic energy has been either extracted in theform of electric power or converted into thermal energy.In our illustrations, QiPEA|. denotes the power residing inthe moving plasma, QHMC is the power dissipated inmagnet coils, and Q,cp is the power developed in the in-duction coil for an arbitrarily prescribed load current of500 A. None of the cases calculated was optimized inany way, and therefore, significant improvement in per-formances can be gained by simple parameter improve-ment.

To determine the economic effectiveness of MHDdecelerators, we compared their cost with the cost ofgraphite energy sinks. It is convenient to arrive at thedecelerator cost by estimating the magnet and channelcosts separately. The cost estimate for a conventionalmagnet (including core, winding, cooling, and some sup-port structure) is based on a unit cost of 100 000 S/m3;

TABLE VII-II

CONSTANT-AREA SUPERSONIC DECELERATOR0.35- by 0.35- by 1.00-m CHANNEL

Case*

1

Plasma Conductivity (mho/m) 1000Magnetic Field (T)Exit Velocity (m/s)Exit Pressure (N/mJ)Exit Mach NumberIdeal Power (W)

aln all cases:

0.13457841.94 X103

1.0075.35 X 104

Instantaneous Mass Flow Rate(kg/s)

Pulse Rate (Hz)Ion Pulse Duration (s)Inlet Velocity (m/s)Inlet Pressure (N/m2)

5101.24 X IO"4

104

105

2

5000.21957671.81 X 105

1.0013.62 x lO 4

3

1000.57658121.95X105

1.0061.01 X 104

111

TABLE VII-III

EFFECTS OF EXIT HEIGHT ON CHANNEL PARAMETERS OF A DIVERGINGAREA SUPERSONIC CHANNEL FOR PLASMA CONDUCTIVITY EQUAL TO 1000 mho/m

CHANNEL INLET, 0.35 by 0.35 m; CHANNEL LENGTH, 1 m

Case1

1

Channel Height, Exit (m) 0.35Mach Number, ExitExit Velocity (m/s)Exit Pressure (N/m2)Ion Pulse Duration (s)Ideal Power (W)

"In all casts:Inlel Velocity (m/s)Magnetic Field (T)Inlel Mach Number

1.00757841.94 X1.24 X5.35 X

I04

0.1341.844

10s

10-"104

2

0.41.18767381.42 X1.20 X5.91 X

105

10-"104

3

0.51.29873201.04 X1.15 X6.98 X

10s

JO"4

104

4

0.61.33575238.40 X1.12 X8.03 X

104

io-4

104

5

0.71.34175637.17 X1.10X9.05 X

104

lO"4

104

the channel costs are estimated at less than SI00 000.The cost of energy sinks is based on the graphite(cheapest energy-sink material) cost of 11 $/kg. A moreaccurate cost model of MHD decelerators has been con-structed and will be used when a more extensive andreliable economic data base becomes available.

The conclusions inferred from our investigation maybe summarized as follows.

• MHD decelerators are a competitive alternative toenergy sinks if the sink erosion rate exceeds ~1 to 2cm/yr.

• Net electric power production with steady-statemagnet operation appears possible only when theplasma conductivity is greater than ~100 mho/m.For lower conductivities, pulsed or superconductingmagnets must be used; however, the low power out-puts obtained at such low conductivities indicatethat these designs may be economically infeasible.

• Power output and cost of the MHD decelerators areboth proportional to their size; therefore, a tradeoffinvestigation must be carried out to determine themost cost-effective configuration.

• Performances of induction and electrode-typedecelerators are approximately equal. Theinduction-type decelerators generate slightly morepower because the average magnetic Reynolds

number along the channel is slightly higher than un-ity for the conditions. However, the electrode-typedecelerators may not be feasible because of limitedelectrode lifetime in the plasma environment an-ticipated for ICF reactors.

High-Temperature-Lithium Boiler Blanket Concept (J.H. Pendergrass)

An artist's rendition of the high-temperature-lithiumboiler concept, described in a preliminary form previous-ly, is shown in Fig. VII-3. The concept's operating princi-ples are as follows. Highly penetrating 14-MeV fusionneutrons give up only a modest fraction of their kineticenergy in passing through cavity walls of moderatethickness, and only a small fraction is absorbed bynuclear interactions. Most of the fusion-neutron kineticand interaction energy is deposited in a boiling, lithium-containing, liquid-metal blanket. Transport of vapor to,and its condensation on, heat-exchanger transfer sur-faces constitutes the primary blanket heat-transportmechanism, which would be self-pumping if provisionswere made for gravity return of the condensate. The con-cept offers two significant advantages: (1) the presenceof lithium in the blanket allows tritium breeding ratios of

112

12

100

Fig. VIII.Effect of conductivity on power allocation for aconstant-area supersonic duct.

1.0 or greater, hence tritium self-sufficiency is achieved;and (2) pure lithium and lithium-containing molten alloysconsidered for use have relatively low vapor pressures(0.4 atm at 1500 K rising to ~15 atm at 2000 K for purelithium) so that structural requirements are not severe.

The preliminary description of the concept3 listed anumber of potential advantages over other high-temperature fusion reactor blanket concepts. We haveattempted to further assess the possibility of achievingthese benefits and have developed conceptional solutionsto important design problems.

Design aspects considered include• The containment of molten lithium-containing al-

loys at elevated temperatures;• The recovery of most neutron kinetic and interac-

tion energy as high-temperature heat;

100 1000

Fig. Vll-2.Effect of conductivity on power allocation for adiverging-area supersonic duct.

• An allowance for differential thermal expansion andfor routine, periodic replacement of neutron-damaged first-wall structures while maintaining allother reactor functions;

• The design of the primary heat exchanger, withemphasis on limiting the tritium escape ratesthrough primary heat exchangers into power-cycleor process loops to acceptable values;

• Neutronics and tritium breeding performance;• The distribution of boiling and steady-state vapor

volume fractions;• The recovery of bred tritium by methods that take

advantage of high blanket temperatures; and• Electric power conversion cycles for topping, which

can take advantage of the high heat deliverytemperatures.

As blanket containment structural materials, we con-sidered ceramics, glass-ceramics, and graphites, all ofwhich possess the required strength and resistance tocreep in the 1500 to 2000 K temperature range. As a

113

RE ACTOR ANO HEAT EXCHANGEROUTER SHELLS AND HOT PIPING

STRUCTURAL DETAIL

LEGEND

BEAM TUBE STRUCTURALDETAIL

Fig. VI1-3.Artist's rendition of lithium boiler high-temperature tritium-breeding blanket and primary heat-tran-sport concept. (A) Outer containment clamshell, (B) boiling lithium, (C) inner lithium-cooled blanketcontainment clamshell including wetted-wall first-wall protection, (D) pellet injection lube, (E) laserbeam tube, (F) supersonic condenser, (G) lithium vapor, (H) vapor conduit, (1) process heat or powercycle heat exchanger, (J) process heat or power cycle stream, (K) condensate return line, (L) first-wall and beam-tube lithium coolant feed.

protection from attack by the liquid metal, we consideredcoating these materials with thin adherent layers of TZM(0.5:0.08:99.42 wt% Ti:Zr:Mo) or protecting the struc-ture with thicker, nonadherent TZM liners. Reflux ex-periments of 1000-h duration at 1923 K indicated thatTZM is not attacked by pure lithium under these con-ditions.10 First walls operating at blanket temperaturecould permit the recovery of x-ray and pellet debrisenergy at high temperature. However, TZM liners, sup-ported by cooled steel walls with insulation between linerand structure and possessing sufficient compressivestrength to transmit internal pressure loads to the load-bearing structure without collapsing, appear more attrac-tive.

A carbon foam, made by sintering together hollowspheres manufactured from petroleum pitch, which has adensity of 0.39 g/cm3, a thermal conductivity of only~10"3 W/cm-K, and a compressive strength of 1500

psia at room temperature, has been identified as a poten-tially attractive insulating material.11 However, lack ofknowledge concerning high-temperature physical proper-ties and behavior under 14-MeV neutron irradiation tohigh fluences compels us to examine other insulationconcepts as well.

Provisions for accommodating differential thermal ex-pansion and periodic, routine replacement of first-wallstructure are indicated in Fig. VII-3.

Most neutrons passing through the first-wall structurecan be absorbed in a liquid, pure-lithium blanket lessthan 2.0 m thick, and tritium breeding ratios greater than1.5 can be achieved with such thicknesses. However, theblanket will actually be thicker because of significantfractional vapor volume. Boiling is expected to be un-usual because energy is deposited volumetrically ratherthan transferred into the blanket through solid surfaces.

114

Because most theory and experiments concerned with or-dinary nucleate boiling are of limited applicability, ex-periments will probably be required to clarify this issue.

With liquid-lithium thicknesses of this order ofmagnitude, significant amounts of gamma energy,produced primarily by (n.y) reactions of neutrons withreactor structure and blanket fluids, and of neutronenergy can escape from the blanket. While deposition ofas much fusion energy as possible in the high-temperature zone may be desirable because lowertemperature heat can be used less efficiently, capture ofthe last few percent of escaping energy is generally moredifficult than capture of all the rest. On the other hand,greater gamma and neutron energy leakage means thatthicker reactor shielding and a greater capacity to coolthe shield will be required. Exploratory neutronicscalculations for 1.5-m-thick, 20-vol%-vapor, pure-lithium blankets indicate that gamma-energy leakage ismore than 10% of neutron wall loading.

In an attempt to retain more gamma and neutronenergy within the blanket without large increases inblanket thickness, we investigated the use of lithium-richboiling lead-lithium mixtures. Preliminary neutronicscalculations for 1.5-m-thick blankets indicate thatgamma- and neutron-energy leakage can be reduced sub-stantially, i.e., by a factor of about 2, that the total ofneutron kinetic plus interaction energy is increased, andthat the tritium breeding ratio is relatively insensitive tolead content up to large lead concentrations. Althoughthere is little compatibility experience for high-temperature materials with lead and much less for lead-lithium mixtures, all indications are that no particularproblems will be encountered. Lead-lithium alloys havelower vapor pressures than lithium and are projected tohave lower tritium solubilities, thereby further loweringcontainment structural stresses and blanket tritium in-ventories.

Conservative calculations, which assume that the onlysignificant resistance to tritium permeation through heatexchangers is that of bulk materials to diffusion, indicatethat the use of double-walled heat-transfer equipmentwith ceramic permeation barriers and removal of tritiumescaping through the first wall by a helium sweep gasbetween the two walls will limit tritium escape rates toacceptable levels, i.e., to 1.0 Ci or less per day. Thehelium sweep gas also aids heat transfer by providing aconductive mechanism operating in parallel with thermalradiation, to which helium is essentially transparent, andpermits ready adjustment of pressure differential loading

between inner and outer walls. Because readily fabricablemetals of acceptable cost creep excessively in thetemperature range of interest, the use of ceramics, glass-ceramics, and graphites for construction of primary heatexchangers will probably be necessary. Protection fromblanket fluid attack by coating or cladding suchmaterials with. e.g.. TZM, is necessary. On the process-fluid side, heat-transfer surfaces must be protected fromattack by oxidizing gas mixtures with oxides, e.g..alumina or silica, as structural materials or protectivelayers. Double-walled heat-transfer equipment construc-tion affords both protection against catastrophic heat-exchanger failure due to sudden rupture and the pos-sibility of detecting small leaks before they assumeserious proportions. Because the largest resistance toheiU transfer to gaseous process or power-cycle streamsis the gas film resistance, a large temperature-droppenalty is not paid for double-walled heat-exchangerconstruction in such cases.

If permeable windows are used for tritium recovery,tritium window performance is enhanced by highertemperatures and their simplicity is attractive. However,high-temperature distillation can also take advantage ofthe high temperatures provided by lithium boiler blanketsand offers other advantages, including the removal ofother volatile blanket impurities such as heliumgenerated by (n,a) reactions of blanket and structuralmaterials, oxygen, and nitrogen. Distillation for tritiumrecovery has been rejected in the past because at lowtemperatures (below ~123O K), the liquid is enriched indeuterium and protium relative to the vapor. At highertemperatures the opposite is true, with the atom ratio ofdeuterium in the vapor to deuterium in the liquid rising to10 at ~185O K, corresrionding roughly to an enrichmentby a factor of 10 for each equilibrium stage, so that onlya few stages would be required. Also, side-streamdrawoffs of only a few percent of the blanket boilup arerequired, heat can be recovered at temperatures only alittle below the blanket operating temperature inoverhead condensers, column shell construction wouldbe similar to blanket containment, and the device wouldbe self-pumping.

Preliminary indications are that thermionic diodedesigns developed for topping fossil-fuel-fired steampower plants12 can be readily adapted for mating withlithium boiler blankets to improve overall plant thermalconversion efficiencies or to provide dc power forelectrochemical processes such as the bismuth-oxidehybrid cycle described below.

115

Fusion-Fission Hybrid Reactor Studies (W. A. Reupke,H. S. Cullingford)

Interest in proliferation-resistant fuel cycles hasstimulated the development of hybrid fusion fission con-cepts based on the resource-efficient thorium fuel cycle.Previous studies have identified a high-power-densityhybrid concept that produces a large amount ofbyproduct electricity13 as well as fissile fuel. Our studieshave turned to quasi-symbiotic concepts that maximizefissile-fuel production and minimize heatproduction—so-called fissile-fuel factories. Such con-cepts shirt the burden of electric power production fromthe hybrid reactor to a host system of fission reactors.We have adopted the following conditions as our designbasis: (I) electrical self-sufficiency. (2) tritium self-sufficiency, and (3) maximum fissile-product output perunit of blanket heat.

Our studies of fissile-fuel factories are based on laser-driven ICF reactor concepts. The specific designs in-vestigated are spherical and use the wetted-wall reactorcavity, which is enclosed by spherical shells consisting ofbreeding materials, coolant, and structure. Liquid lithiumis used as the fertile material for breeding tritium and asthe reactor coolant. The fissile breeding material isthorium metal.

Neutronics calculations were performed with our one-dimensional SN code ONETRAN-DA. Multigroup crosssections and kerma factors were processed fromENDF/B-IV using the NJOY processing system to forma coupied 30/12 group neutron/gamma-ray set. Severalblanket compositions and configurations were in-vestigated. Calculated quantities of interest include thetritium breeding ratio RT, the fissile conversion ratio ornumber of fertile atoms converted per thermonuclearsource neutron RF. the blanket energy deposition perthermonuclear source neutron Q. and the ratio of fissile-equivalent thermal energy to blanket energy 180 RF/Q.A typical reactor geometry and blanket composition isgiven in Table VII-IV. Selected results of the ncutronicscalculations are presented in Table VJI-V. Only casesthat satisfy the requirement for tritium self-sufficiencyare included in Table VII-V. For Case 1, a 4-cm-thickberyllium neutron-multiplier region was included adja-cent to the reactor cavity and the allowable volume frac-tion of thorium, while retaining sufficient lithium coolantfor a tritium breeding ratio of 1.1, was determined. Thefissile breeding region for Case 2 includes the maximumvolume fraction of thorium consistent with requirements

CARBON

Outside Radius(cm)

199.9200.0201.0230.0232.0262.0282.0284.0292.0296.0

TABLE VII-IV

REFLECTED BLANKET

Composition'

VoidLi50% Li. 50% SSb

Li50% Li. 50% SS68.6%Th. 16.5% Li, 14.9% SSC50% SS, 50% LiLiSS

"Compositions in vol%.hSS siainless steel.

for structure and coolant. Sufficient beryllium neutronmultiplier (II cm) was included to obtain a tritiumbreeding ratio of 1.1. Case 3 is similar to Case 2 exceptthat the beryllium-region thickness was reduced to 4 cmand sufficient lithium was added between the berylliumand thorium regions to obtain & tritium breeding ratio of1.1. For Case 4, the beryllium was eliminated and thelithium-region thickness was increased sufficiently to ob-tain an adequate tritium breeding ratio. The effects of us-ing a carbon reflector outside the fissile breeding regionarc illustrated by Case 5 (sec also Table VIMV).

The most attractive designs are Cases 2 through 5.The use of a beryllium neutron multiplier results in smal-ler overall systems (Cases 2 and 3) than the design withsufficient lithium for adequate tritium breeding withoutthe neutron multiplier (Case 4). For Case 5, the thicknessof the fissile breeding region is reduced from 50 to 30 cm,but the fissile fuel production per thermonuclear neutronRF. is almost as large as for Case 4. Thus, the averagerate of production of fissile fuel per unit volume of fissileblanket is larger by ~70% for Case 5 than for Case 4.This may have important implications for optimumprocessing cycles.

116

TABLE Vll-V

SUMMARY OF BLANKET NEUTRONICS EVALUATION

Beryllium BreedingMultiplier Region

Case 'cm) I

BreedingRegion

RT RF Q(MeV) ,180 R,7Q

I234

4114 20 cm Li

29 cm Li

29 cm Li

50 cmthick

12 Th69 Th69 Th69 Th69 Th 30 cm100 C 20 cm

1.1

I.II.II.I

0.300.780.710.67

I.I 0.63

22352827

26

2.54.14.64.6

4.4

The minimum pellet gain requirement for electricalself-sufficiency can be estimated from an energy balance.

i.e..

P = f G £1 L 6 V "where P is the net electric power produced by the fuelfactory, f is the microexplosion repetition rate, G is thepellet gain (thermonuclear energy released/laser energyon target). EL is the electric energy input to the laser perpulse. T]L is efficiency of the laser system, r\B is the beamtransport efficiency. n,h is the thermal-to-electric conver-sion efficiency, and M ( is the blanket energy multiplica-tion factor. For electrical self-sufficiency, P = 0.Inserting estimates of n,L = 0.06, r\M = 0.95, T ^ = 0.33,and Ma = 2, we find G > 30. Thus, electrically self-sufficient fissile-fuel factories can operate with relaxedpellet gain requirements compared to pure commercialfusion for which G > 100.

The number of converter reactors with a thermalpower equivalent to the thermal power of the fissile-fuelfactory at 100% burnup is

Nequiv V

where RF is defined as above, n,, is the efficiency ofrecovery of fissile material from the blanket, En^/E^,, isthe ratio of energy release of fusion events to fissionevents, and C is the reactor conversion ratio. InsertingRF = 0.65, n.R = 0.90, EfjE,* = 180/17.6, M, = 2, andC = 0.9, we find N . ^ = 33.

We have coordinated the investigation of specificengineering features by means of a generic point design.Areas receiving our attention included the fusion driveror core, blanket geometry and construction, materialsselection, heat-removal considerations, and systemlayout.

The metallic thorium is contained in 0.38-mm-thickstainless steel cladding. The outside diameter of the fuelpins is 13 mm; they are supported in hexagonal modulararrays of 37 fuel pins each. The minimum pitch-to-diameter ratio of the fuel pins is 1.05. Each module iscooled by liquid lithium at an inlet temperature of 773 Kand an inlet velocity of 1 m/s. the maximum time-averaged thorium centerline temperature is 780 K, andthe time-averaged outlet temperature is 778 K. Com-pared to the requirements of net-electric-power produc-ing blankets, the heat-removal requirements of electrical-ly self sufficient fissile-fuel factories are moderate.

Figure VII-4 shows the layout of a generic system.Two square modules and eight hexagonal modules con-tain lithium-cooled thorium. The remaining four lithium-cooled square modules of the truncated octahedron arededicated to pellet injection (1), beam transport (2), andcavity exhaust (1). The entire assembly is enclosed in ashielded cell with remote manipulators for blanket-module assembly and disassembly.

Our studies have established that a CO2-laser-drivenlithium wetted-wall ICF fissile-fuel factory concept istechnically feasible. Ncutronics calculations have shownthat lithium prebreeding regions or beryllium multipliersoffer a high output of fissile material with minimum

117

Serviceoble cover

8IOOHI module(being inserted)Reactor enclosure »M

Reactor mompulotof

Lithium aietted firstMi l

Blanket cor* owembliee

Lithium COM lea. piping

Lithium hot leg piping

Reactor support structure

Supers mic spray

Lithium hot leg piping

Opposing loser beam bundle

Fig. ¥11-4.Low-power-density hybrid reactor design using lithium wetted-wall protection concept.

residual heat. Engineering studies have identified a con-sistent system of ICF core, conversion blanket, and heat-removal components.

INTEGRATED PLANT DESIGN STUDIES

Introduction

We are studying three commercial applications oflaser fusion: (1) electric-power generation, (2) synthetic-fuel (hydrogen) production, and (3) fissile-fuel produc-tion. Integrated plant systems for these applications aredescribed by the computer model TROFAN, in whichplant subsystems are described analytically, and impor-tant parameters may be varied for design tradeoff andoptimization studies.

We have previously defined preliminary electricgenerating plant concepts14 based on lithium wetted-walland magnetically protected wall reactor concepts. Ourcurrent efforts include updating TROFAN with moredefinitive analytical models of reactor concepts andelectric generating plant subsystems, and by incor-porating a capital cost data base, being developed undercontract by Burns and Roe, Inc., Oradell, New Jersey.

Our hydrogen production studies are coordinated witha parallel study, also supported by the Office of FusionEnergy, on thermochemicai hydrogen production using atandem-mirror reactor as a fusion energy source. Thisstudy is further suppotted at LASL by the Division ofEnergy Storage for experimental verification of ourbismuth-oxide electrothermochemical cycle. Our studiesinclude integration of the electrothermochemical processwith the high-temperature lithium boiler conceptdescribed above, and the production of hydrogen byhigh-temperature electrolysis.

Reported below are results from our studies ofelectric-power generation and hydrogen production.

Electric-Power Generation (E. M. Leonard, J. H.Pendergrass)

We are extensively modifying our laser fusion systemscode TROFAN. The modifications involve improve-ments in our criteria for determining reactor cavity sizesand structural requirements. Cavity sizing is based on al-lowable temperature rises in cavity structures, allowablerates of material loss from surfaces by evaporation and

118

sputtering, and times required for restoration of cavityconditions to those necessary for initiation of the nextpellet microexplosion. Cavity structural requirements arebased on calculated loads. The loads considered are pel-let debris impact, blast-wave reflections, evaporation andsputtering recoil, blanket thermal expansion, static pres-sures, and thermal stresses.

TROFAN is also being restructured to be compatiblewith a capital cost account structure for laser fusion,which we have developed, and with a capital cost andcost-scaling-law data base being developed for us byBurns and Roe, Inc. The cost data base and cost-scalinglaws are for laser fusion electric-power generating sta-tions incorporating the lithium wetted-wall andmagnetically protected wall reactor concepts. Weadapted a list of cost accounts for laser fusion powerplants from a list of capital-cost accounts, which wasdeveloped for magnetic confinement fusion for uniformuse throughout the magnetic-confinement fusion com-munity to report capital-cost estimates. This cost ac-count system is a modification of earlier fossil-fuel andfission power-plant capital-cost account lists.15'1* Themodifications involved the addition of accounts to coverfeatures peculiar to laser fusion and the elimination of ac-counts that cover items peculiar to magnetic confinementfusion.

Costs are expressed in 1978 dollars for a maturefusion-power-plant construction industry. While costsrepresentative of the time frame of a mature fusion

economy are desirable, uncertainties associated with theprojection of costs so far into the future are too great tojustify such attempts and they were not made. However,other advanced technology is subject to similar cost es-calation, and cost comparisons made in terms of currentdollars are reasonable, especially if periodically updated.Wherever the development of comp'etely new technologyis involved and technological barriers to construction offacilities or manufacture of equipment with requiredcharacteristics at acceptable costs based on present in-dustrial capabilities were perceived, an effort was madeto determine whether the technological barriers are insur-mountable (e.g.. due to obvious materials limitations) orcan be eliminated (e.g.. by constructing manufacturingfacilities of increased capacity).

Design tradeoff and parameter studies, for which thecost data base will be used, require cost-scaling informa-tion, rather than cost estimates of point designs, overwide ranges of principal plant parameter values, some ofwhich are listed in Table VII-VI. The cost-scaling equa-tions were developed to involve as few parameters aspossible consistent with high accuracy for futuristictechnology at the conceptual stage of development toprovide compatibility with a systems code of manageableproportions.

Initially, we hoped that for each reactor concept, asingle cost-scaling law could be developed for each plantcomponent or system, which would cover the entire plantparameter space of interest. Such laws would involve

TABLE VII-VI

RANGES OF SOME PRINCIPAL ELECTRIC POWER PLANTPARAMETERS COVERED BY THE COST DATA BASE

Wetted-WallReactors

Magnetically ProtectedWall Reactors

Number of reactor cavities 1-20Cavity thermal powers (MW) 150-3000Laser amplifier pulse

repetition rate (Hz) 1Laser amplifier pulse

total energy, (MJ) 1-5Nominal plant net

electric-power output (MWe) 1000-2000Reactor materials of Stainless steels and

construction refractory alloys

1-4750-6000

10

1-5

1000-2000Stainless steels and

refractory alloys

119

scaling relative to a single-point plant design for eachreactor concept. However, large changes in plantcapacity and/or in the number of reactor cavities withinthe ranges of interest necessitated qualitative as well asquantitative changes in plant layout, plant equipment,and facility conceptual design. In addition, the ranges ofparameter values for which accepted scaling laws or newscaling laws developed during the present study are ac-curate about point designs are quite limited. Therefore,more than one conceptual point design, plus scalingabout the point designs, for the two reactor conceptsproved necessary for accurate coverage of the entireparameter space of interest. The point conceptualdesigns upon which the scaling laws are based are (1) a3000-M Wt plant with 20 wetted-wall reactor cavities, (2)a 3000-MWt plant with 4 wetted-wall reactor cavities,(3) a 3000-MWt plant with 1 wetted-wall reactor cavity,(4) a 5600-MWt plant with 4 magnetically protected wallreactor cavities, and (5) a 5600-MWt plant with 1magnetically protected wall reactor cavity.

Different approaches were used to develop the costdata for various plant facilities, components, andsystems. For some, especially those peculiar to laser fu-sion and involving technology that is only in the concep-tual stage, more detailed conceptual designs weredeveloped and basic unit material, fabrication, assembly,and installation costs were estimated. For example, raw-materials costs for large refractory-metal reactor vesselswere obtained from suppliers, and fabrication, assembly,and installation costs were based on information fromvarious manufacturers, reviews of published information,and Burns and Roe in-house expertise. Fabrication costsfor such vessels include assembly (primarily by weldingusing advanced technology such as electron beams orlasers) of factory-constructed subassemblies in the reac-tor service building or wing, and final assembly in reac-tor cells. Because the same facilities and equipment arerequired for scheduled maintenance and repair ofdamaged equipment, their cost is included in amaintenance equipment account.

In other cases, e.g., for such relatively conventionalplant equipment as turbine generators and steamgenerators, design and cost information was obtainedfrom manufacturers. Designs for some major plantsystems, e.g., ultimate heat rejection, service coolingwater, chilled water, emergency tritium removal,radwaste, and building service systems, which arerelatively conventional and insensitive to nuclear islandcharacteristics, were adapted from existing facilities anddesigns, such as 1000-MWe light-water reactor and

breeder-reactor power plants, and costs for such systemswere appropriately adjusted.

The scope of the data base includes two steam cycles,one subcritical with maximum temperature of 730 K andpressure of 2270 psia, and the other, an 810 K, 3500-psia supercritical cycle. Cost data were developed foronly a single type of bred-tritium recovery system,refractory-metal tritium windows, and a single type ofultimate heat-rejection system—dry cooling towers.

Synthetic-Fuel (Hydrogen) Production (J. H.Pendergrass)

Based on a preliminary investigation of applications offusion energy to hydrogen production by ther-mochemical water splitting," we conclude that tritiummust be bred in the high-temperature blanket zone if atritium breeding ratio BR > 1 and a fraction of high-temperature thermal energy >0.5 is to be attained. In thisstudy17 of a provisional blanket concept, tritium breedingwas accomplished in a low-temperature zone, and high-temperature heat was generated in a separate graphitezone. The high-temperature heat was transferred to thethermochemical process via a helium coolant loop.

The constraints that all tritium breeding be ac-complished in a low-temperature blanket zone and thatBR >1 result in the conversion of only 33.8% of fusionreactor energy release to stored chemical energy ofhydrogen. In addition, 8.7% of total energy release is ex-ported as electric power if low-temperature heat orelectric power generated from this heat cannot be used todrive the thermochemical cycle.

Although the provisional blanket concept does notpermit a good match between fusion-energy deliverycharacteristics and thermochemical-cycle energy require-ments, indications are that the combination is com-petitive with other nuclear-energy-driven synthetic-fuelproduction concepts. Furthermore, the high-temperaturelithium-boiler concept described above offers thepossibility of a better match between energy deliverycharacteristics and energy requirements.

A fusion-energy synthetic-fuels econometric model17'18

was developed as part of this study to facilitate examina-tion of the tradeoffs arising from considerations oftritium breeding, electric-power production, andhydrogen production. Important conclusions were

• The lowest hydrogen production costs are achievedwith plants that do not export electricity.

120

• Hydrogen production cost depends onthermochemical-cycle efficiency in a roughly hyper-bolic fashion, so that high efficiency in the ther-mochemical portion of the plant is essential for low-cost production.

• Hydrogen production cost increases approximatelylinearly with capital costs of fusion-energy source,electric generating plant, and thermochemicai plant,with the first projected to be the highest.

• If other chemical fuels are available at low cost, or ifelectricity can be freely substituted for hydrogenenergy, a selling price for hydrogen commensuratewith its energy content alone would make electric-power production more attractive in an infinitemarket situation. However, the first circumstance isnot projected to continue for long; free substitutionof electricity for hydrogen as a chemical feedstock isnot possible; and substitution for many hydrogenfuel uses may not be economic.

• Despite inevitable uncertainties, comparison withpublished hydrogen production cost estimates usinglight-water fission reactors (LWRs), high-temperature fission reactors (HTRs), and liquid-metal-cooled fast breeder reactors (LMFBRs) asenergy sources, as well as conventional electrolysis,advanced electrolysis, or thermochemicai cycles forwater-splitting suggests that fusion-driving of ourreference cycle will be competitive, with HTRs com-bined with a pure thermochemicai cycle offering themost formidable competition.

Our reference thermochemicai cycle," illustratedschematically in Fig. VII-5, consists of six basic steps:(1) production of hydrogen (H2O) and aqueous sulfuricacid (H2SO4) by electrolysis of aqueous sulfur dioxide(H2SO3) at low temperature (~300 K); (2) reaction ofsolid bismuth oxide (Bi2O3) with aqueous sulfuric acid toform bismuth oxide sulfate (Bi2O3-3 SO3) and liberationof water (H2O) at low temperature (~300 K); (3) thermaldecomposition of bismuth oxide sulfate to reform thebismuth oxide and release sulfur trioxide (SO3) at highertemperature (maximum, ~125O K); (4) thermal decom-position of sulfur trioxide (SO3) to sulfur dioxide and ox-ygen at highest cycle temperature (maximum, 1450 K);(5) rapid quenching of the dissociated gas mixture to pre-vent back reaction, followed by separation of sulfur diox-ide and oxj'gen; and (6) formation of sulfurous acid bydissolution of the sulfur dioxide in water to close the cy-cle.

A preliminary conceptual process design wasdeveloped for our reference cycle to permit preliminaryestimation of thermal efficiency and future hydrogenproduction cost if this cycle continues to be promising.Except for the electrolyzers, the low-temperature opera-tions are relatively conventional. Similarly, many of thehigh-temperature steps of the cycle also involve standardprocess technology, with a few important exceptions.Inadequate thermodynamic, kinetic, solids-characteris-tics, and materials-compatibility data made necessary anumber of assumptions and approximations, which maynot be very accurate when experimental data from theconcurrent experimental program at LASL becomeavailable.

Although the reference cycle still appears promising,the process design exercise has revealed a few shortcom-ings, but lack of data makes assessment of theseriousness of these shortcomings impossible. Alter-natives, described in Table VII-VII, within the bismuthoxide cycle family, in which fewer than three sulphurtrioxide groups are added to and removed from thebismuth oxide molecule, permit avoidance of some ofthese potential problems at a cost primarily of increasedsolids circulation. Process designs for these alternativecycles are partially completed.

ENGINEERING DEVELOPMENT STUDIES

Introduction

Our engineering development studies include the as-sessment of advanced technologies unique to long-life,high-repetition-rate ICF subsystems, and the analysisand development of conceptual subsystems for inclusionin our integrated plant systems studies. These studies in-clude methods of increasing the overall efficiencies ofCO2 laser systems, laser gas-handling and coolingsystems, pulsed-power systems, pellet injection/trackingand beam firing synchronization, beam transport, andpellet fabrication.

Reported below are results from our studies on in-creased efficiency of energy extraction in CO2 lasers andon lifetimes of cryogenic pellets in a reactor environment,and from a study under contract to United TechnologiesResearch Center (East Hartford, Connecticut) on pelletinjection, tracking, and laser-beam synchronization.

121

D.C.ELECTRICAL

H,0

JoENEIWY I

HYDROGEN GENERATOR(ELECTROLrZER)

soz

BISMUTH OXIDEREACTION

Bi2O33SO3

•ft

BISMUTH SULFATE

DECOMPOSITIONBi20j-3S03 B i jOj^Svj ,

0XY6ENRECOVERY

1 SEPARATION WORKTHERMAL EQUIVALENT

SULFUR TRIOXIDEDECOMPOSITION

L.

so3S 0 2 , 0 2

THERMALENERGY

Fig. Vll-S.Bismuth-sulfate/sulfuric-acid hybrid thermochemical hydrogen cycle.

TABLE VH-VII

REFERENCE ELECTROTHERMOCHEMICAL BISMUTH OXIDE SULFATE CYCLEAND ALTERNATIVES WITHIN THE SAME FAMILY"

Alternative Reactions InvolvingNumber Bismuth Oxide and Bismuth Oxide Sulfates

1 Bi,O, 3SO,(s)—Bi2O,(l)+3SO,(g)Bi;O,(s)+3H2SO4(aq)-»Bi2O, 3SO3(s)+3H2O(l)

2 Bi,O, 3SO,(s)-»Bi,O, SO,(?)+2SO,(g)BijO, SO<(s)+2H3SO4(aqH-Bi,O, 3SOj(s)+2H,O(l)

3 BijO, 3SO,(s)-»Bi2O, 2SO,(s)+SO,(g)BUO, 2SO,(s)+H,SO4(aq)—BijOj 3SO,(s)+HjO(0

4 BijO, 2SO,(s)—Bi2O3(l)+2SOj(g)Bi,O,(s)+2HjSO4(aq)-*Bi;O, 2SO,(s)+2H2O(aq)

5 Bi,O., 2SO,(s)-Bi,O, SO,(?)+SO,(g)Bi2O, SO,(s)+H,SO<(aq)-.Bi2O1 2SO,(s)+H!O(l)

6 Bi,O, SO,—Bi;O,(l)+SO,(g)BUO,(s)+H,SO,(aq)->Bi2O1 SO,(s)+H,O( I)

Sulfuric Acid Required SolidsConcentration Circulation Rate

Required Tor Sulfur (mols Solids Molten Salts MustTrioxide Addition Circulated per be Handled at

Step (wt%) mol H ; Produced) High Temperature

53-59

53-59

53-59

5-20

5-20

5

1/3

2/3

1

2/3

1

1

yes

t

no

yes

yes

Energy Extraction with Long Pumping Discharges (H.C. Volkin)

Pulsed-power systems tend to be cheaper and morereliable as the required pulse duration increases and thedelivered voltage decreases. This consideration favorslong discharge times in pulsed electrical excitation of gas

lasers and is compatible with multiple optical-pulseenergy extraction in CO2 laser amplifiers. In multipulselaser operation, a sequence of two or more light pulsespasses through the amplifier in the time interval betweenthe start of successive electrical discharges. Between op-tical pulses, amplifier gain is restored by vibrational

122

relaxation and (within the discharge time) by further dis-charge pumping. Laser efficiency can be increasedbecause vibrational energy stored in the nitrogen of thelasing gas mixture and in the higher levels of the CO2

asymmetric-stretch (AS) mode can be extracted by laterpulses and because less excitation energy in the AS modeis lost by intramolecular transfer to the other modes ofCO2. The gain recovery time is the time required after anoptical pulse has traversed the amplifier for the gain toreach its maximum value in the absence of any externalpumping. The rate parameters for vibrational energytransfer are temperature-dependent. Hence, gainrecovery time varies somewhat with the translationaltemperature and the modal vibrational temperatures.

A given gas mixture has an optimum ratio of electricfield E to molecular number density N for efficient dis-charge pumping. The pressure-scaling law for electricalexcitation kinetics can be stated as follows. With a givengas mix and value of E/N, if energy deposition per unitvolume Wo and population inversion density AN0 occurat the pressure Po in a time To with the current density Jo,then at pressure P in the time (PJP)TO with the currentdensity (P/P0)J0, the energy deposition is (P/P0)W0 andthe inversion density is (P/PO)ANO. Because an opticalcross section at pressure P differs by the factor (Po/P)from that at Po, the saturation energy at P differs by thefactor P/Po, but the small-signal gain coefficient is thesame. Under this pressure scaling, the efficiency of con-verting electrical energy to energy stored in the inversionremains constant. A known discharge efficiency at somehigher pressure carries over to a lower pressure with theindicated increase in discharge time, along with theproportionate decrease in E and J. However, the smallerinversion density obtained at the lower pressure neces-sitates a correspondingly larger volume of laser gas toproduce a given laser output energy.

To assess the laser-system tradeoffs involved in theseconsiderations, parameter studies were made to deter-mine laser efficiencies, multipulse sequencing, and opticalpulse energies associated with long discharge times. In allcases the fill pressure and temperature were 760 torr and300 K; the discharge time was ~4 us, and the first op-tical pulse occurred when the gain coefficient reached2%-cm"1. We assumed that energy extraction occurredonly on the 10.6-um band, but that a!1 the availableenergy was extracted by each pulse. For each gas mix,current densities of 6,4, and 2 A cm"2 were considered,and three time periods used between successive opticalpulses were 0.2, 0.4, and 0.8 us. With the longer inter-pulse periods, fewer pulses are obtained with generally

higher average energies. Smaller current densities alsolend themselves to fewer optical pulses, because the gaincoefficient reaches 2%-cm"1 later in the discharge andless energy is deposited in the gas. Of four gas mixes in-vestigated, the mixture 3:l:l::He:N2:CO2 gave good ef-ficiency with a 0.5-us discharge lime and 20 A cm"2 at1800 torr. In addition, three helium-free N2:CO2 mixes,namely. 0:1:4. 0:1:2, and 0:1:1, were examined toprovide data for tradeoffs involving the cost reduction in-herent in eliminating helium from the system.

We define a single-pulse efficiency for a given case asthe efficiency obtained under the following circum-stances: the discharge continues until the gain reachesthe peak value that can occur under the prescribed con-ditions, at which point all the energy available on the10.6-um band is extracted by the single optical pulse.The single-pulse efficiencies for the various gas mixes are3.71% (3:1:1), 3.05% (0:1:1), 3.58% (0:1:2), and 3.77%(0:1:4). Typical gain recovery times are shown in TableVII-VIJI.

If no restriction is placed on the number of pulses, the3:1:1 mix is again markedly superior in efficiency to thehelium-free mixes. Limiting the number of pulses by therequirement that the energy of the final pulse be abouthalf that of the initial pulse, we find that efficienciesaround 22% are attainable with the 3:1:1 mix in 12 to 14pulses with an interpulse period of 0.4 or 0.2 us. Thiscorresponds to a sixfold increase over single-pulse ef-ficiency. By comparison, efficiencies around 16% are ob-tained with 0:1:1 in 17 pulses (0.2 us period), and with0:1:2 in 16 pulses (0.2-us period). In all these cases, thedischarge current was 4 A cm"2; the efficiency and pulseenergy (J/fatm) at various pulse times are shown inTable VIMX.

TABLE VII-VIII

GAIN RECOVERY TIMES FOR GASMIXTURES AT VARIOUS TEMPERATURES

He:N,:CO, MixGain Recovery Time

(Us)

3:1:1 0.83 (312 K) 0.71 (330 K)0:1:1 0.27(301 K.) 0.19 (315 K)0:1:2 0.24(301 K) 0.18 (318 K)0:1:4 O.17(3O4K)O.12(322K)

123

TABLE VIMX

SOME EFFICIENT MULTIPLE-PULSE ENERGY EXTRACTION SEQUENCES"

Pulse Number (in Parentheses), Efficiency (%), and PulseEnergy (J/( atm) for He:N2:CO2 Mix

3:1:1

Discharge Current (A cm"2) 4Energy Deposition (J/f-atm) 92.4Interpulse Period (us) 0.2

(1) 5.04,1.99(2) 8.26,1.64(3)10.6,1.50(5)13.9,1.39(7)16.2,1.34(9)17.8,1.31

(11)190,1.29(12)195,1.28(13)203,1.14

(14)215,0.98(15)223,0.83

3:1:1

492.40.4

(1) 5.04,1.99

(2) 8.15,1.94(3)10.4,1.98(4)12.0,2.02(5)13.4,2.06(6)14.5,2.08

(7)15.4,2.11

(8)17.4,1.85(9)19.1,1.54

(12)224,0.83

0:1:1

4162

0.2

(1) 5.12,2.08(2) 8.24,1.93(4)11.7,1.76(6)13.4,1.62(8)14.4,1.51

(10)143,1.42(12)15.1,1.33(14)15.1,1.26(16)15.1,1.18

(17)15.7,0.95(18)16.1,0.72

0:1:2

4134

0.2

(1) 6.04,2.27(2) 9.45,1.93(3)11.5,1.72(5)13.7,1.45(7)14.7,1.30(9)15.2,1.20

(11)154,1.13(13)155,1.07(15)154,1.02

(16)160,0.73

"Efficiency and pulse energy are given for selected optical pulses in a sequence, where the pulse number is shown inparentheses. Pulses in the discharge time occur between the double horizontal lines, and their efficiencies refer tothe cumulative electrical-looptical efficiency up to the time of the pulse.

At 9 or 10 optical pulses, efficiencies of 19 to 20% canbe attained with the 3:1:1 mix by using an interpulseperiod of 0.4 us and a discharge current of 4 A cm"2.With helium-free mixes, an efficiency of 15.5 to 16% isfound with the 0:1:1 mix using a period of 0.2 us and 2Acm~2 , and 13.5 to 14% is found with the 0:1:2 mix us-ing 0.4 us and 4 A cm"2. If four or five optical pulses aredesired, then 2 A-cm"2 and interpulse periods of either0.4 or 0.8 us give efficiencies of 14 to 16.5% with 3:1:1,and 11 to 12% with either 0:1:1 or 0:1:2 mixes.

Cryogenic Pellet Lifetimes and Temperatures (J. J.Devaney)

Lifetimes of several typical cryogenic pellets werecalculated for several customary environments insideand outside laser fusion reaction chambers. Optimum

cooling and resulting temperatures of multishell pelletswere computed. Pellets of 50:50 at.% solid D:T generallycontain 31 to 46% of the DT in the form D2, so that pel-let destruction commences when the D2 , which has thelowest triple point, begins to melt. The triple point of D2

is 18.71 K. Heet mechanisms affecting the pellet are in-ternal tritium radioactivity, external radiation into thepellet, forced convection, and conduction into the pellet.

Although the beta radioactivity of tritium is exceeding-ly weak, with a half-life of 12.26 yr and a mean beta par-ticle energy of only 5.7 keV, attempting to keep a tritium-containing pellet at very low temperatures—where theefficiencies of radiative and conductive cooling as well asthe heat capacities are extremely low—is difficult and insome cases impossible. Radiation, of course, varies asabsolute temperature to the fourth power; specific heatsof crystalline (Debye) solids drop as temperature to thethird or higher power at low temperatures; below 4 K the

124

conduction-electron contributions to the specific heat aresignificant and are proportional to temperature; thespecific heats of many plastics are roughly proportionalto temperature over a wide low-temperature range; andlast, low-temperature thermal conductivities all showdecreases with decreasing temperature, very roughlydirectly proportional to the temperature. In fact,radiation-cooling has dropped to such a low level around4 to 20 K that typical pellet designs containing ap-preciable amounts of solid DT within a vacuum cannotexist at all. Steady-state DT temperatures calculated forsuch pellets at absolute zsro in vacuum range from 88 to114 K, well above the D2 triple point.

Such pellets, therefore, need direct cooling of innerDT-containing structures and have lifetimes away fromsuch cooling of typically 26 to 80 s if initially at 10 K,and of 30 to 95 s if initially at 4.2 K. However, in mul-tishell pellets, even a low-density interstitial gas such ashydrogen or helium is very effective in cooling inner shellstructures. Ironically, too low a coolant temperature willcondense too much of the conductive gas, and so allowinner DT structures to melt. For example, one of our pel-let designs will survive only if the outer shell structure iskept between 7.9 and 18.2 K. In this design the minimuminner DT temperature of 9.4 K is achieved for an outercooling bath of 9.0 K. All other coolant temperatureslead to a higher inner DT temperature.

The high external temperatures of a laser fusion reac-tor cavity ftirther restrict the lifetime of a pellet.However, the pellet designer has considerable controlover the melting of restrictive parts, i.e., solid DT, byrather modest pellet design changes. For example, byaluminizing exposed plastic parts, he can reduce the ac-quisition of radiative energy, or by using low-conductivity materials, he can restrict the heat flow;these changes can account for extensive lifetime changes.We examined the survival of laser fusion pellets in twoextreme reaction-chamber atmospheres, marking the up-per and lower bounds of expected chamber lifetimes forthe pellets investigated, namely, 1017 atom/cm3 of argonat 773 K and a vacuum at the same temperature. A highpellet injection speed of 13 700 cm/s was assumed. Wefound that Rayleigh-Benard instabilities do not developand that all relevant flows are laminar. Our worst case, apellet moving at 13 700 cm/s in 1017 atoms/cm3 at 773K with sensitive areas exposed and with an initialtemperature of 10 K, will last 0.048 s. A protected pellet(aluminized, baffled—best case for this general design)will live 64 s in a 773 K vacuum if initially at 10 K.

Pellet Injection and Tracking (J. J. Devaney)

A preliminary investigation of potential pellet andlaser beam space-time interaction systems for inertialconfinement fusion reactors, with emphasis on CO2

laser-driven reactors, was completed under contract toUnited Technologies Research Center.20

We examined the synchronization requirements;velocity requirements; accelerators, includinggravitational, mechanical, pneumatic, and electrical;tracking, including accuracy required, degrees offreedom, measurement resolution, reference framestability, and some merits of tracking both within andwithout the chamber; guidance, including trajectory cor-rection required, types of guide tubes, and electrostaticguidance; beam pointing, including active-optics techni-ques and capabilities and nonlinear phase conjugation;and pellet velocity tradeoffs. Three systems that mayhave merit were identified: (1) adaptive laser opticsoperated in a predictive mode, with pellet tracking insideand outside the reactor; (2) feedback-controlled elec-trostatic guidance of the pellet, with high-precision track-ing and guidance external to the reactor; and (3) a com-bined adaptive optics-electrical guidance system, withrelaxed operating parameters for both subsystems. Theprimary critical issue was identified as the ability tomake the position and velocity tracking measurementswith the required speed and accuracy. Other criticalissues include reference frame motion, injector accuracy,and injector reliability.

Based on the information developed in this study,system recommendations shown in Table VII-X aremade for 100-m/s injection velocities. Such highvelocities are needed if high repetition rates are desired.The preferred system uses adaptive optics to point andfocus the laser beams on the pellet. Due to insufficientslew rate, the adaptive optics must operate in the predic-tive mode, whereby the pellet trajectory at the focus ispredicted based on precise tracking data, and the laserbeams are aimed and focused before the arrival of thepellet at the focal region. Errors between tracking andadaptive-optics reference frames (due to, e.g., noise andvibration) are assumed to be less than errors in trajectorymeasurement. In practice, this implies active trackingbetween reference frames and precise autoalignment ofall beams. Guidance in this system is provided only bythe rough constraint of inherent angular scatter of thepellet accelerator imposed by a mechanical guide tube.The accelerator is a pneumatic gas gun, and pellets are

125

TABLE VII-X

PELLET/LASER BEAM SPACE/TIME SYSTEM RECOMMENDATIONS(Pellet Velocity, 100 m/s)

BeamPointing

Mode

Guidance

Tracking

ActivePredictiveGuide

tubeInside &

outside

Recommended

NonePredictiveElectro-static

Outside

ActivePredictiveElectro-static

Inside &outside

Not Recommended

ActiveActive

NoneInside &

outside

Phaseconjug.

PassiveGuide

tube

Inside

Propulsion Pneumatic Pneumatic Pneumatic Pneumatic Pneumatic

tracked both outside and inside the reactor for adaptiveoptics control and synchronization of laser fire. Someautoalignment and initial trajectory measurement re-quirements are relaxed if final tracking can be performedinside the reactor through each beam line using sharedapertures.

An alternative system with fixed optics relies onelectrostatic guidance of electrically charged pellets toplace pellets on the correct trajectory to pass through thefocal region. The tracking requirements on this systemare very severe and must be sufficiently precise to func-tion solely on measurements outside the reactor.

A hybrid system that combines adaptive optics andelectrostatic guidance is also possible. This approachrelaxes performance requirements on both systems, butat the cost of increased system complexity.

The above system recommendations are based on theassumption that the injection velocity is 100 m/s. If thereactor could be designed so that the injection velocitywere significantly reduced, then the above conclusionswould be altered. Such a design would change the meansof (1) propulsion, (2) tracking, and (3) pellet irradiation.

If the injection velocity were 25 m/s or less,gravitational acceleration is a reasonable and preferredmeans of target propulsion. (A velocity of 25 m/s re-quires a drop tower of only 32 m in height.) Pellet motionduring a pulse round trip through the laser system (whichtakes 2 us) would be 50 urn or less so that phase con-jugation—passive focusing by nonlinear optics—is notruled out on the basis of pellet motion.

If the injection velocity were ~10 m/s or less, the re-quired mirror slew rates are low enough so that activelytracking adaptive optics could be used for pellettrajectory-beam direction relative angles of less than10°. This case would apply to each beam in the polar il-lumination case when the pellet is injected along thebeam-direction axis. Autoalignment system requirementswould again be eased by shared aperture tracking on allbeams, but in addition, low pellet velocity wouldeliminate the need to predict pellet position at the focusand should produce more accurate pellet irradiation.

Pellet tracking and trajectory measurement require-ments are greatly eased by lower velocities. The time in-terval during which position determination must be madeincreases and the accuracy with which the velocity mustbe measured may decrease. This last factor increases thefeasibility of a system that does not include tracking in-side the reactor for synchronization. At lower velocity,the voltage or the length of the guide-plate region re-quired to electrostatically correct trajectory errors, isalso less.

In summary, many benefits in reducing pellet injectionsystem complexity and cost would be derived from lowerpellet injection velocities. Furthermore, the nature of thepreferred system is a strong function of pellet velocity.

Finally, this preliminary study should be interpreted asa first exploration of conventional state-of-the-art techni-ques, and not be considered a definitive review of all pos-sibilities including highly innovative techniques.

126

REFERENCES

1. L. A. Booth, Comp., "Central Station PowerGeneration by Laser-Driven Fusion," Los AlamosScientific Laboratory report LA-4858-MS, Vol. 1(February 1972).

2. J. J. Devaney, "Magnetically Protected First Wallfor a Laser-Induced Thermonuclear Reaction," LosAlamos Scientific Laboratory report LA-5699-MS(August 1974).

3. "Inertial Fusion Program, July 1—December 31,1978," Los Alamos Scientific Laboratory reportLA-7587-PR (May 1980).

4. R. W. Conn et al., "Solase, A Conceptual Laser Fu-sion Reactor Design," University of Wisconsinreport UWFDM-220 (December 1977).

5. I. O. Bohachevsky and J. F. Hafer, "SputteringErosion of Fusion Reactor Cavity Walls," LosAlamos Scientific Laboratory report LA-6633-MS(December 1976).

6. I. O. Bohachevsky and J. F. Hafer, "Dependence ofSputtering Erosion on Fuel Pellet Characteristics,"Los Alamos Scientific Laboratory report LA-6991-MS (November 1977).

7. "Laser Fusion Program at LASL, January 1—June30 1977," Los Alamos Scientific Laboratory reportLA-6982-PR (April 1978), p. 135.

8. S. Chow and I. O. Bohachevsky, "MHD Decelera-tion of Fusion Reaction Products," Los AlamosScientific Laboratory report in preparation.

9. J. C. Goldstein, I. O. Bohachevsky, and D. O.Dickman, "Ion Motion in Laser Fusion ReactorStudies," Paper 9P7, Bull. Am. Phys. Soc, Series 1121 (October 1976).

10. G. A. DeMastry, "Corrosion Studies of Tungsten,Molybdenum, and Rhenium, in Lithium," Nucl. Ap-pl. 3, 127 (1967).

11. Y. Amagi, Y. Nishimura, and S. Gomi, "HollowCarbon Microspheres from Pitch Material and theirApplications," Materials '71, 16th National Symp.

and Exhibit. Soc. of Aerospace Material andProcess Engineers, Anaheim, California (April 21-23. 1971), p. 315.

12. G. V. Fitzpatrick and E. J. Britt, "ThermionicCentral Station Power—The THX Approach—Topical Report," Rasor Associates, Inc., reportNSR 2-6 (February 1977).

13. T. G. Frank. "A Thorium-Uranium Cycle ICFHybrid Concept," Proc. 3rd Topical Meeting on theTechnology of Controlled Nuclear Fusion, SantaFe, New Mexico (May 1978).

14. L. A. Booth and T. G. Frank, "Commeccial Ap-plications of Inertial Confinement Fusion," LosAlamos Scientific Laboratory report LA-6838-MS(May 1977).

15. "Guide for Economic Evaluation of Nuclear PowerPlant Designs," NUS Corporation report NUS-S31(January 1965).

16. United Engineers and Constructors, Inc., "1000-MW(e) Central Station Power Plants—InvestmentCost Study," WASH-1230, Vols. I-IV (June 1972).

17. L. A. Booth, M. G. Bowman, G. E. Cort, K. E. Cox,D. J. Dudziak, R. A. Krakowski, J. H. Pendergrass.and A. S. Tar, "Production of Electrother-mochemical Hydrogen using a Fusion Source ofHigh-Temperature Process Heat," Proc. 3rd ANSTopical Meeting on the Tech. of Controlled NuclearFusion (to be published).

18. A. Tai and R. Krakowski, "Generalized Energy—Balance and Economic Considerations for Ther-mochemical Hydrogen Production from FusionReactor Blankets," Trans. Am. Nucl. Soc. WinterMeeting, Washington, DC (1978).

19. K. E. Cox, "Thermochemical Processes forHydrogen Production," Los Alamos ScientificLaboratory report LA-6970-PR (October 1977).

20. R. G. Tomlinson et al., "Pellet and Laser BeamSpace-Time Interaction System Study," UTRCR78-954373-1, Los Alamos Scientific LaboratoryContract LP8-2822E (November 1978).

127

VIII. RESOURCES, FACILITIES, AND OPERATIONAL SAFETY

The design and construction of HEGLF facilities continued. Safety policies andprocedures continued to be successful in minimizing the hazards of operating high-energy lasers.

MANPOWER DISTRIBUTION

The distribution of employees assigned to the variouscategories of the DOE-supported Laser Fusion ResearchProgram is shown below.

APPROXIMATE STAFFINGLEVEL OF LASER PROGRAM

December 31, 1978

Tasks

CO, Laser DevelopmentCO, Laser ExperimentsTarget DesignTarget FabricationDiagnostics DevelopmentSystems and Applications StudiesAdvanced TechnologyWeapons Application

TOTAL

DirectEmployees

77115274140

839

320

High-Energy Gas Laser Facility (HEGLF)

Construction Package I. including the laser building,the mechanical building, the office building, and awarehouse is 40% complete, with a scheduled comple-tion date of September 1979.

Construction Package II, including the target building,the power-transmission system, and miscellaneous con-

struction is 42% complete with a scheduled completiondate of August 1979.

In the interest of continuity, we have presented detailsin HEGLF Design and Construction in Sec. II.

New Laboratories

The contract for the High-Voltage and Optical-Evaluation Laboratories was awarded in January 1978;construction is essentially complete, with a scheduled oc-cupation date of January 22, 1979.

OPERATIONAL SAFETY

General

Laser Fusion Research Program activities have nevercaused biological damage to any employee from laserradiation. The excellent lost-time injury rate was alsomaintained, despite increased activities and additionalstaff.

Laser Protective Eyewear Movie

A 16-mm color movie, "Lasers and Your Eyes," wasproduced for indoctrination of new employees and foruse by industry and Academe to demonstrate the poten-tial ocular hazards of lasers and to share informationabout available protective eyewear developed in theLaser Fusion Program for lightweight corrective spec-tacles from a series of colored filter glass.

128

IX. PATENTS, PUBLICATIONS, AND PRESENTATIONS

PATENTS

Serial numbers, filing dates, patent numbers, and issuedates may not be shown for several months after they areassigned. Therefore, for any given period, cases may bemissing from these listings if patent activity occurred latein the reporting period.S. N. 928,027— "Laser Target Fabrication, Struc-

ture and Method for ItsFabrication," E. H. Farnum and R.J. Fries, filed July 25, 1978.

S. N. 942,227— "Enhancement of Laser PulseContrast Ratios Via TransientResponse of Narrow Band Reso-nant Interferometers," R. A. Fisherand B. J. Feldman. Hied September14, 1978.

S. N. 945.376— "Fiber Optic Solid State Switch,"M. E. Thuot, filed September 25,1978.

S. N. 951.203— "Vacuum Aperture Isolator forRetroreflection from Laser-Irradiated Targets," R. F. Benjaminand K. B. Mitchell, filed October 13,1978.

S. N. 951,543— "High Power Laser Amplifier," W.T. Leland and T. F. Stratton, filedOctober 13, 1978.

S. N. 960.410— "Beam Heated Linear Theta-PinchDevice for Producing HotPlasmas," I. O. Bohachevsky, filedNovember 13, 1978.

PUBLICATIONS

E. R. Grilly, "Development of Cryogenic Targets forLaser Fusion," TIC (1977), 13 pp. MN, and Adv.Cryog. Eng. 23 (1978), pp. 676-81.

A. T. Lowe and R. J. Fries, "Plasma Polymerized P-Xylene as a Laser Fusion Target," Sur. Sci. 76, pp. 252-56.

K. B. Mitchell and R. P. Godwin, "Energy-TransportExperiments in 10 Micrometer Laser-ProducedPlasmas," J. App. Phys. 49, pp. 3851-4.

K. B. Reipe and M. Kircher. "Design of the EnergyStorage System for the High Energy Gas Laser Facilityat LASL." Engineering Problems of Fusion Research7th Symp. Proc. 2. M. S. Lubell, Ed.. Knoxville, Tennes-see (1977). pp. 1053-5.

B. E. Newnam and D. H. Gill, "Damage Resistance ofAr-Coated Germanium Surfaces for NanosecondCarbon Dioxide Laser Pulses," TIC (1977), 19 pp.

B. E. Newnam, "Damage Resistance of Coated Opticsfor Pulsed Carbon Dioxide Lasers," SPIE. Proceedingsof Optical Coatings-—Applications and Utilization II,TIC (1978). 18 pp.

S. J. Gitomer. "Laser Induced Fusion, IEEE Minicourse,2nd. Proc., G. H. Miley, Ed., Troy. New York (1977),pp. 10-20.

E. E. Bergmann, I. J. Bigio, B. J. Feldman, and R. A.Fisher. "High-Efficiency Pulsed 10.6 MicrometerPhase—Conjugate Reflection Via Degenerate Four-Wave Mixing," Opt. Lett. 3 (1978), pp. 82-4.

F. D. Wells and D. Remington, "Mirror Position DisplayEquipment for the Target Chamber Mirror Mounts ofthe 8-Beam Laser Fusion Research Facility," CubeSymp. Abst. J. J. Ruminer, Comp. (1978), pp. 81-2.

E. E. Stark, "Overview of Laser Fusion," Los AlamosScientific Laboratory Mini-Review LASL-77-24 (1978).

M. J. Campbell, "Device for Microparticle ArrayPreparation," Rev. Sci. Instrum. 49 (1978), pp. 1488-9.

M. E. Thuot, "Fiber Optics and Microprocessors—AControl-System Solution for the Laser-Fusion En-vironment," TIC (1978), 16 pp.

J. E. Sollid, S. J. Thomas, E. Foley, and C. R. Phipps,"Threshold of Detection for Various Materials at 10.6Micrometers," Appl. Opt. 17 (1978), pp. 2670-1.

J. J. Devaney, "Injection of Laser Fusion Pellets. Part 1.Accuracy Required, Acceleration, and Residual GasDeflections," Los Alamos Scientific Laboratory reportLA-7477-MS (1978).

129

J. E. Sollid. "Diffraction Measurement for LASL's An-tares Laser System," Los Alamos Scientific Laboratoryreport LA-7443-MS (1978).

C. W. Cranfill. and R. More, "IONOS. A Fast, Analytic,Ion Equation-of-State Routine,'" Los Alamos ScientificLaboratory report LA-7313-MS (1978).

R. L. Whitman, R. H. Day, R. Kruger, and D. M.Stupin. "Analysis of Laser Fusion Targets UsingMonochromatic X-Ray Microradiographs," Los AlamosScientific Laboratory report LA-7534-MS (1978).

M. A. Stroscio, D. B. Henderson, and A. G. Petschek."Numerical Computation of the Density ProfileProduced by 10.6 Micrometer Irradiation of a SiliconDioxide Microballoon," Nucl. Fusion 18 (1978). pp.1425-30.

R. K. Ahrenkiel. "Properties of Cobalt ChromiumSulfide Applicable to Faraday Rotators for CarbonDioxide Laser Systems," IEEE Trans. Magn. 14(1978),pp. 454-6.

E. E. Stark, "Lasers and Power Systems for InertialConfinement Fusion Reactors," TIC (1978), 24 pp.

E. E. Stark, "Carbon Dioxide-Laser Fusion," Proc. 13thIntersociety Energy Conversion Engineering Conf.,August 20-25, 1978, pp. 1362-1365.

A. J. Scannapieco and E. W. Cranfill, "Derivation of thePhysical Equations Solved in the Inertial ConfinementStability Code DOC," Los Alamos Scientific Laboratoryreport LA-7214-MS (1978).

W. T. Leland. "Design Engineering of Large High Pres-sure Gas Laser Amplifiers," TIC (1978).

W. T. Leland, "Advances in Laser Technology(Emphasizing Gaseous Lasers)," (Proc. SPIE Seminar,D. Finkleman, Ed., Washington, DC, 1978), pp. 38-45.

R. K. Ahrenkiel, J. F. Figueira, C. R. Phipps, and D.Dunlavy, "New Saturable Absorber for the CarbonDioxide Laser Using Doped KC1," AppJ. Phys. Lett. 33(1978), pp. 705-7.

J. L. Lyman, R. G. Anderson, R. A. Fisher, and B. J.Feldman, "Absorption of Pulsed CO2 Laser Radiation ofSF6 at 140 K," Optics Lett. 3 238 (1978).

J. E. Sollid, C. R. Phipps, Jr., and E. J. McLellan, "ALensless Method of Measuring Gaussian Laser BeamDivergence." Appl. Opt. 17 3527 (1978).

PRESENTATIONS

The following presentations were made at the AnnualMeeting of the Optical Society of America, San Fran-cisco, California. October 30—November 3, 1978.

R. A. Fisher, E. E. Bergmann, B. J. Feldman, and I. J.Bigio. "Phase Conjugate Reflections at 10.6 jjm viaIntractivity Degenerate 4 Wave Mixing in Germanium."

S. C. Stotlar. E. J. McLellan, and A. J. Gibbs, "Ultra-Fast PyroelectricMeasurements."

Detectors for CO, Laser

R. A. Fisher and B. J. Feldman. "Phase Conjugation inInverted CO2."

J. E. Sollid, "A Lensless Measurement of Gaussian LaserBeam Divergence."

W. C. Sweatt and V. L. Zeigner, "Antares TargetSystem Optical Design."

D. M. Stupin, M. A. Winkler. M. J. Campbell, and R.Liepins. "Radiographic Selection of Laser FusionTargets: Hollow Metal Shells and Plastic Coatings."

The following presentations were made at the 1978International Laser Conference, Orlando, Florida,December 11-15, 1978.

R. K. Ahrenkiel, D. Dunlavy, J. F. Figueira, and C. R.Phipps, Jr., "A New 10.6 urn Saturable Absorber: KC1Doped with KReO4."

S. J. Czuchlewski, E. J. McLellan, J. F. Figueira, E.Foley, C. E. Knapp, and J. A. Webb, "A High-Power(~10 GW) Short-Pulse (<1 ns) CO2 TEA Amplifier."

130

1. J. Bigio. B. J. Feldman. R. A. Fisher, and E. E.Bergmann. "High Efficiency Phase-Conjugate Reflectionin Germanium and in Inverted CO2."

E. J. McLellan. "A Reinjection CO, Oscillator."

The following presentations were made at the Electro-Optics Laser '78 Conference. Boston. Massachusetts.September 19-21. 1978.

E. J. McLellan and S. C. Stotlar. "Sub One-hundred psPyroelectric Detector Research and Evaluation Programat LASL."

R. A. Fisher. E. E. Bergmann, B. J. Feldman. and I. J.Bigio. "Phase-Conjugate Reflect on at 10.6 um viaDegenerate 4 Wave Mixing in Germanium."

K. C. Jones. J. L. Munroe. W. H. Reichelt. and J. E. Sol-lid. "Antares CO, Laser System. Optics and Com-ponents."

The following presentations were made at the Twen-tieth Annual Meeting of the Division of Plasma Physicsof the American Physical Society. Colorado Springs.Colorado. October 30—November 3. 1978.

C. W. Cranfill and R. Moore (H-Division. LLL. Califor-nia). "IONEOS: A Fast. Analytic. Ion-Equation of StateRoutine."

S. J. Gitomer and R. J. Mason. "A New CO, LaserTarget with Burnthrough Shields,"

A. J. Scannapieco and H. Brysk. "Symmetry of EnergyDeposition by a Multibeam Laser System."

S. J. Gitomer. "Spectral Analysis Techniques Applied toSubnanosecond Laser."

R. A. Kopp. "High Compression Ablative Targets forthe Helios Laser System."

R. J. Mason. "Laser Generated Hot Electron Transportin Self-Consistent E and B Fields."

A. J. Scannapieco. "Illumination and Energy Depositionin a Multi-Beam Laser-Fusion System."

A. J. Scannapieco and K. A. Taggart, "Stability of LaserFusion Targets."

W. S. Varnum and E. L. Lindman. "Antares DesignConcepts II."

The following presentations were made at the ANS1978 Winter Meeting. Washington, DC, November 12-17. 1978.

J. J. Dcvaney and L. A. Booth. "Laser-Fusion Pellet In-jection Accuracies. Accelerations and Deflections."

I. O. Bohachevsky. D. O. Dickman. and J. C. Goldstein."A New Plasma Model for the ICF Magnetic WallReactor Cavity Concept."

J. H. Pendergrass and C. E. Cort. "Lithium Boiler High-Temperature Fusion Reactor Blanket Concept."

In addition, the following presentations were made atvarious institutions.

R. K. AhrenkicU "A New Saturable Absorber for theCO2 Laser Using Doped KCI." Seminar. University ofRochester. August 21-22. 1978.

J. H. Birely. J. L. Lyman. G. Quigley. R. A. Fisher, andB. J. Feldman. "Multiple Photon AbsorptionMeasurements." National Conference on Nonlinear Op-tics and Lasers. Lenigrad. USSR. July 1978.

R. A. Fisher. B. J. Feldman. I. J. Bigio. and E. E.Bergmann. "Infrared Wavefront Reversal in Ger-manium." National Conference on Nonlinear Optics andLasers. Leningrad, USSR. July 1978.

B. J. Feldman. R. A. Fisher. E. E. Bergmann. and I. J.Bigio. "Infrared Phase Conjugation." Massachusetts In-stitute of Technology. July 1978.

R. A. Fisher, B. J. Feldman, "Infrared Phase Conjuga-tion," Seminar, Sandia Laboratories, Albuquerque, July1978.

R. A. Fisher and B. J. Feldman. "10-um PhotoplasmaFormation in Germanium," Seminar, SandiaLaboratories, Albuquerque, July 1978.

R. A. Fisher, B. J. Feldman, I. J. Bigio, and E. E.Bergmann. "Infrared Phase Conjugation Via 4-WaveMixing in Germanium," Seminar, Physics Department,University of California, Berkeley, November 1, 1978.

131

B. J. Feldman, R. A. Fisher, I. J. Bigio, and E. E.Bergmann. "Infrared Phase Conjugation," Seminar,Lawrence Livermore Laboratory, November 3, 1978.

R. A. Fisher, B. J, Feldman, I. J. Bigio, and E. E.Bergmann. "Phase-Conjugate Optics in the Infrared,"Seminar, Stanford University, Palo Alto, California,November 5, 1978.

R. A. Fisher. E. E. Bergmann, B. J. Feldman, and I. O.Bigio, "Phase Conjugation: the Wave of the Future,"Seminar. Physics Department, Lehigh University,Bethlehem. Pennsylvania, October 18, 1978.

B. J. Feldman. R. A. Fisher, E. E. Bergmann, and I. J.Bigio. "Infrared Phase Conjugation." Seminar. StanfordResearch Institute, Menlo Park. California, October 27,1978.

R. A. Fisher, B. J. Feldman, E. E. Bergmann, and I. J.Bigio, "Infrared Phase Conjugation." Seminar, OpticalSciences Center. University of Arizona, October 17,1978.

B. J. Feldman. R. A. Fisher, [. J. Bigio. and E. E.Bergmann. "Wavefront Reversal in Nonlinear Optics,"Seminar. Westinghouse Research Laboratories, Pitt-sburgh, Pennsylvania. October 16, 1978.

J. P. Hong. "Experiments in Minimum Length Represen-tation of Text." Third Berkeley Workshop onDistributed Data Management and Computer Networks,Berkeley, California, August 29-31, 1978.

J. P. Hong and S. Johnson, "Minimum Length Represen-tation or Text." Seventeenth IEEE Computer SocietyInternational Conference (COMPCON 78),Washington, DC, September 5-8, 1978.

W. H. Reichelt, "Damage Considerations in Short PulseCO2 Laser Systems," 10th Annual Symposium on Op-tical Materials for High-Power Lasers, Boulder,Colorado. September 12-14, 1978.

R. Williamson. "A 60-Inch Annular Pitch Polisher forLASL's Laser Fusion EfTort," Abstract of talk given atOptical Society of America Optical Fabrication andTesting Workshop, Dallas, Texas, November 1-10,1978.

T. F. Stratton. "Antares Overview," Los AlamosChapter of the Laser Institute of America, Los Alamos,New Mexico, November 14, 1978.

A. T. Lowe and C. D. Hosford, "Magnetron SputterCoating of Microspherical Substrates," NationalAmerican Vacuum Society Meeting, San Francisco,California, November 27—December 1, 1978.

I. O. Bohachevsky, "Elements to be Considered in Plan-ning Heavy Ion Fusion Program—A Summary," Proc.Heavy Ion Fusion Workshop. Argonne NationalLaboratory. September 19-26. 1978.

E. E. Stark, "COrLaser Fusion," Thirteenth IntersocietyEnergy Conversion Engineering Conf., San Diego,California, August 20-25, 1978.

T. G. Frank and L. A. Booth, "Commercial Applicationsof Thermionic Conversion Using a Fusion ReactorEnergy Source—A Preliminary Assessment," ThirteenthIntersociety Energy Conversion Engineering Conf., SanDiego. California, August 20-25. 1978.

A. R. Larson and C. D. Cantrell, "An Application ofNonequilibrium Quantum Statistical Mechanics toHomogeneous Nucleation," 11th International Sym-posium on Rarefied Gas Dynamics, Cannes, July 3-8,1978.

C. R. Phipps and S. J. Thomas, "10-um PhotoplasmaFormation in Germanium," Seminar, SandiaLaboratories, Albuquerque, July 1978.

*U.S. GOVERNMENT PRINTING OFFICE: 1980-777-069/10

132