unt digital library/67531/metadc623893/m2/1/high_re… · inertial confinement lawrence livermore...

INERTIAL CONFINEMENTINERTIAL CONFINEMENT LawrenceLivermoreNationalLaboratory

LawrenceLivermoreNationalLaboratory

October—December 1998, Volume 9, Number 1ICF Quarterly Report

UCRL-LR-105821-99-1

Special Issue: Laser Technology for theNational Ignition Facility

Special Issue: Laser Technology for theNational Ignition Facility The NIF Injection

Laser System

Main Amplifier Power Conditioning

Integrated Computer Control System

Multiaperture Optical Switch

Beamlet Experiments

Modeling the Interaction of the NIF Laser Beam with Laser Components

Beam Control and Laser Diagnostic Systems

Design and Performance of Flashlamp-Pumped Nd:Glass Amplifiers

On the Web:http://lasers.llnl.gov/lasers/pubs/icfq.html

This document was prepared as an account of work sponsored byan agency of the United States Government. Neither the UnitedStates Government nor the University of California nor any of theiremployees makes any warranty, express or implied, or assumes anylegal liability or responsibility for the accuracy, completeness, orusefulness of any information, apparatus, product, or process dis-closed, or represents that its use would not infringe privately ownedrights. Reference herein to any specific commercial product, pro-cess, or service by trade name, trademark, manufacturer, or other-wise, does not necessarily constitute or imply its endorsement, rec-ommendation, or favoring by the United States Government or theUniversity of California. The views and opinions of authorsexpressed herein do not necessarily state or reflect those of theUnited States Government or the University of California and shallnot be used for advertising or product endorsement purposes.

The ICF Quarterly Report is publishedfour times each fiscal year by the Inertial ConfinementFusion Program at the Lawrence Livermore NationalLaboratory. The journal reports selected current researchwithin the ICF Program. Major areas of investigationpresented here include fusion target theory and design,target fabrication, target experiments, and laser and opti-cal science and technology. In addition, the Laser Scienceand Technology program element of LLNLÕs LaserPrograms serves as a source of expertise in developinglaser and electro-optics capabilities in support of the ICFmission and goals and also develops new lasers for gov-ernment and commercial applications. To keep our read-ers informed of these new capabilities, the ICF QuarterlyReport now covers additional non-ICF funded, but relat-ed, laser research and development and associated appli-cations. As another improvement, we have added a shortsummary of the quarterly activities within Nova laseroperations and NIF laser design. Questions and com-ments relating to the technical content of the journalshould be addressed to the ICF Program Office,Lawrence Livermore National Laboratory, P.O. Box 5508,Livermore, CA 94551.

The Cover: Several components and resultsstemming from the NIF laser technology and engineeringdevelopments are shown. In the upper left is a picture ofthe 500-kA spark gap switch developed byMaxwell/Physics International, which enables the cost-effective, high-energy, pulse-power modules designedfor the NIF (see the article ÒMain Amplifier PowerConditioning for the NIFÓ on p. 15). On the upper right isa picture of a prototype 4 ´ 1 plasma electrode Pockelscell (PEPC) mounted on its side for convenient testing inthe laboratory. The PEPC, when energized, rotates thepolarization of the light across the apertures of fourbeams to allow four-pass pulse amplification in the mainamplifier cavity (see the article ÒMultiaperture OpticalSwitch for the NIFÓ on p. 33). The contour plot resem-bling a birdÕs face, shown in the center, is actually the cal-culated intensity distribution of laser light after it passesthrough a small (4-µm-diam) zirconia inclusion imbed-ded in fused silica. The light, which propagates from topto bottom, shows local intensifications as large as 1000times that can readily cause optical damage because ofthe focusing action of the inclusion and interference withthe main beam (see the article ÒModeling the Interactionof the NIF Laser Beam with Laser ComponentsÓ on p.63). The three tubular objects compared to a dime at bot-tom are actually conical pinholes of the type tested in theBeamlet laser. The use of high-Z conical pinholes allowstemporally shaped NIF pulses to be spatially filtereddown to 150 µrad. This filtering produces more uniformbeams that allow increased NIF output energy withoutthe threat of plasma generated in the pinhole cutting offthe pulse in time (see the article ÒBeamlet ExperimentsÓon p. 43).

UCRL-LR-105821-99-1OctoberÐDecember 1998

Printed in the United States of AmericaAvailable from

National Technical Information ServiceU.S. Department of Commerce

5285 Port Royal RoadSpringfield, Virginia 22161

Price codes: printed copy A03, microfiche A01.

Work performed under the auspices of the U.S. Department ofEnergy by Lawrence Livermore National Laboratory underContract WÐ7405ÐEngÐ48.

i

In this issue:Foreword iii

The NIF Injection Laser System (J. K. Crane) 1The injection laser system (ILS) is the front-end of the National Ignition Facility (NIF)laser system. The ILS creates the initial pulse that is formatted in time, frequency, andspace; amplified; and fanned out into 192 3-J pulses that seed the NIF main amplifierchains. We describe the development of this complex laser system, its performance, andour work to build engineering prototypes.

Main Amplifier Power Conditioning for the NIF (M. Newton) 15Design and development of the NIF power conditioning system are nearing comple-tion. Tests on the individual components and on a prototype power conditioning mod-ule at Sandia, Albuquerque, have demonstrated the required performance.

Integrated Computer Control System (P. J. VanArsdall) 21The NIF design team is developing the integrated computer control system, which con-sists of over 300 front-end processors attached to 60,000 control points that are coordi-nated by a supervisory system of eight operator consoles. Software for this massive sys-tem is being constructed from object-oriented components provided by a reusable soft-ware framework designed for event-driven control. The framework is interoperableamong different kinds of computers and functions as a distributed, plug-in softwarebus by leveraging a Common Object Request Brokering Architecture.

Multiaperture Optical Switch for the NIF (M. Rhodes) 33This article discusses the NIF optical switch, an electro-optic device of unprecedentedproportions. Based on plasma-electrode Pockels cell technology, the NIF optical switchcontrols multipass operation of the main cavity amplifier in the NIF laser.

Beamlet Experiments (P. Wegner) 43The Beamlet laser is a single-aperture physics prototype of the 192-beam Nd:glass laserdriver for the NIF. In the four-year period since its activation milestone in September1994, Beamlet has produced over one-thousand full-system shots in over twenty experi-mental campaigns addressing a broad range of laser physics and component engineeringissues related to NIF design and operation. This past July, Beamlet completed its NIF mis-sion and has subsequently been transferred to Sandia National Laboratories,Albuquerque, where it will continue service as a high-energy backlighter in the study ofZ-pinch plasmas.

Modeling the Interaction of the NIF Laser Beam with 63Laser Components (M. D. Feit)Using computational modeling, in combination with experimental results, one can predictand ameliorate potential adverse effects of the high-intensity NIF beam on the laser compo-nents. Such effects include pinhole closure in the spatial filters and laser-induced damage ofoptics. In this article, we present models to describe and analyze these two phenomena.

ICF Quarterly Report

INERTIAL CONFINEMENTINERTIAL CONFINEMENT

Scientific EditorHoward T. Powell

Publication EditorJason Carpenter

DesignerPamela Davis

Technical EditorsPat BoydCindy CassadyRobert KirvelAl MiguelJeff MorrisJane Staehle

Classification EditorRoy Johnson

Art StaffTreva CareyFrank MarquezThomas ReasonJudy RiceTony SanchezFrank Uhlig

Cover DesignSandy Lynn

OctoberÐDecember 1998, Volume 9, Number 1

ii

IN THIS ISSUE

UCRL-LR-105821-99-1

Beam Control and Laser Diagnostic Systems (E. S. Bliss) 79Beam control and laser diagnostic systems control propagation of each beam through theNIF laser, correcting for both alignment and wavefront errors. They also characterize thebeam in several locations with respect to pulse energy, power versus time, spatial flu-ence uniformity, and wavefront. A wide range of optomechanical subsystems is requiredto accomplish these tasks, while multiplexing helps to keep costs under control.

Design and Performance of Flashlamp-Pumped Nd:Glass 99Amplifiers for the NIF (A. Erlandson)We have designed and tested flashlamp-pumped Nd:glass for the NIF. The amplifierÕsmodular design improves maintenance and storage efficiency and features active gascooling. Our results showed that gain, wavefront, and thermal recovery measurementsmeet NIF performance requirements.

Nova Update A-1

National Ignition Facility Update B-1

Publications and Presentations C-1

iii

FOREWORD

UCRL-LR-105821-99-1

FOREWORDIn 1994, a group of scientists and engineers at Lawrence Livermore National

Laboratory (LLNL), with input from the other U.S. inertial confinement fusion (ICF) lab-oratories, formed a four-year laser development plan to enable the construction of thelaser hardware for the National Ignition Facility (NIF).* The risk-reduction laser activitiesnecessary to complete detailed Title II engineering for NIF were defined, the manpowerrequirements and overall costs in each area were estimated, the required developmentlaboratories and teams were identified, and preliminary milestones were proposed. Withboth financial and technical help from the French Commissariat a lÕEnergie Atomique(CEA), we began laser development activities in earnest in 1995. Simultaneous with thelaser-component development activities, a NIF Project engineering team was assembledand began the design of the real NIF hardware, interfaces, and assembly plans. Thedevelopment and engineering activities were completely on schedule four years later atthe end of 1998, as marked by the completion of detailed Title II engineering reviews.

The activities in the intervening four years were intense and involved a large numberof people from LLNL, Sandia National Laboratories (SNL), Los Alamos NationalLaboratory (LANL), the French CEA, and many other groups. Laser-development activi-ties covered most of this period while the team continuously redefined the milestonesand conducted frequent laser-technology reviews. This issue of the ICF Quarterly Reportdescribes the NIF laser component designs that resulted from these laser developmentand engineering activities.

In this issue, the article ÒThe NIF Injection Laser SystemÓ by Crane et al. describes theultrastable and ultraflexible laser pulse generation system that provides input to each ofthe 192 beams of the NIF. The article ÒMain Amplifier Power Conditioning for the NIFÓby Newton et al. summarizes the work of the joint LLNL and SNL team, who successful-ly tested a Òfirst-articleÓ 1.7-MJ power conditioning module at Sandia in Albuquerque,NM. The article ÒIntegrated Computer Control SystemÓ by VanArsdall et al. describesthe complex, distributed computer control system for the NIF, which may be a model forother large computer control systems. The article ÒMultiaperture Optical Switch for theNIFÓ by Rhodes et al. describes the design basis and performance of a prototype 4 ´ 1plasma electrode Pockels cell that matches the NIF beam architecture. The articleÒBeamlet ExperimentsÓ by Wegner et al. gives the results of the experimental campaignson the Beamlet laser to define the design and performance limits of the NIF. The articleÒModeling the Interaction of the NIF Laser Beam with Laser ComponentsÓ by Feit andBoley describes the use of target-interaction modeling to understand and improve thelaser performance limits. The article ÒBeam Control and Laser Diagnostic SystemsÓ byBliss et al. describes the wavefront system and the alignment and diagnostic systemresulting from development and engineering. Finally, the article ÒDesign andPerformance of Flashlamp-Pumped Nd:Glass Amplifiers for the NIFÓ by Erlandson et al.describes the work of an integrated U.S.ÐFrench team to design and prototype the large 4´ 2 power amplifiers. This article is reprinted from the last Quarterly issue (Vol. 8, No. 4)for this special issue in order to cover all the NIF laser design and development results.

Howard T. PowellScientific Editor

*See Chapter 1 of the document Core Science and Technology Plan for Indirect Drive ICF Ignition, H. Powell and J. Kilkenny,Lawrence Livermore National Laboratory, Livermore, CA, UCRL-ID-117076 Rev. 1 (December 1995).

More scientifically detailed descriptions of both the U.S. and French work, as well as work of smaller ICF laser researchgroups around the world, can be found in SPIE, Volumes 2633, 3047, and 3492. These volumes were published following theFirst, Second, and Third International Conferences on Solid-State Lasers for Application to Inertial Confinement Fusion heldrespectively in Monterey, CA; Paris, France; and Monterey, CA, in 1995, 1996, and 1998.

1

The injection laser system (ILS), orÒfront end,Ó is the portion of the

National Ignition Facility (NIF) where asingle pulse is produced, modulated, andshaped, then amplified and multiplexed tofeed the 192 main amplifier chains in theNIF.1 The ILSÕs three major subsystems aresummarized in the overview, thendescribed in detail in their own sections.In many cases, the subsystems have beendeveloped and are in an engineering pro-totype phase in which we work with out-side vendors to produce hardware. Wehave also connected two of the subsys-tems, the master oscillator room (MOR)and preamplifier module (PAM) develop-ment labs, to perform integrated perfor-mance measurements on a combined system.

OverviewThe ILS is composed of three major sub-

systems: the master oscillator room (MOR),the preamplifier modules (PAMs), and thepreamplifier beam transport system(PABTS).

The master oscillator room is responsi-ble for generating the single pulse thatseeds the entire NIF laser system. In theMOR, this single pulse is phase-modulatedto add bandwidth, then multiplexed into48 separate beamlines on single-mode,polarizing fiber. Before leaving the MOR,the pulses are temporally sculpted intohigh-contrast shaped pulses designed toproduce ignition of the deuteriumÐtritium

(DÐT) targets. Forty-eight single-modefibers from the MOR serve as inputs to the48 PAMs that are the second major subsys-tem in the ILS.

The preamplifier modules provide thelargest amount of amplification in theentire NIF laser system, >100 dB. In addi-tion to providing amplification, the PAMsspatially shape the Gaussian beam thatemerges from the single-mode fiber toform a square beam that is shaped to com-pensate for the spatial gain profiles of themain slab amplifiers. A third function per-formed in the PAMs is spectral dispersionof the phase-modulated light producedinitially in the MOR. This dispersion ispart of a scheme called smoothing by spec-tral dispersion, or SSD,2 that reduces thespatial coherence of the laser light irradiat-ing the target. The 48 17-J outputs from thePAMs enter the final subsystem of the ILS,the preamplifier beam transport system.

In the PABTS, the 48 beams from the 48PAMs are split into 192 separate beamsthat feed the main amplifier chains. Afterthis four-way split of the beams, each leghas an optical trombone section for pre-cisely adjusting the timing so that all 192beams converge on target simultaneously.

Figure 1 shows a schematic block dia-gram of the ILS and its constituent sys-tems. The requirements for the ILS arederived from the overall laser systemrequirements at the target. Table 1 lists theoverall system requirements at the targetand the resulting parameters for the ILSdetermined from the flowdown.

THE NIF INJECTION LASER SYSTEM

J. K. Crane R. B. Wilcox M. Hermann M. Martinez B. Moran

M. Henesian G. Dreifuerst B. Jones J. E. Rothenberg K. M. Skulina

D. Browning L. A. Hackel L. Kot F. Penko F. Deadrick

UCRL-LR-105821-99-1

2


UCRL-LR-105821-99-1

Master Oscillator RoomThis section describes the subsystems that

make up the MOR. All these systems havebeen demonstrated at the scientific proto-type stage and are being engineered for thefinal NIF version. A simplified version of theMOR architecture is shown in Figure 2.3

The 192 pulses that converge on the tar-get with a combined energy of nearly2,000,000 J originate as a low-power continuous-wave (CW) signal in the MORÕsmaster oscillator. This oscillator is a fiber dis-tributed feedback (DFB) laser, made by

imprinting linear Bragg gratings in an Yb-doped fiber using UV light. This technologyis similar to that used in Er-doped fiberlasers for communications.4 The Bragg grat-ings define the cavity end reflectors, whichare such narrow band reflectors that theyallow only one cavity mode to lase.

When pumped with a 120-mW laserdiode, the oscillator produces about 50 mW CW in a single longitudinal mode.The master oscillator wavelength is pre-cisely tuned to within 0.01 nm of thedesired value by temperature-controllingits mount, which changes the Bragg

Master �Oscillator Room�

�(MOR)

Preamplifier�Modules�

�(PAMs)

Preamplifier�Beam Transport�

System��

(PABTS)

48�beams

Schematic of ILS

� � �

48 �fibers

192 �beams

FIGURE 1. The NIF injec-tion laser system (ILS).(70-00-0299-0358pb01)

TABLE 1. System requirements for the laser at the target and the requirements at the output of the ILS, determined from a flowdown of the system requirements back to the front end.

NIF System Requirementsa Injection Laser System Requirements

Output energy 1.8 MJ Injected energy into main amps 3 J

Peak power 500 TW Peak power at injection 1.2 GW

Wavelength 352 nm Preamp output energy (flat-top beam) 22 J

Pulse duration 20 ns Preamp power energy (shaped beam) 16.9 J

Power balance 8% in 2-ns window Wavelength 1053 nm

Pointing accuracy @ target 6 µrad Pulse duration 20 ns

Power dynamic range >50:1 Output pulse rate 1/20 minute

Prepulse in 20-ns window <108 W/cm2 Prepulse contrast >2 ´ 106

Number of beamlets 192 Square pulse distortion <2.3

Bandwidth 81 GHz

Critically dispersed (SSD) Ð

Spatially shaped for gain compensation Ð

Number of preamplifier modules 48

aReference 1.

wavelength by strain and the thermoopticeffect. Figure 3 shows a picture of apumped, unmounted oscillator, and aschematic of the fiber master oscillator.

Amplification and ModulationThe CW signal from the oscillator is

chopped by an acoustooptic modulator toa pulse width of 100 ns and passedthrough a double-pass fiber amplifier witha gain of 500. The amplifier is internallyfiltered to reduce the amount of noise itadds to the oscillator signal. When the sig-nal has passed through the amplifier once,it reflects off a 2-�-wide passive fiberBragg grating filter. This reduces amplifiedspontaneous emission (ASE) noise fromthe first pass. After a second pass, the out-put signal is filtered through a 5-nm band-pass filter to further reduce ASE. A direc-tional coupler is used to multipass the

amplifier because it has low back reflec-tion, high damage threshold, and nearlythe same loss as a circulator at this wave-length. After passing through this amplifi-er and undergoing several componentlosses, the 100-ns pulse has a peak powerof about 1 W.

At this point, phase modulation isapplied to the single-frequency signal tocontrol its eventual interaction with thefinal NIF optics and the target. Phase mod-ulation at 1.5 GHz produces about 10 opti-cal sidebands, which reduce the spectralintensity in the final optics and preventstimulated Brillouin scattering (SBS),which could damage the optics.5 Phasemodulation at 17 GHz creates a bandwidthof 3 �. When diffracted from a grating inthe PAM, the 3-� bandwidth quicklymoves the focused laser beam on target tosmear out speckle through a process calledsmoothing by spectral dispersion (SSD).

3


UCRL-LR-105821-99-1

Timingaccuracy

Precisiontriggersystem

AWGchassis

1-48

Signalrouter

Signalrouter

Phasemod

Timingsystem

Microwavesystem

Osc.controls

Osc.diagnostics

Mod.diagnostics

Amplifierdiagnostics

Timing range

AWG = arbitrary waveform generatorLD = laser diagnosticTD = target diagnostic

ToLD,TD

Topreamps

Topreamps

SBS, SSDbandwidth

Wavelength

Oscilllator

Fiducial

Sync

Lig

ht

Lig

ht

Pulse durationand shape,

power balance,prepulse

Dist.rack

48

24

24

48

Signalrouter

Amps,splitters

FIGURE 2. The master oscillator room (MOR) architecture. (70-00-0299-0359pb01)

4


UCRL-LR-105821-99-1

Both modulations are applied with lithi-um niobate, electrooptic waveguide modu-lators similar to those used in fiber-opticcommunications. Modulators that producethese modulations require only a fewwatts of microwave power, reducing thecomplexity of the drive circuit andenhancing reliability. Because it is impor-tant that SBS not damage the NIF optics, adoubly redundant, fail-safe circuit moni-tors the presence of 1.5-GHz phase modu-lation and blocks the laser pulse if theamount of modulation is too low. Severalparameters are measured on each pulse toensure that each one has the requiredbandwidth.

To reduce gain saturation in the fiberamplifiers, the 100-ns pulse is reduced inanother waveguide modulator to 30 nsbefore being introduced into the fiberamplifier and splitter array.

The amplifiers in the array are Yb-doped, two-stage, CW-pumped fiber

amplifiers, shown in Figure 4 and devel-oped with JDS Feitel. The first stage is asmall core, Yb-doped, gain fiber with gainof 12, followed by a band-stop filter to pre-vent saturation of the second stage byASE. The second stage, a power amplifier,is made of larger core fiber and has a gainof 3, for a total gain of 36 for each fiberamplifier. The amplifiers for NIF will beproduced to our design and specificationsby a commercial manufacturer.

After each stage of amplification, thepulse is split four ways. In the first stage,three outputs are used, so that the totalnumber of outputs in three stages ofamplification is 3 ´ 4 ´ 4 = 48. The 48 out-puts each produce about 2 W in 30 ns.6

Arbitrary WaveformGenerator

The final temporal pulse shape is sculpt-ed from the 30-ns square pulse after the laser

Power controlcircuit

DFB laser

x

Oscillator System

x

Collimator

PINdetector

(b)

DFB = distributed feedbackWDM = wavefront division ) multiplexer

(a)

Driverdiode

Polarizationcontroller

Pumpdiode

Isolator

Output

WDM

FIGURE 3. (a) Photo of apumped, unmountedoscillator. (b) Schematic ofthe fiber master oscillator.(70-00-0599-1058pb01)

has fanned out to 48 parallel fiber lines.Together with Highland Technology, wedeveloped a high-bandwidth, program-mable, arbitrary waveform generator (AWG)to produce the high-temporal-contrast pulsesthat are needed for fusion ignition.7

An electronic waveform generator supplies a shaped-voltage pulse to an inte-grated optic, lithium niobate amplitudemodulator, which modulates the lightpulse. The electronic waveform generatorworks by adding 300-ps voltage pulsesseparated in time by 250 ps. NIF will have96 300-ps pulse generators, each of whichcan be precisely controlled in amplitude.

By programming these pulse amplitudes,we can create an arbitrary pulse shape.

There are some applications where a shorter pulse is required. This is providedby a second pulse generator through a spe-cial circuit. For example, some targetsemploy a secondary target that is illuminat-ed by a short laser pulse. When the pulsestrikes the target, x rays are produced.These x rays act like a photo flash to cap-ture an instant in the target implosion.Also, timing of the 192 beams to coincide inthe center of the target chamber requires ashort pulse for timing precision. In thesecases, a pulse of about 150 ps is produced.

5


UCRL-LR-105821-99-1

200-MHzdriver

AO

Yb:silicafiber

Directionalcoupler Delay

for 100 ns

Diode-Pumped High-Gain Amplifier

Bragg grating

Pump diode

Acousto-optic chopper

Isolator

Input from oscillator

Output

WDM

FIGURE 4. The Yb-doped, two-stage, CW-pumped fiber amplifiersin the splitter array.(70-00-0599-1059pb01)

6


UCRL-LR-105821-99-1

The electronic waveform generator andoptical modulator are combined in a singlechassis that is being produced by an out-side manufacturer. There will be 48 ofthese units in the final NIF laser system.Figure 5a shows the generator.

Figure 5b shows a typical, high-tempo-ral-contrast pulse generated by the AWG.It is essential that these pulses arrive at thetarget simultaneously to provide uniformtarget implosion. Only 12 ps of timinguncertainty is allowed for the pulses leaving the MOR.

Synchronization of the pulse generatorsis accomplished by triggering each onewith a short optical pulse passively splitfrom a single, high-energy optical pulse. Inthis trigger system, a 100-ps electricalpulse synchronized with the integratedtiming system (ITS) drives an electroopticmodulator to produce a 100-ps opticalpulse. This is amplified in fiber amplifiersand split to trigger the pulse waveformgenerators. A portion of this pulse is alsosent out of the MOR to fiducial laser andtarget diagnostic systems to be used as atiming marker exactly synchronized withthe main optical pulses.

After final temporal shaping, the 48 parallel fiber lines are routed through a central distribution rack. From there they are sent to the 48 PAMs. The central

fiber distribution is carefully designed toensure that the individual lines producethe correct timing delay to each PAM.

We can add or subtract fiber jumpers todiscretely vary the delay of each beamlinefor needed adjustment, or to satisfy therequirements of special experiments. Inaddition to the laser itself, the MOR con-tains auxiliary systems for controls, diag-nostics, and precision timing.

Preamplifier ModulesAs part of the development effort for

the ILS, the various subsystems in theMOR and PAM were assembled and testedin the laboratory, and then combined intoan Integrated ILS testbed.8 Using this com-bined system, we were able to test thosespecifications of the ILS that required thefunctions of both the MOR and PAM.

Each of the 48 fibers from the MOR dis-tribution rack feeds one of the 48 PAMs inthe NIF laser system. The PAM is a high-gain (>100 dB) preamplifier that also spatially shapes the beam for the mainamplifier chain. Figure 6 is a schematic of the PAM showing the three major sub-systems: an ultrastable, high-gain, diode-laser-pumped, Nd:glass regenerativeamplifier; a beam-shaping module; and afour-pass amplifier.

0.10

0.05

0

0.15

0.20

0.25

0.30

0 5 10 15 20Time (s × 10Ð9)

(a) (b)

AW

G o

utp

ut (

V)

AWG output�Specified output

FIGURE 5. (a) The arbitrary waveform generator (AWG) chassis. (b) Plot showing specified (black) and measured (gray) electronic pulseshapes produced by the AWG. (70-00-0299-0360pb01)

Regenerative Amplifier

The regenerative amplifier is the highestgain amplifier in the entire NIF laserchain.9Ð11 The optical layout for the regener-ative amplifier, or ÒregenÓ for short, isshown in Figure 6. The input section to theregen consists of a fiber launch, where thepulse from the MOR is launched from single-mode fiber into free-space via a preci-sion fiber positioner and a short-focal-lengthlens. Next, two Faraday isolators in seriesprotect the single-mode fiber from a high-intensity pulse reflecting, or propagatingback, from the regen output. A second lens,in conjunction with the short-focal-lengthlens in the fiber launch, forms a telescope tomatch the beam size at the fiber output tothe laser cavity mode for efficient couplingof energy into the regen.

A Faraday rotator, a half-wave plate(WP), and a thin-film-polarizer (TFP) forma unidirectional coupler to separate thecounterpropagating input and output laserpulses. The input laser pulse from theMOR is injected into the regen cavitythrough a second TFP.

The regen cavity is a long, asymmetriccavity with a single, diode-laser-pumped,Nd:glass amplifier located at one end.The amplifier has a single-pass gain of G = 1.4 in a 5-mm diameter ´ 50-mm rodthat is end-pumped by a 4-kW diodearray. The cavity transmission is T = 0.77,so the net gain per round trip of the regen is Gnet = G2 ¥ T = 1.5. A cavity

Pockels cell switches polarization of theinput pulse after it has made a singlepass, trapping the pulse in the regen until the Pockels cell is switched back to its original polarization.

Each time the circulating pulse makes around trip in the cavity, the pulse is ampli-fied by Gnet. The total gain of the regenera-tive amplifier is the net gain raised to thepower of the number of round trips thatthe pulse makes in the cavity before beingswitched out. For example, if the numberof round trips in the regen cavity is k = 40,then the total regen gain is

Gtotal = (Gnet)k = (1.50)40 = 1.4 ´ 107 . (1)

Eventually the amount of extractedenergy depletes the available storedenergy in the amplifier (amplifier satura-tion) and the net gain drops below unity.In typical operation we extract 12 mJfrom the regen for 0.8-nJ input, for a total gain of 72 dB. Upon leaving theregen cavity, the amplified pulse passesthrough a second Pockels cell called theslicer, which acts as an optical gate. Theslicer passes the main regen pulse whilerejecting pre- and postpulses that leakout of the cavity due to imperfect polar-ization switching.

The output mode from the regen is a 3.5-mm-diam circular Gaussian-shaped beam.

7


UCRL-LR-105821-99-1

Backside of preamplifier module

Diode-pumped amp

Pulse slicer

Vacuum spatial filter

FR

Mirror/grating

FR = Faraday rotatorpc = Pockels cellWP = wave plate

Throughport

FRWP

WP

Input from MOR:

1 nJ

Regenerativeamplifier

Output22 J<21 ns

Frontside of preamplifier module

5-cm rodamp

Beam-shapingmodulepc

FIGURE 6. Optical layout of the preamplifier module (PAM). (70-00-0299-0361pb01)

8


UCRL-LR-105821-99-1

Beam-Shaping Module

The next subsystem in the PAM is thebeam-shaping module, where the roundGaussian beam is converted to a square-shaped beam that is sculpted in a welldefined manner to precompensate for thespatial gain profile of the main ampli-fiers.12 The goal is to produce a spatiallyflat-top beam at the NIF target chamber toyield the maximum irradiance at the target.

First the round beam is magnified bya 20´ telescope. The expanded Gaussianbeam is shaped by an anti-Gaussian fil-ter that flattens the center of the beam.This truncated Gaussian beam is furthershaped by a gain-compensating maskthat carves out the center of the beam toproduce the final shape. We determinethis final shape based on a complexdiffraction code that models an entirebeamline, including the spatial gain pro-files of the main amplifiers.12 The finalmask in the beam-shaping module, a ser-rated aperture, apodizes the beam sothat it will propagate over a long dis-tance with minimal ringing at the edgesof the beam from diffraction.13

All of the beam-shaping masks in themodule are made by depositing chromeon glass in a standard photolithographicprocess. The shaping masks reduce theenergy from the regen by a factor of 10 to15, from 12 mJ down to <1 mJ. This spa-tially shaped beam is injected into the finalPAM system, the four-pass amplifier.

Four-Pass AmplifierThe final amplifier in the PAM is laid

out in a four-pass configuration asshown in Figure 6. The input beam isspatially filtered by a vacuum relay tele-scope with a pinhole that blocks thehigher spatial frequencies produced bythe high-frequency teeth in the finalapodizing mask. The filtered input beampasses through a combination of Faradayrotator, half-wave plate, and thin-filmpolarizer that acts as a directional cou-pler to separate the input beam from thecounterpropagating output beam exitingfrom the four-pass cavity.

The four-pass cavity contains a Nova, 5-cm, flashlamp-pumped, rod amplifier atthe cavity center. This amplifier can oper-ate with a single-pass gain of 20.

Due to the high, single-pass gain of the5-cm amplifier, the cavity is especially sus-ceptible to unwanted, parasitic oscillation.Parasitic oscillation can occur in a high-gain amplifier if there is sufficient gain andfeedback to cause oscillation. This oscilla-tion is uncontrolled and chaotic and pro-duces a background of intense light thatcan propagate into the main amplifierchain along with the amplified and format-ted pulse from the MOR.

We take several steps to successfullyeliminate unwanted oscillation in the four-pass cavity. These include carefully con-trolling the polarization of the light usingthe Faraday rotator and the quarter-waveplate, and offsetting the beam slightly fromthe optical axis as it propagates throughthe cavity so light is never reflected direct-ly back on itself.14

Figure 7 shows the measured output ener-gy of the four-pass amplifier as a function ofenergy at the input. The solid curve in theplot is the predicted performance basedupon a FrantzÐNodvik model for saturatedgain.15 The dashed line shows the requiredperformance of the four-pass amplifier for anunshaped input pulse (see Table 1).

A second function of the four-pass ampli-fier optical layout is to provide angular dis-persion of the light that has been frequencymodulated in the MOR for SSD. To achievethis, we replace one of the end mirrors in thefour-pass cavity with a diffraction gratingpositioned at the Littrow angle so the firstdiffraction order is reflected back in the same direction as the incoming beam. If thefour-pass output is viewed in the far field(i.e., at the focus of a lens), the focal spot isdithered in one dimension at the sinusoidalfrequency of the rf driver to the modulator(e.g., 17 GHz). The angularly deflected lightfrom the amplifier propagates through therest of the amplifier chain to the target cham-ber and final optics assembly, where it passesthrough a special phase plate.

The combination of the rapid ditheringof the FM light due to dispersion from thegrating, and the smearing of the focal spot

produced by the random phase plate,reduces the spatial coherence (or specklepattern) normally associated with coherentlight and increases the uniformity of thetarget illumination.

The current beam-smoothing requirementfor the NIF is a 3-� bandwidth, critically dispersed in one dimension (see Table 1).ÒCritically dispersedÓ means that the tempo-ral skew produced across the 3-cm beam bythe grating at Littrow angle must be equal tothe period of the sinusoidal rf waveform thatdrives the phase modulator. The temporalskew DT is given by

DT = 2Dtan qL/c = 1/fmod , (2)

where D = 30 mm is the beam dimensionat the grating in the dispersion direction,and fmod is the rf driver frequency. TheLittrow angle is given by qL = sinÐ1(l/2d),where l is the wavelength of the laser andd is the grating groove spacing.

In the current design, the modulation fre-quency for SSD is 17 GHz, and the criticaldispersion requirement is slightly exceededfor a 600-line/mm grating. The 17-GHzmodulator system was not available for theILS subsystem tests, so we used a 3-GHzmodulation frequency. We used an 1800-line/mm grating that exceeded the critical

dispersion requirement and allowed us toclearly resolve individual FM sidebands.

Figure 8a is an image taken by the far-field camera diagnostic at the output of the four-pass amplifier. The angular dispersion produced by the 1800-line/mmgrating cleanly separates the individualspots in the far field. We show a plot inFigure 8b comparing a lineout from thecamera image with a power spectrum for an 81-GHz modulation bandwidth. The power spectrum, P(t), is given by theBessel series:

E(f) Å ·n=0° Jn(m)¥[d(f0 + nfrf)

+ (Ð1)n d(f0 Ð nfrf)] (3)

P(f) Å |E|2 , (4)

where E(f) is the laser field in the frequen-cy domain, Jn is the nth order Bessel func-tion, m is the modulation index, d is thedelta function, and frf is the rf frequencydriving the phase modulator. The locationsof the peaks and their magnitudes fromthe image lineout accurately compare withthe simulated spectral peaks. The 81-GHzFM spectrum is confirmed by the spec-trometer measurements in the MOR and atthe regen output.

9


UCRL-LR-105821-99-1

0

5

10

15

20

25

30O

utpu

t ene

rgy

(J)

NIF requirement for�flattop beamshape

Calculated extraction energy�Measured extraction energy

0 0.2 0.4 0.8 1.0 1.20.6Input energy (mJ)

FIGURE 7. PAM outputenergy as a function offour-pass input energy.Solid curve is aFrantzÐNodvik modelfor saturated gain. (70-00-0299-0362pb01)

10


UCRL-LR-105821-99-1

The results from experiments on theIntegrated ILS Testbed demonstrate thatthe PAM and MOR combined systemsmeet the requirements listed in Table 1. The optical layout and specificationsdemonstrated in the PAM development lab become the basis for the first PAM engineering prototype.

PAM EngineeringPrototype

The 48 preamplifier modules that seedthe main amplifier chains in the NIF lasersystem reside on the preamplifier supportstructure (PASS), a large space frame in themain laser bay. The PAMs are designed tobe line-replaceable units, or LRUs, that can slide into place on precision rails onthe PASS or be quickly removed as a self-contained unit and replaced or repaired as needed.

In our laboratory we set up and assem-bled the first preamplifier module, an engi-neering prototype of the production PAMsthat will form part of the ILS. Figure 9 is aphotograph of the PAM prototype. Wereproduced the three optical subsystems,the regenerative amplifier, the beam-shaping module, and the four-pass ampli-fier that were developed and demonstrat-ed on the Integrated ILS testbed, andmounted these subsystems on the PAM

prototype. The regenerative amplifier andbeam-shaping module are mounted on oneside of a vertical optical support structure.The larger four-pass amplifier is mountedon the opposite side.

The electronics to support the PAMinclude: a diode-laser power supply, stepper motor controllers, ion-pump con-trollers, timing modules, temperature controllers, and a VME-based embeddedprocessor controller that provides overallelectronics control of the PAM. These elec-tronics units are mounted above the opticssupport structure in an electronics baythat runs the length of the PAM. A high-

energy, electrical pulser that drives the six flashlamps in the 5-cm rod amplifier is housed in a separate unit, apart from the PAM.

We have operated the PAM prototypeover the past several months to demon-strate the requirements listed in Table 1.In addition, we developed a high-resolution output beam diagnostic toaccurately measure the wavefront, near-field contrast, and the far-field spot sizeat the output of the PAM prototype.Figure 10 shows processed data takenwith the high-resolution out-put beamdiagnostic for a 17-J shot. The informa-tion that we garner from these compre-hensive tests of the PAM engineeringprototype will enable us to design the 48 production PAMs.

f

Ð40 Ð20 20 4000

0.2

0.4

0.6

0.8

1.0

Frequency (GHz)

(a)

(b)

Image lineout�Model: m = 13.5�� f = 3 GHz

Sign

al (r

el. u

nits

)

FIGURE 8. (a) The far-field camera image at thePAM output showing thebeam dispersed by an1800-line/mm grating.(b) Plot comparing a hor-izontal lineout from (a)and an FM spectrumwhere the rf frequency is3 GHz and the modula-tion depth is 13.5. (70-00-0299-0363pb01)

11


UCRL-LR-105821-99-1

FIGURE 9. PAM engi-neering prototypemounted on mini-PASSstructure. This viewshows the four-passamplifier. (70-00-0299-0364pb01)

Ð15

Ð10

Ð5

0

10–0.2–0.1

0

0 0

0.1

0.2

0.30.4

0.1

–0.1

5

0 10Ð10mm

(b)

mm

Ð0.6

Ð0.4

Ð0.2

0

0.4

0.6

0.2

0 0.4Ð0.4mrad

(a)

mra

d

FIGURE 10. (a) Far-fieldimage at the output of thePAM prototype takenwith a high-resolutionPAM output diagnostic(Strehl ratio = 0.65).(b) Phase map of the out-put wavefront taken witha radial shearing interfer-ometer (peak-to-valleyphase = 0.71l). (70-00-0299-0365pb01)

12


UCRL-LR-105821-99-1

Preamplifier Beam Transport System

The final system in the ILS, or frontend, of the NIF laser is the PABTS, shownin Figure 11. The 16.9-J output from aPAM passes through an isolation modulethat contains a large-aperture Faradayrotator, half-wave plate, and polarizers toisolate and protect the front end from ahigh-energy pulse traveling backwardsfrom the main amplifier chain.

From the isolation module, the beamenters the 1:4 split assembly, where the sin-gle beam is split into four beams that willseed four separate main amplifier chains.The four-way split and balancing of poweramong the four legs is accomplished usingthin-film polarizers and half-wave plates.

Each of the beams passes through a six-element vacuum relay telescope that

relays the beam to the input relay plane ofthe transport spatial filter. This zoom tele-scope has variable magnification from 0.95to 1.02 and can be used to adjust the sizeof the beam in each of the four legs toaccommodate changes in the main ampli-fierÕs optics.

Next a timing section allows foradjustment of the timing in each leg bychanging the optical path length via atranslating mirror. A pair of turning mir-rors directs the beam either into thetransport spatial filter (TSF), or a finaltelescope prior to the TSF, dependingupon the side of the laser space frame inrelation to the main amplifiers.

Detailed mechanical and opticaldesigns of the PABTS are complete, andhardware prototyping is under way. One quad of a PABTS beamline will bebuilt in 1999.

FIGURE 11. The preamplifier beam transport system (PABTS); different subassemblies are designed to be line-replaceable units. (70-00-0299-0366pb01)

M3C

P1 P2 P3

P4

P5

M1D

HWP1

M1B

M4A

M5A

M5B

M4C

M5C

M4D

M5D

M4B

FR1

BD1 BD2 BD3

BD4

M2A

Legend:) Motorized tip/tilt (2 motors)) Motorized translation along Z-axis (1 motor)) Motorized rotation around Z-axis (1 motor)BRS) Back reflection sensorP) PolarizerHWP) Half-wave plateFR) Faraday rotatorM) MirrorL) LensBD) Beam dumpW) WindowQWP) Quarter-wave plateSH) Shutter

P7AHWP5A

SH1A

L1A L2A

W1AL3A L4A

L5A L6A

L5B L6B

L5C L6C

L5D L6D

M2B

P7BQWP1B

SH1BW1BL3B L4B

M2C

P7CQWP1C

SH1CW1C L3C L4C

M2D

P7DQWP1D HWP5D

SH1D

M3D

W1DL3D L4D

L1B L2B

L1C L2C

L1D L2D

HWP5B

HWP5C

M3A

BRS

BRS

BRS

BRS

Single input from PAM

Isolation module

1:4 split assembly

Vacuum relay telescopes

Timing slide assembly (4 places)

Timing injection assembly (2 places)

Turning mirror assembly (8 places)

Vacuum relay telescopes

Outputs seed 4-amplifier chains

HWP2

HWP3A

HWP4A

HWP3C

HWP4C

Verticalsplit

Verticalsplit

Horizontalsplit

P6A

P6C

QWP1A

M3B

SummaryThe ILS, or front end, of the NIF laser

system is nearing the end of its develop-ment phase. In the past year, we per-formed a series of experiments thatdemonstrated or exceeded the perfor-mance specifications listed in Table 1 forthe combined MOR and preamplifierdevelopment systems. Many of the subsys-tems described in this article and demon-strated in our development laboratories are now engineering prototypes. In thecoming year, we will assemble an engi-neering prototype of the entire ILS, includ-ing MOR, PAM, and PABTS, plus the com-puter controls, timing, and diagnostics sys-tems. We will then demonstrate the inte-grated operation and performance of theNIF laser system front end.

AcknowledgmentsWe thank the important contributions

made by several people in various keystages of the ILS development: Ray Beach,Scott Burkhart, Scott Mitchell, Mike Perry,Jim Davin, Brent Dane, Ralph Page, andNick Hopps.* In addition we would like toacknowledge the scientific, technical, anddesign contributions of the following peo-ple: Richard Hackel, Ernesto Padilla, ErnieDragon, Rod Lanning, Regula Fluck, DaveYoung, Jim Crawford, Rob Campbell, JohnBraucht, Ron Tilley, Marcus Bartlow, PeteLudwigsen, Mike Gorvad, Curt Laumann,Dave Wang, Dave Aikens, Bob Powers,Horst Bissinger, Steve Herman, Russ Jones,Brad Golick, Don Bartel, Jim Hauck, RonKorniski, Gloria Zuleta, Don Fleming,Greg Fischer, M. Reta, Kathy Coatney, WillHouse, Annette Springer, Cidelia Sanchez,Lori Dempsey, and Frances Mendieta.

*Atomic Weapons Establishment, Aldermaston, Berkshire, UK

Notes and References1. J. A. Paisner, J. D. Boyes, S. A. Kumpan, W. H.

Lowdermilk, and M. Sorem, ÒConceptual Designof the National Ignition Facility,Ó Proceedings ofthe First International Conference on Solid StateLasers for Application to Inertial ConfinementFusion, Monterey, CA, May 31ÐJune 2, 1995, SPIEProceedings Series 2633, 2.

2. S. Skupsky et al., J. Appl. Phys. 66, 3456Ð3462(1989).

3. R. B. Wilcox, D. Browning, G. Dreifuerst, E.Padilla, and Frank Penko, ÒDevelopment SystemPerformance of the NIF Master Oscillator andPulse-Forming Network,Ó Third InternationalConference on Solid State Lasers for Application toInertial Confinement Fusion, Monterey, CA, June1998.

4. A. Asseh et al., Electronics Letters 31(12) 969(1995).

5. J. R. Murray et al., J. Opt. Soc. Am. B 6, 2402Ð2411(1989).

6. R .B. Wilcox, D. F. Browning, and R. H. Page,ÒFiber DFB Lasers as Oscillators for High-PowerMOPA Systems,Ó Conference on Lasers andElectrooptics (CLEO), Baltimore, MD (1999).

7. S. C. Burkhart, R. B. Wilcox, D. Browning, and F.Penko, ÒAmplitude and Phase Modulation withWaveguide Optics,Ó Proceedings of the SecondInternational Conference on Solid State Lasers forApplications to Inertial Confinement Fusion, Paris,France, Oct. 22Ð25, 1996, SPIE Proceedings Series3047, 610Ð617.

8. J. K. Crane et al., ÒIntegrated Operations of theNational Ignition Facility (NIF) Optical PulseGeneration Development System,Ó ThirdInternational Conference on Solid State Lasers forApplication to Inertial Confinement Fusion,Monterey, CA, UCRL-JC-129871 (June 1998).

9. J. K. Crane et al., ICF Annual Report 1997,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-LR-105821-97, 246Ð259,(1998).

10. M. D. Martinez et al., ÒOptimized, Diode-Pumped, Nd:Glass, Prototype RegenerativeAmplifier for the National Ignition Facility,ÓProceedings on Optoelectronics and High-PowerLasers & Applications, San Jose, CA, Jan. 24Ð30,1998, SPIE Proceedings Series, vol. 3267, pp. 234Ð242.

11. N. W. Hopps et al., ÒOptimization of theAlignment Sensitivity and Energy Stability ofthe NIF Regenerative Amplifier,Ó ThirdInternational Conference on Solid State Lasers forApplication to Inertial Confinement Fusion,Monterey, CA, UCRL-JC-130961 (June 1998).

12. M. Henesian et al., Ò Diffraction Modeling of theNational Ignition Facility (NIF) Optical PulseGeneration System and Integration into the End-to-End System Model,Ó Third InternationalConference on Solid State Lasers for Application toInertial Confinement Fusion, Monterey, CA(June 1998).

13. J. M. Auerbach and V. P. Karpenko, AppliedOptics 33, 3179Ð3183 (1994).

14. B. D. Moran et al., ÒSuppression of Parasitics andPencil Beams in the High-Gain National IgnitionFacility Multipass Preamplifier,Ó Proceedings onOptoelectronics and High-Power Lasers &Applications, San Jose, CA, Jan. 24Ð30, 1998, SPIEProceedings Series, vol. 3264 , pp. 56Ð64.

15. L. M. Frantz and J. S. Nodvik, J. Appl. Phys. 34,2346Ð2349 (1963).

13


UCRL-LR-105821-99-1

15UCRL-LR-105821-99-1

Design of the power conditioning sys-tem (PCS) for the NIF laser is near-

ing completion. Modeling predicts that thedesign will meet all performance require-ments. A first-article NIF test module(FANTM), Figure 1, has been built atSandia National Laboratories (SNL) inAlbuquerque, New Mexico. This powerconditioning module is being tested to ver-ify the design performance and systemlifetime. NIFÕs power conditioning systemwill consist of 192 modules. The finaldesign of the module, to be built to drivethe NIF first bundle consisting of eightbeamlets, will be based on the informationand understanding gained from the opera-tion and testing done on the first-articlepower conditioning module.

The NIF PCS is being designed andbuilt by SNL in collaboration withLawrence Livermore National Laboratory(LLNL) and industrial partners. Its archi-tecture is different from any laser powerconditioning system previously built atLLNL and was chosen as the most cost-effective way to reliably deliver the largeamount of electical energy needed for NIF.

The design of the power conditioningsystems for flashlamp-pumped lasers rep-resents an evolution that has occurred overmany years and several generations oflasers.1Ð3 The direction of this evolutionhas been to build modules that can handle

larger amounts of energy with a smallernumber of components. NIFÕs forerunnersutilize small, independent modules thatstore less than 100 kJ. In comparison, NIFÕsPCS will be able to store nearly 2 MJ in asingle module. The move toward fewercomponents was driven by the need toreduce the cost of supplying energy todrive lasers of this size.

The specifications for the electrical per-formance of the PCS are derived from the

FIGURE 1. The first-article NIF test module (FANTM).(70-00-0399-0763pb01)

*Sandia National Laboratories, Albuquerque, NM**American Controls Engineering***Maxwell Physics International Consultant

MAIN AMPLIFIER POWER CONDITIONINGFOR THE NIF

M. Newton D. Smith* B. Moore* G. Ullery J. Hammon***

M. Wilson* D. Muirhead* T. Downey* S. Hulsey W. Gagnon

H. Harjes* D. Van DeValde* P. McKay* R. Anderson**

16

MAIN AMPLIFIER POWER CONDITIONING FOR THE NIF

UCRL-LR-105821-99-1

Damping�elements

Output

CapacitorsPower supply�

and controls

Ballast inductors

overall performance goals of the laser sys-tem. The PCS must also meet lifetime, reliability, and maintainability require-ments that are derived from the opera-tional requirements for the NIF laser sys-tem. The overall system performance willbe demonstrated through performanceand reliability testing using FANTM.

System PerformanceAn important measure of NIF laser per-

formance is the gain coefficient of the ampli-fier. Gain coefficient is a measure of thechange in light intensity vs the path lengthof the light through the laser glass. Gaincoefficient is dependent not only on thelaser glass, but also on the temporal charac-teristics of the light pulses generated by theflashlamps. The desired temporal character-istics of the flashlamp light translate directlyinto requirements on the shape, amplitude,and timing of the drive pulses supplied bythe power conditioning system.

A computer model developed by KenJancaitis of LLNL calculates the gain coef-ficient of the NIF amplifier for a givenelectrical drive input. This code, GainCalcv1.0, is used to verify that the outputwaveforms of the PCS meet the gain coef-ficient requirement of >5.0%/cm.

Power ConditioningModules

NIFÕs power conditioning equipmentwill be located in capacitor bays adjacentto each of four laser clusters. Forty-eightindependent power conditioning modules(PCMs) such as the one shown in Figure 2will drive the lasers in each cluster.

Each module will shape and deliverpulses of energy to NIF flashlamps. Adescription of an individual module isshown in Table 1.

A schematic of the module is shown inFigure 3. Each module is nominally com-prised of 20 nominally 83.5-kJ capacitorsconnected in parallel through damping ele-ments to limit fault currents. The energystored in a module is switched to the out-put transmission lines through a single gas-discharge switch. The peak current throughthis switch is approximately 500 kA.

Energy is evenly distributed to the 20 paral-lel flashlamp circuits through the ballastinductors and transmission lines. Theseflashlamps are configured as 20 circuits,with each circuit having two flashlamps inseries. The module will deliver nominally68 kJ of energy to each flashlamp pair dur-ing the main discharge pulse.

Each of the 192 modules will have amaximum energy storage capacity of

TABLE 1. Power conditioning module specifications.

Operating voltage 23.5 kV

Minimum delivered energy 34 kJper lamp

Number of lamps per module 40

Pulse duration 360 µs

Peak output current 530 kA

Nominal energy storage* 1.6 to 1.9 MJ

Number of capacitors 20 nominal, 24 maximum

*Modules must each deliver a minimum of 34 kJ per flash-lamp with easy expandability to approximately 40 kJ per flashlamp. To deliver 34 kJ per flashlamp, the module muststore 1.6 MJ; to deliver 40 kJ per flashlamp, the module must store 1.9 MJ.

FIGURE 2. Power conditioning module (the switch isnot shown) overall size is 10 ft high by 11 ft wide by 5 ft deep. (70-00-0399-0764pb01)

17


UCRL-LR-105821-99-1

Ð5

0

5

10

15

20

25

30

0 0.2 0.4 0.6 0.8 1.0Time (ms)

Cur

rent

(kA

)

Ð5

0

5

15

10

25

20

30

35

40

Voltage

Flashlamp�trigger Main�

discharge

Current

Vol

tage

(kV

)

Preionization pulse

Preionizationpower supply

Ballast inductors

Damping element

500-kA switch

C20C2C1

Powersupply

#20

#2

#1Flashlamps

Flexible cables

. . ....

nearly 2 MJ. The potential for faults tocause significant damage increases as theamount of energy in a single moduleincreases. Thus, significant investmentshave been made to develop componentsthat are either robust against all known fail-ure modes, or that fail in a well-controlledfashion and can be easily replaced. The sizeand configuration of the NIF power condi-tioning module represent an aggressive bal-ance between cost and risk.

Each module must deliver a very specificquantity of energy and waveshape to eachof the flashlamps. Each flashlamp must bepreionized with a small amount of energy(500 joules) approximately 300 microsec-onds before the main discharge to ensurebreakdown and to uniform plasma bore-filling of the flashlamps. A minimum of 34 kilojoules must be delivered to eachflashlamp during the main discharge.Typical current and voltage waveformsmeasured for energizing the flashlamps areshown in Figure 4. This pulse format shows

the features for preionization and triggeringin addition to the main discharge pulse.

CapacitorsSelf-healing, or metallized, dielectric

capacitors are used to store the energy fordriving the NIF flashlamps. Unlike film-foilcapacitors used in past laser systems, themetallized dielectric capacitors have a verypredictable lifetime. This characteristicallows the design of a capacitor with a meanlifetime tailored specifically to the NIFrequirement of 20,000 shots, without theadditional cost of excessive design margins.

A summary of the NIF capacitorrequirements is shown in Table 2.

The cost of the 3,840 capacitors that NIFwill require constitutes 30 to 40 percent ofthe total power conditioning system cost,the systemÕs largest single cost component.Capacitor reliability will also have a signifi-cant impact on the overall system reliability.Significant resources have been invested to

FIGURE 3. Power condi-tioning module schematic. (70-00-0499-0826pb01)

FIGURE 4. (a) Flashlampdrive current (black) and(b) voltage across flash-lamp (gray). (70-00-0399-0762pb01)

18


UCRL-LR-105821-99-1

develop capacitors that carefully balancereliability and cost.

Capacitor development for NIF begannearly three years ago, when LLNL beganworking with multiple vendors to developcapacitors that would meet NIF specifica-tions. After multiple iterations of design, fab-rication, and testing, several vendors(Thomson Passive Components, ICAR,Aerovex, and Maxwell Technologies) havebuilt capacitors that meet the NIF require-ments. Development of lower-cost capacitordesigns will continue until 1999, when LLNLbegins the final design-qualification testing.

Capacitor-Charging PowerSupplies

Each module will be charged by its owncapacitor-charging power supply. The powersupply specifications are shown in Table 3.

We have been working with three ven-dors (Dell Global Technologies, EMI, andMaxwell Technologies) to develop thesepower supplies. The specifications forthese supplies are very close to those ofstandard off-the-shelf supplies, with theexception of the regulation and stabilityrequirements. Prototype supplies havebeen built by the three vendors and havebeen operated and tested on the first-article power conditioning module.

Each power supply must charge its mod-uleÕs capacitors to 24 kV in a maximum of80 s. This charge time was chosen because itwas an acceptable compromise betweencapacitor lifetime, switch prefire probability,and power supply/prime power cost. Once

a module is charged, it will be held at itscharge voltage for a maximum of 15 s whileother systems in NIF are synchronized andarmed for the shot.

InductorsThere are two different types of induc-

tors in each module. The ballast inductorensures an equal distribution of deliveredenergy to each flashlamp circuit. The damp-ing element limits fault currents in themodule resulting from capacitor failuresand short circuits across the bus. Thedesign of both inductors is driven by a gen-eral philosophy that all of the inductorsmust remain intact during any of the possi-ble fault conditions. However, the individu-al inductor(s) affected by the fault will bereplaced before the next shot. Efforts areunder way to optimize both designs tominimize the cost while ensuring that theinductors will meet the fault requirements.

The stored energy is discharged througha high-power spark-gap switch (see the sec-tion on switches below) into a distributionnetwork of 20 ballast inductors and trans-mission lines. The purpose of the ballastinductors is to ensure that the moduleÕsenergy is delivered equally to all of the 40flashlamps that it must power.

The two primary requirements of bal-last inductors are:

¥ To provide current shaping over theNIF lifetime requirement of 20,000shots.

¥ To withstand fault currents of 100 kA.

We have successfully demonstrateddesigns that will meet these requirements.Developments are continuing to lower thecost of these units.

TABLE 2. Capacitor specifications.

Type Metallized dielectric

Nominal capacitance 290 µF (+10%, Ð5%)

Operating voltage 24 kV

Peak discharge current 30 kA

Voltage reversal <10%

Lifetime 20,000 shots

Maximum size 36" ´ 18" ´ 18"

Duration of discharge 360 µs

TABLE 3. Power supply specifications.

Maximum operating voltage 26 kV

Charge rate 26 kJ/sec

Regulation ±0.05%

Shot-to-shot stability ±0.05%

19


UCRL-LR-105821-99-1

The second type of inductor, a resistiveinductor or damping element, is connectedin series with each capacitor. The purposeof this damping element is to limit faultcurrents in the event that one of the capac-itors fails internally, or in the event that themoduleÕs output bus is short-circuited. Tominimize damage, the damping elementsmust be able to withstand a fault currentof 400 kA. Considerable effort is beingexpended to ensure reliability of this sys-tem component. Damping element designshave been developed that will meet therequirements, but efforts are ongoing toÒvalue engineerÓ this component as well.

SwitchesEach module must transfer the 2 MJ of

energy stored in its capacitors to 40 flash-lamps. It does this through a switch, whichmust meet the following criteria:

¥ Be highly reliable (no prefires).¥ Have a relatively long lifetime. ¥ Conduct 530 kA. ¥ Operate at 24 kV.

SNL fabricated a test stand to test vari-ous types of switches to identify one thatwould meet these requirements. The typesof switches tested include: spark-gapswitch; ignition switch (a mercury-filledarc switch); rotating arc-gap switch; andtwo solid-state switches.

Maxwell/Physics InternationalÕs Model ST300, a spark-gap switch shown inFigure 5, was selected on the basis of itsperformance and relatively low initial cost,even though its operating cost will behigher. Each ST300 will need to bereplaced every 1,500 to 2,000 shotsthroughout NIFÕs anticipated lifetimerequirement of 20,000 shots. A decisionwas made to utilize this switch instead of arotating arc-gap switch, which remains analternate. Although insufficient testing hasbeen done on the rotating arc-gap switch,it appears that this type of switch wouldmeet the NIF lifetime requirement.However, its cost is five to ten times thatof the spark-gap switch Model ST300.

Data from SNLÕs multiple lifetime test-ing of the Model ST300 spark-gap switch,which has proven to be accurate and

repeatable, was used to predict whenswitches would need to be replaced orrefurbished.

Controls and DiagnosticsEach of NIFÕs 192 power conditioning

modules will have its own embedded con-troller, which will control module func-tions including:

¥ Charging of the main capacitors.¥ Gas-purging of the switch.¥ Charging the preionization circuit.¥ Charging and firing of trigger circuits.

This embedded controller also servesdiagnostics functions. Twelve-bit digitizersare used to collect waveform informationfrom the module during a shot. The con-troller collects information, including 20output current waveforms as well as reflec-tor ground current and facility ground cur-rent. It collects power-supply-chargingwaveform information from the main

FIGURE 5. Spark-gap switch. (70-00-0499-0827pb01)

20


UCRL-LR-105821-99-1

capacitor bank and the preionization capac-itor bank. It monitors the waveforms toensure that they are correct, and it informsthe operator if a shot was acceptable. Ifthere were problems inherent in the shot,the controller pinpoints the problem andalerts the operator.

Low-energy preionization lamp check(PILC) shots evaluate the health of all the flashlamps in the NIF system. Theembedded controller evaluates waveforminformation to detect any broken flash-lamps. This is a critical step that mustprecede flashlamp cooling, a procedurethat could scatter broken glass and otherdebris throughout the amplifier andpotentially cause significant damage tothe amplifier and other laser systemcomponents. If no abnormalities aredetected during the PILC, the controllerwill communicate the status to the supervisory control system, which willthen initiate flashlamp cooling.

The ÒFANTMÓThe first-article NIF test module, or

FANTM, represents the entire power con-ditioning system from Òthe wallplug,Ó orAC power source, to the flashlamps. Itwas built in October 1998 to test andevaluate power conditioning modulemechanical design and electrical perfor-mance. Since then, it has been operatedand tested at SNL Albuquerque, collect-ing data to ensure that the module meetsall NIF requirements, including fault tol-erance and overall reliability. It also hasbeen used to evaluate design improve-ments and to estimate time and effortrequired to build PCMs and install themin the NIF laser system.

Resistors initially were used to simulateflashlamp loads during activation and ini-tial operation. In January 1999, NIF flash-lamps were installed at Sandia as a loadfor the first-article module. Testing isunder way to fully characterize moduleperformance and reliability when drivingNIF flashlamps.

AcknowledgmentsThis article describes the contributions

of many people, too numerous to list as co-authors. People who have made contri-butions to the designs and calculationsdescribed in this paper are: Jake Adcock,Gary Mower, and Dean Rovang (SandiaNational LaboratoriesÐAlbuquerque); Steve Fulkerson, Doug Larson, and ChetSmith (Lawrence Livermore NationalLaboratory); Bob Anderson (AmericanControls Engineering); Jud Hammon(Maxwell/ Physics International); and Bill Gagnon (consultant).

Notes and References1. K. Whitham, D. Larson, B. Merritt, and D.

Christie, ÒOVA Pulse Power Design andOperational Experience,Ó Pulsed Power for Lasers,in Proc. SPIE (Society of Photo-opticalInstrumentation Engineers, Los Angeles, CA,Jan. 13Ð14, 1987), 735, 12Ð17.

2. D. Larson, R. Anderson, and J. Boyes, ÒFaultTolerance of the NIF Power ConditioningSystem,Ó in Proc. of the First Annual Conference onSolid State Lasers for Application to InertialConfinement Fusion (Society of Photo-opticalInstrumentation Engineers, Bellingham, WA,May 30-June 2, 1995), 2633, 587Ð595.

3. W. L. Gagnon, ÒThe Near and Long Term PulsePower Requirement for Laser-Driven InertialConfinement Fusion,Ó in Proc. of the Second IEEEInternational Pulsed Power Conference (Institute ofElectrical and Electronics Engineers, Lubbock,TX, June 12Ð14) 1979, 49Ð54.

21UCRL-LR-105821-99-1

The National Ignition Facility (NIF)design team is developing the inte-

grated computer control system (ICCS),which is based on an object-oriented frame-work applicable to event-driven controlsystems. The framework provides an open,extendable architecture that is sufficientlyabstract to construct future mission-criticalcontrol systems. Supervisory software isconstructed by extending the reusableframework components for each specificapplication. The framework incorporatesservices for database persistence, systemconfiguration, graphical user interface, sta-tus monitoring, event logging, scriptinglanguage, alert management, and accesscontrol. More than twenty collaboratingsoftware applications are derived from thecommon framework.

The ICCS consists of 300 front-end pro-cessors (FEPs) attached to 60,000 controlpoints coordinated by a supervisory sys-tem. Computers running either Solaris orVxWorks operating systems are networkedover a hybrid configuration of switchedfast Ethernet and Asynchronous TransferMode (ATM). ATM carries digital motionvideo from sensors to operator consoles. Abrief summary of performance require-ments follows (Table 1).

The ICCS architecture was devised toaddress the general problem of providing

distributed control for large scientificfacilities that do not require real-timesupervisory controls. The framework uses the client-server software modelwith event-driven communications to distribute control. The framework is inter-operable among different kinds of com-puters and transparently distributes thesoftware objects across the network byleveraging a Common Object RequestBroker Architecture (CORBA).

INTEGRATED COMPUTER CONTROL SYSTEM

P. J. VanArsdall R. A. Saroyan

R. C. Bettenhausen J. P. Woodruff

F. W. Holloway

TABLE 1. Selected ICCS performance requirements.

Requirement Performance

Computer restart <30 minutes

Postshot data recovery <5 minutes

Respond to broad-view <10 secondsstatus updates

Respond to alerts <1 second

Perform automatic alignment <1 hour

Transfer and display digital 10 frames permotion video second

Human-in-the-loop controls within 100 msresponse

22


UCRL-LR-105821-99-1

CORBA DISTRIBUTION

Past architectural approaches to distributed controls have relied on the tech-nique of building large-application programming interface (API) libraries to giveapplications access to functions implemented throughout the architecture. Thispractice results in large numbers of interconnections that quickly increases the sys-tem complexity and make software modification much more difficult. To addressthis problem in the ICCS, software objects are distributed in a client-server archi-tecture using CORBA.

CORBA is a standard developed by a consortium of major computer vendorsto propel the dominance of distributed objects on local area networks and theWorld Wide Web. The best way to think of CORBA is as the universal Òsoftwarebus.Ó CORBA is a series of sophisticated, but standard sockets into which softwareobjects can Òplug and playÓ to interoperate with one another, even when made bydifferent vendors. By design, CORBA objects interact across different languages,operating systems, and networks.

At a greatly simplified level, the major parts of CORBA are shown in the figure below. The interface types and methods provided by the server objects andused by the clients are defined by an industry standard Interface DefinitionLanguage (IDL). The IDL compiler examines the interface specification and gener-ates the necessary interface code and templates into which user-specific code isadded. The code in the client that makes use of CORBA objects is written as if theserver was locally available and directly callable. Each computer on the networkhas an object request broker that determines the location of remote objects andtransparently handles all communication tasks.

�

Software bus

(network)

Server objects�(FEP)

Object�request�broker

Object�request�broker

Server software�templates

Client software�templates

Interface Definition�Language (IDL)

IDL compiler

Client application�(GUI)

Traditional client/server�interactions

�Software bus network implementation for CORBA distribution.(40-00-0298-0341pb01)

23


UCRL-LR-105821-99-1

FIGURE 1. Photographof the NIF computercontrol room.(40-00-1298-2628pb01)

Over twenty distributed software appli-cations built from the common frameworkwill operate the NIF control system hard-ware from a central control room (Figure 1).The ICCS software framework is the key tomanaging system complexity and, becauseit is fundamentally generic and extensible,it is also reusable for the construction offuture projects.

Control SystemArchitecture

The ICCS is a layered architecture con-sisting of FEPs coordinated by a superviso-ry system (Figure 2). Supervisory controls,which are hosted on UNIX workstations,provide centralized operator controls andstatus, data archiving, and integration ser-vices. FEP units are constructed fromVME/VXI-bus or PCI-bus crates of embed-ded controllers with interfaces that attach tocontrol/monitor points (e.g., motors andcalorimeters). FEP software implements thedistributed services needed to operate thehardware by the supervisory system.Precise triggering of fast diagnostics andcontrollers is handled during a two-secondshot interval by the timing system, which iscapable of providing triggers to 30-ps accu-racy and stability. The software is distribut-ed among the computers and providesplug-in software extensibility for attachingcontrol/monitor points and other softwareservices by using the CORBA protocol.

The operator console provides thehuman interface in the form of operator dis-plays, data retrieval and processing, andcoordination of control functions.Supervisory software is partitioned into sev-eral cohesive subsystems, each of whichcontrols a primary NIF subsystem such asalignment or power conditioning. Severaldatabases are incorporated to manageexperimental and operations data. The sub-systems are integrated to coordinate opera-tion of laser and target area equipment.

FEPs perform sequencing, data acquisi-tion, process control, and data reduction. Thesoftware framework includes a standard wayfor FEP units to be integrated into the super-visory system by providing the common dis-tribution mechanism coupled with softwarepatterns for hardware configuration, com-mand, and status monitoring functions.

A segment of the control system accom-modates industrial controls for which stan-dard commercial solutions already exist.The segment is composed of a network ofprogrammable logic controllers that residebelow the FEP and attach to field devicesthat functionally control, for example, vac-uum systems, argon gas in the beam tubes,and thermal gas conditioning for amplifiercooling. This segment also observes theindependent safety interlock system, whichmonitors doors, hatches, shutters, andother sensors to establish and display thehazard levels in the facility. Potentially haz-ardous equipment is permitted to operateonly when conditions are safe. Interlocks

24


UCRL-LR-105821-99-1

function autonomously to ensure safetywithout dependency on the rest of the con-trol system.

There are eight supervisory softwareapplications that conduct NIF shots in col-laboration with 19 kinds of FEPs as shownin Figure 3. The ICCS is partitioned intoseveral loosely coupled systems that areeasier to design, construct, operate, andmaintain. Each subsystem is composed ofa supervisor and associated FEPs. Theeighth supervisor is the shot director,which is responsible for conducting theshot plan, distributing the countdownclock, and coordinating the other sevensupervisory applications.

Seven subsystems comprise the primary supervisory controls. TheAlignment Supervisor provides coordina-tion and supervision of laser wavefrontcontrol and laser component manual and automatic alignment. The LaserDiagnostics Supervisor provides functions

for diagnosing performance of the laser bycollecting integrated, transient, and imageinformation from sensors positioned inthe beams. The Optical Pulse GenerationSupervisor provides temporally and spa-tially formatted optical pulses with thecorrect energetics and optical characteris-tics required for each of the beams. TheTarget Diagnostics Supervisor coordinatesthe collection of data from a diverse andchanging set of instruments. The Power-Conditioning Supervisor is responsible forhigh-level control and management ofhigh-voltage power supplies that fire themain laser amplifiers. The Pockels CellSupervisor manages operation of the plas-ma electrode Pockels cell optical switchthat facilitates multipass amplificationwithin the main laser amplifiers. The ShotServices Supervisor provides monitoringof environmental and safety parameters aswell as control of programmable logiccontroller subsystems.

Operator consoles

Switchednetwork

Database

File servers

Actuators DataAcquisition Utilities Safety

interlocks

PowerPC-basedVME/VXI

cratesProgrammablelogic controllers

Sun SolarisUltraSparc workstations

Front-end processors Industrial controls

Intel-based Windows NT

Timing systems

FIGURE 2. Integratedcomputer control systemarchitecture.(40-00-1298-2627pb01)

25


UCRL-LR-105821-99-1

Computer System and Network

Figure 4 shows the NIF computer andnetwork architecture, which is composedof 30 UNIX workstations, 300 FEPs, andseveral hundred more embedded con-trollers (not shown). The main controlroom contains seven graphics consoles,each of which houses two workstationswith dual displays. The supervisors arenormally operated from a primary console,although the software can easily be operat-ed from adjacent consoles or remotegraphics terminals located near the front-end equipment. File servers providearchival databases as well as centralizedmanagement services necessary for coordi-nating the facility operation.

The network design utilizes both Ethernetand ATM technologies to take advantage ofthe best features of each. ATM providestime-sensitive video transport, whereasEthernet provides connectivity for the largemajority of systems. The design utilizes

155 Mbit/s ATM and Ethernet at both 10and 100 Mbit/s speeds, depending onexpected traffic requirements. The operatorconsoles and file servers have both ATM and100 Mbit/s Ethernet connections, while mostFEPs have either 10 or 100 Mbit/s connec-tions made through Ethernet switches.Given the low cost and performance advan-tages of Ethernet switches relative to sharedEthernet hubs, switches with 100 Mbit/suplinks will be used.

TCP/IP (Transmission Control Protocol/Internet Protocol) is the protocol used forreliable data transport between systems,either over Ethernet or ATM. TCP providesretransmission of packets in the event thatone is lost or received in error. The onlytraffic not using TCP will be digitized videoand network triggers. Video is transferredusing the ATM adaptation layer 5 (AAL5)protocol. Network triggers are broadcast tomany end-nodes simultaneously usingmulticast protocols.

The network supports the transport ofdigitized motion video in addition to themore typical control, status, and shot data.

Shot integration

Supervisory subsystems

Application FEPs

Service FEPs

Shotdirector

Targetdiagnostics

Laserdiagnostics

PowerconditioningAlignment

Opticalpulse

generation

Switchpulser

LaserenergyWavefront

Masteroscillator

Targetdiagnostics

Powerconditioning

Plasmapulser

Laserpower

Automaticalignment

Industrialcontrols

Preamplifiermodule

Precisiondiagnostics

Hartmannimage

processor

Alignmentcontrols

Digitalvideo Timing

Pockels cellShot

services

Beamtransport

Pulsediagnostics

High-resolution

CCD

FIGURE 3. Softwareapplications in the NIFcontrol system.(40-00-0298-0324pb02)

26


UCRL-LR-105821-99-1

The network transports video streams at 10 frames per second (about 25 Mbit/s)between special FEPs that digitize cameraimages and operator workstations that candisplay at least two concurrent streams.Video compression is not used because ofthe high cost of encoding the video stream.

Digitized video is sent via the ATM appli-cation program interface (API) using theATM quality of service capabilities. The APIprovides an efficient method of movinglarge, time-sensitive data streams, resultingin higher frames/second rates with lowercentral processing unit (CPU) utilizationthan alternative approaches, which is animportant consideration for the video FEPsand console workstations. Performance test-ing of the prototype video distribution sys-tem indicates that 55% of the FEP CPU (300-MHz UltraSparc AXI) is used to broad-cast three streams, while 10% of the operatorworkstation CPU (300-MHz UltraSparc 3DCreator) is utilized for each playback stream.

Estimates of the peak traffic require-ments for the various subsystems wereanalyzed as a basis for the network design.The expected peak traffic flows betweensubsystems in terms of messages per sec-ond and message size were specified. Thisdata was combined and analyzed to deter-mine peak throughputs into and out ofeach network-attached device.

A discrete event simulation was createdto evaluate the performance of key net-work scenarios. For example, it was shownthat Ònetwork triggersÓ in the millisecondregime could be reliably broadcastbetween computers in the network via theuser datagram protocol. One applicationusing network triggers is the arming of thevideo FEPs to enable digitizing the laserpulse during shots. The network triggertactic will save $250K over a hardwareimplementation.

If bandwidth requirements increase inthe future, the network architecture allows

Two dual-monitorworkstations per console

Remoteworkstations

Eight supervisory consoles

Front-endprocessors

Front-endprocessors

Users &external

databases

Firewall

Core100-Mbit/s

Ethernetswitch

Automaticalignment

servers

24-cameravideo

digitizers

EdgeEthernetswitch

EdgeEthernetswitch

ATM155-Mbit/s

switch

Servers

13 edge switches

300 FEPs2150 loops 500 CCD cameras

) Legend

ATM OC-3 155 Mbit/s (fiber)

Ethernet 100 Mbit/s

Ethernet 10 and 100 Mbit/s

FIGURE 4. NIF comput-er system and networkarchitecture.(40-00-1298-2629pb01)

27


UCRL-LR-105821-99-1

the integration of Gbit/s Ethernet and 622 Mbit/s ATM technologies in a relative-ly straightforward manner.

Software FrameworkThe ICCS supervisory software frame-

work is a collection of collaborating abstrac-tions that are used to construct the applica-tion software. Frameworks1 reduce theamount of coding necessary by providingprebuilt components that can be extended toaccommodate specific additional require-ments. The framework also promotes codereuse by providing a standard model andinterconnecting CORBA backplane that isshared from one application to the next.

Components in the ICCS framework aredeployed onto the file servers, workstations,and FEPs, as shown generically in Figure 5.Engineers specialize the framework for eachapplication to handle different kinds of con-trol points, controllers, user interfaces, andfunctionality. The framework conceptenables the cost-effective construction of theNIF software and provides the basis forlong-term maintainability and upgrades.

The following discussion introduces theframework components that form the basisof the ICCS software.

ConfigurationÑa hierarchical organiza-tion for the static data that define the hard-ware control points accessible to the ICCS.Configuration provides a taxonomic system

Dual servers Workstation

Network

Ethernet ATM

Workstation

Solaris

Integration services• Configuration server• Central system) manager• Message log server• Alert manager• Central reservation) manager• Machine history) archiver• Shot data archiver• Shot setup server

Sample application• Local system manager• Application object) factories• Application function) objects• Application GUI model• Alert subscriber• Message log subscriber• Local reservation) manager - Application only -• Shot life-cycle subscriber• Local countdown

Sample front-end processor• Local system manager• Device/controller object factories• Device objects) Ð Commands) Ð Status) Ð Messages) Ð Alerts) Ð History• Controller objects• Local reservation manager• Status monitor• FEP function objects

OracleDMBS

User interface• Status displays• Control panels

Shot director• (Sample application) plus• Central countdown• Shot life-cycle publisher

Solaris VxWorks

Solaris

Solaris

Emulateddevice

Realdevice

User interface• Status displays• Control panels

FIGURE 5. Deploymentof ICCS frameworkobjects into a sampleapplication and FEP onnetworked computers.(40-00-1298-2630pb01)

28


UCRL-LR-105821-99-1

used as the key by which clients locatedevices (and other software services) on theCORBA bus. During normal operation, con-figuration provides to clients the CORBAreferences to all distributed objects. Animportant responsibility of configuration isthe initialization of FEPs during start-up.Configuration data are stored in thedatabase and describe how and where thecontrol hardware is installed in the system.Calibration data for sensors, setpoints foralignment devices, and I/O channels usedby devices on interface boards are examplesof static data managed by configuration.During ICCS start-up, this framework col-laborates with an object factory located inthe FEP. Using the data and methods storedin the configuration database, the object fac-tory instantiates, initializes, and determinesthe CORBA reference for each device andcontroller object in the FEP.

Status MonitorÑa device that providesgeneralized services for broad-view opera-tor display of device status informationusing the publisherÐsubscriber model ofevent notification. The status monitoroperates within the FEP observing devicesand notifies other parts of the systemwhen the status changes by a significantamount. Network messages are only gen-erated when changes of interest occur.

Sequence Control LanguageÑa lan-guage used to create custom scripting lan-guages for the NIF applications. The serviceautomates sequences of commands execut-ed on the distributed control points or othersoftware artifacts. After full implementationof the design, operators will create and editsequences by selecting icons that representcontrol constructs, Boolean functions, anduser-supplied methods from a visual pro-gramming palette. The icons are then inter-connected to program the sequence, andany Boolean conditions or method argu-ments needed are defined to complete thesequence script.

Graphical User Interface (GUI)Ñthemethod through which all human interac-tion with the ICCS will take place andwhich will be displayed on control roomconsoles or other workstations in the facility.The GUI is implemented as a framework toensure consistency across the applications.

Commercial GUI development tools areused to construct the display graphics. Thisframework consists of guidelines for lookand feel as well as many common graphicalelements.

Message Log ServerÑa device thatprovides event notification and archivingservices to all subsystems or clients withinthe ICCS. A central server collects incom-ing messages and associated attributesfrom processes on the network, writesthem to appropriate persistent stores, andalso forwards copies to interestedobservers such as GUI windows.

Alert System ManagerÑa componentthat raises an alert for any applicationencountering a situation that requiresimmediate attention, which then requiresinteraction with an operator for the controlsystem to proceed. The alert systemrecords its transactions so that the data canbe analyzed after the fact.

Reservation ManagerÑa system thatmanages access to devices by giving oneclient exclusive rights to control or other-wise alter a control/monitor point. Theframework uses a lock-and-key model.Reserved devices that are ÒlockedÓ canonly be manipulated if and when a clientpresents the Òkey.Ó

System ManagerÑa component thatprovides services essential for the integrat-ed management of the ICCS computer net-work. This component ensures that neces-sary processes and computers are operat-ing and communicating properly. Servicesinclude orderly system start-up, shut-down, and watchdog process monitoring.

Machine History ArchiverÑa systemthat gathers information for analysis aboutNIF operational performance to improveefficiency and reliability. Examples of suchinformation are component adjustments,abnormal conditions, operating service,periodic readings of sensors, and referenceimages.

Generic FEPÑa component that pullstogether the distributed aspects of theother frameworks (in particular the systemmanager, configuration, status monitor,and reservation frameworks) by addingunique classes for supporting device andcontroller interfacing. These classes are

29


UCRL-LR-105821-99-1

responsible for hooking in CORBA distri-bution as well as implementing the cre-ation, initialization, and connection ofdevice and I/O controller objects. Thegeneric FEP also defines a common hard-ware basis including the processor archi-tecture, I/O board inventory, devicedrivers, and field-bus support. The FEPapplication developer extends the basesoftware classes to incorporate specificfunctionality and control logic.

Shot Data ArchiverÑa server workingin collaboration with the System Managerto assure that requested shot data aredelivered to a disk staging area. The ICCSis responsible for collecting the data fromdiagnostics, making the data immediatelyavailable for Òquick lookÓ analysis, anddelivering the data to an archive.

Shot Setup ServerÑa server workingin collaboration with the ICCS shot direc-tor to manage shot setup plans. Theseplans contain the experimenterÕs goals

for a shot and the hardware setup derivedfrom the goals.

Software DevelopmentEnvironment

The ICCS incorporates Ada95, CORBA,and object-oriented techniques to enhancethe openness of the architecture and long-term maintainability of the software. C++ isalso supported for the production of graphi-cal user interfaces and the integration ofcommercial software. Software developmentof an expected 500,000 lines of code is man-aged under an integrated software engineer-ing process (Figure 6) that covers the entirelife cycle of design, implementation, andmaintenance. The object-oriented design iscaptured in the Rose design tool in UML(unified modeling language) notation thatmaintains schematic drawings of the soft-ware architecture.

Requirementspecification

Designdescription

Object-orienteddesign tool

Objectmodel

NIFsoftware

Ada languageeditor

Engineers writecode details

Engineers modelsoftware framework

Interfacespecification

Ada hostcompiler

Ada crosscompiler

Unixtarget

Real-timetarget

Targetarchitectures

Object request brokerdistribution

Architecture neutral

Automaticcode generationof specifications

VxWorks PowerPCSun Sparc

Reverseengineering

Reverseengineering

FIGURE 6. Flowchart ofthe ICCS software engi-neering process incorpo-rating model-drivendesign techniques.(40-00-1298-2631pb01)

30


UCRL-LR-105821-99-1

Developers analyze detailed requirementsand express them as scenarios that helpdetermine how the software will be orga-nized. In essence, the Rose tool is used tomodel the public interfaces and interactionsbetween major software entities (i.e., theclasses). Rose automatically generates Adacode specifications corresponding to theclass interface and type definitions. Thedeveloper fills in the detailed coding neces-sary to implement the private contents ofeach class. Rose generates IDL for classesthat are distributed, which is passed throughthe IDL compiler to generate Ada or C++skeleton code as before. The design descrip-tion is a narrative document that explainsthe object-oriented model and contains otherinformation necessary for implementationand maintenance.

Source code is compiled for a varietyof target processors and operating sys-tems. Current development is either self-hosted to Solaris on Sparc processors

or cross-compiled for VxWorks onPowerPC processors. The models, sources,binaries, and run-time images are version-controlled by the Apex configurationmanagement system, which allows theframeworks and applications to be independently developed by differentengineers, each having a protected view of the other components.

SummaryThe ICCS is being developed using the

iterative approach to software construction2

that is proven effective for projects whoserequirements continue to evolve. Five itera-tions in the ÒLightÓ series (Figure 7) areplanned leading to facility deployment in2000 when the first 8 of the 192 beams willbe operated. Each new release will followan updated plan addressing the greatestrisks to the architecture while increasing thefunctionality delivered to the Project.

Sunlight

Searchlight

Spotlight

Torchlight

Flashlight

Penlight

Nightlight

Full�system

Full�automation

Hardware�integration

Shot�coordination

Framework�enhancements

Production�prototypes

Production�releases

Sep 1999 Apr 2000 Sep 2000 Sep 2002

First bundle First bay NIF

Oct 2003

Cooperative�actions

Apr 1999

Vertical�slices

Tied to first-�bundle availability

Oct 1998

Key

Code�name

Theme

Due date

FIGURE 7. Light seriesof NIF software construction.(40-00-1298-2632pb01)

31


UCRL-LR-105821-99-1

The first release, ÒNightlight,Ó was com-pleted this fall and submitted to indepen-dent test. Nightlight delivered 120,000source lines of tested codeÑabout 20% ofthe anticipated totalÑwhich exercised theframework in initial vertical slices of allsupervisory and FEP applications.Nightlight also established an independenttest process and the procedures by whichquality software could be assured.

Construction of the ICCS incorporatesmany of the latest advances in distributedcomputer and object-oriented softwaretechnology. Primary design goals are toprovide an extensible and robust architec-ture that can be maintained and upgradedfor decades. Software engineering is amanaged process utilizing model-drivendesign to enhance the product quality andminimize future maintenance costs. TheICCS framework approach permits soft-ware reuse and allows the system to beconstructed within budget. As an addedbenefit, the framework is sufficientlyabstract to allow future control systems totake advantage of this work.

AcknowledgmentsThe authors wish to acknowledge the

contributions of the following colleagueswithout whose efforts this work would notbe possible: G. Armstrong, R. Bryant, R.Carey, R. Claybourne, T. Dahlgren, F.Deadrick, R. Demaret, C. Estes, K. Fong,M. Gorvad, C. Karlsen, B. Kettering, R.Kyker, L. Lagin, G. Larkin, G. Michalak, P.McKay (Sandia National Laboratories,Albuquerque, NM), M. Miller, V. MillerKamm, C. Reynolds, R. Reed, W. Schaefer,J. Spann, E. Stout, W. Tapley, L. Van Atta,and S. West.

Notes and References1. V. Swaminathan and J. Storey, Object Magazine,

April 1988, pp. 53Ð57. 2. B. Boehm, IEEE Computer, May 1988, pp. 61Ð72.

33UCRL-LR-105821-99-1

Polarizer

Optical switch

Pockels cell

LM1

LM3

Transportspatial filter

11-slabamplifier

Injection pulse

7-slab amplifier

LM2

Mainamplifier

cavity

This article discusses the optical switchsystem for the National Ignition

Facility (NIF). The NIF laser architecture isbased on a multipass power amplifier toreduce cost and maximize performance. Akey component in the laser design is anoptical switch that ÒclosesÓ to trap theoptical pulse in the amplifier cavity forfour gain passes and then ÒopensÓ todivert the optical pulse out of the amplifiercavity. The optical switch consists of a full-aperture Pockels cell that provides voltagecontrol of the beam polarization and areflectingÐtransmitting polarizer.

Figure 1 shows one of NIFÕs 192 beam-lines. During a NIF shot, an optical pulseenters the cavity from the transport spatialfilter by reflecting off the polarizer. Afterthe optical pulse passes through thePockels cell, a precisely timed voltagepulse is applied to the Pockels cell so thatthe beam polarization is rotated 90¡ when

it returns after reflecting from the LM1mirror. The optical pulse then passesthrough the polarizer, reflects off the LM2mirror, and passes back through the polar-izer. Because voltage is still applied to thePockels cell, the beam polarization is rotat-ed back to its original polarization on thisthird pass through the Pockels cell. Afterthe optical pulse passes, the voltage pulsestarts to turn off. By the time the opticalpulse returns again, no further rotationoccurs. The beam reflects off the polarizerand exits the cavity, having passedthrough the main amplifier four times.

Optical switches are common in manytypes of lasers, such as Q-switched lasersand regenerative lasers. However, the size(40 ´ 40 cm), shape (square), and energydensity (5 J/cm2) of the NIF beams requirean optical switch of unprecedented pro-portions. Commercially available, conven-tional Pockels cells do not scale to such

MULTIAPERTURE OPTICAL SWITCHFOR THE NIF

M. Rhodes C. Ollis

R. Bettenhausen J. Trent

P. Biltoft

FIGURE 1. One of 192beamlines of the NIF,showing the location ofthe optical switch system.(70-00-0299-0367pb01)

34

MULTIAPERTURE OPTICAL SWITCH FOR THE NIF

UCRL-LR-105821-99-1

large apertures or the square shape requiredfor close packing. The NIF optical switch isbased on the plasma-electrode Pockels cell(PEPC). This article first reviews PEPC tech-nology, then discusses the details of the NIFoptical switch system, including the linereplaceable unit (LRU), support structure,vacuum and gas systems, required assem-bly hardware, electronic systems required tooperate the optical switch, and the comput-er control system.

Plasma Electrode PockelsCell Technology

In any Pockels cell, polarization rotationof a transmitted laser beam occurs whenan electric field is applied to an electro-optic crystal, inducing birefringence. In aconventional Òring-electrodeÓ Pockels cell,the electric field is applied to a cylindricalrod of crystal via rings of metal depositedon the end faces of the rod. To obtain suffi-cient field uniformity across the activeaperture, the length of the rod must be oneto two times the diameter. This require-ment poses a problem for Pockels cellswith apertures larger than about 10 cm.

In a PEPC, the electric field is applied ina different way. A cloud of ionized gas(plasma) is formed by electric discharge onboth sides of the crystal. The plasma iselectrically conductive and transparent tothe incoming laser beam. The plasma coatsboth sides of the crystal, allowing directapplication of the required voltage (about17,000 V in our case). The electric field isuniform across the entire aperture regard-less of aperture shape or size.

PEPC technology was first invented at LLNL in the 1980s, but focused development of a PEPC to be used on a high-energy laser began in 1990. We pro-gressed from a 27-cm-aperture PEPC to a32-cm PEPC, then to the 37-cm PEPC that became a working part of theBeamlet laser (the scientific prototype forNIF). With the success of Beamlet and theBeamlet PEPC, we began working on aPEPC for NIF in 1995. The focus was tofind a way to closely pack multiple aper-tures to meet the goals of the NIF concep-tual design. Our work culminated in theNIF PEPC shown in Figure 2.

NIF PEPC LineReplaceable Unit

As with many other NIF components,the NIF PEPC is designed as a linereplaceable unit (LRU) to enable rapidreplacement and off-line repair. This is thesmallest subarray of apertures that will beinstalled or removed from the NIF beam-line. The PEPC LRU, shown in Figure 2, isa 4 ´ 1 module (four apertures high by onewide). A total of 48 PEPC LRUs will berequired to provide optical switching forthe 192 beamlines in NIF. Although theLRU integrates four apertures into onemodule, electrically it is divided into apair of two-aperture PEPCs, one of whichis shown in cross section on the right side of Figure 3. Also shown are the pulsegenerators that operate the PEPC. Eachplasma pulser drives a pulse of current(~1000 A) through the low-pressure heli-um (~65 mTorr) on either side of the KDPcrystals forming the plasmas. Once theplasmas cover both sides of the KDP, theswitch pulser fires, applying the exactvoltage (~17 kV) across the crystalsrequired for 90¡ of polarization rotation.

The NIF PEPC is a complex piece ofelectroÐmechanicalÐoptical hardware thatintegrates various elements, including afour-aperture PEPC, a high-vacuumpumping system, a feedback-controlledgas-feed system, a support frame, and athree-point kinematic mounting system.Each of the four apertures in an LRUrequires three large-aperture optical com-ponents: two fused silica windows andone plate of KDP. Each LRU contains eightwindows and four KDP plates.

The Operational Core: A Four-Aperture PEPC

The operational core of the PEPC LRU is a four-aperture PEPC whose design isbased on nine years of development andtesting. All the PEPCs we have built haveessentially the same construction. They area sandwich structure made of an insulatingmidplane between a pair of housings. Themidplane acts as a carrier for the KDP crys-tals. The housings, in conjunction with thewindows, form sealed volumes covering

35


UCRL-LR-105821-99-1

Cathode�area

Vacuum�ducts

Anode�area

Window

Vacuum�pump

Vacuum�baffles

External�frame

Housings

Midplane

the KDP surfaces. The plasma dischargesrequired for PEPC operation are formedwithin these sealed volumes. The housingsalso act as mechanical members that sup-port the midplane and allow for mountingof cathode-electrode assemblies, anode-electrode assemblies, vacuum baffle assem-bly, and silica windows.

An important difference between theBeamlet PEPC and the NIF PEPC is thechoice of housing material. Housings inthe Beamlet PEPC were fabricated fromsolid polyethylene. Although this materialprovided excellent insulation properties,the polyethylene proved difficult tomachine to close tolerances. For NIF, the

FIGURE 2. Each NIFPEPC line replaceableunit (LRU) is a 4 ´ 1module.(08-00-1198-2240pb01)

36


UCRL-LR-105821-99-1

Plasmapulser

Cathode

Electronemission

Side 1B1-1

B1-1

B1-3

B1-5

B1-7

B1-3

Dischargecurrent

Anode

Plasmapulser

–

+

–

+

+ –Switchpulser

Zout

Rterm

KDP

KDP

Plasma

Cathode

Electronemission

Side 2

Dischargecurrent

Anode

Plasma

KDP chargingcurrent

KDP chargingcurrent

PEPC housings are made from aluminum.Compared to polyethylene (or any otherplastic), aluminum has superior strengthand machinability. We take advantage ofthe higher strength by reducing the hous-ing wall thickness enough to meet theaperture packing requirements for NIF.However, aluminum does not provide therequired insulation between the dischargeanode and cathode structures. They areinsulated from the housing with plastic,and plasma is insulated from the alu-minum walls by an anodic coating on thealuminum (Al2O3 insulator).

We have tested our design of the opera-tional core by constructing an operationalprototype shown in Figure 4. This deviceforgoes the external frame describedbelow, but otherwise is a fully operationalNIF 4 ´ 1 PEPC. We have tested theswitching performance of the operationalprototype in a four-beam polarimeter thatilluminates each aperture with a low-fluence NIF-size laser beam. The experi-mental apparatus measures the fraction oflight in two orthogonal polarizations, from

which we can determine switching effi-ciency. If all the light is perfectly rotatedexactly 90¡, then the switching efficiency is100%. In practice, there are always factorsthat cause less than perfect switching.Some examples are strain in the crystal,variations in applied voltage, and strain inthe vacuum windows. Switching efficiencyonly includes losses due to polarizationerrors and does not include absorption inthe crystal (about 5% per cm) or reflectivelosses from the optic surfaces. The NIFrequirements are that the average switch-ing efficiency across an entire aperturemust be better than 99%, and that the min-imum switching efficiency of any smallspot must be greater than 98%. Resultsfrom the prototype show that each of thefour apertures has an average switchingefficiency of at least 99.8%, and the lowestsmall-area switching efficiency is 99.5%.Thus, the NIF requirements are easily met.We have determined that the dominantfactor causing polarization error in ourprototype is strain in the vacuum win-dows. Such strain is unavoidable because

FIGURE 3. The LRU is electrically divided into a pair of two-aperture PEPCs, one of which is shown in cross section on the right. Alsoshown are the pulse generators that operate the PEPC. (70-00-0299-0368pb01)

37


UCRL-LR-105821-99-1

of the atmospheric pressure differenceacross the windows. However, ensuringthe flatness of the window support shelf inthe housings can minimize the resultingpolarization error.

Vacuum and Gas FeedSystem

The vacuum and gas feed system isrequired for plasma-discharge production.The vacuum system consists of a hybridturbomolecular-drag pump. A singlepump evacuates the entire LRU interior tothe 10Ð6-Torr range. The discharge volumeis pumped through ports built into thecathode bases. These ports connect to thevacuum baffle structure. The baffles areinsulators that keep the switch pulse fromshorting out through the common vacuumsystem. The baffles connect to the turbopump with a series of ducts. All turbopumps in a cluster exhaust to a mechani-cally pumped foreline manifold.

The gas feed system injects the dis-charge gas near the anodes so the gasflows across the crystal faces. A constantpressure is maintained with a feedback-control system. We determined the operat-ing pressure experimentally to optimizeplasma uniformity, and the value is typi-cally 65 mTorr in the NIF PEPC LRU. Agroup of gauges mounted on the vacuum

ducting measures the pressure over various ranges. The high-vacuum base pres-sure is measured with a cold-cathode ion-ization gauge. The pumpdown from atmo-sphere is measured with a 0- to 1000-Torrcapacitance manometer. The operating pres-sure is measured with 0- to 1-Torr capaci-tance manometer. This gauge provides feedback to the pressure controller thatvaries the output of a mass flow controller.

Support Frame andKinematic Mounts

In the PEPC LRU, the operational coremounts into an L-shaped frame, as shownin Figure 5. The bottom of the supportframe is an interface plate that incorpo-rates all of the interfaces required betweenthe PEPC LRU and the rest of the PEPCsystem. The interface plate includesfeedthroughs for 4 plasma-pulse cables, 16switch-pulse cables, 2 switch-pulse voltagemonitors, gas input, foreline vacuumexhaust, vacuum gauge feedthroughs, andaccess to the turbo pump. The interfaceplate also seals to the NIF periscope struc-ture after LRU insertion.

The PEPC LRU attaches to theperiscope structure by a three-point kine-matic mounting system that allows forsafe, precise, ultraclean, repeatable posi-tioning of the PEPC in the NIF beamline.

FIGURE 4. Photographof the prototype used totest our design of theoperational core. Thisdevice has no externalframe, but otherwise is afully operational NIF 4 ´ 1 PEPC.(08-00-1198-2242pb01)

38


UCRL-LR-105821-99-1

A ball mounted to the top of the LRU, asshown in Figure 6, is one of the threemounting points. When inserted into theperiscope, the ball mates with a coneattached to the periscope frame. A closingÒclamshellÓ mechanism captures the balland keeps it in contact with the cone. TheLRU essentially hangs from this device. Theother two mounting points are a pair of Vgrooves, attached to each side of the LRUnear the bottom, that mate with a pair ofballs attached to the periscope frame.

The mounting system provides for pre-cise and repeatable placement of an LRU ina NIF beamline. Each LRU must be inter-changeable to any other location in thebeamline. The requirement for positioning

is ±1 mm in translation and ±120 µrad inrotation about the vertical axis. We haveverified that our design meets theserequirements by performing tests with theprototype assembly. Test results show thatwe achieved ±0.04 mm in translation and±55 µrad in rotation about the vertical axis. The results are well within the NIF requirements.

Assembly FixturesThe PEPC LRUs will be assembled in the

NIF Optical Assembly Building (OAB). Adedicated area in the OAB will be config-ured specifically for PEPC assembly andtesting. Because the PEPC LRUs contain

FIGURE 5. Prototype ofthe PEPC LRU mountedto a mockup section ofperiscope structure.(70-00-0299-0369pb01)

39


UCRL-LR-105821-99-1

Ergotech workpositioner

NIF PEPCassembly fixture

optics (windows and KDP crystals), theassembly procedure is designed to mini-mize contamination of optical surfaces.

We have designed and built a PEPCassembly fixture to facilitate assembly ofPEPCs in a safe, ultraclean, and ergonomicmanner. This device, shown in Figure 7,will be used in conjunction with anErgotech motorized positioner. The

Ergotech lifts and rotates the assembly fix-ture while the assembly fixture facilitatesintegration of the housings, midplane, andsupport frame.

The potting fixture, shown in Figure 8, is designed to control integration of KDPswitch crystals into the glass midplane.The midplane has four apertures for crys-tals, which must be precisely positioned

FIGURE 6. A ballmounted to the top ofthe LRU is one of thethree mounting pointsby which the PEPC LRUattaches to the periscopestructure.(70-00-0299-0370pb01)

FIGURE 7. The PEPC assembly fixture will be usedwith an Ergotech motorized positioner to assemblePEPC LRUs in the Optical Assembly Building.(70-00-0299-0371pb01)

FIGURE 8. The potting fixture controls integration of KDP switch crystals into theglass midplane. (70-00-0299-0372pb01)

40


UCRL-LR-105821-99-1

PEPC�supervisor

Plasma�FEP

Plasma�pulser

Switch�pulser

Valves

Pulse�diagnostics�

FEPPEPC

Turbo pump�controller

Gas�controller Voltage�

monitor

Switch pulser�FEP

Operator�consoles

Archive

Reports

with respect to the midplane. Four gasketsthat hold the crystals in the potting fixtureare custom shaped to achieve precise posi-tioning and provide a seal while potting.The crystals are potted with very-high-grade silicone rubber. The crystals are frag-ile and will fracture with minimal mechan-ical or thermal shock. The fixture and pot-ting process minimize both mechanicaland thermal loading of crystals.

PEPC Control SystemThe function of the PEPC control sys-

tem is to provide remote control and mon-itoring of the PEPC subsystem during NIFoperation. The system will control all thecomponents of the PEPC system, includingplasma pulsers, switch pulsers, vacuumand gas gauges, valves, and turbo androughing pumps.

As with other segments of the NIFcontrol system, the PEPC control systemis based on front-end processors (FEPs)that connect upstream to supervisorysoftware modules and downstream toinput/output interfaces, as shown inFigure 9. Each cluster of 12 PEPC LRUswill require 3 FEPs. Because there are 4clusters in NIF, the PEPC control systemwill have a total of 12 FEPs. The 3 FEPsper cluster divide the control duty forthe PEPC subsystems. The first FEP con-trols the switch pulsers. The second FEPcontrols the plasma-pulser controls,monitors the individual gas and vacuumsystem on each LRU, and monitors thecluster-wide vacuum roughing system.The third FEP controls pulse diagnostics.It contains high-speed waveform digitiz-ers for the acquisition of switch-pulsevoltage and plasma-pulse current.

FIGURE 9. The PEPCcontrol system.(70-00-0299-0373pb01)

41


UCRL-LR-105821-99-1

18000

14000

10000

6000

2000

0 100 200Time (ns)

Reference

Difference

Tolerance

Actual

300 400 500

During a NIF shot, the PEPC control sys-tem turns on the PEPC system, checks fornormal operation of each LRU, fires thePEPC system every 5 s until a NIF shotoccurs, and then turns the PEPC system off.The control system determines whether anLRU is firing properly by comparing theswitch-pulse voltage waveform with a ref-erence waveform, as shown in Figure 10.The actual waveform is subtracted from thereference. Any deviation in pulse timing orshape leads to a difference signal that islarger than the noise tolerance. The controlsystem detects the difference and alerts theshot-control operators that a PEPC LRU isnot working properly.

ConclusionThis article describes the NIF optical

switch system, which enables multipassoperation of the NIF laser amplifier. TheNIF optical switch is based on a novel electro-optic device called the plasmaelectrode Pockels cell (PEPC). WithPEPC technology, it is possible to con-struct Pockels cells with very large aper-tures. After proving the viability oflarge-aperture, PEPC-based opticalswitches on the Beamlet laser, we havedesigned and successfully tested a four-aperture PEPC for the NIF.

FIGURE 10. The PEPCcontrol system deter-mines whether an LRU isfiring properly by com-paring the switch-pulsevoltage waveform with areference waveform.(70-00-0299-0374pb01)

43UCRL-LR-105821-99-1

The Beamlet laser is a single-aperture,nearly full-scale physics prototype of

the 192-beam Nd:glass laser driver for theNational Ignition Facility (NIF).1 Itemploys a multipass amplifier architecturesimilar to the design of a single NIF beam-line. As shown in Figure 1, the laser sys-tem consists of a preamplifier followed bytwo large amplifier stages: a four-pass cav-ity amplifier and a single-pass boosteramplifier. The cavity amplifier containseleven side-pumped BrewsterÕs angle slabssituated at one end of a 36-m-long image-relayed cavity formed by a spatial filterand two end mirrors. The pulse from thepreamplifier is injected near the focal

plane of the spatial filter and makes fourpasses through the slabs before being ejected from the cavity by a full-apertureplasma-electrode Pockels cell and polariz-er. The five-slab-long booster amplifierprovides additional amplification anddelivers typically 12 kJ in 3 ns, the Beamletdesign point. The beam is then spatiallyfiltered and relayed to the final optics,where it is frequency-converted to thethird harmonic and focused to an equiva-lent NIF target plane. Comprehensivediagnostics at the input and output of themain amplifier provide beam data forgauging system performance at both 1.053-µm (1w) and 0.351 (3w) wavelength.2,3

BEAMLET EXPERIMENTS

P. Wegner B. Van Wonterghem

S. Burkhart C. Widmayer

J. Murray

Beamlet deformable mirror

Cavity mirror

SSD Preamplifier

11 slabs

(b)(a)

Cavity amplifier

5 slabsBooster

amplifier

Pockelscell

NIF deformable mirror

Polarizer

Transport filter

Transport mirrors

NIF injection

point

Frequencyconverter

(in vacuum)

Final optics

Oscillator

Cavityfilter

FIGURE 1. (a) Photo of the Beamlet facility taken from the output end of the laser, showing the test mule and focal-plane diagnostic (fore-ground) used for third-harmonic experiments. (b) Diagram of the Beamlet prototype laser showing relative locations of major components.(70-50-1297-2574pb01)

44

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

BeamletÕs design diverges from theNIFÕs in three important and quantifiableways:

1. A smaller clear aperture (36 vs 40 cm). 2. Injection of the Beamlet pulse into the

18-m-long cavity filter (vs injection ofthe NIF pulse into the 60-m-longtransport filter).

3. An adaptive optics system that utilizes a 7-cm deformable mirror atthe output of the preamplifier (vs the40-cm deformable mirror that will bedeployed at the cavity first-mirrorposition on NIF).

In addition, the Beamlet master oscilla-tor and preamplifier are non-prototypicalof NIF, although they share certain mod-ern design features. Nonetheless, Beamlethas proven to be an essential test bed forevaluating laser physics and componentengineering issues related to the NIF.

The primary mission for Beamlet hasbeen the integrated testing of NIF lasertechnologies. Since its activation milestonein September 1994, Beamlet has producedover one thousand full-system shots inover twenty experimental campaignsaddressing a broad range of technicalissues relating to high-power beam propa-gation, high-energy temporally shapedpulses, spatial filtering, wavefront control,final optics, and frequency conversion.Experiments in high-power beam propaga-tion established NIF B-integral limits andspatial-filter requirements for controllingnonlinear ripple growth and beambreakup. Experiments in pulse shapingproduced the high-contrast, high-fluence20-ns shaped 1w pulses required for theNIF ICF mission. Experiments addressingspatial-filter issues established the pres-sure limits for NIF spatial filters, anddemonstrated new pinhole designs thatare effective in mitigating closure and backreflections under NIF operating condi-tions. Experiments in wavefront controldemonstrated ability to meet NIF focusingrequirements and provided a baseline forestablishing finishing specifications forNIF optical components. Final opticsexperiments evaluated prototype UV

components in NIF-like configurations at high fluence and demonstrated high-efficiency frequency conversion to thethird harmonic using crystals fabricatedwith both conventional and rapid-growthtechnology. The following sectionsdescribe each of these activities.

High-Power PulsePropagation

The control of nonlinear ripple growthleading to beam breakup is an essentialpart of the design and operation of high-power solid-state lasers.4 The mechanismof concern is the intensity-dependentrefractive index in the laser components,which enables a weak ripple wave copropagating with a strong pump waveto couple and scatter a third wave conju-gate in angle to the ripple.5Ð7 Subsequentinteraction of the conjugate wave with thepump feeds back and amplifies the ripplewave leading eventually to the unstablegeneration of higher-order ripple modes,self-focusing, and beam breakup.

Methods of mitigating ripple growth arelimited, but critical for producing high-quality beams with safe modulation levels.Optical components must comply withstringent specifications for homogeneity andsurface finish to limit the source of phaseperturbations, which cause small-scaleamplitude modulation. Growth of thesesource terms depends on the B-integral, orintensity-dependent phase retardationallowed between spatial-filter pinholes (DB),and is limited by pinhole sizing.

Near-field modulation experimentswere conducted on Beamlet to establishthe B-integral limits and spatial filteringrequirements for the NIF. By propagating200-ps pulses through the laser under con-ditions equivalent to those expected dur-ing a 20-ns ICF ignition pulse, we obtainedÒsnap shotsÓ of beam quality that mightotherwise be masked by temporal integra-tion effects in the diagnostics. The reducedgain and high B-integral that occur duringthe most stressful period near the end ofthe ignition pulse were simulated by con-ducting the majority of the tests with the

45

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

(a) ∆B = 0.9

(b) ∆B = 2.6

booster amplifiers turned off. This configu-ration mimics the most severe condition inwhich the pulse has extracted all the ener-gy from the booster amplifiers.

A metric for evaluating beam quality isthe beam contrast, defined as the standarddeviation of the near-field irradiance divid-ed by the mean. Contrast was measured forvarious pinhole sizes as a function of the B-integral accumulated between the Pass-4pinhole in the cavity spatial filter and thepinhole plane of the transport spatial filter,DB. Initial tests were conducted without apinhole in the transport spatial filter toevaluate beam contrast at the input lens ofthe spatial filter, where the risk of 1w dam-age is highest. Figure 2 shows examples ofthe resulting near-field irradiance dataobtained at low and high B-integral. Thedependence of contrast on DB is plotted inFigure 3a for two different cavity pass-4pinhole sizes corresponding to acceptanceangles of ±200 µrad and ±130 µrad. Openpoints in the plot represent shot data, andsolid points are the result of numerical sim-ulation. The onset of rapid deterioration inbeam quality occurs at lower DB with thelarger pinhole. At a DB of approximately 2 rad, the 200-µrad data has begun a rapidascent towards beam breakup while the130-µrad data is still in the slowly varyingregion of the curve. Thus for DB < 2, cavitypinhole sizes of 200 µrad or smaller areacceptable, but with smaller pinholes themargin for error is substantially increased.

Figure 3b plots the contrast that results atthe output relay plane of the laser when apinhole is used in the transport spatial filter.In this case there is discrepancy between thetest data and the simulations that is attribut-ed to incomplete characterization of thenoise fields in the preamplifier. Nonetheless,both show that the beam contrast shouldremain small for DB as high as 3 rad if thepinhole size in the filters is reduced from 200to ~100 µrad. Reducing pinhole size is achallenge because it increases the risk of clo-sure for long pulses (see the section SpatialFiltering on p. 47). For DB < 1.8 rad, the contrast is < 0.08 and, within error, is inde-pendent of pinhole size below 200 µrad. Thisresult is the origin of the 0.1-contrast specifi-cation and 1.8-rad DB limit for the NIF.

High-Energy ShapedPulses for ICF

Temporal pulse shaping will be achievedon the NIF using low-voltage waveguideelectrooptic modulation techniques in inte-grated optical circuits.8 This technology hasbeen used on Beamlet to produce 20-ns,NIF-like ignition pulses with a pulse-shaping system consisting of two arbitrarywaveform generators (AWGs) in series. A20-ns AWG with 1-ns resolution formed theelectrical signal corresponding to the long,

FIGURE 2. 1w near-fieldbeam modulation at (a) low and (b) high B-integral through thebooster amplifiers.Images show the central24.4 cm of the beam andwere obtained with out-put sensor lookingthrough the transportspatial filter with thepinhole removed.(70-00-0499-0781pb01)

46

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

0.40

0.30

0.20

0.10

00 1 2 3 4

ÆB booster (radians)

NIF limitC

ontr

ast Æ

I/I

0.20

0.15

0.10

0.05

00 1 2 3 4

ÆB booster (radians)

NIF limit

Con

tras

t ÆI/

I

(a)

(b)

low-intensity foot of the pulse, and a 7-nsfast (250-ps) AWG appended the high-bandwidth features of the complex pulseshape to the end of the foot. The resultingdriver voltage signal controlled the lightamplitude propagated through a LiNbO3modulator.

The required optical pulse shape at themodulator was calculated from the desiredpulse shape at the laser output usingempirical lumped-gain models for the vari-ous Beamlet amplification stages. The rela-tively small amount of saturation in the

regenerative amplifier and preamplifierwas well approximated by energy gains G that decreased linearly with extractedenergy at the rate of 1.4%/mJ and 16%/J,respectively. The gain model for the mainamplifier was based on the curve fit to thedata shown in Figure 4, which plots mea-sured input energy vs output energy forshots spanning approximately one year.The data was best approximated with two polynomials: 2nd order below 6583 Jand 3rd order above. The resulting compos-ite curve produces a better fit than could be

FIGURE 3. (a) Near-field irradiance contrastratio vs B-integral fortwo cavity/transportpinhole configurations:130 µrad/open (squares)and 200 µrad/open (circles). Measurementsand simulations aredenoted by open andfilled symbols, respec-tively. (b) Resultsobtained with a pinholein the transport spatialfilter: 130 µrad/100 µrad(squares) and 200 µrad/200 µrad (circles). (70-00-0499-0782pb01)

47

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

4

3

2

1

00 5 15 20 2510

Time (ns)

Measured�Calculated

Pow

er (T

W)

0

200

400

600

800

1000

1200

1400

1600

1800

0 2000 4000 6000 8000 10000 12000 14000 16000 18000

y = 4.908400e –10x3 – 6.601717e – 06x2 + 7.375619e – 02x – 1.171389e + 02

Crossover occurs at 6583.82

Last 5 datapoints by PROP92

y = 3.345649e – 06x2 + 1.156416e – 02x + 1.216097e + 00

Output energy (J)

Inp

ut

ener

gy (

mJ)

Data series 1Data series 3Data series 4

achieved with a simple single-gain-elementFrantzÐNodvik calculation, and has provenadequate for pulse shaping.

With predictions based on these models,we produced the 15.5-kJ, 1w NIF ignitionpulse shown in Figure 5. Scaled for beamsize, the energy and peak power of thispulse fall on the NIF red-line performancecurve for square pulses shown in Figure 6,demonstrating the primary 1.05-µm laserrequirement for inertial confinementfusion. Successful propagation of thispulse required advances in spatial filter-ing, described in the following section.

Spatial FilteringSpatial filter issues of importance for

the NIF fall into three categories:

1. Pinhole closureÑthe problem ofkeeping the required small pinholesopen for the full duration of the NIFignition pulse.

2. Back reflectionsÑthe need to avoidback reflections from pinholes thatcan damage the injection optics.

3. Background pressureÑthe questionof maximum safe operating pressuresfor NIF cavity and transport spatialfilters (CSFs and TSFs, respectively).

Beamlet experiments addressed each ofthese areas in turn.

Pinhole ClosureThree types of pinholes were tested for

closure on Beamlet, as shown in Figure 7.The first type is a washer design consistingof a hole in a flat plate oriented at approxi-mately normal incidence to the beam. Thesecond type is a leaf design consisting offour azimuthal segments displaced along

FIGURE 4. Gain of themain amplifier plottedas input energy requiredto produce a given out-put energy. Curvedenotes best fit used forpulse shaping.(70-00-0499-0783pb01)

FIGURE 5. Measuredtemporal profile of a 15.5-kJ, 1w ignition pulseproduced on Beamlet.Data is in good agree-ment with predictionsbased on the pulse shap-ing model (gray profile). (70-00-0499-0784pb01)

48

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

Washer 4-leaf Cone

Cone4-leaf

Beam�direction

8

6

4

2

00 5 10 15 20 25

Energy (kJ)

1 ns

1.7 ns

3 ns

3.5 ns4.5 ns eq

5 ns

NIF 11/5, ∆B < 1.8Beamlet square pulse dataBeamlet shaped pulse data

5.5 ns eq

8 ns10 ns

Pow

er (T

W)

the beam axis to eliminate the possibility ofplasma convergence at the center of thepinhole. The third type is a conical pinhole,designed such that low-density plasmas onthe surfaces reflect or refract the incominglight rather than absorb it.9 The design ofthe cone pinhole is parameterized by thecone angle a and the cone length L. For theBeamlet tests we set a = 1.3 amin and L =0.7Lmax, where amin equals the f-number ofthe spatial filter divided by 2 and Lmax =2R0/(a +amin).10 Here R0 is the radius ofthe pinhole aperture, which when dividedby the focal length of the spatial filter inputlens gives the cutoff angle of the pinhole.

Pinhole performance was evaluated witha specialized set of diagnostics. A pulsedMachÐZender interferometer was used tomeasure the phase shift of a 532-nm probebeam passing through the pinhole duringthe passage of the main laser pulse. Thefringe pattern from the interferometer wasrecorded on a streak camera and a 120-psrise-time gated optical imager to obtaintime-resolved phase maps in x-t and x-ythat could be correlated with electron den-sity in the pinhole (see Reference 11 andarticle by Feit on p. 63 of this QuarterlyReport). A second gated optical imagermeasured the transmitted near-field beam

FIGURE 6. NIF 11-5amplifier configurationsafe performance limit,with Beamlet experi-mental data pointsscaled to the NIF beamarea. Square pulse datais plotted as open cir-cles. Shaped pulse data(filled circles) is charac-terized by an equivalentpulse duration definedas the energy divided bypeak power. (70-00-0499-0785pb01)

FIGURE 7. The threetypes of pinholes testedon Beamlet.(70-00-0499-0786pb01)

49

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

0

1

2

3

4

5

10 100Atomic mass

E cl

osu

re (k

J)

Cone 4-leafWasher

AuTaFeC

irradiance during the last nanosecond ofthe pulse. Because pinhole closure affectsthe trailing edge of the pulse first, this diag-nostic gave the most definitive indication ofclosure at near-threshold conditions. A sec-ond streak camera was configured to imagea central strip of the transmitted near-fieldbeam irradiance and measure the time vari-ation of the modulation in that strip. Thisdiagnostic was used to determine when theincreased modulation associated with clo-sure occurred during the pulse.

The majority of the tests were conductedwith 20-ns square pulses to simulate theleading foot of a shaped ignition pulse. Tocompare shots of different energies Epulseand closure times t, we adopted a figure ofmerit Eclosure defined as the energy neededto close a pinhole at the end of a 20-nssquare pulse. Because the closure time isinversely proportional to the plasma expan-sion velocity, which in turn is proportional tothe laser irradiance at the edge of the pin-hole,11 Eclosure is simply Epulset(ns)/20, thevalidity of which has been confirmed bymeasurement. Applying this analysis to thetest data for ±100-µrad pinholes with 20-nssquare pulses yields the results summarizedin Figure 8, which plots Eclosure as a functionof atomic mass for the three pinhole typesand four different materials. Conical pin-holes outperformed washer and four-leafpinholes in all cases. In general, performance

was also better for higher-Z pinholes whichproduce shower closure velocities, with theexception of the Au cone, which did not per-form as well as the Ta cone. It is believedthat an inadequate finish on the interior sur-face of the Au cone caused its lower-than-expected performance.

The conical pinhole design was alsotested using 20-ns shaped ignition pulses.The required ignition pulse at the inputto the NIF frequency converter has anenergy of 19.4 kJ and a contrast of 10:1(Figure 5). However, simple scaling lawsfor pinhole closure show that the pinholemost susceptible to closure is not the out-put pinhole in the f/80 TSF, but the Pass-4 pinhole in the f/31 CSF, where thepulse energy is 14.8 kJ and the contrast is21:1. Tests with this pulse shape closed a±100-µrad stainless-steel (SS) cone pin-hole at 3´ the energy of the correspond-ing 20-ns foot-only pulse, suggesting thatthe closure energy for ignition pulses canbe obtained by multiplying Eclosure inFigure 8 by 3. In this case, none of the±100-µrad pinholes tested would workfor the NIF ignition pulse. The ±100-µradTa cone comes close based on this simple3´ scaling, but additional margin wouldbe needed for the increased angulardivergence associated with beam smooth-ing by spectral dispersion (SSD) and forfinite alignment tolerances.

FIGURE 8. Closure ener-gies for three types of±100-µrad pinholes plot-ted vs atomic mass.(70-00-0499-0787pb01)

50

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

Streaked interferogram

Position

Power

Tim

e

Phase shift (∆φ) at pulse end

Position in pinhole (µrad)–100–150 –50 0

0

0.5

1.0

1.5

50 100 150

At closurethreshold

At end of15.5-kJ pulse

∆φ(w

aves

)

On the other hand, a ±150-µrad SS conepinhole meets NIF requirements with ease,passing a 15.5-kJ, 10:1 ignition pulse with±7.5 µrad of added SSD divergence withno sign of closure. The 10:1 ignition pulseis harder to keep open than the required21:1 pulse with the same total energybecause the intensity in the foot is largerby 2´. Figure 9 shows the interferometrydata for this shot. The interferogramrecorded by the streak camera is shown onthe left. Analysis of the fringe pattern atthe time indicated gives the phase shift ofthe probe pulse as a function of position inthe pinhole, plotted on the right. A secondcurve in this plot shows how the phaseshift would look at closure based on datafrom a different shot. The large separationbetween the curves indicates that the ±150-µrad SS cone pinhole was quite farfrom closure at end of the pulse. This pinhole is currently our baseline choice for NIF.

Back Reflections fromPinholes

Back reflections have been a problemfor staged pinhole geometries, in whichthe pinhole angular acceptance is gradual-ly decreased from the input to the outputof the laser. Because pinhole stagingresults in the best output beam quality, itremains the preferred mode of operation,and thus a solution for back reflections isrequired. All of the pinhole tests were per-formed in this configuration and generat-ed measurable back reflections for pulsesgreater than about 1 TW into the transportspatial filter. However, the energy reflected

from the cone pinholes was down by atleast an order of magnitude from the ener-gy reflected by the other pinhole geome-tries. Furthermore, its back reflectionincreased approximately linearly withpower, whereas the back reflection for bothleaf and washer pinholes increased nonlin-early, indicating a stimulated scatteringprocess at the pinhole. Imaging of the back-reflected light (Figure 10) showed unam-biguously that the back reflections originatefrom the surfaces of the pinhole rather thanfrom an on -axis plasma, confirming theadvantages of the cone geometry over theplanar washer and leaf designs.

Maximum Background GasPressure

Residual gas in the spatial filters atpressures above a certain maximum pthcauses increased modulation in the near-field irradiance of the transmitted pulse,similar to the modulation observed abovethe threshold for pinhole closure.Threshold pressure is known to dependstrongly on spatial filter f-number,12 andas a result the Beamlet gas-pressure testswere conducted with both f/26 and f/78geometries to determine values for pthapplicable to the NIF CSF (f/31) and TSF(f/80), respectively. Figure 11 shows theresults plotted as pressure vs peak intensityat the focus of the spatial filter. Three typesof points are plotted for each f-number todistinguish whether the data came inabove, at, or below pth as gauged by themodulation level in the transmitted near-field beam irradiance. Best-fit pth for thef/26 data is indicated by the upper curve,

FIGURE 9. Interfer-ometric data showing thephase shift of a probepulse at the pinhole dur-ing passage of a 15.5-kJ,10:1 contrast ignitionpulse with ±7.5 µrad ofadded SSD divergence.The pinhole was a ±150-µrad SS cone.(70-00-0499-0788pb01)

51

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

4-leaf pinhole

±200-µrad�cavity pinhole

±100-µrad�transport pinhole

Cone pinhole image (enhanced)

0.0001

0.001

0.01

0.1

f/78

f/26

1

10 10,000 100,000 1,000,0001,000100

NIF TSF2 mTorr

Intensity (TW/cm2)

p th (

Tor

r)

NIF CSF6 mTorr

AboveAtBelow

Threshold condition:

which is well defined out to the maximumintensity expected for the NIF CSF. Thef/78 data, however, was obtained byreducing the aperture of the beam by 3´,which also reduced the maximum poweroutput by 9´. Consequently, the f/78 dataends well short of the expected intensitiesfor the NIF TSF. Maximum and minimumcredible extrapolations are indicated bythe two gray curves, with the dark grayintermediary curve corresponding to ourcurrent best estimate. The results give pthvalues of 6 mTorr for the NIF CSF and 2 mTorr for the TSF. There is roughly a factor of two uncertainty in the f/78 data,because the beam quality was significantly

better for the central subaperture of thebeam used in those tests, and better beamquality has been observed to give lowervalues of pth. Safety considerations dictatethat the maximum allowed NIF operatingpressures be roughly an order of magni-tude below these values.

Measurements with both residual gasand a pinhole in the spatial filter showed nointeraction between the effects of the residu-al gas and pinhole closure. Eclosure remainsessentially constant for pressures from wellbelow to 3´ above pth. The reason there isno interaction is that the two phenomenaaffect different temporal parts of the pulse.Data from the streaked near-field diagnostic

FIGURE 10. Back-reflected laser light fromTSF pinholes, as imagedthrough the pinholes inthe CSF. Back-reflectedenergy from the 4-leaf pinhole was 180 mJ at2.0 TW, as measured atthe input sensor, whilethat from the cone wasonly 10 mJ at 3.5 TW.(70-00-0499-0789pb01)

FIGURE 11. Data fordetermining thresholdpressure pth, at whichresidual spatial filterpressure disturbs thetransmitted pulse. Filledtriangles and squaresdenote data that wasabove or at threshold,respectively. Data belowthreshold is indicatedby circles. Dashedlines show predictedthresholds at expectedNIF filter intensities(70-00-0499-0790pb01)

52

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

Ð100 0.5 1.0 1.5 2.0

0

10

20

Ad

ded

full

angl

e be

am d

iver

genc

e @

80%

(µra

d)

Temperature difference (°C)�[vertical cavity surfacesÑtop reflector]

30

40

showed that pressures at or slightly abovepth affect mainly the leading edge of thepulse, whereas near-threshold pinhole clo-sure affects the trailing edge.

Wavefront Control andBeam Focusability

Effective wavefront control is essentialfor achieving the high-brightness focal-spotconditions specified for NIF targets. Forexample, certain NIF weapons physics tar-get requirements call for delivery of 500 TWof 3w radiation inside a 250-µm-diam focalspot.13 For the 7.7-m focal-length lenses onthe NIF target chamber, this spot size corre-sponds to a half angle of 16 µrad, whichsets a stringent upper limit for the diver-gence of the laser.

There are several sources of divergencein the laser, primarily in the 1w section,that can significantly degrade the qualityof the focal spot unless mitigated or other-wise controlled. These sources fall readilyinto four categories:

1. Static phase errors related to the fin-ishing, mounting, and alignment ofthe optical components.

2. Prompt phase errors related to adeformation of the amplifier slabsduring pumping.14

3. Thermally induced phase errorsrelated to heat accumulation in theamplifiers, including gas turbulenceeffects.15,16

4. Nonlinear phase errors associatedwith the intensity-dependent ripplegrowth and whole-beam self-focusing.

At a given power level, minimumdivergence and maximum beam bright-ness are achieved when the system is coldand thermally induced phase errors areabsent (see Figure 12). In this case, perfor-mance is primarily limited by the fractionof prompt and static phase errors thatremain uncorrected by the wavefront-control adaptive optic system (AOS) as aresult of its limited spatial resolution.

The active component in the BeamletAOS is a 7-cm-square deformable mirror

(DFM) with 39 independent actuators;17 thenumber of actuators and their arrangementis similar in design to the 40-cm mirror thatwill be deployed on the NIF.18 The mirrorresides at the output of the preamplifier(see Figure 1) and conditions the wavefrontof the pulse before it is injected into thecavity amplifier. Wavefront data for closed-loop control of the mirror is provided byeither of two 77-element Hartmann sensorslocated in diagnostic packages situated atthe input and output of the main amplifier.Closed-loop control allows the figure of themirror to be updated continuously (~1 Hzresponse) to maintain a predefined wave-front at the Hartmann sensor, which is typi-cally specified to be either ÒflatÓ or a com-pensating figure determined from thewavefront error measured on a previousshot (termed ÒprefigureÓ). Additional diag-nostics, including radial shear interferome-ters, were used to independently check the operation of the AOS and quantifybeam quality.19 Measurements of system

FIGURE 12. Beamlet measurements correlating 1wlaser divergence with amplifier temperature. TheAOS is unable to correct for gas-density fluctuationsof high spatial frequency caused by amplifier heating.(02-30-1093-3491pb01)

53

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

Wavefront

(a)

(b)

(c)

1.01 λ P-V, 0.19 λ rmsStrehl 0.24

80% energy inside ± 10.1 µrad3.3 × 1022 W/sr-TW peak

2.94 λ P-V, 0.56 λ rmsStrehl 0.05

80% energy inside ± 27.8 µrad 0.7 × 1022 W/sr-TW peak

0.90 λ P-V, 0.13 λ rmsStrehl 0.50

80% energy inside ± 10.3 µrad6.7 × 1022 W/sr-TW peak

Log scale (~5 decades)

MaximumMinimum

Focal spot

±30µrad

wavefront and beam quality both with andwithout an optimized AOS are presented inthe following subsections.

Preamplifier WavefrontQuality

The output of the Beamlet preamplifieris very close to diffraction-limited.Continuous-wave measurements showedthat the AOS improves the wavefront ofthe preamplifier by ~0.25 waves to achievea residual error of 0.32 waves peak to val-ley, 0.06 waves rms, and a Strehl ratio of0.87. Firing the 5-cm rod added ~0.2 wavesof prompt phase error that was not readilyevident unless the AOS was actively cor-recting the static error. With the rodpumped, wavefront measured with andwithout the DFM was qualitatively differ-ent but similar in peak-to-valley and rmserror. The 80% spot size was equivalent forthe two cases (4.3-µrad half angle), but theDFM improved the brightness of the focalspot by ~30%. These results are consistentwith those of Reference 20.

Output Wavefront QualityBeam quality at the output of the sys-

tem is approximately 2.5´ the diffractionlimit with the AOS optimized to correctboth prompt and static wavefront errors inthe main amplifier, meaning that 80% ofthe energy is in a diameter 2.5´ the 80%diameter of a diffraction-limited beam.With the preamplifier pumped and themain amplifiers static (rod shot condition),the residual wavefront error at the outputof the system was ~1 wave peak to valley,0.2 waves rms, and the 80% half angle ofthe focal spot was 10.5 µrad (Figure 13a).The measurement was made with the AOSoperating closed-loop to maintain a flatwavefront at the output Hartmann sensorup until 1 s prior to the shot. Dataobtained under similar conditions, butwith the main amplifiers pumped, yieldedan output wavefront error of ~3 wavespeak to valley, 0.6 waves rms, and a much-degraded focal spot (Figure 13b). The difference between these two wavefrontsgives the prompt distortion caused by

FIGURE 13. 1w wavefront and focal spot measured at the output of the mainamplifier for (a) rod shot with the AOS maintaining a flat wavefront up until 1 sprior to the shot, (b) same AOS condition as (a) but with the main amplifierspumped, and (c) main amplifiers pumped with the AOS maintaining an optimizedprefigured wavefront up until 1 s prior to the shot. (70-00-0499-0791pb01)

54

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

TABLE 1. Results of high-power focal spot measurements conducted with an optimized AOS at output powers of up to 5.3 TW (1w) and3.1 TW (3w) in a 200-ps pulse.

1w focal spot 3w focal spot

50% 80% 50% 80%SB1w P1w P3w half angle half angle Peak half angle half angle Peak

Shot Amplifier (rad) (TW) (TW) (µrad) (µrad) intensity* (µrad) (µrad) intensity*

B7082005 11-0 1.3 1.3 0.6 5.4 11.7 3.7 8.0 13.6 0.67

B7082103 11-0 2.0 2.0 1.2 5.4 11.5 3.9 7.4 12.8 0.73

B7082205 11-0 2.1 2.2 1.3 5.7 11.8 3.0 8.1 13.7 0.70

B7082501 11-0 2.8 2.8 1.8 5.1 11.6 3.8 8.4 14.1 0.67

B7082601 11-5 1.6 3.4 2.3 5.3 12.3 3.3 7.7 12.5 0.92

B7082702 11-5 1.8 3.8 2.2 4.6 11.7 3.9 6.7 12.3 0.83

B7082802 11-5 2.6 5.3 3.1 4.4 12.1 4.7 6.5 14.9 0.82

B7082902 11-5 2.4 5.0 3.2 4.0 11.6 4.8 7.3 15.4 0.77

* 1022 W/sr-TW. Divide by the square of the lens focal length in cm to obtain irradiance (W/cm2-TW)

pumping the large amplifiers. With anappropriate prefigure of the DFM basedon this measurement, it was possible toachieve an output wavefront and focalspot on a low-power system shot that wereequivalent in quality to the data obtainedon rod shots (Figure 13c).

These results demonstrate that the 39-actuator design of the DFM is highlyeffective at correcting the prompt wave-front distortions incurred in the mainamplifier, and that, as a result, the focus-ability of the laser is primarily limited by the static errors in the main amplifierthat are not correctable with the AOS. Inthe case of Beamlet, this residual error has been shown to meet the NIF high-brightness focal-spot requirements at bothlow and high power. Table 1 summarizesthe results of high-power focal spot mea-surements conducted with an optimizedAOS at output powers of up to 5.3 TW (1w) and 3.1 TW (3w) in a 200-ps pulse.Amplifier configuration is denoted 11-0 or11-5, depending on whether the boosteramplifier was pumped. Maximum powerwas achieved with the 11-5 configuration,for which the total B-integral accumulatedin the amplifiers was 2.6 rad (SB), and the corresponding 80% power half anglesof the 1w and 3w focal spots were 12 and 15 µrad, respectively. Scaling the power in the 3w focal spot (0.8 ´ 3.1 TW) by the

ratio of NIF to Beamlet beam sizes (1240 cm2/1050 cm2) and multiplying bythe number of beams (192) and the trans-mission of the final optics (0.94) results in a NIF-equivalent performance of 540 TWinside ±15 µrad. Thus if the quality andassociated static errors of the NIF opticsare held to Beamlet levels, and the NIFAOS behaves equivalently, the NIF focus-ing requirements will be achievable.

Static Errors and OpticsFinishing Specifications

The static wavefront errors in theBeamlet amplifier were quantified by calculating the difference between theinput and output wavefronts, as measuredon a rod shot with a flat mirror in place ofthe DFM. The result, shown in Figure 14,has proven useful for correlating optics-finishing specifications with focal-spotperformance. For the central core of thefocal spot, corresponding to divergenceangles less than ~30 µrad, the finishingeffects of importance are long-wavelengthfigure errors, for which the appropriatespecification is the rms gradient of thetransmitted wavefront.21 Applying a low-pass filter with a cutoff frequency of0.03 mmÐ1 to the difference data in the fig-ure and calculating the rms gradient of theresult yields a value of 1300 �/cm for all

55

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

Phase gradient (waves/cm)

Phase (waves)

Nor

mal

ized

occ

urre

nce

0

σx = 0.087σy = 0.088

–0.2–0.4

–0.75 0.76

0.20

2

4

6(a) (b)

0.4

of the optics combined. Assuming incoher-ent addition of phase between different ele-ments, and accounting for multiple coherentpasses through sections of the amplifier, theaverage rms gradient per optic is estimatedto be 1300/17.2 = 75 �/cm. Simulationsusing an average gradient distributionbased on this result, and nominal powerspectral densities for the high-frequencyaberrations22 predict focal spots that areconsistent with the Beamlet measure-ments.23 Thus, to ensure focal-spot perfor-mance equivalent to Beamlet, specificationsfor NIF optics currently limit the rms gradient of the transmitted wavefront to 70 �/cm for spatial scale lengths >33 mm.

Final Optics andFrequency Conversion

The NIF final optics perform severalcritical functions in a compact assembly:

1. Frequency converting the 1w pulsefrom the laser amplifier to 3w withhigh efficiency.

2. Focusing the 3w energy onto the target.

3. Diverting the unconverted energyaway from the target.

4. Providing a full-aperture sample ofthe 3w beam for diagnostics.

5. Randomizing the spatial coherenceof the laser energy at the target.

6. Shielding upstream optics fromdebris and high-energy x rays andneutrons emitted from the target.

The designs of the components thataccomplish these tasks have beendescribed elsewhere.24,25

To test and validate various aspects ofthe final optics design on Beamlet, we constructed a Òtest muleÓ (term borrowedfrom the auto industry describing a flexi-ble prototype) at the output of the mainlaser amplifier that allowed us to field 37-cm-aperture versions of the opticalcomponents in a NIF-like configurationwithout the cost and complexity of activat-ing a complete final optics assembly(FOA). A NIF-like 1w window at the inputto the test mule isolated the vacuum envi-ronment of the final optics from the mainamplifier, allowing safe operation at full3w fluence without risk of damage andpotential fracture of the input window.Temperature-controlled water flowingthrough heat exchangers on the surface ofthe test mule maintained a constant ther-mal environment of 20.0 ± 0.1¡C for thefrequency converter; similar passive cool-ing will be used for the NIF FOA. Beamalignment into the test mule and the 3wdiagnostics package was accomplished

FIGURE 14. (a) Staticwavefront distortion ofthe main amplifier. mea-suring 1.51l P-V, 0.34lrms. (b) Horizontal (gray)and vertical (solid) gradi-ent distributions after fil-tering with a cutoff fre-quency of 0.03 mmÐ1.Total rms gradient is therss of sx and sy.(70-00-0499-0792pb01)

56

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

using four 45¡ high-damage threshold 1wmirrors. The requirement for slight out-of-plane orientation of the mirrors resulted inmixed ÒsÕÓ and ÒpÓ reflections and theassociated risk of beam depolarization,which could negatively impact frequencyconversion. Fractional depolarized energywas measured both with and without themirrors and found to be acceptable at lessthan 1%. Test mule experiments address-ing frequency conversion and high-fluenceoperation of the final optics are describedin the following subsections.

Frequency ConversionExperiments conducted with the test

mule played an important role in provingthe design of the NIF frequency converterand validating the physics model onwhich a detailed error analysis of its per-formance is based. The model is the mostcomprehensive yet developed for the con-verter, including details such as spatially

varying birefringence in the crystals that hasonly recently been quantified using orthogo-nal polarization interferometry techniques.26

Several Beamlet test configurations, summa-rized in Table 2, were used to evaluate bothsecond-harmonic generation (SHG) andthird-harmonic generation (THG), with converters consisting of both conventionallygrown and rapidly grown potassium dihy-drogen phosphate/potassium dideuteriumphosphate (KDP/KD*P) crystals.27 The crystals were tested in a prototype37-cm-aperture final optics cell (FOC),which also contained the final focus lens, as shown in Figure 15a. Precision-machinedsurfaces in the FOC supported the opticsaround their perimeter and registered themwith microradian tolerance. Compliantclamps held the optics in place; a load of 0.6 lb/in. was found to adequately constrainthe crystals while providing good surfacefigure (Figure 15b). The configuration of the1w laser was the same for all tests: elevencavity amplifiers, five booster amplifiers, a200-µrad C pinhole in the Pass-4 cavity spatial filter pinhole, and a 150-µrad SS conepinhole in the transport filter. Pulse formatwas 1.5 ns square. Estimated accuracy of the energy-conversion efficiency mea-surements was ±6% (3w).

SHG efficiencies measured with a con-ventionally grown Type-I doubler from a NIF production boule are plotted inFigure 16a. Maximum energy efficiencywas 73% (aperture-averaged, time-integrated) at an input 1w irradiance ofapproximately 4 GW/cm2 (aperture-averaged, peak-in-time). Similar tests of arapidly grown Type-I doubler achieved70.5% efficiency at similar drive irradi-ance. The measured performance of thesecrystals was in good agreement with mod-eling based on measured 1w pulse param-eters and measured crystal refractive-index variations. The effects of the latterwere verified by measuring the 2w near-field fluence distributions with the crystaltilt biased well away from exact phasematching, a configuration very sensitive tophase mismatch. As shown in Figure 17,the resulting nonuniformities in the datawere well reproduced in the model.

THG efficiencies measured with arapidly grown doubler and tripler areplotted in Figure 16b. Maximum energy

Table 2. Beamlet test configurations used to evaluate both SHG and THG, with converters consisting of both conventionally grown and rapidly grownKDP/KD*P crystals.

Conventional Rapid growth

Parameter SHG THG SHG THG

1w laserBeam size (cm) 34 34 34 30

DoublerSerial number 345-1 345-1 RG8B-2 RG8B-2Thickness (mm) 11.09 11.09 11.10 11.10Dq distribution (µrad int, 1 s) 22.3 18.8 17.5 27.8Surface loss (% before/after)

S1 (1w) 0.91/Ð Ð/1.49 1.03/Ð 1.64/1.67S2 (1w) 0.91/Ð Ð/1.49 1.03/Ð 1.64/1.67S2 (2w) 1.70/Ð Ð/2.53 1.48/Ð 2.85/3.30

TriplerSerial number Ð LL1-37-1 Ð RG8A-1Thickness (mm) Ð 9.48 Ð 9.41Deuteration level (%) Ð 70 Ð 85Dq distribution (µrad int, 1s) Ð 36.2 Ð 67.7Surface loss (% before/after)

S1 (1w) Ð 0.60/Ð Ð 2.75/3.05S1 (2w) Ð 1.82/Ð Ð 1.30/2.20S2 (3w) Ð 0.10/Ð Ð 0.37/2.21

Measured performanceMaximum energy efficiency 73 75 70.5 73.5At 1w irradiance (GW/cm2) 4.0 3.8 3.9 3.6

57

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

Crystals

Focus lens

Cell

37 cm

Crystal 328-62.0 lb/inch

Crystal 345-10.6 lb/inch

2.9 µm P-V

6.3 m P-V

Surf

ace

def

orm

atio

n (µ

m)

–2.4

3.9

(b)(a)

efficiency was 73.5% at an input 1w irradi-ance of approximately 3.6 GW/cm2. In comparison, the model, with an input fieldbased on near-field 1w irradiance data andan eleven-time-slice approximation of themeasured 1w pulse shape, predicted anenergy conversion efficiency of 77%, and apeak-power conversion efficiency of 79.5%.Including the 30-GHz bandwidth of thedrive pulse and the measured depolariza-tion in the Beamlet laser lowers the calculat-ed energy efficiency to 75%. Incorporatingthe additional losses caused by the degrada-tion of the sol-gel antireflection coatingsover the course of the experiment furtherreduces the efficiency to 71.5%, suggestingthat the model is accurate to within theuncertainty in the component transmissions.Calculated and measured near-field fluencedistributions for both the third-harmonicand residual second-harmonic fields were infairly good agreement as a result of havingthe orthogonal-polarization interferomerydata incorporated in the model. The energybalance in the model (the ratio of total ener-gy out of the converter to total energy intothe converter) was ~3% higher thanobserved, consistent with the actual trans-missions of the components in vacuum

being lower than the initial values modeled.Based on these results, peak-power 3w con-version efficiencies approaching 80% shouldbe achievable at NIF ICF drive irradiances,provided that high-quality antireflectioncoatings are maintained on the frequency-converter optics.

High-Fluence OperationHigh-energy operation of the final optics

was investigated as high-damage-thresholdfused-silica components became available.Third-harmonic fluences of up to 8 J/cm2

and NIF-equivalent energies of up to 9.6 kJin 3-ns square pulses were tested in a seriesof three campaigns (Figure 18) that pro-duced valuable data for extrapolating com-ponent performance and lifetime for theNIF.28 The tests culminated in a limitednumber of full-fluence shots through anintegrated final optics configuration,including frequency-conversion crystals,focus lens, and a diffractive optics packagecontaining a color separation grating (CSG)and beam sampling grating (BSG) on a sin-gle silica plate, a kinoform phase plate(KPP), and a debris shield. Time constraintsimposed by Beamlet shutdown allowed

FIGURE 15. (a) Cut-away view of the finaloptics cell showingmounting scheme forthe crystal and focuslens. (b) Measured sur-face figure of a mounteddoubling crystal. Darkbands are caused byinterference of the reflec-tions from the near-parallel front and backsurfaces of the crystal. (70-00-0499-0793pb01)

58

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

very little on-line characterization of theindividual diffractive optics prior to the inte-grated test. A low-damage-threshold versionof the KPP was tested previously and itsperformance reported elsewhere.29 The CSGconcept was used successfully at low fluencein an experiment conducted for the FrenchCommissariat a lÕEnergie Atomique. Results

of the high-fluence tests revealed problemswith CSG-induced beam modulation anddamage associated with the sol-gel coatingbeing thicker than l/4 near the step edges, aneffect previously identified as being responsi-ble for reducing CSG diffraction efficiency.Improved CSG designs under developmentare expected to eliminate this problem.

2ω c

onve

rsio

n ef

fici

ency

3ω c

onve

rsio

n ef

fici

ency

1.0(b)

(a)

0.8

0.6

0.4

0.2

0

1.0

0.8

0.6

0.4

0.2

0

0 1 2 3

I1ω (GW/cm2)

4

0 1 2 3

I1ω (GW/cm2)

54

Data, time-integratedPlane-wave modelPlane-wave model with x-t fillFull model w/diffraction β3ω, ∆θ, and ∆υ

Data, time-integratedPlane-wave modelPlane-wave model with x-t fill

FIGURE 16. (a) Plotcomparing measuredand calculated SHG effi-ciency versus 1w irradi-ance for conventional-growth doubler #345-1.(b) Comparison of mea-sured and calculatedTHG efficiency forrapid-growth doubler#RG8B-2 and tripler#RG8A-1.(70-00-0499-0794pb01)

59

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

1

0

2

4

3

6

5

7

8

9

10

0

FOC (crystals and lens)�FOC, debris shield�FOC, DOPs, debris shield

2 4 6 8 12 14 1610E1ω(kJ)

E3ω

(kJ)

Optics configuration:

0

0.85

Measured Model

(b)

∆θdblr = 690 µrad

F 2ω(J

/cm

2 )

0

0.75(a)

∆θdblr = 715 µrad

F 2ω(J

/cm

2 )

FIGURE 17. (a) Comp-arison of measured andmodeled 2w near-fielddistributions for conven-tional-growth doubler#345-1 at a drive irradi-ance of 3.9 GW/cm2.Measured and modeledconversion efficiencieswere 6.7% and 6.5%respectively at an angular detuning of 715 µrad (internalangle). (b) Similar com-parison for rapid-growthdoubler #RG8B-2 at adrive irradiance of 4.2 GW/cm2. Measuredand modeled conversionefficiencies were 6.6%and 7.0% respectively atan angular detuning of690 µrad. Sharp featuresin (b) are the boundariesbetween {101} (pyrami-dal) and {100} (prismat-ic) growth regions in thecrystal; conventionalgrowth material is allpyramidal.(70-00-0499-0795pb01)

FIGURE 18. 3w energyat the output of the finaloptics vs 1w energy atthe output of the laseramplifier, based on dataobtained during theBeamlet high-energy testmule campaigns scaledto a NIF beam size of1240 cm2. Pulse durationwas 3 ns square.(70-00-0499-0796pb01)

60

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

SummaryBeamlet has contributed to NIF technol-

ogy development in many areas anddemonstrated important aspects of the NIFdesign. Near-field modulation experi-ments at high power have established theB-integral limits and spatial-filter pinholesizes needed for controlling nonlinear rip-ple growth at high power and assuringsafe operation without damage. High-contrast 20-ns pulses have been demon-strated at NIF-equivalent energy andpower using prototypical pulse shapingtechnology. Pinhole designs have beendeveloped and tested for spatial filtering.These experiments resulted in a ±150-µradSS cone baseline pinhole for the NIF, oper-ating with neither closure nor back reflec-tion. Detailed wavefront and far-field irradiance measurements demonstrated (1) that the Beamlet 3w pulse meets spot-size criteria for NIF high-brightness mis-sions, and (2) have provided a baseline forvalidating NIF propagation codes andestablishing NIF optics specifications. Inaddition, efficient frequency conversion to the third harmonic was demonstratedwith conventional and rapid-growthKDP/KD*P crystals in a prototypical NIFfinal optics configuration.

AcknowledgmentsThe authors gratefully acknowledge the

many contributions from the large team oftechnicians, scientists, engineers, and admin-istrative staff who made this work possible.

Notes and References1. B. M. Van Wonterghem et al., Appl. Opt. 36, 4932

(1997).2. S. C. Burkhart, W. C. Behrendt, and I. Smith, ICF

Quarterly Report 5(1), 68, Lawrence LivermoreNational Laboratory, Livermore, CA, UCRL-LR-105821-95-1 (1994).

3. J. A. Caird et al., ÒBeamlet focal plane diagnos-tic,Ó Proc. Soc. Photo-Opt. Instrum. Eng. 3047, 239(1996).

4. J. B. Trenholme, Laser Program Annual Report,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-50021-75, 237 (1975).

5. V. I. Bespalov and V. I. Talanov, JETP Lett. 3, 307(1966).

6. R. L. Carman, R. Y. Chiao, and P. L. Kelley, Phys.Rev. Lett. 17, 1281 (1966).

7. E. S. Bliss, D. R. Speck, J. F. Holzrichter, J. H.Erkkila, and A. J. Glass, Appl. Phys. Lett. 25, 448(1974).

8. S. C. Burkhart, R. Wilcox, D. Browning, and F.Penko, ÒAmplitude and phase modulation withwaveguide optics,Ó Proc. Soc. Photo-Opt. Instrum.Eng. 3047, 610 (1996).

9. P. M. Celliers et al., Appl. Opt., 37, 2371 (1998).10. J. E. Murray, D. Milam, C. D. Boley, and K. G.

Estabrook, ÒSpatial filter issues,Ó presented atSolid State Lasers for Application to InertialConfinement Fusion, Monterey, CA (1998).

11. D. Milam et al., ÒPinhole closure measure-ments,Ó presented at Solid State Lasers forApplication to Inertial Confinement Fusion,Monterey, CA (1998).

12. J. E. Murray et al., ÒSpatial filter issues,Ó Proc.Soc. Photo-Opt. Instrum. Eng., 3047, 207 (1996).

13. J. Murray et al., ICF Quarterly Report 7(3), 99,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-LR-105821-97-3 (1997).

14. A. C. Erlandson, M. D. Rotter, D. N. Frank, andR. W. McCracken, ICF Quarterly Report 5(1), 18,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-LR-105821-95-1 (1994).

15. M. D. Rotter, R. W. McCracken, A. C. Erlandson,and D. Brown, ÒThermal recovery measure-ments on multi-segment amplifiers,Ó Proc. Soc.Photo-Opt. Instrum. Eng. 2633, 70 (1995).

61

BEAMLET EXPERIMENTS

UCRL-LR-105821-99-1

16. R. Hartley et al., ÒWavefront correction for staticand dynamic aberrations to within 1 second ofthe system shot in the NIF Beamlet demonstra-tion facility,Ó Proc. Soc. Photo-Opt. Instrum. Eng.3047, 294 (1996).

17. J. T. Salmon et al., ÒAn adaptive optics systemfor solid state lasers used in inertial confinementfusion,Ó Proc. Soc. Photo-Opt. Instrum. Eng. 2633,105 (1995).

18. C. La Fiandra, P. Mehta, G. Goldstein, R. A.Zacharias, and S. Winters, ÒNIF deformable mir-ror,Ó presented at Solid State Lasers forApplication to Inertial Confinement Fusion,Monterey, CA (1998).

19. P. J. Wegner et al., ÒWavefront and divergence ofthe Beamlet prototype laser,Ó presented at SolidState Lasers for Application to InertialConfinement Fusion, Monterey, CA (1998).

20. B. M. Van Wonterghem et al., ÒA compact andversatile pulse generation and shaping subsys-tem for high energy laser systems,Ó Proc. Soc.Photo-Opt. Instrum. Eng. 1870, 64 (1993).

21. J. K. Lawson et al., ÒNIF optical specificationsÑthe importance of the RMS gradient,Ó presentedat Solid State Lasers for Application to InertialConfinement Fusion, Monterey, CA (1998).

22. J. K. Lawson, D. M. Aikens, R. E. English Jr., andC. R. Wolfe, ÒPower spectral density specifica-tions for high-power laser systems,Ó Proc. Soc.Photo-Opt. Instrum. Eng. 2775, 345 (1996).

23. W. H. Williams et al., ÒModeling characteriza-tion of the National Ignition Facility focal spot,Ópresented at LASE Õ98, San Jose, CA (1998).

24. R. E. English Jr., J. Miller, C. Laumann, and L.Seppala, ICF Quarterly Report 7(3), 112, LawrenceLivermore National Laboratory, Livermore, CA,UCRL-LR-105821-97-3 (1997).

25. V. Karpenko et al., ICF Quarterly Report 7(3), 166,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-LR-105821-97-3 (1997).

26. P. J. Wegner et al., ÒFrequency converter devel-opment for the National Ignition Facility,Ó pre-sented at Solid State Lasers for Application toInertial Confinement Fusion, Monterey, CA(1998).

27. N. P. Zaitseva et al., J. Cryst. Growth 180, 255(1997).

28. M. Kozlowski et al., ÒLaser damage perfor-mance of fused silica optical components mea-sured on the Beamlet laser at 351 nm,ÓSymposium on Optical Materials for HighPower Lasers, Boulder, CO (1998).

29. P. J. Wegner et al., ÒThird-harmonic performanceof the Beamlet prototype laser,Ó Proc. Soc. Photo-Opt. Instrum. Eng. 3047, 370 (1996).

63UCRL-LR-105821-99-1

Large laser systems with high peakpower, such as the National Ignition

Facility (NIF), contain a great variety ofoptical components. These include lenses,mirrors, potassium dihydrogen phosphate(KDP) crystals, spatial filters, diffractiveoptics plates, debris shields, and so forth.The NIF beam has sufficient intensity toproduce adverse effects either on thesecomponents themselves or on their func-tionality as part of the laser. Laser-interaction modeling like that normallyused in treating ICF targets plays animportant role in ameliorating theseadverse effects.

We illustrate such modeling here withtwo examples. These arise from the needto avoid pinhole closure in the spatial fil-ters and the need to avoid laser-induceddamage to optical elements such as lenses.

Spatial filters remove high spatial fre-quency noise from the beam1 by focusingit through a pinhole. The material formingthe pinhole removes the most divergentrays, which are found at the outermostpart of the focused beam. At NIF intensi-ties (see below), the laser intensity at thepinhole edge can be sufficiently large tocreate a plasma that can expand into thepinhole and degrade the quality of lightpassing through the pinhole later in thepulse. This degradation, called pinhole closure, is examined below for variouspinhole designs.

To contain cost, large laser systems likeNIF necessarily operate near laser-induceddamage thresholds for some optical ele-ments. On the other hand, because of the

very large number of components, it isnecessary to establish safe operational lim-its and tolerable levels of damage risk. Itwould be difficult to establish such limitsfrom experimentation alone because of thelarge number of materials, coatings, andenvironments, and the need to scale fromresults of small-scale experiments to full-sized optical elements. Thus, it is useful toestablish theoretical models of laser-induced damage to aid understanding,interpretation, and application of empiricalresults; for example, there exists a goodunderstanding of the deleterious effects ofintensification due to nonlinear propaga-tion (self-focusing).2 In this article, we dis-cuss mechanisms of laser damage initia-tion on fused silica. (Except for the fre-quency-doubling and -tripling crystals, alloptics in the NIF Final Optics Assemblywill be fabricated from fused silica.) Theresults presented here are part of an exten-sive experimental-theoretical effort tounderstand 3w fused silica damage, inorder to ameliorate it and to devise qualityassurance tests of damage vulnerabilitysuitable for NIF.3 This 3w fused-silica laserdamage effort, led by M. R. Kozlowski, hasbeen documented in numerous BoulderDamage Symposium papers during thepast few years.

Pinhole ModelingIn a spatial filter, a pinhole subtending

a half-angle a removes incoming rays mis-directed beyond this angle, as illustratedschematically in Figure 1. Equivalently,

MODELING THE INTERACTION OF THE NIFLASER BEAM WITH LASER COMPONENTS

M. D. Feit

C. D. Boley

64

MODELING THE INTERACTION OF THE NIF LASER BEAM WITH LASER COMPONENTS

UCRL-LR-105821-99-1

the pinhole removes noise of spatial wavelength shorter than l/a. If the angleis too small, however, the wings of thebeam at focus can deposit substantial ener-gy on the pinhole material.4 For typicalNIF pulses, the intensities near the pinholeedge will range from several GW/cm2 toseveral TW/cm2, depending on the pin-hole size, the pulse shape, and the beamalignment. At these intensities, the materi-al will generate a plasma that expands intothe pinhole and that can cause difficulties.At lower electron densities, the plasmawill induce aberrations on the beam, whileat higher electron densities, the plasmawill deflect and absorb the beam.

We have modeled two types of pinholes,shown in Figure 2. The first type, which wehave treated more extensively than the sec-ond, consists of four azimuthally spacedblades, staggered in the longitudinal direc-tion (parallel to the beam), with the bladessuccessively rotated by 90¡. The use ofstaggered blades avoids plasma conver-gence at the centerline, in contrast to a cir-cular ÒwasherÓ type pinhole. The longitu-dinal separation of adjacent blades (typical-ly 2.4 mm) is sufficient to minimize interac-tions among the plasmas during the pas-sage of the beam. At the same time, the

blades are close enough so that each filtersthe beam in the far field. This is possiblebecause an aberrated beam has an extend-ed range near focus. The blades can be sit-uated either horizontally/vertically(ÒsquareÓ orientation) or at a 45¡ angle(ÒdiamondÓ orientation) relative to the far-field pattern of the square laser beam. Thelatter orientation performs better than theformer, because it allows more room forthe diffractive lobes of the beam at focus,which extend horizontally and verticallyfrom the central spot.

The second type of pinhole has a conicalshape, which is designed to refract the fil-tered light away from the beam rather thanto absorb it.5 Experiments have shown thatthis design is superior to the blade design.6

Cones pass at least twice the energy ofblades, and they also avoid a back reflec-tion problem encountered with blades.However, they are more difficult to modelthan blades. One reason is that their longi-tudinal length is 3 to 4 times that of theblade design, with the result that the inten-sity distribution at the entrance differs sig-nificantly from that at focus. In the workdescribed here, cones are treated via a com-paratively primitive model, which alsoapplies to blades.

f

Pinhole

r = fα

α

FIGURE 1. Schematicgeometry of a spatial filter (not to scale),showing incoming raysmisdirected by a givenangle brought to a com-mon point in the focalplane. (70-00-0499-0960-pb01)

FIGURE 2. Typical 4-leaf and cone pinholes.The size of the squareopening in this 4-leafpinhole is 2.7 mm.(70-00-0499-0961pb01)

65


UCRL-LR-105821-99-1

We expect the fourth-pass pinhole in theNIF cavity spatial filter (CSF) to present thegreatest problems. To compare this with the pinhole on the NIF transport spatial filter (TSF), we employ simple scaling argu-ments.6 Assuming that the plasma closurespeed is constant, one can show that the clo-sure time is sensitive to the f-number of thespatial filter, scaling approximately as f3,and that it also scales roughly with theinverse of the beam power. On NIF, the CSFand TSF have f-numbers of 31 and 80,respectively, while the power into the TSF isat most three times that into the fourth passof the CSF. Applying the anticipated scaling,we expect the fourth-pass CSF pinhole tohave a closure time about 1/6 that of theTSF pinhole. The former pinhole, therefore,should present more serious problems thanany other NIF pinhole. Its behavior wassimulated by experiments on Beamlet,which has two spatial filters of f-number 26,close to the value of the NIF CSF. TheBeamlet TSF was chosen because of its high-er beam power and its proximity to outputdiagnostics. The pinhole experiments wereperformed during 1997 and 1998 (Ref. 6).

In addition to experiments on Beamlet,pinhole-related experiments were conduct-ed with the Optical Sciences Laser (OSL)7 tounderstand the dependence of plasmaspeed on irradiance and material composi-tion. Here a blade was illuminated with apulse of intensity similar to that expected onpinhole edges, and the electron density ofthe ablated plasma was probed in time viainterferometry. This allowed for the study ofphenomena in an off-line setting. We usedthe results to explore parameter dependen-cies and to test our numerical models.

In the following discussion, we considerfirst the OSL experiments and then theBeamlet experiments.

Modeling OSL ExperimentsIn the OSL experiments,7 a strong pulse

of peak intensity 50Ð600 GW/cm2 andduration 5 to 15 ns struck a material blade(knife edge) at a right angle. The materialsof interest for pinholes were stainless steeland tantalum (Ta). These were chosenbecause we found that high-Z materialsgenerally exhibit slower closure rates than low-Z materials. The geometry is

illustrated in Figure 3. A probe beam waspassed over the blade and through themain beam, and its phase shift was mea-sured as a function of distance above theblade and time. In modeling this experi-ment,8 we set up the two beams and theblade, with many thin zones near the sur-face of the blade. These zones, which fol-lowed the material, expanded into the vac-uum as the material vaporized.

To describe the plasma, we employedLASNEX, a 2D Lagrangian radiationhydrodynamics code typically used in ICF

Ð0.90

Ð0.95

0

0 .05

0.10

Ð0.05 0 0.05

Main beam

Probe beam

z (cm)

r (c

m)

Blade FIGURE 3. Schematicillustration of knife-edgeexperiments on OSL.The intensity profile ofthe main beam is indi-cated. Both this beamand the probe beamhave circular profiles. (70-00-0499-0962pb01)

66


UCRL-LR-105821-99-1

target calculations.9,10 LASNEX calculatesthe absorption and refraction of the raysand the plasma properties. While it isbelieved to be the best available code forthe present purposes, this application rep-resents a low-energy extrapolation of itsnormal range. The code does not treat indetail, for example, the thermodynamics ofvaporization. Of primary interest here isthe electron density, which produces phaseshifts in the probe beam by decreasing thelocal index of refraction according to

(1)

where nc is the critical density (about 1021 cmÐ3 for light of wavelength 1053 nm).The optical path difference in waves along aray path is proportional to the line integralof the change in the index of refraction:

(2)

with the sign chosen to make f positive.Figure 4 shows typical LASNEX predic-

tions for the electron density midwaythrough a pulse of maximum intensity

175 GW/cm2, incident on a Ta blade. Notethat the expansion is nearly symmetricalabout the tip of the blade. The electron density decreases almost exponentiallywith distance, with the critical electrondensity located a few micrometers fromthe surface. The maximum electron tem-perature is about 40 eV, while the maxi-mum ionization state is about 19. Each ofthese quantities is moderately uniformwithin the bulk plasma but decreasessharply near the blade.

To find the phase change of the probebeam, we employed the postprocessingcode HOLOX.11 For typical plasma sizes atthe midpoint of a pulse, an electron densi-ty of about 1018 cmÐ3 produced a shift of asingle wave with insignificant deflection.Note that this is three orders of magnitudebelow the critical density.

In both experiment and simulation, thephase profile exhibited a regular behavior asa function of distance above the blade andtime. At a given time, the phase decreasedexponentially with distance above the blade.As time increased, the phase profiles flat-tened out in a regular manner. This is illus-trated in Figure 5, which shows the phaseprofiles at particular times for a stainless-steel blade, which we model as iron. Notethat the calculated phase change generally

n n n( ) / ,/

x x= − ( )[ ]11 2

e c

φλ

= − −( )∫ ndl

1 ,

Ð0.08

0

0.10

Ð0.08 0 0.08

1.000E + 18�1.468E + 18�2.154E + 18�3.162E + 18�4.642E + 18�6.813E + 18�1.000E + 19�1.468E + 19�2.154E + 19�3.162E + 19�4.642E + 19�6.813E + 19�1.000E + 20�1.468E + 20�2.154E + 20�3.162E + 20�4.642E + 20�6.813E + 20�1.000E + 21

A�B�C�D�E�F�G�H�I�J�

K�L�

M�N�O�P�Q�R�S

FIGURE 4. Electron density at t = 10 ns, for apulse of maximum inten-sity 175 GW/cm2 on a Tablade on OSL, as calcu-lated by LASNEX. Thecontour labels give theelectron density in cmÐ3.Dimensions are in cm.(70-00-0499-0963pb01)

67


UCRL-LR-105821-99-1

exceeds the measured phase change, but thatthe trends are similar. For both experimentand model, the overall behavior is summa-rized by a function of the form

(3)

where x is the distance above the blade.The parameters depend on intensity andmaterial. This functional form is valid fortimes greater than t0, the time required forthe electron density to reach an exponen-tial profile, which typically is a fewnanoseconds. Setting the numerator tozero, we see that c1 is the Ò1-wave speedÓ(the speed of the point with f = 1). Thespeed c0 is related to the rate at which theprofiles flatten with respect to position. Wealways have c1 > c0. The speed of the nthwave is c1 Ð c0 ln n (valid as long as it ispositive, which typically holds for n ² 10).

Figure 6 shows the calculated and mea-sured 1-wave speeds for a range of edgeirradiances. For stainless steel, the calculatedspeed reaches approximately 2.3 ´ 107 cm/s

at 1200 GW/cm2. This exceeds experimentby a constant offset of about 0.2 ´ 107 cm/s.The data for Ta cover a more limited rangeof intensities, reaching only 300 GW/cm2,and the speeds are smaller because of thelarger atomic weight. Again the calculatedspeeds exceed experiment by a fixed offset.Because of the offsets, the calculated speedsdo not appear to aim toward zero at lowintensities, which is clearly nonphysical.Indeed, LASNEX was not intended tomodel this regime.

From the OSL experiments, we learnedthat the plasma model gives results ofabout the right magnitude. Although thecalculated phase speeds exceed experi-ment, they do so in a systematic way. Thusthe model is conservative, in the sense thatit overstates the plasma effects.

Modeling BeamletExperiments

We extended the plasma calculations to 4-leaf pinholes,8 using a beam profileappropriate to Beamlet. Since the far-fieldrange of the beam exceeded the longitudinal

10Ð1

100

101

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

2

2 4

4

Time (ns)

6

6

Fe�Epulse = 12 J�Pedge = 344 GW/cm2

x (mm)

Phas

e (w

aves

)

FIGURE 5. Phase profiles versus distance from the blade, for a pulse of peak intensity 345 GW/cm2

illuminating a stainless-steel blade on OSL. The points denote measurements, while the lines give calculations. (70-00-0499-0964pb01)

φ x tc t t x

c t t, exp ,( ) =

−( ) −−( )

1 0

0 0

2.5

2.0

1.5

1.0

0.5

0

(×107)

0 200 400 600 800 1000 1200

Ta

Fe

Intensity (GW/cm2)Sp

eed

(cm

/s)

FIGURE 6. Speed of the 1-wave contour on OSL, forblades of stainless steel and Ta. The points denotemeasurements, while the lines give calculations. (70-00-0499-0965pb01)

68


UCRL-LR-105821-99-1

length of the pinhole, the plasma calculatedfor a single blade was replicated and rotatedto describe four blades. For a given pinholesize, the electron density in the square orien-tation was markedly more peaked than thatin the diamond orientation, because the far-field lobes intercepted the blades at a higherpower level.

To find the effects on the beam, a sepa-rate propagation calculation was neces-sary. We propagated a beam through theentire spatial filter, including the plasmas,via the code PROPN1 (Ref. 12). The plas-mas were used to set up phase screens forthe code, and the edge of each bladedefined a transmission mask. While thisscheme should be iterated, the first itera-tion gave interesting and useful results.

Figure 7 shows lineouts of the calculat-ed near fields, at the image plane located adistance twice the focal length from theoutput lens, for a Ta pinhole of half-angle100 µrad in the square orientation. The cal-culation is for a 20-ns pulse of energy 1 kJ.One sees that the high-frequency noise vis-ible on the incoming wavefront is indeedremoved for early time slices (less than 10 ns), when the blades filter the beamwithout generating significant plasmas.The intensity has been filtered to theexpected scale of l/a ~ 1 cm. For latertime slices, though, the beam breaks upinto successively larger regions, and thebeam contrast (the normalized variance ofthe intensity) steadily increases. The

characteristic size of the regions alsoincreases, reaching about 5 cm at 19 ns.This trend is also seen in experiment.Closure, defined as a 20% increase in con-trast, was observed at about 10 ns, whichis close to the calculated time of 11 ns.

When calculations are done for this pin-hole in the diamond orientation, however,the pinhole remains open. Such a trendwould be expected from the considera-tions given earlier. In experiment, the pin-hole also remaind open at this energy. Itfinally closed for a pulse of energybetween 2 and 3 kJ.

If Ta is replaced by stainless steel, then the diamond pinhole is predicted to closerapidly after 10 ns, in about the same man-ner as the Ta square orientation. The con-trast rises by an order of magnitudebetween 10 ns and 19 ns. The reason for theadverse behavior is the increased plasmaspeed. Experimentally, stainless steel alsowas observed to close more rapidly than Ta.

Although the calculations for 4-leaf pinholes are instructive and reveal trendssimilar to experiment, they are not readilyextended to cone pinholes. In addition, weare interested in describing misalignmenteffects, for which the plasma and propaga-tion calculations are even more burden-some than those described above. Hence asimple pinhole model has been devisedthat applies to both 4-leaf and cone pinholes and that can also describe mis-alignment.6 The model was developed wth

0.10

0.08

0.06

0.04

0.02

0–25 0 25 –25 0

x (cm)25 –25 0 25

Inte

nsit

y (G

W/

cm2 )

FIGURE 7. Horizontal lineouts, along the beam center, of the calculated near field for a 100-µrad Ta pinhole, in the square orientation. Thepulse duration is 20 ns, and the beam energy is 1 kJ. The lineouts give the incoming pulse (left) and the outgoing pulse at 8 ns (center) and13 ns (right). The horizontal dimension is in cm. (70-00-0499-0966pb01)

69


UCRL-LR-105821-99-1

J. E. Murray. A schematic view of the focalplane, as pictured in the simple model, isshown in Figure 8. We envision a roundbeam within a round pinhole, with the center of the beam displaced a distance rmis = fa from the center of the pinhole,where f is the focal length and a is the mis-alignment angle. The radius rb of the beam isarbitrarily chosen to enclose 99% of theazimuthally averaged energy. To apply themodel to a 4-leaf pinhole, we choose aneffective pinhole radius. For the square ori-entation, this is the pinhole half-width, whilefor the diamond orientation it is the averagedistance of a blade from the beam center.

with rpin the pinhole radius. In the denomi-nator, note that the speed cn depends on theintensity, which in turn depends on the pin-hole radius and misalignment. The depen-dence of speed on intensity was obtainedfrom OSL measurements, as describedabove, supplemented with LASNEXresults. Since the speed is not known for the glancing angle appropriate to a cone pinhole, we assume that it is the 90¡ speed, multiplied by a parameter g.Thus the model has two parameters: band g (with g = 1 for a 4-leaf pinhole).

Figure 9 shows the closure predictions for a diamond Ta pinhole of half-angle 100 µrad. The upper and lower lines corre-spond to beam powers of 135 GW and 160 GW, respectively, for a temporally con-stant pulse shape. The closure time decreas-es with misalignment, with the sharpest rateof decrease occurring for the smallest mis-alignment. Also shown are three data pointsfrom the Beamlet TSF, with powers in therange 135 to 160 GW. These closure timesdecrease from 20 ns at 2.5 µrad to about 13 ns near 12 µrad, although the error barsare appreciable. The model matches thisbehavior and predicts a closure time ofabout 7 ns at a misalignment of 20 µrad. In the fit we use b = 0.9, which follows from a fit to all data for 4-leaf pinholes.

Closure times for the stainless-steelcone, of half-angle 100 µrad, are displayedin Figure 10. The lines give predictions forbeam powers of 140 GW and 195 GW,with b = 0.9 and g = 0.1. The data are

rpin

rmis

rbeamβr beam

Plasma

FIGURE 8. Schematic geometry of beam and pinhole,as envisioned in the model of Eq. (4). (70-00-0499-0967pb01)

We suppose that the pinhole closeswhen the plasma ablated from the nearestedge, traveling at a particular speed, pene-trates to a fraction b of the beam radius.The choice of this speed is somewhat arbi-trary. It might seem reasonable to choosethe 1-wave speed, but the model worksmore satisfactorily with a faster speed, cor-responding to a smaller wave index. Inpractice, we have chosen the 0.1-wavespeed. From these assumptions, we canwrite the closure time in the form

(4)τβ

γ=

− −

−( )[ ]r r r

c I r rn

pin mis b

pin mis

,2010

Misalignment (µrad)0

25

20

15

10

0

5

Clo

sure

tim

e (n

s)

FIGURE 9. Closure timevs misalignment for adiamond Ta pinhole ofhalf-angle 100 µrad,according to model andBeamlet experiment, fora beam power in therange 135Ð160 GW. Thepulse is constant in time.Since the experimentalpoints vary somewhat inbeam power, the modelcurves are given for thelargest and smallestpowers. (70-00-0499-0968pb01)

70


UCRL-LR-105821-99-1

distributed within this range. As in the pre-vious case, the calculated closure timedecreases most rapidly at small alignments.Again the model is comparable with exper-iment, although it is unfortunate that thedata tend to congregate near a commonpoint. Additional data would be desirableto test the model more generally.

Finally, we use the model to predict clo-sure as a function of both pinhole size andmisalignment for a stainless-steel cone pin-hole. Here we attempt to simulate a tempo-rally shaped ignition pulse (Haan pulse) of14.8 kJ on NIF. On Beamlet, a Haan pulse of8.3 kJ was observed to close (i.e., barely pass)

at 20 ns. For the same misalignment, a tem-porally flat pulse closed at 2.2 kJ. Assumingsimple scaling, we suppose that a 14.8-kJHaan pulse can be simulated by a flat pulseof 3.9 kJ. Figure 11 shows how the closuretime increases as the pinhole half-angle isincreased from 100 to 150 µrad. With no mis-alignment, closure occurs after 20 ns (thepoint is barely visible in the corner). NIFspecifications allow for 10% misalignment,based on the pinhole radius. The calculationindicates that a 120-µrad pinhole providesample margin for a closure time of 20 ns. Inthis case, the edge power is 34 GW/cm2. Onthe basis of experiments and this modeling,the current NIF design calls for a 150-µradstainless-steel cone pinhole.

Initiation of Surface Laser Damage

The NIF final optics will operate at awavelength of 351 nm (3w or third harmon-ic of 1-µm light). Optical damage at 3w is amore severe problem than at first-harmonicwavelength. Damage in fused silica is typi-cally initiated on the surface or in a near-surface layer, usually most visible on theexit side of the optical element for reasonsdiscussed below. Experiments indicate thatdamage is initiated at subwavelength-sizedsites. Local heating of absorbing nanoparti-cles can result in material fracture due tothermal-induced stress. Also thermal explo-sion, in which the absorbing region growsin size with heating, can play a role.Damage in fused silica due to bulk defects,e.g., bubbles or inclusions, can also occurbut is expected to be ultimately less signifi-cant than surface damage because thesedefects are expected to be relatively rare.

This section discusses our effort tounderstand quantitatively the interaction ofhigh-power lasers with damage-initiatingdefects. Our theoretical description mustinclude optical propagation, absorptionand ionization of material, hydrodynamicsand shock wave propagation, thermal andradiation transport, elastic-plastic materialresponse and material failure. No compu-tational model at present contains all ofthe necessary physics. However, we have

2010Misalignment (µrad)

0

25

20

15

10

0

5

Clo

sure

tim

e (n

s)

150110 120Pinhole half-angle (µrad)

130 140

0%

20%

30%

40%

100

25

20

15

10

0

5

Clo

sure

tim

e (n

s)

10% �misalign-�ment

FIGURE 10. Closuretime versus misalign-ment for a stainless-steel cone of half-angle100 µrad, according tomodel and Beamletexperiment. The beampower is in the range140 to 195 GW. (70-00-0499-0969pb01)

FIGURE 11. Calculatedclosure time versus pinhole half-angle, ascalculated for variousfractional misalign-ments, for a stainless-steel cone. The closuretime for a 100-µrad pin-hole with perfect align-ment is about 25 ns. Thepulse simulates a 14.8-kJignition pulse.(70-00-0499-0970pb01)

71


UCRL-LR-105821-99-1

used several powerful computer codeswith numerical diffraction and scatteringmodels to extend our understanding of theinitiation of laser damage.13 In combinationwith carefully designed experiments, mod-eling can identify significant physicaleffects and scaling behavior.

First, we discuss several importantphysical effects accompanying laser inter-action with metallic surface contaminants.Comparing the results of experiment andmodeling, we address the following ques-tions. What difference does it make if thecontaminant is on the entrance or exit sur-face? What is the connection between plas-ma generation and damage? What is theeffect of the surrounding environment? Wethen discuss damage initiation at subsur-face absorbing nanoparticles embedded bythe finishing process. Finally, we point outthe danger posed by even nonabsorbingbulk defects.

Surface ContaminantsOptical surfaces can be contaminated by

small particles, say from tens of nanometersto hundreds of micrometers in size. Thesearise, for example, from dust, condensation,or debris from light interaction with chamberwalls and targets. Various types, sizes, andshapes of contaminant have been studied.The simulations presented here refer to artifi-cial ÒparticlesÓ deposited onto a silica sub-strate. These 1-µm-thick particles of C, Al, orTi were sputter-deposited through a maskand were either round or square. Damageinitiation experiments were carried out in the3w Laser Damage group labs and at the OSL.

For metallic particles, laser light isabsorbed in a thin skin depth leading tostrong heating and plasma formation. Thetemperature of the resulting plasma can beas high as 20eV resulting in multiple ioniza-tion of the material. Such a hot plasma is astrong radiator of UV and soft x rays. In thiscase, radiation transport dominates thermalconduction as a means of transporting ener-gy. This radiation is strongly absorbed in any surrounding air, causing heating andfurther ionization. In this case, the ionizationfront in the air can expand supersonically.Figure 12 shows our results for irradiating a

0

5

10

15

20

Ð400 Ð300 Ð200 Ð100 0 100

Tem

pera

ture

(eV

)

z (µm)

t = 1 ns

5 ns

13 ns

FIGURE 12. Temper-ature resulting fromabsorbing front-surfacemetallic particle onfused silica in air (8.5-nspulse on 1-µm-thick Al).Air interface at z = 0. Air (to left of expandingmetallic plasma) is heat-ed by absorption of ultraviolet radiationfrom plasma.(70-00-0499-0971pb01)

front-surface absorber in air. The hot Òshoul-derÓ seen at the left at late times is air thathas been heated through absorption of UVradiation. UV emission from the plasma canalso be absorbed in the substrate, creatingcolor centers that increase absorption seenby subsequent pulses.14 The strongly local-ized energy deposited in the substrate pro-duces strong shock waves that can causemechanical damage. Finally, the plasmaradiation can induce electronic defects in theglass that permanently change its character-istics and decrease the damage threshold forsubsequent pulses.

Laser damage usually is easier to induceon the exit surface. Conventional wisdompoints to Fresnel reflection and interferenceinside the material as the source of thisasymmetry.15 However, the predicted ratioof thresholds does not always hold, andthis effect should vanish for antireflection-coated optics. Another difference betweencontaminated entrance and exit surfaces isthat with absorption on the entrance sur-face, the plasma formed expands andshields the particle from the incoming laserlight; that is, further laser energy isabsorbed in the plasma itself. Consequently,the pressure pulse launched into the sub-strate is on the order of 10 kbar. With theabsorber on the exit surface, however, theplasma is formed at the interface of twosolid materials and confined. The high den-sity of the plasma means higher heat capac-ity, lower temperature, and much higherpressures (say, 60 kbar) leading to a lowerlaser damage threshold. Figure 13 comparesthese two cases.

72


UCRL-LR-105821-99-1

Subsurface AbsorbingNanoparticles

Experimental observation shows thatmicropit damage spots appear near thelaser damage threshold, particularly on theexit side of nominally clean fused silicasamples (see Figure 14). All micropits havecomparable sizes, with depth comparableto the width. Pits tend to be elongated,and cracks open preferentially normal tothe electric field polarization direction.

It is natural to think that such pits areinitiated by subwavelength particulateabsorbers in the subsurface layer. Suchan absorber might be a small contami-nant particle, for example, due to the

polishing process. Polishing a brittle sur-face can result in a thin layer of nearlyinvisible mechanical damage (micro-cracks). These cracks can serve as dam-age initiation centers, particularly ifsome absorbing material is trapped inthem. This appears to be the case for cer-tain ceria-containing polishing com-pounds. Because ceria is strongly absorb-ing at 3w, we will consider ceriananoparticles here. Heating the materialaround the absorber can result in furtherabsorption increase, thermal explosion,material ejection, and crater formation.

HeatingThe absorption cross section s for a

particle small compared to an opticalwavelength is given in Refs. 16 and 17.The fraction of light incident on the geo-metric cross section that is absorbed isgiven by a = s/pa2. Starting from a verysmall radius a, a will initially grow fasterthan linearly with a, and then, if theabsorberÕs dielectric coefficient is of largemagnitude, tends to saturate at large par-ticle size. Because of diffraction, thisabsorption fraction can actually be largerthan unity, i.e., more laser energy isabsorbed than that intercepted by theparticleÕs geometric cross section.

0

10

20

30

40

–40 –20 0 20 40

Pres

sure

(kba

r)

z (µm)

Air

Particleon front

Particleon back

SolidFIGURE 13. Pressureprofiles resulting fromfront- or rear-surfacecontaminants are differ-ent due to plasma con-finement for the rear-surface case. Highertemperatures were alsocalculated for the rear-surface case. (70-00-0499-0972pb01)

Beam direction

E E

S P2 µm

Beam direction

2 µm

FIGURE 14. Micropitlaser-induced damageobserved on initially pit-free fused silica surfaceshows characteristic subwavelength size andorientation with respectto polarization of laserelectric field (courtesy of F. G�nin).(70-00-0499-0973pb01)

73


UCRL-LR-105821-99-1

The size of the absorber is also typicallymuch smaller than a thermal diffusionlength. This simplifies treatment of thetemperature in the surrounding material,which can be treated as stationary. InFigure 15 we present the peak temperatureT0 for a ceria particle with refractive indexn = 2 + 0.2i. The light intensity wasassumed to be 3 GW/cm2, and the pulseduration was taken as 3 ns. The thermalconductivity was taken as 0.014 W/cm K.In the figure, ka = 0.5 corresponds to aceria particle with radius about 28 nm.Here k = 2p/l is the wave number of thelight, and a is the radius of the particle.The peak temperature grows as the squareof the particle size divided by the wave-length for small particles. This resultdemonstrates why the NIF third-harmoniclight is more dangerous than the funda-mental. The absorption efficiency dependson the particle size measured in wave-lengths, so it is effectively three times aslarge at the third harmonic.

Thermoelastic StressWhen the temperature distribution is

known, the thermoelastic stresses can becalculated.18 For example, Figure 16shows the xx component of stress for aspherical ceria particle embedded in fusedsilica. The peak temperature was assumedto be 1000 K, and the coordinates aregiven in terms of the particle radius.Notice that the radial stress (equatorial inthe figure) is compressive while the tan-gential or hoop stress (at the poles) is ten-sile. For an infinite medium, the stress dis-tribution is symmetric as shown. For anear-surface particle, the presence of thefree surface modifies the distribution, andmaterial failure most likely occurs initiallyaround the equator of the particle (theÒpoleÓ is directed toward the surface).

Glass DamageThe DYNA2D time-dependent mechan-

ical response code13 was used to modelthe damage to fused silica due to heatingof a near-surface subwavelength ceria par-ticle. A tensor damage model was used todescribe the mechanical damage of brittlematerial. For low loading, elastic waves

can propagate. At higher loads, plasticdeformation and tensile and compressivefailure can occur. The fact that materialstrength is larger for loads applied only ashort time is taken into account usingparameters determined from experimentson high-velocity projectile impacts on glass.

We considered a ceria particle embed-ded in glass. It was assumed that the glassoutside the particle is nonabsorbing. Theceria particle was described with the sametype of damage model as the glass, butwith different parameters.

In the runs presented below, we consider100-nm-radius particles at distances of 300 nm and 150 nm from the surface.Energy was deposited at a constant rate for3 ns, which corresponds to a laser fluence of10 J/cm2. Figure 17 shows the damage dis-tribution in ceria and glass for particlesplaced 300 nm under the surface at 1, 2, and2.5 ns, respectively. The figure indicates the

1.0

0.5

1.5

2.0

2.5

3.0

0.2 0.3 0.4 0.5

Tem

pera

ture

, eV

ka

FIGURE 15. Calculatedpeak temperatureincreases at a ceria parti-cle of radius a embeddedin fused silica. Laser flu-ence was 9 J/cm2 at 3w.(70-00-0499-0974pb01)

–2

–1

0

y (µ

m)

1

2

–2 –1 0x (µm)

1 2

FIGURE 16. Variation of xx component of stressnear embedded ceria par-ticle in silica. Peak tem-perature is 1000 K anddrops off as 1/r outsideparticle. Radial stress iscompressive and hoopstress is tensile at particlesubstrate interface.(70-00-0499-0975pb01)

74


UCRL-LR-105821-99-1

amount of damaged material. Because ther-mal expansion in ceria is much larger thanthat of fused silica (130 ´ 10Ð7 KÐ1 vs 7 ´ 10Ð7 KÐ1), thermal expansion generatesthe initial stresses and damage in the sur-rounding material. Simultaneously, theshock reaches the free surface and reflectsback, generating tensile stresses that easilydamage the material. As noted above,because of the free surface, fracture propa-gates from the equator toward the free sur-face as additional material fails as stressesare redistributed. The damaged regionforms a characteristic conical shape. At theclosest and farthest parts of the region, theparticle stresses are mainly compressive andinitially do not damage the material. Laterincreases in pressure and arrival of thereflected wave crush most of the materialwithin the cone. All mechanical resistanceto shear is destroyed within the cone. Thecrushed material is finally ejected, forming aconical pit similar to that observed in exper-iments. The velocity of ejection is not high,only 150 to 200 m/s. It takes a comparative-ly long time to evacuate the pits. If the com-pletely damaged material is taken as theeventual pit boundary, the estimated pitdiameter is about 800 nm and the depthabout 400 nm.

This model assumed fixed absorptionand does not take into account thermalexplosion, i.e., growth of absorption withcontaminant heating. We estimate thiseffect in the next section.

Thermal ExplosionWhen the temperature around the inclu-

sion reaches a critical value, a thermal explo-sion takes place.19,20 This involves the rapidexpansion of the heated region into theglass, which is then ionized. It occursbecause the plasma produced by the initiallyabsorbed light radiates UV, which is stronglyabsorbed in the matrix resulting in heatingand an increase in the absorption coefficientof the glass. The situation is very similar tolaser-supported ionization waves,20,21 themain difference being that absorption occursin a volume instead of just at a front.

Our analysis shows that the radius ofthe absorbing region tends to grow expo-nentially: a = a(0) exp G, where the growthfactor G is given by

500 nm

(a)

(b)

(c)

t = 1 ns

300 nm

t = 2 ns

t = 2.5 ns

FIGURE 17. Growth ofmaterial damage infused silica due toburied 100-nm ceria particle 300 nm belowthe surface. Damage ini-tiates near the particleequator and grows to the surface to form aconical region (a) after 1-ns irradiation, (b) 2-ns,(c) 2.5-ns. The materialinside the conical regionis eventually completelycrushed and ejected (calculation by D. Faux).(70-00-0499-0976pb01)

75


UCRL-LR-105821-99-1

(5)

For absorption at 3w, solid-state densityelectron ne, ionization potential I0 of 10 eV,fluence F of 10 J/cm2, and a scattering rate0.5 of the optical frequency, the growthfactor G is about 10. The plasma ball willrapidly grow to a size comparable to awavelength, after which exponentialgrowth ceases. For the parameters usedabove, this final radius a is about 500 nm.Any nanoparticle strong absorber will ini-tiate damage of at least this size.

Nonabsorbing Bulk DefectsThe danger posed by absorbing defects in

a transparent substrate has long been recog-nized and understood.22 Such defectsabsorb energy, and they heat and expand,thereby thermally and mechanically stress-ing the surrounding material. It is alsoknown that pure diffractive effects (e.g.,clipping at the pinhole) in high-power lasersystems can lead to laser-induced damageby causing intensity modulations that seednonlinear self-focusing.23,24 In this case,nonlinear refraction increases the local beamintensity level above the damage threshold.

For high-power laser systems, intensitymodulations due to purely transparentdefects may be capable of inducing damagewithout invoking any nonlinear effects. Bothnegative index defects (e.g., voids) and posi-tive index defects (e.g., high-refractive-indexinclusion) scatter light strongly, causingstrong localized intensity modulation. High-refractive-index inclusions are especially dangerous since they act like efficient focusing lenses.

The situation of interest here is in theborderline area of wave optics and geo-metric optics. Very small defects (size com-parable to a wavelength) with refractiveindex not very much different from that ofthe surrounding material can be treated byperturbative methods (Born approxima-tion, WKB method, etc.) or treated byparaxial wave propagation. In the presentcase, we are interested in defects up tomany wavelengths in size with very largedifferences in refractive index (e.g., Ð0.5 for a void up to +0.6 for a pure zirconia

inclusion). This situation cannot be treatedby paraxial optics since it involves strongreflections including total internal reflec-tions inside an inclusion.

The vector theory of electromagnetic scat-tering was worked out by Mie; it is describedin the classic book of Van de Hulst.17 To bedefinite here, we choose to model sphericaldefects for which the solution can be calcu-lated in a convenient form. In the case oflarger (compared to wavelength) spheres, itis adequate to use the scalar approximationfamiliar from spherical scattering in theSchroedinger equation. That is, we wish tosolve the scalar wave equation

(6)

where k0 is the free space wave number2p/l, and h(r) is the spatially dependentrefractive index. The refractive index isassumed to have the value h1 inside asphere of radius a, and the value h2 outsidethis sphere. For convenience, we define thematerial wave numbers k1,2 = k0 h1,2. Then,the solution can be written as a superposi-tion of spherical waves of the form

(7)

Here jl and nl are spherical Bessel func-tions and Pl is a Legendre polynomial. Theeffect of the scatterer centered at r = 0 isgiven by the phase shifts dl, which vanishidentically for no scatterer. The phase shiftsare determined from a transcendentaleigenvalue equation

(8)

where

(9)

GF

n I c=

+10

120

ωεe

Im

∇ + ( ) =202 2 0E k r Eη

E l i i

j k r n k r

P

ll

l

l l l l

l

= +( ) ( )

× ( ) ( ) − ( ) ( )[ ]× ( )

=

∞

∑ 2 10

2 2

exp

cos sin

cos

δ

δ δ

θ

tan δγγl

l l l

l l l

k j k a j k a

k n k a n k a( ) =

′( ) − ( )′( ) − ( )

2 2 2

2 2 2

γ ll l l

l l

k j k a

j k a=

′( )( )

76


UCRL-LR-105821-99-1

and the primes denote differentiation withrespect to argument. Although there are aninfinite number of terms in the summationin Eq. 7, the number of partial wave phaseshifts dl appreciably affected by the scatter-ing is proportional to the size of the scat-terer, i.e., roughly equal to ka. This is whyit is computationally difficult to treat verylarge defects.

The calculations reported here are forsmall defects (wavelength scale and less)since these are most likely to occur. Theintensifications important for damage initi-ation only become larger and persist overlonger distances for larger defects. On theother hand, spherical defects probablyexhibit the largest magnitude effect, espe-cially for large refractive index inclusions.Actual defects need not be perfectly spher-ical, of course. Our results serve to pointout the large intensifications, thus the seri-ous consequences, which can result fromtransparent defects.

VoidsThe distribution of intensity around a

typical small spherical void (refractiveindex unity) in fused silica is shown inFigure 18. A plane wave of intensity unity is incident from above. Because therefractive index in the void is lower thanthat of fused silica, the void acts like athick diverging lens. From a geometricoptics picture, light rays are bent stronglyaway from the axis, leaving a ÒshadowÓregion behind the void. It is the intensitymodulations on the edge of this shadowthat concern us here. Intensity maxima oftwice the incident intensity occur in thevicinity of the void. Further away, thesemaxima tend to die out. These modula-tions die out more slowly for larger voids (over a distance comparable with theRayleigh range ka2).

InclusionsDefects with larger refractive index

than the surrounding material are moredangerous since they act as concentratinglenses. These are not simple lenses, ofcourse, because they are Òthick,Ó i.e., thereis a large variation in optical path lengthover the incoming beam. Consider a spher-

ical inclusion of radius a. The phase varia-tion experienced by straight-ahead rayspassing through the sphere is given by

(10)

at transverse position x < a. Here Dn/n isthe relative change in refractive index.Expanding the square root yields a simpleestimate for the effective focal length asa/(2Dn/n). This estimate is reasonablyborne out by the wave-optical calculations,especially in that the focal length is propor-tional to the size of the sphere. The full cal-culation has to be carried out, however, todetermine the intensity at the (aberrated)focus.

The intensification factor can be large,even when the change in refractive indexis small. For example, the peak axial inten-sity (in units of the input intensity) down-stream from a 4-µm sphere of index 1.51 ina silica (n = 1.5) substrate is about 1.5. Amodest increase in refractive index of theinclusion to 1.6 increases the maximumintensity to 10 times the initial intensity.

Ð10

Ð5

0

5

10Ð10 Ð5 0

x (µm)

1 2 30Intensity (a = 3 µm)

5 10

z (µ

m)

FIGURE 18. Distribution of intensity in vicinity ofsmall spherical void in glass. Light incident from top. Note shadow region behind void and intensitymodulations. (70-00-0499-0977pb01)

∆ ∆φ = 2 1 2 1 2Ka n n x a( / )[ ± ( / ) ] /

77


UCRL-LR-105821-99-1

For a high-index inclusion, e.g., ZrO2, arefractory used in fabrication of fused sili-ca, the intensification is nearly 1000(Figure 19). Bulk inclusions of high refrac-tive index are thus expected to lead todamage even though they are unlikely tobe perfect spheres. Because of the highvulnerability of surfaces, bulk damageproduced by inclusions may be accompa-nied by associated exit surface damage.

ConclusionsIn this paper, we have discussed models

that describe two of the ways in which theNIF beam can interact adversely with lasercomponents. These interactions can lead topinhole closure in the spatial filters andlaser-induced damage to optical elementssuch as lenses. We have discussed a com-plex model for 4-leaf pinholes and a sim-ple model for pinholes in general, includ-

ing the cone design. The models have beentested via data taken on OSL and Beamletand appear to be adequate to make projec-tions for NIF. For the description of laser-induced damage, our modeling hasdemonstrated the importance of verysmall absorbing particles as well as nonab-sorbing particles in initiating the damage.We expect that such laser-interaction mod-eling, which involves a number of areas ofphysics and extensive numerical codes,will continue to contribute to the develop-ment of advanced laser technology.

AcknowledgmentsIt is a pleasure to acknowledge the con-

tributions of our colleagues to the workpresented here. These colleagues include J.M. Auerbach, K. G. Estabrook, D. R. Faux,F. Y. G�nin, M. R. Kozlowski, D. Milam, J.E. Murray, A. M. Rubenchik, S. Schwartz,and L. M. Sheehan.

Notes and References1. J. T. Hunt, J. A. Glaze, W. W. Simmons, and P. A.

Renard, Appl. Opt. 17, 2053Ð2057 (1978).2. See, e.g., J. T. Hunt, K. Manes, P. A. Reynard,

Appl. Opt. 32, 5973Ð82 (1993); for assessments forthe 1w chain, see M. D. Feit, C. C. Widmayer, W.H. Williams, R. A. Sacks, P. A. Renard, and M.A. Henesian, ÒNIFÕs assessment of the opticaldamage threat for flat in time pulses,Ó and M. D.Feit, W. H. Williams, C. C. Widmayer, R. A.Sacks, and P. A. Renard, ÒAssessment of thedamage threat for the two baseline ICF shapedtemporal pulses,Ó both in NIF Laser SystemPerformance Ratings, Lawrence LivermoreNational Laboratory, Livermore, CA, UCRL-ID-131115 (July 1, 1998).

3. M. R. Kozlowski, R. P. Mouser, S. M. Maricle, P.J. Wegner, and T. L. Weiland, ÒLaser damageperformance of fused silica optical componentsmeasured on the Beamlet laser at 351 nm,Ó p.428; S. Schwartz, M. D. Feit, M. R. Kozlowski,and R. P. Mouser, ÒCurrent 3w large optic testprocedures and data analysis for the qualityassurance of National Ignition Facility optics,Ó p. 314; M. D. Feit, A. M. Rubenchik, M. R.Kozlowski, F. Y. G�nin, S. Schwartz, and L. M.Sheehan, ÒExtrapolation of damage test data topredict performance of large-area NIF optics at355 nm,Ó p. 226; all in SPIE 3578: XXX AnnualSymposium on Optical Materials for HighPower Lasers, Boulder, CO, Sept. 28ÐOct 1, 1998.

4. J. M. Auerbach, N. C. Holmes, J. T. Hunt, andG. J. Linford, Appl. Opt. 18, 2495Ð2499 (1979).

5. P. M. Celliers et al., ÒSpatial filter pinhole forhigh-energy pulsed lasers,Ó Appl. Opt. 37,2371Ð78 (1998).

FIGURE 19. Light intensification near zirconia sphere(n = 2.1) in fused silica. Laser incident from top.(70-00-0499-0978pb01)

10

5

0

0 5x (micron)

z (m

icro

n)

Ð5

Ð5

Ð2 0 2Log (intensity) a = 4 and lambda = 0.355

78


UCRL-LR-105821-99-1

6. J. E. Murray, D. Milam, C. D. Boley, and K. G.Estabrook, ÒSpatial filter pinhole development forthe National Ignition Facility,Ó in preparation.

7. D. Milam et al., ÒPinhole Closure Measurements,ÓThird Annual International Conf. on Solid StateLasers for Application to Inertial ConfinementFusion, Monterey, CA, June 7Ð12 (1998).

8. C. D. Boley, K. G. Estabrook, J. M. Auerbach, M.D. Feit, and A. M. Rubenchik, ÒModeling ofLaser Knife-Edge Experiments,Ó Third AnnualInternational Conf. on Solid State Lasers forApplication to Inertial Confinement Fusion,Monterey, CA, June 7Ð12 (1998).

9. G. B. Zimmerman and W. L. Kruer, Comments inPlasma Physics and Controlled Fusion 2, 51Ð61(1975).

10. J. A. Harte, W. E. Alley, D. S. Bailey, J. L.Eddleman, and G. B. Zimmerman, ICF QuarterlyReport 6(4), 150Ð164, Lawrence LivermoreNational Laboratory, Livermore, CA, UCRL-LR-105821-96-4 (1997).

11. M. K. Prasad et al., Phys. Fluids 8, 1569Ð1575(1992).

12. J. M. Auerbach, Lawrence Livermore NationalLaboratory, Livermore, CA, internal communica-tion (1998).

13. Laser absorption, ionization, and hydrodynam-ics were modeled with the HYADES code; cf. J.T. Larsen, ÒHYADESÑA RadiationHydrodynamics Code for Dense PlasmaStudies,Ó in Radiative Properties of Hot DenseMatter, edited by W. Goldstein, C. Hooper, J.Gauthier, J. Seely, and R. Lee (World Scientific,Singapore, 1991). Mechanical strength was mod-eled with Lawrence Livermore NationalLaboratoryÕs DYNA code; cf. R. G. Whirley, B. E.

Engelmann, and J. O. Hallquist, DYNA2DÑANonlinear, Explicit, Two-Dimensional Finite ElementCode for Solid Mechanics User Manual, LawrenceLivermore National Laboratory, Livermore, CA,UCRL-MA-110630 (1992).

14. A. P. Gagarin, L. B. Glebov, and V. G.Dokuchaev, Kvant. Electron. 4, 1996 (1977);English translation in Sov. J. Quant. Electr. 7, 1136(1977).

15. N. L. Boling, M. D. Crisp, G. Dube, Appl. Opt. 12,650Ð60, (1973).

16. L. D. Landau and E. M. Lifshitz, Electrodynamicsof Continuous Media (Pergamon Press, New York,1960).

17. H. van de Hulst, Light Scattering by SmallParticles (Dover Publications, New York, 1981).

18. A. P. Boresi, Elasticity in Engineering Mechanics(Prentice Hall, New Jersey, 1965).

19. M. F. Koldunov, A. A. Manenkov, I. L. Pokotilo,Sov. J. Quant. Electr. 18, 345 (1988).

20. S. I. Anisimov and V. A. Khokhlov, Instabilities inLaserÐMatter Interaction (CRC Press, Boca Raton,Florida, 1995).

21. M. von Allmen and A. Blatter, Laser BeamInteractions with Materials (Springer, Berlin, 1994).

22. See, e.g., ÒModeling laser damage caused byplatinum inclusions in laser glass,Ó J. H. Pitts,Laser Induced Damage in Optical Materials: 1985,Boulder Damage Symposium, NBS SpecialPublication 746 (1986).

23. J. A. Fleck, Jr., and C. Layne, Appl. Phys. Lett. 22,467Ð9 (1973).

24. N. B. Baranova, N. E. Bykovskii, B. Ya.ZelÕdovich, and Yu. V. Senatskii, Kvant. Electr. 1,2435Ð49 (1974), translated in Sov. J. Quant. Electr.4, 1354Ð61 (1975).

79UCRL-LR-105821-99-1

The beam control and laser diagnosticsystems for the National Ignition

Facility (NIF) align the laser beam, diag-nose the beam, and control the beamÕswavefront. Accomplishing these tasksrequires approximately 12,000 motors andother actuators, 700 cameras and otherdetectors, and 192 wavefront sensors anddeformable mirrors.

System control components are locatedthroughout each beamline as illustrated inFigure 1. Each of NIFÕs numerous lasercontrol systems serves one or more of thefollowing three functions: laser beamalignment, beam diagnostics, or wavefrontcontrol. We designed many of the systemsto perform multiple functions and sharecomponents to reduce the costs and space

requirements for NIF. For instance, theinput sensors both align and diagnose theinitial laser pulse.

The specific requirements for align-ment include positioning the 192 beamswithin the 40-cm apertures of the lasercomponents, focusing them accuratelythrough the far-field pinholes of theamplifier-chain spatial filters, and delivering them to the precise locationsspecified on the target. All alignmentfunctions are accomplished automatical-ly by recording video images of referencelight sources and beams, calculatingwhat adjustments will achieve thedesired relative positions of the imagedobjects, and sending the correspondingcommands to system motors.

BEAM CONTROL AND LASERDIAGNOSTIC SYSTEMS

E. S. Bliss A. A. Grey R. D. KykerF. R. Holdener R. D. Demaret T. J. McCarvilleJ. T. Salmon C. M. Estes V. J. Miller KammJ. R. Severyn M. Feldman R. A. SacksR. A. Zacharias A. J. Gates C. E. ThompsonS. A. Boege W. E. Rivera R. L. Van AttaR. D. Boyd C. F. Knopp S. E. Winters

CSF tower

TSF pinhole & alignment towerTSF diagnostics tower

Deformable mirrorOutput sensor

Input sensorOptics inspection system

CSF = cavity spatial filter LM = laser mirrorPAM = preamplifier module TSF = transport spatial filter

Chamber centerref. system

Target alignment sensor

Precision diagnostics

1ω

3ω

Pulse sync system

Final optics assembly

PAM

AlignmentDiagnosticsWavefront

12,000 motors & actuators700 cameras & detectors2,000 closed loops

LM1

LM2

LM3

FIGURE 1. Major beamcontrol components fora single NIF beamline. (40-00-1097-2260pb02)

80

BEAM CONTROL AND LASER DIAGNOSTIC SYSTEMS

UCRL-LR-105821-99-1

The laser diagnostic systems measurethe beam energy, power vs time, and thenear-field transverse profile. The detectorsthat monitor these parameters arecalorimeters, fast photodiodes, streak cam-eras, and charge-coupled device (CCD)video cameras. Requirements for accuracyand reliability are very high.

Wavefront systems measure opticalaberrations on the laser output beams anduse a deformable mirror in the four-passamplifier of each beamline to compensate.The resulting improvement in beam quali-ty leads to higher frequency-conversionefficiency and provides better focusingcharacteristics in the target chamber.

For the sake of clarity, this article is orga-nized by functionÑlaser beam alignment,beam diagnostics, and wavefront control.The input sensor components that handlealignment are described in the laser beamalignment section, whereas the input sen-sorÕs diagnostic components are describedin the following section on beam diagnos-tics. Discussion of the front-end processors,another important element of laser control,appears in the ÒIntegrated ComputerControl SystemÓ article (see p. 21).

Laser Beam AlignmentIn the NIF, laser beam alignment is per-

formed in the following areas:

¥ Input sensor.¥ Spatial filter towers.¥ Output sensor.¥ Target area.

We discuss the design of the alignmentcomponents in each of these areas below.

Input SensorThe input sensor is located at the output

of the preamplifier module (PAM) of theoptical pulse generator. For alignment pur-poses, the input sensor must measure beampointing and centering, provide alignmentreferences, and provide a beam for align-ment through the rest of the system. Thesensor must monitor the output at threepoints within the PAM: the regenerativeamplifier, the beam shaping module, andthe multipass amplifier (for the design ofthese components, see ÒThe NIF InjectionLaser SystemÓ article (p. 1). The opticaldesign for this alignment system was driv-en by three factors: (1) resolution and field-of-view requirements, which are directlyrelated to performance of the sensorÕs func-tion; (2) cost, which limited the number ofoptical elements and control points; and (3) space, which required that the sensorpackage fit within the preamplifier trans-port-optics design footprint. Figure 2 is aphotograph of the instrumentation side of aprototype input sensor with beam pathshighlighted. The main beam passes on theother side of the package.

The sensor includes a CCD camera thatmeasures both the near-field and far-fieldintensity profile of the beams. For the farfield, the camera measures beam pointingand provides a pointing reference for thePAM control points. The same camera isused to produce a near-field image for

Alignment�laser

Input to�fiber bundle

Near- and�far-field�camera

Integrating�sphere and�

energy diode

Multipass�beam

Regen beam

Beam shaper�output

FIGURE 2. Photographof a prototype of theNIF input sensor pack-age that is being testedin the laboratory.Instruments and indica-tions of beam pathshave been labeled. (40-00-0499-0879pb01)

81


UCRL-LR-105821-99-1

beam-centering alignment. We designedthis camera to perform its near-field imag-ing without disrupting the far-field point-ing reference by adding insertable lenses tothe fixed far-field lens element and camera.

The continuous-wave (CW) alignmentlight source shown in Figure 2 is used toalign the rest of the system. It provides abeam of the same wavelength and size asthe regenerative amplifier (regen) beam.Although the regen beam itself can also beused, it is not always available betweenshots, and then only at the rate of onepulse per second. The CW alignment beam goes through the input-sensor beam shaper, where it is shaped into a27.5-mm-square beam.

Spatial Filter TowersThe spatial filter towers are line-replace-

able units (LRUs) that are installed fromabove at the center of the cavity spatial fil-ter (CSF) and transport spatial filter (TSF)(Figure 3). These towers hold multiple com-ponents that serve several functions foreach beam, as shown in Figure 4. Platformsholding the components for two beams aremounted at each of the four levels in thetowers. They provide a stable base for

injection optics, diagnostic optics, pinholeassemblies, and beam dumps. Figure 5 is aphoto of a prototype alignment tower plat-form with some of the components for onebeam mounted. This prototype has beenshown to meet precision requirements.

Each platform in the CSF and each align-ment platform in the TSF has a pinholeassembly that rotates around an axis parallelwith the beam propagation direction toaccurately position one of the multiple setsof pinholes in the focal plane of the corre-sponding spatial filter lenses. The same plat-forms also carry pointing references towhich the alignment beam from the inputsensor is aligned between shots. In the CSF,the reference takes the form of a reticle thatis positioned by selecting a particular posi-tion of the rotating pinhole assembly. A localilluminator, also mounted on the platform,provides light for viewing the reticle directlywith a camera on the same platform. In theTSF, as shown in Figures 4 and 5, the point-ing references are provided by fiber-opticlight sources moved into the pinhole planeby their own precision translation stage.Images of the TSF references are viewed inthe output sensor. Baffles, one of which isshown in Figure 5, control stray light. Fixedand prealigned beam dumps absorb energy

Cavity spatial filter

Transport spatial filter

�Alignment tower

Alignment tower�Diagnostics tower

FIGURE 3. Location of the three NIF spatial filter towers. (40-00-1097-2304pb02)

82


UCRL-LR-105821-99-1

from faults or leakage from the Pockels celland from target back reflections. The point-ing references and other alignment opticsidentified in Figure 4 are removed from thebeam path to fire the laser.

Output SensorThe output sensor and relay optics pack-

ages are located beneath the TSF center vessel, as shown in Figure 6. The sensor

3ω KDP sampledbeam reflecteddown to output

sensor

1ω SF4 sampled beamreflected down to

output sensor

Special splitter/cube reflectorshuttles to 4 positions:(1) direct Pass #1(2) direct Pass #4(3) direct view reticle-KDP(4) behind proective baffles at shot time

1

23

4

Pass #1

Pass #4

Alignment beamreflected down tooutput sensor

Input to Pass #1 from PABTS injectedfrom below TSF at 80¡ angle

Pass #3

Alignment sensor CCD

Special lens for viewingfour pinholes simultaneously

Tapered cone—see Ref. 1

Pinholes #1and #4

Pinholes #2 and #3

Beam dump on Pass #5 pathFast withdrawal (1 s before a shot)

Insertable half-waveplate pair on common

stage Passes #2 & #3

Alignment reticleilluminator

Pass #2

Pass #5

(b) Schematic top view of TSF components

(a) Schematic side view of CSF components

Beam dump forPasses #2 & #3

contained withinTower #1

Switches for homeand desired positions

Permanent bafflesbetween Passes #1 and #4

KDP beamdump—see Ref. 5

To target

3ω targetchamberalignmentfiber lightsource

x,y fine

In/out

1

2

3

4

m

m

m

m

m

m

mm

mKDP 1ωviewing screen

Insertable lightsource—see Ref. 4

SF4 Lens ReflectionSF4 Lens ReflectionSF4 lens reflection

Switches for home anddesired positions

Pinhole wheel—see Ref. 2

Pass #4 Pass #1

Pinhole wheel—see Ref. 3

FIGURE 4. Schematicviews of the (a) CSFtower and (b) TSF align-ment and diagnostictower components forone beam.(40-00-0499-0880pb01)

83


UCRL-LR-105821-99-1

and relay optics view the following, foralignment purposes:

¥ Injection beam at Pass 1 for injectionpointing.

¥ Pass 1 in near field for beam center-ing at the final optics.

¥ Pass 4 in far field for output pointing.¥ Pass 4 in near field for beam centering.¥ Reflection from the final optics at the

pinhole plane to adjust the angle of theKDP frequency conversion crystals.

These systems are required to centerbeams within 0.5% of the beam dimen-sion, locate the focused beams on shotpinholes within 5% of the pinhole diameter, position beams on target within 50-µm rms, and adjust the KDPcrystal angle within ±10 µrad over a fieldof view of ±200 µrad. Light for thesetasks is intercepted by a moving beam-splitter cube pickoff near the TSF pinhole plane on the alignment tower(see Figure 4b).

Light for all these functions is thendirected down to the output sensor along asingle path per beam. The pickoffs, relayoptics, and transport mirrors are staggeredto multiplex the eight beams spatially foreach bundle. Two beams per bundle ÒtimeshareÓ each output sensor, using beamsplitters and shutters in the transportpaths. The nominal beam size in the relaysis 20 mm.

Figure 7 is a photograph of the proto-type NIF output sensor package withbeam paths added for illustration. Onlythe 1w camera is used for alignment. It hastwo lens systems for near-field and far-field viewing and a motorized focusadjustment. The near-field lens system isshared with diagnostics (see the diagnos-tics discussion on p. 85), and the far-fieldlens system performs focused beam andreticle viewing functions. Both camerashave motorized continuously variableattenuators.

Target AreaWe have two main alignment systems

in the NIF target area: the chamber centerreference system (CCRS) and the targetalignment sensor. Figure 8 shows the lay-out of these two systems within the targetchamber. The target alignment sensorinserter uses the same positioner design asthe target inserters. These inserters, theCCRS modules, and target diagnostics aremounted on the same platform to mini-mize the relative motion among them. At

FIGURE 5. A prototype alignment tower platformwith the principal components for one beam.(40-00-0499-0881pb01)

Main�beams

Electronic�racks

Output sensor LRUs

Injection�beams

1ω + 3ω �diagnostic�beams and �relay optics

1ω alignment�beams and �relay optics

FIGURE 6. The output sensor packages and associated relay optics are locatedbelow the center part of the TSF. (40-00-1097-2249pb02)

84


UCRL-LR-105821-99-1

shot time, the alignment sensor isremoved and protective baffles are posi-tioned in front of the CCRS port windows.

The CCRS must provide a stable posi-tion reference system in the target bay andbe able to position targets repeatably with-in the narrow field of view (FOV) of thetarget diagnostics. Its long-term stabilitymust be ²30 µm, and its FOV must be ±5 cm from the target chamberÕs center.

The system has two identical viewinginstruments that have been adjusted toidentify a common position in the centerof the target chamber, and each has twosimultaneous modes of operation. Onemode measures a componentÕs targetchamber position by imaging the positionof a reticle on the component. The othermode measures the componentÕs orienta-tion by monitoring the direction of lightreflected from the reticle.

The target alignment sensor collectsimages of alignment beams in the targetplane and provides data for controlling beam

Target alignment sensor

Target alignment�sensor assembly�

No. 1

Target alignment�sensor assembly�

No. 2

Target chamber�(upper half removed)

Chamber center�reference system No. 2

Chamber center�reference system No. 1

Target�positioner�assembly

FIGURE 7. This photo-graph shows the princi-pal components andbeam paths of the proto-type NIF output sensorpackage being tested inthe laboratory. The out-put sensor performsboth alignment anddiagnostic functions.(40-00-0499-0882pb01)

FIGURE 8. Layout ofthe target area align-ment systems. (40-00-1097-2251pb02)

85


UCRL-LR-105821-99-1

positions and spot size. The total deviationfor all beams must be ²50 µm rms; a singlebeam must deviate no more than 200 µm;and the sensor must achieve a specified cen-tral lobe size to ±50 µm. The sensor mustoperate everywhere within 5 cm of the targetchamber center. Figure 9 shows a schematicof the setup for the target alignment sensorand a photo of the prototype. The sensor hastwo CCDs, which see both the target and thebeams. The assemblies that reflect the beamswere designed for minimal deflection andhigh natural frequency.

Beam DiagnosticsNIFÕs beam diagnostics characterize the

beam at key locations in the beamline(Figure 10). These systems include the:

¥ Input sensor.¥ Output sensor.¥ Calibration calorimeters and final

optics diagnostics.¥ Temporal diagnostics.¥ On-line optics inspection system.¥ Target chamber diagnostics.¥ Supplemental diagnostics:

Ð Roving assemblies.Ð Trombones.Ð Precision diagnostics.

We briefly discuss the requirements anddesign activities of each of these diagnosticsystems below.

Input SensorThe input sensor, in addition to provid-

ing certain alignment functions (see p. 80

and Figure 2), characterizes the PAM bysampling at the output of the regenerativeamplifier, the beam shaping aperture, andthe multipass rod amplifier. The sensormeasures beam energy, near-field images,and temporal pulse shape. The imagingresolution is 1% of the beam dimensionand 2% of the maximum fluence.

Diagnostic samples are obtained from a1% partial reflector for the regenerativeamplifier and through leaky mirrors for thebeam shaper and multipass amplifier. Thecoatings on these mirrors provide adequatesignal levels to the energy diagnostics, aswell as to the alignment diagnostics.

The energy from the regen, beamshaper, or four-pass amplifier is measuredwith an integrating sphere and photodi-ode followed by a charge integrator anddigitizer. Shutters select which sample ismeasured and a CCD camera obtains near-field images for each beam prior to or dur-ing a shot. An optical fiber bundle sends asample of the multipass output to thepower sensor (see Figures 10 and 13),where it is time multiplexed with othersignals. The PAM output beam can also bediverted to a calorimeter to periodicallycalibrate the multipass energy diagnostic.

Output SensorThe output sensor performs many diag-

nostic tasks in addition to its alignmentfunctions (for alignment discussion, see p. 82). The sensor characterizes the 1w out-put (energy, near-field fluence profile, tem-poral pulse shape, and wavefront) and 3woutput (near-field fluence profile and tem-poral pulse shape). The 1w output energy

(a) (b)

CCD

CCD

Incoming�beams

CCDs see both the�target and the beams��The target is viewed�from three sides��No light hits the target��Design will vary for�different targets

¥��¥��¥��¥�

FIGURE 9. (a) A sche-matic illustrating thesetup concept for targetalignment. (b) Photo-graph of the prototypetarget alignment sensorcurrently under test.(40-00-0499-0883pb01)

86


UCRL-LR-105821-99-1

must be measured within 3%, and the 1w and 3w temporal pulse shape must bemeasured within 2.8%. The 1w beamwavefront must be measured within 0.1 wave, and the 3w output beam must be imaged at the plane of the conversioncrystal with a spatial resolution of 3.2 mm.Figure 7 shows the output sensorÕs princi-pal components and beam paths, andFigure 11 shows a schematic for the outputsensorÕs diagnostic functions.

Reflections from optics in the main beamsupply samples of the 1w and 3w beams.The 1w sample reflects from the front

surface of a beam splitter at the output ofthe TSF. This surface is coated with a solgelantireflection coating (nominally 0.1% R),and the splitter is tilted by 0.8 mrad to off-set the reflected sample from the path ofthe existing beam. The 3w sample reflectsfrom the flat entrance surface of the targetchamber lens in the final optics assembly.This lens is also tilted by 0.6 mrad to theopposite side to offset the reflected samplebeam from the exit path, and its coating issimilar to that on the SF4 lens samplingsurface. Pickoffs for the beam samples arenear the focus in the TSF. Pickoff optics for

Input sensor OutputsensorPAM

3ω energy

FO inspection& pulse sync.

Back-refl.sensor

Portable sensor

Precision diagnosticsLaser optics inspection

Calibration calorimeters

CSF

TSF

Output beamsplitter

FOA

AMP

AMP

Rovingassemblies

Pol. LM2

LM4LM8

LM3

LM1

Power sensor

TCC

Insertablecalorimeter

FIGURE 10. Location of the laser diag-nostics for NIF. (40-00-1097-2254pb02)

3ω

Alignment

1ω

3ω NF imagesVar. atten.

NF telescope

1ω fiber bundle�to power sensor

3ω fiber bundle�to power sensor

NF

FF

1ω NF and FF images

Shutters

Wavefront�sensor�Energy�

PD

Shot-time atten.�in relay optics

3ω power

1ω power

CCD

Atten. diffuser/�integrator

FIGURE 11. Output sensor schematicfor beam diagnostics. NF and FF meannear field and far field, respectively.PD means photodiode. (40-00-1097-2256pb01)

87


UCRL-LR-105821-99-1

1w and 3w sample beams are located onTSF Tower #2, as shown in Figure 4. Relayoptics transport the beams from the TSF tothe output sensors beneath the TSF centervessel. Each side of the output sensor col-lects data for two beams as illustrated inFigures 7 and 11.

The output sensor has three CCD cam-eras, each with continuously variable atten-uators. One cameraÑshared with thealignment functionsÑimages the 1w near-field profile, the second images the 3wnear-field profile, and the third is the detec-tor array for a Hartmann wavefront sensor(see p. 93 for a discussion of the Hartmannsensor). The 1w energy is measured by anintegrating sphere with a time-integratedphotodiode, which is inserted at shot time.Power samples are sent to the power sensor (see Temporal Diagnostics at rightcolumn) using two optical fiber bundles.

Calibration Calorimeters andFinal Optics Diagnostics

Calorimeters, which measure beamenergy, are used in three areas of NIF: theinput sensor, the output sensor, and thefinal optics diagnostics. Each input sensorhas a port for manual mounting of a 5-cmcalorimeter to calibrate the sensorÕs energydiode without opening the beamline(Figure 2). The output energy diodes arecalibrated using two groups of eight rov-ing bundles of 50-cm calorimetersÑonefor each laser bay. Each group can beremotely positioned to intercept the out-puts from one eight-beam bundle at a

time. Finally, 192 10-cm calorimeters in thefinal optics diagnostics measure a fractionof each beamÕs 3w energy as it propagatestoward the target.

We use calorimeters similar to those onNova. The 5-cm calorimeters are an off-the-shelf design, and the 50-cm and 10-cmcalorimeters are scaled versions of the 40-cmones used in the Nova target chamber. Thesecalorimeters can meet the NIF requirements.

The final optics diagnostics uses adiffractive splitter to obtain a sample forthe 3w calorimeter (Figure 12). Thiscalorimeter calibrates the 3w power foreach shot. It must operate in a vacuum,and have a damage threshold >3 J/cm2 at351 nm, a 10- ´ 10-cm aperture, a 1- to 60-Jenergy range, a repeatability of <1% at the30-J level, and a linearity of <1% over arange of 50:1. The sampling grating on theflat surface of the focus lens diverts <1% ofthe 3w energy and focuses the sample infront of the calorimeter.

Temporal DiagnosticsTemporal diagnostics includes two

portable sensors, rack-mounted powersensors near the input and output sen-sors, and a back-reflection sensor. NIFwill have two portable streak cameras,mounted on carts, which can each beused in place of a normal power sensorfor one beam at a time. The streak cam-eras must have a time resolution of 10 ps,a dynamic range of 1000:1, and multiplechannels, and be easily movable amongthe other sensor packages. Each camera

To thetarget

12°

3ω calorimeterand enclosure

Diffractive splitterin a removable cassette

FIGURE 12. The finaloptics diagnostics uses adiffractive splitter andvolume-absorbingcalorimeter. (40-00-1097-2257pb02)

88


UCRL-LR-105821-99-1

can handle 19 sample inputsÑ1w or3wÑthrough fiber-optic bundles. One offour sweep times can be selectedÑ1.5 ns, 5 ns, 15 ns, or 50 nsÑwith resolutionsfrom 10 ps (for 1.5 ns) to 250 ps (for 50 ns).

Each power sensor must have a dynam-ic range of 5000:1, a record length of 22 ns,an accuracy of 2.8% averaged over a 2-nsinterval, and a rise time of 450 ps. It takessamples on fiber bundles from the outputand input sensors, and time multiplexessignals as shown in Figure 13 to minimizecosts. The transient digitizer in the sensoris a commercial technology with a longrecord length. Each 1w photodiodereceives signals from four input sensorfiber bundles and eight output sensor fiberbundles. Each 3w photodiode receives sig-nals from four output sensor fiber bundles.Time separation is achieved using thepropagation time through the laser andoptical fiber delay lines for signals close intime. The dynamic range of the eight-bitdigitizer is extended using four channels,each with a different sensitivity.

On-Line Optics Inspection System

One on-line optics inspection system islocated in each switchyard, and one more attarget chamber center. These systems haveaccess to each beamline through a set oftranslatable mirrors described later (seeSupplemental Diagnostics on p. 90). The

requirement for main-beam optics is todetect flaws ³3 mm. Our goal is that thesesystems detect defects ³0.5 mm so that theoptics can be removed for refinishing whiledefects are still small. We considered anumber of issues when designing thesesystems. First, we selected dark-field imag-ing since damage spots appear best againsta dark background. This allows us to detectspots below the resolution limit. Our imag-ing scheme is to backlight the large opticswith apodized and collimated laser illumi-nation sources. Undisturbed light is inter-cepted by a stop in the dark-field optics,but light diffracted from damage spots isimaged onto a CCD camera. Second, thesystemsÕ resolution will be limited by oneof two factors: the far-field aperture in theTSF or the number of pixels in the camerasÕCCD arrays. We must properly account forthese limitations in the design. Third, depthof field will be short enough to isolate allbut the most closely spaced optics. Fourth,use of different pinhole combinations in theCSF and careful image processing canÒstrip awayÓ any overlaid images.

The laser bay optics are inspected withthe high-resolution cameras located in eachswitchyard. Figure 14 shows the layout ofa switchyard inspection package. Foroptics in the main laser, the illuminationsource is the alignment laser located in theinput sensor. This alignment beam is inject-ed into the TSF along Pass 1, and LM1 isaligned to return the beam along Pass 4.

Fiber bundles

Optical delays

PhotodiodeVar. atten.

Lenses

Transient digitizers�at 4 sensitivities

FIGURE 13. Schematicdiagram of a power sensor in which time-multiplexed opticalpulses are combined ona common photodiode,and the electrical out-puts drive multiplechannels in transientdigitizers.(40-00-1097-2258pb02)

89


UCRL-LR-105821-99-1

The inspection package captures dark-fieldimages at each plane containing an optic.Image subtraction software will help detectthe changes from previous inspections.

For inspecting switchyard and targetoptics up to the first surface of the finalfocus lens, the switchyardsÕ 1w source willbe injected toward the target using the outward-looking roving mirror (p. 91). Thefirst surface of the final focus lens will bealigned to retroreflect a portion of thisbeam. Dark-field images will be capturedby the inspection system in the switchyardat each plane containing an optic.

For inspecting the final optics, we willimage through the kinoform phase plate at3w using a damage inspection packageinserted at the target chamber center(described further in the next section). Thisviewer will have a sufficiently short depthof field to discriminate between the closelyspaced final optics elements.

Target Chamber DiagnosticsTarget chamber diagnostics include the

pulse synchronization detector module andthe target optics inspection system. Both are

located at the center of the target chamberat the end of the diagnostic instrumentmanipulator, as shown in Figure 15a.Similar adjustments are required for boththe pulse synchronization and final opticsinspection modules. They can be orientedto any beam position. Translation com-mands for the diagnostic manipulator andangle commands for the modules comefrom the CCRS (see Target Area on p. 83).The maximum move time to intercept lightfrom a different four-beam quad is 5 s; thetypical time is <0.8 s.

Figure 15b shows the components ofthe synchronization module. The pulsearrival times at target chamber centermust be set with 20-ps relative accuracy.The module must simultaneously capturesignals from the four beams of a finaloptics assembly quad and position eachfocused beam on the end of a separatefiber bundle with an accuracy of <100 µm.The fiber bundles carry the optical signalsto a streak camera where their relativetimes of arrival are compared. The signalis obtained by firing a rod shot and capturing the leakage of 1w radiationthrough the conversion crystals.

Laser(1, 2, 3ω) Atten.

Up-collimator

Up-collimator

Apodizer

Insertable mirror

Beamline

Insertablecorner cube

FF lens

Insertable stop

NF lens

Var. atten.

Hi-res. CCD (1000 × 1000)

Optics inspection sensor package FIGURE 14. Inspectionpackage used for all butthe final optics. Onesuch inspection systemresides in each switch-yard. NF and FF meannear field and far field,respectively. Light isdirected into or from thepackage by an insertablemirror.(40-00-1097-2259pb02)

90


UCRL-LR-105821-99-1

The final optics damage inspectionmodule, also located at target chambercenter, is shown in Figure 15c. The moduleexamines the closely spaced final opticsthat are not observable by other means.The final optics are backlit with collimated3w radiation. A high-pass filter, with a central schlieren stop placed at the focus,removes those rays that are not deviatedby flaws. A CCD camera captures datafrom the full-aperture image. The camera,lenses, schlieren stop, and filters rotate as aunit in the azimuthal direction.

Supplemental DiagnosticsAs described above, the NIF design

includes significant diagnostic capabilitieson each beamline. However, it will beimportant to be able to calibrate some ofthese measurements, to verify that they areoperating correctly, or to collect moredetailed information. For this purpose,additional diagnostics that can be used onone or a few beams at a time are located in each switchyard. They include full-aperture calibration calorimeters and asuite of precision diagnostics.

Figure 16 shows the layout of compo-nents related to supplemental diagnosticsin one of the switchyards. An array of

eight calorimeters, the Òroving calorimeterassembly,Ó travels on horizontal rails toany of the 12 bundle locations. If desired,each of the eight calorimeters can collectthe output from the corresponding beamin that bundle. However, any of the eightcalorimeters can also be rotated towardthe laser on its outside vertical edge sothat it allows the beam to pass. Beams thatare allowed to pass continue on to the tar-get chamber or are directed to the preci-sion diagnostics as described below.

The optical-mechanical system designedto intercept any one beam in each switch-yard and send it toward the precision diag-nostics is called the Òroving mirror system.ÓIt comprises an additional pair of parallelhorizontal rails, two pair of parallel verticalrails, and three translatable mirrors. The x-ytop mirror (Figure 17a) in combination withthe y mirror picks off a beam and diverts ittoward the precision diagnostics and opticsinspection package. The x-y bottom mirror,combined with the y mirror, provides apath from the optics inspection package tothe target chamber (Figure 17b). Both rov-ing systems are inside of a large enclosurethrough which the laser output beams pass.

Each of the roving calorimeter and x-ymirror assemblies weighs about 1000 lb.The horizontal motion of the assemblies is

Chamber

Incomingbeams

(a) (b) (c)

Diagnosticinstrumentmanipulator

Connection to the diagnostic instrumentmanipulator

Pulse synchronizationmodule/final optics

damage inspection module

Connection tothe azimuthdrive motor

Fiber couplingplate

Firstlens

Azimuthalbearings

Second lens

2K × 2KCCD camera

Schlierenstop

Elevationtrunnion

Elevationmirror

Elevationmirror

Azimuthbearings

0.8 mFIGURE 15. (a) The pulse synchronization detector module and final optics damage inspection module can be attached to the end of thediagnostic instrument manipulator and inserted into the center of the target chamber; (b) exploded view of the pulse synchronization mod-ule; (c) exploded view of the final optics damage inspection module. (40-00-1097-2261pb02)

91


UCRL-LR-105821-99-1

Laser bay/switchyard wall with some of the 96 beam penetrations shown

m

mm

mm

mm

m

m

m

m

m m

mmm

m

mmmm

m m

mm

mm

m

mm

m

m m m

Trombonelength

adjustment

Precision�diagnostics (3ω)

Precision diagnostics (1ω)

Optics inspection systemPM3

x-y bottom mirror in storage

y mirror

Roving�mirrors

Roving calorimeters

Insertable calorimeter

Final optics assembly�with removable IOM �

x-y top mirror�in use

FIGURE 16. Supple-mental diagnostics forcalibration and for avariety of detailed mea-surements are located inthe switchyard. (40-00-1097-0381pb02)

Optical path frominspection diagnostics

x-y bottom translatable mirror (injectsinspection beam towardtarget chamber)

x-y toptranslatable

mirror

y translatablemirror

Optical path to

targetchamber

(b) Inspection diagnostics to target chamber

y translatablemirror

Beam comingfrom laser bay

Beam going todiagnostics and

optics inspectionpackages

x-y bottomtranslatable mirror

x-y top translatablemirror (intercepts beam

from laser)

(a) Laser bay to precision diagnostics FIGURE 17. Two rovingmirror system configura-tions provide (a) a pathfrom the laser bay to theprecision diagnostics sta-tion and (b) a path frominspection diagnostics tothe target chamber. (40-00-1097-2262pb03)

92


UCRL-LR-105821-99-1

belt driven by motors fixed at one end ofthe enclosure. The average time to movefrom one beam to an adjacent beam is esti-mated at 16 s. The enclosure maintains anargon gas environment, and the mecha-nisms must be designed to avoid produc-tion of particles that might contaminate thenearby optics. The path length from theroving mirror to the precision diagnosticstations for each beam is matched to thatbeamÕs path length to the target chamberby adjustment of an optical Òtrombone.Ó

The precision diagnostic stations areshared diagnostics that measure laser outputperformance one beam at a time using moreextensive instrumentation than that found inthe output sensors (Table 1). During installa-tion and activation, the precision diagnosticswill be used to verify the performance ofeach beam, including its dedicated diagnos-tic packages. The 3w precision diagnostics,illustrated in Figure 18, will measure fre-quency conversion characteristics of theselected beam using a separately selectedintegrated optics module (IOM). Each IOMcomprises the set of optics, including fre-quency conversion crystals and final focus

lens, that is normally mounted at eachbeamÕs entrance to the target chamber. Theprecision diagnostics provide the only capa-bility for simultaneously measuring high-power 3w beam properties at the full 40-cmnear-field aperture and in a far-field planeequivalent to the target chamber focus. OnceNIF is operational, the station will be avail-able for diagnosing beamline and compo-nent problems and for performing laser sci-ence experiments.

The precision diagnostic station will beable to measure the following aspects ofthe 3w laser pulse:

¥ Energy, with an accuracy of 2.7%.¥ Power vs time, with an accuracy of

3.1% and a rise time of 450 ps for allparts of a 22-ns pulse having a 50-to-1contrast ratio.

¥ Focused spot size and smoothness,with 30-µm spatial resolution and 20-ps temporal resolution for a selected1.5-ns period.

¥ Near-field spatial profile in the fre-quency conversion crystal plane,with resolution as high as 300 µm.

TABLE 1. The 1w measurement capabilities of the precision diagnostic station compared to those of the output.

Measurement Precision diagnostic Output sensor

EnergyRange To 22 kJ To 22 kJAccuracy Better than 1.5% 3%

Power resolution <40 ps or 100 ps 450 ps

Far-field imaging FOV ±180 µrad (best) None

Near-field imaging resolution (in main beam) ~1.4 mm and/or ~300 µm 3.2 mm

WavefrontHartmann precision Better than l/20 l/10Radial shearing interferometer precision Ð0.1 l (16´ reference) None

SchlierenEnergy balance Better than 15% NonePower resolution <40 ps or 100 ps NoneFar-field imaging FOV ±500 µrad NoneNear-field imaging resolution ~1.6 mm (in main beam) None

Prepulse sensitivity 2 ´ 104 W/beam for 200 ns Nonepreceding foot pulse

Potential added measurements Versatile Fixed

93


UCRL-LR-105821-99-1

Wavefront ControlThe three main components of the

wavefront control system are the:

¥ Hartmann wavefront sensor.¥ Deformable mirror.¥ Computer controller.

The use of lasers as the driver for iner-tial confinement fusion experiments andweapons physics applications is based ontheir ability to produce high-energy shortpulses in a beam with low divergence. Thefocusability of high-quality laser beams farexceeds that of alternate technologies, andthe challenge for NIF is to ensure that thepotential of high output beam quality isrealized. Although considerable effort hasbeen expended to minimize aberrationsdue to beamline components, other designconstraints, including cost, have causedthe residual error to be significant. Awavefront correction system is required tocompensate for these errors.

During preparations for a pulsed shot,the wavefront control system monitors thewavefront of each alignment laser at thebeamline output and automatically com-pensates for measured aberrations using afull-aperture deformable mirror. In the lastfew minutes before a shot, the controlledwavefront is biased to include a precorrec-tion for the estimated dynamic aberrationscaused by firing the flashlamp-pumpedamplifiers. One second before a shot,

closed-loop operation is interrupted, andthe Hartmann wavefront sensor is config-ured to measure the pulsed wavefront. Themeasured pulsed wavefront error providesadditional information for setting precor-rection wavefronts prior to the next shot.

Requirements for the system includeoperation at 1 Hz closed-loop bandwidth,reduction of low spatial frequency anglesin the beam to less than 20 µrad, and arange of at least 15 waves for correction ofsimple curvature measured at beamlineoutput. Figure 19 shows the location ofsystem components.

Hartmann SensorThe Hartmann sensor includes a 2D

array of lenslets and a CCD video cam-era. The output sensor (Figure 7) deliv-ers a demagnified image of the outputbeam to the lenslet array. Each lensletcollects light from a specific part of thebeam and focuses it on the CCD. Thefocal length must correspond accuratelyto the distance from the lenslet to theCCD; this result is obtained using anindex-matching fluid sealed between thelenslets and an optical flat. The lateralposition of the focused spot is a directmeasure of the direction of the lightentering the lenslet. Directional datafrom the 77 hexagonally packed lensletsof the NIF sensor are processed to deter-mine the output wavefront with an accu-racy of ²0.1 wave averaged over the

Target plane

Lens

Integrated optics module

Frequency converters

FOV = field of viewWFOV = wide field of view

Multiplewavelengthcalorimeter

Energy balance (1, 2, 3ω)Pulse shape (1, 2, 3ω)

Alignment (1ω)Far-field imaging (3ω FOV 4 × 4 mm at focal plane)Apertured calorimetry (3ω)

Near-field imaging (3ω resolution: <100 µm)WFOV calorimetry (3ω)

Beam dump

1-m apertureprimarymirror

7.7 m

FIGURE 18. The preci-sion diagnostics mea-sure the characteristicsof a full-power NIFbeamline. The largervacuum chamber is nec-essary to avoid high-intensity air breakdownand to expand the beamenough that it doesnÕtdamage beam-splittingoptics. (70-00-0796-1536pb02)

94


UCRL-LR-105821-99-1

lenslet spacing and a spatial resolutionof 43 mm when mapped into the 400-mmbeamline aperture.

Deformable MirrorEach NIF beam includes a large-aper-

ture deformable mirror for wavefront con-trol in the main laser cavity. The requiredoptical clear aperture is approximately 400 ´ 400 mm square. The mirror shape isdetermined by small push and pull dis-placements of 39 actuators spaced 74 mmapart on the back side of the mirror. Afterthree deformable mirror designs were built

and tested, including two from commer-cial sources, an LLNL design was judgedto offer the best performance. Its PMN(lead magnesium niobate) actuators can bereplaced in the event of failure, and theepoxy joints used to attach the actuatorsare protected from flashlamp light. Theceramic part of each actuator is kept undercompression by preloading to minimizethe growth of microcracks.

Figure 20 is a photograph of the proto-type mirror built from the chosen design.It has a BK7 glass substrate with a harddielectric coating having a reflectivity of³99.5% and a 1w transmission of >0.2%.

Deformablemirror

Fiber-opticwavefront#

reference source

Wavefrontsample frombeam splitter

CW laser forclosed-loopoperation

Hartmann sensor inoutput sensor packageWavefront

controller rack

FIGURE 19. Generallocation of the primarywavefront control com-ponentsÑthe Hartmannsensor, the deformablemirror, and the wave-front controller. A diag-nostic beam splitterimmediately followingthe spatial filter outputlens provides a wave-front sampling surface. (40-00-1097-2265pb02)

FIGURE 20. An LLNL40- ´ 40-cm prototypedeformable mirror hasbeen successfully assem-bled and tested. (40-00-0499-0886pb01)

95


UCRL-LR-105821-99-1

The transmission is required to allowviewing of centering references locatedbehind each mirror.

When installed in the laboratory fortesting, the LLNL prototype mirror hadsufficiently low residual error at nominaldrive voltages that response functionswere readily measured, and subsequentmodeling of the deformable mirror in aNIF beamline has used the measured func-tions. The mirror was also operated inclosed-loop mode and was able to matchthe nominally flat interferometer referencemirror to within 0.3 µm peak-to-valley and0.03 µm rms. Subsequently, with an aber-rated optic in the beam to simulate someof the general features of a NIF beamline,the closed-loop system was able restore aflat wavefront to similar accuracy.Although additional testing will be doneon the LLNL mirror, it is now consideredto be NIF qualified.

Figure 21a shows the predicted longspatial-wavelength aberration of a NIFbeamline resulting from a combination ofthermal distortions of the amplifier slabs,inaccuracies in optics fabrication, align-ment errors, optic distortions due to coat-ing stress, and the bending of parts due tomounting and gravity effects. The proper-ties of the focused 3w spot that wouldresult from such a wavefront were calcu-lated using the extensive modeling capa-bilities developed for the NIF system overthe last several years. The results are pre-sented in Figures 21b and 21c. The beam

focuses very poorly, with a 37-µrad spotradius required to include 80% of thebeam energy.

When a deformable mirror with themeasured characteristics of the LLNL pro-totype mirror is included in the beamlinemodel as part of a wavefront correctionsystem, the mirror assumes the shape ofthe surface in Figure 22a, and the focuscharacteristics change to those presentedin Figures 22b and 22c. The 80% spotradius is reduced by more than 3.5´ to10.5 µrad, which means an increase in thefocused fluence of more than an order ofmagnitude. Of course, system aberrationsthat have higher spatial frequencies thanthe deformable mirror can address willdegrade this result, but the wavefront cor-rection capability plays an essential role inachieving NIF performance goals.

Wavefront ControllerThe wavefront controller function is

accomplished by systems that are modularat the eight-beam-bundle level. Eachwavefront controller comprises computerhardware and software to periodically cal-ibrate the associated Hartmann wavefrontsensors and deformable mirrors, operatethe automatic wavefront correction loopsduring preparations for a shot, and cap-ture pulsed wavefront measurement dataduring a shot. In the moments immediate-ly prior to a shot, the system is generallyoperated under closed-loop control to an

4

1.0

0.8

0.6

0.4

0.2

00

0 20 40 60

200 400 600

3

21

0

00

–1 10

10

Wav

es

Frac

tion

insi

de

rad

ius

Wavefront at laser output

(c)(a)

Radius (µm) (focus = 770 cm)

–10–10

Focal spot half-angle (µrad)

x (cm)y (cm

)

100

–100

–100 100–50 500

–50

50

0

Focused spot (µrad)

(b)

50%: 18.1 µrad80%: 37.0 µrad95%: 67.7 µrad

3.7×103

3.7×104

3.7×105

3.7×106

3.7×107

Size of enclosed fraction (half-angle):

FIGURE 21. Performance of a NIF beamline is degraded by optical aberrations: (a) uncorrected output wavefront, (b) focal spot correspond-ing to these aberrations, and (c) radial distribution of energy. (40-00-0499-0884pb01)

96


UCRL-LR-105821-99-1

offset wavefront value. This is because theflashlamp-pumped amplifiers introduce adynamic wavefront change when they arefired, and the wavefront system must beset to anticipate that change.

Since the Hartmann sensor data is invideo format, the controller incorporatesimage processing capabilities appropriatefor recognizing and tracking the positionof the 77 focused spots from eachHartmann image. The image processingcode attains maximum accuracy by auto-matic adjustment of software parametersfor gray scale and brightness. The con-troller also measures and applies the influ-ence matrix for the deformable mirroractuators and the amplifier precorrectionfile in accordance with the mirror controlalgorithm. When operating in closed-loop,the controller is intended to maintain aclosed-loop bandwidth of approximately 1 Hz on each beam.

The NIF wavefront controller hardwarewill use the VME industrial computer busand multiprocessor architecture, which isdesigned to make maximum use of stan-dard components and to accommodatereplacement of modular elements asmicroprocessors and other computer elec-tronics continue to evolve. The systemwill be attached to the integrated comput-er control system network using CORBA(see p. 21).

SummaryAll of the laser control systems have com-

pleted 100% design reviews, and we are pro-ceeding with prototype testing, Title IIIdesign, and procurement. Activities are nowshifting to completion of detailed drawingsand ordering of production hardware. Asthis hardware arrives, it will be thoroughlytested before installation in the laser, switch-yard, and target bays of the NIF facility. Thebeam control and laser diagnostics systemswill be among the first to be activated sothat their functionality can be used to con-trol and monitor subsequent additions toeach beamline. When NIF is completed,these systems will operate in conjunctionwith the integrated computer controls toprovide the level of automatic operation thatis essential for successful facility operation.

AcknowledgmentsThe authors specifically acknowledge

the significant contributions of the talent-ed technology associates, technicians, anddesigners in beam control and laser diag-nostics to the designs described above. Inaddition, because the NIF Project is now inits fifth year, the design has benefited fromthe efforts of some who are no longer onthe staff, but whose contributions are notforgotten.

0.4

100

–100

–100 100

1.0

0.8

0.6

0.4

0.2

0–50 500 0

0 20 40 60

200 400 600

–50

50

0.20

–0.2–0.4–0.6–0.8–1.0

00

0

–110

10

Wav

es

Frac

tion

insi

de

rad

ius

Focused spot (µrad)Deformable mirror shape

(b) (c)(a)

Radius (µm) (focus = 770 cm)

50%: 5.7 µrad80%: 10.5 µrad95%: 29.9 µrad

1.91×104

1.91×105

1.91×106

1.91×107

1.91×108

–10–10

Size of enclosed fraction (half-angle): Focal spot half-angle (µrad)

x (cm)y (cm

)

FIGURE 22. Performance of a NIF beamline when the wavefront correction system is in operation: (a) shape of the deformable mirror, (b) improved focal spot, and (c) radial distribution of energy. (40-00-0499-0885pb01)

97


UCRL-LR-105821-99-1

Notes and References1. Present baseline design uses a tapered cone

on Pass #4 only preventing pinhole closure for a20-ns pulse. Passes #1, #2, and #3 use stainless-steel washer-type pinholes.

2. The CSF pinhole wheel rotates to multiple preci-sion set points to locate an alignment reticle, aselection of the shot pinholes, several opticsinspection pinholes, or a 35-mm clear aperture.

3. The TSF pinhole wheel positions multiple shotpinholes or an 80-mm clear aperture.

4. Two insertable light sources mounted on a com-mon translation stage provide references forbeam alignment and wavefront calibration.

5. Insertable 45¡reflector with attached dump for3w diagnostic beam is withdrawn from beam foralignment and optical inspection.

99

The National Ignition Facility (NIF)will use a 192-beam, 1.8-MJ laser for

inertial confinement fusion (ICF) experi-ments.1 A major laser component of theNIF, the flashlamp-pumped Nd:glassamplifiers must provide sufficient gainand stored energy to meet requirementsfor laser energy and power while addingminimal wavefront distortion to the laserbeams. The NIF amplifiers differ fromthose used in previous ICF laser systemsmainly in their overall scale and packag-ing. Figure 1 shows a two-slab-long, 4 ´ 2NIF amplifier module, where the n ´ mdesignation denotes the number (height ´width) of parallel amplifying channels orbeam apertures that are combined. TheNIF amplifiers use 40-cm-square apertures,which approximate the practical size limitimposed by amplified spontaneous emis-sion (ASE) that depumps the laser slabsand limits gain. Large-aperture sizereduces system costs by reducing the num-ber of laser beams needed to produce therequired energy on target. Previously, thelargest amplifiers constructed were the 2 ´ 2 Beamlet amplifiers, which combinedonly four 40-cm-square apertures.2

Amplifiers with combined beams werefirst proposed by Lawrence LivermoreNational Laboratory (LLNL) in 1978 as away of reducing the cost of MJ-class fusion

laser systems.3 Combining beams in a sin-gle enclosure reduces costs in three ways:(1) by making amplifiers more compact,thereby reducing the size and cost of thebuilding; (2) by increasing pumping effi-ciency, thereby reducing the size and costof the power-conditioning system; and (3) by reducing the number of internalamplifier parts. The NIF design achievesconsiderable cost savings by making theNIF amplifiers larger than the Beamletamplifiers. Similar amplifiers will be usedin the Laser Megajoule (LMJ), a 240-beamlaser system now being developed anddesigned by the French Commissariat alÕEnergie Atomique (CEA). Much of thedevelopment work for the LMJ amplifiersresulted from collaborative efforts betweenscientists and engineers at LLNL and theFrench CEA.

UCRL-LR-105821-99-1

DESIGN AND PERFORMANCE OFFLASHLAMP-PUMPED ND:GLASS AMPLIFIERS

FOR THE NIF*

A. C. Erlandson C. Marshall M. Rotter O. Carbourdin** G. LeTouze**

T. Alger E. Moor S. Sutton E. Grebot** X. Maille**

J. Horvath L. Pedrotti L. Zapata J. Guenet** S. Seznec**

K. Jancaitis S. Rodriguez J. Beullier** M. Guenet**

*This article also appears in the Vol. 8, No. 4 issue of the ICFQuarterly Report; it is included in this issue for completeness.**Commissariat a LÕEnergie Atomique (French AtomicEnergy Commission)

FIGURE 1. A two-slab-long, 4 ´ 2 NIF amplifiermodule. (70-00-0199-0056pb02)

100

DESIGN AND PERFORMANCE OF FLASHLAMP-PUMPED ND:GLASS AMPLIFIERS FOR THE NIF

UCRL-LR-105821-99-1

This article describes the design and per-formance of the NIF amplifiers. First, wedescribe the NIF amplifier design. Next, wedescribe the prototype amplifier we testedand discuss the equipment used to measureits optical performance. We then comparemodel predictions with measurementresults, and show performance predictionsfor the NIF amplifiers that are based on ourtest results and validated models.

NIF Amplifier Design

Pump Cavity Design andPerformance

The NIF amplifiers provide optical gainat the 1.053-µm wavelength by usingneodymium-doped, phosphate glass, rect-angular laser slabs oriented at BrewsterÕsangle with respect to the beam, to elimi-nate reflection losses. The slabs haveabsorbing glass edge claddings to preventinternal parasitic laser oscillation. Eachslab holder supports four slabs, onestacked above the other. Central flash-lamps cassettes pump slabs in both direc-tions, while side flashlamp cassettes with

large silver reflectors pump slabs in onedirection. The glass blastshields betweenthe flashlamps and the laser slabs servethree purposes: (1) they prevent acousticwaves generated by the flashlamps frompropagating into the beam path and caus-ing wavefront distortion; (2) they providea contamination barrier between the flash-lamp cavity and the critical slab cavity;and (3) they form one wall of the channelused for flowing cooling gas around theflashlamps. Figure 2 shows a plan view ofa single NIF amplifier slab column thatillustrates the arrangement of the slabs, flashlamps, blast shields, and reflectors.

The NIF amplifiers must meet require-ments for gain, gain uniformity, wavefrontdistortion, and thermal recovery rate. Highgain is required for the laser to meet its out-put energy and power requirements, espe-cially because of the limited injected-energyfrom an affordable pulse generation system.At the same time, the amplifier mustachieve its gain requirements efficiently to reduce the size and cost of the power-conditioning system used to drive the flash-lamps. Accordingly, the NIF amplifiers useseveral features to increase efficiency. Theside flashlamp arrays use silver reflectors

Side flashlamp cassette

AR-coatedblast shield

Lamp Ag

Central flashlamp cassette

FIGURE 2. Plan view of the NIFamplifier pump cavity. (70-00-0199-0057pb02)

with involute reflector shapes. They aredesigned to reflect flashlamp light towardthe laser slabs, while returning little flash-lamp light back to the absorbing flashlampplasma.4 Compared with flat reflectors, theinvolute reflectors reduce the flashlampelectrical energy required to meet the gainrequirement by ~15%. Additional reduc-tions in flashlamp electrical energy areachieved by using sol-gel antireflective (AR)coatings on both sides of the blast shields(7%) and by preionizing the flashlampswith weak electrical pulses delivered sever-al hundred microseconds before the mainpulse (10%). Preionization causes the flash-lamp arc to develop more uniformly, and itincreases the electrical-to-optical conversionefficiency of the flashlamp plasma. Overall,the predicted storage efficiency of the NIFamplifiers is 3.8%, which is significantlyhigher than those in previous ICF lasers(3.0% and 1.8% for the Beamlet and Novaamplifiers, respectively).2 Storage efficiencyis defined as the total extractable energystored in the laser slabs divided by the elec-trical energy delivered to the flashlamps.

In addition to the above criteria, wewant the amplifiers to produce uniformgain distributions across their apertures, so that uniform beam output fluence canbe readily obtained. Uniform output flu-ence is essential for maximizing outputenergy while maintaining acceptable opti-cal damage risk. With significant gain vari-ations, it becomes necessary to compensateby increasing the input fluence distributionin regions of the aperture with lower gain,as was done on the Beamlet laser. In largeslabs, such as the 4.1- ´ 45.8- ´ 80.9-cmslabs used in the NIF amplifiers, the transverse gain uniformity is determinednot only by the distribution of flashlamplight across the slab, but also by internalamplified spontaneous emission. ASE preferentially depletes gain near the slabÕs ends because this position has thelongest path length for internal amplifica-tion.2,5 The skewed diamond-shapedreflectors in the central flashlamp arraysreduce gain variations in the NIF pump-cavity design by directing flashlamp lightpreferentially near the slab ends.

We used a 3D ray-trace code to evaluatereflector shapes and to predict gain distri-butions across the aperture. The 3D code

calculates pump rates for the neodymium-ion inversion by using a reverse ray-tracetechnique where rays were propagatedbackward from the point of interest in theslab to the flashlamp plasma. The ray-tracemodel tracked the change in the spectralcontent of each ray as it interacts with var-ious reflecting and transmitting media pre-sent in the pump cavity. The absorptionand reflectance properties of cavity com-ponents and the emission properties of thelamp plasma were determined experimen-tally.6 Once the pump distribution was cal-culated as a function of flashlamp power,the peak gain-coefficient distribution wascalculated by solving the differential equation for stored energy density as afunction of time at each point in the slab.In addition to radiative and nonradiativespontaneous decay processes, the modelalso tracked spatially and temporallydependent ASE decay rates throughoutthe volume of the laser slabs. Both thephysical assumptions in the model and the calculational techniques used in thecomputer programs that implement thegain model have been described in detailin earlier publications.7,8,9

Wavefront distortion is produced bywaste heat deposited in the laser slabs byflashlamp pumping processes. Wavefrontdistortion has a prompt component thatis caused by the flashlamp heating thelaser slabs during the shot; it also has aresidual thermal component that iscaused by thermal energy remaining inthe pump cavity from previous shots. Theprompt component is primarily causedby the uneven heating of the two sides ofthe slab. This causes the slab to warpbefore the laser beam passes through the slabs when peak gain is attained.10

Residual thermal distortions are causedby thermal gradients in the laser slabsthemselves, as well as thermal gradientsin the gas that is convectively heated bythe laser slabs. The residual slab distor-tions, like the prompt pump-induced dis-tortion, tend to be low order. Thedeformable mirror system now plannedfor the NIF is expected to correct theresidual slab distortions. The gas distor-tions, however, have higher spatial fre-quencies and are more rapidly varying,making them difficult to correct.

101


UCRL-LR-105821-99-1

102


UCRL-LR-105821-99-1

To accelerate the recovery from residualthermal distortion, the NIF flashlamps willbe actively cooled by flowing gas. Flash-lamp cooling is effective because some60% of the pump waste heat resides in theflashlamp envelopes immediately after ashot. Although water cooling has beensuccessfully used to accelerate the shotrate of the OMEGA Laser at the Universityof Rochester,11 gas cooling was chosen forthe NIF to eliminate water-jacket tubesand to leave more room for efficiency-enhancing reflectors. NIF flashlamp cool-ing system will provide gas flow rates ofup to 20 cubic feet per minute per flash-lamp, with the gas flow direction alternat-ing between flashlamp cassettes as shownin Figure 3. The inlet temperature of thecooling gas will be controlled over a ±5¡Crange centered about the ambient tempera-ture with ±0.3¡C accuracy.

We used several computer codes to cal-culate the wavefront distortion induced byflashlamp pumping processes. The prompttemperature rise, which was assumed to beproportional to the time-integrated localpump rate, was calculated using a 2D ray-trace code, which was a precursor to the3D ray-trace code described above.12 Theresidual temperature component was calculated using a finite-element heattransfer program, with starting tempera-tures set equal to values measured on a

test amplifier shortly after firing the flashlamps.13 The calculated temperaturedistributions were used as input for finite-element calculations of the deformationand stress in the laser slabs.14 The finalstep in the calculation sequence was to use the calculated temperature, deforma-tion, and stress distributions in a ray-tracealgorithm to evaluate the optical pathlength variations across the aperture.15

The effects of refractive index responses to temperature and stress were included inthe calculations.

Mechanical DesignThe large scale of the NIF amplifiers

requires new mechanical designs thatallow convenient assembly and mainte-nance. Accordingly, the NIF amplifiers usea modular design in which the most criti-cal components, including the flashlamps,laser slabs, and reflectors, are mounted inline-replaceable units or cassettes that canbe readily inserted or removed withoutdisturbing their neighbors.16 See Figure 4.

The design uses sealed maintenancecarts that access the bottom of the ampli-fiers in order to install and remove flash-lamp cassettes and four-high slab cassettes.The cassettes are inserted and removedfrom their enclosure, called the frameassembly unit, which is supported by topplates mounted to an overhead supportstructure. Plenums distribute electricityand cooling gas to the flashlamps from thetop, through holes on the top plate.

Each NIF 4 ´ 2 beam bundle uses an 11-slab-long main amplifier that the beampasses four times, and a five-slab-longpower amplifier that the beam passestwice. Assembly of these amplifiers beginsin an off-line clean room, where the frameassembly units are cleaned and the topplates and blast shields with antireflectivecoatings are installed. A flashlamp-light-resistant polymer is used to bond the blastshields to a metal frame, and siliconeinflatable seals are used to seal the metalframe to the inside of the frame assemblyunit. These seals reduce leak rates betweenthe flashlamp cavity and the slab cavity.After the blast shields have been installed,frame assembly units are bolted togetherto form 5- and 11-slab-long units that are

Supply Exhaust Supply

Flashlamps

FIGURE 3. Alternating the direction of cooling gas flow through the flashlmp cas-settes enabled us to put cooling-gas connections at the top of the amplifier, therebysimplifying the amplifier mechanical design. (70-00-0199-0058pb02)

transported to the laser bay where they aremounted to an overhead support structureby their top plates.

The laser slabs, reflectors, and metalparts that comprise the slab cassettes arespray-cleaned and assembled in an elevat-ed, class-100 clean room. We have to main-tain high cleanliness levels because smallparticles resting on the laser slabs cancause damage when heated by flashlampand laser light. Current specifications callfor the laser slabs and metal surfaces to bemaintained at cleanliness levels of 50 and100, respectively. The levels correspond toparticle-size distributions in which thereare only one 50- or 100-µm particle persquare foot of surface area, respectively.Following assembly, the slab cassettes arelowered into a specialized clean cart. Oncethe cart has been moved to the laser bay, itdocks to the bottom of a frame assemblyunit and establishes a hermetic seal tomaintain cleanliness. After the top cover of

the cart and the bottom cover of the frameassembly unit have been pressed togetherto trap residual particles, the cover pair ismoved to the side to open a passagewayfor the slab cassette to be raised into theframe assembly unit. Rollers mounted inthe corners of the cassette guide the cas-sette during insertion and prevent metal-on-metal rubbing, which generates parti-cles. A fail-safe mechanism in the cart acti-vates latches that hold the slab cassette inplace. Slab cassettes can be removedfor occasional refurbishment by reversingthis installation process. Figure 5 shows aprototype slab cassette cart, which has suc-cessfully completed some 50 slab-cassettetransfers in our laboratory. Similar cartswill be used to install and remove the NIFflashlamp cassettes and blast shields.

103


UCRL-LR-105821-99-1

Plenums

Top plates

Frame assembly�units

Blast shields

Side flashlamp�cassette

Slab cassette

Central flashlamp�cassette

�

FIGURE 4. The NIF amplifiers use a modular design.(70-00-0199-0059pb02)

FIGURE 5. The prototype slab cassette cart shown inserting a slabcassette into the NIF prototype amplifier. (70-00-0199-0060pb02)

104


UCRL-LR-105821-99-1

NIF Prototype Amplifier Design

We have built and tested a prototypeamplifier at LLNL to validate key require-ments prior to deployment in NIF. Thisprototype amplifier was extremely close tothe NIF amplifier design, as it used thesame size flashlamps and laser slabs andnearly the same reflector shapes. The pro-totype amplifier was slightly more com-pact in the direction transverse to the laserbeam because the insertion clearancesbetween the slab cassettes and the frameassembly units are smaller, and the blastshields are thinner. In addition, the proto-type was slightly more compact in thedirection parallel to the beam due to small-er gaps between frame assembly units.Ray-trace codes that we have developed topredict amplifier gain show that the differ-ences in gain and storage efficiencybetween the prototype amplifier and theNIF amplifiers are negligible.

The prototype amplifier was tested as aone- , two- , and three-slab-long amplifier.Both the ÒdiamondÓ and ÒXÓ configurationsof the two-slab-long amplifier were tested.Amplifiers of different lengths were tested inorder to infer the gain and wavefront distri-butions of the interior slab; this was impor-tant because the 5- and 11-slab-long NIFamplifiers have many interior slabs in addi-tion to X and diamond end slabs. Figure 6shows the diamond, X, and three-slab-longamplifier configurations we tested.

To measure temperatures at approxi-mately 80 different locations inside theamplifier, we used thermocouples. Theywere placed on three laser slabs in one ofthe slab cassettes with holes drilledthrough the slabs to establish good

thermal contact. Thermocouples were alsoplaced on flashlamps, blast shields, andreflectors near the slab cassette. The pur-pose of the temperature measurementswas to determine the initial temperaturerise of the pump cavity components and to determine cooling rates after shots.

The laser slabs in the prototype ampli-fier were finished from LG-770 glassmade by Schott Glass and LHG-8 glassmade by Hoya. Both types of laser glasshad Nd ion concentrations of 4.2 ´ 1020

ions/cm3. The flashlamps were made byEG&G and ILC Technology.

Apparatus for Gain andWavefront Measurements

We constructed an optical diagnostic system to perform time-resolved gain andwavefront measurements over the entireaperture of our prototype amplifier.Measurements were performed using apulsed, injection-seeded, single-longitudinalmode, Nd:YLF probe laser operating at1.053 µm. This laser produced 80-mJ, 20-ns-long pulses at a repetition rate of 13 Hz. The probe laser beam was expandedand image relayed by a series of telescopes.After passing through the amplifier once,the beam reflected back through the ampli-fier and telescopes a second time. Afterreturning from the amplifier, a portion ofthe beam was sampled with a beam splitterand reflected to scientific-grade charge-coupled device (CCD) cameras for gain andwavefront measurements. Figure 7 shows aschematic of our equipment.

Amplifier gain distributions were deter-mined by calculating the ratio of theimages produced by two CCD cameras(the gain reference camera and the gain

Flat�Ag�

reflector

Probe�beam

Probe�beam

XDiamond 3-slab-longFIGURE 6. Plan view ofthe prototype amplifier inthe diamond, X, and 3-slab-long configurations.(70-00-0199-0061pb02)

probe camera); they recorded the laserbeam intensity before and after the beampassed through the amplifier. Measuredbackground contributions from the flash-lamp light and amplified spontaneousemission were subtracted from the gaincamera image. A correction factor for pas-sive transmission losses was determinedby firing the probe laser without firing theamplifier flashlamps. Crosshair imageswere used to ensure proper registration ofthe gain and reference camera images.

Wavefront distributions were measuredby interfering the probe beam with a refer-ence beam, which was generated with an~25-m-long single-mode optical fiber cut sothat its optical length approximatelymatched the probe laser path length. Toincrease the fringe contrast ratio, a half-wave plate and a polarizer were used toattenuate the probe laser to approximatelythe same intensity as the reference beam.Two cameras were used to record interfero-grams on successive pulses of the 13-Hzprobe laser. A Fourier-transform techniquewas used to calculate wavefront distribu-tions from the interferograms.17 Wechecked the calibration of the interferome-ter by measuring the wavefront of a knownlens. From this check, we estimate the accu-racy of our wavefront measurements to be±0.02 waves rms, for both cameras. Promptpump-induced wavefront distortion was

determined by subtracting the wavefrontmeasured at the time of peak gain from thewavefront measured 75 ms earlier.

Measurement Results andComparison with ModelPredictions

Gain measurements were in close agree-ment with model predictions. For example,Figure 8 shows measured and predictedgain-coefficient distributions for the two-slab-long X configuration, with the flash-lamp pulses close to the 34-kJ per lamp,360-µs-long pulses anticipated for the NIF.Although the distributions in Figure 8characterize the aperture next to the bot-tom, distributions measured in the otherapertures were nearly the same, due to thevertical translation symmetry of the pumpcavity. Gain was lower on the right-handside of the aperture because the right-handsides of both slabs were close to the endsof the amplifier where pump-light fluenceswere lower. Differences between the modelpredictions and measurements were lessthan 1% rms for all other combinations ofapertures and configurations (diamond, X, and 3-slab-long). This agreement wasachieved with a single value of theadjustable parameter used in the model to scale the flashlamp pumping rate.

105


UCRL-LR-105821-99-1

Amplifiers

Conjugate image �plane mirror

Object �plane

Fiber reference

Pulsed, �single-�frequency �laser

Fringe �camera #150/50Fringe�

camera #2

Hartmann�sensor

Image �plane

λ/2

Polarizer

FIGURE 7. We constructed alarge-aperture diagnostic sys-tem to measure gain and wave-front distributions produced bythe prototype amplifier. (70-00-0199-0062pb02)

106


UCRL-LR-105821-99-1

Prompt pump-induced wavefront distor-tion measurements were also in close agreement with model predictions. Figure 9compares the measured and predictedprompt pump-induced wavefront distor-tion for the X configuration and for thestandard flashlamp energy of 34 kJ/lamp.The results have been normalized to a single-slab pass. The close agreementbetween measurements and predictionswas obtained using a scaling factor thatadjusts the prompt temperature rise in theslab relative to the calculated pumping rate.Equally close agreement was obtained for

other apertures and for time delays of up to1 ms after the time of peak gain, using asingle value for the scaling factor.

In addition to the optical characteriza-tions, we also performed temperaturemeasurements to determine the efficacy ofthe flashlamp cooling system. Figure 10shows average slab temperature versustime after a flashlamp shot, for two cases:(1) with flashlamp gas set at ambient tem-perature; and (2) with the flashlamp cool-ing gas set 0.5¡C below ambient tempera-ture. Our data show that the average slabtemperature recovers to within ~0.1¡C of

(a) Measured (b) Predicted (c) Horizontal lineout of measured� and predicted gain distributions

6.0

5.3

4.6

3.9

3.2

2.5

20

10

Ð10 0 10

0

Ð10

Ð20

�

x (cm)

y (c

m)

<α> = 4.700%/cm

6.0

5.3

4.6

3.9

3.2

2.5

10

Ð10 0 10

Ð10

Ð20

x (cm)

y (c

m)

<α> = 4.802%/cm

0

Gai

n co

eff (

%/

cm)

Gai

n co

eff (

%/

cm)

Average gain: 1, 4, 5, 6 C LG770 20 x

FIGURE 8. Measured and predicted gain distributions for the two-slab-long X configuration, with flashlamps energized with 34-kJ-per-lamp, 360-µs-long pulses. Measurements were made in the next-to-bottom aperture. (70-00-0199-0063pb02)

0.20

0.10

0

010

10

20

200

Ð0.10

Ð10

Ð10Ð20

Ð20Ð30

Ð0.20

Wav

es

PÐV approx. 0.269207 waves

Vertical position�(cm)

Vertical position�(cm) Horizontal position�

(cm)Horizontal position�

(cm)

0.20

0.10

0

010

10

20

200

Ð0.10

Ð10Ð10Ð20 Ð20

Ð0.20

Wav

es

(a) (b)

Est. errror = 0.00703810 waves�PÐV approx. 0.28983023 waves

FIGURE 9. (a) Measuredand (b) predicted promptpump-induced wavefrontdistributions for the two-slab-long ÒXÓ configura-tion, with flashlamps ener-gized with 34-kJ-per-lamp, 360-µs-long pulses.Measurements were madein the next-to-bottom aperture. (70-00-0199-0064pb01)

the ambient temperature in about eighthours using ambient-temperature coolinggas, and in about four hours using slightlychilled gas. Measured gas distortions hadhigh frequency components, with the rmsamplitude varying linearly with the differ-ence between the average slab tempera-ture and the ambient temperature.Measured slab distortions were low orderand recovered to less than 0.04 waves(peak-to-valley) per slab per pass within3Ð4 hours after each shot. This measure-ment is consistent with the magnitude ofthe slab distortions, which we expect tocorrect with a deformable mirror on NIF.

Performance Predictionsfor the NIF Amplifiers

Figure 11 shows the beamline-averagedgain-coefficient distribution for a NIFlaser beam, predicted by our validated3D ray-trace code. The code took intoaccount several features of the NIFamplifiers that were different from the prototype amplifier: protectivelycoated Ag reflectors with stable, higherreflectances; improved, two-layer sol-gelantireflective coatings on the blastshields; use of equal numbers of LG-770and LHG-8 laser slabs; and slight reflector-shape differences. The predict-ed aperture-averaged gain coefficientwas 5.3%/cm, and the distribution

shows the expected gain roll-off near theedges due to ASE. Propagation codemodeling of the NIF beamline showsthat the gain has sufficient magnitudeand uniformity for the NIF laser to meetits beam power and energy require-ments. To achieve a uniform output flu-ence distribution and maximum beamoutput energy, however, the input flu-ence distribution of the injected laserbeam has to be tailored to compensatefor the anticipated gain variations.

107


UCRL-LR-105821-99-1

1.0

0.8

0.6

0.4

0.2

00 2 4 6 8

Ave

rage

sla

b te

mp

erat

ure

�ri

se a

bove

am

bien

t (°C

)

Ð0.5°C�Cooling�model

Ð0.5°C�Measurements

Ambient�cooling�model

Time (h)

FIGURE 10. Using slightly chilled(Ð0.5¡C) cooling gas accelerates slabthermal recovery rates. (70-00-0199-0065pb02)

0.06

0.04

0.02

0

100

Ð1010

0Ð10

y (cm)

x (cm)

FIGURE 11. NIF beamline-averaged gain-coefficient distributionpredicted with the 3D ray-trace code. (70-00-0199-0066pb02)

108


UCRL-LR-105821-99-1

Figure 12 shows the total promptpump-induced wavefront distortion for aNIF beamline, predicted with our codesusing the source-term scaling factorderived from AMPLAB experiments. Thedistribution is a low-order ÒMÓ shape,with a peak-to-valley value of 5.9 waves.This distribution possibly overestimates

the prompt pump-induced wavefront dis-tortion that the NIF beamlines will actual-ly produce, because it is conceivable thatstray flashlamp light may have distortedthe optics, which are part of the diagnos-tic. Nonetheless, the predicted distortionsare within the range that now appears tobe largely correctable with the NIFdeformable mirror.

To estimate the effect of the gas distor-tions on the NIF laser beams, we scaledthe gas distortions measured on the proto-type amplifier to account for the greaterpath length and greater number of slabs inthe NIF amplifiers. In addition, we used abeam propagation code to calculate thebeam focal spot. Our estimate shows thatthe gas distortions will meet the NIFrequirement (less than 5 µrad added beamdivergence) within 3 hours after the shot,provided the temperature of the flashlampcooling gas is ~1¡C below the ambient tem-perature. Figure 13 shows model results forflashlamp cooling with gas at ambient tem-perature and at 1¡C below ambient tempera-ture. The conservative and aggressive esti-mates correspond to scaling the AMPLABresults by length (for aggressive estimates)and by number of laser slabs (for conserva-tive estimates).

Conclusion and SummaryAdvances in amplifier technology devel-

oped for the NIF include a modular designand bottom-access carts to improve main-tenance; features to improve storage effi-ciency that reduces the size and cost of thepower conditioning; and active gas coolingto accelerate the laser shot rate. Further,gain, wavefront, and thermal recoverymeasurements performed on the NIF pro-totype amplifier are consistent with meet-ing NIF performance requirements.

AcknowledgmentsWe thank numerous contributors to our

design, analysis, and experimental efforts:Richard McCracken, Charles Petty,Lawrence Smith, Alex Drobshoff, Ian M.Thomas, John Trent, Drew Hargiss, RandyAceves, Jerome Hall, Rene Neurath, JoeSmith, Wil Davis, Jean Marie Morchain,*Beatrice Bicrel,* Patrick ManacÕh,*

20

10

0

Ð10

Ð20Ð20

Ð100

1020

8

6

4

2

0

Vert. pos. (cm)

Horiz. pos. (cm)

Wav

efro

nt (w

aves

/sl

ab/

pass

at 1

µm

)

NIF System

FIGURE 12. NIF beamline prompt pump-induced wavefront distor-tion predicted with ray-trace and finite-element codes (70-00-0199-0067pb02)

FIGURE 13. Predicted gas-distortion phase distrubances in NIF forambient and chilled gas cooling. Aggressive and conservativeassumptions are used to bracket the problem. (70-00-0199-0068pb02)

Gas

rm

s ph

ase

dis

tort

ion

(wav

es)

Conservative (β = 0.5)

Aggressive (β = 1)

NIF criteria

00.1

0.2

0.3

0.40.5

0.6

0.70.8

0 2 4 6 8Time (h)

Ambient coolingChilled cooling

Thierry Adjadj,* Gary Ross, John Toeppen,Larry Morris, Jean Lobit,* Douglas Swort,Ronald Bettencourt, Michael Weddle,Loretta Kleinsasser, Dzung Nguyen, JohnH. Campbell, Daniel Walmer, WilliamSteele, Jerry Britten, Curly Hoagland,Gregory Rogowski, Steve Letts, CindyAlviso, Irving Stowers, Joe Menapace,Douglas Larson, Scott Hulsey, Phil Test,Steve Fulkerson, George Pollock, AndyHinz, Anthony Lee, Janice Lawson, WadeWilliams, Mark Newton, Ken Manes,Richard London, Yuri Zakharenkhov, PeterAmendt, Jean-Francois Mengue, FabrisseLaniesse,* John Hunt, John Murray,Richard Hackel, Jeff Paisner, Steve Payne,and Howard Powell.*Commissariat a LÕEnergie Atomique (French Atomic EnergyCommission)

Notes and References1. W. H. Lowdermilk, Second Annual Conference on

Solid State Lasers for Application to InertialConfinement Fusion, in Proc. SPIE, Bellingham,WA, Vol. 3047, p. 16Ð37 (1997).

2. A. C. Erlandson, M. D. Rotter, D. N. Frank, andR. W. McCracken, ICF Quarterly Report 5(1),18Ð28, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-LR-105821-95-1 (1995).

3. W. F. Hagen, Laser Program Annual ReportÐ1977,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-50021-77, pp. 2-228 to 2-231 (1978).

4. W. T. Welford and R. Winston, High CollectionNonimaging Optics, Academic Press, New York,NY, p. 266 (1989).

5. A. C. Erlandson, K. S. Jancaitis, R. W. McCracken,and M. D. Rotter, ICF Quarterly Report 2(3), 105Ð114,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-LR-105821-92-3 (1992).

6. H. T. Powell, A. C. Erlandson, and K. S.Jancaitis, Conference on Flashlamp-pumpedLaser Technology, in Proc. SPIE, Bellingham,WA, Vol. 609, pp. 78Ð94 (1986).

7. H. T. Powell, A. C. Erlandson, K. S. Jancaitis,and J. E. Murray, High-Power Solid State Lasersand Applications, in Proc. SPIE, Bellingham, WA,Vol. 1277, pp. 103Ð120 (1990).

8. K. S. Jancaitis, S. W. Haney, D. H Munro, G.LeTouz�, and O. Cabourdin, in Proc. SPIE, Vol.3047, pp. 106Ð111 (1997).

9. G. LeTouz� et al., Ò3-D Gain Modeling of the LMJand NIF Amplifiers,Ó Solid State Lasers forApplication to Inertial Confinement Fusion, ThirdInternational Conference 1998 (to be published).

10. J. E. Murray, W. F. Hagen, and B. W. Woods, LaserPrograms Annual Report, UCRL-50021-82, LawrenceLivermore National Laboratory, Livermore, CA, p. 7Ð72, (1982).

11. J. H. Kelly, M. J. Shoup III, M. D. Skeldon, and S.T. Bui, Solid State Lasers III, edited by G. J.Quarles, in Proc. SPIE, Bellingham, WA, Vol.1627, p. 252Ð259 (1992).

12. G. LeTouz�, O. Cabourdin, J. F. Mengue, M.Rotter, and K. Jancaitis. ÒShaped Reflectors forPump Cavities,Ó 2nd Annual Conf. Solid StateLasers for Application to ICF, Limeil, France(1996).

13. A. Shapiro, TOPAZ3DÑA 3-D Finite ElementHeat Transfer Code, UCID-20484, LawrenceLivermore National Laboratory, Livermore, CA(1985).

14. B. N. Maker, NIKE3DÑA Nonlinear, Implicit 3-DFinite Element Code for Solid and StructuralMechanics, UCRL-MA-105268, Rev. 1, LawrenceLivermore National Laboratory, Livermore, CA(1995).

15. S. Doss and R. Gelinas, Laser Program AnnualReport, UCRL-50021-86, Lawrence LivermoreNational Laboratory, Livermore, CA, p. 7Ð132(1986).

16. A. C. Erlandson et al., Second Annual Conferenceon Solid State Lasers for Application to InertialConfinement Fusion, in Proc. SPIE, Bellingham,WA, Vol. 3047, p. 16Ð37 (1997).

17. M. Takeda, H. Ina, and S. Kobayashi, J. Opt. Soc.Am., 72, 156Ð160 (1982).

109


UCRL-LR-105821-99-1

A-1

Nova Operations performed 216experiments, which surpassed our

goal for the quarter. These experimentssupported efforts in Inertial ConfinementFusion (ICF), Defense Sciences, universitycollaborations, Laser Science, and Novafacility maintenance.

The process of planning the decommis-sioning of the Nova facility continued. InDecember, a notice appeared in theCommerce Business Daily to advertise theavailability of Nova components. After theFebruary 1999 deadline for proposals, acommittee organized by the Department ofEnergy will determine the disposition ofNova components. It has already beendecided that parts of Nova will be reusedin the National Ignition Facility (NIF). Forexample, inspections have indicated thatthe FR5 rotator glass in the Nova Faradayrotators meets the specifications requiredfor use on NIF.

The Petawatt project continued to inno-vate with the installation of a preformedplasma beamline. The center 15 cm ofBeamline 8, which is unused by Nova, isalready being used as a 4w probe in theten-beam target chamber. Hardware was

installed in December that will permit thisbeam to be diverted to the Petawatt targetchamber to create a preformed plasma inwhich the Petawatt beam will interact.

Hardware was installed and tested toallow the use of cryogenic targets in theNova ten-beam target chamber for equation-of-state experiments that will be conductednext quarter. The ten-beam experiments willfollow the two-beam cryogenic campaign,which served to map out the single-shockHugoniot for deuterium from 200 kbar to 3 Mbar. The NIF will use multiple shocks tocompress the hydrogen fuel to much higherdensity than can be achieved with a singleshock, and there is no fundamental theorythat allows us to calculate a priori whatshock compression will be achieved for agiven drive. In the cryogenic ten-beamexperiments, we are developing diagnosticcapability to tune multiple shocks for NIFignition implosions and measuring the equa-tion of state of hydrogen after two shocks.We are also simulating the plasma condi-tions that will occur in a NIF hohlraum withcryogenic dense HeÐH gas.

NOVA UPDATEOCTOBERÐDECEMBER 1998

UCRL-LR-105821-99-1

B-1

Overall AssessmentOverall progress on the NIF Project

remains satisfactory for the first quarter ofFY99. The current top-level assessment ofProject status remains similar to that statedat the end of the fourth quarter 1998: thatthere will be no change to the fourth quarter2001 Level 2 milestone for the End ofConventional Construction, nor to thefourth quarter 2003 Project Completion date. However, the NIF Project Office nowanticipates that based upon the status ofConventional Facilities, CSP-4, work on theLaser Bay Core, and the status of SpecialEquipment design and procurement, therecould be an impact of 6 to 8 weeks in thefourth quarter 2001 completion of the Level 4 milestone for start-up of the firstNIF bundle. The impact of current field conditions on this important milestone,which is to be completed in three years, continues to be evaluated on a weekly basis.

The first quarter 1999 was a productivequarter on the NIF site as the totalConventional Facilities work completedapproached 42%. With the completion ofPhase I work by Nielsen DillinghamBuilders (NDBI), CSP-4, in December, thecoordination between prime subcontrac-tors on site has been reduced significantly.Hensel Phelps (CSP-9) now has unrestrict-ed access and sole occupancy of the LaserBuilding. NDBI (CSP-6/10) has sole pos-session of the Target Building, and NDBI(CSP-5) has sole possession of the OpticsAssembly Building (OAB). Coordination ofaccess to the work areas, laydown space,

and coordination of interface between sub-contractors in the OAB Corridor are nowthe focus of the overall site coordination.

In Special Equipment, overall progressis satisfactory. Ten 100% reviews were con-ducted in the quarter, bringing the totalpercentage complete to 89%. In addition tothe ongoing work to close out the SpecialEquipment design reviews and the contin-uation of procurement efforts, an expand-ed effort has been under way in the plan-ning area. The primary focus is improvingthe ability to summarize the IntegratedProject Schedule (IPS), providing easiermethods to focus on particular areas orfunctions, and validating the installationlogic. In addition, all Special Equipmentprocurement contract awards greater than$100K in the NIF Planning System musthave equivalent contract milestones in theIPS. The Special Equipment part of the IPShas been reviewed and updated to ensurethis compliance.

In Optics, facilitization is complete atHoya and Schott with all contracts placedand pilot activities already started. Hiringand training for pilot runs is mostly com-pleted. Both laser glass vendors are doinga great job in supporting the NIF sched-ules. The Zygo facility was completed infirst quarter 1999 and first-bundle optics(LM3) are now being shaped. Pilot produc-tion contracts were finalized at two potas-sium dihydrogen phosphate (KDP) ven-dors and the first pilot runs were started,and six 1000-L crystallizers continued tooperate at Lawrence Livermore NationalLaboratory (LLNL).

NATIONAL IGNITION FACILITY UPDATEOCTOBERÐDECEMBER 1998

UCRL-LR-105821-99-1

B-2

NATIONAL IGNITION FACILITY UPDATE

UCRL-LR-105821-99-1

Key Assurance activities to support theProject are all on schedule, including con-struction safety support, litigation support tothe Department of Energy (DOE) for the set-tlement of 60(b) (e.g., quarterly reports), andthe Final Safety Analysis Report (FSAR).

At the end of FY98 there were 10 of 95DOE/OAK (Oakland Office) PerformanceMeasurement Milestones remaining to becompleted, and by the end of first quarter1999, were completed. The only remainingFY98 milestone is the ÒCooling TowersOperational,Ó which will be accomplished,as projected last quarter, in March 1999.

The FY99 DOE/OAK PerformanceMeasurement Milestone plan includes 124milestones. In the first quarter 1999, 11milestones were due, and seven were com-pleted. The cumulative variance for FY99is four milestones. It is anticipated thattwo of the open first-quarter 1999 mile-stones will be completed in January andthe remaining two in February.

In Special Equipment, three 100% reviewswere conducted in October includingAlignment Control, Target Area Alignment,and Final Optics Assembly. Two 100%reviews were conducted in November,including Wavefront Control Systems andRoving Mirror Diagnostic Enclosure. Five100% reviews were conducted in Decemberincluding Electronic Rack Cooling, TimingSystem, Pulse Generator Alignment,Precision Diagnostics, Laser Optics DamageInspection, Roving Mirror and RovingAssemblies.

Table 1 lists Engineering Change Requests(ECRs) that were resolved by the Level 4

and Level 3 Change Control Boards duringthe quarter. No ECRs were over $250K. Thecost impact of these ECRs and all otherchanges (i.e., Cost Transfer Requests) arereported in the change log at the end of eachquarter. Schedule impacts are described inthe individual ECRs. There are no impacts toany Level 0, 1, 2, 3 milestones.

In system integration, the followingaccomplishments were achieved duringthe first quarter 1999.

¥ Final preparations were completedfor the design review of the rackcooling system in December, anddetailed preparations began for theoverall cable plant design review.Development of interface controldocuments (ICDs) defining cableplant requirements made someprogress, but was not completed asrequired. Preparations for procure-ment of electrical racks and cableplant also continued.

¥ Precision Survey staff is completingthe applicable Design Basis Booktemplates in preparation for thedesign review. Working reviewshave been scheduled to review thedraft and final viewgraphs andDesign Basis Book inputs. IntegratedContractor Orders (ICOs) are beingplaced with Argonne, Fermi, andLawrence Berkeley to fund theirattendance and participation in thereview. Stanford Linear AcceleratorCenter (SLAC) is already fundedunder an existing ICO.

TABLE 1. ECRs resolved by the Level 4 and Level 3 Change Control Boards during the first quarter 1999.

ECR Title Resolution

325 Replacement of TBD in Angular Positioning Approved337 Increase Size of TB Fire Sprinkler Ring Approved298 SDR 003 Title II Update Approved209 SSDR 1.8.7 Rev. D from C - FOA Approved344 SSDR Changes for Beam Control and Laser Diagnostics Approved376 Additional Laser Bay Embedded Grounding Plate Approved402 Revise Conduit Routing to MOR Panel Approved

UCRL-LR-105821-99-1 B-3


¥ All cable tray design and lengthdrawings for the Laser Bays are com-plete except for the preamplifier sup-port structure (PASS) area. Thesedrawings are being recorded in theProject Database Management (PDM)system. The PASS structure areadesign effort began in December.Switchyard cable trays for 0Õ0Ó, 50Õ,and Ð21Õ9Ó levels are complete. The7Õ10Ó level has been completely mod-eled to determine possible interfer-ence problems, but final tray layoutshave not been completed.

¥ The Special Equipment rack coolingDesign Review was presented inDecember. Comments and consider-ations offered at the Design Revieware being discussed, and points ofconcern are being addressed. Heatexchanger fans have been orderedfor planned lifetime tests. Reliability,availability, and maintainability(RAM) analysis efforts have begunto determine the reliability figuresfor the overall design. A letter ofintent has been mailed to five ven-dors; no responses have beenreceived at this time.

Site and ConventionalFacilities

Despite the holidays, December wasanother busy and productive month on theNIF site as the total Conventional Facilitieswork completed approached 42%.

With the completion of Phase I work byNDBI, CSP-4, in December, the coordina-tion between prime subcontractors on sitehas been reduced significantly. HenselPhelps (CSP-9) now has unrestrictedaccess and sole occupancy of the LaserBuilding. NDBI (CSP-6/10) has sole pos-session of the Target Building, and NDBI(CSP-5) has sole possession of the OAB.Coordination of access to the work areas,laydown space, and coordination of inter-face between subcontractors in the OABCorridor are now the focus of the overallsite coordination.

The decision to close up the hold-opensalong grid 28 and at the capacitor bays,after the laser bay mat slabs are poured,

will allow a significant amount of work tobe completed this winter that was original-ly scheduled for next winter. Most of thegirts and siding, plenum wall, and rackpiping work can now be completed beforeturning each laser bay over to SpecialEquipment forces.

The Laser Building core continues to bethe area of the project where the scheduledelays are the major concern. The cast-in-place concrete walls and roof in the corefrom Grid 12 to 14 are scheduled to be com-pleted in January (see Figure 1.) The instal-lation of the air handlers between Grid 12and 14 are the first units that must beplaced in the Laser Building, and untilthese units are placed, the remaining unitscannot be installed. Unfortunately, these airhandlers contain the largest fans, which todate have not met the specified require-ments for vibration testing.

Laser Bay 2 overhead and undergroundwork has progressed well and is basicallyback on schedule. Laser Bay 1 is getting offto a slow start but should be completedfaster than originally planned based on theexperience from Laser Bay 2. CapacitorBay 4 slab on grade placement and finegrading of Capacitor Bay 2 and 3 slabswere completed in December, which willopen up significant areas for overheadrough-in to start. The building is dried-into the extent possible considering the hold-opens required, and the site is ready forthe upcoming winter weather.

FIGURE 1. Reinforcing for the concrete walls in theLaser Building mechanical mezzanine. (40-60-1198-2218#5Apb01)

B-4


UCRL-LR-105821-99-1

The OAB is on schedule (see Figure 2),and the variances in the Target Buildingwill be remedied when NDBI concentratesits resources on the critical areas of thebuilding.

Laser SystemsThe past quarter represents a shift in the

emphasis for most of Laser Systems fromdesign to procurement. Several large pro-curement actions were initiated (FrameAssembly Units [FAUs] and plasma elec-trode Pockels cell [PEPC] assembly fixture),and documentation was prepared fornumerous others (flashlamps and maincapacitor charging supplies, for example).Subsystem prototype testing continued toprovide valuable information for valueengineering and procedure development aswell as design validation. Detailed planningdocumentation was completed for FY99.

¥ Oscillator stability measurements at the input to the RegenerativeAmplifier were completed using pro-totype optical pulse generation (OPG)hardware. The results show an rmsinstability of 3.1%, compared with 5%required. However, these tests do notinclude the smoothing by spectral dis-persion (SSD) modulator, which isexpected to increase the instability.The high-frequency variations are dueto the unstabilized laser relaxationoscillations as well as contributionsfrom other components. A feedback

control loop is being fabricated to sup-press these. The low-frequency varia-tion is a result of several sections ofpolarization-maintaining fiber in thesystem that have a polarization-statetemperature dependence. This slowvariation will be minimized with thepolarizing fiber system in the designfor the NIF. This prototype oscillatorsystem has been used during the pastquarter to conduct integrated testsincluding the oscillator and preampli-fier module (PAM). The PAM was operated over a range of outputenergies spanning 0.07 to 22.1 J in aspatially shaped beam. This surpassesthe energy performance requirementsfor the NIF system. Future tests will include SSD, steady-state opera-tion, and demonstration of the tempo-ral pulse-shaping capabilities of theintegrated Master Oscillator Room(MOR)/PAM system. The drawingsfor the long-lead preamplifier beamtransport components were complet-ed and checked during the past quar-ter and are in the process of being pro-cured for the integrated testing sched-uled for late this calendar year.

¥ The amplifier effort remained onschedule during the first quarter ofthe Title III effort. Two large amplifi-er procurements were issued forbids in November as planned: frameassembly units (FAUs) and theEuclid alignment system. Bids arecurrently under evaluation for theFAU procurement. The amplifierteam has also made progress indeveloping an improved correlationbetween observed cleanliness levelsin the amplifier and component pro-cessing. This effort will lead toimproved cleanliness proceduresand component cleanliness specifi-cations for the NIF amplifier.

¥ The PEPC assembly fixture was pro-cured during the past quarter, andthe parts have arrived at LLNL. Thefixture will be assembled for func-tional tests beginning in January. Avalue-engineering effort was initiat-ed in an attempt to simplify andreduce the cost of the PEPC line-replaceable unit (LRU) and controls

FIGURE 2. Optics Assembly Building interior, showingspecial equipment installation at floor level and ductwork at the ceiling level. (40-60-1198-2279#15Apb01)

UCRL-LR-105821-99-1 B-5


hardware. Several cost-reducing andreliability-enhancing features havebeen evaluated on the prototype cellduring the past quarter. Examplesinclude a smaller, lower-cost vacuumpump, circuit boards to mount elec-trical components on the LRU hous-ing, a cold-cathode vacuum gauge toreplace the hot cathode ionizationgauge, and numerous improvementsin the controls design and packaging.Procurement packages are currentlybeing prepared for the PEPC LRUsfor the first bundle.

¥ The primary focus of the power conditioning effort over the pastquarter was validating the design ofthe power conditioning systemmodule through testing of the first-article prototype at Sandia. Over7000 shots have been accumulatedto date. Data indicates that the mod-ule meets specifications when firinginto the resistive dummy load, and simulations predict perfor-mance will meet requirements whendriving flashlamps. The lifetimetests uncovered problems with the design of the commerciallyavailable energy dump resistor,which will be corrected in the NIFdesign. Analysis of the several catas-trophic failures revealed problemswith the design of the module enclo-sure. The design is being modifiedto contain any shrapnel associatedwith electrical faults, while reducingthe pressure pulse contained withinthe enclosure. Capacitor qualifica-tion tests are now under way withsamples from four suppliers. Testshave now been completed on at leastfive samples from each supplier.

Beam Transport SystemFabrication contracts proceeded into

full production, and first deliveries of fin-ished structures have arrived at LLNL.The Switchyard 2 structure contract wasawarded, and the Laser Bay concreterequest for proposals (RFP) was releasedfor bid. In the next quarter, a convoy oftruck-borne structures and vacuum vesselswill begin to arrive at LLNL and occupy

the staging areas. Remaining design activities are nearing completion. Theintensity and level of detail of installationplanning will continue to increase.

¥ Production and shipment of all spa-tial filter stainless plate is complete.Vacuum vessel fabrication of endvessels is proceeding at Ranor.Fabrication and welding of the firstfour vessels is complete and await-ing the start of machining. Fabrica-tion of weldments is progressing onschedule with the next group of fourvessels. Stadco work on the centertransport spatial filter (TSF) has thefirst vessel in machining and the sec-ond vessel in finish welding. Thefirst vessel completion is expected inMarch, and the final unit is sched-uled in late May. March metal workon the center cavity spatial filter(CSF) has the first vessel at Allied formachining, and the second vessel isin final welding.

¥ Initial design development anddetailing is complete on the LaserBay interstage docking frames of thespatial filter (SF) end vessels.Engineering review and designchecking is progressing with theassistance of personnel from thenewly integrated InfrastructureGroup. Completion of design, engi-neering, analysis and specificationsis in progress to support procure-ment review and release in February.

¥ Review of the drawings for theVacuum System was completed, andrequired changes are in progress. It isexpected the drawings will go intoPDM by the end of February.Development of the Beam TransportSystem vacuum control screens hasbegun.

¥ Tests were completed on a method ofprecisely controlling pressure in theamplifier slab cavities. This methodused weight balanced on top of a flex-ible bellows to maintain even pres-sure. Results showed that there is toomuch spring in the bellows to meetthe delta-P requirements. Test equip-ment is being retooled for another can-didate method.

B-6


UCRL-LR-105821-99-1

¥ The Switchyard 2 structure procure-ment RFP was issued for bid. Allprospective contractors attended aprebid conference and constructionsite visit at LLNL. The contract wasawarded in December to AGRA CoastLtd. in Vancouver, BC. The designteam is now continuing on Switch-yard 1 drawing details. Switchyard 1is similar, but not identical to Switch-yard 2. A completely new set of draw-ings is required, equivalent in size tothe Switchyard 2 set.

¥ The Laser Bay concrete pedestal sup-port structure design drawings andanalyses are complete. The specifica-tion is completed and has been incor-porated into the bid package that hasbeen sent out to the bidders. A bid-ders site visit has been scheduled forJanuary. Trial assemblies wereaccomplished on all power amplifierlower structures, which were subse-quently shipped to LLNL.

¥ The transport mirror mount designvalidation team made good progresstoward resolving the Title II actionitem regarding the mounting-induced wavefront effects of trans-port mirrors. Preliminary analysis ofa new design approach was com-pleted in October. Hardware for thenew design was received in mid-November; and assembly of the testmirror in the mount began. Testingof the new mount then began inDecember The effect of gravity- andmount-induced aberrations (predict-ed from detailed finite-element anal-ysis) on propagation for six beam-lines was assessed. The combinationof finite-element analysis and beampropagation analysis indicates thatthe new design will meet therequirements for laser system per-formance. The interferometric mea-surements under way are intendedto validate the finite-element analy-sis. Preliminary results are encour-aging, but show discrepancies thatare not fully understood. Additionalon-line and off-line tests were donein late December (and planned tocontinue during January) to resolvethese discrepancies.

Integrated Computer Control System

All baseline Title II design reviews, concluding with the 100% review of theIntegrated Timing System in December,are complete. ÒNightlight,Ó the first ofseven planned incremental releases of con-trol system software, is complete.Nightlight was a confidence builder. Theengineering team felt that they gained cru-cial experience and that the release suc-cessfully met planned goals. Independenttesting of the software has also been com-pleted. Sixty-nine test incidents werereported, of which twenty-one are catego-rized as software defects; the testing pro-gram is proving its worth at improving thequality of NIF software. The contract forthe design and fabrication of the timingdistribution subsystem was placed.

¥ Seven incremental releases of software(the ÒLightÓ series) are planned dur-ing NIF construction. A repeatableengineering process is used withineach stage to ensure that the product isbeing correctly built and meetsrequirements. To begin the cycle, aportion of the productÕs requirementsis selected for implementation. Eachrelease will be measured against thisdocumented set of goals. As therelease is delivered, a demonstration isheld, and the source code is placedunder configuration management.The software is then subject to a 6-week period during which it is for-mally tested by the independent testteam, which also provides for ÒadhocÓ testing by additional users.Defects that are found result in testincidents that are traced to determinethe root cause. A thorough assessmentof results and the engineering processis made and documented by theimplementation teams. Targeting spe-cific goals, a software implementationplan is written for the next release,which also estimates work needed infuture increments. Measurements aretaken on the source code, packagecount, and number of public methodsin order to assess software complexityand rate of progress. Using this incre-

UCRL-LR-105821-99-1 B-7


mental process, a new part of the NIFsoftware is delivered about every sixmonths. After each release, the resultsare analyzed to determine correctiveactions.

¥ The Nightlight software delivery andtest are complete. In a few cases,some software tasks were deferreduntil the next release so that assess-ment and planning for Penlightcould be completed for maximumbenefit. Nightlight testing goals wereto verify that planned features oper-ate as expected, to establish an inde-pendent test process, and to developthe repeatable processes by whichthe development and independenttest teams could effectively worktogether. Nightlight established test-ing activities consistent with indus-try standards, including: test readi-ness reviews, implementation of atest incident tracking system, adop-tion of standard software defectseverity levels, standardization ofdocumentation methods for test pro-cedures and checklists, and configu-ration management of test productssuch as procedures and results.

¥ The planning process for the secondrelease of the Penlight software wascompleted, culminating in reviews ofthe software applications and sup-porting frameworks. A detailed plan-ning document was prepared thatcontains a review of the Nightlightrelease followed by sections describ-ing each application subsystem and asection covering the supervisoryframework. In addition, the plancoordinates device support, con-troller support, graphical user inter-faces, database coverage, and frame-work coverage.

¥ A procedure for informal code walk-throughs (a line-by-line review of soft-ware) was developed after a piloteffort was completed late last year. Theprimary purposes of the walk-throughs are early error detection,promotion of a team programmingstyle, and dissemination of techniques.The procedure involves assigning twoor three people as reviewers, havingreviewers meet informally with the

code developer, and using a simple,but managed, reporting mechanism.

¥ Automatic alignment softwaredemonstrated concurrent alignmentof 96 beams using software to emu-late the time required to executeevery controlled action. Under thenormal conditions emulated (i.e., nocommanded actuator failed to per-form its nominal requirement), theentire 96-beam laser completedalignment in just over 11 minutes.

¥ The video front-end processor (FEP)delivered for Nightlight exceededperformance goals of three simulta-neous streams of video, each at 10frames per second. The performanceof the newly coded Ada softwareexceeded the C language prototypeseven though the new code is object-oriented. These results help allayconcern that object-oriented soft-ware might not have adequate per-formance over past approaches.

¥ The 100% review of the integratedtiming system was completed inDecember. The meeting went verywell and included a presentation bythe target physics users that conclud-ed the timing system specificationswill lead to a system that can meetNIF performance requirements.

¥ In the integrated safety system, lifetesting of the fiber-optic permissivetransmitters and receivers to be usedin the capacitor bays has made goodprogress. The units have currentlycompleted over 20,000 simulatedcharging cycles.

¥ A new version of programmablelogic controller communications soft-ware was obtained from RockwellSoftware that was directed at fixing apreviously reported incompatibilityproblem. Initial testing of this versionindicates that the problem has beenresolved.

Optical ComponentsOptomechanical Systems. Emphasis in

all areas has shifted to preparing procure-ment packages and awarding contracts forfirst-bundle hardware. Awards were madefor fused silica blanks for polarizers and

B-8


UCRL-LR-105821-99-1

lenses for the OPG system, the CSF lenscassettes, and major procurements for theline-replaceable unit (LRU) assembly veri-fication system.

The team assembled to resolve the opticalmounts action item regarding validation ofthe transport mirror mount design madegood progress. The specification of opticalcomponents for the OPG optical system wascompleted in December (48 optical compo-nent drawings). The 100% review Title IIDesign Review was held in November.

¥ The optical component drawings areawaiting check-off in Sherpa (theDOE/OAK Performance Measure-ment Mile-stone will be completed inJanuary). Subsystems include the fiberlaunch/ regenerative amplifier andmultipass amplifiers (both are part ofthe PAM), the input sensor telescope,the 1:4 split, the preamplifier beamtransport system, and the injectionsystem. This accomplishment resolvesa previously reported problem.

¥ Excellent progress was made on theredesign and documentation of thepreamplifier beam transport system(PABTS). The main part of this designis a six-element, adjustable telescopeto accommodate the various pathlengths, timing, magnification, relayplane, and collimation requirementsof the part of the system just prior toinjection in the TSF. The nominal con-figuration required for the OPGdemonstration has been specified;final analysis and specification of theconfigurations and sensitivities forthe other bundles are in progress.

¥ The 1w and 3w diagnostic relay tele-scope systems (eight per bundle) wereredesigned for lower cost. The newdesigns have fewer elements but moreaspheric surfaces. In addition toreducing cost, these new designsgreatly simplify the ghost and straylight management for the system.

¥ The Title II design review was held inNovember for the KDP crystals. Thereview focused on the ability of rapidand conventional KDP growth tomeet performance specifications andprogress towards facilitization and

demonstration of the KDP finishingequipment and processes. Thereviews demonstrated that DKDP(deuterated) grown by conventionalgrowth meets third-harmonic laserdamage specifications and that rapid-growth processes can produce KDPof the size and quality required forPockels cells and second-harmonicgenerators. There were no significantissues raised during the review.

¥ Castings of first-cluster polarizermaterial occurred in November withan expected delivery of annealed andshaped parts at the end of FY99. Thefirst-cluster BK7 transport mirrormaterial was ordered in December.The material will be stored at the ven-dorÕs warehouses in the United Statesuntil the blanks are needed. Pilotmaterial for first-bundle laser mirrorLM3 is currently being formed andshaped with the first set of blanks sentfrom overseas in December.

¥ An evaluation of the suitability ofthe current sol coating process forproduction of the NIF 1w diagnosticbeam-splitter reflecting surface wascompleted. The analysis indicatedthat current coating materials andpractices will likely meet NIFrequirements.

An ammonia-treated spatial filter(SF7/input) lens was installed in a Novabeamline. This lens is a full-scale demon-stration of a hardened and passivatedcoating in relatively high fluence. Closureon this experiment will come with removaland evaluation of the lens coating proper-ties at the shutdown of the Nova laserlater in FY99.

Laser ControlPreparations and presentations for all

remaining 100% design reviews were com-pleted. Prototype testing began to pay offby validating performance in some casesand identifying areas for improvement inothers. Vendor qualification activitiesincreased at the end of the quarter in prepa-ration for scheduled procurement requisi-tions. Completion of detailed drawing

UCRL-LR-105821-99-1 B-9


packages to support the procurementschedule is currently the largest challenge.

¥ All first-bundle build-to-specifica-tion components for the TSF align-ment tower are in fabrication or pro-curement except for a translationslide requiring manufacturer designmodifications and fiber-optic split-ters and collimators. Requisitionsfor these items will be sent to pro-curement early in the next quarter.Life testing of a commercial posi-tioner for the TSF diagnostic towershowed that longer life limit switch-es must be incorporated. The build-to-specification procurement pack-age for the diagnostic tower is beingheld until the positioner vendormakes an appropriate substitution.Detail drawings for build-to-printcomponents on all tower platformsare scheduled for completion in thenext quarter.

¥ The input sensor was assembled,and most of the sensor test standcomponents were also received. Inaddition, the hardware and softwarefor running the input sensor motorsand shutters were completed.Testing is expected to begin near theend of January. Characteristics of theturning mirror that transmitsapproximately 1% of the PAM out-put beam to the input sensor weremeasured as a function of humidity.Significant transmission variation inthe relative humidity range betweenzero and about 35% suggests thatthe operating atmosphere for thisoptic must either be carefully held atnear-zero humidity (dry nitrogen,for example) or in the 40±10% rangeto avoid unacceptable changes in theinput sensor energy calibration.

¥ Instrumentation for measurement ofoptical coating properties wasdesigned, assembled, and applied tokey components. Effects of contami-nation on sol-gel coatings for diag-nostic beam splitters and the stabilityof antireflection (AR)-coated, beam-sampling gratings in a vacuum envi-

ronment were among the propertiesmeasured. The grating sample effi-ciency was found to change by 25%over an eight-hour period when thetest volume was evacuated. Thisbehavior will be studied further, butmeasurements so far strongly suggestthat an uncoated sampling gratingmay be required.

¥ Fibers received in November fromVavilov Institute were tested for dis-persion and attenuation at 350 nm.The 10-m sample of large core, 435-µm-diam fiber was found to havelow dispersion as the Russiansclaimed, but a longer sample isrequired for an accurate measure-ment. A change of deliverables for thesecond half of their current contractwas proposed to Vavilov so that alonger sample can be provided. If suc-cessful, a single, large-core fiber couldreplace an entire 19-fiber bundle. Arecent sample of 100-µm core fiberwas also tested and found to havehigh attenuation, 400 dB/km as com-pared with the predicted 150 dB/km.This discrepancy will be pursued.

¥ The two vendor-produced deform-able mirror prototypes were bothretested after repairs by their respec-tive companies. In both cases LLNLpersonnel participated in someaspects of the repair work. The testing of one of the rebuilt mirrorswas completed in December, and it was declared performance quali-fied, because the test results wereessentially indistinguishable fromthose of the LLNL prototype mirror.However, some concerns about relia-bility remain. Testing of the othervendor prototype began right beforethe holidays and is continuing intoJanuary. Initial indications are thatperformance will fall short unlessadditional changes are made.

¥ The Nightlight prototype version ofwavefront controller software wasreleased after internal testing in thewavefront laboratory as part ofclosed-loop operation of the prototypedeformable mirrors. After release, the

B-10


UCRL-LR-105821-99-1

controller was tested by the indepen-dent software-testing group, and nodefects beyond those already identi-fied in the laboratory were found.

¥ Full beamline propagation modelingincluding complete measurement-based models of the prototypedeformable mirrors has become amature tool for system optimization.During the quarter, the latest esti-mates of pump-induced aberrationsand gravity and mounting aberrationsfor switchyard mirrors were added tothe baseline propagation model. Themodel was also used to performwavefront correction sensitivity stud-ies with actuator spacing and beamalignment error as variables. Thesestudies showed that a 5% increase infocused energy in the target chambercould be obtained by a small reduc-tion in actuator spacing and that theresulting increase in sensitivity toalignment error was negligible. Thisimprovement, the energy equivalentof adding nine complete beamlines tothe NIF, is the direct result of achiev-ing a convincing wavefront systemmodeling capability.

Target Experimental SystemsThe majority of target chamber subsec-

tion welding was completed in the firstquarter 1999. The vacuum chamber seamswere welded at the LLNL site, see Figure 3.Precision Components Corporation contin-ued to fabricate the weld neck ports.

In Place Machining Corp. has complet-ed and demonstrated a final mockup ofthe equipment that will be used for boringthe holes in the chamber. The machineswill use two single-point tool bits that arehydraulically driven.

¥ The target chamber move date hasbeen changed from March 25, 1999, toJune 1, 1999, and the chamber leakcheck will be performed while thechamber is in the fabrication build-ing. This cuts the Pitt-Des MoinesSteel downtime, or Target Buildingstay-out period, helping Pitt-Des

Moines Steel complete the targetchamber on schedule, and providesbetter weather and drier ground con-ditions for crane site preparation andthe target chamber move.

¥ The analysis of the beam-dump pro-totype placed on Nova was nearlycomplete in November. The datareduction and analysis for the titani-um samples was completed. As pre-liminary results indicated, there wereno showstoppers. The stainless-steellouvers perform well for both firstwall and beam dumps. X-ray fluores-cence tests were performed on theNova contaminated fused silica sam-ples, and the results coordinated wellwith the damage tests performed onthe optics. The results of this workalso did not indicate problems withthe stainless-steel louvers for firstwall and beam dumps.

¥ The target positioner changes havebeen completed to eliminate the separate carriage to which thegraphite-fiberÐreinforced compositeboom was to be attached. This willmake the boom a single unit with themounting pads for the linear bearingsextending from hoops bonded to

FIGURE 3. Target chamber with temporary ties hold-ing the 18 plates together prior to full welding.(40-60-1298-2447#17pb01)

UCRL-LR-105821-99-1 B-11


the boom with an intermediary elec-trically insulating layer. Changing thedouble-walled vacuum vessel to a single-wall vessel having the samecross-sectional moment of inertia isnearly complete. This redesignrequired changes in the linear railbeds and track mounting beams andallowed improvements to the ballscrew intermediate supports.

¥ The interface control document (ICD)for the Diagnostics Data AcquisitionSystem (DAS) cable definition hasbeen completed. The interface defini-tions in the ICD are done but finalcable determination is not available atthis time. The design is currently atthe 35% level. The standards andguidelines document is presentlybeing revised to include the prelimi-nary Diagnostic CommunicationProtocol (DCP). The DCP will be thecommunications interface betweenthe diagnostic controller and the FEP.The DCP messages and their requiredarguments have been defined, andpreliminary DCP application inter-faces should follow soon. Softwaregoals for the Penlight version of theTarget DAS FEP have been deliveredfor publication in the PenlightSoftware Implementation Plan. Workon a Connection Manager daemon iscontinuing. Ada packages to write alog file for the Connection Managerand time-stamp log file entries arecomplete. Currently, an Ada packageto read and parse a configuration fileis under development.

¥ It has been decided that existing Novadiagnostics will satisfy the DiagnosticInstrument Manipulator (DIM) laserverification experimental require-ments. These are the gated x-rayimager and the six-inch manipulator-based streak camera. The resourcesthat were planned for these two diag-nostics will be redirected toward thework required to finish the design ofthe DIM. Design continues on the con-trol and signal cables to support vari-ous target configurations on the frontend of the target positioner. The

mechanical design of the carriage andboom assemblies is changing fromtwo mating components to a singlecomposite assembly that will be morerigid. This new design will have someimpact on the attachment of the cabletracking system but will not affect theschedule.

¥ Design changes are being incorpo-rated into the mirror frames as aresult of meetings with several fabri-cators. These modifications will helpreduce the component cost.

¥ The design layouts for the retractableand removable beam tubes for thelower level that are adjacent to thefloor in the upper and lower mirrorrooms have been prepared and willbe reviewed by the infrastructuresand operations groups before pro-ceeding with detailed design.

¥ Work has been delayed on the tritiumenvironmental protection systemsuntil 2004 to be able to prepare for theintroduction of tritium into the facilityin 2005. The exception to this delay isthe stack monitoring system, whichwill be in place for the first bundle.

¥ The 100% review Title II DesignFinal Optics Assembly (FOA)Review was held in November. Thereview had been delayed by approx-imately one month to allow moretime for scientific review of the lastBeamlet tests using the final opticscell (FOC). The following topicswere included in the review: physicsoverview, subsystem reviews (FOC,integrated optics module, actuationsystem, debris shield cassette, vacu-um isolation valve, 3w calorimeterchamber, thermal control system,vacuum venting system), and proto-type testing results.

¥ Preparations proceeded well forfirst-article procurements of the inte-grated optics module (IOM), theFOC, and 3w calorimeter chamber.The detail drawings for the IOM arecomplete and awaiting check. TheFOC drawing package was releasedunder Configuration Management.

B-12


UCRL-LR-105821-99-1

Operations SpecialEquipment

Title II design is proceeding well. Thishas been a productive quarter with FY99budget planning process, intensivedetailed design continuing, and fast-trackprototyping and testing going well in thesupport laboratory. Several key internalmilestones were completed.

¥ The alpha prototype release of theOperations Special EquipmentControls Supervisor (OSECS) wascompleted and demonstrated inDecember in Bldg. 432. The alphaprototype is the first vertical slice ofOSECS and implements theTransport and Handling (T&H)cover removal operation for the TopLoading (TL) delivery system. The subsystems included in thisrelease are the supervisory softwaremodule and the TL supervisorygraphical user interface, both imple-mented in Java using CORBA, andthe TL safety interlock system. Theacceptance test procedures for thisrelease have been completed, andthe testing will be conducted inJanuary. The detailed design of theLaser Bay Transport System (LBTS)interfaces with OSECS was complet-ed, and the vendor (AGVP) is pro-ceeding on schedule with their soft-ware development tasks.

¥ The detailed design for the BottomLoading (BL), TL, and Side Loadingdelivery systems is progressing. TheT&H team met with the OpticsAssembly Building (OAB) team anddiscussed the OAB docking portrequirements versus the existingdesign. The decision was made toswitch the two docking port (univer-sal and amplifier) locations. Theamplifier docking port will now belocated closer to the elevator. Thiswill allow the canisters to dock to theOAB without rotating the canistersafter a line-replaceable unit (LRU) isinside.

¥ The design effort for the BLAmplifier Slab Delivery System con-tinues. The primary concentrationwas on the cover removal mecha-nism and docking mechanism forthe system. A peer review of this sys-tem was completed this quarter. Allinterface partners were in atten-dance at this review. One of the pro-totype NIF frame assembly unitswill be available for testing of theLRU insertion/removal process.This testing is scheduled forDecember 1999.

¥ Fabrication of the hardware for theLBTS is continuing. The frame isbeing fabricated at Alkab, Inc. Themajority of the parts that make up theweldments have been machined.Welding of the parts has begun, andAGVP is on schedule and ready toreceive the transporter from RedZone.

¥ Detailed optical assembly and align-ment design, prototyping, and pro-curement are progressing, with mostof the prototype equipment now pro-cured, and Phases II and III of OABinstallation are complete. Specialequipment installation into the OABwas completed in November. Thosesystems include optic insertionpedestals, Ergotech supports, andgranite final supports. Jib cranes forairlocks and BL and TL dockingports were also installed in the OAB.An intensive acceptance test wascompleted in December for verticaland airlock lifts, and then installationinto the OAB followed. Specialequipment that will go inside theOAB and mount onto the raised flooris still being specified and procured.

Start-Up ActivitiesOverall progress was good. Integrated

computer control system (ICCS) Nightlighttesting was successfully completed, planningwas initiated for OSEC Alpha release testing,and preparation of detailed test flows andtest summaries for several priority areas ofthe Project continued in support of theMaster Test Plan development effort.

UCRL-LR-105821-99-1 B-13


¥ Two test plans were generated by theNIF independent test team (ITT): onefor the ICCS Nightlight release (NIF-5001097) and one for the OSECAlpha release (NIF-5001272). Severalkey test processes were establishedand demonstrated during ICCSNightlight release testing. These pro-cesses, consistent with industry stan-dards, included use of standard doc-umentation methods and electronicreview/approval for test plans andprocedures, test readiness reviews,release of software to the ITT underconfiguration management, adop-tion of standard software defectseverity level classifications, creationof an electronic test log and test inci-dent tracking database, and configu-ration management of test productssuch as test procedures and summa-ry reports.

¥ Tests and demonstrations were com-pleted for all thirteen ICCSNightlight slices. Overall, softwarequality was found to be good. Atotal of 21 software defects wereidentified, and 20 software enhance-ments were recommended fordevelopment staff consideration as aresult of formal testing by the ITTand ad hoc testing by Nova opera-tions personnel. All software issueshave been converted to softwarechange requests (SCRs) and arebeing tracked by the developmentteam. A summary test report (NIF-5001322) was prepared and issuedfor approval and release. A test pro-cedure is being prepared by the ITTfor the OSEC Alpha release.

¥ To facilitate development of theMaster Test Plan (MTP), the NIF sys-tem was broken down into approxi-mately 40 Òsubsystems,Ó and pointsof contact were assigned from theNIF Operations team to begin inves-tigating test plans being developedby the lead engineers and their sup-port personnel for all high-priorityareas. Information collected is beingdocumented using test flow dia-grams and test summary spread-

sheets, and these products will formthe basis for the MTP. To date, pre-liminary test flows and spreadsheetshave been completed for 15% of thesubsystems (PAM, injection optics,spatial filter LRUs, transport mir-rors, PEPC, and deformable mirror),and another 25% are in progress. Inaddition to individual subsystemtest plans, development of thesequence for subsystem integrationwas initiated. This effort is approxi-mately 35% complete. A draft out-line for the MTP was prepared andissued for review and comment.

¥ A preliminary plan for a system-wide Component Installation Planhas been prepared. It is initially aseries of connected boxes represent-ing cleaning, packaging, transport,or cleanliness validation steps neces-sary for each NIF subsystem. Eachcleaning process is represented by aspecific procedure. Creation of theseComponent Installation Plans willidentify similarities as well as con-scious differences between subsys-tems. The identification of similari-ties will help to ensure that similarsubsystems are cleaned by similarprocesses, using similar equipmentand vendors, and are validated toconsistent standards. The commonprocesses for cleaning and validat-ing the cleanliness of subsystemsidentified during the construction ofthe Component Installation Planswill be documented in engineeringprocedures referred to as MechanicalEngineering Letters (MELs). Sixteenof the documents have already beenwritten and a few have reached astage of maturity in which they havebeen put under configuration con-trol. The contamination control teampresented a status report entitledOrganic Cleanliness and Effects on Sol-gel Coatings and Aerosol GenerationStatus Report in December. Thisreview completed the major work onorganics and their effect on thetransmission of sol-gel coatings. Theconclusion is that all vacuum com-

B-14


UCRL-LR-105821-99-1

ponents need high-pressure or ultra-sonic/Brulin washing; elastomersneed bakeout at 170¡C; and all nitro-gen and argon-filled components willbe high-pressure or ultrasonic/Brulinwashed to protect sol-gel coatingsfrom >0.1% transmission loss in 1000hours. The status of cleanliness controlprocess maps in support of the sys-temwide Component InstallationCleanliness Plan was reviewed inDecember. The discussion includedthe cleaning process selection chart,components for which specific cleanli-ness plans are required, and rules forpreparing the process flow maps. Thestatus of the flowcharts will bereviewed again in February 1999.

¥ A baseline approach for a laser oper-ations model has been defined, large-ly using parts of existing propagationcode, BTGain. Interfaces of thismodel with operational and opticsdatabases, as well as the NIF ICCS,have been explored. A preliminaryeffort analysis to implement thismodel has been completed and willbe used to decide on an action plan.

¥ The final revision of the FY98 oper-ability model report was issued.This included incorporation ofreview comments from the initialrelease. The model results were pre-sented as part of the NIF RAM andOperations Modeling AssessmentsReview held in October.

ES&H and SupportingR&D

Assurances. Key Assurance activitiesduring November to support the Projectincluded: construction safety support, QAprocedures, audits, and surveillance plan-ning, and preparation for upcoming inde-pendent external audits (DOE OAK,Independent Project Assessment). Otheractivities included: litigation support to theDOE for the settlement of 60(b) (e.g., quar-terly reports), the overall litigation againstthe Stockpile Stewardship and ManagementProgrammatic Environmental Impact Report,

interface with the Institutional surveill-ance for buried hazardous/toxic and/orradioactive materials, Final Safety AnalysisReport preparation, and support for envi-ronmental permits and the Pollution Pre-vention and Waste Minimization Plan. All are on schedule.

¥ The Congressionally mandatedIndependent Project Assessment isdue to begin in January 1999. Theproposed presentation material andsupporting documents were pre-pared for the Project ManagerÕsreview. The DOE Field Manager hasprovided the proposed Task A reviewagenda from the Independent ProjectAssessment Team. In a related action,the Management Position Descrip-tions were updated completly andare being edited for the ProjectManagerÕs approval.

¥ The paleontological finds were pre-viously accurately located and re-ported to the University of Calif-ornia Museum of Paleontology. InDecember, these bones were careful-ly covered with sand and then amarker level of gravel. The coveringwas witnessed by a representative of the Environmental ProtectionDepartment.

¥ Elemental analyses have been com-pleted for the first significant neu-tron shielding concrete pours in theTarget Bay. The results show signifi-cant, but manageable, levels ofimpurities. Activation calculationshave shown that the shielding meetsthe design criteria and specifications.Future Target Bay concrete pourswill be sampled as well to ensurecontinued compliance with specifi-cations.

¥ The DOE Oakland Annual Manage-ment Assessment was completed indraft form and factual commentsreturned to the assessment team. Thereport will be issued in January 1999.

Optics Technology. Facilitization is allbut complete at Hoya and Schott with allcontracts placed and pilot activities

UCRL-LR-105821-99-1 B-15


already started. Hiring and training forpilot runs is mostly completed. Both laserglass vendors are doing a great job in sup-porting the NIF schedules.

The Zygo facility was completed in thefirst quarter 1999, and first-bundle optics(LM3) are now being shaped. Spectra-Physics has completed all major facilitywork and is in the process of installingequipment, such as the clean line. CurrentlySpectra-Physics is in the process of install-ing the new planetary, conditioning sta-tions, and preparing the interferometer areafor installation of the new interferometer inthe second quarter. The Laboratory forLaser Energetics (LLE) at the University ofRochester has completed the coating andcleaning facilities and is near completion ofthe metrology facilities that will also becompleted in the second quarter.

¥ The Schott laser glass grinding andpolishing has been installed withassistance from Zygo Corporation,and initial results look very good.The flatness specification and pro-cess flow time should not be anissue. All fine annealing ovens havebeen delivered and set up.

¥ At Hoya, full-scale furnace and lehr are complete. Hoya is still onschedule to start the Laser GlassPilot run mid-January 1999. Two120Ó Lapmaster grit grinder andpolishers have been delivered andset up. First-bundle laser glasscladding strips were produced inJapan and delivered on time.

¥ Six 1000-L crystallizers continued tooperate at LLNL. A KDP boule wasgrown to a size that will yield nearly20 Pockels cell switch crystals. Thetwo DKDP crystals, one horizontaland one vertical, continued to growwithout significant problems. Amathematical model was formulat-ed and tested for the leaching ofimpurities from the growth tank andtheir incorporation into the boule.

¥ Corning has completed the debug-ging of all facilitization equipmentincluding the boule/blank extrac-tion saw, 84Ó Blanchard, overarm

lapper, 24Ó Zygo interferometer andfixturing. Pilot production of NIFfused silica blanks has begun. Allprocess and quality assurance proce-dures have been reviewed andapproved by LLNL. The first opticswere shipped to Tinsley inDecember.

¥ TinsleyÕs lens and window finishingfacility building was completed onschedule, and beneficial occupancywas obtained in November. Mostmanufacturing equipment wasinstalled into the new building. Afew pieces of equipment remain tobe installed. They are expected to beinstalled prior to the optics beingready to be placed on thosemachines. The facility has beenmade ready for start of pilot produc-tion. Early pilot production of NIFoptics has continued at Tinsley. Thefirst three of seven focus lenses arenearing completion. They will beavailable for ultraviolet laser dam-age testing in February. The finish-ing of two cavity mirrors and twodebris shields is still in process. Mainpilot is scheduled to award atTinsley in January. Work has begununder the Flexible Finishing Facilitycontract awarded to Kodak in lateSeptember, and LLNL has begunreceiving the Kodak MonthlyProgress Reports and detailedschedules. All LLNL or government-furnished equipment (GFE) haseither been ordered or alreadyreceived at Kodak. The PilotProduction Option associated withthe Kodak Flexible Finishing Facilitycontract was exercised and awardedduring mid-December.

¥ The automated aqueous cleaning sys-tems from JST Custom Fabricationwere completed and ready for deliv-ery to LLNL at the end of December.Chemat Technology Inc. has complet-ed approximately 80% of the mechan-ical assembly and control softwaredevelopment on the sol dip-coatingsystem.

B-16


UCRL-LR-105821-99-1

¥ Zygo facilitization is complete, andshaping of first-bundle optics hasbegun.

¥ The LLE 54Ó planetary design wascompleted to allow coating of onlyNIF optics in the 72Ó chamber andOmega optics in the 54Ó chamber. Themetrology facilities are under con-struction. All of the major re-modelingwork is complete, but utilities andhigh-efficiency particulate air (HEPA)filter installation need to be complet-ed. LLE conducted an experimentwith an in-situ interferometer toexamine the surface flatness of anoptic during heating and cooling toimprove their model of the appropri-

ate rapid cooldown cycle. This workwill also aid in understanding reflect-ed wavefront distortion due to coatingversus other contributions.

¥ Spectra-Physics received beneficialoccupancy for the remaining facilityin November, contingent on someminor fire sprinkler relocations andlandscaping issues. The metrologylabs are starting to be populatedwith equipment. The Large AreaConditioning Station assembly start-ed in December. The isolation padfor the interferometer was sealed toprevent humidity penetration intothe dry enclosure, and the vibrationisolation table was installed.

C-1

AAlbrecht, G. F., Sutton, S. B., George, E. V.,Sooy, W. R., and Krupke, W. F., Solid StateHeat Capacity Disk Laser, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-130963; also inLaser & Part. Beams 16(4), 605Ð625 (1998).

BBack, C. A., Alvarez, S., Spragge, M.,Bauer, J., Landen, O. L., Turner, R. E., andHsing, W., High Energy DensityExperimental Science in Support of SBSS,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-131975 ABS.Prepared for WomenÕs Technical andProfessional Symp, San Ramon, CA, Oct 15,1998.

Back, C. A., Landen, O. L., Turner, R. E.,Koch, J. A., Hammel, B. A., MacFarlane, J.J., and Nash, T. J., SpectroscopicMeasurements to Investigate and DiagnoseRadiative Heating of Hohlraum Walls,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-131993 ABS.Prepared for 8th Intl Workshop on RadiativeProperties of Hot Dense Matter, Sarasota, FL,Oct 26, 1998.

Belian, A.P., Latkowski, J. F., and Morse, E.C., ÒExperimental Studies of ConcreteActivation at the National Ignition FacilityUsing the Rotating Target NeutronSource,Ó Fusion Tech. 34(3), Pt. 2, 1028Ð1032(1998).

Berger, R. L., Cohen, B. I., Langdon, A. B.,MacGowan, B. J., Rothenberg, J., Still, C.W., Williams, E. A., and Lefebvre, E.,Simulation of the Nonlinear Interaction ofFilamentation with Stimulated Backscatter,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-132598 ABS.Prepared for 1999 Centennial Mtg of theAmerican Physical Society, Atlanta, GA, Mar20, 1999.

Berger, R. L., Langdon, A. B., Rothenberg,J. E., Still, C. H., Williams, E. A., andLefebvre, E., Stimulated Raman and BrillouinScattering of Polarization Smoothed andTemporally Smoothed Laser Beams, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-132259; also inPhys. of Plasmas 6(4), 1043Ð1047 (1999).

Berger, R. L., Still, C. H., Williams, E. A.,and Langdon, A. B., On the Dominant andSubdominant Behavior of Stimulated Ramanand Brillouin Scattering Driven byNonuniform Laser Beams, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-132259; also inPhys. Plasmas 5(12), 4337Ð4356 (1998).

Bibeau, C., Beach, R. J., Mitchell, S. C.,Emanuel, M. A., Skidmore, J., Ebbers, C.A., Sutton, S. B., and Jancaitis, K. S., HighAverage Power 1 mm Performance andFrequency Conversion of a Diode-End-PumpedYb:YAG Laser, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-129612; also in IEEE J. Quant.Electr. 34(10), 2010Ð2019 (1998).

PUBLICATIONS AND PRESENTATIONSOCTOBERÐDECEMBER 1998

UCRL-LR-105821-99-1

Bittner, D. N., Collins, G. W., Monsler, E.,and Letts, S., Forming Uniform HD Layers inShells Using Infrared Radiation, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-131371. Preparedfor 40th Annual Mtg of the Div of PlasmaPhysics, New Orleans, LA, Nov 16, 1998.

Boege, S. J., and Bliss, E. S., Beam Controland Laser Characterization for NIF, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-129854; also inFusion Tech. 34(3), Pt. 2, 1117Ð1121 (1998).

Brereton, S. J., Yatabe, J. M., and Taylor, C.A., The Safety and Environmental Process forthe Design and Construction of the NationalIgnition Facility, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-130860; also in Fusion Tech. 34(3),Pt. 2, 760Ð766 (1998).

Brereton, S. J., Latkowski, J., Singh, M.,Tobin, M., and Yatabe, J., DecommissioningPlan for the National Ignition Facility,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-130859; also inFusion Tech. 34(3), Pt. 2, 1041Ð1046 (1998).

Britten, J. A., and Summers, L. J.,Multiscale, Multifunction DiffractiveStructures Wet-Etched into Fused Silica forHigh-Laser Damage Threshold Applications,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-130115; also inAppl. Optics 37(30), 7049Ð7054 (1998).

CChatterjee, I., Hallock, G. A., and Wootton,A. J., Observation of Multiple Modes of InteriorDensity Turbulence in the TEXT-U Tokamak,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-132555.

Colvin, J. D., Amendt, P. A., Dittrich, T. R.,Tipton, R. E., and Vantine, H. C., Analysis ofLaser Indirect Drive Double-Shell CapsuleImplosion Experiments as a Testbed forBenchmarking Compressible Turbulent MixModels, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-132571 ABS. Prepared for 1999 CentennialMtg of the American Physical Society, Atlanta,GA, Mar 20, 1999.

Cook, R., Buckley, S. R., Fearon, E., andLetts, S. A., New Approaches to thePreparation of Pa MS Beads as Mandrels forNIF-Scale Target Capsules, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-129955; also inFusion Tech. 35(2), 206Ð211 (1999).

Cook, R. C., McEachern, R. L., andStephens, R., Representative Surface ProfilePower Spectra from Capsules Used in Novaand Omega Implosion Experiments, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-129961; also inFusion Tech. 35(2), 224Ð228 (1999).

DDe Yoreo, J. J., Yan, M., Runkel, M., Staggs,M., Liou, L., Demos, S. G., Carman, L.,Zaitseva, N. P., and Rek, Z. U.,Characterization of Optical Performance andDefect Structures in Nonlinear Crystals forLarge Aperture Lasers, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-132557 ABS-S. Prepared for Confon Lasers and Electro-Optics Õ99/QuantumElectronics and Laser Science Conf Ô99,Baltimore, MD, May 23, 1999.

Dimonte, G., Nonlinear Evolution of theRayleighÐTaylor and RichtmyerÐMeshkovInstabilities, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-132439. Prepared for 40th Annual Mtg of theDiv of Plasma Physics, New Orleans, LA,Nov 16, 1998.

Dimonte, G., and Schneider, M., DensityRatio Dependence of RayleighÐTaylor Mixingfor Sustained and Impulsive AccelerationHistories, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-132570. Submitted to Phys. of Plasmas.

Dimonte, G., and Schneider, M., DensityRatio Dependence of RayleighÐTaylor Mixingfor Sustained and Impulsive AccelerationHistories, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-132570 ABS Rev. Prepared for 7th IntlWorkshop on the Physics of CompressibleTurbulent Mixing, St. Petersburg, Russia,Jul 5, 1999.

C-2

PUBLICATIONS AND PRESENTATIONS

UCRL-LR-105821-99-1

Dimonte, G., New Buoyancy-Drag Model forRayleighÐTaylor Mixing and Comparison withExperiments, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-132695 ABS. Prepared for 7th Intl Workshopon the Physics of Compressible TurbulentMixing, St. Petersburg, Russia, Jul 5, 1999.

Dittrich, T. R., Haan, S. W., Marinak, M.M., Pollaine, S. M., Hinkel, D. E., Munro,D. H., Verdon, C. P., Strobel, G. L.,McEachern, R. L., Cook, R. C., Roberts, C.C., Wilson, D. C., Bradley, P. A., Foreman,L. R., and Varnum, W. S., Review ofIndirect-Drive Ignition Design Options for theNational Ignition Facility, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-131383.Submitted to Phys. of Plasmas.

Dittrich, T. R., Haan, S. W., Marinak, M.M., Pollaine, S. M., and McEachern, R.,Reduced Scale National Ignition FusionCapsule Design, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-131383; also in Phys. Plasmas5(10), 3708Ð3713 (1998).

EErlandson, A., Alger, T., Horvath, J.,Jancaitis, K., Lawson, J., Manes, K.,Marshall, C., Moor, E., Payne, S., Pedrotti,L., Rodriguez, S., Rotter, M., Sutton, S.,Zapata, L., Seznec, S., Beullier, J.,Carbourdin, O., Grebot, E., Guenet, J.,Guenet, M., LeTouze, G., and Maille, X.,ÒFlashlamp-Pumped Nd:Glass Amplifiersfor the National Ignition Facility,Ó FusionTech. 34(3), Pt. 2, 1105Ð1112 (1998).

Erlandson, A. C., Jancaitis, K., Marshall,C., Pedrotti, L., Zapata, L., Rotter, M.,Sutton, S., Payne, S., Seznec, S., andBeullier, J., Recent Advances in Flashlamp-Pumped Neodymium-Glass Amplifiers,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-132461 ABS-S.Prepared for Conf on Lasers, and ElectroOptics/Quantum Electronics and LasersScience (CLEO/QELS Ô99), Baltimore, MD,May 23, 1999.

Estabrook, K. G., Remington, B. A., Farley,D., Glenzer, S., Back, C. A., Turner, R. E.,Glendinning, S. G., Zimmerman, G. B.,Harte, J. H., and Wallace, R. J., Nova LaserExperiments of Radiative Astrophysical Jets,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-132258 ABS.Prepared for 193rd Mtg of the AmericanAstronomical Society, Austin, TX, Jan 5,1999.

GGlendinning, S. G., Amendt, P., Cline, B.D., Ehrlich, R. B., Hammel, B. A., Kalantar,D. H., Landen, O. L., Turner, R. E., Wallace,R. J., Weiland, T. J., Dague, N., Jadaud, J.-P., Bradley, D. K., Pien, G., and Morse, S.,Hohlraum Symmetry Measurements withSurrogate Solid Targets, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-129773; also inRev. Sc. Instrum. 70(1), 536Ð542 (1999).

Glendinning, S. G., Laser-Drive ICFFundamentals and Recent ExperimentalResults, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-132158 ABS. Prepared for WomenÕsTechnical and Professional Symp, San Ramon,CA, Oct 15, 1998.

Glenzer, S. H., Alley, W. E., Estabrook, K.G., De Groot, J. S., Haines, M., Hammer, J.H., Jadaud, J.-P., MacGowan, B. J., Moody,J. D., and Rozmus, W., Thomson Scatteringfrom Laser Plasmas, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-131357-DR. Prepared for 40thAnnual Mtg of the Div of Plasma Physics,New Orleans, LA, Nov 16, 1998.

Gresho, P. M., Some Aspects of the Hydro-dynamics of the Microencapsulation Route toNIF Mandrels, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC- 129963; also in Fusion Tech.35(2), 157Ð188 (1999).

C-3


UCRL-LR-105821-99-1

HHammel, B. A., Bernat, T. P., Collins, G. W.,Haan, S., Landen, O. L., MacGowan, B. J.,and Suter, L. J., Recent Advances in IndirectDrive ICF Target Physics at LLNL, LawrenceLivermore National Laboratory, Livermore,CA, UCRL-JC-130302. Prepared for 17th IntlAtomic Energy Agency Fusion Energy Conf,Yokohama, Japan, Oct 19, 1998.

Hammer, J. H., Tabak, M., Wilks, S. C.,Lindl, J. D., Bailey, D. S., Rambo, P. W., Toor,A., Zimmerman, G. B., and Porter, J. L.,High Yield ICF Target Design for a Z-PinchDriven Hohlraum, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-131389. Prepared for 40th AnnualMtg of the Div of Plasma Physics, NewOrleans, LA, Nov 16, 1998.

Hammer, J. H., Tabak, M., Wilks, S., Lindl,J., Rambo, P. W., Toor, A., Zimmerman, G.B., and Porter, J. L., High Yield Inertial FusionDesign for a Z-Pinch Accelerator, LawrenceLivermore National Laboratory, Livermore,CA, UCRL-JC-132152. Prepared for 17th IntlAtomic Energy Agency Fusion Energy Conf,Yokohama, Japan, Oct 19, 1998.

Hammer, J. H., and Ryutov, D. D., LinearStability of an Accelerated Wire Array,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-132255. Preparedfor 40th Annual Mtg of the Div of PlasmaPhysics, New Orleans, LA, Nov 16,1998.

Ho, D. D.-M., Chen, Y.-J., Pincosy, P. A.,Young, D. A., and Bergstrom, P. M.,Hydrodynamic Modeling of the Plasma PlumeExpansion from the Targets for X-RayRadiography, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-132181-ABS. Prepared for 1999 Particle AcceleratorConf, New York City, NY, Mar 29, 1999.

KKalantar, D. H., Remington, B. A.,Chandler, E. A., Colvin, J. D., Gold, D. M.,Mikaelian, K. O., Weber, S. V., Wiley, L. G.,Wark, J. S., and Hauer, A. A., High PressureSolid State Experiments on the Nova Laser,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-129810. Preparedfor 1998 Hypervelocity Impact Symp,Huntsville, AL, Nov 16, 1998.

Kalantar, D. H., Remington, B. A., Colvin,J. D., Gold, D. M., Mikaelian, K. O., Weber,S. V., and Wiley, L. G., Shock CompressedSolids on the Nova Laser, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-132687 ABS.Prepared for 11th Topical Conf on ShockCompression of Condensed Matter, Snowbird,UT, Jun 27, 1999.

Kane, J. O., Arnett, D., Remington, B. A.,Glendinning, S. G., Bazan, G., Drake, R. P.,and Fryxell, B. A., Scaling SupernovaHydrodynamics to the Laboratory, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-132438. Preparedfor 40th Annual Mtg of the Div of PlasmaPhysics, New Orleans, LA, Nov 16, 1998.

Kastenberg, W. E., Industrial EcologyAnalysis Final Report, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-CR-132271.

Kauffman, R., Inertial Confinement FusionMonthly Highlights, September 1998,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-TB-128550-98-12.

Kauffman, R., Inertial Confinement FusionMonthly Highlights, October 1998,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-TB-128550-99-1.

Kauffman, R., Inertial Confinement FusionMonthly Highlights, November 1998,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-TB-128550-99-2.

Key, M. H., Cowan, T. E., Hammel, B. A.,Hatchett, S. P., Henry, E. A., Kilkenny, J. D.,Koch, J. A., Langdon, A. B., Lasinski, B. F.,and Lee, R. W., Progress in Fast IgnitorResearch with the Nova Petawatt LaserFacility, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-132178. Prepared for 17th Intl AtomicEnergy Agency Fusion Energy Conf,Yokohama, Japan, Oct 19, 1998.

Kozioziemski, B., Collins, G. W., Bernat, T.P., and Sater, J. D., Helium Bubble Dynamicsin D-T Layers, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-130038 ABS Rev. Prepared for 1999 Centennial Mtg, Atlanta, GA, Mar 20, 1999.

C-4


UCRL-LR-105821-99-1

Kruer, W., Inertial Confinement FusionQuarterly Report, JanuaryÐMarch 1998,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-LR-105821-98-2.

LLanden, O. L., Amendt, P. A., Turner, R. E.,Glendinning, S. G., Kalantar, D. H.,Decker, C. D., Suter, L. J., Cable, M. D.,Wallace, R., and Hammel, B. A., Time-Dependent, Indirect-Drive Symmetry Controlon the Nova and Omega Facilities, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-130109.Prepared for 17th Intl Atomic EnergyAgency Fusion Energy Conf, Yokohama,Japan, Oct 19, 1988.

Landen, O. L., Amendt, P. A., Suter, L. J.,Turner, R. E., Glendinning, S. G., Haan, S.W., Pollaine, S. M., Hammel, B. A., Rosen,M. D., and Lindl, J. D., Simple Time-Dependent Model of the P2 Asymmetry inCylindrical Hohlraums, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-132445-DR. Prepared for 40thAnnual Mtg of the Div of Plasma Physics,New Orleans, LA, Nov 16, 1998.

Lane, M. A., Van Wonterghem, B. M., andClower Jr., C. A., Phased NIF Start-Up,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-129271; also inFusion Tech. 34(3), Pt. 2, 1127Ð1134 (1998).

Lasinski, B. F., Langdon, A. B., Hatchett, S.P., Key, M. H., and Tabak, M., PICSimulations of Ultra Intense PulsesPropagating through Overdense Plasma forFast-Ignitor and Radiography Applications,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-131245.Submitted to Phys. of Plasmas.

Latkowski, J. F., Neutron Activation of theNIF Final Optics Assemblies and Their Effectupon Occupational Doses, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-129515; also inFusion Tech. 34(3), Pt. 2, 767Ð771 (1998).

Liang, E. P., Wilks, S. C., and Tabak, M., PairProduction by Ultraintense Lasers, LawrenceLivermore National Laboratory, Livermore,CA, UCRL-JC-124499 Rev 1; also in Phys.Rev. Lett. 81(22), 4887Ð4890 (1998).

Lowenthal, M. D., Greenspan, E., Moir, R.,Kastenberg, W. E., and Fowler, T. K.,ÒIndustrial Ecology for Inertial FusionEnergy: Selection of High-Z Material forHYLIFE-II Targets,Ó Fusion Tech. 34(3), Pt.2, 619Ð628 (1998).

MMacGowan, B. J., Berger, R. L., Cohen, B.I., Decker, C. D., Dixit, S., Glenzer, S. H.,Hinkel, D. E., Kirkwood, R. K., Langdon,A. B., and Lefebvre, E., Laser BeamSmoothing and Backscatter SaturationProcesses in Plasmas Relevant to NationalIgnition Facility Hohlraums, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-132191.Prepared for 17th Intl Atomic EnergyAgency Fusion Energy Conf, Yokohama,Japan, Oct 19, 1988.

Marshall, C. D., Erlandson, A., Jancaitis,K., Lawson, J., Sutton, S., and Zapata, L.,Performance of National Ignition FacilityPrototype Laser Amplifier Facility, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-131986 ABS.Prepared for Lasers and Electro-OpticsSociety 1998 11th Annual Conf, Orlando, FL,Dec 1, 1998.

McEachern, R., Matsuo, J., and Yamada, I.,Cluster Ion Beam Polishing for InertialConfinement Fusion Target Capsules,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-130683.Prepared for 12th Intl Conf on IonImplantation Technology, Kyoto, Japan, Jun22, 1998.

Meier, W. R., Systems Modeling and Analysisof Heavy Ion Drivers for Inertial FusionEnergy, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-130954; also in Fusion Tech. 34(3), Pt. 2,326Ð330 (1998).

Moran, M., Brown, C. G., Cowan, T. E.,Hatchett, S., Hunt, A., Key, M., Pennington,D., Perry, M. D., Phillips, T., and Sangster, T.C., Measurements of MeV Photon Flashes inPetawatt Laser Experiments, LawrenceLivermore National Laboratory, Livermore,CA, UCRL-JC-131359. Prepared for 40thAnnual Mtg of the Div of Plasma Physics,New Orleans, LA, Nov 16, 1998.

C-5


UCRL-LR-105821-99-1

NNewton, M., and Wilson, M., MainAmplifier Power Conditioning for the NationalIgnition Facility, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-130975; also in Fusion Tech. 34(3),Pt. 2, 1122Ð1126 (1998).

Norton, M. A., Boley, C. D., Murray, J. E.,Sinz, K., and Neeb, K., Long-Lifetime, Low-Contamination Metal Beam Dumps for NIFSpatial Filters, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-129768. Prepared for 3rd AnnualIntl Conf on Solid State Lasers for Application(SSLA) to Inertial Confinement Fusion (ICF),Monterey, CA, Jun 7, 1998.

PPayne, S. A., and Bibeau, C., PicosecondNonradiative Processes in Neodymium-DopedCrystals and Glasses: Mechanism for theEnergy Gap Law, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-128623; also in J. Luminescence79(3), 143Ð159 (1998).

Petzoldt, R. W., IFE Target Injection andTracking Experiment, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-LR-120192; also in Fusion Tech.34(3), Pt. 2, 831Ð839 (1998).

Porcelli, F., Rossi, E., Cima, G., and Wootton,A., Macroscopic Magnetic Islands and PlasmaEnergy Transport, Lawrence LivermoreNational Laboratory, Livermore, CA, UCRL-JC-132568. Prepared for Workshop on FusionRelated Physics and Engineering in SmallDevices, Trieste, Italy, Oct 6, 1998.

RRemington, B. A., Hydrodynamics Experimentson Intense Lasers, Lawrence LivermoreNational Laboratory, Livermore, CA, UCRL-JC-131249 ABS Rev. Prepared for 1999Centennial Mtg, Atlanta, GA, Mar 20, 1999.

Rhodes, M. A., Fochs, S., and Biltoft, P.,Plasma Electrode Pockels Cell for the NationalIgnition Facility, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-129710; also in Fusion Tech. 34(3),Pt. 2, 1113Ð1116 (1998).

Roberts, C. C., Letts, S. A., Saculla, M.,Hsieh, E. J., and Cook, R., Polyimide Filmsfrom Vapor Deposition: Toward High Strength,NIF Capsules, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-132194; also in Fusion Tech. 35(2),138Ð146 (1999).

Rosen, M., Physics Issues That DetermineInertial Confinement Fusion Target Gain andDriver Requirements: a Tutorial, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-131381.Submitted to Phys. of Plasmas.

Rossetter, E., Electronics Cooling in aDiagnostic Instrument Manipulator,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-ID-132225.

SSanchez, J. J., and Giedt, W. H., ThermalControl of Cryogenic Cylindrical Hohlraumsfor Indirect-Drive Inertial ConfinementFusion, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-132168. Submitted to J. Appl. Phys.

Sater, J., Kozioziemski, B., Collins, G. W.,Mapoles, E. R., Pipes, J., Burmann, J., andBernat, T. P., Cryogenic D-T Fuel LayersFormed in 1 mm Spheres by Beta-Layering,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-128031. Preparedfor 40th Annual Mtg of the Div of PlasmaPhysics, New Orleans, LA, Nov 16, 1998.

Sawicki, R., Bowers, J., Hackel, R., Larson,D., Manes, K., and Murray, J., Engineeringthe National Ignition Facility, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-131522; also inFusion Tech. 34(3), Pt. 2, 1097Ð1104 (1998).

Small IV, W., Celliers, P. M., Da Silva, L. B.,Matthews, D. L., and Soltz, B. A., Two-Color Mid-Infrared Thermometer Using aHollow Glass Optical Fiber, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-128082 Rev 1;also in Appl. Opt. 37(28), 6677Ð6683 (1998).

C-6


UCRL-LR-105821-99-1

Still, C. H., Langer, S. H., Alley, W. E., andZimmerman, G. B., Scientific Programming-Shared Memory Programming with OpenMP, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-131544; also in Comput. Phys. 12(6),577Ð584 (1998).

TTatartchenko, V., and Beriot, E., Growth ofLarge KDP Crystals in the Form of Plates,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-132196.Prepared for 3rd Annual Intl Conf on SolidState Lasers for Application (SSLA) to InertialConfinement Fusion (ICF), Monterey, CA,Jun 7, 1998.

Turner, R. E., Amendt, P., Landen, O. L.,Glendinning, S. G., Bell, P., Decker, C.,Hammel, B. A., Kalantar, D., Lee, D., andWallace, R., Demonstration of TimeDependent Symmetry Control in Hohlraumsby the Use of Drive Beam Staggering,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-132330.Submitted to Phys. Rev. Lett.

WWeber, S. V., Kalantar, D. H., Colvin, J. D.,Gold, D. M., Mikaelian, K. O., Remington,B. A., and Wiley, L. G., Nova ExperimentsExamining RayleighÐTaylor Instability inMaterials with Strength, LawrenceLivermore National Laboratory,Livermore, CA, UCRL-JC-132739 ABS.Prepared for 7th Intl Workshop on thePhysics of Compressible Turbulent Mixing, St.Petersburg, Russia, Jul 5, 1999.

Wegner, P. J., Henesian, M. A., Salmon, J.T., Seppala, L. G., Weiland, T. L., Williams,W. H., and Van Wonterghem, B. M.,Wavefront and Divergence of the BeamletPrototype Laser, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-129724. Prepared for 3rd IntlConf on Solid State Lasers for Application(SSLA) to Inertial Confinement Fusion (ICF),Monterey, CA, Jun 7, 1998.

Wegner, P. J., Auerbach, J. M., Barker, C. E.,Burkhart, S. C., Couture, S. A., De Yoreo, J.J., Hibbard, R. L., Liou, L. W., Norton, M.A., and Whitman, P. A., FrequencyConverter Development for the NationalIgnition Facility, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-129725. Prepared for 3rd IntlConf on Solid State Lasers for Application(SSLA) to Inertial Confinement Fusion (ICF),Monterey, CA, Jun 7, 1998.

Wharton, K. B., Kirkwood, R. K., Glenzer,S. H., Estabrook, K. G., Afeyan, B. B.,Cohen, B. I., Moody, J. D., MacGowan, B.J., and Joshi, C., Observation of ResonantEnergy Transfer between Identical-FrequencyLaser Beams, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-JC-130320-DR. Prepared for 40th Annual Mtgof the Div of Plasma Physics, New Orleans,LA, Nov 16, 1998.

Wharton, K. B., LaserÐPlasma InteractionsRelevant to Inertial Confinement Fusion,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-LR-132566.

Wilks, S. C., Liang, E. P., and Tabak, M.,Pair Production by Ultraintense Lasers,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-JC-124499 Rev 1.Submitted to Phys. Rev. Lett.

Williams, W., Lawson, J., and Henesian,M., PHOMGLASSÑInterferometry AnalysisSoftware for Laser Glass Slab Blanks,Lawrence Livermore National Laboratory,Livermore, CA, UCRL-CODE-99001.

ZZaitseva, N., Carman, L., Smolsky, I.,Torres, R., and Yan, M., Effect of Impuritiesand Supersaturation on the Rapid Growth ofKDP Crystals, Lawrence LivermoreNational Laboratory, Livermore, CA,UCRL-JC-132677 Rev 1. Submitted to J. ofCrystal Growth.

C-7


UCRL-LR-105821-99-1

ICF/NIF and HEDES ProgramLawrence Livermore National LaboratoryP.O. Box 808, L-475Livermore, California 94551

Address Correction Requested

unt digital library/67531/metadc623893/m2/1/high_re… · inertial confinement lawrence livermore...

Documents