Science on High-Energy Lasers: From Today to the NIF Page... · 2011. 3. 22. · June 1995 UCRL-ID-119170 LAWRENCE LIVERMORE NATIONAL LABORATORY University of California • Livermore,

June 1995

UCRL-ID-119170

LAWRENCE LIVERMORE NATIONAL LABORATORYUniversity of California • Livermore, California • 94550

Science on High-Energy Lasers:From Today to the NIF

Richard W. Lee,Richard Petrasso,Roger W. Falcone

This report has been reproduceddirectly from the best available copy.

Available to DOE and DOE contractors from theOffice of Scientific and Technical Information

P.O. Box 62, Oak Ridge, TN 37831Prices available from (615) 576-8401, FTS 626-8401

Available to the public from theNational Technical Information Service

U.S. Department of Commerce5285 Port Royal Rd.,

Springfield, VA 22161

DISCLAIMER

This document was prepared as an account of work sponsored by an agency of the United States Government. Neitherthe United States Government nor the University of California nor any of their employees, makes any warranty, expressor implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of anyinformation, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights.Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, orotherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United StatesGovernment or the University of California. The views and opinions of authors expressed herein do not necessarily stateor reflect those of the United States Government or the University of California, and shall not be used for advertising orproduct endorsement purposes.

Work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory underContract W-7405-Eng-48.

Science on High-Energy Lasers:From Today to the NIF

R. W. LeeR. PetrassoR. W. Falcone

January 1995

UCRL-ID-119170

This page intentionally left blank

UCRL-JC-119170

i

Table of Contents

Acknowledgments ..................................................................................................................... ix

Introduction................................................................................................................................. xi

Section I—High-Energy Lasers in the Context of Current Science ................................... 1

Section II—Overview of Science on High-Energy Lasers ................................................... 3A. Astrophysics and Space Physics ................................................................................. 3B. Hydrodynamics ............................................................................................................. 4C. Material Properties ........................................................................................................ 5D. Plasma Physics .............................................................................................................. 6E. Radiation Sources .......................................................................................................... 7F. Radiative Properties ....................................................................................................... 8

Section III—Experimental Capabilities .................................................................................. 9A. Definition of Existing Facility ...................................................................................... 9

The Laser..................................................................................................................................................... 9Laser Beams........................................................................................................................................... 9Beam Pointing, Aiming, and Synchronicity ...................................................................................... 10Laser Pulse Shapes .............................................................................................................................. 10Laser Shot Rate ................................................................................................................................... 10Beam Diagnostics ............................................................................................................................... 10

Diagnostic Capabilities ........................................................................................................................... 10Overview of Target Chamber Diagnostics ......................................................................................... 11X-ray Imagers ...................................................................................................................................... 11X-ray Spectrometers ........................................................................................................................... 13Optical Diagnostics ............................................................................................................................ 14Hohlraum Radiation Sources.............................................................................................................. 15

B. NIF Facility Definitions............................................................................................... 16Laser Beams ............................................................................................................................................. 17Beam Pointing, Aiming, and Synchronicity ........................................................................................ 17Laser Pulse Shapes .................................................................................................................................. 17Laser Shot Rate ........................................................................................................................................ 17Beam Diagnostics .................................................................................................................................... 17

C. NIF Facility Modifications ......................................................................................... 18Higher Beam Irradiance Requirements ............................................................................................... 18Variations in Lens-Focusing Requirements ......................................................................................... 18Target Chamber Insertion Mechanisms ............................................................................................... 18Added Target Alignment Capabilities ................................................................................................. 18Reduced Time Between Experiments ................................................................................................... 18Flexible Time Intervals Between Experiments .................................................................................... 19Short-Pulse Beams ................................................................................................................................... 19

D. References..................................................................................................................... 19

ii

Section IV—Astrophysics and Space Physics ..................................................................... 21A. Opacity.......................................................................................................................... 21B. Equation of State .......................................................................................................... 24C. Plasma Spectroscopy in High-Energy Astrophysics .............................................. 26

The Bowen Mechanism .......................................................................................................................... 26Low-Mass X-ray Binaries ....................................................................................................................... 27Shock-Wave Ionized Media ................................................................................................................... 29

D. Supernova Instabilities ............................................................................................... 29E. High-Velocity Cratering ............................................................................................. 32F. Thermonuclear Reaction Rates in Stars .................................................................... 32G. Electron-Positron Plasmas ......................................................................................... 34H. References .................................................................................................................... 36

Section V—Hydrodynamics .................................................................................................... 37A. Techniques ................................................................................................................... 37

Radiography—A Typical Configuration ............................................................................................. 37Shock Planarity ........................................................................................................................................ 39

B. Stable Hydrodynamics................................................................................................ 41Planar Geometry...................................................................................................................................... 41Spherical Geometry ................................................................................................................................. 42

C. Unstable Hydrodynamics .......................................................................................... 44Planar Geometry...................................................................................................................................... 44

Imposed Perturbation—Rayleigh-Taylor ........................................................................................... 44Imbedded Random Surface—Rayleigh-Taylor ................................................................................... 46Imbedded Interface—Richtmyer-Meshkov ......................................................................................... 49

Spherical Geometry ................................................................................................................................. 51

D. Future NIF Experiments ............................................................................................ 54Stable Flow ............................................................................................................................................... 54

Shock–Shock Interactions ................................................................................................................... 54Shock–Boundary Interactions ............................................................................................................. 55Hypersonic Flow ................................................................................................................................. 57Impact Cratering ................................................................................................................................. 57Scaled Radiative Energy Coupling ..................................................................................................... 58

Unstable Flow .......................................................................................................................................... 59Classical Instabilities .......................................................................................................................... 59Other Instabilities ............................................................................................................................... 59Turbulent Flow and Vortex Dynamics............................................................................................... 61

E. References ..................................................................................................................... 62

Section VI—Material Properties ............................................................................................ 63A. Equation of State ......................................................................................................... 63

Directly Driven High-Pressure Shocks ................................................................................................ 64Indirectly Driven Colliding Foil Experiments..................................................................................... 64Indirectly Driven Shock Experiments on Polystyrene ....................................................................... 67

B. Opacity .......................................................................................................................... 68

iii

C. Strength of Material .................................................................................................... 71Simple Compression Studies ................................................................................................................. 72Plasticity and Adiabatic Shear Bands ................................................................................................... 72Metastable Phases ................................................................................................................................... 73Phase-Change Measurements ............................................................................................................... 73Rear-Surface Studies ............................................................................................................................... 74

D. Future NIF Experiments ............................................................................................ 76The Principal Hugoniot .......................................................................................................................... 77Multiple Shock States ............................................................................................................................. 80Isentropic Release States......................................................................................................................... 82Isochoric Heating Experiments ............................................................................................................. 84X-ray Diffraction Studies........................................................................................................................ 85Other Material Properties....................................................................................................................... 87

E. References ..................................................................................................................... 87

Section VII—Plasma Physics .................................................................................................. 89A. Interpenetrating Plasmas ........................................................................................... 89B. Plasma Streaming in Magnetic Fields ........................................................................ 91C. Laser-Plasma Instabilities............................................................................................ 91

Stimulated Raman Scattering ................................................................................................................ 93Stimulated Brillouin Scattering ............................................................................................................. 94Filamentation Instability ........................................................................................................................ 95

D. Future NIF Experiments ............................................................................................ 96Laser-Plasma Interactions in Large, Hot Plasmas .............................................................................. 97

Long Scale-Length Plasma Production............................................................................................... 97Stimulated Brillouin and Raman Scattering ...................................................................................... 98Filamentation ...................................................................................................................................... 99

Short-Pulse High-Intensity Laser-Plasma Interactions.................................................................... 100Study of Very High Magnetic Fields .................................................................................................. 101Generation of Large Fluxes of High-Energy Electrons .................................................................... 102Interpenetrating Plasmas and Turbulence......................................................................................... 102Nuclear Reactions in Large-Volume, High-Temperature Plasmas................................................ 102

E. References ................................................................................................................... 103

Section VIII—Radiation Sources ......................................................................................... 107A. Spectrally Continuous Sources ............................................................................... 107

Conversion Efficiency of Gold Burn-Through Foils ......................................................................... 107Spectral Character of Gold Burn-Through Foils ............................................................................... 107Temporal Behavior of Gold Burn-Through Foils ............................................................................. 109Spectral Character of Other Foils ........................................................................................................ 109Angle of Incidence vs Conversion Efficiency ................................................................................... 112Laser Intensity vs Conversion Efficiency ........................................................................................... 114Hohlraum Radiation Temperature Scaling ....................................................................................... 114

B. Spectrally Narrow Sources ....................................................................................... 116Effect of Z on Emission ......................................................................................................................... 116Comments on Sparseness of Data ....................................................................................................... 117

iv

C. X-ray Laser Sources................................................................................................... 117Studying X-ray Lasers........................................................................................................................... 117Two-dimensional Plasma Imaging ..................................................................................................... 118Development of X-ray Lasers as Probes ............................................................................................. 121XUV Interferometry .............................................................................................................................. 125

D. Future NIF Experiments .......................................................................................... 127Incoherent X-ray Sources ..................................................................................................................... 127

Broadband Sources ............................................................................................................................ 127Narrow or Line Sources .................................................................................................................... 128

Coherent X-ray Sources ......................................................................................................................... 128X-ray Lasers ...................................................................................................................................... 128High-Order Harmonic Generation ................................................................................................... 130Applications of X-ray Lasers ............................................................................................................. 131

Particle Sources ...................................................................................................................................... 131

E. References ................................................................................................................... 132

Section IX—Radiative Properties ......................................................................................... 133A. Atomic Physics and Isolated-Emitter Spectroscopy ............................................ 133B. Plasma-Emitter Radiative Properties ...................................................................... 136C. Dynamics Properties ................................................................................................. 137D. Plasma Spectroscopy ................................................................................................ 138E. Radiative Transfer ..................................................................................................... 139

LTE Radiation Flow .............................................................................................................................. 141Non-LTE Radiative Transfer ............................................................................................................... 142

F. Future NIF Experiments ........................................................................................... 144Spectroscopy of High-Z Elements....................................................................................................... 144The Transparency Window ................................................................................................................. 146Strongly Correlated Effects .................................................................................................................. 149Plasma Spectroscopic Topics ............................................................................................................... 153

Spectral Line Shifts, Level Shifts, and Continuum Lowering .......................................................... 153Ion Dynamics.................................................................................................................................... 155Continuum Measurements ............................................................................................................... 156

Population Kinetics ............................................................................................................................... 156Radiation Transfer and Line Formation............................................................................................. 157Strong Magnetic Field Effects .............................................................................................................. 159

G. References................................................................................................................... 159

v

Figures and Tables

Section III—Experimental CapabilitiesFigure III-1. Schematic view of a hohlraum .................................................................................... 15

Figure III-2. Schematic view of several different hohlraums ....................................................... 15

Figure III-3. Measured radiation temperature vs input laser power ........................................... 16

Table III-1. Diagnostics available on Nova, with their abbreviated names ............................... 12

Section IV—Astrophysics and Space PhysicsFigure IV-1. Comparison of the regimes of temperature and density ........................................ 22

Figure IV-2. Variation of temperature and density with depth for models ofstellar structure ............................................................................................................. 23

Figure IV-3. Effects of improved opacity and equations of state on understandingCepheid pulsations ...................................................................................................... 25

Figure IV-4. Grotrian diagram for the Bowen mechanism ........................................................... 26

Figure IV-5. Illustration of the process of dielectronic recombination........................................ 28

Figure IV-6. Rayleigh-Taylor mix in supernova modeling ........................................................... 30

Figure IV-7. Schematic setup of an astrophysical mix experiment.............................................. 31

Figure IV-8. Reaction rates that form the CNO bi-cycle................................................................ 34

Figure IV-9. Trajectories in temperature and density parameter space for some big-bangnucleosynthesis models............................................................................................... 35

Section V—HydrodynamicsFigure V-1. Side view of the jet experimental arrangement ......................................................... 37

Figure V-2. Microscope photographs of the cylindrical carbon foam piece ............................... 38

Figure V-3. Comparison of experimental data with simulation at t = 19.5 ns (early time) ...... 39

Figure V-4. Comparison of experimental data with simulation at t = 25.8 ns (late time) ........ 39

Figure V-5. Schematic of a shock breakout experiment ................................................................ 39

Figure V-6. Two candidate mounting schemes to study planarity ............................................. 40

Figure V-7. Breakout data for the two different mounting schemes, standard and tophat ..... 40

Figure V-8. Schematic of a test bed for hydrodynamic simulations............................................ 41

Figure V-10. The WAX, a 12-frame x-ray gated imager ................................................................ 43

Figure V-11. Typical implosion sequence for a 1-ns square drive pulse..................................... 43

Figure V-12. Time history of an implosion...................................................................................... 44

Figure V-13. Schematic of a setup to study large-growth Rayleigh-Taylor instability ............. 45

vi

Figure V-14. Face-on streak-camera image of the backlight source absorbed by the foil ........ 46

Figure V-15. Intensity traces for the accelerated foil ...................................................................... 46

Figure V-16. Diagram of target for study of imbedded random surface .................................... 46

Figure V-17. Schematic of point projection spectroscopy technique ........................................... 47

Figure V-18. Spectrometer images from point projection spectroscopy ..................................... 48

Figure V-19. Radiograph of a sulfinated plastic ablator driving a molybdenum foil, above,with three intensity traces .......................................................................................... 49

Figure V-20. Axial distribution of mass areal densities ................................................................. 50

Figure V-21. Gated x-ray image of planar Richtmyer-Meshkov mix target .............................. 50

Figure V-22. A 100-µm-wide vertical trace of the centerline of the image in Fig. V-21 ........... 50

Figure V-23. Vertical trace of the image in Fig. V-21 normalized by the backlight ................... 51

Figure V-24. Machined “bumpy” ball used in implosions ........................................................... 51

Figure V-25. Expanded view of an implosion image taken using the WAX.............................. 52

Figure V-26. Measured trace quantifying the limb brightening seen in Fig. V-25 .................... 52

Figure V-28. Streak camera spectra from a microsphere with an iron dopant .......................... 53

Figure V-29. Schematic of a shock–shock interaction experiment ............................................... 55

Figure V-30. Graph of electron density vs time in µs for two cases............................................. 55

Figure V-31. Propagation of a shock in an inhomogeneous medium ......................................... 56

Figure V-32. Setup to explore the hydrodynamic response of a simulated soil ........................ 57

Figure V-33. Experimental results of hydrodynamic response of soil ........................................ 58

Figure V-35. The results of a simulation of the generation of vorticesin an incompressible flow ........................................................................................... 61

Section VI—Material PropertiesFigure VI-1. Schematic of direct laser-driven shock experiment ................................................. 64

Figure VI-2. Streak image of direct laser-driven shock experiment ............................................ 65

Figure VI-3. Schematic of radiation-driven shock using a 2-step target foil .............................. 65

Figure VI-4. Streak camera image of radiation-driven shock using a 2-step target .................. 66

Figure VI-5. Schematic setup with target foil shielded and stepped ........................................... 67

Figure VI-6. Overview of arrangement for radiation-driven shock experiments ..................... 68

Figure VI-7. Schematic of point projection spectroscopy method for measuring opacity ....... 69

Figure VI-8. Absorption (opacity) of the aluminum/niobium sample ....................................... 70

Figure VI-9. Schematic of x-ray diffraction technique ................................................................... 72

Figure VI-10. Schematic of a rear-surface shock breakout experiment ....................................... 74

Figure VI-11. Data from the rear-surface shock breakout experiment........................................ 75

Figure VI-12. Spectra from the probed shocked rear surface of a crystal ................................... 75

Figure VI-13. Plot of stress vs strain rate domain........................................................................... 77

Figure VI-14. The principal Hugoniot defined ............................................................................... 78

Figure VI-15. Relationships of the Hugoniot, an isotherm, and an isentrope............................ 78

Figure VI-16. Double-shock compression experiment................................................................... 80

Figure VI-17. Schematic of double-shock compression experiments .......................................... 81

vii

Figure VI-18. Model of pressure history vs time ............................................................................ 82

Figure VI-19. Model of pressure history vs volume....................................................................... 83

Figure VI-20. The three shock regimes in solids ............................................................................. 84

Figure VI-21. The Seeman-Bohlin camera ....................................................................................... 86

Section VII—Plasma PhysicsFigure VII-1. Simulation of an interpenetrating plasma ............................................................... 90

Figure VII-2. Schematic of colliding plasma experiment .............................................................. 90

Figure VII-3. X-ray image of colliding plasma experiment........................................................... 90

Figure VII-4. Optical image of a plasma streaming into a magnetic field .................................. 91

Figure VII-5. Schematic of regions where laser-plasma instabilities can occur ......................... 93

Figure VII-6. Wave-matching condition for stimulated Raman scattering ................................ 93

Figure VII-7. Contour plot of intensity of Raman-scattered light ................................................ 93

Figure VII-8. Wave-matching condition for stimulated Brillouin scattering ............................. 94

Figure VII-9. Diagram of physical mechanism that generates the reflected wave ................... 94

Figure VII-10. Spectra for the back-reflected light due to stimulated Brillouin scattering ...... 95

Figure VII-11. Image of laser-induced filamentation ..................................................................... 96

Figure VII-12. Schematic of the formation of a large-volume hot plasma .................................. 98

Figure VII-13. Time history of emission of a gasbag ..................................................................... 99

Figure VII-14. Schematic of concept of fast-ignitor scheme for heating ................................... 101

Table VII-1. The number of reactions for 3He + 3He.................................................................... 103

Table VII-2. Selected reactions having rates larger than the 3He + 3He rate ............................ 103

Section VIII—Radiation SourcesFigure VIII-1. Total x-ray emission, measured as

conversion efficiency vs target thickness ............................................................. 108

Figure VIII-2. Measurement of transmitted laser energy............................................................ 108

Figure VIII-3. Time history of spectrum vs energy in eV........................................................... 109

Figure VIII-4. Spectrum of the gold M-bands .............................................................................. 110

Figure VIII-5. Temporal history of spectrally integrated emission from the rear side ........... 110

Figure VIII-6. Spectral character of various elements from Z = 56 to Z = 63 ........................... 111

Figure VIII-7. Absolutely measured spectra emitted by laser-irradiated targets .................... 112

Figure VIII-8. Absolute conversion efficiency of laser-irradiated targets................................. 113

Figure VIII-9. Effect of angle of incidence of laser on x-ray conversion efficiency ................ 113

Figure VIII-10. Equivalent radiation temperature vs time for several gold hohlraums ........ 115

Figure VIII-11. Conversion efficiency of various discrete transitions ....................................... 116

Figure VIII-12. Schematic of double-pulse irradiation technique.............................................. 118

Figure VIII-13. Schematic of traveling-wave x-ray laser ............................................................. 119

Figure VIII-14. Schematic of grating technique ............................................................................ 119

Figure VIII-15. X-ray line emission in 3rd and 4th order with traveling wave ....................... 120

viii

Figure VIII-16. Time history of hard x-ray and XUV emission spectra.................................... 120

Figure VIII-17. Time history of hard x-ray (4–2 transitions) andJ = 2–1 x-ray laser transition ................................................................................. 121

Figure VIII-18. Schematic of experimental setup for plasma imaging...................................... 121

Figure VIII-19. X-ray-laser-backlit image of accelerated foil ..................................................... 122

Figure VIII-20. Moiré deflectogram of laser-irradiated plastic target ...................................... 123

Figure VIII-21. Schematic of x-ray microscope ............................................................................. 124

Figure VIII-22. Image of resolution test pattern obtained with x-ray laser microscope ........ 124

Figure VIII-23. X-ray microscope images of rat sperm nuclei ................................................... 125

Figure VIII-24. Experimental setup for x-ray laser interferometry............................................ 126

Figure VIII-25. Interferogram obtained using x-ray laser interferometer................................. 126

Figure VIII-26. Collisionally pumped x-ray laser wavelengths ................................................ 129

Figure VIII-27. Streak image of coherent XUV radiationgenerated by high-order harmonics .................................................................... 130

Section IX—Radiative PropertiesFigure IX-1. Schematic of dot spectroscopy technique ................................................................ 135

Figure IX-2. Example of the spectrum of lines of holmium XXXIX ........................................... 135

Figure IX-3. Ratios of various line emission intensities ............................................................... 135

Figure IX-4. Schematic of a spectroscopy experiment ................................................................. 138

Figure IX-5. Temperature measurements from a spectroscopy experiment ............................ 139

Figure IX-6. Ratio of the chromium to titanium helium-like 1s2–1s3p intensities .................. 140

Figure IX-7. A spectrum of the chromium and titanium produced by irradiatinga large bag of gas ......................................................................................................... 140

Figure IX-8. Schematic of radiative flow experiment .................................................................. 141

Figure IX-9. Possible experimental packages for radiation flow experiment .......................... 142

Figure IX-10. Streak-camera record of backlight .......................................................................... 142

Figure IX-11. Schematic of setup to study NLTE phenomena ................................................... 144

Figure IX-12. Spectrum from absorption of untamped boron nitride ....................................... 145

Figure IX-13. Experimental results for absorption of a gold backlight ..................................... 146

Figure IX-14. Schematic of proposed experimentto obtain uniform temperature and density ......................................................... 147

Figure IX-15. Temperature and density vs time for an overtamped sample ........................... 148

Figure IX-16. Temperature and density vs time for an untamped sample............................... 149

Figure IX-17. Transparency window (calculated) ........................................................................ 150

Figure IX-18. Temperature and density parameter configurations for variouslaser experiments ...................................................................................................... 151

Figure IX-19. Temperature and density diagram for a specific element (aluminum) ............ 152

Figure IX-20. Effects of line shapes, line merging, and continuum lowering .......................... 154

Figure IX-21. Temperature and density vs time for an undertamped sample ......................... 158

ix

Acknowledgments

The purpose of this document is to present toa general audience the many possible technicalareas that can be explored on a high-energy laser.The nature of a document of this kind impliesthat a certain level of detail will be lost. Weaccept this, but feel nonetheless that presentingthe many areas of past, current and futureinterest will be informative and will provide astarting point for further discussions.

Since a document of this kind must rely oninformation from numerous individuals, wewould like to make an acknowledgment of thesecontributions here. The difficult work of scientistsfrom the other institutions in the US should berecognized and we take this opportunity to thankall those working at the high-energy laserfacilities in the United States—at the NavalResearch Laboratory (NRL), the Laboratory forLaser Energetics (LLE) and Los Alamos NationalLaboratory (LANL), and the Lawrence LivermoreNational Laboratory (LLNL)—for the manydiscussions and contributions that have beenmade to the generation of much of the workreported. We also would like to acknowledgethose that helped directly with the presentdocument, and will do this by scientific area:

Astrophysics and Space PhysicsD. Arnett (U of AZ), John Castor (LLNL),D. Dearborn (LLNL), M. Horanyi (U of CO),S. Kahn (UCB), L. Klein (Howard), R. Klein(LLNL/UCB), K. Kumar (Howard), E. Liang(Rice), J. Luhmann (UCD), G. Mathews (LLNL),B. Remington (LLNL), D. Savin (UCB), D.Schramm (U of Chicago), T. Weaver (LLNL),J. Weisheit (Rice), S. Woosley (UCSC).

HydrodynamicsR. Cauble (LLNL), L. Da Silva (LLNL),P. Demotakis (Cal Tech), G. Dimonte (LLNL),B. Hammel (LLNL), J. Kilkenny (LLNL), L. Klein(Howard), O. Landen (LLNL), P. Marcus (UCB),

Paul Miller (LLNL), T. Perry (LLNL), T. Peyser(LLNL), B. Remington (LLNL), Ö. Savas (UCB).

Material PropertiesN. Ashcroft (Cornell), R. Cauble (LLNL),R. Christensen (UCD), N. Ghoniem (UCLA), NeilHolmes (LLNL), J. Keady (LANL), M. Kreisler(U of MA), T. Perry (LLNL), J. Wark (Oxford),N. Woolsey (Oxford).

Plasma PhysicsC. Back (LLNL), B. Barletta (LBL), H. Baldis(LULI), R. Berger (LLNL), S. Cramer (UCD),L. Goldman (U of Rochester), A. Kermin (MIT),B. Kruer (LLNL), C. Liu (U of MD), N. Luhmann(UCD), B. MacGowan (LLNL), J. Moody (LLNL),T. Peyser (LLNL), P. Rambo (LLNL).

Radiation SourcesC. Back (LLNL), L. Da Silva (LLNL), R. Freeman(Bell), C. Jacobsen (Stony Brook), M. Key(Rutherford), J. Koch (LLNL), D. Matthews(LLNL), J. Molitoris (LLNL), J. Nilsen (LLNL),M. Richardson (CREOL), D. Umstadter (U of MI).

Radiative PropertiesC. Back (LLNL), L. Collins (LANL), J. Castor(LLNL), U. Feldman (NRL), H. Griem (U of MD),C. Hooper (U of FL), W. Hsing (LANL), S. Kahn(UCB), J. Koch (LLNL), J. Moreno (LLNL),A. Osterheld (LLNL), G. Pollak (LANL), J. Reader(NIST), D. Savin (UCB), J. Seely (NRL), R. Ward(LLNL), J. Weisheit (Rice), B. Young (LLNL).

Facility & DiagnosticsC. Clower (LLNL), G. Tietbohl (LLNL), J. Colvin(LANL), J. Kilkenny (LLNL).

Finally, we would like to thank J. Azevedo,J. Olivera, and S. Jennings for all their help inputting this together. It is by virtue of the hardwork invested by all these individuals that thisdocument exists.

Experimental Capabilities x Section III


xi

Introduction

There is considerable interest in the generalscientific community and in those area of thedefense program of the Department of Energythat are concerned with high-energy-densityscience in scientific studies using high-energylasers. These communities envision programs ofwork that are at once scientifically stimulatingand accessible to a broadly based constituency.The ideas in this document are the result of longconsultation with many individual researchers,from all over the world and from many differentfields of science. Their ideas focus on futureapplications at the National Ignition Facility(NIF).

This document presents both a concisedefinition of the current capabilities of highenergy lasers and a description of capabilities ofthe NIF. Five scientific areas are discussed(Astrophysics, Hydrodynamics, MaterialProperties, Plasma Physics, Radiation Sources,and Radiative Properties). In these five areaswe project a picture of the future based oninvestigations that are being carried on today.Even with this very conservative approach wefind that the development of new higher energylasers will make many extremely exciting areasaccessible to us.

In the area of astrophysics we find that ahigh-energy laser can generate such extremeconditions that it becomes possible to studynumerous previously inaccessible areas. Amongthese are the radiative opacity of the outerenvelopes of stars, key to our understanding ofhow stars evolve. Another is the equation ofstate for stellar material, a major determinantof how dwarf stars behave. There is a closeconnection between the powerful tools of plasmaspectroscopy and the astronomer’s inference ofconditions in the violent inflowing spiral ofmaterial surrounding a neutron star or a blackhole that may lie at the center of a quasar.

This very hot material can be recreated in ahigh-energy-density laboratory experiment. Inother experiments we can study a fundamentalphysical process—mixing at hydrodynamicallyunstable material interfaces—that is importantin many areas of astrophysics, such as insupernova envelopes.

For the future, high-energy laser facilitieswill furnish excellent opportunities to conductadvanced fluid dynamics experiments. Theability to study fluids has been amply demon-strated with current experiments, and expandedpossibilities can be envisioned for the NIF.Because a high-energy laser has the capabilityto deposit a large quantity of energy inmaterials over large spatial scales, over longtimes, and at high energy densities, it willbe able to generate hydrodynamic flowconditions more extreme than those obtainedwith other types of machines. This is especiallytrue with the creation of very high-density,high-temperature volumes of material, whichmay be used to drive strong shocks and high-velocity flows.

The capabilities of the future high-energylasers for studying condensed matter physics atextreme conditions will make them the mostimportant new tools in high-pressure physics.For example, we will be able to experimentallyvalidate material behavior for pressuresbetween 1 and 100 TPa. The NIF will providethe combination of drive energy and uniformityof drive, with the concomitant large time andlength scales for real advances. The NIF willallow us to substantially improve our under-standing of dense fluids, pressure ionization,phase boundaries, and energy partition inshocked systems. The basic experimentalmethods for these studies have existed for sometime—the ability to exploit them at ultra-highpressures is still awaiting a NIF-class laser.

xii

The study of plasma physics has beenstimulated over the past four decades by itsclose connection with the goal of creating fusionas an energy source and with studying aspects ofastrophysical plasmas of various types. Theadvanced capability of the NIF will allow us toproduce hot, dense plasmas that are both largeand homogeneous, allowing detailedcharacterization. Thus, for example, we will beable to determine electron and ion temperature,charge state, electron density, and flowvelocities of plasmas. This information wouldallow a wide range of quantitative experimentsto be performed. One extremely interestingavenue of research would allow betterquantitative measurements on certain nuclearcross sections, and would be made possible by theability of the NIF to create large volumes ofheated plasma in a controlled manner.

In the area of x-ray lasers the NIF willallow us to extrapolate existing collisionalx-ray lasers down to wavelengths of ~20 Å. Atthese wavelengths x-ray laser interferometrycan be used to measure electron densities inplasmas exceeding solid density. Short-pulse

capabilities on the NIF, with pulse lengths inthe neighborhood of~100 fs, will allow us todevelop inner-shell pumped x-ray lasers as aviable source. This new class of x-ray laser willhave the potential to be extended down to aslow as 1 Å in wavelength and have a short pulseduration. Further, a NIF-class laser will makepossible a wide variety of x-ray and particlesources suitable for addressing numerous basicand applied physics questions. The NIF will becapable of producing intense broadband thermalx-rays from high-Z radiation converter targets,coherent amplified x-rays (x-ray lasers) fromhigh-gain linear plasmas, intense neutron pulsesfrom implosion plasmas, and intense pulses ofhard x-rays produced by fast electrons.

The contributions future high-energy laserscan make to science that deals with hot anddense matter will create a watershed forprogress in this complex regime so important inour universe. This is true whether the researchderives from the fields of astrophysics,material properties, plasma physics, or energygeneration.

Richard Lee, Livermore, CARichard Petrasso, Cambridge, MARoger W. Falcone, Berkeley, CAJanuary, 1995

High-Energy Lasers/Context of Current Science 1 Section I

Section I

High-Energy Lasers in the Context of Current Science

Interest in the development of a broad-basedscientific community for the use of high-energylasers can be placed in the wider context of thealready existing scientific user communities inthe other parts of the world. In the UnitedKingdom, France, Germany, and Japan, thereexist high-energy lasers, supporting diverseprograms of work based, to various degrees, onscientific user communities. For a combinationof reasons, the development of the same diversescientific user base has not occurred to any largedegree in the USA, although we have hadworld-class high-energy lasers since theinception of these facilities in the early 1970s.

First and foremost is the inescapable factthat the Inertial Confinement Fusion (ICF)program in the USA has had a substantialclassified component, and this has created animpediment to the free transmission ofinformation. Although there have beenunclassified components of the ICF program inthe USA, the residual effect of classification wasto stunt growth in the scientific community. Asmentioned above, this must be seen against aworldwide background where scientific-basedcommunities have flourished. Thus, a lack ofdevelopment here cannot easily be explained asbeing due to a lack of relevance of the high-energy laser to basic or applied science or itsincompatibility with scientific investigations.

Recent changes have modified theprevailing atmosphere and will, hopefully, bebeneficial to the development of a scientific usercommunity. The first of these changes is veryrecent (December of 1993), with thedeclassification of the concept of indirect-drive(i.e., radiation-drive) ICF, which was thedominant program. This declassification is a

result of a number of politico-economicadjustments, including the easing of the East-West tension and a commitment from theDepartment of Energy (DOE) to make atransition into this less-than-Cold-War era.Second, there is the active pursuit of a “dual-use” strategy, which in the present case cansimply be seen as a concerted effort by the DOEand other government agencies to have theircapabilities utilized more broadly. Third,important changes are now taking place in theICF program in response to this new climate.There is now a commitment to scientificapplications of high-energy lasers. Theseapplications set the stage for this document.

The changes that have been occurring in theICF program, the DOE, and in the USAgenerally, only lay the foundation for this questfor scientific participation; successfuldevelopment of a user community is notguaranteed by our commitment alone. Thisdocument represents an attempt to catalog themany possible uses for a high-energy laserfacility. Therefore, this effort is in support of thebroad concept of the development of a usercommunity, which has been addressed by theorganization of workshops and extendedindividual discussions.

The workshops and discussions haveproduced a number of intriguing possibleapplications, which will be found below. Themethod used to evaluate the many proposedtopics of interest was a straightforward analysisof the applicability of the suggested research tothe high-energy laser facilities of centralimportance here. We have relied extensively onthe many contacts mentioned above.

Section I 2 High-Energy Lasers/Context of Current Science


Overview of Science on High-Energy Lasers 3 Section II

Section II

Overview of Science on High-Energy Lasers

A. Astrophysics andSpace Physics

The reason for considering experiments onhigh-energy-density plasmas in connection withastrophysics and space physics is simply thatsuch plasmas are a dominant aspect of manyastronomical objects, whether in our solarsystem, in other stars, or in other galaxies. Theastrophysicist or space physicist is faced with theproblem of deciphering the few clues theobservations provide in order to learn about theobjects’ nature. The clues most often consist ofplasma radiation, and a full familiarity with whatplasmas actually radiate is needed to learn aboutthe astronomical objects. In other cases thetheoretical astrophysicist or space physicistbuilds up a complete conceptual structure of theobject, and then tests aspects of the structureagainst the observations. In this case thedynamics of plasmas form one of the basicbuilding blocks of the theory, but these dynamicsrequire observational support themselves.

The goal of astrophysically motivatedlaboratory experiments is to create in thelaboratory a sample plasma similar to whatwould be found in the astronomical body, thento study its properties—its radiation or itsthermodynamic state, whichever is critical inthe astrophysical picture. The properties thatare studied will be found under the other projectheadings of this report: Hydrodynamics, MaterialProperties, Radiative Properties, and so on.Astrophysics does not, in general, provide newplasma properties that need to be studied, but

rather gives a specific motivation for studyingcertain of the ones already known.

The discussion in Section IV will highlight afew instances in which there is a specificastrophysical motivation for studying a particularplasma. Among these is the radiative opacity ofthe outer envelopes of stars, key to ourunderstanding of how stars evolve. Another isthe equation of state of stellar material, which isa major determinant of the behavior of dwarfstars—the white dwarfs left over at the end ofstellar evolution and the brown dwarfs that arethe failed almost-stars.

There is a close connection between thepowerful tools of plasma spectroscopy and theastronomer’s inference of conditions in theviolent inflowing spiral of material surrounding aneutron star or a black hole that may lie at thecenter of a quasar. This very hot material can bere-created in a high-energy-density laboratoryexperiment. In other experiments we can study afundamental physical process—mixing athydrodynamically unstable material interfaces—that is important in many places in astrophysics,such as in supernova envelopes, and in terrestrialapplications as well.

Astronomy provides some of the drivinginterest in these and other laboratory experi-ments, and it also provides another laboratoryand an additional data set. Merging the data fromlaboratory experiments with astronomicalobservation will improve our knowledge of theuniverse, as well as our understanding of thebehavior of matter found here on earth. (SeeSection IV for a full discussion of astrophysicsand space physics on high-energy lasers.)

Section II 4 Overview of Science on High-Energy Lasers

B. Hydrodynamics

Interest up to the present in performinghydrodynamic experiments can be divided intotwo broad areas. First, there is the interest instable hydrodynamics. This includes the study ofpressures obtainable, shock characteristics, andother aspects of the hydrodynamics problem asthese affect various aspects of, for example, ICF.It includes looking at how these experimentspermit testing of hydrodynamics simulationcapabilities.

Secondly, there is a keen interest in unstablehydrodynamic flow. This interest includesunderstanding flow in the linear growth regimes,through the nonlinear regimes, to turbulent flow.The instabilities that have been tackled includeRayleigh-Taylor, Richtmyer-Meshkov, andKelvin-Helmholtz. These experiments haveattacked the unstable flow problem in a high-energy-density regime, and that in itself is aunique aspect of the work.

In addition, the dimensionality of the flowstudied can be changed in the experiments. Inpreliminary experiments, such as thoseillustrated in Section V, the desire is forsimplicity, but this is a goal and is not alwaysachieved. On the other hand, there areexperiments in converging spherical geometry,such as those required for ICF, as contrasted withmore idealized planar geometry experimentsperformed for basic information. Examples ofboth stable and unstable hydrodynamics inspherical and planar geometry are shown inSection V.

For the future, high-energy, laser-basedfacilities can furnish excellent opportunities toconduct advanced fluid dynamics experiments.This has been amply demonstrated with currentexperiments illustrated here, and expandedpossibilities for further investigations can beenvisioned for the NIF. The capability of a laserto deposit a large quantity of energy in materialsover large spatial scales, over long times, and athigh energy-densities means that it will be able togenerate hydrodynamic flow conditions moreextreme than those obtained with the morefamiliar machines such as wind tunnels or shocktubes. This is especially true concerning the

creation of very high-density, high-temperaturevolumes of material, which may be used to drivestrong shocks and high-velocity flows.

The extensive set of diagnostics available tostudy laser-produced flows facilitatesmeasurements of the quantities of interest,including interfacial geometries, velocities,densities, and/or temperatures. The NIF will beof great interest for the experimental study of thephysics of high Mach number flow and theinstabilities associated with the presence ofcertain material conditions and interfaces.Typically, large Mach numbers have beendifficult to attain in the laboratory, becauseachieving the high flow velocities requires a veryhigh-enthalpy source, such as very high-temperature, high-pressure gas, which isproblematic. The NIF, however, should excel inproducing a high-enthalpy source (i.e., ahohlraum). Using the hohlraum as the source,the NIF will be able to generate very strongshocks and large Mach number flows.

This possibility of testing conventional theoryunder extreme conditions opens up a largeperspective toward the understanding of variousastrophysical phenomena involving thepropagation of blast waves and the plasma flowsuch as those generated by supernova explosions.Compressibility of the materials studied and itsrelationship to the onset of turbulence is also animportant connection to be probed under theunprecedented conditions attainable with thesehigh-energy-density sources. Finally, because ofthe larger spatial scales, longer times, and higherenergies provided by the NIF, it will permit fluiddynamics investigations to be

• Larger, thereby increasing relative spatialresolution.

• Longer, thereby allowing more growthtime for the development of features.

• More energetic, thereby helping tomaintain drive conditions and reducedecompression effects.

Thus, it will permit the detailed study ofphenomena that to the present have been studiedexperimentally exclusively by astronomicalobservation, or only by computer theoreticalsimulation. (See Section V for a full discussion ofhydrodynamics on high-energy lasers.)


C. Material Properties

The interest in material properties at extremeconditions has developed from both fundamentaland applied interest. For, example, thechallenges of understanding the equation of state(EOS) at extreme conditions requires thedevelopment of a model that can treat stronglycoupled systems, with all the difficulties of hot,dense matter. On the other hand, we know thatICF and astrophysical studies require basic EOSdata in high-energy-density regimes. Theevolution of the study of material properties hascome to include, not only EOS studies performedin both directly and indirectly driven targets,but opacity measurements, and strength ofmaterials studies.

The capabilities of the NIF for studies ofcondensed matter physics at extreme conditionsof pressure and temperature will make it themost important new tool in high-pressurephysics. While we will describe in the followingsections vital contributions that only the NIF canmake to this field, we want to make it clear thatstudies in this field will make an essentialcontribution to other fields as well. A broad andprecise knowledge of condensed matterproperties for pressures between 1 and 100 TPa(between 10 and 1000 Mbar) and temperatures upto a few hundred electron volts will be essentialfor the design, simulation, and interpretation ofexperiments for other scientific applications ofthe NIF (such as hydrodynamics, planetaryphysics and astrophysics, and radiativeproperties) as well as for the ICF program. Forexample, we have yet to understand the behaviorof hydrodynamic instabilities in the pressurerange for which pressure ionization becomesimportant. We must have a “caloric” equation ofstate to interpret the entropy fluctuations and todetermine the appropriate invariants for high-Mach-number flows under these conditions.

Presently, no models of material behavior forpressures between 1 and 100 TPa have beenexperimentally validated. Impact and staticexperiments provide a wealth of data at lowpressures, below a few Mbar, and statisticalmechanical models such as Thomas-Fermitheories are believed to be valid at very high

pressures. The upper bound is approximate, andincreases with Z.

At tens of TPa (hundreds of Mbar), themodels that do exist only agree within about 30%in density, and vary by as much as a factor of twoin temperature along a shock adiabat (Hugoniot).These models differ quantitatively and infunctional form, and the discrepancies betweenthe models are functions of density, pressure, andtemperature. Although a few shock-waveexperiments have been carried out in thispressure range in the former Soviet Union(driven by a nuclear device), none have beensufficiently accurate to test theory. Over the lastdecade, a few laser-driven experiments haveprobed this region of pressure with evenless accuracy.

The NIF will provide the combination ofdrive energy, uniformity of drive, and theconcomitant large time and length scales for realadvances. This effort will substantially improveour understanding of dense fluids, pressureionization, phase boundaries (such as inproposed plasma phase transitions), and energypartition in shocked systems. The basicexperimental methods for these studies haveexisted for some time—the ability to exploit themat ultrahigh pressures is still awaiting a NIF-class laser.

We need to know material properties overwide ranges of density and pressure. Mostprevious experiments have been shock-waveexperiments on the principal Hugoniot. Thepulse-shaping capabilities of the NIF are essentialto reach off-Hugoniot states, especially near theisentrope. Cold compression, along the room-temperature isentrope, will be limited by theeventual insulator-metal transition in diamond,which will occur at about 1 TPa.

While experiments and theory haveemphasized shock properties in the past, theseare merely the simplest and most fundamentalmeasurements. There is yet to be anyexperimental data on transport properties atthese extreme conditions, such as thermal andelectrical conductivity, and we note that thehydrodynamics itself will depend on radiativetransport away from the shock front at thehighest pressures of interest.


Finally, we note that for the first time anintense source of neutrons will be available fromthe ignition process. This will allow us to heatalong isochores, or paths of constant density, totens of electron volts in many metals, allowingaccess to a region of state space previouslyunreachable in any experimental environment.(See Section VI for a full discussion of materialproperties on high-energy lasers.)

D. Plasma Physics

The study of plasma physics has been stimu-lated over the past four decades by its closeconnection with the goal of the creating fusionas an energy source and with astrophysicalplasmas of various types. That is, the luminousmatter in the universe is composed almostentirely of plasma, and this makes the scientificinvestigation of plasma of the utmost importance.Furthermore, the NIF will allow us to makewide-ranging and detailed studies of universalplasma conditions, an opportunity, to the best ofour knowledge, that will not be afforded on anyother facility.

The rather large number of experiments thathigh-energy lasers permit in the area of plasmaphysics indicates the close connection amongmany of the applications. A number of theinteresting phenomena that take place in plasmas(such as parametric instabilities) rely on a largedegree of homogeneity in plasmas that are bothhot and dense. Most plasma physics experimentsto date have been complicated by the largegradients and small scale of the plasmasproduced by current lasers.

The advanced capability of the NIF will allowus to produce hot, dense plasmas that are bothlarge and homogeneous, allowing detailedcharacterization. Thus, for example, it will bepossible to determine electron and iontemperature, charge state, electron density, andplasma flow velocities. This in turn wouldprovide a medium in which a wide range ofquantitative experiments can be performed.

Some of the laser-plasma interactionexperiments that will be possible on the NIF willbe developments of those already attempted onsmaller, less well-characterized plasmas.

However, it will be possible to go beyond suchexperiments, which were geared to the needs ofthe current ICF program, and investigate a widerange of scientifically important interactions.

There is general interest in having anadvanced high-energy laser facility, whichshould include a beamline for short-pulse (i.e.,pulse durations less than 1 ps) high-powerexperiments. This capability is especiallyimportant for many basic plasma physics studies.These include

• Relativistic, ultrahigh-intensity regimesof laser-matter interaction

• Ponderomotive effects• Relativistic self-focusing and

filamentation• Laser beam channeling• Intense harmonics generation• Ultrahigh B-field generation with

ancillary studies of the resultant physicalprocesses

• Strongly driven instability regimes• Generation and transport of high electron

fluxes in a plasma• High-gradient accelerator schemes• New x-ray lasers• High temporal resolution diagnostics

Many of these will be discussed below.Another new area of focus will be the study

of magnetic fields, as large as 100 MG, which aregenerated by temperature and density gradientspresent in specially produced plasmas. Faradayrotation of an optical probe beam and/or Zeemansplitting of spectral lines should make it possibleto investigate and quantify these magnetic fields.The generation of large currents of fast electrons,produced by SRS and Two-Plasmon decay inplasmas at 0.25 of the critical electron density,will be a related area of study. Coupling theseelectrons into a high-Z metal converter shouldallow the production of short-duration bursts of50- to 100-kV x-rays with numerous potentialradiographic applications. Further, the feasibilityof performing nuclear reaction experiments in thelarge, hot, well-characterized plasmas that will beproduced on the NIF has been demonstrated insmaller experiments at the Nova laser facility.

In general, these larger plasmas will translateinto longer interaction lengths, greater


homogeneity, higher temperatures, longer timescales, and reduced velocity gradients. Therefore,the NIF laser will make strong contributions tofundamental plasma physics. (See Section VII fora full discussion of plasma physics on high-energy lasers.)

E. Radiation Sources

The wavelength of current lasers occurs inapproximately the visible wavelengths. Theconversion of this visible wavelength into other(usually shorter) wavelength radiation has beenone of the dominant themes of high-energy laserexperiments. This generation of radiation hasmany practical applications beyond the intrinsictheoretical interest in production mechanisms, asabsorption sources, x-ray heating sources, x-raylasers, etc.

The development of these light-sourcecapabilities will be broken down into threegeneric areas. First, there is the effort to use thehigh-energy laser to make spectrally continuousor quasi-continuous sources for use as absorptionsources and as heating sources for the applicationof radiatively heating matter. Second is the workon the conversion of the high-energy laser intonarrow-band radiation sources. Here the effort isusually quantified by the conversion efficiencyinto a single-line transition or into a small, closelygrouped set of lines. Finally, there is the effort tofurther develop the x-ray laser, which work,of course, overlaps with the applications ofx-ray lasers.

A NIF-class laser will make possible a widevariety of x-ray and particle sources suitable foraddressing numerous basic and applied physicsquestions. The NIF will be capable of producingintense broadband thermal x-rays from high-Zradiation converter targets, coherent amplifiedx-rays (x-ray lasers) from high-gain linearplasmas, intense neutron pulses from implosionplasmas, and intense pulses of hard x-raysproduced by fast electrons. Accurate energyspectra and absolute measurements of theconversion of laser energy into all types ofradiation and particle fluxes will play animportant role in benchmarking our basic

understanding of laser plasma interactions andatomic physics.

Broadband x-rays generated from laser-produced plasmas can be used to produce andcharacterize large and uniform plasmas relevantto ICF and astrophysics. Previous experimentsusing this technique have yielded importantresults which can be extended to hightemperatures and densities. The hightemperatures and densities produced duringimplosion and subsequent ignition will be anunmatched source of continuum x-raysextending from the XUV to MeV with pulsedurations less than 100 ps.

X-ray lasers, besides being importantcoherent radiation sources, offer a critical test ofour atomic modeling. On the NIF, x-ray laserswill allow us to extrapolate existing neon-likeand nickel-like collisional x-ray lasers down towavelengths of ~20 Å. At these wavelengthsx-ray laser interferometry can be used to measureelectron densities in plasmas exceeding soliddensities. Short-pulse capabilities on the NIF,with pulse lengths ~100 fs, will allow us todevelop inner-shell pumped x-ray lasers as aviable source. This new class of x-ray laser willhave the potential to extend down to 1 Å inwavelength and have a short pulse duration.High-order harmonic generation, an alternativesource of coherent radiation, can be used toproduce tunable coherent XUV radiation.

Neutron production will exceed 1018 in single100-ps pulses, making the NIF a potentiallyuseful source for producing uniform high-densityand low-temperature plasmas (i.e., stronglycoupled plasmas). Fast electrons with hundredsof keV in energy will be generated by parametricinstabilities, providing another potential sourceof high-energy x-rays for plasma backlightingand probing.

These radiation sources are supplemented bythe possibility of using radiation enclosures, orhohlraums, for the generation of radiationenvironments and x-ray drive fluxes. Hohlraumswill be able to produce far in excess of the 200-eVequivalent radiation temperature sources nowavailable on the current high-energy laserfacilities. These sources will not only be of highereffective temperature but will also be able to


provide uniform x-ray drive over far larger areasthan can now be imagined on present-daysystems. Thus, the advantages of using x-rayheating for the study of hot, dense matter will begreatly enhanced with the advent of the NIF. (SeeSection VIII for a full discussion of radiationsources on high-energy lasers.)

F. Radiative Properties

The importance of the study of radiativeproperties in high-energy-density plasmasderives from three separate factors. The first isthat the radiative property can be the bestindicator of the level of scientific knowledge in aparticular area. So, for example, when one isinterested in the development of newdescriptions of atomic structure, the mostimportant measurements to be made are theaccurate determination of transition energies.Moreover, as with the case of spectral linebroadening, one can develop sophisticatedplasma theoretical constructions, kinetic theories,or scattering theories and find that the simplestplace to test the theory is through a prediction ofa line shape, which can then be measured.

Second, radiative properties provide theclassic example of the non-interfering probe.Thus one can obtain fundamental information onthe plasma by looking directly at the emission orthe absorption spectrum. Spectral formation in itsmost sophisticated forms can provideinformation on the plasma ionization balance, therate processes, the densities, the temperatures

and the fluctuation levels. The radiativeproperties are therefore a powerful diagnostic ofthe plasma state.

Third, although radiative properties providean important test bed for theoretical developmentand provide diagnostic information, they are alsonecessary as primary data for numerous otherstudies. For example, spectral line lists areinadequate for many of the charge states ofheavier elements. Therefore, the categorization ofthe energies of highly ionized species becomes anend in itself, as data tables are necessary for themyriad uses to which spectra are put. In this way,line shapes, line widths, and line shifts must bemeasured for use in radiation-hydrodynamicsimulations of entities as small as an ICFmicrosphere implosion and as large as a galaxy.

These factors together explain the interest inradiative properties of hot, dense matter. It isclear that the largest part of the community ofscientists outside the x-ray laser area that areinvolved in high-energy laser experiments areconcerned with measuring radiative properties. Itis a fertile area with much to be gained from theexisting high-energy laser facilities and evenmore to be gained with the advent of the NIF.

Some of the possible applications forexperiments on high-energy lasers are presentedin the Section on Radiative Properties; here wehave only broadly summarized a few of the ideasto provide an indication of the type ofexperiments being considered. (See Section IX fora full discussion of radiative properties to bederived from high-energy laser experiments.)

Experimental Capabilities 9 Section III

Section III

Experimental Capabilities

Before illustrating the experimentalpossibilities of high-energy lasers, we will discussthe definition of a high-energy laser facility.

We note that the repetition rate of theexperiments is relatively low when compared totabletop laboratory experimental facilities. On theother hand, the system has, for a high-energy-density production facility, a fairly rapidturnaround time. The inherent problem in usinga facility to obtain both high temperatures andhigh densities is that it tends to cause varioushigh levels of destruction, which increases thetime it takes to ready the facility for a new shot.For this reason, the high-energy laser has arelatively low duty cycle (i.e., one shot per hour).However, even when compared to the duty cycleof tabletop facilities (i.e., greater than one shotper hour), the high-energy lasers provide benefitsthat can help offset their comparatively lowduty cycle.

The most obvious factor in support of thehigh-energy laser for research work, in spite ofthe relatively low duty cycle, is the fact that thediagnostics complement that can be brought tobear on a single experiment is very large. Second,as a result of the diagnostics and the properties oflasers, is the fact that the high-energy lasers canperform experiments that cannot be performed atany other type of facility. Thus, the high-energylaser can be seen as a distinct breed of facility thatpermits novel regimes to be investigated, butwhich allows these investigations to occur inreal time.

A. Definition of Existing Facility

The following definition of an existing high-energy laser facility is based on the Nova laser

facility at LLNL. Although this is currently thelargest facility, it is generically similar to theother facilities currently extant. Note that thedefinition of the NIF laser, which is containedbelow in Subsection B), is being developed to alarge degree with the assistance of the LLNLLaser Science program, and most of the facilityinformation has been obtained from individualsassociated with that program.

The Laser

Laser BeamsThe Nova facility has ten separate beam lines,

all of which are roughly equivalent. The laser isbased on the master oscillator power amplifier(MOPA) Nd:YLF configuration, where the initialseed laser is generated in a desired temporallyshaped pulse. The amplification occurs in a longchain of amplifiers separated by spatial filteringpinholes in increasingly larger diameter sections.The 1-ω0 (1.053-µm) wavelength beam arrives atthe target chamber with approximately 10kilojoules of energy in a beam diameter of 74 cm.This assumes a pulse that is temporally a 1-ns-square pulse—that is, the pulse rises in 100 ps toa uniform intensity and persists for 1 ns until itfalls in 100 ps. The frequency of each beam isthen either doubled (using a KDP crystal) to a2-ω0 (green) beam, or tripled to a 3-ω0 (blue)beam, for final focusing down into the targetchamber. The conversion efficiency of thedoubling process is approximately 60%, yielding~4000 joules of energy in the beam to be focusedinto the target chamber. The ten beams arebalanced to produce ≤5% rms deviationin energy.

The 1-ω0 beam is of high spatial uniformity;however, the nonlinearity of the up conversion to

Section III 10 Experimental Capabilities

2 ω0 or 3 ω0 amplifies any amplitude modulationin the beam. Several techniques can be used tosmooth the beam, a major consideration fordirectly irradiating a surface where one desiresuniformity. First, a random phase plate can beused to introduce a randomness in the phases ofthe beam. This creates an effectively smootherintensity distribution by overlapping many smallbeamlets of random phase created from the large-aperture beam. Second, the beam can be split upinto several pieces using wedges that allow aneffectively flat intensity distribution over thefocal spot. Third, it is possible to introduce abandwidth on the normally monochromaticbeam; when combined with a diffraction grating,this will be used to move the focal spot, resultingin a temporally smoother intensity.

The limiting factor in the firing of the beam isthe cooling-down time for the large diskamplifiers. This limits a particular laser beam linefrom firing more than once every hour. However,the complexity of the diagnostic setup andalignment of the target can be, and often is, thelimiting factor in the turnaround time ofthe experiments.

Beam Pointing, Aiming, and SynchronicityAs the beam enters the target chamber, it is

focused into the center using an f/4.3 lens thatwill produce a minimum spot size of 150 µm indiameter. The other option at the present time forthe beam is the insertion of f/8 lenses. This is arather new option.

The accuracy of the focusing, or pointing, is±30 µm rms, with a maximum displacement ofthe beam centerline from target chamber centerof 2 cm. The beams can be sent to three differenttarget areas: a one-beam chamber, a two-beamchamber with the beams coming in opposingports, and a ten-beam chamber. In the ten-beamchamber the beams come into the chamber in twoclusters of five beams, with the clusters enteringon opposite sides of the midplane. The fivebeamlines in a cluster are placed in a five-sidedsymmetry, and come into the chamber in a conewith a 50° half angle. The synchronicity is ±20 psat target chamber center.

Laser Pulse ShapesPulse durations can vary from 100- to 600-ps

Gaussian pulses to 1- to 5-ns-square pulses. Thereis also flexible pulse-shaping capability, so thatthe energy can be delivered in 2- to 3-ns shapedpulses with intensity contrast between peak andfoot of the pulse greater than ten. Two of thebeams can be delayed, independently of eachother, by up to 25 ns after the main beam. Thislast capability has been found useful forproviding a source of backlighter x-rays.

Laser Shot RateThe laser currently fires at a rate of

approximately 115 shots per month. This isconsistent with the fact that the laser fires aboutseven shots per day utilizing a two-shift system.The week has four scheduled experimental daysand one day maintenance.

Beam DiagnosticsThere are several diagnostics of beam

performance available. First, the pulse shape ismonitored in the master oscillator for temporalsignature and energy. Next, the propagated beamcan be monitored for the total 1-ω0 energy andthe temporal variation of the 1-ω0 energy.Further, images can be made of the near and farfield of the 1-ω0 beam.

The up-converted beam, in either 2 ω0 or3 ω0, can be monitored in the chamber forabsolute intensity as well as temporal intensity,and images of the beam front can be produced inthe far or near field.

Diagnostic Capabilities

In contrast to smaller facilities, thediagnostics on Nova are engineered to beroutinely run on any shot by facility staff. Makingthis possible involved applying professionalengineering standards to the diagnostics. Thishas avoided the more usual operating procedureof several Ph.D.-level scientists working for manyweeks to get their principal diagnostic installedfor their series of experiments, and after anintense campaign of experiments removingtheir diagnostic.


In the long run, this is an efficient use ofresources. Also it is necessary for efficient facilityscheduling, as the user will run in the mode ofdoing different sets of experiments in any oneweek. This mode of scheduling precludes theexclusive use of a target chamber for blocks of aweek or more, but has the advantage of givingexperimentalists time to analyze a set of databefore proceeding with an experimental series.

To achieve a high standard of engineering forthe diagnostics, experimentalists work withengineers and technicians for many months todesign, build, and implement the diagnostics,with the goal of implementing diagnostics thatcan be routinely run by facility staff for any user.

The chamber can be quite large—for Nova itis 2.4 m in radius. However, the chamber willpump to vacuum in less than one hour. It is notroutinely vented during the four- or five-dayshooting week. For shots where the targetgeometry does not change, this allows shots to befired without realignment of the laser beams. Asa result, diagnostics are designed to be operatedwithout venting the target chamber.

For flexibility, all of the x-ray imagers andx-ray spectrometers that operate close to thetarget have been made to be interchanged usingone of six standardized vacuum-load lock andmanipulator devices, known as SIMs (for six-inchmanipulators). These manipulators provide astandard for several facilities. In excess of twentydiagnostics are now mounted on the carts that fitinto the SIMs. This standardization allows keydiagnostics to be moved from one location on thetarget chamber to another location over a periodof about 1 hour, so that the configuration oftarget diagnostics can be rapidly changed fordifferent experimental campaigns. Out of the 30or so shots that are fired most weeks, diagnosticsare typically moved every 2–4 shots, the numberof consecutive shots that are fired in anyone campaign.

Overview of Target Chamber DiagnosticsTable III-1 lists the diagnostics available on

Nova. The list includes the dozen different x-rayimaging diagnostics, x-ray spectrometers,neutron diagnostics, and optical diagnostics.

The x-ray imagers are reflective grazingincidence x-ray microscopes, pinholes or coded

apertures, or point-projection systems. Because itis important to have time resolution, the detectorsfor the imagers are x-ray streak cameras, gatedmicrochannel plate (MCP) detectors, or simplyx-ray film with flash backlighting.

The x-ray spectrometers available are crystalspectrometers (Bragg or Laue), broad-band diodearrays, or grazing incidence grating spectro-meters. Again, the detectors are x-ray streakcameras, gated MCP detectors, or x-ray film.

The neutron diagnostics are yield (byactivation), high- and low-resolution energyspectra, emission-time diagnostics, andneutron imaging.

The optical diagnostics are streaked opticalspectrometers, used mainly for stimulatedBrillouin and Raman scattering in either onelocation or in an array of different angularlocations; optical and UV imagers that either goonto streak cameras or gated optical imagers; andarrays of calorimeter modules.

Note that in Table III-1 we have listed theabbreviated names of the diagnostics, as theseabbreviations are used in many of the figures.The various uses of the diagnostics have beencovered in the section above.

X-ray ImagersNova has had for several years reflective

grazing incidence x-ray microscopes, two of theKirkpatrick-Baez design and one of the Woltherdesign. The detectors for one of the four-channelKirkpatrick-Baez x-ray microscopes are film. Theother microscope has four time-gated x-raydetectors using gated MCP (microchannel-plate)technology.1

The Wolther microscope has a spatialresolution of approximately eight microns,although in one direction there is a large scattercaused by the surface roughness of the x-rayoptic. This instrument has been widely used onNova.2 The collecting solid angle of the optic islarge enough that only a small sector of theannular imaging region need be used to form animage. Five separate 5° sectors of the annularoptic are used to make it into a five-channelsystem, with an x-ray streak camera on onechannel and gated MCP detectors on each of theother four channels.3 This instrument has beenwidely used on x-ray backlighting experiments


Table III-1. Diagnostics available on Nova, with their abbreviated names. These include x-rayimagers and spectrometers, neutron diagnostics, and optical imagers and spectrometers.

Name Abbrev.

X-ray imagers

Wolther x-ray microscope 22X 1

Gated x-ray pinhole camera GXI, WAX,GACS, GSIX

4

Gated soft x-ray framing camera SXRFC 1

Streaked soft x-ray imager NSDSS 1

Ring-aperture microscope RAM 1

Streaked slit/array imager SSC/SMP 2

Kirkpatrick-Baez microscopes 8X 2

Axial pinhole cameras APH 2

Large-area backlighting

Point-projection spectroscopy PPS 2

Soft x-ray microscope 1

XRL beam divergence camera Cube 1

XRL spatial-coherence diagnostic 1

Neutron diagnostics

Yield Cu, In scin 3

Bang time NETMCP,GaAs

3

Burnwidth NTD & GaAs 2

High-resolution high-sensitivityspectrometer

LaNSA 1

Medium-resolution neutronspectrometer

NTOF 3

Ultrahigh-resolution spectrometer fNTOF 1

High-resolution high-sensitivityspectroscopy

LANL Ti 1

Neutron imager NPAM

Name Abbrev.

X-ray spectrometers

Streaked crystal spectrometers NSCS,Keanetech

2

High-resolution streakedspectrometer

HICKS 1

Static crystal spectrometers Henway,POS

6

High-resolution crystal spectrometer HOPS 1

Gated crystal spectrometer TOPS 1

Gated imaging XUV spectrometer IXUVS

Laue spectrometer HETS 1

Low-resolution x-ray diode array Dante 2

Low-resolution high-energyfluorescers

FFLEX 2

Spatial coherence diagnostic 1

Grazing-incidence spectrometer COFFIN 1

High-resolution spectrometer HIRES 1

Time-resolved soft x-rayspectrometer

SFFD 1

Gated grazing-incidencespectrometer

McPIGS 1

Optical spectrometers and imagers

Streaked/gated imager SOP 2

Streaked optical spectrometers SOS, BSS 4

Multiple streaked spectroscopy MATRES 1

Spatially discr. streaked opticalspectroscopy

SDOSS 1

Calorimeter array EBM 1

Full-beam backscatter SOS5 1

on Nova where accuracy, ease of alignment, andsensitivity are important. Note that the solidangle/channel of the Wolther microscope is3 × 10–6 sr compared to about 10–7 sr for aKirkpatrick-Baez microscope.

For higher x-ray energy imaging, simplersystems are also being used on Nova. Crucialinformation on symmetry and the mix of pusher

material into a fuel is obtained from imagingimplosion cores at higher photon energy andbetter spatial resolution. The higher energyimplies high sensitivity, and the requirement ofgood spatial resolution implies good timeresolution to freeze plasma motion.

To tackle this problem, gated pinhole andcoded aperture systems have been used.


To ensure adequate sensitivity, the 5- to 10-µmpinholes of the instruments are brought as closeas 5 cm from the target. Because the instrumentsare used to record signals from ≥3 keV, relativelyrobust x-ray filters can be used, allowing thepinholes to survive for a few shots the targetshrapnel from the 20–40 kJ on target.

Time resolution is achieved by gating0.5-mm-thick coated MCPs. At present, severalcoating configurations are routinely used:a four-channel open-ended strip configuration, a16-channel serpentine arrangement, and 16channels onto four parallel strips. The motionalblurring is given by

δ δx = v tand so the 80-ps gate time that is easily achievedis adequate for 10-µm resolution with velocitiesof approximately 107 cm/s. For improvedresolution, 30–40-ps gate times are becomingavailable with thinner MCPs.

X-ray backlighting has been in use for laser-produced plasmas for several years, but forimplosions and in-flight measurements of higher-density planar foils, higher-energy backlighting isrequired. Several developments have madehigher-energy (~5–7 keV) streaked x-raybacklighting possible. The most important are:

• Random phase plates are now used onthe backlighting beams to produce asmooth, virtually monochromatic sourceof x-rays by irradiating, for example, aniron disc at ~1015 W/cm2.

• A slit imager onto a close-in x-ray streakcamera with the photo cathode about50 cm from the plasma allows foradequate sensitivity for recording thisbacklighting source.

For soft x-ray studies, two gated and streakedsoft x-ray imaging instruments are in use. Aschematic of the gated soft x-ray framing camerais shown in Fig. VI-10 (see Section V,Hydrodynamics). For two soft channels, acombination of grazing incidence mirrors andfilters defines two bands of soft x-rays. The thirdand highest energy channel has its passbanddefined by a filter. Images in x-ray bands at500 eV, 1 keV and 2.5–3 keV are recorded with100-ps gate time at four different times, givingtwelve images in total. This instrument allows the

spatial evolution of the soft x-ray emitting regionof a laser plasma to be accurately measured.

A major diagnostic technique for obtainingspectral and spatial information simultaneouslyis point projection spectroscopy.4 In thistechnique, one or two of the backlighting beamsare focused onto one or two 15- to 20-µm coatedfibers, to produce a flash of quasi-continuousx-rays from a point source. This “point” projectsa shadow of the object after the x-rays have beenreflected off a Bragg crystal, forming an image.Because reflection only occurs when Bragg’s lawis satisfied, the shadow has spectral resolution aswell as spatial resolution.

The x-ray laser researchers are characterizingthe coherence, both spatial and temporal, of x-raylasers, and beginning to demonstrate biologicalapplications. The spatial coherence of laboratoryx-ray lasers has been measured using a diagnosticbased on partially coherent diffraction from auniformly redundant transmission slit array.5

X-ray SpectrometersIn the crystal regime, there are several static

survey instruments, called Henways, that useconvex crystals at a distance of about 1 meterfrom the plasma for survey work from 1.5 keV toabout 10 keV.

There are also streaked crystal spectro-meters.6 On the ten-beam target chamber, theseuse flat crystals on the front of a close-in x-raystreak camera at anywhere from 3 to 6 degreesBragg angle, or at a much larger Bragg angle forhigh resolution. For implosion work there hasbeen an emphasis on high-energy, high-resolution spectroscopy because of the need forthe line emission to penetrate the compressedpushers. Argon dopants have been successfullyused to measure density from line broadening,but for future implosions, higher energy x-rayemission with higher spectral resolution will berequired. With a large Bragg angle and a PETcrystal, resolving powers of about 2000 have beenachieved in streak mode at photon energies ofabout 6 keV.

For accurate absolute measurements ofbroad-band spectra, an array of fifteen calibratedx-ray diodes, called Dante, is used with filtersand grazing incidence mirrors to isolate channelsbetween 100 eV and >1.5 keV.7 This was one of


the core diagnostic systems of Nova, andalthough originally designed for 1-ns recording,the photocurrent is now commonly recordedwith fast oscilloscopes (such as the TektronixSCD 5000), giving a time resolution FWHM of135 ps with some small ringing at 300–400 ps.This allows time-resolved measurements of thex-ray conversion efficiency to be routinely made.

There are several soft x-ray instrumentsenumerated in Table III-1. A particularly high-resolution grazing-incidence spectrometer,HIRES, can achieve a spectral resolution of 8,000–26,000 in the 100- to 200-Å range when used withan x-ray streak camera, and up to 35,000 whenused with soft x-ray film. This instrument wasrecently used to measure the line width of thedominant laser transition in neon-like seleniumx-ray laser plasmas, and to observe the scaling ofthe line width with amplifier gain-lengthproduct.8

The Streaked Flat Field Spectrometer (SFFS)is a time-resolved XUV spectrometer that utilizesa Harada varied line space grating.9 The gratinghas a flat focal plane that lies almostperpendicular to the x-ray propagation direction,allowing the use of a transmission photocathodex-ray streak camera. The instrument is sensitivefrom about 400 Å (30 eV) to ~10 Å (1000 eV).With a 1200-line-per-mm grating with highefficiency in higher orders, up to 6th order hasbeen seen at 44.83 Å. In addition, the instrumentis used to view short-wavelength x-rays, such as44.83-Å x-ray laser emissions, in multiple orders,extending the dynamic range of the spectrometerand allowing a resolution of λ/∆λ ~ 500 to beachieved in some measurements (e.g., 44.83 Å in5th order).

For shorter-wavelength spectra, the SFFS canbe used with a 2400-line-per-mm grating. Bothgratings are installed in the spectrometer andeither can be selected by translating them intoposition. The wavelength range is determined byan iridium-coated mirror, midway between thedetector and the grating and almost parallel tothe grating surface. The mirror relays thespectrum from the grating to the photocathode.Unlike similar instruments that rely on movingthe detector to vary the wavelength coverage, theSFFS keeps the detector fixed and moves the

relay mirror. This feature allows the spectrometerto be re-entrant, increasing sensitivity as thespectrometer moves closer to the targets.

Like all the streaked diagnostics on the two-beam target chamber (Keanetech, XCSS, XRLpower meter, XRL coherence measurement,SOS-3, SOS-5, PCDs, Dante, HSI, High Flux,UVImager, High Res, 2-Beam Incident Streak,Laser Diagnostic Station), the SFFS has a fiber-optic timing fiducial that allows the recordedspectra to be referenced to the incomingirradiation and other diagnostics to a precision of~30 ps. The timing fiducial also allows for thetesting and timing of streak cameras withoutfiring the entire laser chain.

The Microchannel Plate Intensified GrazingIncidence Spectrograph (MCPIGS) is a grazing-incidence XUV spectrometer with a gatedmicrochannel plate detector. The instrument is aRowland circle geometry spectrometer with a 2°angle of incidence on the 998.8-mm radiusgrating. The detector is a 150-mm-long curvedmicrochannel plate that lies along the Rowlandcircle. It has several configurations of strips thatare gated off at intervals using a silicon Austonswitch triggered by the fiducial laser beam. Withseven gratings available, a broad range of XUVspectra can be surveyed with this instrument.There are two MCPIGS spectrometers on the two-beam chamber. One lies in the equatorial plane toview x-ray laser output emitted in a narrow beamperpendicular to the plane of the Nova foci. Theother MCPIGS spectrometer views the plasmafrom an off-axis direction.

Optical DiagnosticsThe optical and UV diagnostics are also listed

in Table III-1. One of the most ubiquitousdiagnostics is the streaked optical imager, calledthe SOP (for streaked optical pyrometer). Thisinstrument is an f/10 Cassegrain telescope with amagnification of 8.47, and a resolution that isclose to diffraction-limited. The instrument ismost commonly used to record shock breakoutfrom witness plates, for which purpose the imageis recorded onto a UV-sensitive streak camera.Alternatively, the image can now be recordedonto a gated optical imager.

There are several streaked opticalspectrometers at fixed locations on the two target


chambers (SOS1, SOS2, SOS3, SOS4 and SOS5).These are primarily used to measure thestimulated Brillouin and Raman spectra to derivedensity and laser-induced plasma turbulencelevels.10 To make it possible to map out theangular dependence of the scattered light, amultichannel streaked spectrometer that allowsthe angular position of the detectors to be easilyvaried has been implemented on the Nova ten-beam target chamber.

For measurements of the absorption of laserlight there is a set of energy balance modules foruse on both the ten-beam and the two-beamtarget chambers.

Hohlraum Radiation SourcesAs an overview of the capabilities of a laser

system of this sort we provide the following.First, a unique feature of the system is thehohlraum. A hohlraum is, in the jargon of theInertial Confinement Fusion (ICF) program, ahigh-Z—usually gold—cylinder with end caps.The laser enters the cylinder through the endsand is focused on the interior walls in an annularpattern. The laser light is converted to x-rays bylaser–matter plasma creation. A typical hohlraumis 2400 µm long and has an 800-µm radius. Theradiation temperature equivalent, found bymeasuring the emission from the hohlraum andequating it to σT4, is ~200 eV.

The temperature and the size and scale of thehohlraums can, of course, be changed toaccommodate experimental needs. See Fig. III-1

TargetLaser beams~ 20 kJ at 3530 Å

Au cylindrical hohlraum - scale 1

Laser pulse shapes Targets

Observation hole

2400 µm

• Up to 28 kJ in 10 beams• Square: 1 ns, flat top, 100-ps rise• Shaped: 3 to 1 contrast, 2.2 ns

• Gas-filled microspheres• Tamped opacity samples• Cylinders for convergent hydro

1600

µm

Figure III-1. Schematic view of a hohlraum.The one shown here is the nominal hohlraum.Called a scale-1, it generates a radiationtemperature of ~200 eV.

for an example of a hohlraum. The design of ahohlraum can affect the radiation temperatureand conditions of the targets placed inside.

In Fig. III-2, three different types ofhohlraums are shown, each of which has adifferent utility. Note that the radiationtemperature is changed by the hohlraum scale.Figure III-3 shows the scaling of the measuredradiation temperature vs incident laser energies.In the figure there is also a simple fit to the datausing empirically determined parameters, theconversion efficiency of the laser energy to x-rayenergy, η, and the amount of loss of laser energydue to heating of the walls, α ; these aredetermined to be 0.7 and 0.8 respectively for the

4800

µm

Scale 3

Scale 1Scale 0

7500 µm 1/3 scale 3 500 × 500 µm

Tr ~60 eV Tr ~200 eV Tr ~330 eV

Figure III-2. Schematic view of several different hohlraums. The different types are used to create aradiation environment for various purposes. Numerous configurations exist for different tasks.


PL (TW)

300

280

260

240

220

200

180

403020100

Data

η P l = ( [1– α w + A h) σT4] A

T r (

eV)

Figure III-3. Measured radiation temperature vs input laser power. Temperature is derived fromequating the measured energy output to σT4. The data points are indicated, along with a dashed lineshowing a simple fit to the data using empirically determined parameters.

scale-1 geometry. The two areas that play a roleare the area of the walls, Aw, and the area of theholes in the wall, Ah. The formula is given as

η α σP A A TL w h= −[ ] +( )1 4 .

The ten beams of the laser typically deliver20–30 kJ of 0.35-µm light into a cylindrical goldhohlraum that is 2.5 mm long and has a diameterof 1.6 mm, producing a uniform quasi-thermalsource of x-rays. X-ray drive produces highablation pressures that are spatially very smooth.The radiation can be used to drive capsuleimplosions, drive planar shocks in materialsamples for hydrodynamics, provide a localthermodynamic equilibrium (LTE) environmentfor opacity studies, and provide drive forequation-of-state studies. The shock pressurecreated by radiation drive is a strong function ofthe temperature:

P Mbar T keV( ) = × ( )2 3 104 3 5. ..

For example, shocks of approximately 100 Mbarcan be driven in plastic with 200-eV drive.

The laser is also used for direct illuminationof targets. The direct illumination of low-densitytargets, for example foams or gases, can produceplasmas with an electron temperature of a few

keV. Ablation pressures generated in directlydriven targets are:

P(Mbar I Wcm m) / .– /

= ( ) ( )[ ]12 1014 2 2 3λ µ

Thus, for laser light with a wavelength ofλ = 0.53 mm (frequency-doubled light) and anintensity of I = 2 × 1014 W/cm–2, the ablationpressure generated is about 10 Mbar. Directlydriven shocks can be made reasonably planarover about 1 mm by smoothing the beam profileswith random phase plates and SSD.

To see the radiation environment of ahohlraum quantified, refer to Section VIII,Radiation Sources.

B. NIF Facility Definitions

The following definition of a proposed high-energy laser facility is based on ongoing work atLLNL. This facility will be capable of producing1.8 MJ of temporally shaped laser energy on avariety of targets. This laser system is beingdeveloped by the LLNL ICF program, whichprovided much of this information.


Laser Beams

The laser will have four separate beam-linebundles, each with a 4 x 42 array of beamlets, fora total of 192 beams. The laser is based on themultipass amplifier Nd:YLF configuration wherethe initial seed laser is generated in a desiredtemporally shaped pulse. Amplification occurs ina series of multipass and single-pass amplifiersseparated by spatial filtering pinholes inincreasingly larger diameter sections. The 1-ω0(i.e., 1.053-µm) wavelength beam arrives at thetarget chamber with approximately 10 kJ ofenergy in a beam 40 cm square; this assumes apulse that is temporally a 5-ns square pulse. Avariety of other pulse shapes will be available atother energies.

The beam is then frequency doubled using aKDP crystal to a 2-ω0 (green beam) and tripled, toa 3-ω0 (blue beam), for final focusing onto thetarget. The crystal conversion efficiency will yieldapproximately 10 kJ of 3-ω0 energy in the beam tobe focused into the target chamber. The 192beams will be balanced to produce no more than8% rms deviation in power.

The 1-ω0 beam will be of high spatialuniformity; however, the nonlinearity of the upconversion to 2 ω0 or 3 ω0 amplifies anyamplitude modulation in the beam. Severaltechniques can be used to smooth the beam, amajor consideration for directly irradiating asurface where uniformity is desired. First, arandom phase plate can be used to introduce arandomness in the phases of the beam. Thiscreates an effectively smoother intensitydistribution by overlapping many small beamletsof random phase created from the large aperturebeam. Second, the beam can be split up intoseveral pieces using wedges which allow aneffectively flat intensity distribution over thefocal spot. Third, a bandwidth can be introducedon the normally monochromatic beam; whencombined with a diffraction grating, this will beused to move the focal spot, resulting in atemporally smoother intensity.

On the NIF, temporal smoothing will also beachieved by overlapping four separatewavelengths, separated by 5–10 Å, andsmoothing each one by spectral dispersion(adding bandwidth).

The limiting factor in the firing of the beam isthe cooling-down time for the large diskamplifiers. This limits a particular laser beamlinefrom firing more than once every eight hours.Diagnostic setup and target alignment will beperformed within this time.

Beam Pointing, Aiming, andSynchronicity

The beam as it enters the chamber is focusedinto the target chamber center using an f/8 lensthat will produce a minimum spot size of 500 µmin diameter.

The accuracy of the focusing, or pointing,will be ±50 µm rms, with a maximumdisplacement of the beam centerline from thetarget chamber center of 7 cm. The beams comeinto the chamber in 48 clusters of four beams,with half entering on each of opposite sides of themidplane. Multiple laser beams will enter oneach side of the hohlraum along two concentriccones, with cone half-angles of approximately 27°and 53° , and with two-thirds of the beams on theouter cone and the remaining one-third on theinner cone. The pulse shape can be different onthe two cones.

Laser Pulse Shapes

There will be a flexible pulse-shapingcapability, so that the energy can be delivered inup to a 220-ns shaped pulse with >50:1 intensitycontrast between peak and foot of the pulse.Approximately 20% of the beams can be delayed,independently of each other, by up to 30 ns afterthe main beam. This last capability will be usefulfor providing a source of backlighter x-rays.

Laser Shot Rate

The NIF will have a capability of 3 shots/24hours for 100 shots/year with energy of 100 kJ,35/year for energy of MJ, and 10/year for energyof 20 MJ.

Beam Diagnostics

There will be several diagnostics of beamperformance available. First, the pulse shape willbe monitored in the master oscillator fortemporal signature and energy. Next, the


propagated beam can be monitored for the total1-ω0 energy and the temporal variation of the1-ω0 energy. Further, images can be made of thenear and far field of the 1-ω0 beam.

The up-converted beam, in either 2 ω0 or3 ω0, can be monitored in the chamber forabsolute as well as temporal intensity, andimages of the beam front can be produced in thefar or near field.

C. NIF Facility Modifications

A central issue that was addressed by thescientific community was the desiredmodifications to the original NIF design. To alarge degree the requests of the scientificcommunity were very similar to those of othercommunities. In particular, the group desiredsmaller spot sizes and pulse variation from 100 psor less to 20 ns. In addition, they identified asimportant the need to establish line foci toperform x-ray laser experiments and the need tohave an expanded subpicosecond capability.However, the most important request by far wasthe many demands for improved turnaroundtimes. This took many forms, but the messagewas the same—a science-based university-intensive effort that seeks to train studentsrequires more than one to three experiments aday. The suggestions to improve the turnaroundincluded:

• Additional target areas.• More rapid cooling of the amplifiers.• More independence of the beams, so that

a portion of the beams could be usedflexibly.

• The introduction of a smaller (e.g., Nova-sized) beam for independentbacklighting.

Each of the special requirements that thescientific community asked for is presentedbelow with a brief discussion of the requirement.

Higher Beam Irradiance Requirements

Some physics experiments will require higherbeam irradiances than those outlined in the NIFdesign documents. The present laser spotdiameter requirement for ignition targets,

utilizing kinoform phase plates for beamconditioning, is 500 µm. This will be achievedwith adaptive optics with a number of actuatorsconsistent with this requirement. The NIFFunctional Requirements and Primary Criteriaspecify the maximum spot size to be 500 µm. Forhigher irradiances that will lead to increasedhohlraum temperatures and the subsequentpressures desirable for a subset of the scienceexperiments, smaller spot sizes are needed—50 to150 µm, or as small as possible and practical (seethe discussion in Section VII, Plasma Physics).Higher irradiances than for the present designmay be obtained by increasing the number ofactuators in the adaptive optic. To facilitate point-projection x-ray backlighting, beams focused tothe minimum spot size should be able toeffectively irradiate a 25-µm-diameter fiberor equivalent.

Variations inLens-Focusing Requirements

The NIF should have the capability of puttingcylindrical lenses on at least 20 beamlets forexperiments utilizing x-ray lasers. These lensesshould nominally produce a focus of~150 µm × 5 cm. These lenses should be capableof being easily inserted and removed. Thebeamlets should also be able to operate at 2ω atpower levels consistent with those required forx-ray lasers.

Target Chamber Insertion Mechanisms

The target chamber should have two targetinsertion mechanisms. The second target inserterdoes not have to have cryogenic capability.

Added Target Alignment Capabilities

The scientific community envisions the use oftargets that may be, in general, more complicatedthan ICF targets, so additional target alignmentcapabilities may be needed. This is most simplyspecified by requiring the capability to view allsides of the targets(s) during target alignment.

Reduced Time Between Experiments

To accommodate the expected large numberof users, the time between shots should be


reduced below the maximum specified eighthours. The facility should incorporate theimprovements that are developed from theseprograms to ensure greater access to theuser communities.

Flexible Time Intervals BetweenExperiments

Finally, it is important to note that the largestpossible timing and pulse duration flexibilityneeds to be incorporated for the NIF to reach fullutilization. Individual beamlet pulse-lengthflexibility, as presently designed, is highlydesirable, with range-of-pulse widths from<0.1 ns to >20 ns. NIF design should permit eachcluster of four beams to be independently timed,with arbitrary time delays among the clusters.The time interval between the first and last beamsshould be >200 ns. If possible, delays of up to4000 ns would be useful to some users.

Short-Pulse Beams

There was a decided interest in having theNIF contain as many short-pulse beams aspossible. Because of its size, the projected singleshort-pulse beam was considered interestingenough to warrant involvement by thecommunity. However, scientists interested in theshort-pulse experiments felt that a single beamwas far too limiting. As can be witnessed from

the discussions of short-pulse experiments inboth Section VII, Plasma Physics, and SectionVIII, Radiation Sources, there are a large numberof topics in this category, and the possibility thatone or more may become of primary interest inthe future is thought by the participants to berather large. Hence, additional short-pulsebeamlines would be strongly supported.

D. References

1. J. D. Kilkenny, Laser and Particle Beams 9, 49(1991).

2. J. D. Kilkenny, Phys. Fluids B 2, 1400 (1990);B. A. Remington, et al., Phys. Rev Lett. 67,3529 (1991).

3. R. J. Ellis et al., Rev. Sci. Instr. 61, 2759 (1990).4. J. Balmer et al., Phys Rev. A 40, 330 (1989);

T. S. Perry et al., Phys. Rev. Lett. 67, 3784(1991).

5. J. E. Trebes et al., Phys. Rev. Lett. 68, 588(1992).

6. B. A. Hammel et al., Rev. Sci. Instrum. 61, 2774(1990).

7. H. N. Kornblum et al., Rev. Sci. Instrum. 57,2179 (1986).

8. A. Koch et al., Phys. Rev. Lett. (1992).9. T. Kita et al., Appl. Opt. 22, 512 (1983).10. R. E. Turner et al., Phys. Rev. Lett. 57, 1725

(1986).



Astrophysics and Space Physics 21 Section IV

Section IV

Astrophysics and Space Physics

The level of feasibility of possible high-energy laser-based experiments discussed hereranges from a straightforward extrapolation ofwell-established Nova techniques, to ratherspeculative possibilities that will probablyrequire technology beyond the presentlyproposed NIF capabilities before useful resultscan be obtained. Nevertheless, we discuss in thissection the important scientific issues inapproximately descending order of feasibility ofapplication on the NIF facility.

A. Opacity

The radiative opacity of the material in stellarinteriors plays a key role in determining howstars evolve—what the maximum mass of astable star is, how hot and how luminous the staris while it burns its hydrogen fuel, what pulsa-tional instabilities it may fall prey to, whether itloses much of its outer envelope in outbursts ofone kind or another, and during which stage ofits evolution it might explode as a supernova.The star’s energy budget dictates these otherconsequences, and because the radiative flux isinversely proportional to it, opacity sets the flowof heat through the star. In particular, theinstability of the star is especially sensitive tonuances in opacity, and is most sensitive in thetemperature range of a few hundred thousanddegrees Kelvin. Opacity is very complex in thisregion, and has stood as a formidable challengeto theoretical atomic physics. A brief discussionof recent advances is presented in Section VI,Material Properties, Part B.

Some of the regimes of temperature anddensity for which a star’s structure and evolution

are very sensitive to opacity are reachable in alaboratory high-energy-density facility such asthe NIF, as illustrated in Fig. IV-1—temperaturesfrom tens of electron volts to several hundredelectron volts, and densities from normal soliddensity to a few hundred times less.

The figure shows two regimes possible on theNIF, one with the laser only (shown in light gray)and the other using ignition within a deuterium-tritium capsule (shown in black). These arecompared with conditions for four phases withinstars (shown in dark gray): hydrogen burning;helium burning; carbon burning; and neon,oxygen, and silicon burning. The phases shownprogress toward increasing temperature becausethe Coulomb barrier that keeps the nuclei apartgets higher as the nuclear charges increase.

This progression toward increasingtemperature is also an evolutionary sequence forstars. A star first gets hot enough in its interior toburn hydrogen (the main sequence starts), andonly when the hydrogen is used up does theinterior get heated and compressed sufficiently toignite helium burning. At this point the star maybe a red giant.

When the helium is exhausted, the interioragain is heated and compressed until carbonburning begins via a suite of reactions beginningwith 12C + 12C → 24Mg as well as other possibleproducts. Among these products are α particles,which can be captured by the 12C to build up anumber of other nuclei, including oxygen, neon,and silicon.

When the carbon is exhausted and the star isa red or blue supergiant, it may undergo a phaseof burning in which the 20Ne is photodissociatedto produce oxygen and α particles; the α particlesare quickly captured again by the nuclei, and the

Section IV 22 Astrophysics and Space Physics

He-burning

C-burning

Ne - O - Si burning

1030

1020

101010–1 100 101 102 103

Temperature (keV)

Den

sity

(cm

–3)

NIF laser only

H-burning

Burning NIF capsule

Figure IV-1. Comparison of the regimes of temperature and density that will be attained on the NIFwith typical conditions found in stars as they progress through their successive nuclear burningstages. Two NIF regimes are shown. The broader regime (light gray) is for the NIF laser only, andextends up to about 14 keV. It shows the temperatures that may be reached in a laser-heated hohlraumon the NIF, or within an imploding capsule, even if there is no nuclear ignition. The second NIFregime (in black) shows the conditions that might exist in a deuterium-tritium capsule after ignition,with temperatures between between 9 and 60 keV. The dark gray areas show conditions within starsfor four phases: hydrogen burning; helium burning; carbon burning; and neon, oxygen, and siliconburning. These phases progress toward increasing temperature. Although it doesn’t always happen,this sequence of phases is also an evolutionary sequence for stars, from an initial hydrogen-burningphase, through the helium-burning and carbon-burning phases, to a phase where neon, oxygen, andsilicon are burned.

oxygen reacts with itself to make a number ofelements between magnesium and sulfur. This isindicated by the neon-oxygen-silicon box.

Whether this last phase occurs in a given star,or even whether carbon burning occurs non-explosively at all, differs from star to star. Inlower mass stars the nuclear evolution is cut offby a hydrodynamic instability that causes all thestar’s outer layers to be expelled, after which thestar lives out its life as a white dwarf. In massivestars a thermonuclear instability turns the starinto a supernova after the carbon-burning phase,or perhaps after the oxygen-burning phase.

The radiative opacity of interest in stars is aharmonic mean of the spectral-dependentopacity, which means that the opacity of interestis dominated by those windows in the frequencyspectrum that lie between the strong absorption

features. The kind of experiment that addressesthe opacity is a spectral absorption measurementof a thick sample, centered on these frequencywindows. The ability of the NIF to provide astable temperature environment for a longenough period, combined with high-quality time-resolved x-ray spectrometers, makes it perfect forthis kind of experiment.

The variation of temperature and densitywith depth for some detailed stellar models isshown in Fig. IV-2, which illustrates thestructures for stars in late stages of evolution.Four late stages are shown: a possible MACHO(massive compact halo object), a 1-solar-masswhite dwarf, a 25-solar-mass pre-supernova, anda star undergoing core carbon burning as well ashelium and hydrogen burning.


10-4 10-3 10-2 10-1 100 101 102

Temperature (keV)

1015

1017

1019

1021

1023

1025

1027

1031

1029

Den

sity

( c

m-3

)

A B C

D

Figure IV-2. Variation of temperature and density with depth for models of stellar structure for fourlate phases. A) A possible MACHO—massive compact halo object. B) A 1-solar-mass white dwarf.C) A 1-solar-mass star at the tip of the red giant branch. D) A 25-solar-mass pre-supernova, a starundergoing core carbon burning as well as helium and hydrogen burning. Each curve shows severaldifferent dominant energy mechanisms—the dashed line for convective transfer, the solid line forradiative regions, and the gray line for regions in which the hydrogen has been completely depletedby nuclear burning.

Opacity has the most effect on stellarstructure in the intermediate zone between theconvective outer envelope and the hydrogen-depleted core. The temperature here ranges from106 to 108 degrees K (100 eV to 10 keV)depending on the position in the star and thestellar mass. The corresponding range of densityis from 10–4 to 102 g/cm3.

The stellar material is rich in hydrogen andhelium, which do not dominate opacity in theseranges, but there are also significant amounts ofcarbon, nitrogen, oxygen, and iron. The carbon,nitrogen, and oxygen ions are importantcontributors at 100 eV, but iron is very importantover the whole temperature range. A firstmeasurement of the opacity of iron has alreadybeen made with Nova, as described in Section VI,Material Properties, SubsectionB. With the NIF,the temperature range up to a few keV withoutignition, or to more than 10 keV with ignition,will be available, which means that conditions atthe centers of the stars will be reachable.

Another area in which NIF-basedexperiments could be useful is with regard toopacities and x-ray cross-sections relevant tocompact x-ray sources such as cataclysmicvariables, x-ray binaries, and quasars. Directmeasurements of inner- and outer-shell photo-absorption cross-sections of carbon, nitrogen, andoxygen could be made in well-characterizedlaser-produced plasmas via a backlightermethod. These cross-sections have not beenmeasured in other ways because of the difficultyof producing a sufficiently large column densityof the desired ions. Such measurements wouldcomplement theoretically calculated cross-sections and aid in the modeling of x-ray ionizedmaterial in and around compact sources.

In the last few years a group at LLNL hasmade new calculations of the radiative opacityfor stellar interiors, and this work has beenincorporated into theoretical models of stellarevolution and pulsation by a number ofastrophysicists around the world. As a result,


some outstanding astrophysical problems havebeen solved.

The problems solved are ones that appearedwhen stellar structure and evolution calculationsmade with the tables of opacity generated usingthe then state-of-the-art opacity codes at LANLwere compared with observations. Previousstellar evolution calculations for the mostmassive stars had predicted that the range ofsurface temperature for the hydrogen-burningphase of evolution was narrow—that the starswere all quite hot at this time, but lowertemperatures were observed. With the newopacity calculations, the range of temperaturespredicted is extended to lower temperaturesas observed.

When the massive stars evolve from bluesupergiants to yellow supergiants they pulsatefor a while as Cepheid variables. The formerevolution calculations led to a relation betweenstellar luminosity and mass that gave the wrongpulsation periods; this problem is solved with thenew opacities. The evolution calculations of thelow-mass stars in globular clusters formerly ledto inferred ages for the clusters that were greaterthan the age of the universe; with the newopacities, this problem has now disappeared.Even though these and other pieces ofastrophysical evidence support the new LLNLopacities, the opacities embody manyapproximations, like the LANL opacities from the1960s and 1970s that preceded them. They shouldbe confirmed in as many ways as possible,including by direct experiment.

B. Equation of State

Stellar structure and evolution are governedby three physical properties of the star’smaterial—the opacity just discussed, the nuclearreaction rates, and the equation of state. The lastof these means the relations between the densityand temperature of the material on the one handand its pressure and internal energy, specificheats, etc., on the other hand. Under manycircumstances, the equation of state in stellar

interiors is quite simple—most of the gas consistsof hydrogen and other light elements, and thesehave lost most of their electrons. Furthermore,the gas density is low enough that the atomicions seldom interact, so ideal gas physics is agood approximation.

This situation is a happy one for theastrophysicist, but unfortunately it is not alwaysthe case. In the very centers of stars, particularlythose in the later stages of their evolution, densityis quite high and the plasma becomes stronglycoupled. That is, the various ions interactstrongly and no longer behave as free particles.This is often accompanied by electrondegeneracy—the plasma electrons tend to fill upthe states allowed by Pauli’s exclusion principle,which forces some electrons to become veryenergetic. These energetic electrons dominate thepressure and internal energy; however, the ionsstill play a major role in the specific heats.

The theory of stellar evolution is affected byuncertainties in the equation of state in a fewareas. The white dwarf stars are the remnants—the nuclear ashes—of stars that have run throughmost of their evolution and have at some pointejected their outer layers containing all theunburned nuclear fuel. They are the compressedformer cores of their parent stars. A white dwarfof the sun’s mass is the size of the earth. Withno more nuclear heat source, the white dwarfsimply cools off, a process that takes manymillions of years.

Through most of the white dwarf, thepressure of degenerate electrons supports thematerial against gravity, but near the surface theelectrons become less degenerate and the ionsbecome important. The ions are also important insetting the specific heat and thus the rate atwhich the white dwarf can cool.

White dwarfs are also sometimespulsationally unstable. Pulsation is a processthat is sensitive to specific heats, and indeedwhite dwarfs are often in several pulsationmodes simultaneously. The precise measurementof the pulsation frequencies gives a sensitiveprobe (asteroseismology) of the interior structureof the star.


See Fig. IV-3 for an example of theimportance of improved opacity and equations ofstate on the understanding of stellar evolution.The figure shows the effects of opacity onCepheid pulsations. Until the early 1990s, thereremained a discrepancy between the predictedmass of pulsating Cepheid stars and observation.The mass should be predicted from the ratio ofthe first harmonic, P1, to the fundamental, P0;however, until the opacity of the stars wascorrected the discrepancy remained. The figureshows calculated results both with older opacities(upper, lighter lines) and with opacities corrected(solid lines), and the observed ratios (dots).

6 M

0.74

0.72

0.70

1.0 3.0 5.0 7.0

4 M5 M

6 M

7 M

5 M

4 M

P1/Po

Po

Figure IV-3. Effects of improved opacity andequations of state on understanding Cepheidpulsations. Until the early 1990s, there remaineda discrepancy between the predicted mass ofpulsating Cepheid stars and observation. Themass should be predicted from the ratio of thefirst harmonic, P1, to the fundamental, P0;however, until the opacity of the stars wascorrected the discrepancy remained. The ratio ofP1/P0 is presented as a function of thefundamental P0. The upper (lighter) linesrepresent the results with the older opacities,the solid lines are the older results with theopacity corrected, and finally, the dots are theobserved ratios. The results of the comparisonindicate that the new opacities resolve theproblem, putting the observations and thetheory in agreement. The ability to verifyopacity at astrophysically relevant conditionswill be greatly enhanced with the NIF.

The results of the comparison indicate that thenew opacities resolve the discrepancy, puttingthe observations and the theory in agreement.The ability to verify opacity at astrophysicallyrelevant conditions will be greatly enhancedwith the NIF.

Another kind of dwarf star—the browndwarf—is poorly understood, owing in part toour imperfect understanding of the equation ofstate. Brown dwarfs are stars with masses so lowthat they never get hot enough inside to burnhydrogen, but simply condense from theinterstellar medium and shine dimly for awhileas they get rid of the heat formed in condensing.Brown dwarfs are very difficult to detect, but it ispossible that they are so numerous that theycomprise as much as half of all the mass in ourgalaxy. This is mass that is inferred to exist fromthe dynamics of the galaxy, but that has not beendiscovered in luminous form. Their internalstructure and their cooling time depend on thedetails of the equation of state at densitiesapproaching solid density and temperatures of afew eV; these conditions are easily reached in alaser experiment.

Equation-of-state (EOS) experiments fall intotwo classes. One class might be thought of as“direct” EOS experiments, in which for a certainexperimental setup the experimental data yielda quantity that can be compared with thetabulated functions of pressure, energy, etc. Thesecond class is of “indirect” EOS experiments,which measure some of the detailed intermediatedata that enter into theoretical expressions forthe thermodynamic functions. Experiments ofthe second kind serve as benchmark tests of theEOS theory. Experiments of the first kindare described in detail in Section VI,Material Properties.

The indirect kind of experiment could, forexample, seek to validate the model of plasmascreening—very important in dense plasmas—bydiagnosing the occupation probabilities of atomicstates close to the ionization limit. This isdiscussed at greater length in Section IX,Radiative Properties.


C. Plasma Spectroscopy inHigh-Energy Astrophysics

High-energy astrophysics refers to x-ray andgamma-ray, and to some extent optical and UV,observations of stars and galaxies for which thesource of the radiation is a violent event, or aplasma very far from thermodynamicequilibrium. Among these are quasars and otherexploding galaxies, and neutron stars that areingesting material from a companion star. Thetechnique used to study them is to investigatetheir spectra—in the x-ray or gamma-ray range,or whatever is available—to find the signaturesthat reveal the nature of the emitting object. Theastronomer’s problem is to obtain a spectrumshowing as much detail as possible. However, asecond problem is to have the knowledge base onnon-equilibrium spectra that allows inferences tobe drawn from the data. Here is where the NIFhas a role, working together with first-principlesplasma-modeling tools.

An area where our modeling of plasma isrelatively weak concerns low-temperature

radiation-dominated plasmas. These are plasmasirradiated by dilute radiation sources—that is, theilluminating radiation field is much weaker inabsolute intensity than the blackbody thatmatches the spectral shape. The radiation, whilestrong in absolute terms, is still weak incomparison with the very high impliedtemperature. Such plasmas are very far fromlocal thermodynamic equilibrium (LTE). Oneuncertainty is of the plasma energetics. Perhaps alot of energy is tied up in states that are notevident in the spectrum.

The Bowen Mechanism

Because the plasmas are radiation-dominated, the thermodynamic state of thematter and the appearance of the spectrum itselfdepend strongly on the flow of radiation throughthe plasma. The re-absorption of radiationemitted elsewhere plays a major role. Aparticularly notable example of this is the Bowenfluorescence mechanism, which is discussed insome detail below and shown schematicallyin Fig. IV-4.

He II 2 2P02

1 2S

λ303.78

O III2p3d 3P02

2p3p 3P2,1

2p3p 3S12p3p 3D

3,2,1

2p3s 3P02,1,0

2p2 3P2,1,0

λ303.62λ303.80

2p 2P1/2

λ374.441

N III 3d 2D5/2

λ374.436

3p 2P

Figure IV-4. Grotrian diagram for the Bowen mechanism. The observed fluorescence occurs when theemission from the HeII Lyman α transition pumps the 2p2 3P state of OIII to the state 2p3d 3P. Thisleads to the enhanced emission from the transition down to the states 2p3p 3P, 3S, and 3D. There isalso enhanced emission from these states to the 2p3s 3P level. In addition, it has been pointed out thatthe emission from the OIII 2p3s 3P2 to the 2p2 3P can pump the ground state of NIII 2p 2P to the2d 2D5/2 level, which then can decay via the 3p 2P level with enhanced emission in the visible. Thislatter part of the mechanism is not yet confirmed. The NIF would be required in order to generateboth the large volumes of plasma necessary and a radiation source sufficient to create the plasma in anappropriate state to provide experimental verification of the Bowen mechanism.


Ira Bowen found, in 1924, that in somenebulae (the planetary nebulae that are shellsejected by red giant stars) the emission lines ofOIII arising from the upper states 2p3d 3P0

2 weregreatly enhanced relative to other members of thesame multiplet. The lower levels of thesetransitions, namely 2p3p 3P2,1 , 2p3p 3S1, and2p3p 3D3,2,1 fed transitions to 2p3s 3P0

3,2,1 thatwere also enhanced. Bowen found that there wasa coincidence between the HeII Lyman α line atλ303.78 and the OIII line λ303.80 that excites 2p3d3P0

2. Furthermore, the decay of 2p3s 3P01

produces λ374.436, which can pump the lineλ374.441 in NIII that excites 3d2D5/2, producingstill more fluorescence.

The NIF would be required in order togenerate both the large volumes of plasmanecessary and a radiation source sufficient tocreate the plasma in an appropriate state toprovide experimental verification of the Bowenmechanism. The uniform heating of material toextreme conditions, when the plasmas requirex-ray heating, is energy-intensive, becauselarge proportions of the flux are needed forvolumetric heating.

The Bowen mechanism was analyzed in the1960s and 1970s, not with a full collisional-radiative model, but with some radiative transfer.The NIII branch is not considered well-established, but at least the first lines in the OIIIbranch are. The line coincidences havemismatches of 16 km/s between HeII and OIII,and 4 km/s between OIII and NIII. The HeII/OIIIdifference is larger than the Doppler width,which is about 9 km/s for helium at 20,000°K.Most nebulae also have an internal velocitydispersion of about 10 km/s due to generalexpansion, or to turbulence, and there is opacitybroadening, especially in the HeII Lyman α line.These two factors combine to ensure coincidence.

The OIII lines, at λλ 3100–3400, are seen inx-ray binary stars like Sco X-1 (a low-mass x-raybinary) and AM Her (a cataclysmic-variable-likewhite dwarf binary), and also in some Seyfertgalaxies (spiral galaxies that are relativelyordinary but with a black-hole-like object in thecenter that makes x-rays and excites a lot of lineemission in the nearby interstellar gas—mini quasars).

The Bowen mechanism still does not have asecure theoretical footing, and needs someexperimental verification. The full network ofatomic kinetic rates connecting these levels ofHeII, OIII and NIII is complex enough to be hardto calculate. An experimental exploration of theBowen mechanism will create a large volumeof plasma with a temperature of 2–3 eV and inwhich oxygen and nitrogen are mostly doublyionized. This may be illuminated by a secondregion of hotter plasma that serves as a lightsource for the first. The column density of thelarge plasma, perhaps 1020 cm–2, would beample to provide the necessary optical depth inHeII Lyman α , and perhaps also in OIII λ374.436as well.

Low-Mass X-ray Binaries

Low-mass x-ray binaries (LMXRB) consist ofa neutron star in orbit with a 1- to 2-solar-massstar; the x-ray spectra are relatively hard—10 keVor so. Considerable modeling work has beendone on these, with particular attention to theiron spectrum—first the K-shell lines, and morerecently the L-shell emission. This is a formidabletask, especially the dielectronic recombination,and experimental confirmation is very desirable.The iron K and L lines are seen in spectra thathave been obtained of x-ray binaries like Sco X-1using BBXRT and two Japanese satellites. Similarlines are hinted at in the few decent x-ray spectraof Seyfert galaxies, which, if confirmed, wouldcorroborate the theory of an accretion disksurrounding a black hole in the nucleus ofthe galaxy.

Figure IV-5 illustrates the relevant process ofdielectronic recombination. The dielectronicrecombination process is shown by the twoarrows. A free electron in the 2s continuum (onthe left), whose energy is indicated by the blueline, is captured into one of the 2pnl states (on theright) at exactly the same energy, as shown by thefirst arrow. This state can decay by ejecting thefree electron again, or, as shown by the secondarrow, by radiating a photon and making thetransition to a singly excited 2snl state, typicallywith the same n. This two-step process isdielectronic recombination.


2s continuum

2s4l

2s3d2s3p2s3s

2p2

2s2p

2s2p

2s2

2p3l

2p4l

2p continuum

2p2

Figure IV-5. Illustration of the process ofdielectronic recombination for therecombination of C+3 to C+2. On the left side arethe levels of C+2 converging to the series limit,which is the 2s state of C+3. Most of these levelsbelong to the configurations 2snl, with n = 3,4, ... . On the right side are the levels of C+2 thatconverge to the 2p state of C+3. These belong tothe configurations 2pnl, again with n = 3, 4, ... .Because the 2s and 2p states of C+3 have thesame principal quantum number, their energydifference is slight, and therefore many of the2pnl states lie below the 2s series limit andsome of the 2pnl states lie just barely above the2s limit. The dielectronic recombination processis shown by the two arrows. A free electron inthe 2s continuum, whose energy is indicated bythe blue line, is captured into one of the 2pnlstates at exactly the same energy, as shown bythe first arrow. This state can decay by ejectingthe free electron again, or, as shown by thesecond arrow, by radiating a photon andmaking the transition to a singly excited 2snlstate, typically with the same n. This two-stepprocess is dielectronic recombination.

When the available doubly excited states liehigh up in the ground-state continuum, thenumber of free electrons with the right energy issharply limited by the Boltzmann distributionunless the electron temperature is high. In caseslike the one illustrated, though, the ∆n = 0transition of the upper ion allows doubly excited

states very close to the ionization limit, whichmakes them accessible at even quite low electrontemperatures. The total rate of dielectronicrecombination for ions like this is very large.

In the next five years there will be two newhigh-resolution x-ray astronomical observatories:XMM and AXAF. These will, for the first time,obtain spectral detail of the iron K and L featuressufficient to distinguish different types ofexcitation. The line shapes in these objects, and,for the binaries, their time variations, can beanalyzed to provide information about the radialand azimuthal structure of their accretion disks,provided the atomic model itself is sufficientlywell understood.

An experimental approach to the iron Lspectrum would be of great interest. Theproblem, in laboratory sources, is to avoid largeoptical depths in these lines, which is undesirablewhen basic line intensity data are being sought.The laser-produced plasma has the advantageover other sources, such as the Electron Beam IonTrap (EBIT), in that the temperature can be keptas low as the astronomical situation indicates.

One of the key issues involving iron is thedielectronic recombination through the near-threshold resonances that comes about with∆n = 0 capture (Koster-Kronig transitions)—forexample, the capture onto lithium-like iron inwhich the 2s electron is excited to 2p. Thisdielectronic rate is very large, and a propercalculation of it requires a careful calculation ofthe energy of each resonance and itsautoionization width; large departures from thegeneral formulae for dielectronic recombinationrates are found. A schematic of the process isshown in Fig. IV-5. In beryllium-like iron thestates 2pnl with n greater than about 11 are auto-ionizing. The spacing of states differing by one inn is about 11 eV, so there are certain to be a fewautoionizing levels within 10 eV or so of theionization threshold.

It would be quite interesting to produce aplasma with abundant lithium-like iron but withan electron temperature less than 50 eV, in orderto study this process. The color temperature ofthe radiation would need to be a few hundredeV in order to produce this ion, which has anionization potential of 2048 eV. We note that it


is true that in accretion disks of x-ray binariesand Seyfert galaxies this ion of iron is probablyfound where the electron temperature is hotter,of order of a few hundred eV to 1 keV, so thenear-threshold dielectronic recombination isrelatively unimportant.

Shock-Wave Ionized Media

Another radiation transport study possiblewith the NIF involves shock waves in the targetchamber to study photo-ionized nebulae. A widerange of experiments can be done to simulate theeffect of energetic events in circumstellar andinterstellar gases by utilizing a low-density gasfill in the 10-m-sized laser target chamber. Thedeposition of laser light onto a small target at thecenter of the chamber will lead to a bright UVand/or x-ray source (80–90% of the laser light canbe converted to x-rays). In a low-opacity gas aphoto-ionized nebula will be created.

Detailed study of such processes will serve tobenchmark models that are used in astrophysicalcontexts, such as quasars. Processes such asphotoionization and excitation, radiative anddielectronic recombination, and line trappingwould be studied. For a high-opacity gas, thex-rays will shortly be re-absorbed, leading to thecreation of a shock wave in the chamber gas,similar to a supernova remnant in the interstellarmedium. Similar work has already been doneon a small scale at the Naval ResearchLaboratory.1 Fundamental studies of thepropagation of such shocks, including effects ofan applied magnetic field, can be done.

D. Supernova Instabilities

Hydrodynamic instabilities play a major rolein determining the efficiency and performance ofinertial confinement fusion implosions. In laser-driven implosions, high-performance capsulesrequire high aspect ratios (the ratio of the radiusto the shell thickness). These capsules aresusceptible to hydrodynamic instabilities of theRayleigh-Taylor, Richtmyer-Meshkov, andKelvin-Helmholtz varieties, which can inprinciple severely degrade capsule performance.Recently the growth of the ablatively driven

Rayleigh-Taylor instability has been studied byindirect-drive experiments using a shaped x-raydrive pulse.2

The x-ray and gamma-ray observations of thefamous supernova 1987A have indicated thatradioactive cobalt is far more thoroughlydistributed among the explosive debris in theenvelope than was predicted by modelcalculations of thin-shell nucleosynthesis in thepre-supernova star. These observations of 1987Ahave strongly suggested the occurrence of large-scale mixing in the ejecta during the explosion.3,4

A larger scale mixing in the ejecta has also beensuggested for Type Ib/Ic supernovae by lightcurve modeling and spectrum analysis.

The most promising mechanism forexplaining mixing in the ejecta is a combinationof the Rayleigh-Taylor and Kelvin-Helmholtzinstabilities. The Rayleigh-Taylor instability canarise in the supernova envelope when theoutwardly moving shock wave from the initialexplosion propagates through layers of the starwith radial stratification of the heavy elements.As the shock passes through the compositioninterfaces (i.e., oxygen/silicon,helium/carbon+oxygen and hydrogen/helium),a rarefaction front moves back into the star,resulting in an effective reversal of gravity aslow-density composition is pressure-acceleratedinto the underlying high-density composition.

Any perturbation at the interface(i.e., velocity perturbation or spatial perturbation)will get amplified by the Rayleigh-Taylor andRichtmyer-Meshkov instabilities and result in theoverturning of light and heavy elements. Thisresults in the mixing of heavy elementsthroughout the envelope of the supernovaremnant, with associated observationalconsequences in the light curve.

A further mixing will occur as the dense“tongues” of the heavy elements experiencedifferential shear with the lighter elements,resulting in Kelvin-Helmholtz instabilities. Thusthe fingers of heavy and light fluid thatdeveloped initially get far more distorted and themixing layer increases its width.

Eventually these instabilities become sononlinear that the mixing layer appears tobecome fully turbulent. The properties of


turbulently mixed layers may be equallyimportant in understanding how interstellarclouds get reprocessed back into the interstellarmedium. Efficient mixing of cloud and inter-cloud matter has been shown to occur afterclouds get crushed by the interaction of strongshocks from supernova remnants.5,6,7

Experiments at the NIF would be ideally suitedto detailed studies of this important astrophysicalphenomenon. Densities and temperaturesreached in the supernova-shocked envelope arewell within the range attainable with the NIF.

The experimental techniques to study thedevelopment of Rayleigh-Taylor-induced mixingwell into the nonlinear regime using streakedx-ray radiography of the unstable region havealready been demonstrated by Remington et al.2

The NIF is ideally suited to producing theextremely strong shocks with Mach numbersM >> 1 that occur in the supernova envelope. Theshocks would be produced by shaped laserpulses ablating a foil or cylinder consisting of asequence of material interfaces with densitydifferences matching those in the supernovaenvelope. The shock wave would traverse thecomposite interfaces, which can be suitablyperturbed, resulting in instabilities and mixinglike those thought to occur in the supernovaenvelope. Shocks of sufficient strength are far outof the regime that is currently possible instandard shock-tube experiments, but will berelatively easy to achieve at the NIF.

Recent advances in high-resolutionmultidimensional hydrodynamics using adaptivemesh refinement (AMR) techniques now make itpossible to accurately model the experimentswith detailed three-dimensional simulations.Using newly developed AMR algorithms,8 recentcalculations6,7 have studied the development ofshock-induced Rayleigh-Taylor and Richtmyer-Meshkov instabilities far into the nonlinearmixing regime. Calculations such as these,combined with NIF experiments, will allow athorough study of the mixing in supernovaexplosions as well as studies pertaining to mixingin the interstellar medium.

Figure IV-6 (reproduced from Müller et al.9)shows the typical scenario for how Rayleigh-Taylor mix enters into supernova modeling. The

expanding inner core of nickel and silicon isbeing decelerated by the surrounding layer oflower density carbon and oxygen, which in turnis being decelerated by the outer layer of helium.At each interface one effectively has a heavy fluid“sitting on top of” a light fluid. Because of this,the interfaces are unstable to the Rayleigh-Taylorinstability, and imperfections will grow. Typicalgrowth factors are 100–1000.

Figure IV-7 shows how a similar situationcould be set up on the NIF. Since the usualexperiments focus the lasers into the hohlraum,the hohlraum in this case would be shot in aninside-out fashion. The drive lasers wouldimpinge upon a thin (approximately 10-µm)cylindrical nickel hohlraum with sinusoidalperturbations imposed on the outer surface. Theouter surface of nickel would be coated with≈100 µm plastic, and the whole package wouldreside in a ≈10-atm fill of helium.

As the drive turns on, a strong shock(≈100 Mbar) would be launched radially

Figure IV-6. Rayleigh-Taylor mix in supernovamodeling. The four quadrants show contours oftotal density times the mass fraction of 4He atfour instants during a supernova explosion. Thefour times proceed counterclockwise beginningwith the upper right section. Note that theoriginal sphere of helium is already breakingup in the first instant shown.


outwards through the various layers. Eachinterface would be unstable to the Richtmyer-Meshkov instability, and perturbation growthwould begin. Then as the hohlraum explodesradially outwards, the nickel core would bedecelerated by the lower density plastic layer,which in turn would be decelerated by thehelium fill, in much the same fashion as for thesupernova shown in Fig. IV-6.

Diagnosis would be from face-on, edge-on,and end-on radiography, in the standard fashioncurrently utilized on Nova. A large-area copperbacklighter, with Kα emission at 8.4 keV, mightbe suitable. If this “direct-drive” approach toaccelerating the nickel hohlraum proves toimprint an intolerable level of non-uniformityon the nickel-plastic interface, a hybrid indirect-drive approach would be the fall-back. Here,

Drive beams

Backlighter beamCu backlighter disk

Io

Gated x-ray imager

Filters

I = I0e–ρΚZ(x) ZX

I

He fill (500 µm)

CH (100 µm)

Ni (10 µm)

Polyimide (0.5 µm)

2 m

m

5 mm

Figure IV-7. Schematic setup of an astrophysical mix experiment to mimic the mixing in a supernovaexplosion. The drive lasers would impinge upon a thin (approximately 10-µm) cylindrical nickelhohlraum with sinusoidal perturbations imposed on the outer surface. The outer surface of nickelwould be coated with ≈100 µm plastic, and the whole assembly would reside in a ≈10-atm fill ofhelium. As the drive turns on, a strong shock (≈100 Mbar) would be launched radially outwardsthrough the various layers. Each interface would be unstable to the Richtmyer-Meshkov instability,and perturbation growth would begin. Then as the hohlraum explodes radially outwards, the nickelcore would be decelerated by the lower density plastic layer, which in turn would be decelerated bythe helium fill, in much the same fashion as α in a supernova. Diagnosis would be from face-on, edge-on, and end-on radiography.


the lasers would hit an inner solid high-Zcylindrical rod, generating an x-ray drive.The x-ray drive would in turn drive the nickelhohlraum outwards.

Note that already on Nova with a ≈20–30-kJdrive, growth factors of nearly 100,10 shockstrengths of ≈ 50 Mbar, and Mach numbers of≈35 have been achieved.11 On the NIF, with1.8 MJ of drive energy, achieving growth factorsof 100–1000 with 50- to 100-Mbar shocks, assuggested by the supernova modeling, shouldbe straightforward.

E. High-Velocity Cratering

One of the interesting issues in lunar andplanetary physics is that of crater formation—surface damage—by high-velocity impacts. Thevelocities of projectiles in the solar system rangeup to 50 km s–1 (for projectiles in retrogradeorbits), which is out of the range for gas guns.This kind of high-velocity projectile impact studycan easily be done using the laser-driven flyerplate technology possible on the NIF, asdiscussed in the Hydrodynamics and MaterialProperties sections. The physics issues to bestudied include the systematics of craterdimensions with projectile shape and size—thenature of the damage mechanism. An examplewould be the dependence of the cratering on thetype of material forming the target.

A further set of issues is related toelectromagnetic effects caused by the impact. It isknown that electromagnetic noise is produced atthe time of impact. It is also found that there aresubstantial magnetic anomalies in the rockunderlying lunar impact craters. The mechanismsof these phenomena can be studied byperforming the flyer plate experiments withweakly magnetized targets.

The suitability of this project to the NIF arisesfrom the ease of attaining the high impactvelocity and the availability of larger projectilesizes. On Nova the projectiles with a suitablylarge velocity are so small that edge effectsobscure the phenomena to be studied.

F. Thermonuclear Reaction Ratesin Stars

Although nuclear reaction rates have beenthe subject of intense studies for decades, itremains true that few direct experimental studieshave been made at the actual energies at whichreactions occur in stellar environments.Figure V-1 summarizes a number of astro-physical thermonuclear environments and theirassociated temperatures. The actual energiesrelevant for the nuclear cross-section s are givenby the so called Gamow window, which repre-sents the interplay between the Maxwellian tailof the reactant velocity distribution and therapidly diminishing reaction cross-section atenergies below the Coulomb barrier.

The relative merits of the NIF for obtaininginformation on nuclear reaction rates, as opposedto the more usual accelerator technique, can bedetermined as follows. Consider the problemcount rate. For a conventional acceleratorexperiment, the reaction rate for projectile a plustarget nucleus b to produce some product nucleusc can be written:

r a b Accelerator I n xa ab b+( ) =, ,σ ∆where Ia is the incident projectile current, σab isthe cross-section for the reaction betweenprojectile a and target nucleus b, while nb and ∆xare the target density and thickness, respectively.Consider a typical reaction, the 12C(p,γ)13Nreaction, which plays a role in the CNO cycle.The hydrogen burning temperatures for thisreaction are from about 1 to about 3 keV. At3 keV, the Gamow window for this reaction is atapproximately 40 keV, for which one expects across-section of order 0.01 picobarn.12

For optimistic accelerator conditions with anincident current of 100 µA on a carbon target ofdensity 1023 cm–3 and a thickness of 10–4 cm (theapproximate range of protons in carbon for thisenergy), the reaction rate would be at most fiveevents per day, quite difficult to pick out of thecosmic ray background.

On the other hand, the reaction rate in athermalized NIF capsule can be written:

r a b NIF N na b+( ) = ⟨ ⟩, ,σν


where <σν> is the Maxwellian-averaged productof cross-section times relative velocity, nb is thedensity of nucleus b , and Na is the total numberof nuclear species a that can react with b. At thetemperature of 8 keV that might be attained in acapsule imploded using the NIF, the expectedMaxwellian averaged reaction rate13 would beNa<σν> ≈10–5 cm3 mole–1s–1. At a hydrogendensity of 102g cm–3, reacting with 1017 carbonatoms (2 µg) for ≈100 ps, ≈104 reactions would beexpected in a single event. This number ofradioactive 13N nuclei would have a half-life t1/2of ≈10 min, so it would be possible to count themafter an event. Alternatively, the pulse of ≈2-MeVgamma rays might be detected with scintillatorspositioned around the target. The fact that theevents are produced all at once avoids the usuallow signal-to-noise ratio associated withconventional accelerator experiments, in whichthe reactions occur one at a time and aretherefore difficult to distinguish from ambientroom background.

Of course, such experiments would not bewithout their own difficulties. The various statevariables are changing rapidly and it is necessaryto have accurate diagnostics and simulations ofthe thermodynamic evolution of an event,including departures from equilibrium. Thereaction rates are so sensitive to temperature thatit will be difficult to know the burn temperatureto sufficient accuracy. It is possible to envision,however, a relative measurement of severalspecies at once, making it possible to fix thetemperature with the most sensitive reaction.

Also, these are explosive environments noteasily amenable to conventional nuclear countingtechniques. Backgrounds are also certain todevelop. Nevertheless, it is possible to develop aclass of nuclear experiments in which a catcherblanket that could collect radioactive productswould be placed at some distance from thecapsule explosion. The catcher material couldlater be removed from the chamber and countedfor accumulated radioactivity. The final stepwould be to compare the nuclear yields producedin the experiment with the yields predicted byapplying thermonuclear network codes to theconditions one believes occurred in the event.This comparison could be used to test both the

codes and the reliability of the thermonuclearreaction rate data. It could also aid as a diagnosticof the capsule history, because the thermonuclearreaction rates are quite temperature sensitive.

Interesting nuclear reactions to study besides12C(p,γ)13N would be 3He(3He,2p) 4He, and3He(4He,γ) 7Be. These latter two are part of theproton–proton chain of hydrogen-burningreactions in solar-type stars, while the first is partof the CNO cycle. The competition between theproton–proton reaction and CNO cycle for thedestruction of 3He determines how much 7Be isproduced (and ultimately how much 8B), andtherefore how great the flux of neutrinos shouldbe, as detected in the Homestake Mine solarneutrino experiment.

The reaction rates that take place in the CNObi-cycle are illustrated in Fig. IV-8. The primaryloop is shown as the outer loop, the secondaryloop is shown within. In both loops, the energygeneration is the energy gained by creating theHe4 from the four H1, with the residual loss ofthe two neutrinos, ν, divided by the time it takesto circle the cycle. The times for the radiativedecays of the N13, O15, and F17 are fixed but theremainder of the rates will depend on thetemperature and density of the constituents. Thedensities accessible to the NIF will permit thecritical determination of these rates in regimesrelevant to astrophysical situations.

This branching ratio continues to be asignificant uncertainty in the interpretation of thesolar neutrino experiment, and a refinement ofour knowledge here would be of great value. Theparticle decay rate of the 3He(3He,2p) 4Hereaction is some 105 times greater than that of theradiative decay rates of 12C(p,γ)13N and3He(4He,γ) 7Be. In order to diagnose 3He(3He,2p)4He it will be necessary to measure the activationproduced by the ≈5-MeV protons in a witness foilplaced near the target, or in the wall of thehohlraum itself.

Figure IV-1 showed the way in which NIFconditions overlap various stellar burningphases. Temperatures up to carbon-burning areachievable with ignition. Even with the driveralone (i.e., without ignition), hydrogen burningand helium burning temperatures are obtainable,


C12 N13

C13

N14O15

+ H1 → γ +

→ ν + e +

+

+ γ

← H

1 +

+ γ ← H1+

+ H

1 → H

e4 +

+ ν + e + ←

O16

F17

O17

+H1→γ+N15 +H1→γ+

→γ+e +

+ν

+H 1→He 4+

Figure IV-8. Reaction rates that form the CNO bi-cycle. The primary cycle is shown as the outer loop,cycling around C12, N13, C13, N14, O15, and N15. The secondary loop connects the N15, O16, F17, and O17.In both loops the energy generation is the energy gained by creating the He4 from the four H1, withthe residual loss of the two neutrinos, ν , divided by the time it takes to circle the cycle.

as are the density regimes for both hydrogen andhelium burning.

Figure IV-9 shows the evolution of the earlyuniverse through the big bang. Lines are drawnfor both baryons and electrons for baryondensities corresponding to both 0.01 and 1.0times the present closure density of the universeΩb. It seems that with an appropriate capsuledesign, it would be possible to nearly reproducevarious conditions in the early universe duringmuch of the critical epoch (first three minutes) ofprimordial nucleosynthesis.

G. Electron-Positron Plasmas

Thermally produced electron-positronplasmas are thought to play an important role inthe evolution of the cores of massive stars,neutron-star and black-hole accretion disks,pulsars, quasars, astrophysical gamma-raybursters, and in the big bang. Although it is notlikely that the NIF could be used to produce a

pair plasma with the currently proposed designparameters, it is possible that some pairs will beproduced in a way that is similar to the formationof pairs in such astrophysical environments.Hence, experiments with the NIF may shed lighton pair-plasma formation and radiation transportissues. We therefore include here a briefdiscussion of some of the physics issuesregarding pair production.

In the last few years, discoveries of intensebroadened 511-keV annihilation lines fromseveral galactic black-hole candidates suggestthat in addition to transient-pair production,steady-state thermal pair plasmas may exist.Since pairs annihilate on short time scales,maintaining such steady-state conditions requiresthe copious production of pairs, in order tobalance the annihilation rate. Such pair-balancedsteady plasmas represent a new state of matterwith unique radiative and thermodynamicproperties different from ordinary plasmas.

For a plasma to be in a steady state, theheating rate must be balanced by the cooling rate,


0.1 1 10 100 1000Temperature (keV)

1030

1025

1020

1015

1010

Den

sity

(cm

–3)

Primordialnucleosynthesis

NIF laser only nb

Ωb = 1.00

Ωb = 0.01

100106 104 102108Time (seconds)

BurningNIF

capsule

ne– + ne+

e+–e– pairs annihilate

Figure IV-9. Trajectories in temperature and density parameter space for some big-bangnucleosynthesis models. The boxes indicate the overlap between cosmological models and plasmasthe NIF can produce. The models are for two different baryon densities. The thick lines indicate thepresent baryon density of the universe that would be needed to close the universe, while the thin linesindicate a model with 0.01 of the critical density to close the universe. The dashed lines are the baryondensity and the solid lines are the sum of the electron and positron densities. The upper abscissa is thetime from the big bang. Note that the baryon curves start at the point of “weak freezeout” at 1 MeV,below which the weak interaction processes depart from equilibrium and the neutron-proton ratiobecomes locked in.

which consists of bremsstrahlung, inverseCompton scattering, and pair annihilation. Itturns out that for a pair-balanced plasma thereexists a fundamental limit to the temperatureof ≈10 MeV for hydrogen, above which paircreation can no longer be balanced by annihila-tion and pair density will exponentiate rapidly,leading to a pair-dominated plasma and netcooling of the system. This limiting temperature,referred to as the BKZS limit, is much too high forNIF-based experiments; however, the limitingtemperature can be lower for a high-Z or highmagnetic-field plasma.

The laser intensities available at the NIFmay permit the formation of some pairs out ofthe suprathermal electrons during a high-temperature deuterium-tritium (DT) burn. Inequilibrium, at a temperature of 50 keV, thevacuum pair abundance is suppressed by onlya factor of exp(–mec2/kT) ≈ 10–5, implying a pair

density as high as ≈1021 cm–3. This factor will bereduced, however, by the ratio of the burn time(of order ≈10 ps) to the equilibration time. Theequilibration time is of order of the time scale forpair creation.

The most promising avenue to the produc-tion of a pair plasma seems at this time to be touse the petawatt laser under development as anadjunct to Nova. One of the Nova beams isintercepted partway through the amplificationchain and is subjected to pulse compression,producing a pulse of 1.06-µm light containingabout 1 kJ of energy in a pulse length of 1 ps.If this can be focused to a spot size limited onlyby diffraction, then an intensity of 1021 W/cm2

can be produced. The paper by Wilks et al.1shows how such a laser pulse can producerelativistic electrons and confine them througha combination of the optical fields and theponderomotive force.


At very high laser intensities I, satisfyingI λµ2 > 1.4 × 1018 W cm–2 µm2 where λ µ is thelaser wavelength in microns, the electron jittermomentum becomes relativistic and scalesroughly as the square root of the intensity. If thejitter energy E of the electrons is larger than 2mc2,as it is for I λ µ2 >> 1.4 × 1018, then collisions of theelectrons with ions of charge Z in the plasma willproduce pairs; the cross-section for this process isof order 10–30Z 2 cm2, weakly dependent onelectron energy in this range. This would besufficient in a critical-density gold plasma toproduce in 1 ns a pair density of order 10–5–10–4

times that of the electrons.This fraction would be increased if a density

substantially above the critical density could beused, as is suggested by Wilks et al .14 It may beimportant in the experiment design to providelaser illumination from opposing directions; thisallows the pairs that are created to be trapped bythe optical field and be themselves accelerated,producing more pairs in turn. The pairs wouldannihilate over a time likely to be longer than thelaser pulse, so the result would be a temporallybroadened annihilation line in the gamma-rayspectrum. The line profile plus the Comptonizedbremsstrahlung continuous spectrum will formthe most useful diagnostics of the pair plasma(e.g., Liang and Dermer15).

An extension of the peta-watt concept to theNIF—providing compressed pulses anddiffraction-limited focusing capability on somenumber of the NIF beams—would extend thespatial and temporal scales for this relativisticplasma, and perhaps make creation of a pair-dominated plasma a reality.

H. References

1. B. H. Ripin, A.W. Ali, H. R. Griem, J. Grun,S. T. Kacenjar, C. K. Manka, E. A. McLean,A. N. Mostovych, S. P. Obenschein, and J. A.Stamper, in Laser Interactions and RelatedPlasma Phenomena 7, G. Miley and H. Hora,Eds. (Plenum Press, New York, 1986), p. 857.

2. B. A. Remington, S. V. Weber, S. W. Haan,J. D. Kilkenny, G. G. Glendinning, and R. J.Wallace, Phys. Fluids B 5, 2589 (1993).

3. Kumagai T. Shigeyama, K. Nomoto, M. Itoh,J. Nishimura, and S. Tsurata, Ap. J. 345, 412(1989).

4. D. Arnett, J. Bahcall, R. Kirschner, andS. Woosley, Ann. Rev. Astron. Astrophys. 27,629 (1989).

5. R. I. Klein, C. F. McKee, and P. Colella in TheEvolution of Interstellar Medium, L. Blitz, Ed.,Astronomical Society of the PacificConference Series 12, 117 (1990).

6. R. I. Klein, C. F. McKee, and P. Colella, Ap. J.420, 213 (1994).

7. R. I. Klein, J. Bell, R. Pember, and T. Kelleherin the proceedings of the 4th InternationalConference on Compressible Turbulence,Cambridge, UK (in press).

8. J. Bell, M. Berger, J. Saltzman, andM. Welcome, SIAM J. Sci. Comp. 15, 127(1994).

9. E. Müller, B. Fryxell, and D. Arnett, Astron.Astrophys. 251, 505 (1991).

10. B. A. Remington, S. W. Haan, G. G.Glendinning, J. D. Kilkenny, D. H. Munro,and R. J. Wallace, Phys. Rev. Lett. 67, 3259(1991), and B. A. Remington, S. W. Haan,G. G. Glendinning, J. D. Kilkenny, D. H.Munro, and R. J. Wallace, Phys. Fluids B 4, 967(1992).

11. B. A. Hammel, D. Griswold, O. L. Landen,T. S. Perry, B. A. Remington, P. Miller, T. A.Peyser, J. D. Kilkenny, Phys. Fluids B 5, 2259(1993); B. A. Hammel, J. D. Kilkenny, D. H.Munro, B. A. Remington, H. N. Kornblum,T. S. Perry, D. W. Phillion, and R. J. Wallace,Phys. Plasmas 1, (1994—in press); andG. Dimonte and B. A. Remington, Phys. Rev.Lett. 70, 1806 (1993).

12. D. D. Clayton, Principles of Stellar Evolutionand Nucleosynthesis (McGraw-Hill, New York,1968).

13. M. Harris, W. Fowler, G. Caughlann, andB. Zimmerman, Ann. Rev. Astron. Ap. 21, 165(1983).

14. S. C. Wilks, W. L. Kruer, M. Tabak, and A. B.Langdon, Phys. Rev. Lett. 69, 21383 (1992).

15. E. P. Liang and C. D. Dermer, Ap. J. 325, L39(1988).

Hydrodynamics 37 Section V

Section V

Hydrodynamics

A. Techniques

Experience gained by researchers on theexisting high-energy lasers has demonstrated thata variety of fluid-dynamics experiments may becarried out on such facilities. In particular, twotypes of experiments have emerged as theleading hydrodynamic experimentalapproaches—those that involve radiography andothers that utilize self-emission.

To provide enough background tounderstand the belief that the high-energy lasersprovide a vehicle to open novel areas ofhydrodynamics not possible at other facilities, weprovide the following illustrations of technique,which show clearly the current limitations as wellas the advanced state of the techniquesdeveloped. These particular examples areselected from work conducted on Nova, becausemany of their features are shared with otherexperiments and thus they permit some generalobservations to be drawn.

Radiography—A Typical Configuration

In a series of experiments, we can investigatethe quasi-two-dimensional flow of a jet formedby the passage of a shock over a hemisphericalprotrusion on a density interface. A side view ofthe configuration for these experiments, shownend-on to the hohlraum, is depicted in Fig. V-1.

The experimental package—the object to beobserved—consists of a beryllium tube with aninsert of bromine-doped plastic joined to carbonfoam, two materials with different densities. Theinner diameter of the beryllium tube is 700 µm,and its length is 2000 µm. There is ahemispherical feature at the interface between thebromine-doped plastic and the carbon foam, to

seed the instability.The hemispherical feature isproduced by using a lathe to machine ahemispherical hole in a bromine-doped plasticcylinder and a matching hemispherical bump(which has a radius of 150 µm). A photograph ofthe carbon part is shown in Fig. V-2.

The tube is abutted to the center of theexterior of a gold hohlraum, which provides thex-ray drive. A hole exposes the end of thepackage to the hohlraum interior. A large-areatitanium x-ray backlighter provides back-illumination for an x-ray framing camera, whichimages the materials during transmission of theshock wave.

Au hohlraum(end-on)

Br-doped CH

Carbonfoam

Be tube

Au grid (side-on)Ti x-raybacklighter

foil

X-rayframingcamera

Figure V-1. Side view of the jet experimentalarrangement, shown end-on to the hohlraum.The experimental package (target) is abutted toa gold hohlraum. The target is composed of aberyllium tube in which bromine-doped plasticis joined to a carbon foam cylinder with ahemispherical feature on the interface. A large-area titanium x-ray backlighter provides back-illumination for an x-ray framing camera, whichimages the materials in transmission.

Section V 38 Hydrodynamics

Figure V-2. Microscope photographs of thecylindrical carbon foam piece, showing thediamond-turned hemispherical feature. Thehemisphere radius is 150 µm.

The temporal sequence of events in theexperiment is as follows. Eight of the Nova laserbeams are focused into the hohlraum, whichheats up rapidly to a temperature of 220to 240 eV. The bromine-doped plastic end ofthe experimental package, exposed to theinterior of the hohlraum, rapidly ablates. Theablation causes a strong shock wave to belaunched into the plastic. The shock travelsoutward through the experimental package, andencounters the density interface between theplastic and the carbon foam. Upon crossingthe interface, the shock-induced accelerationproduces an instability between the twodifferent-density materials, seeded by the150-µm-radius hemisphere.

Shortly before the time chosen to image theflow, the ninth and/or tenth Nova beams areused to illuminate an x-ray backlighter foil.Titanium and iron are typically used as foilmaterials; they produce 4.7- and 6.7-keV x-rays,respectively.

The x-rays are used to radiograph theexperimental package sideways, using either atwo-dimensional framing camera or an x-raystreak camera. The brominated plastic is moreopaque to the backlighter x-rays than theundoped plastic, so it appears darker in theimages. Shocked, or compressed, material alsoappears darker because of its larger density.

When the opacities of the variouscomponents are known, the transmission can beunfolded to provide material distributions. Thisknowledge of the opacity is central, as it is one of

the fundamental limitations of the experimentaltechnique for providing quantitative data. Thediagnostic technique is discussed in more detailin Hammel et al.1

The passage of the shock from the plasticablator section over this “bump” generates the jetby the same mechanism as the Richtmyer-Meshkov instability. As the jet propagates, itexperiences shear on its edges, leading to aKelvin-Helmholtz roll-up.

Experiments were conducted that imaged thejet at 19.5 ns and 25.8 ns after the initiation of thedrive beams. Comparison of the data andsimulated radiographs from two-dimensionalhydrodynamic simulation are shown in Figs. V-3(for 19.5 ns) and V-4 (for 25.8 ns). The dataquality is limited by several factors, including thesmall size of the jet, the short gating time of theavailable detector, and the availability of only asingle backlighter beam at late time (that is, at atime greater than 8 ns), when these data wereacquired. The picture at 25.8 ns is among datawith the longest time delay between firing themain laser and taking an image that was evercarried out on Nova.

The comparison between the data (left) andthe calculation (right) for early time is shown inFig. V-3. The data shows two dark bands,associated with the swirls, of similar size andlocation as the features in the simulation. Thecomparison at the later time shown in Fig. V-4 isnot as striking. The experimental picture shows asingle, wide, dark band below a large U-shapedfeature. The calculation displays much morestructure, and is a little narrower and moreadvanced down the tube. The effect of the shockscan be seen in the simulations, but not in thedata. It is noted that the comparison of “swirls”and other gross features in the experiments is notcommensurate with the need to obtainquantitative data on the underlying physicalprocesses involved in the hydrodynamics.However, the development of these techniques,which are extremely complex, can be undertakenin preparation for future quantitative work.

Figures V-3 and V-4 show that radiography isquite advanced, but also (and most important)that there is a real need for the advantages to beaccrued from the NIF, as these will allow much


Figure V-3. Comparison of experimental datawith simulation at t = 19.5 ns (early time) for thejet experiment shown in Figs. V-1 and V-2. Onthe left is the x-ray image showing the evolvingjet. On the right is the simulated x-radiographshowing the features of the jet. The comparisonshows relatively good agreement between theactual data and the simulation.

better data to be obtained. The relaxation of sizelimitations, the duration of drive pulses, and theuse of numerous beams for backlighting, all to beavailable on the NIF, will make these types ofexperiments more realistic.

Shock Planarity

To investigate possible methods for makingshocks planar for future studies, a number ofhohlraum/package variations have been tested.These illustrate the flexibility of the high-energylaser experiments and also some of thelimitations of the current methods.

An experimental package was constructedfor the shock planarity experiments. The packageused a thickness of 150 or 300 µm. Eight beams ofthe laser were focused onto the inside of a goldhohlraum to create an x-ray drive. That driveimpinged on the breakout target, and an opticalstreak camera was used to view the package fromits end, rather than its side. With this scheme, theabsolute time of the break and the uniformity ofthe shock that traverses the target can bemeasured. See Fig. V-5 for a schematic ofthe experiment.

The principle of the experiment is that shockbreakout through the rear surface of the packageis accompanied by heating, and the resultingradiation is detected by the optical streak camera.If the shock is curved with the central portion

Figure V-4. Comparison of experimental datawith simulation at t = 25.8 ns (late time) for thejet experiment shown in Figs. V-1 and V-2. Onthe left is the x-ray image showing the evolvingjet. On the right is the simulated x-radiographshowing the features of the jet. The comparisonshows agreement that is not as striking as forearlier time (Fig. V-3).

4 Novabeams 4 Nova

beams

Au hohlraum

Breakouttarget

Optical streakcamera

Narrowbandpassfilter

Figure V-5. Schematic of a shock breakoutexperiment to measure the planarity of a shockfront. Eight beams of the laser are focused ontothe inside of a gold hohlraum to create an x-raydrive. That drive impinges on the breakouttarget, which is imaged onto a streaked opticalpyrometer (streak camera), viewing the packagefrom its end rather than its side.

leading when it reaches the rear surface, thebreakout occurs first at the center then proceedsradially as the rest of the shock front reaches thesurface. When the optical streak camera is used toimage a diameter of the rear surface, the resultingstreaked image of a curved shock is also curved.


Two of the candidate mounting schemestested to determine the effects of mounting on theplanarity of a shock are shown in Figs. V-6, andtheir corresponding data from the streakedoptical pyrometer are shown in Fig. V-7. Onescheme, the standard scheme, consists of simplyabutting the package against the flat of thehohlraum, over the hole. The second is a versionof a design that is called a “tophat” because of itsappearance. The standard scheme will beshadowed by the hohlraum wall, whereas thetophat mount provides for a uniformillumination of the target.

It is clear from the data (Fig. V-7) that thebreakout in the standard mount is not flat—thebreakout takes over 1.7 ns from first arrival to theedges of the package. The tophat mount, on theother hand, has a very flat shock that breaks outover the entire diameter in 400 ps, with breakoutover the central 80% occurring in only 100 ps.Note that the central region, on the position scalebetween 0.2 and 0.6, represents the breakout data,while the bright sidebands are artifacts of theexperiments, not related to the shock breakout.

CH + Br

Standard mounting scheme

Hohlraum wall

CH + Br

Tophat mounting scheme

Hohlraum wall

Figure V-6. Two candidate mounting schemesto study planarity of the shock generated byx-ray drive impinging on the brominated plasticfrom above. The upper scheme is the standardscheme, where the package is simply abuttedagainst the flat of the hohlraum, over the hole.The second scheme is called a “tophat” mountbecause of its appearance.

Tim

e (n

s)

Tim

e (n

s)

–1.0 0.0 1.0 2.0

4

5

6

7

Position (mm)–1.0 0.0 1.0 2.0

3

4

5

6

7

Standard Tophat

Position (mm)

Figure V-7. Breakout data for the two different mounting schemes, standard and tophat, shown in Fig.V-6. Note that the central region, on the position scale between 0.2 and 0.6, represents the breakoutdata, while the bright sidebands are artifacts of the experiments, not related to the shock breakout. Thebreakout in the standard mount is not flat—the breakout takes over 1.7 ns from first arrival to theedges of the package. The tophat mount, on the other hand, has a very flat shock that breaks out overthe entire diameter in 400 ps, with breakout over the central 80% occurring in only 100 ps.


B. Stable Hydrodynamics

Planar Geometry

A series of experiments has been performedto investigate the hydrodynamic motion ofplasmas. The goal is to design an experimentwhere the hydrodynamics could be rigorouslyinvestigated without other complicating factors.In particular, the desire is to avoid any problemsarising from laser–plasma interactions orstrength-of-material complications. Therefore, itis considered essential to deposit the initialenergy in a well-defined manner. Theseconsiderations led to the use of a short pulse ofpenetrating x-rays to provide the initial heatingfor the samples.

The x-rays were produced by focusing one ormore of the beams of Nova onto a thin gold foilapproximately 2000 Å thick. The laser pulselength was 1 ns long. It was found that about 10%of incident laser light was converted into goldM-band x-rays with a photon energy between2 and 3 keV (see Section VIII, Radiation Sources,for more information on the x-ray sources).

To filter out x-rays of lower photon energy(radiation below the critical density for solidmatter), a 50-µm-thick plastic foil was placed at

a distance of 200 µm from the gold foil betweenthe gold x-ray converter and the sample to beheated. We placed a 3-µm aluminum foil at adistance of 1000 µm away from the gold foil. Thisfoil absorbed approximately half of the incidentgold M-band radiation.

The heating spectrum has been accuratelycharacterized using the x-ray filtered diodesinstrument at Nova. This is a ten-channeldetector that uses absolutely calibrated filteredx-ray diodes. X-ray crystal spectrometers werealso used to further spectrally characterize theradiation. With this arrangement an aluminumfoil was heated to a few eV. At thesetemperatures the x-ray absorption coefficients ofthe aluminum are very well known and energydeposition in the aluminum is well characterized.The schematic layout of these experiments isshown in Fig. V-8.

The expansion of the aluminum foil wasobserved by using a time-gated x-ray imager.A 2000-µm tantalum foil was placed perpen-dicular to the direction of the motion at a distanceof ~1 cm from the aluminum foil, thus allowingobservation across the foil. The tantalum wasilluminated by one of the Nova laser beams,which had been smoothed by a random phaseplate, and this produced a uniform x-ray

5 overlapped beams

800 µm200 µm

1500-Å Auburn-through foilconverts laser lightto x-rays

50-µm CHfilters out the low-energy x-rays Backlight beam

Large-area Ta backlight

Gated x-ray imager—GXI

3 µm

Al i

s vo

lum

etric

ally

hea

ted

by x

-ray

Figure V-8. Schematic of a test bed for hydrodynamic simulations. The aluminum sample isuniformly heated by the x-rays created from a 5-beam illuminated gold burn-through foil. The plasticfilter ensures that there is no radiation below the critical density for solid matter. The measurementsare made using a multiple-framed x-ray imaging camera (a gated x-ray imager) looking at thetransmission of the large-area backlight passing the target.


backlighter source. A filter between thebacklighter and the aluminum foil minimizedheating of the sample foil by the backlighter.Imaging was by a 10-µm pinhole array and agated microchannel plate x-ray detector with atotal spatial resolution of 15 µm. Four images canbe taken on each experiment, with a temporalresolution of better than 100 ps. The images aretypically taken at 1-ns intervals.

An example of the reduced data for one suchimage is shown in Fig. V-9. At approximately5 ns the aluminum foil and the plastic foil beginto collide, which is the limiting time for thisconfiguration. Comparison of these experimentswith calculations showed reasonable agreement,as was to be expected for such a simple geometry.One such comparison is shown in Fig. V-9. Thetheoretical prediction of the movement andexpansion of the foil would be critical tests ofthe EOS and opacity of the hydrodynamicsimulation. The agreement here is reasonableand would suggest more complicatedexperimental configurations could be tried. Withthese initial conditions verified, it will be possibleto proceed to more complicated geometries, such

as stepped targets, to provide further tests of thehydrodynamics.

Spherical Geometry

For the spherical geometry case, implosionimaging of ICF capsules, or microspheres, will bepresented. A typical implosion sequence isviewed with a 5.3× magnification spanning∆t = 1.4 to 1.95 ns. The instrument used to makethe measurement is a 12-frame gated pinholecamera denoted WAX. The active element of thismulti-frame x-ray pinhole camera consists of amicrochannel plate (MCP) overcoated with aserpentine (S-shaped) gold microstrip2 (seeFig. V-10).

The MCP gain is switched on and off by a170-ps FWHM 800-V pulse that travels along theserpentine strip. Because a voltage is required toaccelerate the electron produced to the rear of themicrochannel plate, this imager provides asimple means of obtaining temporal resolution bythe choice of voltage pulse shapes.

The WAX has been characterized in situ withshort x-ray bursts produced by irradiating golddisk targets with tightly focused, 100-J, 20-ps,

35

30

25

20

15

10

5

0

Den

sity

(m

g/cc

)

0.60.40.20.0-0.2-0.4-0.6

Distance (mm)

Experiment at 2.5 ns Theory at 2.5 ns

Figure V-9. An example of the data obtained from the experimental test bed for hydrodynamicsimulations. The agreement here between experiment and theory is reasonable and would suggestthat more complicated experimental configurations could be tried.


Target

Pinholearray

1-kV200-pspulse

Phosphor/fiber-opticfaceplate

25-Ω lines

Microchannelplate

12.5-Ωgold microstrip

Film pack

Figure V-10. The WAX, a 12-frame x-ray gatedimager. The instrument is capable of takingsequential images on a serpentine (S-shaped)gold strip coated onto a microchannel plate tomake it sensitive to x-rays. Because a voltage isrequired to accelerate the electron produced tothe rear of the microchannel plate, this imagerprovides a simple means of obtaining temporalresolution by the choice of voltage pulse shapes.

2-ω0 Nova laser pulses. The measured temporalresolution is 80 ±5 ps FWHM, the voltage pulsepropagation speed is 0.55 c, and the transversesensitivity variations are ≤10%.

The target consisted of a 470-µm diameterpolystyrene shell overcoated with a 55-µm plasticablator and filled with 50 atm deuterium, 0.1 atmargon, and 0.02 atm xenon. This fill, as is indi-cated below, is used to provide a spectroscopicindicator of the composition of the implosioncore. A 1-ns square, 17-kJ laser drive in a scale-1gold hohlraum was used. The interframe spacingis 45 ps, or 70 ps between frames separated by abend. The camera is filtered to pass emission onlyabove 3.5 keV, at the argon free–bound andxenon bound–bound photon emission energies.

Figure V-11 shows a typical implosionsequence for a 1-ns square drive pulse, using theWAX diagnostic. The image is filtered to showemission above 3.5 keV. In the figure, the capsule,which converges to 50 µm diameter, already

Time

Time

Time

Time

45 ps

Figure V-11. Typical implosion sequence for a 1-ns square drive pulse, using the WAX diagnostic. Theexperiment had a 17-kJ drive in scale-1 gold hohlraum-driven 470-µm diameter polystyrene shellovercoated with a 55-µm plastic ablator and filled with 50 atm deuterium, 0.1 atm argon, and 0.02 atmxenon. The image is filtered to show emission above 3.5 keV. The capsule, which converges to 50 µmdiameter, already appears in emission in frame 1 (at ∆t = 1.4 ns), peaks in emission in the 5th frame(∆t = 1.6 ns), and is disassembling by frame 11 (∆t = 1.9 ns). Residual diffuse gold M-band emissionfrom the cooling hohlraum plasma is also observed through the 50% laser entrance hole, decreasing inintensity with delay, as would be expected after the end of the laser pulse.


appears in emission in frame 1 (∆t = 1.4 ns), peaksin emission in the 5th frame (∆t = 1.6 ns), and isdisassembling by frame 11 (∆t = 1.9 ns). Residualdiffuse gold M-band emission from the coolinghohlraum plasma is also observed through the50% laser entrance hole, decreasing in intensitywith delay, as would be expected after the end ofthe laser pulse.

Quantitative results for a similar shot with ahigher drive of 29 kJ are presented in Fig. V-12.This experiment used a 1-ns square pulse, 29-kJdrive in scale-1 hohlraum. Time = 0 ps, which isdefined as peak emission time, occurs 0.3 ns afterthe laser pulse is switched off.

In Fig. V-12, A) shows the evolution of theFWHM of the emitting regions in two orthogonalplanes. The 10% m = 2 asymmetry observed maybe due to low-Z patches near the hohlraumequator creating a few percent lower flux at thetarget equator. B) is a plot of the evolution ofthe intensity of the imploding core emissionand hohlraum plasma emission. Afterdeconvolving the instrument time response,the FWHM of the x-ray emission from theimplosion is measured at 150 ps, in agreementwith streaked spectral measurements that havehigher temporal resolutions.

C. Unstable Hydrodynamics

Planar Geometry

There are numerous possible areas ofinvestigation for the study of planar instabilitiesin high-energy-density matter. The two that arediscussed here should only be considered a starton two aspects of the problem, and not thedefining experiments. In the first case we discussthe growth of an initially imposed perturbationunder the influence of a continuous acceleration(i.e., the Rayleigh-Taylor case), while in thesecond case we treat an interface that has randomperturbations and is shocked (i.e., theRichtmeyer-Meshkov case).

Imposed Perturbation—Rayleigh-TaylorThe first experiment shown is one intended

to study large-growth Rayleigh-Taylor instabilityusing a radiation source as the driver.3 See

0

20

40

60

80

100

120

–200 –100 0 100 200 300

Time (ps)

Horizontal FWHMVertical FWHM

400

FW

HM

(µm

) xx

x xx

x x x

x

x x

x

A) X-ray emission

X-r

ay y

ield

Background

0.4

Implosion

0

0.1

0.2

0.3

–200 –100 0 100 200 300

Time (ps)

400

X

X

X X

X

X

X

X

X

X

B) Background emission intensity

Figure V-12. Time history of an implosion.A) Evolution of spatial FWHM of x-ray emis-sion in two orthogonal planes. B) Plot of theevolution of target and hohlraum back-groundemission intensity. This experiment uses a 1-nssquare pulse, 29-kJ drive in scale-1 hohlraum.

Fig. V-13 for a schematic of the experiment. Inthis type of experiment, planar foils offluorosilicone are accelerated by x-ray ablation.The foil trajectory is measured using edge-onradiography to check the bulk movement of theablating sample.

In separate experiments, which employedface-on radiography, contrast in the optical depthis measured as a function of time. From thiscontrast, the evolution of the 50-µm wavelengthinitially impressed sinusoidal perturbations canbe deduced. Measurements of the growth of the


Backlighter beam

Drive beams

Backlighter disk

Io Through hole

Accelerating foil

Filters

X-ray imager

Hohlraum

Drive beams

Figure V-13. Schematic of a setup to study large-growth Rayleigh-Taylor instability. The schematicshows the face-on radiography of a target that has been impressed with a sinusoidal pattern. Thetarget is mounted on the wall of a cylindrical gold hohlraum with the surface perturbation facinginwards. The laser beams enter the hohlraum via holes in the end, generating an x-ray drive. As thefoil accelerates by x-ray ablation towards the x-ray imager, another laser strikes the backlighter disk.This generates a back illumination of x-rays, which pass through the hohlraum and the acceleratedfoil. Modulations in the foil areal density translate to modulations in exposure at the imager.

perturbation were made using both an x-raystreak camera, which shows the continuous one-dimensional growth of the perturbation duringthe experiments, and a two-dimensional imagewith a frame time of 100 ps. The two-dimensionalimage shows the full foil used, in order to verifythe planarity of the foil.

Figure V-14 A) shows the streak cameraimage, with time shown in the vertical andposition across the foil shown in the horizontal.The image shows the optical depth modulation ofthe accelerated foil and indicates that the foilbecomes, overall, increasingly bright due tothinning. The thinning is due in turn to theformation of the “bubble-and-spike” shapecharacteristic of the nonlinear Rayleigh-Taylorregime. Figure V-14 B) shows a two-dimensionalimage of the foil, taken at 2.6 ns, for a frameduration of 100 ps. This data is essential to ensurethat there is no transverse distortion of the foil.

The quantitative data is obtained by takingintensity traces transverse to the grooved

structure at different times. In Fig. V-15 we showa sample of such a series of traces. Note that thecurves, which represent different times in theevolution, are offset vertically for ease ofcomparison. At early times the contrast is smalland still sinusoidal, indicating that the Rayleigh-Taylor instability is in the linear regime. Late intime, the contrast is larger and distinctly non-sinusoidal, exhibiting the bubble-and-spike shapecharacteristic of the nonlinear Rayleigh-Taylorregime. The rapid flattening of the modulationsin the top two curves results from the burn-through when the bubbles break out of the backside of the foil. At this point, the spikes are stillbeing ablated away, but no longer can be replen-ished by matter flowing down from the bubbles.

These experiments have been extended fromthe single mode example shown here to multiplemode experiments,4 and to buried interfaces withimposed mode structures.5 The limiting case of arandom set of perturbations at a buried interfacerequires a somewhat different technique.


A) One-dimensional image B) Face-on two-dimensional image

PositionPosition

Tim

e

Pos

ition

Figure V-14. Face-on streak-camera image of the backlight source absorbed by the foil with initialperturbation. A) A one-dimensional streak camera image with time increasing in the vertical directionand position across the foil in the horizontal direction. The end of the bright spot occurs when thebacklight ceases to radiate effectively. B) The face-on two-dimensional image of the foil indicatinglittle, if any, distortion of the foil. This image was taken with a 100-ps frame time and occurs at 2.6 nsafter the start of the drive.

2.7

1.9

2.3

1.5

1.1

0.7

0.3

–0.1

t (ns)

50-µm increments

Burn-through

Spike Bubble

Non

-line

arLi

near

ln (

expo

sure

)

Figure V-15. Intensity traces for the acceleratedfoil with an initial perturbation of 50 µm wave-length and 1.9 µm amplitude sinusoidallyimposed on the surface. The curves are offsetin the vertical direction to allow simplecomparison. The backlight intensity variationacross the foil and as a function of time is notedby the dashed line at the initial time and laterin the pulse.

Imbedded Random Surface—Rayleigh-TaylorThe second case we discuss under planar

unstable flow concerns an imbedded interfacewith random perturbation and an impulsiveshock acceleration. Here the technique of pointprojection spectroscopy is used. The experimentis illustrated in Fig. V-16. A planar target with anablator section and a foil section, shown inFig. V-16, is radiatively accelerated normal to itssurface. The drive was produced by focusing6−10 kJ of 0.35-µm light in 1 ns from eight of thebeams of Nova into a hohlraum. A planar shockof ~60 Mbar pressure with a rise time of less than50 ps propagated through the layered target,accelerating it to about half of its final velocityand driving an instability between the ablatorand foil materials.

See Fig. V-17 for a schematic of the geometryof the point projection spectroscopy using abacklight. After a suitable delay, the target wasbacklit by a flash of x-rays produced by focusinganother beam, with 200 J of 0.35-µm light in apulse length of 100 to 200 ps, onto a 22-mmcoated fiber. The fibers were coated with 1-µmthick co-sputtered mixtures of gold, bismuth, andplatinum, chosen to give bands of x-ray emissionin the 2- to 3-keV region. Because the flash camefrom a small source, it will project a shadow ofthe target. However, it was first Bragg-reflected


200 µm

60 µm

1.5 µm

Sulfinated plastic ablator

Mo foil

Figure V-16. Diagram of target for studyof imbedded random surface using pointprojection spectroscopy. The target consistsof an ablator section of low-density sulfinatedplastc and a high-density molybdenum section.The dimensions of the low-density sulfinatedplastic/high-density molybdenum targetare shown.

from a potassium pthallate crystal before it wasused to project a shadow of the target in motiononto x-ray film. Since the crystal reflects onlywhen Bragg’s law is satisfied, the shadow hasspectral as well as spatial resolution in one direc-tion, the x direction in Fig. V-17. Spatial resolu-tion is also obtained in the z direction, normal tothe disk, so that the z distance moved by variousmaterials in the package could be measured.

The materials in the target are chosen to havephotoabsorption edges in the wavelength rangeof the shadow of the package—e.g., the sulfur Kand the molybdenum LIII and LII shown inFig. V-18. The measured change in absorptionacross the edge and the known increase inopacity allow the areal density of the material tobe evaluated. Mix of the target materials isdetectable by the overlap, if any, of thephotoabsorption edges. This is schematicallyillustrated in Fig. V-18, where the no-mix targetshows minimal density mismatch at the interface,while the mixed case shows a density mismatch.

A serendipitous feature of this technique isthat high-opacity absorption lines form close tothe cold edges as material is heated and ionized,acting as a sensitive, albeit qualitative, indicatorof the position of the materials.

The accelerated target consisted of a 200-µmdiameter, ~1.5-µm thick disk of molybdenumaccelerated by a 60-µm thick, lower densityablator of sulfinated plastic. Manufacturingresulted in a finish on the molybdenum/plastic

KAP crystal

Hohlraum x-ray drive

Backlight fiber

XZ-motion

Package

Backlighting laser

To

spec

trom

eter

Figure V-17. Schematic of point projectionspectroscopy technique, adapted for study ofthe mixing of interfaces. The layered target isaccelerated by the x-ray drive from a hohlraum,driving an instability between the ablator andfoil materials. After a suitable delay, a pointbacklight is created, and the backlight x-rayspass through the accelerated sample. Thebacklight x-rays impinge on a crystal, whichprovides spectral dispersion. The dispersedx-rays are then imaged on x-ray film.

interface of typically 0.1 µm rms, with randomwavelengths of 1–5 µm. The high initial-densitymismatch in this experiment, measured by theAtwood number, was ~0.8 at the ablator–foilinterface, and was derived from the densitymismatch between the molybdenum (10.4 g/cc)and the ablator plastic (1.36 g/cc).

Most of the detailed information about thetarget is obtained from the absorption spectra ofthe materials in the target. Figure V-19 is anexample of an image in which the target wasbacklit 7 ns from the start of the 1-ns drive. Theimage shows absorption as a function ofwavelength and position in the horizontal (x)direction, and as a function of position only in thevertical (z) direction. The target image capturedin flight can clearly be seen to have movedupward, away from the edge of its support. Aspatial fiducial made of 25-µm wire provides areference for the position of the package whenthe radiograph is taken. It also demonstrates the20-µm source size which, when combined withthe 10-µm motional blurring, gives a limitingspatial resolution of ~30 µm.


S K-edgeMo LII & L III edges

Mix widthAblator

Foil

For mixing

hν and X

Z

For no mixing

Figure V-18. Spectrometer images from point projection spectroscopy of imbedded random surface.On the left are the spectral signals for an unmixed (or no-mix) target, where little mix should occur,and on the right are signals for a mixed target. The film contains the x–z image of the target, with theadded advantage that the spectrum provides an indicator of where, in the z direction, the elements ofinterest are found. The materials in the target are chosen to have photoabsorption edges in thewavelength range of the shadow of the package (i.e., sulfur K and the molybdenum LIII and LIIshown). Mix of the target materials is detectable by the overlap, if any, of the photoabsorption edges.Note that the no-mix target shows minimal density mismatch at the interface, while the mixed caseshows a density mismatch.

The traces at z = 413 µm in Fig. V-19 show aclear sulfur K edge and a weaker molybdenumLIII edge. The middle trace at z = 463 µm showsboth sulfur K and molybdenum LIII edges,whereas the upper trace at z = 496 µm shows alarge molybdenum LIII jump and a small sulfur Kjump. These three traces demonstrate that the foiland the ablator are intermixed over a substantialspatial region.

The axial distribution of areal density ofthe materials of the package comes from the Kand LIII edge-jump measurements shown inFig. V.-19. If Tb(a) is the spectral transmissionbelow (above) the edge then the areal densityof the material along the viewing direction isgiven by

∫ρ(l)dl = ln(Tb/Ta) / ∆(µ/ρ)

where ∆(µ/ρ) is the change in the massabsorption coefficient across a photo-absorptionedge (2220–217 cm2/g and 1990–545 cm2/g forthe sulfur K and molybdenum LIII edges,respectively) and l is the viewing length.

From Fig. V-19, the diameter of the packageat the observation time is ~360 µm, which islarger than the initial 200-µm diameter because ofdecompression. Because the sulfur K andmolybdenum LIII edges are close to the center ofthe package in Fig. V-19, this gives an upper limiton l of 360 µm. The measurements that depend

on the sulfur K and molybdenum LIII edges of theaxial distributions of areal density of sulfur andmolybdenum are shown in Fig. V-20 A). Thelimitation on the lowest density measurable is~0.03 mg/cm2 for the penetration of thesulfinated plastic into the molybdenum. Thislimit arises from errors in the background, as theoverall transmission is high (~0.7). For themolybdenum penetration into the sulfinatedplastic, the lower limit on detectability is0.2 mg/cm2, which arises from the low overalltransmission, ~0.05. Clearly an overlap region of~150 µm exists that is much larger than thesystem resolution.

One potential problem in interpretationwould be the loss of planarity, due to eitherbowing or tilting of the target. However, thisapparent mixing of the ablator and the high-density foil is not due to bowing, as can beinferred from Fig. V-19. Nor is it due to tilting,as another view at 44° is used to check targetorientation. Other effects, such as the ablatorblowing by the molybdenum foil, could beproducing the apparent mix, so to minimize thispossibility control experiments were performedin which little mix should occur.

The target for these no-mix shots hadthe same ablator as the high-mix shots, ofdensity 1.36, but the molybdenum foil wasreplaced by a 15-µm chlorinated plastic foil,


Z, foil motion

Fiducial wire

A A'

B B'

C C'

300 µm

600 µm

1

A

A'

z = 496 µm0

0

1

B

B'

z = 463 µm

–100 0 100 2000

1

C

C'

z = 413 µm

x, distance from center of foil (µm)

Tra

nsm

issi

on

Mo LIIMo LIII S K

Figure V-19. Radiograph of a sulfinated plastic ablator driving a molybdenum foil, above, with threeintensity traces. Radiograph was taken at 7 ns with a backlight pulse length of 150 ps. The traces weretaken at three positions. Zero on the z (vertical) axis is the rear surface of the package at t = 0, and onthe x axis zero is the original center of the foil. The image clearly shows the sulfur K edge, themolybdenum LIII lines, and the edge of the package.

made from polychlorostyrene of density 1.23. Theinitial density mismatch is very small, having anAtwood number of –0.05, so little mixing isexpected. The mixing of the foil and ablator inFig. V-20 (B) is less than 40 µm, which isapproximately the experimental resolution.

Imbedded Interface—Richtmyer-MeshkovTo show another possible configuration that

is suitable for instability studies, we describe an

experiment that addresses techniques to studythe Richtmyer-Meshkov instability. The experi-mental configuration is the same as depicted inFig. V-1, but the density interface is planar, with-out the hemispherical protrusion. The inter-facehas a roughness machined into a high-densitybrominated plastic, and it is this roughness thatseeds the Richtmyer-Meshkov instability.

The experiments to develop the techniquesare in progress, but preliminary data is shown in


Fig. V-21. This image was taken about 8 ns afterthe initiation of the drive. The shock is movingfrom top to bottom, and the reference grids usedto provide spatial fiducials in the experiments arevisible on either side of the image. Note that thisimage does not take into account the spatialprofile of the backlighter, which is roughlyGaussian, and the effect of the cylindricalpackage geometry on the transmission.

In Fig. V-22 we show a 100-µm-wide verticaltrace taken on the centerline of the image inFig. V-21 (the jagged curve) together with a

0.8

Are

al d

ensi

ty (

mg/

cm2 )

0.6

0.4

0.2

0.0

S-plasticablator

Mo foil

A) High-mix case

z position (µm)200 300 400 500 600 700

200 300 400

S-plasticablator

1.5

1.0

0.5

0.0

Cl-plastic foil

B) No-mix case

z position (µm)

Are

al d

ensi

ty (

mg/

cm2 )

Figure V-20. Axial distribution of mass arealdensities for A) the high-mix case of Fig. V-19and B) the no-mix shot, both measured at thepackage centers. The mass areal densities of thesulfur in the plastic, the molybdenum, and thechlorine in the plastic are plotted. The fact thatthere are traces of the two differentspectroscopic markers in the same z-positionindicates that mixing has occurred.

Gaussian fit to the backlighter spatial profile (thesmooth curve). The fit to the backlight spatialdependence was found using the unshockedregion in the image.

The trace was normalized by the backlightspatial dependence to remove most of thebacklighter non-uniformity, and the results aredisplayed in Fig. V-23. In Fig. V-23 the levelregion from ~570 to 825 µm demonstrates that theGaussian profile is a reasonable approximation ofthe spatial shape of the backlight. The two arrowsdenote the locations of the calculated interfaceand shock locations.

–250.0 0.0 250.0 500.0

800.0

600.0

400.0

200.0

Radial position (µm)

Total absorption

Total transmission

Partial absorption

Axi

al p

osi

tio

n (µ

m)

Figure V-21. Gated x-ray image of planarRichtmyer-Meshkov mix target at about 8 ns.

0

5

10

15

20

25

30

35

-200 0 200 400 600 800 1000

Y (µm)

Exp

osur

e

Figure V-22. A 100-µm-wide vertical trace of thecenterline of the image in Fig. V-21 (thin jaggedline), and the accompanying fit to the backlight(thick line).


It is apparent from Fig. V-23 that a simulationcan reproduce the gross features of theexperiment. Sharp drops in the exposuredelineate the edge of the higher-opacity regionsin the shock-compressed foam and then in thebrominated-doped plastic. The width of thefeatures is the distance over which thesetransitions occur. For a perfectly flat shock, themeasured shock width should be negligible. Wemeasure a shock width of ≈15–20 µm, which is ofmarginal interest because the diagnosticresolution is 12–15 µm.

The NIF would provide larger scales andrelieve the severe constraints on spatialdimension that are faced with the present-dayhigh-energy lasers. On the other hand, thecalculations are in substantial disagreement forthe width of the mixed region for this interface.The two-dimensional simulations indicate a mixwidth of less than 15 µm, while the measurementindicates a width of 25–30 µm. Again, the NIFwill provide a test bed for experiments of thistype. In the interim, experimental techniques willcontinue to be developed to obtain a deeperunderstanding of the physical processes involvedin the experiments and how these limit ourability to measure the quantities of interest.

0

0.25

0.50

0.75

1.00

1.25

200 300 400 500 600 700 800 900

Axial position (µm)

Calculated shockCalculated interface

Figure V-23. Vertical trace of the image in Fig.V-21 normalized by the backlight.

Spherical Geometry

The most obvious method to study sphericalunstable flow is to attempt to make the usuallystable implosion unstable in a controllable way.To do this, the use of perturbed microsphereswas perfected. The microsphere, or capsule,shown in Fig. V-24 was a “bumpy” ball equippedwith micron-size perturbations imposed on theouter plastic ablator surface, with a chlorine-doped pusher (i.e., the material at the interfacewith the gas core) and argon-doped 50-atmdeuterium gas.

Fig. V-25 shows a single 12.5× magnification,12-µm resolution image of an implosion, usingthe bumpy ball, taken 0.2 ns after peak emission.The image was taken by the WAX instrumentdiscussed in Section III, ExperimentalCapabilities, Subsection A. Note that the figureshows the positions of the beams on thehohlraum wall relative to the image. Theexperiment used 1-ns square, 17-kJ laser drive

Figure V-24. Machined “bumpy” ball used inimplosions. The bumps, which are random, areplaced on the surface of the microsphere bylaser irradiation. (The cylindrical object to theupper right side is there to hold the microspherefor processing.)


in a 2550-µm-long gold hohlraum. A concavepentagonal shape, the m = 5 mode, is clearlyobserved. The nodes, which correspond toregions of lower flux, occur between laser spots.The m = 5 symmetry is observed on almost all1-ns square implosions.

In addition, Fig. V-25 shows evidence ofhigher emission from the edges of the capsulethan from the center, commonly defined as limbbrightening. The measured polar-averaged radialtrace in Fig. V-26 shows the analyzed data thatsuggests that the evolution of the pusher–gasinterface can serve as a measure of the degree ofmix of the gas and the pusher, and might bediagnosed by time-resolved limb imaging.

As a further monitor of the mixing of theinterface between the solid-density pusher andthe gas fill in an implosion, doping of the pusherwith a high-Z material can both provide infor-mation on the pusher conditions and act as adiagnostic for mixing of the pusher into the gasregion.6 Using the same type of bumpy ball asshown in Fig. V-24, observations have been madeof the emission from various dopants included

100 µm

Beams 5,6

Beams 3,4

Beams 1,2

Beams 9,10

Beams 7,8

Figure V-25. Expanded view of an implosionimage taken using the WAX. The image shows aconcave pentagonal (m = 5) asymmetry, as wellas limb brightening (higher emission from theedges of the capsule than from the center).

into the innermost region of the pusher. Fig. V-27shows x-ray streak camera data from bothbumpy and smooth microspheres. The dopant inthe pusher is 1% chlorine by number. Themicrosphere with the bumpy surface hassignificantly brighter chlorine emission andweaker argon emission than the microspherewith the smooth surface. This is an indicationthat some of the pusher material has been mixedinto the hot gas.

Further extensions of the technique forinferring mix into gas cores permit higher-ZK-shell emitters to be used, as these spectra aresimpler to interpret. Advances in target fabri-cation have expanded the choices of possiblepusher dopant to include several metals.Figure V-28 shows the results of another set ofexperiments with and without bumpy surfacefinishes, using iron as the dopant in the pusher.In the figure the microsphere with imposedbumps is shown to have a substantial helium-likeiron 1s2−1s2p emission feature, while the smooth-surfaced microsphere shows no evidence of theiron emission.

0

0.2

0.4

0.6

0.8

1.0

0 20 40 60Radius (µm)

Experiment

X-r

ay s

igna

l

Figure V-26. Measured trace quantifying thelimb brightening seen in Fig. V-25. Thecomparison is only qualitative because absolutetiming is not known and the target surfaceroughness is not fully characterized.


ClCl

He-αLy-α

Ar

Bumpy Smooth

Tim

e

He-αLy-α

Ar He-α

He-αLy-α

Figure V-27. Streak camera records of spectra from a radiatively driven microsphere using chlorine asthe dopant in the pusher, showing mixing of the pusher into the gas fill. The argon is mixed into thegas core. The left-hand spectrum shows the bumpy-surface experiment and the right-hand side showsthe smooth-surface experiments.

Fe He-α Fe He-α

Ni Kα

Tim

e

Bumpy Smooth

Ni Kα

Figure V-28. Streak camera spectra from a microsphere with an iron dopant in the pusher, showingmixing of the pusher into the gas fill. The argon is mixed into the gas core. The left-hand spectrumshows the bumpy-surface experiment, and the right-hand side shows the smooth-surfacedexperiments.

Finally, further improvements in thespectroscopic diagnostics have been made byusing complex spectral features such asUnresolved Transition Arrays (UTAs).7

However, theoretical further development of areasonable non-local thermodynamic equilibrium(NLTE) theoretical model of these UTAs requiresfurther development.


D. Future NIF Experiments

In the following, we discuss a number oftopics that are among the most promising forexperimental investigation under conditions thatcan only be created at the NIF. We haveorganized these topics according to the type offlow to be considered and the material conditionspresent. Thus, the issues to be considered thatinvolve stable flow are hydrodynamic flow athigh Mach numbers, shock–shock andconverging shock propagation, shock motionthrough inhomogeneous media, and shock–boundary interactions in the strong shock regime,including the relationship to cratering.

Unstable flow problems that are uniquelyaddressable with the NIF involve, for example,the study of the onset and development ofinstabilities such as Rayleigh-Taylor, Richtmyer-Meshkov, or Kelvin-Helmholtz. The occurrenceof these instabilities in blast waves, and insupernova explosions, where they play animportant role in mixing the products fromdifferent regions, can also be studied inlaboratory experiments. Other experiments thatcould possibly be considered concern the studyof radiation condensation and secondaryinstabilities. Finally, the onset of turbulence andthe influence of vortex dynamics in a medium offinite compressibility is an emerging field withimportant possibilities.

Stable Flow

This first section on stable flow will discussthe study of strong-shock hydrodynamics in thepresence of material boundaries,inhomogeneities, and another strong shock.These studies, including shock–shockinteractions, Mach stems, and possibly non-equilibrium gas dynamics analogies and similarphenomena, are all classical problems influid dynamics.

The possibility of performing preciseexperiments in this field is of great interestbecause of the prospect of benchmarkingnumerical models; understanding certain spacephysics, chemistry, and cratering phenomena;and checking exact solutions that are valid in the

strong-shock limit. The shock interactions wouldrepresent an extension of work done in shocktubes and similar facilities, but potentially atsignificantly higher Mach numbers.

The analogies with non-equilibrium flowswould depend upon the burn process of anignition target, and are not possible on presentlasers. The shocks for stable flow studies arecreated by the interaction of a strong radiationfield with a material. The interaction is created bydirect drive, indirect hohlraum heating, oracceleration and subsequent impact of a flyerfoil.8 The transition from radiation energydeposition to hydrodynamic shock motion istherefore also a topic of considerable interest.Note that there are instabilities that can occur atthe material interfaces for some of the topicsdiscussed here, and these are related topics ofinvestigation that will be discussed separately inthe section on unstable flow.

Finally, the general area of magnetohydro-dynamics holds the potential of some interestingphenomena that could be investigated using theNIF’s capability to deposit large amounts ofenergy into small volumes or areas in very shorttimes. Possible topics for future considerationinclude generation of electromagnetic inter-ference, enhancement of magnetic fields throughcompression, and a “photo-voltaic battery.”

Shock–Shock InteractionsShocks created by a two-beam laser incident

onto aluminum targets that propagate towardeach other through a nitrogen atmosphere havebeen recently studied.9 Figure V-29 shows aschematic from an experiment to investigate theinteraction of two blast waves (see Fig. V-30 forthe data). In the experiment, two lasers arefocused onto two 2.5-mm diameter aluminumrods in a nitrogen atmosphere. The ends of theirradiated rods blow outward into the nitrogen,creating two shocks that interact in the medium.Density and temperature in the shock–shockinteraction region were determined by spectralline broadening and intensity ratio methods. Inaddition, an attempt was made to determine thephysical processes occurring as the shocks collideand then either pass through each other or reflect.


Laser

Laser

Al rod2.5 mm

N2

Expanding blast waves 30 mm

Figure V-29. Schematic of a shock–shockinteraction experiment. Two lasers are focusedonto two 2.5-mm diameter aluminum rods in anitrogen atmosphere. The ends of the irradiatedrods blow outward into the nitrogen, creatingtwo shocks that interact in the medium.Spectroscopy of NII and NIII transitionsprovides in-situ measurements of the electrondensity through line broadening. The timedependence of the electron density in theshock-shock interaction region is determined byobserving the midplane with a streak cameracoupled to a visible spectrometer.

Figure V-30 is a graph of electron density vstime in µs for two cases, a single wave (datapoints denoted with x) and a colliding wave(denoted with circles). The curve for the single-wave case illustrates the rapid onset of theshocked density enhancement and subsequentrapid decrease. The curve for the colliding-wavecase illustrates the density enhancement of theshock–shock interaction and the slow decay ofthe density after the shock.

On the NIF it should be possible to extendthese quantitative techniques to obtaining data atextreme shock conditions. Further, the NIF willmake it possible to access the increased shockstrength needed for different nonlinear regimes.

The interaction of strong shocks is a typicalnonlinear flow problem in shock hydrodynamicsand for extreme shock conditions, numericalsolutions have been the only source ofinformation in the past. Increasing the shockstrength will permit different nonlinear regimesto be accessed. A similar study, equallyimportant for hydrodynamic theory, is that of aconverging strong shock caused by focusedacoustic energies such as those formed in pellet

Ne

(101

8 c

m–3

)

0 1 2

0.1

0.2

0.3

0.4

0.5

0.6Colliding wavesSingle wave

Time (µs)

Figure V-30. Graph of electron density vs timein µs for two cases, a single wave and acolliding wave. The curve with the data for asingle shock wave (data points denoted with x)shows the rapid onset of the shocked densityenhancement and subsequent rapid decrease.The curve showing data from the collidingwave (data points marked with circles)illustrates the density enhancement of theshock–shock interaction and the slow decay ofthe density after the shock.

implosion or bubble collapse.10 This should alsobe possible with a large-spatial-scale laser-energy incident.

Shock–Boundary InteractionsThere are a series of possible experiments on

the interaction of a shock wave in a medium thathas high acoustic impedance and is bounded by amedium of low acoustic impedance. Examplesare underwater or underground blast waves thatimpact a planar surface with a gas, or a planarshock passing over a bubble. If the shockpropagates in the strong-shock regime, therarefaction generated either by the shock at thepoint of contact with a planar surface or by aplane shock at the point of contact with aspherical bubble can, in certain geometries,overtake and distort the impinging shock. This isanalogous to the transition from a regular


reflection to a Mach reflection for flow over awedge, except here the reflected wave is ararefaction instead of a shock. This effect hasbeen predicted,11 but the experimental evidencehas all been produced in the gas phase with less-than-strong shocks.12

Self–similar solutions show that the rarefac-tion generated at the interface by an expandingspherical shock can also overtake the shock bothalong the surface and straight down into themedium, so the effect is interesting both forspherical shocks generated by point sourcesand for plane shocks. An important applicationof these studies is in the understanding of range-to-effect of the shock waves created by under-ground or underwater plate tectonics. Anotherapplication is found in the possibility ofexperimentally measuring shock propagationin inhomogeneous media, which is relevantto the astrophysical problem of the hydro-dynamic interaction of shock waves withinterstellar clouds.13

Figure V-31 shows the results of simulationsof an interstellar cloud being shocked, illustrating

propagation of a shock in an inhomogenousmedium. The first frame shows the initial planarshock moving at velocity vb toward a cloud ofradius a0. The cloud is 10 times as dense as theinterstellar medium, and the shock in the cloudhas a Mach number of 10. As the shock transits it“crushes” the cloud, and in re-expansion acomplicated combination of instabilities occurs—Richtmyer-Meshkov due to the impulsiveacceleration of the initially perturbed cloud-intercloud boundary as well as Kelvin-Helmholtzand Rayleigh-Taylor instabilities.

At a time of 8 a0/vb (the second frame) thecloud has clearly suffered large-scale unstableflow. Importantly, by time 12 a0/vb (third frame)the cloud is distorted and axially flattened. Atthis later time there is a plume of fragments thatcontains 70% of the cloud mass. This process ofdistortion and axial flattening continues until thecloud is largely dispersed, with axial dimensionsreaching five time the original size and atransverse dimension reaching two times theoriginal size. This kind of experiment, incompressible inhomogenous media where

z/a 0

5

4

3

0 1 2r/a0

3

2

10 1 2

r/a0

1

0

–1vb

1r/a0

0

t = 0 t = 8 t = 12

Figure V-31. Propagation of a shock in an inhomogeneous medium is illustrated by these simulationsof an interstellar shock propagating through a cloud. The first frame shows the initial planar shockmoving at velocity vb toward a cloud of radius a0. As the shock transits the cloud it “crushes” it, and inre-expansion a complicated combination of instabilities occurs. At a time of 8 a0/vb (the second frame)the cloud has clearly suffered large-scale unstable flow and by time 12 a0/vb (third frame) the cloud isdistorted and axially flattened. This process of distortion and axial flattening continues until the cloudis largely dispersed, with axial dimensions reaching five time the original size and transversedimension reaching two times the original size.


radiation plays a role and time scales fordevelopment of the full complement ofinstabilities are possible, requires the scale of theNIF, where energy is available to provide notonly high-Mach-number shocks butdiagnostic capability.

Hypersonic FlowHypersonic flow problems with Mach

numbers above 20 can be studied with laser-produced plasmas. This is a regime unattainablewith present-day wind tunnels or shock tubes.Three-dimensional flow at these speeds isassociated mainly with the reentry of spacevehicles or the flight of a projected ramjetairplane. Theoretical calculation of this flowinvolves difficult questions of flow pastparticular surface shapes (noting that cones14 areof particular interest, of course). The chemistry ofthe wakes created by a body moving through afluid with hypersonic speed is important for theunderstanding of the effects of flight or reentrythrough the atmosphere, particularly withrespect to the creation of plasmas or the inter-action with constituents such as the ozone layer.

Impact CrateringMany phenomena of interest in impact

cratering occur on temporal and spatial scalesthat are large compared to those of the impactingobject. As a result, the impact can be modeled asa point source of high energy and momentumdensity and the mass of the object can beneglected.15 This situation can be modeled by thedeposition of focused laser energy in smallspheres of high-Z material.

Alternatively, flyer foils that can generateprodigious shocks and post-shock pressures—onthe order of gigabars—could be used. Thus,although the formation of a crater by the impactof a meteor on a surface seems to be entirelydifferent from the formation of a crater by asurface blast, there are possibilities for using thismethod to model impact craters.

A study of concentrated impacts onsurfaces16 shows that scaling laws apply to theshape of both craters formed by impact and thoseformed by surface energy deposition. Therefore,the cratering created by an intense radiationsource or flyer foil would be useful in studying

the scaling of the crater shape and inferring theshape of impact craters.

A series of proof-of-principle experimentsusing focused laser energy to deposit energy in agrout mixture (which simulates the soil sample)below a planar interface with the atmosphere hasbeen performed,17 with results that suggest thatfurther research in this area with higher energydensities would be of great interest. Further, itshould be pointed out that the effect of gravitywas also simulated, as well as geological layeringeffects in separate experiments. The experimentalschematic of these experiments is shown inFig. V-32. The experiment was set up to explorethe hydrodynamic response of a simulated soil.

The experiment demonstrates several facts:• The energy of the laser is captured in a

small volume inside the grout block usedto simulate soil.

• The mass and energy involved in thedeposition did not escape through thelaser entrance hole.

• The energy density provided by 4 kJ oflaser energy in 1 ns is sufficient tovaporize the gold target and thesurrounding grout.

The ability of the laser to deposit largeamounts of energy in a small volume withoutcreating residual gases (which would be abyproduct of the same experiment performedwith high explosive) indicates its utility as asimulation source. On a NIF-sized laser, the

Au hollow sphere

Soil sample

Laser 14˚

Figure V-32. Setup to explore the hydro-dynamic response of a simulated soil. Theexperiment demonstrates employs one laserbeam focused onto the rear of the gold hollowsphere embedded with soil, or grout, sample.


amount of energy deposited and the in-situdiagnostic potential make it possible toinvestigate hydrodynamic response possible inreal-time experimental configurations.

Figure V-33 shows images showingthe results of the experiment to simulatehydrodynamic response of soil described inFig. V-32. Image (A) is a photograph of the postexperiment box of grout formed of a 16 x 16 x 16cm cube, built up of squares of 3/16-inchaluminum plate. The initial cavity had a 1.5-mmradius and was buried 6 cm deep. It capturedapproximately 4 kJ of energy from the laser,enough for the requisite vaporization of the goldsphere. Less than 200 J of laser radiation wasobserved to escape the target.

The top surface is crazed, as seen in imageA), and slightly bowed up. Flash radiographswere taken, and image B) shows a top view,clearly showing the cavity and the entry cone.Note the profusion of radial cracks and the faintbut definite indication of tangential (spherical)cracks. The final cavity was approximately 2 cm,which is consistent with cube-root scaling fromstrong explosion data. This diagnostic is an

example of the detail that can be learned fromscaled experiments that would be very difficult,if not impossible, to get from full-scaleexperiments.

Scaled Radiative Energy CouplingThe coupling of explosive energy into

hydrodynamic motion is quite differentdepending on whether the energy originatesfrom a chemical reaction or from the impact of aflyer foil, or is deposited by a radiative energysource. The study of the physical mechanismsinvolved in the latter case (the coupling of energydeposited by a strong radiative source above, on,or below the surface of dense matter and fluids)is of special interest. This energy coupling variesfrom zero for sources at large heights above thesurface to 100% for deeply buried sources.

The shape of the coupling curve as a functionof source position, of course, has to do withquestions of range-to-effect and cratering, but isalso of scientific interest for the study of the exactmechanisms determining the coupling. It hasbeen proposed that the coupling of energyproceeds through three stages: first, the stripped-

A) B)

Figure V-33. Experimental results of hydrodynamic response of soil . Image A) shows the postexperiment box formed of a 16 x 16 x16 cm cube, built up of squares of 3/16" aluminum plate. Theinitial cavity had a 1.5-mm radius and was buried 6 cm deep. It captured approximately 4 kJ of energyfrom the laser, enough for the requisite vaporization of the gold sphere. Less than 200 J of laserradiation was observed to escape the target. The top surface is crazed, as seen in image A), and slightlybowed up. Flash radiographs were taken, and image B) shows a top view, clearly showing the cavityand the entry cone. Note the profusion of radial cracks and the faint but definite indication oftangential (spherical) cracks.


sphere stage of complete ionization, then thermalwave propagation, and finally a momentum-conserving transition to shock propagationand cratering.

A low-temperature thermal wavepropagation has been observed by monitoringthe ionization front produced by the irradiationof a foam material with an x-ray source producedby laser irradiation of a gold foil.18 Similarstudies have investigated shock propagation19

and the possible formation of a Marshak wave.20

A study of these mechanisms that play a role inultimately transforming the incident radiantenergy into kinetic energy of ground motion is ofgreat interest.

Unstable Flow

In this section, we will describerepresentative fluid dynamics experiments thatare designed to study the growth of flowinstabilities and are appropriate for study withthe NIF laser. Areas of unstable fluid dynamicsthat are readily, and uniquely, addressable by theNIF include the study of Rayleigh-Taylor,Richtmyer-Meshkov, and Kelvin-Helmholtzinstabilities in both planar and non-planargeometries, for the linear, weakly nonlinear, andturbulent regimes.

Some experience has been acquired inprevious experiments on this topic, but its largerscales, longer times, and longer drive duration,which are constant themes for hydrodynamicexperiments, make the NIF well-suited to extendthose investigations. See Subsections A, B, and Cfor descriptions of several state-of-the-artexperiments to illustrate how the NIF will affecthydrodynamics experiments.

These instabilities, and the ensuing mixingthat usually follows their development, are ofinterest because of their occurrence in inertialfusion, astrophysics, energy conversiontechnology, and the medical applications oflasers, to name a few areas. Another area ofunstable flow which has recently become ofsignificant interest is the study of compressibleturbulence and mixing in high-speed jets andshear layers. Recent work indicates thatturbulence may be qualitatively different at highMach numbers and/or high compressibilities.

The NIF will provide an excellent facility togenerate high-Mach-number flows, typicallyeither jets or shear layers, which may be used inthese investigations.

Classical InstabilitiesThe instabilities to be discussed in this

section are those that refer to the growth ofperturbations at a fluid interface. The type ofinstability in the flow is determined by the originof the perturbation. If it is related to the action ofshear or constant acceleration, the instability isgenerally referred to as Kelvin-Helmholtz orRayleigh-Taylor, respectively, and the growth isexponential in time. An impulsive accelerationcaused by a shock normal to the interface causesthe Richtmyer-Meshkov instability, which, in theabsence of shear or acceleration, then growslinearly in time.

These phenomena have been extensivelystudied in the past to understand the resultinginterface motion and material mix, which areimportant effects in inertial confinement fusion21

and for clumping in supernova explosions.22

Nevertheless, important questions remain,because often experiments do not agree with thetheory or with the numerical simulations incertain flow domains, especially at highcompression.23 Experiments on the NIF with longdrive times and large spatial scales are expectedto be especially important for the understandingof these instabilities.

Other InstabilitiesThere are a number of problems associated

with other instabilities, and with combinedinstabilities, that are also appropriate for studywith laser-driven flow. The most obviouscandidate is the instability associated with thepropagation of a strong shock through a uniformgas, which may play an important role in thestructure of supernova blast wave propagationand in the formation of stars and galaxies.Theoretical disagreements have generatedconsiderable controversy in the past, but therecent observation of these instabilities24 in alaser-driven blast wave indicates that moreresearch is needed on this topic. Figure V-34shows a schematic of an experiment performed toinvestigate an unstable Taylor-Sedov blast wave,


or shock, along with the data generated bythe experiment.

The experiment consists of a 200-J laserincident on a 6-µm thick plastic foil. The laser is5 ns in duration and is focused to an 880-µmdiameter spot. The irradiated plastic will ablateand expand at greater than 107 cm/s into thesurrounding fill gas. The gas fill is chosen tochange the adiabatic index, γ, from 1.3 to ~1.0 fornitrogen and xenon fills, respectively. B) showsan optical dark-field shadowgraph of the shockwave at 243 ns in the nitrogen fill. The image hasa frame time of 5 ns or less, and effectivelyfreezes the shock front. The image shows that theshock front, the dark circle, is uniform and thematerial behind the shock is homogeneous. In C),the xenon fill shock front shows a completelydifferent front, not uniform, with inhomogeneityin the material behind the shock. Theconsiderable controversy on this subject, theextension to denser systems, and the higherenergy densities available on the NIF will makethese types of instabilities ripe for the NIF.

The NIF is uniquely suited to producing thestrong shocks with large spatial extent that are

required for these studies. In addition, combinedRayleigh-Taylor, Kelvin-Helmholtz, andRichtmyer-Meshkov instabilities have beentheoretically described in cases where an obliqueshock is preceded or followed by an accelerationand/or shear.25 Such combined instabilities,which occur whenever horizontal and verticalshear are simultaneously significant, areimportant for describing flow in winds and oceancurrents, and for the fanning of smoke plumes.Experiments that could cleanly examine thedominant features of these phenomena are highlydesirable and well-suited to the larger-scale flowsthe NIF will be able to generate.

There is a wide class of two-dimensionalflows that become unstable in the transition tothree-dimensional flow. These secondaryinstabilities occur because two-dimensional finiteamplitude waves can be exponentially unstableto three-dimensional perturbations. This isanother classical fluid dynamics problem. Thestudy of these instabilities by Poiseuille or planeCouette flow has generally been performed in aregime where material compressibility has noinfluence. The effect of material compressibility

1 cm

6 µm CH

N2 or Xe

Gated optical imager

A)

N2 Xe

B) C)

Laser

Figure V-34. Experimental setup and data on the instability of a Taylor-Sedov shock. In A) theschematic of the experiment shows a 200-J laser incident on a 6-µm thick plastic foil. The laser is 5 nsin duration and is focused to an 880-µm diameter spot. The irradiated plastic will ablate and expand atgreater than 107 cm/s into the surrounding fill gas. The gas fill is chosen to change the adiabatic index,γ, from 1.3 to ~1.0 for nitrogen and xenon fills, respectively. B) shows an optical dark-fieldshadowgraph of the shock wave at 243 ns in the nitrogen fill. The image has a frame time of 5 ns orless, and effectively freezes the shock front. The image shows that the shock front, the dark circle, isuniform and the material behind the shock is homogeneous. In C), the xenon fill shock front shows acompletely different front, not uniform, with inhomogeneity in the material behind the shock.


on the development of these secondaryinstabilities is, in itself, of great interest.Experiments with laser-driven flow will enablethe compressible regime to be investigated, andthis will have important implications for theunderstanding of wall-bounded shear flow.

Finally, a significant subject for investigationis radiative condensation instability, a classicalastrophysical problem with applications toTokamak plasmas and z-pinch plasmas as well.This instability is a general feature of the radia-tive cooling of an optically thin plasma. This is, ofcourse, easily observed in interstellar and inter-galactic clouds, solar prominences, and Tokamakor z-pinch plasmas, all of which have cool, denseplasma structures that radiate into hotter, rare-fied plasmas in the surroundings. Experimentsto provide a greater understanding of thisinstability are feasible in the controlled condi-tions characteristic of laser-generated plasmas.

Turbulent Flow and Vortex DynamicsBoth improvements in computational

techniques and the emergence of chaos theoryhave contributed to a renewed interest in thesubject of the transition to turbulent flow. Inaddition, the study of compressible turbulentshear layers has recently been the subject ofincreasing interest in the aeronauticalengineering community.26

The understanding of compressible turbulentshear layers, which differs considerably from themore familiar transition to turbulence forincompressible flow, is of key importance in thefield of scramjet engine design. An experimentalprogram in this area would increaseunderstanding of the applicability of the coherentstructure models that have been developed forincompressible shear layers in the past. There isalso a great need for experimental evidence oflarge-scale structure in compressible shear layers,comparable to that found in incompressible cases.It is for this reason that the larger spatial scalesassociated with NIF experiments could beimportant for this developing field of study.

Vortex dynamics27 are also an important partof this field, since they are associated with thefundamental mechanism for the evolution ofhydrodynamically unstable surfaces toturbulence. The generation of vortices can be

simulated in great detail, and the results of onesuch simulation are shown in Fig. V-35. Thefigure illustrates the results of a simulation of thegeneration of vortices in an incompressible flow,showing the density contours from a simulationof a planar wave. The wave is impinging on aninterface that is at 60˚ to the wave velocity. Thecase is the classically unstable case where thedensity of the medium of the initial wave is afactor of three smaller than the density of theshocked material. The Mach number of the shock

x0 180

A) Initial position and time of 91

0

y

62

0

y

62

x135 315

B) Late time of 620

Figure V-35. The results of a simulation of thegeneration of vortices in an incompressibleflow, showing the density contours from asimulation of a planar wave. The wave isimpinging on an interface that is at 60° to thewave velocity. In (A) the initial position of thewave and the interface are shown as graydashed lines. Further, the scaled time for (A)is 91, when the shock has just passed the initialinterface. Here the vortices are not yet formed.In (B), which is at a late time of 620, the vorticesare fully formed. At present these types ofinvestigations are limited to incompressibleflows; however, compressible-flow experimentswill be possible with the large scales associatedwith NIF experiments.


is 1.2 and the units of the calculations are scaledso that the sound speed in the initially movingmedium is unity.

In (A) the initial position of the wave and theinterface are shown as gray dashed lines. Further,the scaled time for (A) is 91, when the shock hasjust passed the initial interface. Here the vorticesare not yet formed. In (B), which is at a late timeof 620, the vortices are fully formed.

At present these types of investigations arelimited to incompressible flows; however,compressible-flow experiments will be possiblewith the large spatial scales associated withNIF experiments.

E. References

1. B. A. Hammel, D. Griswold, O. L. Landen,T. S. Perry, B. A. Remington, P. Miller, T. A.Peyser, J. D. Kilkenny, Phys. Fluids B 5, 2259(1993).

2. P. M. Bell et al., in “Ultrahigh- and High-Speed Photography” and “Videography,”SPIE Proc. (SPIE, 1989), Vol. 1155.

3. B. A. Remington et al., Phys. Fluids B 4, 967(1992).

4. B. A. Remington et al., Phys. Fluids B 5, 2587(1992).

5. G. Dimonte and B. A. Remington, Phys. Rev.Lett. 70, 1806 (1993).

6. B. A. Hammel et al., JQSRT 51, No. 1 (1994).7. See, for example, J. Bauche et al. in Advances

in Atomic and Molecular Physics, D. Bates andB. Bederson, Eds. (Academic Press, 1988),Vol. 23, p. 131.

8. R. Cauble, D. Phillion, R. Lee, and T. Hoover,JQSRT 11, 433 (1994).

9. R. Elton, D. Billings, C. Manka, H. Griem,and B. Ripin, Phys. Rev. E 49, 1512 (1994).

10. C. Wu and P. Roberts, Phys. Rev. Lett. 70, 3424(1993); Setchell, Storm, and Sturtevant,J. Fluid Mech. 56, 505 (1972).

11. M. Holt, Ann. Rev. Fluid Mech. 9, 187 (1977);M. Kamagai, Lawrence Livermore NationalLaboratory, Livermore, CA, UCRL-96675;J. Grove and R. Menikoff, J. Fluid Mech 219,313 (1990).

12. B. Skews, J. Fluid Mech. 29, 705 (1967); L.Henderson, J. Fluid Mech. 26, 607 (1966) and198, 365 (1989).

13. R. Klein, C. McKee, and P. Colella, Ap. J. 420,213 (1994).

14. V. Kazakov, A. Legostaev, and S. Peigin, HighTemperature 31, 717 (1993).

15. K. Holtsapple and R. Schmidt, J. GeoPhys. Res.85, 7247 (1980), and 87, 1849 (1982).

16. G. Nutt and L. Klein, Phys. Fluids 24, 2143(1981).

17. N. Byrne, SAIC; T. Geffen, SRII; andT. Peyser, Lawrence Livermore NationalLaboratory; private communication.

18. T. Afshar-rad et al., Phys. Rev. Lett. 73, 74(1994).

19. T. Endo et al., Phys. Rev. Lett. 60, 1022 (1988).20. R. Sigel et al., Phys. Rev. Lett. 65, 587 (1990).21. B. A. Hammel, D. Griswold, O. L. Landen,

T. S. Perry, B. A. Remington, P. Miller, T. A.Peyser, J. D. Kilkenny, Phys. Fluids B 5, 2259(1993).

22. E. Muller, B. Fryxell, and D. Arnett, Astron.Astrophys. 251, 505 (1991).

23. G. Dimonte and B. Remington, Phys. Rev.Lett. 70, 1806 (1993).

24. J. Grun et al., Phys. Rev. Lett. 66, 2738 (1991).25. K. Mikaelian, Phys. Fluids 6, 1943 (1994).26. M. Zhuang et al., AAIA J. 28, 1728 (1990);

J. Hall et al. , AIAA J. 31, 2247 (1993).27. Hawley and Zabusky, Phys. Rev. Lett. 63, 1241

(1989); Zabusky et al. , Phys. Rev. Lett. 67, 2469(1991); P. G. Saffman, Vortex Dynamics(Cambridge University Press, New York,1992).

Material Properties 63 Section VI

Section VI

Material Properties

A. Equation of State

Material pressures of millions ofatmospheres (megabars or Mbar) to hundredsof millions of atmospheres are common inastrophysical objects, and are observed orpredicted in the laboratory. Direct illuminationof metals by a Nova-class terawatt laser canproduce pressures of 100 Mbar or greater.Gigabar (Gbar) pressures are predicted inspherically compressed capsules typical ofinertial confinement fusion targets andradiatively driven samples. The thermo-dynamics and hydrodynamics of these systemscannot be predicted without a knowledge of theequation of state (EOS) for the high-pressureregime, because the EOS of a material largelydescribes how that material reacts to pressure.

Although the EOS of matter in the limitingcase of extremely high pressure is expected to bedescribed by a Thomas-Fermi model, the regimeof applicability and approach to this limit arenot known; EOS data is sparse at pressures above3−4 Mbar because of the difficulty in producinghigh-pressure conditions while simultaneouslymeasuring the relevant parameters. Theultimate goal is to use a material of known EOSas an EOS standard material to obtain EOS dataon materials of interest.

In this type of experiment the shockvelocity us and initial density ρ0 are measured ineach of two materials effectively adjacent to oneanother. Shock impedance is the product (ρ0us).If the EOS of material I is known, then it ispossible to derive a Hugoniot point of materialII (i.e., the locus of final states obtained frompassage of shocks of varying strengths through

the material) from the measured shockimpedances.

At present there is no material with an EOSthat is sufficiently accurate to be used as astandard at the pressures of interest. The EOSstandards would be qualified by a series ofexperiments that use paired materials, both ofwhich have equations of state currently undertheoretical development in the pressure range ofinterest. The goal would be to provide accurateexperimental data to test theory over a widerange of pressures, densities, and temperatures.The materials to be qualified as standards couldbe, for example, aluminum and copper, becausethe theory is under development and because ofthe relative ease of fabricating targets fromthese materials.

To perform this type of primary dataexperiment in the pressure range of 10–50 Mbar(i.e., 1–5 TPa) requires, for direct illumination,intensities up to ~3 × 1015 W/cm2, while forx-ray drive, a temperature range of about 110–190 eV will be required. To perform much higherpressure experiments would require asubstantially larger laser system such as theNIF. This is a difficult regime, due not only tothe accessible pressure regions, but also to therigid constraints one must observe with respectto pre-heat, shock velocities measurements, andshock planarity.

There is exploration and development intothree techniques that will allow us to obtainEOS data in the Mbar and near-Gbar regimes.All three use high-energy lasers as the drivingsource to produce intense shocks in materials.Indeed, laser deposition is presently the onlyway to produce these conditions in thelaboratory. As discussed above, the shock brings

Section VI 64 Material Properties

the material into a high pressure state, albeit atransient one; measurements can then be made onthe shocked matter to obtain EOS data points.

A primary objective of this work is toproduce spatially large, planar shocks that areeasy to diagnose and interpret. The shock mustbe spatially large enough that the data areunambiguous. Although very intense shocks canbe generated by focusing a laser to a very smallspot, these shocks generate side motion andpotential development of plasma instabilitiesthat require multi-dimensional modeling.

Directly Driven High-Pressure Shocks

The long history of attempts to produce verystrong shocks in materials by direct irradiationwith lasers has produced conditions inferred intargets of from 2 to 100 Mbar. However, thesehad problems with two-dimensional effects orextremely nonuniform spatial profiles of thelaser, either of which can lead to laser-inducedlocal hot spots and instabilities.

To ameliorate the laser nonuniformities, thelaser spot has been smoothed with the use of arandom phase plate. A random phase plate is afilter with a random pattern of small holes. Itproduces a far-field pattern that is even inintensity (within 10%) except on scales smallerthan about 10 µm. In addition, steering wedgeshave also been used. These steer different partsof the beam in the near field to different spotsin the focal plane to produce a more flat-topped distribution.

As an example, employing phase plates andsteering wedges on a 1-ns beam of Nova providedan intensity of 3 × 1014 W/cm2 of 2ω0 light in afocal spot that was more than 1 mm in diameterand had a uniformity of ±10% over the spotdiameter. The beam was focused on a 25-µmthick aluminum disk, and the rear side of thedisk (i.e., the side away from the beam) wasimaged with a UV streak camera. The resultwas an extremely uniform shock breaking out ofthe rear of the target.

Figure VI-1 shows a schematic of the setupfor the experiment. The streak image in Fig.VI-2 shows that the shock, after traversing the25-µm thickness of aluminum in approximately650 ps, varies in breakout time across the entire

Smoothed laser beam

25-µm thick by 500-µmdiameter aluminum disk

Streak camera

Figure VI-1. Schematic of direct laser-drivenshock experiment. A spatially smoothed beamof Nova irradiates an aluminum disk, 25 µmthick by 500 µm in diameter. The schematic alsoshows the position of the streak camera(streaked UV imager).

diameter by less than 20 ps. This implies thatthe shock was planar across the disk by betterthan 1°.1 Pressures inferred in the shockedaluminum, using stepped targets and wedges,were 20−30 Mbar.

Indirectly Driven Colliding FoilExperiments

To reach a regime of much higher pressurewithout sacrificing spot size, and thus one-dimensionality, a very different technique hasbeen employed—a variation andminiaturization of the well-known flyer-platetechnique.2 In this method, the flyer (a foil inthe present case) stores kinetic energy from thedriver over an acceleration time and delivers itmuch more rapidly as thermal energy incollision with another foil. In addition, theflyer acts as a preheat shield so that the targetremains on a lower adiabat than if it wereexposed to the driver. These attributes make itpossible to achieve much higher pressures usinga flyer-impact foil or an indirectly drivencolliding foil than when using a directlydriven configuration.

In the experiments, an x-ray drive wasproduced by focusing the ten beams of the Novalaser into a millimeter-scale cylindrical goldhohlraum and utilizing the radiation escapingfrom a hole in the cylinder. The experimentalarrangement is illustrated in Fig. VI-3. The


Tim

e

Shock breakout onrear side planar to 1

Figure VI-2. Streak image of direct laser-driven shock experiment. Image is from the diagnostic,showing an extremely planar shock emerging from the rear side of the target.

50-µmCH ablator

2-µm / 6-µmgold target foil

3-µm gold flyer foil

Streak camera1-mm long by 700-µm

gold sleeve

50-µm void

Drivebeams

Drivebeams

Figure VI-3. Schematic of radiation-driven shock using a 2-step target foil. The diagram shows thecolliding foil arrangement (not to scale). The 1-mm-long by 700-µm-diameter experimental package isshown facing the hohlraum x-ray drive. Components of the package are a 50-µm plastic ablator towhich is attached a 3-µm gold flyer foil, a 50-µm void, and a two-thickness gold target foil, all held ina gold sleeve.


hohlraum x-rays ablated a 50-µm layer ofpolystyrene to which was attached a 3-µm-thick gold foil. This flyer foil acceleratedthrough a 50-µm void region and, near the end ofthe laser pulse, collided with a stationary goldtarget foil composed of two thicknesses (2 µmand 6 µm), launching a compression wave intothe target foil. The shock on the rear side ofthe target foil was imaged with an opticalstreak camera.

The experimental packages were more than0.5 mm in diameter, and the cylindrical targetassembly was mounted across a hole in the wallof the hohlraum with the ablator facing thehohlraum interior. The approximately 100-µmlong target assembly was placed at the bottom ofa 1-mm gold sleeve so that the assembly wascompletely shielded from unfocused,unconverted laser light. Additional shieldingprevented the streak camera from viewingheated areas of the sleeve and the hohlraum.

Fig. VI-4 shows a typical streak cameraimage. The image shows shock breakout at twotimes corresponding to the two thicknesses of thetarget foil; the time interval between thebreakout times measures the shock speed in thetarget, assuming the shock speed is constant.

Here, the interval measured between breakouttimes on the two thicknesses is 57 ±5 ps,corresponding to an average shock velocity of70 ±6 km/sec. From the SESAME equation-of-state tables, this shock speed corresponds to adensity of 90 g/cm3 and a pressure of 0.74 Gbar inthe gold target, by far the highest inferredpressure obtained in the laboratory.

Any slight spatial imbalance in the drive orany unpredicted edge effects (from interactionsbetween the flyer foil and the sleeve, forexample) could cause the flyer to tilt or curve,which would drive a non-planar shock into thetarget. However, the relatively large diameterof the foils would allow any non-planarity inshock breakout to be observed. In addition, thestep in the target was at the center of the largefoil, where the effects of edge-inducednonuniformities would be minimized.

If the target foil were preheated by high-energy x-rays from the hohlraum before theflyer-target collision, the measurement would becompromised. To test for this possibility, thex-ray drive was altered in one experiment sothat overall drive intensity was identical to theother experiments, but the intensity of high-energy x-rays (those ≥2.5 keV) was reduced by

700 µm

Shock breakout on 2-µm step

Shock breakout on 6-µm step 57 ps later

Tim

e

Figure VI-4. Streak camera image of radiation-driven shock using a 2-step target. Image shows shockbreakout on the rear surface of the two-step target foil at two distinct times corresponding to the twothicknesses of the target. This time interval provides the shock travel time across the 4-µm stepdifference. The interval in this case is about 57 ps, indicating a shock speed of 70 km/sec.


more than a factor of five. The result was wellwithin the uncertainties of the otherexperiments, so the results are not compromisedby preheat.

A straightforward extension of thistechnique can be made to obtain EOS data forthis regime. If the flyer foil can be shielded sothat the flyer does not significantly heat ordecompress, EOS data points can be found bymeasuring the speed of the flyer foil. Since ithas been demonstrated that the target foils dostay intact, this can be accomplished bymodifying the target foil by shielding it, asshown in Fig. VI-5.

With the target foil shielded and stepped,shock speed and flyer speed can be measuredsimultaneously. (The speed of the materialbehind the shock can be inferred from the flyerspeed.) These two measurements will allow thecalculation of an EOS primary data point in theGbar regime. Thus, in addition to the shockspeed discussed in reference to the single-stepped target shown in Fig. VI-4, the flyerspeed can be simultaneously obtained byrecording shock breakouts from two identicalfoils placed at different distances along theflyer path. One experiment was performed witha three-step gold target foil. Because thecondition of the flyer foil was unknown, an EOSdata point could not be evaluated, but observed

Ablator

Target foil

Flyer foil

Thickness 1

Thickness 2

Figure VI-5. Schematic setup with target foilshielded and stepped to allow simultaneousmeasurement of shock speed and flyer speed.(From the flyer speed, the speed of the materialbehind the shock can be inferred.) These twomeasurements will allow the calculation of anEOS primary data point in the Gbar regime.

shock breakout from the three steps indicatesthe viability of the approach.

To proceed with EOS measurements, thecondition of the flyer foil must be known; indeedthe accuracy of the result may depend on theflyer remaining solid. The development of ahigh-magnification x-ray laser radiographtechnique to determine the conditions ofaccelerated foils has been undertaken and this isdescribed in the section on x-ray laserapplications in Section VIII, Radiation Sources.

Indirectly Driven Shock Experimentson Plastic

Plastics are very different materials thanmetals, but they are just as pervasive in high-energy laser experiments as major constituents ofdiverse targets. Since plastics, unlike metals,are largely transparent to high-energy x-rays,x-rays can be used to backlight relatively thicksamples of plastic and provide information onthe sample as a function of time.

In particular, the shock front in a sample ofplastic can be seen on a streak camera bysimultaneously imaging the transmission of anx-ray backlighter through both the shocked andthe unshocked material. Transmission throughthe denser, shocked plastic is significantly lessthan through normal-density plastic—that is,the transmission is about 70% less for acompression of 4 in a 0.7-mm thick sample usinga 7-keV backlighter energy. If, at the sametime, motion of the material behind the shockcan also be imaged, a data point in the EOSis obtained.

An experimental arrangement thataccomplishes this, adapted from an experimentto measure material mix at interfaces,3 isdepicted in Fig. VI-6. The material motionbehind the shock is viewed as the difference intransmission of the backlighter throughshocked, doped plastic and shocked, undopedplastic; this is the interface shown in the figure.The drive, an x-ray source, is provided by a holein the side of a gold hohlraum.

Here only eight beams of Nova are used forthe drive, saving two beams for the backlighter.Code simulations predict a pressure of ≤40 Mbarnear the interface. Two experiments have been


Streak camera

Area backlighter

Be tube

X-ray drive

CH with dopant

CH without dopant

Interface

Shock front

Laser

Figure VI-6. Overview of arrangement for radiation-driven shock experiments in transmissive mediasuch as plastic (CH), showing relative positions of the package (which is mounted across a hohlraum),backlighter, and streak camera. The drive, an x-ray source, is provided by a hole in the side of a goldhohlraum. Differences in transmission of the backlighter through the package allow measurement ofboth the shock speed and the speed of the interface between doped and undoped plastic. This allowsdirect measurement of the equation of state of plastic in the multi-Mbar regime.

performed, both with 2% bromine doped into thefirst section, which was 300 µm long. A 3-ns longbacklight beam was initiated at about the timethe shock was expected to pass through theinterface. Although the interface was clearlyvisible on the streak records, the shock front wastoo weak to be seen. However, this techniquewill allow EOS data to be obtained on plastic inthe multi-Mbar regime.

B. Opacity

Knowledge of the x-ray opacities (x-rayabsorption) of hot matter is essential forunderstanding its state and radiative transport.The Nova laser has been used to obtainexperimental, high-quality opacitymeasurements. This not only required thedevelopment of high-resolution spectroscopictechniques to measure the x-ray transmission ofthe opacity samples, but also required thedevelopment of techniques to accurately andsimultaneously measure the temperature anddensity of the samples. Simply, to provide abenchmark of the ability to measure a local

thermodynamic equilibrium (LTE) opacity,measurements must be performed on relativelysimple atomic species first, so that thetechnique can be evaluated with some degreeof faith in the predictions. This requiresexacting procedures that provide the tempera-ture, density, and gradient scale lengths on asingle experiment.

The x-ray absorption of opacity samples wasmeasured by the method of point projectionspectroscopy, which is illustrated in Fig. VI-7.In this experiment, eight of the laser beams atNova are used to heat the opacity sample.Then, a point source of x-rays is produced bytightly focusing one of the remaining laserbeams onto a small backlight target of high-Zmaterial. X-rays from the backlight passthrough the opacity sample onto an x-raydiffraction crystal and are then recorded onx-ray film. Other x-rays from the same pointbacklight bypass the sample, but are stilldiffracted from the crystal onto the film record.

The ratio of the x-ray spectrum attenuatedby the sample to the unattenuated x-rayspectrum provides the x-ray transmissionspectrum of the sample. Proper collimation


Io —

point source of x-rays

Hohlraum heated by

8 laser beams

Bragg crystal

Backlight laser

e–τ — tamped opacity sample:1/2 without Z 1/2 with Z

x

y

Io e–τ:Absorption spectrum

Io spectrum

x

y, λ Film

Occluded area:emission & fog

Figure VI-7. Schematic of point projection spectroscopy method for measuring opacity. The laser-produced backlight x-rays pass through the target (the opacity sample) onto an x-ray diffractioncrystal and are then imaged on x-ray film. Because the crystal disperses the spectrum, the result isan image that is both spatially and spectrally resolved. Temporal resolution is provided bybacklight duration.

allows a highly quantitative analysis of thespectrum. Backgrounds from film chemical fog,sample emission, and crystal x-ray fluorescencecan all be separately determined from the x-rayfilm record.

Careful attention was paid to the sampleconditions. It was required that the sample beuniform throughout in both temperature anddensity. Uniformity of temperature wasattained by heating the sample in a specialhohlraum. The hohlraum was constructed sothat no laser light impinged on the sampleeither directly or on first reflection; thus, thesample was heated only by x-rays. Thehohlraum also kept a uniform temperaturethroughout the sample, and this, along withthe relatively high density of the sample,ensured that the sample was in localthermodynamic equilibrium.

Further, the samples were tamped byplastic so that the density throughout thesample was constant as the sample expanded.The thickness of the tamper was determined byhydrodynamic calculations, and density

uniformity was checked in the experiments.This experimental technique allowed themeasurement of the opacity of aluminum andlaid the basis for using aluminum opacity as anindependent temperature diagnostic.

In further experiments, the density of thesamples was determined by imaging thesamples. This was done by using a second point-projection spectrometer to image the expansionof the samples. This spectroscopic record wasalso used to verify that the samples haduniform density throughout the sample. Thetemperature of the sample was determined bymixing the sample with aluminum. A thirdpoint-projection spectrometer was used tomeasure the absorption of the aluminum n = 1to n = 2 transitions. The relative intensities ofthe transitions from the different ion speciesgave the ion balance in the aluminum which(coupled to the density measurement) gave thetemperature of the sample.

These techniques allowed density to bemeasured to an accuracy of ±20% andtemperature to an accuracy of ±3%. With these


accuracies it was possible to make quantitativecomparison between the experimental resultsand the theoretical calculations of opacity.4

In particular, one experiment measured theopacity of niobium. The niobium samplecontained 14% aluminum by weight for thetemperature measurement described above.Figure VI-8 shows the transmission of thealuminum (upper) and the niobium (lower).The dashed lines overlaying the experimental

data are the calculations. In general there isexcellent agreement.

This experimental result defines a milestonein that we have proven the ability to measurethe opacity of the aluminum sample withsufficient accuracy that it can now serve as an in-situ temperature diagnostic for the sample. Theaccuracy of the temperature, measured to be48 eV (±2 eV), indicates an important advanceon previous attempts to measure temperatures ofhigh-energy-density matter.

1.0

0.8

0.6

0.4

0.2

0.0

Energy (eV)

2100 2200 2300 2400 2500 2600 2700 2800

Niobium

Tra

nsm

issi

on

Energy (eV)

1.0

0.8

0.6

0.4

0.2

0.0

1520 1540 1560 1580 1600

Aluminum

Tra

nsm

issi

on

Figure VI-8. Absorption (opacity) of the aluminum/niobium sample. The experimental data are shownas a solid gray line and the LTE opacity prediction calculated using OPAL shown as a dashed line. Theupper graph shows the transmission of aluminum, the lower graph shows the transmission of niobium.The spectrum of the aluminum Kα lines, which have been previously verified to yield an accuratetemperature, are measured on the same experiment as in the niobium spectrum in the lower graph.


Other experiments on opacity haveconfirmed the astrophysically important iron∆n = 0, n = 3 absorption feature, which causes alarge increase in opacity above olderpredictions.5 The work on iron has been furtherextended to measurement of the spectrum insufficient detail to evaluate the Rosseland meanfor comparison with theoretical predictions.6

Both of these studies on iron opacity aremotivated by astrophysical considerations andindicate that, in at least this one area, theexperimental capability has become mature.

C. Strength of Material

Laser generation of multi-kilobar shocks hasseveral applications to the study of highexplosives, high-velocity impacts, and thealteration of the mechanical properties ofcertain alloys and ceramics. There have beenseveral important advances in the field sincethe early 1970s. Some of the most commonmethods of studying such shock waves includeinterferometric methods and the use ofpiezoelectric transducer gauges. However,these methods yield little information on thedynamics of the near-front face, where theshock wave is initiated.

It is possible to measure shock strength (interms of compression ratio) and spatial densityprofile using an x-ray diffraction technique forstudying laser-induced shocks within variouscrystalline materials. This can be used forpressures from 1 kbar up to approximately0.5 Mbar, where shock melting occurs. This x-raydiffraction technique is ideally suited toobservation of the dynamics of the front surfaceof laser-shocked materials. The technique canmake a significant contribution to ourunderstanding of shock launching, and because itis a direct measurement of the relevantparameters of the near-front surface, it canenable conclusive investigations of the responseof novel materials. It is also possible to extendthis technique to probing of the rear-surfacebreakout of the shocked material.

The technique is based on the fact that whena crystal is subject to strain, its x-ray diffraction

properties are altered—most notably the x-raydiffraction peak of the crystal is broadened.The physical reasons for this are easy tounderstand: a strained crystal essentially hasa distribution of interatomic spacings at whichthe Bragg condition can be satisfied, and there-fore the shape of the diffraction peak isa function of the spatial strain profile. Thisfact has been used to study strain profiles incrystals where the strain is caused by dopants orby laser annealing. (The transient alteration ofthe diffraction peak in a laser-shocked crystalhas also been suggested as a means of obtainingsubnanosecond x-ray switching.)

The experiment to measure compression inthe crystal proceeds as follows (see Fig. VI-9). Ashock is launched into the crystal of interest. Atsome point during the shock launching a short(100-ps) pulse of x-ray line radiation is producedby a synchronous, but delayed, laser beamirradiating a separate target. This pulse isBragg-diffracted off the front surface of theshock-compressed crystal. The compression ofthe crystal changes the Bragg conditionaccording to the formula

∆d/d = – Cotθ ∆θ.Thus, the line radiation is diffracted to ahigher angle than that for the unperturbedcrystal, and this angular shift directly indicatesthe change in the interatomic spacing withinthe shocked region. Because there will be adensity gradient at the crystal surface, therewill be, in addition to the overall shift, a rangeof angles at which diffraction takes place. Thus,information can be obtained about the densitygradient as well as about peak density. Indeed,it is possible to extract a density profile fromthe data contained in the diffracted pulse.

The primary interest in this techniquewould be to find a method of studying thematerials under extreme conditions. The presentset of known experiments is restricted to simplestructures with small heated areas. To developlarger perturbed areas and study the materialsat greater depth would require a substantialeffort in the development of larger lasers andx-ray sources for probing.


100-ps backlight

Beam block X-ray film

Spectrum

Crystal

laser beam

250-µm Si Si Si

25-µm CH1000-Å Al

Targets in order of increasing shock stability

Figure VI-9. Schematic of x-ray diffraction technique used to study the dynamic structure of a crystal.The crystal is compressed using a laser-induced shock from a monochromatic x-ray source. During theshock transit through the crystal, a short (100-ps) backlight pulse is Bragg-diffracted off the frontsurface of the shock-compressed crystal, and the resulting spectrum is recorded on the x-ray film.

Simple Compression Studies

Silicon has been shocked and transientstrains of up to 10% have been observed. Themethod of Bragg diffraction relies on the factthat the crystal remains “single”; i.e., allthe planes running parallel to the shock frontare compressed uniaxially and the periodicityof the crystal as a whole is maintained.7 Insome ways it is quite surprising that suchan ordered state exists so far above theHugoniot elastic limit, which for silicon in the(111) direction corresponds to an elasticcompression of about 2.6%. Thus, one simpleexperiment is the investigation of the maximumobservable compression for a series of differentcrystalline materials, and comparison of thiswith shock melting data obtained by moreconventional means.

It is interesting to note the utility of high-energy lasers for the following two reasons.First, they provide for a relatively high shockstrength within the material of interest(~1 Mbar). Although smaller lasers could attainlocal pressure of this magnitude, it could only doso over a spot size on the order of a squaremillimeter. A large spot is needed so that the

shocked region of the crystal subtends a largerange in angles to the x-ray source.

Second, they provide harder x-rays. Smalllasers may not have sufficient backlight energyto produce hard enough x-rays for accuratediagnosis. It has been found that at largecompressions, such as the 10% compressions insilicon, the ∆Θ becomes large for softer x-rays.Soft x-rays diffracted from the shocked portionof the crystal get smeared out on the x-rayfilming, making diagnosis difficult. The changein Bragg angle for harder x-rays is not so great,facilitating diagnosis. For example, for the 10%compressions in silicon, the helium-liketitanium resonance line was an appropriatex-ray source. Much harder x-rays may berequired to measure compressions that are closeto shock melting (30−40%) and for investigatingcrystals with shorter interatomic spacings, suchas lithium fluoride and quartz.

Plasticity and Adiabatic Shear Bands

Other questions that should beexperimentally ripe include investigating thebehavior of a single crystal on the lattice levelwhen it is shocked above the Hugoniot elastic


limit, and determining how long the crystaltakes to start to deform plastically.8

There is evidence that some materials,when shocked above the Hugoniot elastic limit,form small bands of micron or sub-microndimension where the dislocation density isextremely high; such areas are known asadiabatic shear bands. Evidence for suchlocalized yielding is mainly based on recoveryexperiments rather than on in-situ observation.

It is of interest to attempt to observe thesebands by a short-pulse x-ray topographyexperiment, using a point source of laser-plasma-produced x-rays. Point sources of 5 µmdimensions have routinely been produced usinglasers, and these could be used to image shearbands. Although the source size is larger thanthe width of the bands, the bands can runseveral microns into the crystal, and imagingshould be possible in the same way thatdislocations of Ångstrom size can be imagedusing conventional x-ray topography.

It is also of interest to determine the shapeof the unit cell as the crystal is compressedabove the Hugoniot elastic limit. This can bedone by using simultaneous Bragg and Lauediffraction with thin crystals. Bragg diffractiontakes place from the surface of the crystal, thesurface being parallel to the shock front, andmeasures the elastic compression. Observing theLaue diffraction through the crystal at the sametime, interrogating the planes runningperpendicular to the shock front, providesinformation about the plastic componentof strain.

Experiments have been performed using50-µm-thick silicon (111) crystals, with Lauediffraction off a (220) plane through the crystal.Although both Bragg and Laue lines wereobserved when no shock was present, theintensity of the Laue line was not sufficient toenable diagnosis when the crystal wascompressed. Improvements in crystalpreparation and an increase in backlight energyshould permit measurements of this type.

Metastable Phases

It is believed that many crystals, whenshocked above the phase change point, proceed

to a metastable phase before completion. Oneexample is potassium chloride, which undergoesa phase change from sodium chloride structure tocesium chloride structure at a pressure of about20 kbar. Investigations of the multiple wavestructure in sodium chloride have beeninterpreted as transformation to a metastablephase within 2 ns, with completion to thecesium chloride structure taking between 25 and500 ns, depending on crystal orientation.

Investigations of potassium chloride havebeen performed, and it was found that thediffraction lines from the sodium chloride phasedisappeared upon shocking. However, the exactstrain at which the disappearance occurred wasnot determined. Time-resolved x-ray diffractionfrom the shocked crystal could permitmeasurement of the strain at which diffractiondisappeared, which is expected to be at thepoint where the phase change occurs. Thus, onewould simply “streak” the diffraction.

Disappearance of the diffraction recordsuggests that the new phase is polycrystalline,as would be expected in an order-disordertransition. Thus, the ability to obtaindiffraction information from powders would beof importance.

Phase-Change Measurements

The first, and only, in-situ x-ray diffractionevidence for a shock-induced phase transition isthat obtained from shocked boron nitride, whichtransforms to a wurtzite structure. It is adisplacive transition without a change in unitcell volume or occupancy. Such transitionsshould take place on an inverse phonon time;i.e., a picosecond time scale. Obviously it wouldbe interesting to repeat and extend thesemeasurements with the improved timeresolution and synchronization available withhigh-power laser technology.

In addition to shocking boron nitride, itwould be interesting to investigate TeO2. Thismaterial undergoes a transition from tetragonalto orthorhombic at a pressure of 9 kbar. This isinteresting because the transition proceeds by anacoustic shear wave phonon traveling along(110) at an unusually slow sound speed of about1 µm per ns.


Rear-Surface Studies

Dynamic tension can be produced within thebody of a material by subjecting the front surfaceof a planar target to impulsive loading. Tensilestrain is then induced when the duration of thepressure pulse, τ, is less than cd, where c is theshock velocity within the material and d itsthickness. In this case two rarefaction wavescross; one of them produced at the front surfaceas the pressure loading falls off, and the otherproduced at the rear surface as the incidentcompression wave is reflected. For sufficientlyhigh tensile stresses, a section of the rearportion of the target separates, or “spalls,” fromthe bulk.

At laser irradiances >109 W/cm2, absorptionof the laser light produces a high-pressureplasma at the target surface. As this plasmaexpands away from the surface, momentumbalance launches a compression wave into thebulk of the target. The temporal profile of thecompression wave is similar to that of thelaser pulse.

Direct measurements of tensile elastic strainin silicon (111) wafers have been obtained atstrain rates above 108 s–1. Tensile elastic strainsof 3.4% have been directly observed by in-situpicosecond x-ray diffraction.9 This mayconstitute a novel method of addressing theideal strengths of materials.

Figure VI-10 shows a schematic of a rear-surface shock breakout experiment to probe theshock structure of a material using amonochromatic x-ray source. Prior to thebreakout, the probed rear surface will be incompression, during the breakout there will alsobe tension due to a rarefaction wave, andfinally the rear surface will go into tension withno compression.

The experimental results are shown inFig. VI-11, where the x-ray streak camerarecord shows the entire history of the rearsurface from unshocked, to the initiation of thebreakout, through to the time when the crystalis in pure tension. From Fig. VI-11 it can be seenthat at early times, for the first ~200 ps, thex-rays are probing the crystal in a state of com-pression, with the angular shift in diffractioncorresponding to a maximum compression of

6.2 ±0.2%. Later the x-rays are being diffractedfrom both regions of compression and regions oftension, and finally, at late times the x-rays arebeing diffracted from a region of pure tension,with a maximum observed tension of 3.4 ±0.2%.

These three regimes correspond to the timeswhen the reflected wave has not yet collidedwith the rarefaction wave due to release ofpressure at the front surface; the time duringcollision of the rarefaction waves; and the timewhen the reflected wave has passed the frontsurface rarefaction wave by at least the probedepth of the x-rays.

Figure VI-12 compares the spectra from ashocked crystal with those from an unshockedcrystal at various stages of an experiment(Fig. VI-11 shows similar stages in a streakcamera record). Thus, within the target thestrain has changed from 6.2% compression to3.4% tension within ~600 ps (i.e., at a strain rateof 1.6 × 108 s−1). The maximum observed tensilestrain 3.4% corresponds to a uniaxial tensilestress of approximately 70 kbar, comparableto the largest fracture stresses observed instatic measurements.

5-µm Ti foil

1000-ÅAl layer

40–120-µm Sisingle crystal (111)

X-raystreakcamera

Block to stop directx-ray illumination ofx-ray streak camera

Laser beam forbacklightproduction

Shock-driving laser beam

X-rays fromregions of compression

X-rays fromregions of tension

X-rays fromunshocked silicon

Figure VI-10. Schematic of a rear-surface shockbreakout experiment to probe the shock structureof a material using a monochromatic x-raysource. Prior to the breakout, the probed rearsurface will be in compression, during thebreakout there will also be tension due to ararefaction wave, and finally the rear surfacewill go into pure tension with no compression.


10 2Time (ns)

6%

– A

ngle

Com

pres

sion

Ten

sion

Figure VI-11. Data from the rear-surface shock breakout experiment in Fig. VI-10, showing a shocktransiting through the material. The streak camera record shows the entire history of the rear surface,including the compression phase at early time as the shock nears the rear surface, the compression andtension after the shock breaks out, and the late-time phase of pure tension.

0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

–2 –1 0 1Shift in diffraction angle (degrees)

2 3

3.4 ±0.2% tension 6.2 ±0.2% compression

Tension alone

Unshocked

Tension & compression

Figure VI-12. Spectra from the probed shocked rear surface of a crystal. The stages shown graphicallyhere are comparable to the stages visible in the streak camera record in Fig. VI-11. The three genericperiods are shown—the pure unperturbed (unshocked) crystal in a dashed line; the period of shockbreakout, when there is tension and compression, in a solid line; and the pure tension late-time phase inthe dotted line.


Finally, it is of interest to point out thatwith the use of a powder diffraction camera anda bright source (very bright indeed), one shouldbe able to extend all these studies to non-crystalline materials. This would open the wayfor a new source of studies of the detailedresponses of a much wider range of materials.


The study of the properties of materials atextreme conditions of pressure, temperature, anddensity has been a goal of laser experimentssince the advent of high-energy facilities. Themost common application has been ininvestigations of equations of state (EOSs)—therelationship between pressure, density, andinternal energy.

Impacts using guns or high explosives togenerate shock waves have provided nearly allof our current database in this area, startingwith investigations during World War II. Thosemethods have provided a great deal of precisedata on a wide range of materials, but arelimited to pressures below 1 TPa.10 Staticcompression experiments have been performedextensively in diamond-anvil high-pressurecells; these experiments will ultimately belimited to pressures below roughly 1 TPa by theinsulator-metal transition of diamond. Higherpressure experiments have been carried out atpressures from 1 TPa to over 100 TPa in thevicinity of nuclear explosions, but the data areof low precision above 10 TPa.11

Relevant to the present discussion we notethat high-intensity lasers can achieve similarhigh pressures, as was recently demonstrated.12

In the higher pressure regime, many newphenomena will become apparent, such as pres-sure ionization, strongly correlated fluid states,and so on. The strain rates involved can range upto 1020 bar/s, and the behavior of materialsunder these conditions is unknown, especially asapplied to melting and other changes of phase.

Figure VI-13 is a plot of stress vs strain ratedomain. It shows the regions accessible tovarious experimental methods. High-stress,high-strain-rate tensile states can be attainedin solids when a shock breaks out of a free

interface of a sample. Experimental fracturestrengths of brittle materials show a 1/3 powerdependence with the tensile loading strain-rate.Extrapolating these fracture stress measure-ments, which were taken at strain rates below105 s–1, to strain rates up to 1010 s–1, tensile stressequal to the theoretically predicted tensilestrengths can be reached.

Strain rates as high as 109 s–1 are alreadypossible using high-power lasers and in-situx-ray diffraction. With NIF-like technology,the peak tensile stresses will be far in excess ofthose presently realizable and will be atultrahigh strain rates. This capability willmake it possible to investigate the fractureproperties and ultimate strengths of manymaterials, including novel ceramics.

We have essentially no bulk or atomic-scaleknowledge of any materials under theseconditions of pressure loading. Precisemeasurements are necessary to discriminateamong competing models or theories. The NIF isimportant in that it is the first large laser ofsufficient energy and pulse length to allow forthe possibility of such measurements being madewith sufficient accuracy.

Data and validated theoretical models arevital for understanding and modeling suchdiverse phenomena as inertial confinementfusion, planetary interiors, and ultrahigh-velocity meteorite impacts. Data and validatedtheoretical models are also vital for most of thescientific experiments to be performed at theNIF, especially hydrodynamics and radiativeand transport properties in condensed phasesystems at high pressure (P) and temperature(T). We need to know the properties of materialsover wide ranges of pressure and temperature.The detailed sections that follow describe theproblems and some proposed experiments todetermine the EOSs for:

• The principal Hugoniot.• Multiple shock states.• Isentropic release states.• Isochoric heating experiments.• X-ray diffraction studies.• Other material properties.

In doing so, we will describe applications toother fields.


Fra

ctur

e st

ress

(kb

ar)

101 102 103 104 105 106 107 108 109 101010–1

100

101

102

103

104

105

106

Strain rate at fracture ( s–1 )

Recoveryand otherindirect

methods

In-situ methodspossible with NIF

GraniteQuartz

Limestone

Oil shale

Concrete

Conventional techniques

Figure VI-13. Plot of stress vs strain rate domain, showing the regions accessible to variousexperimental methods. High-stress, high-strain-rate tensile states can be attained in solids when ashock breaks out of a free interface of a sample. NIF technology will make it possible to investigate thefracture properties and ultimate strengths of many materials, including novel ceramics.

The Principal Hugoniot

The principal Hugoniot is the locus of singleshock states for shocks of varying strength,starting from material at normal density,atmospheric pressure, and 300°K. These are notthermodynamic paths, but a true locus of states(see Figure VI-14).

The Rankine-Hugoniot relations,13 whichexpress conservation of mass, momentum, andenergy across the shock front, make theprincipal Hugoniot particularly useful. Thefinal shock state (pressure, density, and totalenergy) can be determined from a knowledge ofthe initial state and two dynamic variables.The two dynamic variables are

• The velocity at which the shockpropagates through the undisturbedmedium, referred to as either us or D.

• The velocity of the material behind theshock front, referred to as either up orU and known as the mass orparticle velocity.

For steady shocks, us is directly accessible tomeasurement, in principle.

Shock compression of both fluids and solidscan be visualized as a discontinuous change instate variables. This assumption is used inderivation of the Hugoniot-Rankine relations.With this view only the initial and final statesof a shock-compressed material exist. In Fig.VI-14 we see that for an initial state (point A),there exists an infinite number of end states.A specific end state is reached when a materialis subjected to a shock of specific strength. Byperforming a number of experiments, each witha different shock strength, a locus of end pointsis mapped out (points B, B’, and B”). This locusof end points is known as the Hugoniot.

The straight lines joining the initialcondition A and each of the end points B in thePV plane are the Rayleigh lines. The slope ofa Rayleigh is proportional to the shockvelocity—higher strength compressions areassociated with faster shock velocities.


Figure VI-15 illustrates the relationships ofthe Hugoniot, an isotherm, and an isentrope allgoing through the initial point A. A single pointon the Hugoniot, which is the locus of end states,is reached after shock compression. Shockcompression involves an increase in entropy,and the Hugoniot will always lie above the

isentrope. Because any increase in entropy isassociated with a rise in temperature, theHugoniot also lies well above the isotherm.However, there may be situations where this isnot true (for example, during a phase change)and in these cases a shock is thermodynamicallyimpossible. With the capability of the NIF to

Reduced volume (V/V0)10

Hugoniot

AB

B"

B'

Rayleigh lines

Pre

ssur

e

Figure VI-14. The principal Hugoniot defined. Shock compression of both fluids and solids can bevisualized as a discontinuous change in state variables. For an initial state (point A), there exists aninfinite number of end states. A specific end state is reached when a material is subjected to a shock ofspecific strength. Points B, B’, and B” represent a locus of end points from experiments with differentshock strengths. This locus of end points is known as the Hugoniot. The straight lines joining the initialcondition A and each of the end points B in the PV plane are the Rayleigh lines.


Pre

ssur

e

HugoniotIsentrope

Isotherm

Figure VI-15. Relationships of the Hugoniot, an isotherm, and an isentrope all going through the sameinitial point. Shock compression involves an increase in entropy, and the Hugoniot will always lieabove the isentrope. Because any increase in entropy is associated with a rise in temperature, theHugoniot also lies well above the isotherm.


provide more energy and long tailored pulses,yielding large scale-lengths and increasedplanarity, several of the current problems inachieving measurements will be overcome. Thiswill open new horizons by permitting uniformityand constancy of shocks over time scales relevantto shock studies.

The standard method is to measure thetransit time ∆t across a step of known extent ∆x,and then us = ∆x/∆t. Although this is astraightforward concept, there are severalpractical difficulties in current experiments,difficulties the NIF can alleviate.

First, it is assumed that the velocity isconstant. This has been rather difficult toachieve in ablatively driven shocks. Withcareful target design, the long, temporallytailored pulse of the NIF can solve this problem.

Second, the velocities are large—for alumi-num at 10 TPa, us ≈ 70 km/s. Such a shock wavewill cross a 100-µm step in only 1.4 ns, and the1% accuracy we desire implies an uncertainty inthe measurement of only 14 ps. This will strainthe ability of current diagnostic systems.

The NIF will allow greater spatialdimensions, which will allow higher fractionalaccuracy for measurements of both length andtime. Further, the use of a hohlraum drive willhelp ensure that the shocks are planar and oflarge transverse extent, which is necessary ifthe larger target samples are to be free of two-dimensional effects. Accurate measurements area requirement, because the uncertainty inshocked density varies roughly as the product ofthe compression with the uncertainty in shockvelocity, and compressions greater than fourwill be common.

It is, in principle, possible to determinecompression directly using x-ray backlighting; inpractice, we believe that to be extremelydifficult. What remains is to measure the massvelocity. Again, the scale of targets posited forthe NIF makes possible a method proposedyears ago, but never realized. The idea is todirectly measure the velocity of the shockedmaterial by observing the motion of an interfacein the shocked material using x-ray sidelightingwith an x-ray streak camera. The velocities ofthe shocked material are large as well, but the

spatial dimensions over which the measurementcan be made are improved at the NIF by virtueof the relatively long time scale possible for theexperiment. For it to become possible to directlymeasuring the velocity of the shocked material,a point-backlighter beam is essential to avoidparallax effects.

A further possible set of experiments toproduce information on the principal Hugoniotis the use of what have become known asimpedance-match experiments. In these experi-ments the equation of state of an unknownmaterial is compared to that of a knownmaterial—perhaps by the methods describedabove.14

This method has been used in laserexperiments before, but the scale of theexperiments was such that only limited accuracywas possible.15 The method can be exploited atthe NIF because of the planarity and uniformityof the drive in hohlraums. One suggestion isthat a simple material such as aluminum or goldbe developed as a “standard” material over thebroadest pressure range on the principalHugoniot, and then used as a point of comparisonwith other materials of interest.

On the principal Hugoniot, there are anumber of materials that must be studied.Hydrogen is the chief component of the Jovianplanets. The current upper limit on single shockexperiments in the fluid is 0.02 TPa,16 whichsamples only a fraction of the Jovian interior.Solid D2 or H2 could be shocked to many TPain well-characterized experiments, perhapsencountering the proposed plasma phasetransition.17

Further, metals of various Zs need to bestudied. For aluminum in the range 10–100 TPa,successive pressure ionization steps arepredicted to occur, and they may lead to wideexcursions of the density on the Hugoniot ascompared to calculations with purely statisticalmodels.18 In low-Z metals such as beryllium,these excursions are expected to occur below10 TPa.

The Rankine-Hugoniot conditions relate thedynamic variables and initial state to yield thetotal shock energy, and do not explicitly givethe temperature of the shocked medium. This is


an important issue, since temperature, ratherthan total energy, is often an explicit variablein theoretical models. When we have recourse tocompeting models in the range of a few tens ofTPa (several hundred megabars), we find thatat a given pressure, the densities vary by about30%, while temperatures are known to no betterthan a factor of two.

This is no mere academic problem, since theproblem of high-pressure turbulenthydrodynamic flows explicitly involvestemperature and entropy, and understandingthese flows may prove intractable without moredetailed information about the distribution oftotal energy into external and internal degrees offreedom. The solution to this problem may be todetermine temperature by x-ray absorptionspectroscopy, using absorption line ratios todetermine the temperature of the shocked fluid.We note that the value of temperaturemeasurements is not limited to the principalHugoniot; these comments apply to other high-pressure thermodynamic states and processesas well.

Multiple Shock States

The principal Hugoniot is a single locus ofstates; we need to know the properties ofmaterials over wider ranges of pressure,especially planetary or other isentropes. Thisregion can be probed by the use of multiple-shockexperiments. Of this type of experiment, we candistinguish two main possibilities.

The first is colliding shocks, initiallytraveling in opposite directions. This methodcan be used to generate extremely high pressures,since relative velocity is doubled, causingpressure to be increased by a factor of about four.

A view of the multiple shock trajectories inthe PV space together with an idealizedschematic of a multiple shock experiment isshown in Fig. VI-16. For single-shockcompression there is a limiting compression offluids and solids which is asymptoticallyreached at high pressures. For large increases inpressure, small compressions are achieved at theexpense of significant material heating.

Multiple shocking of a material allowsgreater compressions at lower temperatures and

pressure to be attained. In Fig. VI-16, forexample, a material is shocked from its initialstate A to state B on the principal Hugoniot. Ifthe material is shocked again while in state B,the new end state C will lie on another Hugoniotthat is defined by the new set of initialconditions B. Again a limiting compression isdefined that is some fraction of the specificvolume of state B.

The schematic in Fig. VI-17 shows a double-shock compression experiment, with a particleat a position A embedded in the material to beshocked. In the top picture, a piston drives ashock (small arrow) from the right into thecompressible material (solid). In the middlepicture, the shock, traveling toward the wall,passes the particle at point A. As the shockpasses it, the particle moves from state A tostate B. In this change of state it acquires aparticle velocity that is towards the wall indirection, but its velocity is lower than theshock velocity. The shock hits the immovable

Pre

ssur

e


HugoniotA

C

B

Rayleigh lines

Figure VI-16. Double-shock compressionexperiment. Multiple shocks allow access to awide range of conditions, making it possible toattain greater compressions at lowertemperatures and pressure. In the figure, amaterial is shocked from its initial state A tostate B on the principal Hugoniot. If thematerial is shocked again while in state B, thenew end state C will lie on another Hugoniotthat is defined by the new set of initialconditions B.


us

Drive

us

Drive

us

Drive

A

B

C

Figure VI-17. Schematic of double-shockcompression experiments. A particle at aposition A is embedded in the solid material tobe shocked—a compressible material (solid)adjacent to an infinite-impedance, immovablewall (heavy black line). In the top picture, apiston drives a shock (small arrow) from theright into the compressible material. As theshock (middle picture), traveling toward thewall, passes the particle at point A, theparticle moves from state A to state B. In thischange of state it acquires a particle velocitythat is towards the wall in direction, but itsvelocity is lower than the shock velocity. Theshock hits the immovable wall and is reflectedinto the solid, now traveling toward the right(bottom picture). The shock is now moving backinto the compressed material, which is underpressure (from the piston), has been heated, andhas a velocity—this is state B. As the shockpasses it the second time, the particle moves tostate C on a second Hugoniot, acquiring a newparticle velocity.

wall and is reflected into the solid, nowtraveling toward the right (bottom picture). Theshock is now moving back into the compressedmaterial, which is under pressure (from thepiston), has been heated, and has a velocity—this is state B. As the shock passes it the secondtime, the particle moves to state C on a secondHugoniot, acquiring a new particle velocity.

Situations akin to this occur naturally insolids—for example as a solid is shock-compressed beyond the Hugoniot elastic limit(in the weak shock regime), or during a phasetransition. The flexible pulse technology of theNIF will enable advanced methods of obtainingmultiple shock states.

In the case of relatively oblique incidence,we have the opportunity to study shockinteractions that may lead to hydrodynamicinstabilities, since there will be a shearcomponent to the flow. This type of experimentis only practical with the use of large high-energy lasers, because the problems of timingand balanced shock generation are intractablefor other drivers. The topic of instabilities istouched upon in Section V, Hydrodynamics.

Of greater interest is the possibility ofmultiple or reverberating shocks in a materialinitially at rest. For a given final pressure,multiple shocks will achieve higher finaldensities and lower temperatures. Withmultiple shocks, ten-fold compressions can beachieved.

The limit of very weak multiple shocksapproximates adiabatic compression; allmultiple-shock states of this type fall betweenthe principal Hugoniot and the 0°K isentrope.This is an important regime for planetaryphysics, because all of the major planets arebelieved to have interior distributions oftemperature and pressure that are bestapproximated by isentropes. As we noted in theintroduction, current technology can sample onlya very small fraction of those interiors. Usingliquid or frozen hydrogen targets, experiments onthe NIF will be able to probe states that aretypical of the deep interiors of the Jovianplanets, as well as of “brown” dwarfs in general.

Multiple shocks can be generated byreflecting a simple shock wave at a boundaryof high shock impedance (the product of densityand shock velocity). These multiple shocksmay be characterized by the methods describedabove. However, the flexible pulse technologyproposed for the NIF will enable a more directmethod for characterizing multiple shocks. Inthis case, the driving pulse (for either director indirect drive) is stepped in intensity. With


appropriate target dimensions, we can arrangefor the sample to be shocked in succession bythese pulses.

Isentropic Release States

Isentropic release states are characterizedby higher temperatures and lower densitiesthan on the principal Hugoniot. These states canbe reached by strong shocks followed by arelease to lower pressure. When a shock in amaterial encounters a boundary with a mediumof lower shock impedance, the final state of thematerial is on an isentropic release from theinitial state. At a given pressure, these stateswill have higher temperatures and be morehighly ionized than on the Hugoniot. If wethink of the total pressure as arising from a“cold” compression part (that is, the 0°Kisentrope) and from thermal contributions, thesestates are dominated by the thermalcontributions to the equation of state.

When a strong shock in a material releasesinto a vacuum, it forms a rarefaction fan—agradient of density and temperature. This isdifficult to characterize. It is simpler to mount amaterial of very low shock impedance againstthe initially shocked material, and thenmeasure the shock in the low-impedancematerial. The final state in that material willmatch, in mass velocity and pressure, therelease state of the initially shocked material,so the release state can be characterized bysimple Hugoniot measurements in the low-impedance material.19 The large target scalespossible on the NIF will make this kind ofexperiment possible for the first time in a laser-driven environment.

This process is shown in schematic form inFigs. VI-18 and VI-19, which model the sameshock experiment. Figure VI-18 models theexperiment in terms of pressure history (vstime), and Fig. VI-19 shows the shock in terms ofpressure vs volume.

Figure VI-18 shows the pressure history ofthe particle in the weakly shocked solid. Themodel is idealized and assumes that the solidbehaves in a perfect-elastic, perfect-plasticway. The first shock wave (the elasticprecursor) shocks the particle from normal

pressure, density, and temperature to state B atthe Hugoniot elastic limit (HEL). The particlesits at this state (B,B’) until a second, slowerwave (the plastic wave) shocks the solid tostate C.

Peak pressure C,C’ is maintained until therelease wave, or rarefaction, reaches theparticle at the release elastic limit (REL). Therelease wave is led by an elastic rarefaction(with a speed of u3), which releases the solid tostate D,D’. This release is immediatelyfollowed by a slower-moving plastic rarefactionto the final state E.

E

D'D

C'C

B'B

Au1

u2u3

u4

HEL

REL

Pre

ssur

e

Time

Figure VI-18. Model of pressure history vs timeof a particle in a weakly shocked solid. In theinitial state (A) the particle is at rest and atnormal pressure, density, and temperature. Theparticle is taken from state A to state Bdiscontinuously at the Hugoniot elastic limit(HEL), shown by the wide gray line at B,B’, bythe elastic precursor traveling at speed u1. Asecond, slower wave (the plastic wave) shocksthe solid from state B,B’ to state C. Peakpressure C,C’ is maintained until a release wave(or rarefaction) reaches the particle. Therelease wave, led by an elastic rarefaction,releases the solid to state D,D’, at the releaseelastic limit (REL), shown by the narrower grayline. This release is immediately followed by aslower-moving plastic rarefaction to the finalstate E.


Elastic and plastic rarefactions travel fasterthan the plastic shock (u3 > u2), because thematerial these waves travel into is compressedand hot. With time, the release overtakes theshock, eroding peak compression. It is vital toensure that the duration of peak compressionC,C’ is long enough to prevent decay ofcompression during the experiment, and this iswhere the NIF plays a significant role.

Figure VI-19 shows the Hugoniot, thelocus of end points of shock compression,and the release isentropes for the aboveshock experiment.

At compressions up to the HEL, the solidresponds to shock loading elastically—thelattice compresses in the same direction as theloading stress. The crystal lattice can support ashear stress, and the maximum shear stress thelattice can support defines the HEL.

When compressed beyond the HEL, materialbehavior changes significantly. At compressionsabove the HEL no further change in shear occurs(in the perfect-elastic, perfect-plastic model),and the material is said to yield. From the pointof view of the material, a solid compressedbeyond the HEL behaves in a fluid-like mannerand starts to flow. This dramatic change inmaterial response is observed in the principalHugoniot as a cusp (B,B’). This change inmaterial behavior in the weak shock regime isakin to a phase transition; hence the transitionfrom state B to state B’. (No change in materialproperties is observed in the strong regimebecause only one shock, the plastic wave, exists.(See Fig. VI-20 for a model of the three shockregimes in solids.)

On a microscopic scale (which is still notfully understood), after state C,C’ there isrelaxation in the crystal lattice through dislo-cation movement and nucleation. The solidreleases from state C’ in two stages, first withan elastic rarefaction that moves it from peakcompression state C to some intermediatestate D at the release elastic limit, along anisentrope. Next, a second, plastic, rarefactionreturns the solid to the starting pressure,following a different isentrope to the end stateE. Because shock compression is an irreversible

process, the final state E is different from theinitial state A.

Figure VI-20 shows the three shock regimesin solids—the elastic regime, the weak shockregime, and the strong shock regime. In theelastic regime, the solid is shocked to less thanthe Hugoniot elastic limit (state B). This shockwave consists of a single elastic front. In theweak shock regime, the shock moves the

Hugoniot

1st releaseisentrope

2nd releaseisentrope

HEL

A

B,B'

C,C'

D,D'

E

Volume

Pre

ssur

eREL

Figure VI-19. Model of pressure history vsvolume, showing Hugoniot, end points of shockcompression, and release isentropesfor the shockexperiment above. As the elastic precursor takesthe particle from state A to state B,B’ at theHugoniot elastic limit (HEL), there is adramatic change in material response, when nofurther change in shear occurs and the materialis said to “yield.” This is observed in theprincipal Hugoniot as a cusp (B,B’). On amicroscopic scale, after state C,C’ there isrelaxation in the crystal lattice, and the solidreleases from state C’, first with an elasticrarefaction that moves the solid from peakcompression (state C) to some intermediate stateD at the release elastic limit (REL), along anisentrope. Next, a second, plastic, rarefactionreturns the solid to the starting pressure,following a different isentrope to the endstate E. The final state E is different from theinitial state A because shock compression is anirreversible process.


Hugoniot

HELA

B,B'

C

Volume

Rayleigh linestrong shock

regime

Rayleigh lineweak shock

regime

Figure VI-20. The three shock regimes insolids—elastic, weak shock, and strong shockregimes. Shock compression for a weak shockand a strong shock is drawn on a Hugoniot andthe trajectories are shown as Rayleigh lines.Shock velocity is proportional to the slope;stronger shocks have higher shock velocities. Inthe elastic shock regime, the solid is shocked toless than the Hugoniot elastic limit (HEL)(state B). This wave consists of a single elasticfront. In the weak shock regime, the shockmoves the material continuously to the HEL(point B), at which point the materialdiscontinuously goes to the intercept (point C).The strong shock regime is defined by Rayleighlines that have slopes equal to and greater thanthe straight line joining the HEL and the initialstate. Here the double shock structure neverdevelops, and a single discontinuous front isformed, taking the solid from state A to a stateof pressure higher than for the weak shock.

material continuously to the Hugoniot elasticlimit (point B), at which point the materialdiscontinuously goes to the intercept (point C).The final regime is the strong shock regime.Here the plastic waves acquire a velocitygreater than the elastic precursor (u2 > u1), thedouble shock structure never develops. A singlediscontinuous front is formed, taking the solidfrom state A to a state of pressure higher thanfor the weak shock.

The conditions in a release experiment arefound in nature both in high-energy-densityablation phenomena and in ultrahigh-velocitymeteorite impacts. Thus, it is clear thatunderstanding these states is important tounderstanding the shock generation process atthe NIF itself. Since much of the radiationtransport in laser-generated plasmas is in hot,expanded material, these states may provide afruitful way to generate useful states forradiation transport studies.

Isochoric Heating Experiments

The use of isochoric heating (heating atconstant density) on the NIF will enable anentirely new class of experiments at highpressure and temperature. This implies that theheating pulse will be absorbed in a time that isshort compared to hydrodynamic time scales ofexpansion. It also implies a nearly constantdeposition of energy throughout the samplevolume. Isochoric heating will allow us to reachunique high-temperature and high-pressurestates, and the high density will lead tostrongly coupled physics. Note that otheraspects of strongly coupled plasma are discussedin Section IX, Radiative Properties.

The unique states attainable by isochoricheating are in the center of a region for which nolimiting equation of state theories apply and forwhich there is simply no data at all! Thesestates fall outside the limits of low ρ, high Twhere the activity expansions are valid, andoutside the limits of high T where the Thomas-Fermi model is valid.

A promising method is to exploit theneutrons generated in a very short (<10-ps)pulse at ignition, with a total predicted energyof up to 45 MJ, mostly in 14-MeV neutrons. Weshould be able to reach temperatures in thinfoils of up to 50 eV at solid density. Even inmaterials with small neutron cross sections, itwill be possible to add small amounts of fissilematerial to increase the interaction with theneutrons without significantly changing thematerial equation of state. A free-standing foilheated in this manner will be an excellent testmaterial for spectroscopy.


Another possible method to obtain isochoricheating is to use a set of the NIF beams toirradiate metal targets at high intensity in ashort pulse. This will produce a copious amountof hard x-rays. We calculate that a 10-µm-thickaluminum foil could be heated to tens of electronvolts in this manner. In this case, it could bemounted to a low-Z stepped tamping foil;measuring the shock velocity in the tamper willallow determination of points on a uniquerelease isentrope from the initial state. Again,such a well-characterized sample will be idealfor spectroscopy of high temperature anddensity states.

X-ray Diffraction Studies

The NIF will be able to push time-resolvedin-situ x-ray diffraction into new experimentalterritory. Its higher energies and longer pulselengths will allow creation of harder, brighterx-ray sources with longer durations. These x-raysources will make available greater probedepths and experimental windows over whichthe shock compression process is measured. Thus,extensions of present techniques can then be usedto study higher-Z materials.

For example, it is not possible at present toexperimentally investigate medium-Z metalslike iron and copper, where an extensive single-crystal literature exists. Iron is especiallyimportant for geological reasons, and althoughit has been thoroughly investigated, there isstill controversy over the phase diagram.20

The study of shock-induced polymorphicphase transitions, such as those observed inpotassium chloride, is an area where long-duration x-ray sources (i.e., 10–50 ns) that arebright enough to be time-resolved are ofimportance.21 Of further technologicalimportance is understanding the effects ofshock compression on superconducting properties,shock synthesis, and decomposition of high-temperature superconductors22 and ceramics.23

All of these studies will require theextensive development of polycrystalline x-raydiffraction techniques.24 For x-ray diffractiontechniques to work successfully on polycrystal-line material, the probing radiation needs to beextremely bright. This is because diffraction

from a randomly oriented powder or poly-crystalline solid is much less efficient than froma single crystal. The lower diffraction efficiencyis caused by the random orientation of theconstituent microcrystals, which results in mostof the material not presenting the Bragg angleto the incoming x-ray source. For polycrystallinematerial, a reduction of two orders of magnitudein diffracted intensity is expected whencompared to the same material in the singlecrystal form.

In conventional x-ray diffraction analysisthis reduced diffraction efficiency can beovercome by using long time integration to makea measurement, but even so, focusing cameras arenecessary in some cases. In shock-wave physics,on the other hand, measurements must betaken on nanosecond time scales, requiringultrabright x-ray sources and the use of x-rayfocusing cameras.

In addition, the probing radiation needs tohave a large penetration depth in order tosample a large volume. The standard method isto use a Seeman-Bohlin powder diffractionspectrometer, which allows the use of a largepolycrystalline sample. Because of the longpenetration depths of the hard x-rays thatallow large volumes to be probed, powerful lasersystems with NIF capabilities will be needed tocreate the ultrabright x-ray source at the shortwavelengths preferred.

Figure VI-21 illustrates a Seeman-Bohlincamera (powder diffraction spectrometer). Withthis kind of camera, as shown schematically inthe figure, x-rays from a divergent x-ray sourceare diffracted from a large-area powder orpolycrystalline target. These diffracted x-raysare focused to a single line on the detector,helping to ease the intensity demands on thex-ray source. This increases the diffractionsignal from a powder or polycrystalline sample.

Because a large-area sample is required,large-area shocks must be generated to compressthe target. These shocks need to be uniform,which will require extremely large energies tocreate them. The need for high-flux, high-energy x-ray sources and large areas of shockedmaterial indicate that experiments of this typewill require the capability of the NIF.


R

Pointx-ray source

Powder

4ϑ

2ϑ

X-ray block

Film

Figure VI-21. The Seeman-Bohlin camera. Diffraction from a randomly oriented powder orpolycrystalline solid is much less efficient than from a single crystal. The Seemann-Bohlin cameradiffracts x-rays, which come from a divergent source, from a large-area powder (or polycrystallinetarget). These diffracted x-rays are focused to a single line on the detector (the film), helping to easethe intensity demands on the x-ray source. In a large area sample an enormous number of randomlyoriented microcrystallites are illuminated by the source, and those that are correctly aligned to theincident x-ray beam diffract onto the detector. The detector integrates (and increases) the signal.

In high-temperature superconductors, x-rayabsorption is high because of the high Z content(e.g., YBCO, BSCCO), which makes themdifficult to study. Detailed quantum mechanicalmodeling of brittle fracture in ceramic titaniumcarbide exists, and shock breakout measurementshave been shown to be one of the most successfulways of attaining and measuring ultimatetensile strains.25 This technique could beextended to study novel polycrystalline high-strength materials of the future.

The greater laser energies of the NIF can beused to generate larger compressions, makinggreater portions of the solid-state Hugoniotaccessible to the x-ray diffraction technique.Because the compressions are large (up to 50%for a few megabars for many metals) theshocked area must subtend a large angle to thex-ray source. These shocked regions will needto be 2 cm in diameter or more. Such large-areashocks must be uniform to ensure uniaxialcompression of the solid. Although uniformcompression is not possible by direct radiation,it is possible with x-ray driven ablation.26

Using long compression pulses is important toprevent the rarefaction from overrunning theshock front and eroding the compression. Withx-ray-driven shocks, experimental investigationof shock front thicknesses in the strong shockregime can be made. Knowledge of the shockfront, the thickness and shape, is extremelylimited, and beyond the temporal and spatialresolutions of most experimental techniques.

With the next-generation facility’s rapidrise time, long-duration compression wavesshould be possible using long-duration shapedlaser pulses. Using such well-defined shocksallows us to use x-rays to probe only thematerial in the shock front that has strainsbetween zero and peak compression. Diffractionat these intermediate angles must be related tothe shock front. Using x-ray streak cameraswith 1-ps time resolution, measurements withresolutions of approximately 30 atomic spacingsare feasible. At these resolutions, issues such asshock planarity and target fabrication maydominate. These critical measurements of theshock front are vital for a fundamental under-standing of shock compression.


Other Material Properties

While the emphasis above has been on basichigh-pressure properties, other properties arevery important scientifically. The electricalconductivity of materials at extreme conditionsis crucial to understanding the origin ofplanetary magnetic fields, the transient fieldspredicted in the expansion plume of meteoriteimpacts, and MHD problems in general in hotand dense matter. Thermal conductivities areessentially unknown at pressures above ≈10 GPa.(Note: we have referred to radiation transporthere and in other sections of this report—itsimportance is described in more detail inSection IX, Radiative Properties, Subsection E.)

All materials will be in hot fluid states atextreme shock pressures. For the group IVelements in particular, we note that theseliquids are believed to be semiconductors ormetallic. Their local structure isuncharacterized, but it is believed that formodest pressures above melt, the structure is nota close-packed liquid. This points out the needfor techniques to determine the distributionfunction in space of ions or atoms in the post-shock or other transient high-pressure states.

Another broad area of inquiry is the problemof non-equilibrium properties. This can beviewed as a special case of departure from localthermodynamic equilibrium (LTE), as appliedto excitations of the system in general. Weassume that areas of interest to materialproperties, such as those above, will rapidlyevolve with the advent of the NIF. Until thattime these are systems that cannot be consideredripe for investigation.

E. References

1. R. Cauble et al., ICF Quarterly Report 3, 131(1993).

2. R. Cauble et al., Phys. Rev. Lett. 70, 2102(1993).

3. B. A. Hammel et al., Phys. Fluids B 5, 2259(1993).

4. T. S. Perry et al., Phys. Rev. Lett. 67, 3784(1991).

5. L. Da Silva et al., Phys. Rev. Lett. 69, 438(1992); and D. Guenther, Nature 359, 585(1992).

6. P. Springer et al., Phys. Rev. Lett. 69, 3735(1992).

7. J. Wark et al., Phys. Rev. B 35, 9391 (1987).8. R. R. Whitlock et al., in Shock Compression

of Condensed Matter (North Holland,Amsterdam, 1991).

9. J. S. Wark et al., J. Appl. Phys. 68, 4531(1990) and Phys. Rev. B 40, 5705 (1989).

10. See, for example, Stanley P. Marsh, LASLShock Hugoniot Data (Univ. of Calif. Press,Berkeley, 1980); High-Velocity ImpactPhenomena, R. Kinslow, Ed. (Academic,New York, 1970); Ya. B. Zel’dovich and Yu.P. Raizer, Physics of Shock Waves andHigh-Temperature HydrodynamicPhenomena, (Academic, New York, 1966).

11. E. Avrorin et al., JETP Lett. 43, 308 (1986);Sov. Phys. JETP 66, 348 (1987).

12. R. Cauble et al., Phys. Rev. Lett. 70, 2102(1993).

13. Ya. B. Zel’dovich and Yu. P. Raizer, Physicsof Shock Waves and High-TemperatureHydrodynamic Phenomena (Academic, NewYork, 1966).

14. A. C. Mitchell, W. J. Nellis, J. A. Moriarty,R. A. Heinle, N. C. Holmes, R. E. Tipton,and G. W. Repp, J. Appl. Phys. 69, 2981(1991).

15. N. C. Holmes et al., in Proceedings of 8thAIRAPT and 19th EHPRL Conference, C. M.Backman, T. Johannisson, and L. Thenér, Eds.(Uppsala, 1981).

16. W. J. Nellis et al., J. Chem. Phys. 79, 1480(1983).

17. D. Saumon and G. Chabrier, Phys. Rev. A 44,5122 (1991); Phys. Rev. A 46, 2084 (1992).

18. E. Avrorin et al., J. Sov. Phys. JETP 66, 348(1987).

19. N. C. Holmes and J. A. Moriarty, ResearchMonthly, Lawrence Livermore NationalLaboratory, Livermore, California (to bepublished).

20. C. S. Yoo, N. C. Holmes, M. Ross, D. J. Webb,and C. Pike, Phys. Rev. Lett. 70, 3931 (1993).


21. D. B. Hayes, J. Appl. Phys. 45, 208 (1974);E. B. Zaratskii, G. I. Kanel, P. A.Mogilevskii, and V. E. Fortov, Sov. Phys.Dokl. 36, 76 (1991).

22. Y. Syono and M. Kikuchi, in Shock Waves inMaterial Science, A. B. Sawaoka, Ed.(Tokyo, Springer-Verlag, 1993), p. 101.

23. T. Mashimo, in Shock Waves in MaterialScience, A. B. Sawaoka, Ed. (Tokyo,Springer-Verlag, 1993), p. 113.

24. N. C. Woolsey, Time resolved, in situ, x-raydiffraction from laser shocked solids,

University of Oxford (1994); N. C. Woolsey,J. S. Wark, and D. Riley, J. Appl.Crystallogr. 23, 441 (1990).

25. D. L. Price, B. R. Cooper, and J. M. Wills,Phys. Rev. B 46, 11359 (1992).

26. T. Löwer, R. Sigel, K. Eidmann, B. I. Földes,S. Hüller, J. Massen, G. D. Tsakiris,S. Witkowski, W. Preuss, H. Nishimura,H. Shiraga, Y. Kato, S. Nakai, and T. Endo,Phys. Rev. Lett. 72, 3186 (1994).

Plasma Physics 89 Section VII

Section VII

Plasma Physics

The creation of and interaction with plasmasby high-energy lasers has a long history.1 Theirradiation of a surface with a high-energy laserproduces a blow-off plasma that can be quite hotand that contains densities suitable for theexcitation of many parametric processes. Hence,high-power lasers represent a convenient way ofproducing plasmas with which to perform laser-plasma interaction experiments.

The areas of plasma physics that have beenaddressed on high-energy lasers arise from twoindependent avenues of research. First, there arestudies of the phenomena that are created by thelaser interacting with a plasma. In this area onewould have the instabilities enhanced by laser-plasma coupling, such as stimulated Brillouinscattering, stimulated Raman scattering, etc. Inthe second there are attempts to use the laser toemulate other phenomena occurring in nature.Here the study of interpenetrating plasmas andplasma flow in a magnetic field are examples tobe presented.

A. Interpenetrating Plasmas

Plasma interpenetration occurs in beam-plasma systems and in laser ablationexperiments, and in the context of geophysicsand astrophysics. Therefore, interpenetration ofcounter-propagating plasmas can be considered aproblem of fundamental interest in plasmaphysics. Single-fluid computer models arefundamentally unsuited for the study ofinterpenetration of counter-propagating plasmasbecause such codes enforce stagnation, resultingin unphysical temperatures and shock formation.

Although methods for modeling the collisionlessregime are well developed, the intermediateregime is more difficult to model.

Recently, new computational methods havebeen investigated and applied to the simulationof colliding plasmas formed by laser ablation ofparallel discs or foils. One modeling technique isto follow multiple fluids that are coupled bymeans of self-consistent electromagnetic fields aswell as through collisional interaction.

A different approach is to augment kineticparticle-in-cell calculations with algorithms forcollisions. A typical result from this latter methodis presented in Fig. VII-1, which shows a calcula-tion of two laser-heated parallel plastic foils. Thefigure shows the ion density for each foil at 800ps after initiation of the laser, clearly illustratingthe interpenetration of the plasmas into oneanother. The particle phase space view (A) showsthe slowing down and heating in the center as thetwo plasma flows gently stagnate. The particlespace result (B) shows that the interpenetration isstopped by collisions.

Figure VII-2 shows the schematic for anexperiment performed on two solid targets. Thetwo opposing laser beams impinge on the targets(one of aluminum and one of silicon) and createplasmas that expand toward each other. (Thetargets are of aluminum and silicon to permitspectroscopic identification of the plasmapositions.) Measurements are made with a four-frame gated microchannel plate coupled to afiltered pinhole, so the images represent a two-dimensional map of the emission. Figure VII-3shows an x-ray image of the experiment shownschematically in Fig. VII-2.

Section VII 90 Plasma Physics

1.2

0.6

0

–0.6

–1.2

–1000 0 1000

x (µm)

V x (

108

cm/s

)

1.0

0.8

0.6

0.4

0.2

0

–1000 0 1000

x (µm)n i (

1020

cm

–3)

B) Particle space resultA) Phase space view

Figure VII-1. Simulation of an interpenetrating plasma showing two exploding carbon foils,originally 3.2 µm thick, separated by 1500 µm. A) The phase space view, showing the stopping of theinterpenetration by collisions. B) The particle space result, showing ion density vs position, measuredfrom the center. The overlapping of the two curves clearly shows interpenetration. Both views are at800 ps after initiation.

CH substrate

Al dot Si dot

Gated x-ray pinhole

Figure VII-2. Schematic of colliding plasmaexperiment using two opposing beams toirradiate two facing solid targets. One target isaluminum (on the left) and one is silicon (on theright). The dots have a radius of 250 µm, thelaser spot has a radius of 500 µm, and thedistance between the dots is 800 µm.

Figure VII-3. X-ray image of colliding plasmaexperiment shown in Fig. VII-2. The image wastaken with a four-frame gated microchannelplate coupled to a filtered pinhole.


B. Plasma Streaming inMagnetic Fields

The propagation and stability of collimatedstreams or jets of plasmas in magnetic fields iscentral to a number of important physicalproblems, including beam heating ofmagnetically confined thermonuclear plasmas,2

the interaction of the solar wind with planetarymagnetospheres,3 and the formation andequilibrium properties of extragalactic jets.4

High-energy laser experiments conducted attotal beam energies below 0.3 kJ havedemonstrated the feasibility of studying high-energy jet phenomena with laser-producedplasmas.5 A cylindrical glass hohlraum, 3000 µmlong by 1000 µm diameter, was irradiated withone or two beams of a laser operating at 1.06 µmwavelength. The laser had a pulse length of 2 nsand energies ranging from 0.03 to 0.3 kJ. Thehohlraum was located at the center of a 24-cmdiameter Helmholtz coil, which provided aspatially uniform and temporally constantmagnetic field up to 10 kG over regions andtimes of interest. The hohlraum was orientedwith its axis perpendicular to the appliedmagnetic field.

The laser-produced plasmas from the insideof the cylindrical glass hohlraum were observedto evolve into plasma structures that werestrongly collimated in the direction transverse tothe hohlraum axis and the magnetic field, but jet-

like in the direction parallel to the hohlraum axis.These jets were observed to propagate across themagnetic field lines at their initial velocities viathe mechanism of an E × B drift. Shear in thevelocity field of the jet led to the evolution of aclassic Kelvin-Helmholtz-like instability on theedges of the plasma.

Figure VII-4 shows an optical image taken ofone such expansion of a laser-produced plasmastreaming into a 10-kG magnetic field. Theplasma is emerging from a hohlraum irradiatedwith a single beam at 0.1 kJ. The image was takenat 850 ns after the laser pulse started, and wasobtained using an ultrafast optical framingcamera with 2-ns time resolution. In thisexperiment, the hohlraum end opposite tothe laser-entrance hole was closed using anepoxy plug.

At 850 ns after the laser was incident on thetarget, the maximum plasma jet velocity is≈8 cm/µs and a dramatic vortex structure hasbeen produced. Both the jet formation and theobserved instability may be relevant to a range ofastrophysical phenomena.

C. Laser-Plasma Instabilities

The coupling of high-intensity laser light toplasmas has been the subject of experimentalinvestigations for many years.6 Theseexperiments have focused on measuring a broad

B-field

Hohlraum edge

Laser

Figure VII-4. Optical image of a plasma streaming into a magnetic field. The plasma jet is emergingfrom a hohlraum on the right into a B-field that is coming out of the plane of the figure. The imagewas taken at 850 ns after the laser pulse started, using an ultrafast optical framing camera with 2-nstime resolution. Note the dramatic vortex structure.


range of phenomena such as resonance andcollisional absorption, filamentation, densityprofile and particle distribution modification, andthe growth and saturation of various parametricinstabilities. These phenomena depend on boththe properties of the laser (intensity, wavelength,pulse length coherence, etc.) and the compositionof the plasma.

Experimental studies of laser-plasmainstabilities have become particularly importantin recent years as a result of the vigorous researcheffort in laser-driven inertial confinement fusion(ICF). The success of ICF depends partlyon mitigating the undesirable effects oftwo particular parametric instabilities,stimulated Raman scattering and stimulatedBrillouin scattering.

These two instabilities are of particularimportance because both degrade the targetcompression efficiency in a spherical implosionexperiment. Electron Landau damping from thestimulated Raman scattering instability producesfast electrons that can preheat the core of animploding sphere prior to the arrival of thecompression shock front.7 The stimulatedBrillouin scattering instability can scatter asubstantial fraction of the incident laser light,causing an overall reduction in the laser-to-x-ray drive efficiency and modifying the x-raydrive symmetry.

In addition to its importance for ICF, high-intensity laser-plasma coupling presents anextraordinarily rich topic in the study of high-energy-density physics. For example, laser-produced plasmas provide a unique environmentfor the study of collisional and resonanceabsorption of laser light. Numerous experimentsusing various kinds of target materials andwidely varying laser parameters have verifiedthe general features of collisional absorption,such as the dependence of absorption on plasmatemperature, scale length, laser wavelength andintensity.8,9

Experiments investigating resonanceabsorption, which occurs at the critical densitysurface, show an expected dependence on theangle of incidence and polarization of thelaser.10,11 However, these experiments also show

some discrepancies in the absorbed energy thatmay be attributed to rippling of the criticaldensity (Ncr) surface. Density profilemodification has been observed in experimentswhere resonance absorption is the dominantcoupling mechanism.This profile modificationcan lead to harmonic generation in back-reflectedlaser light.12

Laser-produced plasmas provide a uniqueenvironment for the study of parametricinstabilities. These instabilities can most simplybe described as the resonant coupling of theincident laser light into two other plasma waves.Experiments have been done on stimulatedRaman scattering instability (discussed above inthe context of ICF), where an incident photondecays into an electron plasma wave and ascattered photon. These experiments haveshown some important trends, such as thegeneration of hot electrons from stimulatedRaman scattering and the dependence of theinstability on plasma scale length, laser intensity,and electron collisionality.

Figure VII-5 shows the region in a densitygradient where stimulated Raman scattering andBrillouin scattering play a role. The figure showsthat occurrence of Brillouin scattering,filamentation, and inverse bremsstrahlung ispossible at places below the critical density,while the wave-matching conditions indicate thattwo-plasmon decay and stimulated Ramanscattering occur at and below the quarter criticaldensity point.

The occurrence of the stimulated Ramanscattering and the two-plasmon decay isseparated into two density regimes. The firstregion is near quarter critical density, where thestimulated Raman scattering and two-plasmondecay can occur together, with both beingabsolute instabilities. This means theseinstabilities can grow, saturate, create ion waves,and create density increases. Then the wavescease to satisfy the wave-matching conditions. Atthis point the profiles relax, allowing the processto re-initiate. However, in the lower-densityregime the stimulated Raman scattering is not sostrongly damped and the process is convective.


Ne /

Nc

r

Distance

0.25

1.0

Brillouin, filamentation, inverse bremsstrahlung

Two-plasmon decay

Raman

Figure VII-5. Schematic of regions where laser-plasma instabilities can occur in an electron-density gradient. Electron density (in units ofthe critical density) for an idealized gradient isshown as a function of distance. The occurrenceof Brillouin scattering, filamentation, andinverse bremsstrahlung is possible in regionsbelow the critical density, while the wave-matching conditions indicate that two-plasmondecay and stimulated Raman scattering occur atand below the quarter critical density point. Theoccurrence of the stimulated Raman scatteringand the two-plasmon decay is separated intotwo density regimes. The first is near quartercritical density, where the stimulated Ramanscattering and two-plasmon decay can occurtogether, with both being absolute instabilities.In the lower-density regime the stimulatedRaman scattering is not so strongly damped andthe process is convective.

Stimulated Raman Scattering

The wave-matching conditions for stimulatedRaman scattering are shown schematically inFig. VII-6. Here the incident laser interacts withthe plasma, stimulating a scattered light waveand an electron plasma wave.

Figure VII-7 shows an example of the time-resolved stimulated Raman scattering spectrumarising from a CH2 exploding-foil target. The foilwas irradiated with a temporally square 1.0-nspulse of 0.35 µm light with an intensity of~3 × 1015 W/cm2. The laser burned through thetarget during the pulse, producing anapproximately parabolic density profile. This isseen in the figure, where the maximum density atwhich the stimulated Raman scattering occurs

Incident laser(ωL, kL)

Scattered light wave(ωR, kR)

Electron plasma wave(ωpe, kpe)Plasma

ωL = ωR + ωpekL = kR + kp e

Figure VII-6. Wave-matching conditionfor stimulated Raman scattering. The incidentlaser interacts with the plasma, stimulatinga scattered light wave and an electronplasma wave.

1.5

1.0

0.5

0.0

–0.5400 500 600 700 800

Fiducial

Tim

e (n

s)

Wavelength (nm)

Figure VII-7. Contour plot of intensity ofRaman-scattered light vs wavelength and timefor a CH2 exploding-foil target. It can beinferred that the maximum density at which thestimulated Raman scattering occurs drops from~0.22 to 0.09 Ncr as the foil explodes. The insertshows the data image from a streak cameracoupled to a visible spectrometer. Theexperiment has a thin plastic foil irradiated by a0.35-µm laser at 1015 W/cm2. Note the timingfiducial used to indicate time relative to theincident laser pulse.

drops from ~0.22 to 0.09 Ncr as the foil explodes.Further, the time dependence of the short-wavelength cutoff implies a heating of theelectrons followed by cooling as the foil expands.However, the details of this instability (such asthe frequency spectrum, angular distribution of


the scattered light, saturation mechanisms,threshold intensities, and starting noise level forthe instability) are not well understood.

Two-plasmon decay instability occurs whenthe incident laser light decays into two electron-plasma waves. Experiments utilizing Thomsonscattering have shown that the wave number formaximum growth is the same as that predictedby linear theory.13 These same experimentsshowed large levels of (short-wavelength) ionfluctuations that correlated with the two-plasmon-decay Langmuir waves. Meanwhile,experiments have examined the saturation andlong-time-scale evolution of this instability.14

Stimulated Brillouin Scattering

The stimulated Brillouin scattering instabilityoccurs when the incident laser decays into an ionacoustic wave and a scattered light wave. Therehave been a number of interaction experimentsthat have produced substantial backscattered andsidescattered light that is attributed to thestimulated Brillouin scattering instability.8,15

The intensity of the scattered light seems todecrease as the wavelength of the interactionlight decreases.

The temporal growth rate, spatial behavior,and saturation characteristics of stimulatedBrillouin scattering have been investigated usingThomson scattering and measurements of theback-reflected light.16 The exponential growth intime of the scattering was found to be in goodagreement with linear theory. Experimentsinvestigating the nonlinear effects of stimulatedBrillouin scattering have shown ion acousticwave harmonic generation and a rapidly varying100% modulation of the ion wave.17 Thestimulated Brillouin scattering process has beenstudied in the past because of its creation of alarge reflected wave, which diminishes thecoupling of an intense laser.

Figure VII-8 shows the wave-matchingconditions for stimulated Brillouin scattering. Thephysical mechanism that generates the reflectedwave is shown in Fig. VII-9. An impinging laserof electric field, EL, interacts with initial densityfluctuation, δn, to produce a transverse currentproportional to δnEL. This transverse currentproduces a reflected light wave with a field ER.

Incident laser(ωL,kL)

Scattered light wave(ωR,kR)

Ion sound wave(ωia ,k ia)Plasma

ωL= ω

R + ωia

kL = kR+ kia

Figure VII-8. Wave-matching condition forstimulated Brillouin scattering. The stimulatedBrillouin scattering process creates a largereflected wave, which diminishes the couplingof an intense laser.

J ∝ e δn EL

ER

F ∝ ∇ EREL

Ion wave ∝ δn

Figure VII-9. Diagram of physical mechanismthat generates the reflected wave. An impinginglaser of electric field, EL, interacts with initialdensity fluctuation, δn, to produce a transversecurrent proportional to δnEL. This transversecurrent produces a reflected light wave with afield ER. Then the ponderomotive force due tothe incident and reflected light waves, ELER/8π,can in turn enhance the density fluctuation, δn.This enhancement will grow as long as thewave-matching conditions shown in Fig. VII-8are met.

Then the ponderomotive force due to the incidentand reflected light waves, ELER/8π, can in turnenhance the density fluctuation, δn. Thisenhancement will grow as long as the wave-matching conditions are met.

Stimulated Brillouin and Raman scatteringare three-wave processes that occur when theincoming light photon (of frequency ω0 and wavenumber k0) couples to a scattered light photon


and a plasma wave. The conservation equationsthat govern the interactions are:

ω0 = ωscat + ωpw (energy conservation)k0 = kscat + kpw (momentum conservation)

The subscripts scat and pw refer to the scatteredlight photon and the plasma wave respectively.

Brillouin scattering with its reflectedwave has many interesting features that willbe addressable at the NIF. Among these aresaturation mechanisms; ion heating; andinstability competition between, for example,stimulated Raman and stimulatedBrillouin scattering.

Some typical spectra of back-reflected lightfor a glass microsphere and a plastic diskirradiated with 1.06-µm light are shown inFig. VII-10. Here the spectrum is red-shifted forthe higher-Z target shown, while the spectrum isblue-shifted for the low-Z plastic target. Thespectral shift has been measured with differentangles of incidence in order to estimate soundspeed and plasma expansion velocity. Note thatexperiments using spatial and temporal laser-beam smoothing techniques, mentioned in thedescription of the laser facility, have shown asignificant reduction in the level of backscatteredemission from both stimulated Brillouin andstimulated Raman scattering instabilities.18

More than one parametric instability mayoccur at the same location in a plasma, and thesecan interact with each other. One experimentinvestigating the interaction of stimulated Ramanscattering instability and stimulated Brillouinscattering showed a strong correlation betweenthe quenching of stimulated Raman scatteringinstability plasma waves and the initiation ofstimulated Brillouin scattering ion waves.19

Another experiment showed that thespectrum of backscattered light from stimulatedRaman scattering instability depended on thepresence of stimulated Brillouin scattering.20 Thedata in these experiments suggest an explanationfor the temporal evolution of the “Raman gap,”an observationally missing stimulated Ramanscattering instability signal that would beassociated with regions of electron density from~0.18 to 0.25 Ncr. The data showed that the“Raman gap” could be explained in terms of thestimulated Brillouin scattered light. Although

–10 0 10

Cou

nts

20

–20 0 20–40

B) Low-Z plastic target

A) Higher-Z target

Red shift

150

100

50

0

300

200

100

0

500

400

Shift in Å from laser line

Blue shift

Shift in Å from laser lineC

ount

s

Figure VII-10. Spectra for the back-reflectedlight due to stimulated Brillouin scattering. InA) the material is a glass microsphere (thehigher-Z target), while in B) it is a parylene-Cdisk (the low-Z plastic target). The spectrum forthe higher-Z target is red-shifted, while for thelow-Z target it is blue-shifted.

these experiments do not completely explain thegap, the results are consistent with theassumption that an interaction exists between thetwo instabilities.

Filamentation Instability

Filamentation instability occurs when a smallhot spot in the laser-intensity profile undergoesself-focusing as a result of thermal, pondero-motive, and/or relativistic effects. The result isthat electrons, and eventually ions, are expelledfrom the filament, lowering plasma density andcausing the laser light to focus more tightly. Thiscreates an unstable feedback loop where a tighterfocus leads to higher intensity and more densitydepletion. Eventually, the instability saturates


through diffraction effects, thermal absorption,or laser-light scattering by stimulated Brillouinand Raman scattering instabilities.

Filamentary structures have been inferredfrom density perturbations measured usinginterferometry, shadowgraphy, and dark-fieldimaging, and by imaging second and three-halves harmonic emission.21,22 Other experimentsdesigned to study the growth of a well-definedintensity perturbation show an increase in theperturbation of spatial density that is consistentwith ponderomotive filamentation.23

Figure VII-11 shows an example of laser-induced filamentation results for an experimentusing a long scale-length plasma. The plasma wasformed by focusing a 4-ns pulse of 1.05-µm lightonto a plastic foil. The pulse had a spot size of~1 mm and a peak intensity of ~6 × 1012 W/cm2.The plasma was then irradiated with a second,more intense, 300-ps pulse of 1.05-µm light, thisone with a peak intensity of ~8 × 1014 W/cm2,creating a second-harmonic emission. The figureshows an image of the pattern of the second-harmonic emission.


In the past, laser-produced plasmas have notbeen sufficiently uniform, and so the interactionphysics, which is extremely sensitive togradients, has been clouded. Some of the plasma-physics phenomena that will be studied with alarge laser like the NIF are grouped here fordiscussion. In some cases, knowledge of thebehavior of the phenomena is quite extensive anddeeper understanding is sought. In other cases,the phenomena have not been addressedexperimentally at all. This difference in treatmentis due to the level of sophistication we now havein the various areas.

The experiments envisioned in the presentdiscussion would require good focal-spotfocusing capability—as small as 50 µm FWHM—and a dedicated interaction beam line with veryhigh or low f/number as well as the potential tooperate with laser wavelengths at first, second,and third harmonics. The experiments wouldalso need some type of Thomson-scattering probebeam plus collection optics. These experiments

20 µm

Figure VII-11. Image of laser-induced filamentation. The second-harmonic emission is created byirradiating a preformed plasma with a second, intense 1.05-µm laser. The left image (with the scale)shows a magnified section of one of the filamentary structures. The image on the right shows thetarget surface at the top together with an entire spatial record.


require the energy of the NIF to produce the largepreformed plasma needed to obtain largeinteraction lengths.

The NIF’s proposed 1.8-MJ laser will allowthe production of a plasma at an electrontemperature of 3 keV with an 11-mm diameterand a plasma at 5–6 keV with a 6-mm diameter.(With smaller plasmas, the NIF will be able toproduce even higher electron temperatures.) Thelarger plasmas will translate into longerinteraction lengths, greater homogeneity, highertemperatures, longer time scales, and reducedvelocity gradients. Thus, the NIF is central to theexperimental possibilities discussed below.

Laser-Plasma Interactions in Large,Hot Plasmas

Long Scale-Length Plasma ProductionSeveral kinds of diagnostics can be used to

study laser-plasma interactions in large, hotplasmas. One primary diagnostic forfilamentation would be stimulated Brillouinscattering signal, which is indicative of lowerdensity in the filaments that could be coupledwith high spatial resolution (on the order of1 µm). Other diagnostics would be XUV imagingand high-resolution optical probing with 4-ω0light. Further, the study of the angulardistribution of scattered light would provide anindication of filamentation.

Thomson scattering could be used for theseinvestigations, thus making highly localizedmeasurements of plasma temperature anddensity by scattering from thermal electron-density fluctuations in the plasma. In addition tousing Thomson scattering for thermalmeasurements, it is possible to reach far deeperunderstanding of the physics of laser-plasmainstabilities by using Thomson scattering tomeasure the coherent motion of electronsinvolved in ion acoustic and electron plasmawaves. This measurement can provide atemporally and spatially resolved measure of thecoherent fluctuation amplitude in a specificdirection determined by the angle made by thedetector, scattering volume, and scatteringlight source.

Measurement of the background fluctuationlevels would provide information about theinitial level of coherent fluctuations, theamplification of these levels with the applicationof a pump or interaction beam, and the saturationlevel of the fluctuations. In addition, it couldprovide useful information about the couplingbetween stimulated Raman and stimulatedBrillouin scattering by means of simultaneousmeasurement of the ion acoustic and electronplasma waves at the same spatial location.

Measurement of optical-wavelengthThomson scattering is well-suited for studyinglow electron densities such as are encountered inthe corona, whereas x-ray wavelength Thomsonscattering is better-suited for studying plasmadensities that exceed the critical density forstandard visible and near-UV lasers. Thepossibility of developing an x-ray Thomsonscattering measurement to study coherentplasma motion in high-density plasmas is anexciting one.

Present open-geometry plasmas, called“gasbags,” have produced large (millimeter-scale) plasmas at interesting electrontemperatures. A schematic of the experimentsthat produced these large-scale plasmas is shownin Fig. VII-12, along with the diagnosticcomplement that is available. The gasbag isformed of two plastic films of thickness 0.6 µmclamped by a hoop to form a bag. Pressure in thegasbag is stabilized using a pressure transducerat densities such that the fully ionized speciesyield electron densities of ~1021 cm–3. There is aseparate interaction beam (indicated by the lightblue arrow) to drive the instabilities in acontrolled way. For the NIF this geometry shouldbe directly scalable to approximately 4 cm3, using9/10 of the NIF laser to form the plasma and1/10 to create interactions. The uses for this typeof large uniform system range from plasmainteraction to nuclear reaction rate studies.

Figure VII-13 shows the time history, in two-dimensional x-ray images, of the emission of agasbag. The sequence of images demonstrates theuniformity of the plasma in a gasbag. Imageswere taken at three times during the heatingpulse, using a gated x-ray imager, and represent


frame times of 100 ps each. At initiation of thelaser the plasma is a glowing emission of theouter shell of the plastic membrane bag. At425 ns, after the laser has heated the plasma, it isuniform. By 875 ps, the gas can be seen to beextremely uniform. This gasbag technique holdsmuch promise with large laser systems like theNIF because scaling of this kind of experiment tolarge systems is straightforward.

There have been many plasma physicsexperiments that have been virtually impossibleto interpret quantitatively. Coupling thecapability to make large, uniform hot plasmaswith the ability to control perturbations of theplasma (for example, by the use of higher densitymaterials nearby or controlled flow velocities)

will allow us to study the fundamental aspects ofthese experiments. As an example one can look atthe plasma-physics experiments that focus on thehohlraum environment.

Stimulated Brillouin and Raman ScatteringAs mentioned above, the study of laser-

plasma interactions and parametric instabilitiessuch as stimulated Raman scattering (SRS) andstimulated Brillouin scattering (SBS) are of greatinterest to many researchers.24,25

Plasmas can support different types ofwaves. In stimulated Brillouin scattering, an“ion-acoustic” wave is created, which is thecollective motion of the ions and electrons in asound wave. These have substantial inertia.

Gated spectrometer(SXRFC)

Gated imaging(GXI-2)

Streaked imaging(SSC)

Gated imaging(WAX)

Streaked spectrometer (SSC)

Interaction beam

Laser

Streaked high-resolutionspectrometer

(HIX)

Gated imaging(GXI-3)

Laser

Figure VII-12. Schematic of the formation of a large-volume hot plasma. The figure shows both thegasbag used to create a large-scale plasma, and the diagnostic complement that is available. Thegasbag (light gray sphere) is formed of two plastic films of thickness 0.6 µm. The hoop (shown indarker gray) holds the two plastic films clamped to form the bag. The gas bag is filled through twotubes in the hoop. Pressure in the gasbag is stabilized using a pressure transducer at densities suchthat the fully ionized species yield electron densities of ~1021 cm–3. The roughly spherical volume ofthis plasma is 0.066 cm3, with a radius of 2.5 mm. The gas in the volume is heated to temperatures thatare on the order of 3000 eV. The diagnostic complement access is indicated by the light gray arrows,while the heating lasers are shown on either side in black. There is a separate interaction beam(indicated by the medium gray arrow pointing in) to drive the instabilities in a controlled way. For theNIF this geometry should be directly scalable to approximately 4 cm3, using 9/10 of the NIF laser toform the plasma and 1/10 to create interactions.


In stimulated Raman scattering, an “electron-plasma” wave is created, which is the collectiveoscillation of the plasma electrons at the plasmafrequency, oscillating about the stationary ions.Beating between the incident and scattered lightwave serves to enhance the plasma wave,producing more scattering of the incident light.The result is a feedback loop that, in times on theorder of picoseconds, produces an exponentialincrease in scattered signal. This signal saturatesat anywhere from 10 to 100% of the incident light.

Processes such as Brillouin and Ramanscattering are obviously important from the pointof view of transporting laser light effectivelythrough a plasma, as is the case in inertialconfinement fusion. The scattering of light atlarge angles is detrimental to the symmetry ofICF experiments because it results in energybeing deposited in the wrong place. Anotherexample is the effect of SRS on an electron-plasma wave. The by-product of stimulatedRaman scattering is the conversion of theelectron-plasma wave energy into extremely fastelectrons, from 50 keV to 1 MeV energies. Thesefast electrons not only can transport energy

away from the experiment, but can also deposit itwithin a target, giving rise to undesirableheating.

Because laser-plasma instabilities are thesubject of numerous studies, and instabilitiesform a challenging area of plasma physics, theNIF’s large uniform plasmas and laser-pulseshaping coupled to the diagnostics possible onthe NIF make it an ideal site for studyingthese processes.

Sidescatter is of major interest to plasmaphysicists. This is because in strong gradientswith large laser spots, sidescatter is predicted todominate backscatter. Further, the behavior ofsidescatter will be different in flowing andstationary plasmas. Theoretical and experimentalunderstanding of sidescatter is incomplete, sothis is an area that will benefit greatly fromexperiments on the large uniform, controllableplasmas that the NIF will afford.

FilamentationFilamentation instability occurs when a small

hot spot in the laser intensity profile undergoesself-focusing as a result of thermal,ponderomotive, and/or relativistic effects.26 The

1 mm

0 ps 425 ps 875 ps

Figure VII-13. Time history of emission of a gasbag, demonstrating the uniformity of a large scale-length plasma. X-ray emission images were taken at three times during the heating pulse, using agated x-ray imager, and represent frame times of 100 ps each. The plasma starts off at laser initiation (0ps) as a glowing emission of the outer shell of the plastic membrane that forms the bag. At 425 ns, afterthe laser has heated the plasma, it is uniform, with a slight enhancement of emission from themembrane. At 875 ps the gas is extremely uniform. The white strip in the middle of the gasbag is thehoop that holds the plastic membrane together (see Fig. VII-12). This gasbag technique holds muchpromise with large laser systems like the NIF because scaling of this kind of experiment to largesystems is straightforward.


result is that electrons, and eventually ions, areexpelled from the filament, causing the laser lightto focus more tightly. This creates an unstablefeedback loop whereby a tighter focus leads tohigher intensity and more density depletion.Eventually, the instability saturates due todiffraction effects, thermal absorption, or laserlight scattering by stimulated Brillouin andRaman scattering instabilities.

The growth of filamentary structures can bedetermined from the width and length of hotspots, known as speckles, in the incident beam.The filamentation can be described in terms of agrowth rate along the length of a speckle; longerspeckles are more likely to produce focusing ofthe hot spot in a filament.

The length of a speckle is given by 8f2λ,where f is the f/number of the beam and λ is thewavelength of the laser light. Because the lengthof a speckle is dependent on them, the f/numberand wavelength of the incident laser beam arepowerful levers for modifying filamentation,provided that the speckle lengths are smallerthan the scale length of the plasma. By usinglarge, uniform plasmas on the NIF, it will bepossible to produce very large filaments,allowing the study of filamentation over a widerange of wavelengths and incident-beamf/numbers.

Understanding the scattering of incident laserlight by parametric instabilities and the processof filamentation is of interest to ICF. However,there are many aspects of these instabilities thatare not adequately understood and are of interestto scientists for reasons beyond those generatedby fusion studies. Understanding of thesaturation mechanisms for these instabilities,together with the effects of different types ofbeam smoothing on filamentation, are subjectsthat challenge our ability to model these plasmas.

The nonlinear processes that determinesaturation of the plasma-wave amplitudes andthe way in which depletion of the incident beamlimits filamentation and plasma-wave growth aresubjects that will benefit from study over abroader range of parameter space (temperatures,densities, and gradient scale lengths) than thatexpected from ICF targets alone. It is expectedthat on the NIF it will be possible to extend the

parameter space in terms of laser intensity spatialdistribution, f/number, pulse length, andwavelength beyond that possible at any existinglaser facility. We can explore filamentation over abroad range of experimental parameters byvarying the color and f/number of thefilamentation beam and by varying the plasmaconditions (e.g., temperature, density, andaverage Z).

Short-Pulse High-IntensityLaser-Plasma Interactions in Large,Hot Plasmas

There is widespread agreement that the NIFshould include a beam line for short-pulse, high-power experiments. This capability is especiallyimportant for many basic plasma-physics studies.These include relativistic, ultrahigh-intensityregimes of laser-matter interaction; generation ofastrophysically relevant plasmas; fast-ignitorphysics27; high-gradient accelerator schemes28;new x-ray lasers; and high-temporal-resolutiondiagnostics.

The fast-ignitor laser-fusion conceptcombines many interesting high-power, short-pulse phenomena. These include generation andtransport of high electron fluxes in a high-densityplasma,29 intense harmonic generation at thecritical density, ultrahigh B-fields, pondero-motive effects, relativistic self-focusing andfilamentation, laser-beam channeling, and holeboring.30

A schematic of the fast-ignitor scheme forheating compressed matter is shown inFig. VII-14. In the first step of the fast-ignitorscheme, an implosion of a sphere filled with gasresults in a core of gas with very high densities(600 g/cc). Next, a laser beam creates a channel inthe sphere by pushing the critical density surfacetoward the core, after which the final heater orignitor beam is turned on. The ignitor beaminteracts with the density gradient generated bythe channeling beam and generates hot electronsat MeV energies, which penetrate into thecompressed gas core and cause an instantaneousrise in the local temperature of the core. Theentire concept requires a NIF-like laser because itrequires a high compression and two additionallaser systems that are both significant. Adding


e-

C) Ignitor laser beamB) Channeling laser beamA) High-compression implosion

Compressedmaterial

Pusher

Figure VII-14. Schematic of concept of fast-ignitor scheme for heating. First, in step A there is a high-compression implosion of a sphere filled with gas, resulting in a core of gas that will be at densities of600 g/cc. Step B requires that a laser beam with a pulse duration of 100 ps and an intensity of 1018

W/cm2 create a channel by pushing the critical density surface toward the core. Finally (step C), thefinal heater or ignitor beam is turned on. This beam is a 5-ps beam that has an intensity of 1020 W/cm2.The ignitor beam interacts with the density gradient generated by the channeling beam and generateshot electrons at MeV energies. These electrons penetrate into the core of the compressed gas and causean instantaneous rise in the local temperature of the core.

the need to diagnose the system makes clear theneed for the NIF in order to perform theexperiment.

The high-gradient accelerator schemesinvolve relativistic self-focusing andfilamentation,31 and parametric instabilities suchas stimulated Brillouin and Raman scattering instrongly driven and transient regimes.32

This short-pulse regime is also moreamenable to detailed simulation (that is, 2Dparticle-in-cell codes can be run for times relevantto the entire experiment and include the entireregion of interaction).33 It is also more amenableto systematic exploration of linear and nonlinearbehavior. Short-pulse experiments would requirethat a short-pulse beam with the capability forgood focusing (i.e., on the order of 20-µmdiameter) be added to the NIF. Theseexperiments require the NIF in order to producethe large preformed plasma needed to accesslarge interaction lengths.

The high-gradient accelerator schemes usethe fact that a plasma can support much higherelectromagnetic fields than can conventionalaccelerators such as those used in elementaryparticle physics experiments. As a result the high-gradient accelerator would be much morecompact and potentially cheaper.

A number of novel high-gradient acceleratorschemes34 have been proposed and studied atlaser powers not high enough to produce the

desired relativistic electron quiver velocity(velocity induced by the oscillating fields of thelaser). Relativistic self-focusing is central to theacceleration scheme to maintain the high laserintensity over many Rayleigh lengths. Otherprocesses such as stimulated Raman andBrillouin scattering or filamentation can break upthe laser beam and degrade the efficiency of thescheme. If these high-gradient acceleratorschemes prove successful, applications to tunablesources of x-rays are also envisioned.

Study of Very High Magnetic Fields

A further area of focus will be the study ofmagnetic fields generated by the temperature anddensity gradients present in plasmas. In laser-produced plasmas the large magnetic fields canbe generated in many ways.35 For example, oneproposed scheme is to generate extremely largemagnetic fields by intentionally creating acirculating current path for Raman-generated hotelectrons.36 This might be done by using anumber of beams to irradiate a preformedplasma, such as in a gasbag. Simple estimatesshow that currents up to 108 amperes might begenerated using the NIF, and magnetic fields aslarge as 108 gauss could result. Even if thecurrents and fields are limited to substantiallysmaller values, this scheme might generate alaser-induced toroidal pinch with intriguingapplications.


Calculations have indicated that magnetic fieldsgenerated by laser-produced plasmas can be aslarge as 100 MG.37 We may be able to useFaraday rotation of an optical probe beam andZeeman splitting of spectral lines to investigateand quantify these magnetic fields. By inhibitingthermal conduction, magnetic fields influencetemperature distributions and thereby influencestimulated Brillouin and Raman scattering,filamentation, and propagation of laser beams inplasmas. We may also be able to study the effectof magnetic fields on atomic energy levels bymeans of simultaneous x-ray spectroscopy. Oneinteresting possibility is to use the Weibelinstability driven by intentionally inducedanisotropy in the electron velocity distribution.

These experiments would require the NIFin order to create significant (and hencediagnosable) volumes of plasma at the sametime that extremely large magnetic fieldsare generated.

Generation of Large Fluxes ofHigh-Energy Electrons (50–100 keV)

On the NIF it would be possible to study thedeliberate production of fast electrons throughstimulated Brillouin scattering and two-plasmondecay mechanisms in plasmas that are at 0.1 to0.25 of the critical density.38 In particular, theseplasmas could be preformed from gas-filledtargets or foams.

It would also be possible to study theconversion of high-energy electrons (50–100 keV)to x-rays. Appropriate high-Z layers would beplaced near the target. The fast electrons couldthen deposit their energy in the high-Zmaterial and the resultant x-ray emission wouldbe studied.

It might be possible to develop a compacthigh-energy x-ray source that could be used forbacklighting/probing of other experiments.

Finally, the enormous currents created bythese methods would generate magnetic fields,and the study of these magnetic fields would alsobe of interest—see Section IX, RadiativeProperties for a discussion of strong magneticfield effects.

Interpenetrating Plasmasand Turbulence

The study of the interaction, orinterpenetration, of two high-temperatureplasmas has received limited attentionexperimentally.39 Apart from its intrinsic interest,the interaction of plasma streams is part of thestandard ICF experiments because of thegeneration of plasma motion inside a hohlraum.One particular interest is the equipartition ofenergy between the two groups of ions, and inthis regard it has been calculated that ions canachieve very high temperature through collisions.

Moreover, the inter-streaming plasmas canprovide the ideal system for generating andstudying turbulence in plasmas. This wouldmake it possible to study the generation of two-stream ion-ion turbulence or the ion Weibelinstability. The ion Weibel instability is of interestbecause it creates magnetic fields with λ ~ c/ωpibut has a slow growth rate of ~ωpiu/c, where ωpiis the ion plasma frequency and u is the relativevelocity of the interpenetrating flows.

Producing high-temperature plasmas thatcan collide with sufficiently high density forinteractions to take place requires large amountsof laser energy. Nova and other present-day lasersystems are severely limited in the amount oflaser energy available. Current laser facilities arealso not capable of creating plasmas large enoughto permit using diagnostics to probe the region ofinterpenetration with sufficient spatial resolution.Although some preliminary work could beperformed on present-day lasers, the energiesavailable in the NIF will make these studiesfar more attractive, with quantitative resultsbeing feasible.

The study of ion turbulence will require adiagnostic such as Thomson scattering that willpermit the direct measurement of ∆n/n of the ionsas a function of time. The diagnostic should allowthis measurement to be made in differentdirections, thus permitting us to follow therandomization of the ion fluctuations as well asthe anisotropy of ion acoustic properties. TheNIF, with the much larger plasmas it will beable to generate, will make these kinds ofdiagnostics possible.


Nuclear Reactions in Large-Volume,High-Temperature Plasmas

A great deal of study has been done tocharacterize the behavior of ICF capsules. Asdiscussed earlier in Section IV, Astrophysics andSpace Physics, these capsules are expected toachieve conditions suitable for testing certainnuclear reaction rates of astrophysical interest.However, compressional heating of ICF-likecapsules may not be the only way to explorenuclear reaction rates at temperaturescorresponding to stellar interiors. We note thatmost Main Sequence stars have core tempera-tures between 1 and 2 keV and that Novaexperiments are already raising millimeter-sizegas samples to temperatures near 2 keV.

While the temperatures and densities of gas-bag experiments are lower than what might beachieved in an imploded capsule (such as an ICFcapsule), they are sufficient to examine manylight-element reaction rates, and there are definiteadvantages to direct heating of gas samples. Withdirect heating, the large sample size guaranteeshomogeneous conditions that can be held forrelatively long (nanosecond) periods. Thiseliminates much of the difficulty and uncertaintyin converting an observed number of reactionsinto an energy-dependent cross section. With 60times the energy, the NIF will be capable ofraising larger samples to higher temperatures.

An example of a reaction rate that might bestudied in this way is the 3He + 3He → 2p + 4Hereaction. This reaction rate is important for thebranching ratio between the branches of theproton-proton chain (PPI and PPII) in the sun,and therefore has direct consequences on theexpected neutrino rate. Table VII-1 gives thenumber of reactions expected in a nanosecondfrom a 0.6 cm3 gas sample at 3 atm.

Another important cross section forastrophysics involves the destruction of 7Li.Primordial lithium abundances are a criticalconstraint on big-bang nucleosynthesis andcosmology. Typically the primordial abundancesare obtained from the lithium abundances seen inthe oldest stars. However, the observedabundances must be corrected for any

destruction that occurs, and observations of oldclusters indicate that more lithium is destroyedthan expected. The main reaction for thisdestruction, 7Li + p → 8Be → 2 4He, is more than5000 times larger than the 3He rate at 2.1 keV.Table VII-2 shows other reactions that have ratesexpected to be higher than the 3He + 3He rate at2.1 KeV, and which are thus candidates forinvestigations on the NIF.

Table VII-1. The number of reactions for 3He+ 3He in a 0.6-cm3 3-atm sample.

Temperature (keV) Reactions per ns

2.1 22.6 163.4 2804.3 5500

Table VII-2. Selected reactions having rateslarger than the 3He + 3He rate at 2.1 keV.

D + p → 3He + γD + D → n + 3HeD + D → p + TT + T → 2 n +4He

3He + D → p + 4He3He + T → D + 4He3He + T → n + p + 4He4He + T → 7Li + γ6Li + p → 7Be + γ6Li + p → 4He + 3He7Li + D → n + 2 4He7Li + T → 2 n + 2 4He9Be + p → D + 2 4He9Be + p → 4He + 6Li10B + p → 4He + 7Be11B + p → 3 4He


E. References

1. See the discussion in H. A. Baldis, E. M.Campbell, and W. L. Kruer, “Laser PlasmaInteractions,” Handbook of Plasma Physics,M. N. Rosenbluth and R. Z. Sagdeev, Eds.,Volume 3: Physics of Laser Plasmas, A. M.Rubenchik and S. Witkowski, Eds. (ElsevierScience Publishers, B. V., 1991).

2. E. Ott and W. M. Mannheimer, Nuclear Fusion17, 1057 (1977).

3. D. J. Southwood, Planet. Space Sci. 16, 587(1968).

4. L. L. Smarr, M. L. Norman, and K. A.Winkler, Physica 12D 83 (1984).

5. T. A. Peyser et al., Phys. Fluids B 42, 448(1992).

6. Handbook of Plasma Physics, M. N. Rosenbluthand R. Z. Sagdeev, Eds., Volume 3: Physics ofLaser Plasmas, A. M. Rubenchik and S.Witkowski, Eds. (Elsevier Science Publishers,B. V., 1991), Chapter 9 and references therein.

7. F. Ze, L. J. Suter, S. M. Lane, E. M. Campbell,W. C. Mead, J. D. Lindl, M. D. Rosen, D. W.Phillion, C. W. Hatcher, R. P. Drake, J. S.Hildum, and K. R. Manes, Comments PlasmaPhys. and Controlled Thermonuclear Fusion 10,33 (1986).

8. D. Montgomery and E. M. Campbell,Comments Plasma Phys. and ControlledThermonuclear Fusion (1990).

9. E. M. Campbell in: Radiation in Plasmas,B. McNamara, Ed., Volume I (1984), pp. 579–621.

10. K. R. Manes, V. C. Rupert, J. M. Auerbach,P. Lee, and J. E. Swain, Phys. Rev. Lett. 39, 281(1977).

11. A. G. M. Maaswinkel, K. Eidmann, andR. Sigel, Phys. Rev. Lett. 42, 1625 (1979).

12. R. L. Carman, C. K. Rhodes, and R. F.Benjamin, Phys. Rev. A 24, 2649 (1981).

13. H. A. Baldis and C. J. Walsh, Phys. Fluids 26,3426 (1981).

14. H. A. Baldis, J. C. Samson, and P. B. Corkum,Phys. Rev. Lett. 41, 1719 (1978).

15. M. D. Rosen, D. W. Phillion, V. C. Rupert,W. C. Mead, W. L. Kruer, J. J. Thomson, H. N.Kornblum, and M. N. Rosenbluth, Phys. Rev.Lett. 29, 565 (1979).

16. J. E. Bernard and J. Mayer, Phys. Fluids 29,2313 (1986).

17. C. J. Walsh and H. A. Baldis, Phys. Rev. Lett.48, 1483 (1982).

18. J. D. Moody, H. A. Baldis, D. S. Montgomery,K. Estabrook, S. Dixit, and C. Labaune,Journal of Fusion Energy 12, 323 (1993).

19. C. J. Walsh, D. M. Villeneuve, and H. A.Baldis, Phys. Rev. Lett. 53, 1445 (1984).

20. H. A. Baldis, R. P. Drake, W. L. Kruer, K. G.Estabrook, E. A. Williams, andT. W. Johnston, Phys. Rev. Lett. 62, 2829(1989).

21. C. Joshi, C. E. Clayton, A. Yasuda, and F. F.Chen, J. Appl. Phys. 53, 215 (1982).

22. J. A. Stamper, R. H. Lehmberg, A. Schmitt,M. J. Herbst, E. C. Young, J. H. Gardner, andS. P. Obenschain, Phys. Fluids 28, 2563 (1985).

23. P. E. Young, H. A. Baldis, R. P. Drake, E. M.Campbell, and K. G. Estabrook, Phys. Rev.Lett. 61, 2336 (1988).

24. For stimulated Raman scattering see: W. Sekaet al., Phys. Fluids 27, 2181 (1984); R. P. Drake,Phys. Fluids B 1, 1082 (1989); R. P. Drake andR. E. Turner, in Laser Interaction & RelatedPlasma Phenomena, Volume 9, H. Hora and G.H. Miley, Eds. (1991); K. Estabrook, W. L.Kruer, and B. F. Lasinski, Phys. Rev. Lett. 45,1399 (1980); S. P. Batha et al., Phys. Rev. Lett.66, 2324 (1991); R. E. Turner et al., Phys. Rev.Lett. 57, 1725 (1986).

25. For stimulated Brillouin scattering see P. E.Young, Phys. Fluids B 3, 1245 (1991) and thereferences therein.

26. P. E. Young, Phys. Rev. Lett. 61, 2336 (1988);C. Labaune et al., Phys. Fluids B 4, 2224 (1992).

27. M. Tabak, J. Hammer, M. E. Glinsky, W. L.Kruer, S. C. Wilks, J. Woodworth, E. M.Campbell, and M. D. Perry, Phys. Plasmas 1,1626 (1994).

28. J. M. Dawson, Sci. Am. 260, No. 3, 54 (1989);“Advanced Accelerator Concepts,” J. S.Wurtele, Ed., AIP Conf. Proc. No. 279 (AIP,NY, 1993).

29. M. E. Glinsky, R. J. Mason, and M. Tabak,Bull. Am. Phys. Soc. 38, 2080 (1993).

30. S. C. Wilks, W. L. Kruer, M. Tabak, and A. B.Langdon, Phys. Rev. Lett. 69, 1383 (1992);X. Liu and D. Umstadter, Phys. Rev. Lett. 69,


1935 (1992); M. P. Kalashnikov, P. V. Nichels,Th. Schlegel, M. Schnuerer, F. Billhardt,I. Will, W. Sandner, and N. N. Demchenko,Phys. Rev. Lett. 73, 260 (1994).

31. P. Sprangle et al., Appl. Phys. Lett. 53, 2146(1988).

32. T. Antonsen and P. Mora, Phys. Rev. Lett. 69,2204 (1992); S. C. Wilks et al., Phys. Rev. Lett.69, 1383 (1992); W. B. Mori et al., Phys. Rev.Lett. 72, 1482 (1994).

33. M. E. Glinsky, R. J. Mason, and M. Tabak,Bull. Am. Phys. Soc. 38, 2080 (1993).

34. C. E. Clayton, K. A. Marsh, A. Dyson,M. Everett, A. Lal, W. P. Leemans,

R. Williams, and C. Joshi, Phys. Rev. Lett. 70,37 (1993) and references therein.

35. M. G. Haines, Can. J. Phys. 64, 912 (1986);this contains a review of the different waysof generating B-fields in a plasma.

36. W. L. Kruer, “Laser Induced CirculatingCurrents,” to be published (1994), in Bull.Am. Phys. Soc., DPP meeting 1994,Minneapolis, Minnesota.


38. For two-plasmon decay see H. A. Baldis andC. J. Walsh, Phys. Fluids 26, 1364 (1983);A. Simon et al., Phys. Fluids 26, 3107 (1983).

39. R. Bosch et al., Phys. Fluids B 4, 979 (1992).



Radiation Sources 107 Section VIII

Section VIII

Radiation Sources

A. Spectrally Continuous Sources

There are two broad categories of quasi-continuous radiation sources. First there is theradiation source generated by a hohlraum andused to heat a sample either inside the hohlraum(as in the case of an implosion microsphere or anopacity sample), or abutted to it (as with theradiation-flow experiments or the hydrodynamicmixing interface experiments). With this wegroup the gold burn-through foils used to heatthe non-local thermodynamic equilibrium(NLTE) tamped samples. Second, we have theeffort to develop backlights that are quasi-continuous.

In the first category the work done on thegold will be used to illustrate the cogent features.These burn-through foils have been studied indetail and provide a good sense of what to expectfrom very similar processes that occur in ahohlraum, noting, of course, that in a hohlraumone is using the front surface emission and anenclosure that will increase the energy in theradiation field.

Conversion Efficiency ofGold Burn-Through Foils

The efficiency with which gold burn-throughfoils convert visible laser light to x-rays can beroughly quoted as being in the region of ≥40% ofthe laser energy on target, noting that manypapers use a conversion efficiency of x-rayenergy to absorbed laser energy. We choose theformer method because it gives a better idea ofthe x-rays available from a particular laser.

Figure VIII-1 shows the conversion efficiencyfrom the front and rear sides for gold foils ofvarying thicknesses, for the two laser intensities

of 1012 and 1013 W/cm2.1 Conversion efficiency isobserved from both the front and rear surfaces.Efficiency observing from the front is shown inFig. VIII-1 as data points rising with increasedthickness, and from the rear is shown as datapoints falling with increasing foil thickness. Thelimiting case of a solid target is approximated bya 10,000-Å foil, where the conversion efficiencytransmitted in x-rays is very small, while thefront-side conversion rises to ~50% or more. Notethat these data have been obtained both withfiltered x-ray diode arrays and with spectro-meters, and both data sets are shown.

Figure VIII-1 illustrates two importantpoints. First, the conversion efficiency is notstrongly dependent on intensity above 1012

W/cm2, and second, the thinnest foils convertless. The reason for this is obvious as thetransmission of laser energy through the foilincreases dramatically as the foil becomes thin.The transmission is presented in Fig. VIII-2,where the ratio of the transmitted laser energy toincident laser energy is shown. Note that in thefigure the higher intensity shows more burn-through, but the fraction is less than or equal to0.01 for 1500-Å foils. This burn-through is due toinhomogeneities—hot spots or intensityfluctuations—in the beam profile. These hot spotscan be smoothed out and, indeed, have been inmore recent experiments.

Spectral Character ofGold Burn-Through Foils

The spectral character of the emission foundfrom a gold target similar to ones used inFig. VIII-1 is shown in Fig. VIII-3.2 Here the time-dependent spectrum is shown vs energy. Thespectrum was taken from the rear side of a gold

Section VIII 108 Radiation Sources

0.6

0.4

0.2

00 200 1000100

Front

Free-standing

1012 W/cm2

Rear

0.6

0.4

0.2

00 200 1000100

Supported

10 12 W/cm2

0.6

0.4

0.2

00 200 1000100

Free-standing

1013 W/cm2

0.6

0.4

0.2

00 200 1000100

Supported

1013 W/cm2

Gold layer thickness (nm)

X-r

ay o

utpu

t nor

mal

ized

to la

ser

ener

gy in

put

Figure VIII-1. Total x-ray emission, measured as conversion efficiency vs target thickness, fromthe front and rear surfaces of free-standing and supported gold targets. The gold foils vary inthickness from 500 Å to 10,000 Å, and irradiation is 1012 and 1013 W/cm2. The front surface always hasa larger conversion efficiency than the rear. Note that the method of support and the intensity do notaffect the results.

0 100 15050 200

1013 W/cm2

0 100 15050

1012 W/cm2

Au layer thickness

100

10–1

10–2

10–3

10–4

100

10–1

10–2

10–3

10–4

Tra

nsm

itted

to in

cide

nt la

ser

ener

gy

Figure VIII-2. Measurement of transmitted laser energy as a fraction of incident laser energy vs targetthickness, for two different intensities. This is from the same experimental setup used to obtain thedata in Fig. VIII-1.


2.01.51.00.50.0Energy (keV)

100 ps 200 ps 500 ps 600 ps

2500

2000

1500

1000

500

0

Flu

ence

J/(

KeV

-ns)

into

4π

2500

2000

1500

1000

500

0

2.01.51.00.50.0Energy (keV)

700 ps 900 ps 1000 ps 1100 ps

Flu

ence

J/(

KeV

-ns)

into

4π

Figure VIII-3. Time history of spectrum vsenergy in eV from the rear side of a 1000-Å goldfoil supported on a 5000-Å plastic layerirradiated at 2 × 1014 W/cm2. (The front and backspectra of a target of this thickness will be verysimilar, with no major differences in thespectral character.) The results were measuredby a 12-channel filtered x-ray diode that coversthe range from 80 through 2000 eV.

foil that was 1000 Å thick, mounted on a 5000-Ålayer of plastic. (The front and back spectra of atarget of this thickness will be very similar, withno major differences in the spectral character.)The location of the two “bumps” in the spectrumcorrespond to the oxygen and nitrogen bands, at~300 eV and ~800 eV respectively.

There is also an M-band contribution, whichdoes not show up on soft x-ray instrumentsbecause it spans the energies between 2000 eVand 3600 eV. Figure VIII-4 shows the M-bandcontribution for a gold disk irradiated at2 × 1014 W/cm2.3 This band-like structure isobtained for all high-Z elements. The line drawnon the figure indicates the spectral region overwhich the M-band contribution is defined.

Temporal Behavior ofGold Burn-Through Foils

The time history of the emission in the harderx-ray regimes, such as the M-band, will followthe laser pulse, and there is much data to supportthis supposition. However, the soft x-rayspectrum will last for time longer than the laserpulse, and duration will increase as thewavelength of the radiation increases. Thus, forthe softest x-rays measured using the filteredx-ray diodes in Fig. VIII-3 (i.e., ~80 eV), the timeduration will be two to three times the pulseduration. Because the pulse is 1 ns in duration,this indicates emission on the order of several ns.

In Fig. VIII-5, the total energy emitted perunit time from a source on the rear of gold foils ofvarious thicknesses is shown. The foils wereirradiated at 2 × 1014 W/cm2. Normalization ismade to a sphere (i.e., “into 4π”), but themeasurements reflect only the intensity of aspecific ray. Thus, the correct units should bejoules per ns per steradian, and the fact that thesesources tend to look Lambertian (i.e., they havecos(ø) dependence) provides the correct intensity.Here ø is the angle of observation relative to thetarget surface normal.

Spectral Character of Other Foils

Next we illustrate the fact that the spectralcharacter of the soft x-ray emission can bemodified by the choice of target material.4

Figure VIII-6 A) shows the soft x-ray spectrumfrom pure targets of several elements from Z = 56to Z = 63, while B) shows the soft x-ray spectrumfrom targets consisting of a compound thatcontains approximately 1% of the same element.

The spectrum in each pair of cases shows thatas the opacity of the sample is increased by using


2.8

2.4

2.0

1.6

1.2

0.8

0.4

02.0 2.4 2.8 3.2 3.61.6

Gold M-band spectrum

Photon energy (keV)

Flu

ence

(ke

V/K

eV-s

pher

e) 1

016

Figure VIII-4. Spectrum of the gold M-bands, which cover the spectral region from 2000 to 3600 eV,from a thick target with intensities of 2 × 1014 W/cm2. This band-like structure is obtained for all high-Z elements. The line indicates the spectral region over which the M-band contribution is defined. Theunits used here are unique to the inertial confinement fusion (ICF) program at LLNL and can be easilyconverted.

1500

1000

500

0

1.41.21.00.80.60.40.20.0Time (ns)

5000 Å

2000 Å

1500 Å

700 Å1000 Å

J/ns

into

4π

Figure VIII-5. Temporal history of spectrally integrated emission from the rear side of gold foils, forvarious foil thicknesses irradiated at 2 × 1014 W/cm2. Note that the duration of the laser, 1 ns in thiscase, is not closely followed by the softer x-ray emission. The time delay for the thicker foils occursdue to the lag in getting the softer x-ray to the back surface. Also note that the hard x-ray flux (e.g., thegold M-band contribution) shows no timing differences with different foil thickness.


Ba (56)

La (57)

Ce (58)

Pr (59)

Nd (60)

Eu (63)

Inte

nsity

(ar

b. u

nits

)

70 90 110 130Wavelength (Å)

A) Pure element

70 90 110 130

Wavelength (Å)

Ba (56)

La (57)

Ce (58)

Pr (59)

Nd (60)

Eu (63)

B) Element with 1% impurity

Figure VIII-6. Spectral character of various elements from Z = 56 to Z = 63. In cases A) the targets aremade of the pure element, while in cases B) the element is a 1% impurity in the target material. Theincreased optical depth in the cases with the pure element leads to an enhancement of the soft x-rayfeatures compared to the cases with the impurity, where the optical depth is low and the evidence ofindividual lines becomes apparent. Note also that the dominant features move toward higher energyas Z increases.

a pure Z target, the softer x-ray feature andthereby the quasi-continuous nature of theemission are increased. Moreover, the ability totailor the source in energy by changing Z canbe observed.

Figure VIII-7 shows the same type of spectraldata as does Fig. VIII-6, but for a much widerrange of Zs.5 The figure shows the front-surfacespectra measured in 1012 ergs per cm persteradian for an irradiance of 3 × 1013 W/cm2.

The experiments used the frequency-doubledlaser at a wavelength of 0.53 µm and pulse lengthwas 3 ns. Note that the gold, oxygen, andnitrogen bands are seen while the M band isbeyond the spectral coverage.

Figure VIII-8 shows the absolute conversionefficiency of x-ray production in absolutelymeasured spectra emitted by laser-irradiatedtargets as a function of target Z. Note that the netincrease in x-ray conversion is quite marked forthe high-Z elements.


1.0

0.5

02

1

02

1

020

10

020

10

0

12

6

012

6

08

4

08

4

08

4

0

Be Z = 4

C Z = 6

Al Z = 13

Ti Z = 22

Cu Z = 29

Mo Z = 42

Sn Z = 50

W Z = 74

Au Z = 79

Pb Z = 82

K

K

L

L

LN O

N O

N O

M

M

0 100 200Wavelength (Å)

0 100 200Wavelength (Å)

(1012

erg

/cm

× s

r) →

d2E

dλ

dΩ

Figure VIII-7. Absolutely measured spectra emitted by laser-irradiated targets of different elements.Intensity is 3 × 1013 W/cm2, pulse duration is 3 ns, and wavelength of the laser is 0.53 µm. Thespectroscopic structure K, L, M, N, and O refer to the shells from which the emission arises. Note thatthe intensity axis changes in the various plots.

Angle of Incidence vsConversion Efficiency

The next pertinent issue concerns the effectthat the angle of incidence of the laser withrespect to the target surface normal has onconversion efficiency. There are two obviousgeometries that can be used to irradiate a samplewith the x-rays from the laser-conversion target.

The first is to place the sample at the rear of ahigh-Z foil that is thin enough to let x-raysthrough, but thick enough to stop laser light from

being transmitted (see Figs. VIII-1 and VIII-2 forinformation on effects of target thickness). Thesecond is an angled arrangement used to enhanceradiation, based on the fact that the front-surfacex-ray conversions are higher. For this secondgeometry, the effect of the angle of incidenceis important.

Studies have been made of the effect of angleof incidence on conversion efficiency. InFig. VIII-9 we show two quantities as a functionof angle of incidence of the laser.6 First, the left-hand axis refers to absorption of the laser light by


4 6 10 20 40 60 1000.01

0.02

0.04

0.06

0.1

0.2

0.4

0.6

1

2

4

6

10

20

40

Be C Al Ti Cu MoSn W Au Pb

dE dΩ (

J/sr

)

Atomic number (Z)

Con

vers

ion

effic

ienc

y (%

)

Figure VIII-8. Absolute conversion efficiency of laser-irradiated targets of different elements fromberyllium through lead. Net increase in x-ray conversion is quite marked for the high-Z elements.

X-r

ay c

onve

rsio

n ra

te E

x/E a

bs (%

)

100

50

00 10 20 30 40 50 60 70 80

η abs

η x

Abs

orpt

ion

(%)

100

50

0

Laser incidence angle

Figure VIII-9. Effect of angle of incidence of laser on x-ray conversion efficiency for intensity of 1014

W/cm2 on a solid gold target. The left-hand axis refers to the open circles, which measure absorption ofthe laser light. The right-hand axis refers to the filled circles and represents the x-ray conversionefficiency relative to the absorbed laser energy.


the target, represented by the open circles. Theright-hand axis refers to x-ray conversionefficiency measured relative to the absorbedenergy, represented by the filled circles.

To convert this efficiency relative to absorbedenergy to conversion efficiency relative to theincident energy, multiply the Ex/Eabs by theabsorption percent. This gives, for example, forthe 20° result, a conversion efficiency of 46%,which is in keeping with the 40% figure notedabove. Efficiency does not drop off until the angleof incidence becomes greater than 45°. Thisalleviates experimental design constraints. Theactual change in spectral shape as the angle isincreased can also be shown not to deviate forangles less than 45°. These facts indicate theuniformity of the spectrum at angles less than 45°and give an indication that angles up to as highas 50° preserve a similar x-ray spectral character.

Laser Intensity vsConversion Efficiency

The next important question concernsconversion efficiency from the laser light tox-rays as a function of intensity on target. InFig. VIII-1 it can be seen that the conversionefficiency is roughly constant from 1012 to1013 W/cm2. However, the question arises: Whatis the behavior at higher or lower irradiance?There is data from the rear side of the target toindicate that the conversion stays roughlyconstant, at least to 1014 W/cm2. Beyond thisirradiance, the onset of laser-driven parametricprocesses will start to play a role and the benignnature of the x-ray conversion process is notguaranteed. Note that with the advent ofsmoothed beams this statement should be seen asworthy of investigation.

On the lower irradiance side, the broaddefinition of the soft x-ray region—from 50 eVthrough 1250 eV—allows a wide range oftemperatures over which the plasma will emitx-rays in this spectral band. Clearly the sourcewill peak at lower energies and the totalenergy in the x-ray band will drop as laserintensity drops.

However, although we have no data onthe conversion efficiency below 1012 W/cm2,a reasonable estimate is to assume that:

• Absorption stays high at irradiancesabove 109 W/cm2, the threshold forplasma formation.

• The absorbed laser light will radiate as ablackbody radiator.

Therefore, we equate the laser intensity to thepower radiated; i.e., the Stefan-Boltzmannrelationship: Tev

~(10–5 I laser )0.25 . Here Tev is theplasma temperature in eV and Ilaser is the laserintensity in W/cm2. Thus, for an intensity of 1012

W/cm2, the temperature of the plasma will be~18 eV, indicating that the peak of the emission—were it a blackbody emitter—would be at 50 eV.The total energy emitted above 50 eV would behalf the laser energy. Since this is, in the spirit ofthe approximation, 40%, this rough estimate ofconversion efficiency can be extended down tointensities of ~1010 W/cm2, keeping in mind thatthere is not yet any data at these intensities.

Hohlraum RadiationTemperature Scaling

To illustrate various radiation environmentsand their scaling with hohlraum size, data takenat Nova are presented in Fig. VIII-10. Equivalentradiation temperatures vs time are shown forthree different variations of the gold hohlraum,as well as two different laser-pulse durations. Allexperiments shown used a total energy of 18 kJ.The hohlraums were all cylindrical, with laserentrance holes in the sides for the 500-µm- and1000-µm-length hohlraums, and through the endsin the 2800-µm-length hohlraum.

As can be seen from the empirical scaling law(shown in Fig. III-3), both the area of the interiorwalls and the area of holes affect the radiationenvironment. The holes for the two longer(therefore larger) hohlraums are 500 µm indiameter for the laser entrance holes and 800 µmfor the x-ray diode observation line of sight. Forthe smaller hohlraum, each laser entrance hole is300 µm and the line of sight is 400 µm.

The observations in Fig. VIII-10 were madewith a 12-channel x-ray-filtered diode array thatis time-resolved and absolutely calibrated. Tokeep the laser light from emerging from the lineof sight and to assist in maintaining itsunobstructed area, a 1500-Å aluminum foilsupported on each side by 5 µm of plastic was


3.02.52.01.51.00.50.0

T (

keV

)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Time (ns)

1-ns pulse—500 µm × 500 µm

2-ns pulse—1000 µm × 1000 µm

1-ns pulse—1600 µm × 2800 µm

Without M-band

With M-band

Figure VIII-10. Equivalent radiation temperature vs time for several gold hohlraums (schematics forthe hohlraums are shown in Fig. III-2). The different curves for each hohlraum indicate thereproducibility of the data. The contribution of the gold M-band to radiation temperature is indicatedby including data both with and without the M-band contribution.

inserted in the line of sight. The x-ray diodemeasures the specific intensity at the detector.The specific intensity is converted into anequivalent radiation temperature by integratingthe source emissivity over all channels andequating this to the blackbody emissivity, givenby σTR

4. This is done for each of the times ofinterest. The contribution of the gold M-band, at~2–3 keV, is most pronounced in the smallhohlraums.

To indicate the contributions of this highlynon-Planckian part of the spectrum we also showin the figure the equivalent radiation temperaturewithout the M-band contribution, indicatedby the gray line. Note that the smaller thehohlraum, the higher the peak TR but theshorter the duration of the x-ray flux. This isdue to closure of the holes. Also, the relativeM-band contribution to the TR is larger in thesmaller hohlraums.

The 1000-µm cylindrical hohlraum was usedto investigate a longer duration pulse length. Inthis case, a 2-ns pulse was used, and the highpeak x-ray flux is maintained for approximately1 ns, as can be seen in the figure.

The two different sets of curves for the1000-µm case (i. e., with and without M-bands)represent the effect of the angle of observationwith respect to the laser-irradiated inner surfaceof the hohlraum. The two angles are 72° and 0°with respect to the cylinder axis. The curveswith lower TR are for the viewing angle of 72°,and these show that when the detector observesless non-LTE gold emission in the hohlraum, ithas a smaller M-band contribution as well aslower total flux. Thus, the radiation field in thehohlraum is a complicated function ofviewing angle.

Finally, the figure shows three distinctexperiments for the largest hohlraum, indicating


the level of reproducibility of the x-rayproduction. These largest hohlraums are thestandard configuration used in most inertialconfinement fusion (ICF)-type experiments.In this case the M-band contribution is smallbecause of the combined effects of the relativelylarge size of the hohlraum and the viewing angle,which does not allow a view of the laser-irradiated spots in the hohlraum.

B. Spectrally Narrow Sources

The conversion of the laser radiation todiscrete lines or narrow bands of lines representsa distinct field of study. The usual motivation forthe study of conversion into narrow spectral lineshas been the development of backlight (i.e.,absorption) sources for use as probes of hot,dense matter. Because this has been themotivation, there is also an interest in the spatialextent of the source. In a manner similar to thatused to estimate the conversion efficiency of thequasi-continuous source, it can be broadlyassumed that it is possible to attain conversionefficiencies into a single spectral line of 0.1% to1%, with the lower efficiency being for higherx-ray energies than 4 keV, and the higherefficiencies being relevant for transition energiesbelow 4 keV.

Effect of Z on Emission

The effect of Z on the emission is that itchanges the character of the emission from singlespectral lines to bands. As shown in Fig. VIII-4,the gold M-band is an example of the type ofspectral structure that can expected. Conversionefficiency into these band-like structures isapproximately 5% for 0.53-µm light.

Figure VIII-11 shows a compilation of theresults of experiments on several elements.7 Theexperiments irradiated a planar target of thematerial of interest with intensities of 1014 to 1015

W/cm2. The dashed line indicates the resultswith 1.06 µm light, while the solid line connectsthose points with 0.53 µm light.

Although there are some differences in theperformance of the targets under intensityvariation, it is not significant in this broad outline

of the results. Included in the figure are theresults for aluminum silicon, titanium, nickel,and zinc K-shell emission, and since this spans Zsof 13, 14, 22, 28, and 30 respectively, we have afairly good indicator of levels of conversion.

Figure VIII-11 also shows conversionefficiency into the gold (Z = 69) M-bands. Tochange the conversion unit of this figure intofractional conversion in output joules, multiplythe left-hand value ξ x by the x-ray line energygiven on the bottom axis by 1.6 × 10–16 J/keV. Forexample, the gold M bands are at an energy of~2.5 keV and the ξx is ~1.3 × 1014, yielding~5% conversion.

1013

1012

1011

1014

0 2 4 6 8 10X-ray line energy (keV)

Au M

Al k

Al k

Au M

Si k

Ti k

Ni k

Zn k

λL - 1.06 µ

λL - 0.53 µm

X-r

ay p

hoto

n / i

ncid

ent l

aser

joul

e

Figure VIII-11. Conversion efficiency of variousdiscrete transitions, showing the K-shell linesof aluminum, silicon, titanium, nickel, and zincas well as the M-band conversion of gold.Conversion efficiency is shown as a function oftransition energy for intensities on the order of~1014 W/cm2 on solid targets. The dashed linestie together data for a 1.06-µm laser, and thesolid line ties together the data for the 0.53-µmlaser.


Comments on Sparseness of Data

In this discussion of spectrally narrowsources, we have given some simple rules fordetermining the conversion efficiency of laserlight into x-rays. There are however, many gapsin the data, and particularly interesting is the lackof data from lower irradiances. For lowerirradiances (that is, down to 1010 W/cm2) the sizeof the source region can become large and theirradiation characteristics will be of interest. Thelevels of uniformity, angular dependence,temporal duration of the converted light, andspectral character are of interest.

Although this data does not exist, it is feltthat the simple rules above can be used. Furtherwork can, of course, be undertaken but thisremains a very large field and it might be wise tohave a first attempt at a design for the specificexperiments of importance before seeking toobtain further quantitative data.

C. X-ray Laser Sources

There are three distinct paths of research forx-ray laser studies. First, there is research into thedetails of the mechanisms that have made thefirst viable laboratory x-ray laser possible.Second, there is an effort to use the currentlydeveloped lasers as a source to study otherphenomena. Third, there is the continuing workto develop the x-ray laser as a light source forfuture applications. Although these threeavenues are not completely distinct it is simplerto present the following discussions in this way.

Studying X-ray Lasers

Since the first demonstration of soft x-raylasing using the collisional excitation mechanismin neon-like selenium, many other neon-like ions,ranging from copper (Z = 29) to silver (Z = 47),have been made to lase.8 However, attempts toproduce lower-Z neon-like x-ray lasers had beenunsuccessful. In the effort to develop a tabletopx-ray laser that would require smaller high-energy laser drivers than Nova and could beused for applications such as biological imaging,nonlinear optics, holography, etc., a prepulsetechnique has been developed. This technique

has been used successfully to produce lasing inmany lower-Z neon-like ions such as titanium(Z = 22), chromium (Z = 24), iron (Z = 26), andnickel (Z = 28).9 The use of this prepulsetechnique has opened up a new class of neon-likex-ray lasers for investigation.

In studying the neon-like lasers it wasnoticed that the ions with odd Z do not lase aswell as those with even Z. Since elements withodd Z have a nuclear spin and a nuclear momentwhile those with even Z tend to have no nuclearspin, one possible explanation for this anomalousbehavior is that hyperfine splitting is playing animportant role in the gain of the neon-like laserlines for elements with odd Z.

To test this, we measured the line shape ofthe J = 0 to 1, 3p–3s transition in neon-likeniobium and zirconium using a high-resolutiongrazing incidence grating spectrometer. Weobserved a 28-mÅ splitting between the twolargest hyperfine components in the niobium(Z = 41) line at 145.9 Å, in good agreement withtheory. As predicted, no splitting was observedin zirconium (Z = 40).10

This 131-cm–1 splitting is the largesthyperfine splitting ever observed and is also theshortest wavelength transition (145.9 Å) and mosthighly ionized plasma (31 electrons) in which thehyperfine effect has been directly observed, withthe exception of some very recent acceleratorexperiments in hydrogen-like bismuth done atGSI Darmstadt.

Further studies have been performed on theuse of low-density foams as possible x-ray lasertargets. Currently, x-ray laser experiments useeither thin foils or thick slab targets at soliddensity. These targets are illuminated by Nova,causing them to heat and expand to reach the lowdensities required for lasing. This hydrodynamicexpansion introduces large electron-densitygradients in the plasma, which in turn result inlarge refraction effects as the x-rays propagatedown the laser axis.

If targets could be fabricated from uniformlow-density materials at the density required forlasing (i.e., <3 mg/cm2), then beam propagationand ultimately laser coherence could beimproved. Successful observations have beenmade of lasing in zirconium aerogel foam at


90 mg/cm3. At lower density, molybdenum-doped agar foam at 3 mg/cm3 and selenium-doped agar at 8 mg/cm3 have been tried withoutsuccess. However, the aerogel foams have muchsmaller cell sizes (500 Å compared with 1–2 µmfor agar) and look very promising if the densitycan be lowered.

Very recent experiments are using atraveling-wave geometry to produce short-pulsex-ray lasers for use in imaging experiments onICF targets. The traveling wave geometry enablesus to drive the x-ray laser targets with shortpulses (~100 ps) while still using a long enoughtarget (~3 cm or greater) to produce strong laseroutput. We have recently observed a hundred-fold contrast in neon-like germanium x-ray lasersobserved with and against the direction of thetraveling wave.

Two-dimensional Plasma Imaging

For two-dimensional plasma imaging, spatialresolution is currently limited by the pulse lengthof the x-ray laser, or alternatively by the gatetime of the detector. Gated detectors with largeactive areas have temporal resolutions that are, atbest, in the range of 100 to 200 ps. Forcharacteristic expansion velocities of 107

cm/s,this corresponds to a spatial resolution of10−20 µm, which is well below the resolutionpossible with available imaging optics. Therefore,there exists a real need to produce x-ray laserswith pulse durations shorter than the 200-psFWHM routinely produced with explodingfoil targets.

Demonstrated x-ray lasers are eitherrecombination x-ray lasers, which rely ontransient inversion to achieve gain, orcollisionally pumped systems in whichinversions are produced in steady-stateconditions. The gain duration in conventional(or quasi-static) recombination x-ray lasers isdetermined by cooling and recombination rates,and typically results in pulse widths >200 ps.This allows little control over the output pulselength without significantly reducing thetotal gain.

Collisionally pumped x-ray lasers haveachieved large gain-length products and havebeen shown to operate over a wide range of

pump conditions and a variety of targets. Inaddition, the wide-wavelength range over whichcollisionally pumped systems operate will makeit possible to take advantage of, or alternativelyavoid, line absorption in plasma-probingexperiments.

To produce saturated x-ray lasers with pulsedurations shorter than 20 ps and imagingresolutions of ~2 µm, experiments have beenperformed that irradiate an exploding foil withmultiple pulses. The concept is illustrated inFig. VIII-12. The first pulse heats and explodes athin foil target to produce a plasma with lowdensity gradients. The second pulse then ionizesthe preformed plasma to produce conditionssuitable for high gain. The energy of the firstpulse can be significantly lower than that of thesecond pulse, because high electron temperaturesare not necessary. This improves the efficiency ofthis approach by reducing radiation and thermallosses during hydrodynamic expansion.

To test this technique and illustrate theimportance of density gradients in x-ray laserperformance, experiments have been performedwith yttrium x-ray laser targets. A key modifica-tion made to the laser beams was to introduce atilt into the phase front of the beam. This allowedus to pump the foil in a traveling-wave con-figuration as shown in Figs. VIII-13 and IX-14.The traveling-wave configuration is important

LaserLaser

Thin foil

100 ps

300 ps

Figure VIII-12. Schematic of double-pulseirradiation technique for producing short-pulsex-ray lasers.


Phasefront

C

V

θ

Phasefront

Figure VIII-13. Schematic of traveling-wave x-ray laser, illustrating how optical beam is tiltedto generate traveling-wave pump alongexploding foil.

given the short gain duration and finitepropagation time of the x-ray laser across the foil(i.e., ~100 ps for 3-cm foil). The traveling-wavepump along the exploding foil was generated bytilting an optical beam. The grating techniqueused to tilt the phase front of the traveling-wavebeam is illustrated in Fig. VIII-14.

In the experiment, a 3-cm-long foil consistingof 2000 Å of yttrium on 1000 Å of plastic (lexan)was irradiated from both sides with two 100-pspulses of 0.53-µm laser light. The pulses wereseparated by 300 ps and the foil was pumpedusing a traveling wave. The total intensity on thetarget was 2.4 × 1014

W/cm2 for the first pulseand 1.6 × 1014

W/cm2 for the second pulse. Themeasured soft x-ray spectrum in Fig. VIII-15shows bright J = 2–1, 155-Å x-ray line emission in3rd and 4th order of the grating. The effects oftraveling wave pumping are clearly evident asthe output is a factor of ~200 higher when thewave is traveling towards rather than away fromthe spectrometer. The two cases, wave travelingtoward and wave traveling away from thespectrometer, are shown in Fig. VIII-14.

Figure VIII-16 shows the time-resolvedemission spectra, in both hard x-ray and XUVrange, of the traveling-wave experiment. Thecharacteristic neon-like and fluorine-like 3–2 and

θin

θout

Dout

Din

θ ph

θout

θinDout = Din

cos( )cos( )

tan( ) =θph θouta cos( )λ

sin( ) + sin( ) =θin θoutλa

Figure VIII-14. Schematic of grating techniqueused to tilt phase front of traveling-wave x-raylaser beam to necessary angle.

hard x-ray emission spectrum (A) shows 4–2transitions that indicate that appropriateionization conditions for gain are producedduring both pulses. In contrast, the XUVspectrum (B) shows strong J = 2–1 x-ray laseremission at 155 Å during the second pulse only.The most likely reason for this is that during thefirst pulse the high density gradients refract thex-ray laser emission out of the gain region.

Figure VIII-17 shows the temporal evolutionof both the x-ray laser and hard x-ray emission.The x-ray laser emission has a time duration of45 ps FWHM, making this the shortest x-ray laserpulse width demonstrated.

Although in this experiment comparableintensities were used for both pulses, it would beclearly more efficient to reduce the intensity ofthe first pulse and increase the pulse separationto allow for expansion. Alternatively, a betterpulse shape may consist of a low-intensitypedestal that irradiates the foil until the high-intensity short-pulse irradiation begins. Usingthis approach we should be able achievesaturation x-ray laser output intensities of4 × 1011

W/cm2 with less energy by a factor of ~4.Extending this technique to shorter output pulseswill require shorter optical laser irradiation or areduction in irradiation intensity to shorten theduration of neon-like ions.


8

12

16

460 520 580 640

8

12

16

λ (Å)

3rd order, J = 2–1 (155 Å) 4th order

Figure VIII-15. X-ray line emission in 3rd and 4th order with traveling wave. Top shows the casewhere the beam is going away from detector and bottom shows where beam is going toward detector,where the output is increased by a factor of two.

4.0 4.5 5.0 5.5 6.0 6.5

–100

0

100

200

300

400

500

λ (Å)

160 190 220 250 280 310

λ (Å)

A) Hard x-ray spectrum B) XUV spectrum

Tim

e (p

s)

Figure VIII-16. Time history of hard x-ray and XUV emission spectra in neon-like yttrium. The hardx-ray emission spectrum (A) shows characteristic neon-like and fluorine-like 3–2 and 4–2 transitionsthat indicate that appropriate ionization conditions for gain are produced during both pulses. Incontrast, the XUV spectrum (B) clearly shows strong J = 2–1 x-ray laser emission at 155 Å during thesecond pulse only.


0

20

40

60

80

100

120

0 100 200 300Time (ps)

400 500 600 700 800

4–2 transitionJ = 2–1 x-ray laser

45 ps110 ps

Rel

ativ

e in

tens

ity

Figure VIII-17. Time history of hard x-ray (4–2transitions) and J = 2–1 x-ray laser transition inneon-like yttrium. The x-ray laser emission hasa time duration of 45 ps FWHM.

Development of X-ray Lasers as Probes

Because the collisionally pumped soft x-raylasers now operate over a wavelength rangeextending from 35 to 300 Å, and because thesesources have high peak brightness, it is possibleto begin using use them for x-ray imaging andplasma interferometry. The attraction of x-raylaser light sources is their extraordinarybrightness, which was estimated to be equivalentto a Planckian of temperature of 40 keV.11

It is possible to demonstrate applicationsof the x-ray laser probe to long-scale-lengthplasmas using Moiré deflectometry and softx-ray imaging. (Progress in the development ofshort-pulse x-ray lasers using a double-pulseirradiation technique that incorporates atraveling wave pump is described above.)

The first of the x-ray laser probe applicationsis diagrammed in Fig. VIII-18. Here a series ofexperiments has begun to characterize driveuniformity of thin-foil targets irradiated withlaser beams that are smoothed with randomphase plates. Targets of 10-µm plastic foilsovercoated with 3 µm of aluminum are irradiatedat 1014

W/cm2 with a 1-ns-square, 0.53-µm-wavelength beam from the Nova laser. Theseexperiments image the accelerated foils by usingthe x-ray laser as a backlight. In this experiment,10-µm plastic foils overcoated with 3 µm ofaluminum are irradiated at 1014

W/cm2 with a

Multilayerimagingmirror

Multilayercollimatingmirror

X-ray lasertarget

Sampletarget

Laser beam

Detector

Laser beam

Figure VIII-18. Schematic of experimental setupfor plasma imaging using the x-ray laser as abacklight. This setup is used to characterizedrive uniformity of thin-foil targets. Targets of10-µm plastic foils overcoated with 3 µm ofaluminum are irradiated at 1014

W/cm2 with a1-ns-square, 0.53-µm-wavelength beam from theNova laser. The accelerated foil is backlit with a~150-ps x-ray laser pulse from an explodedyttrium x-ray laser target.

1-ns-square, 0.53-µm-wavelength beam fromthe Nova laser. The beam’s focal spot has aneffectively flat-top distribution over a 700-µmdiameter region. The accelerated foil is backlitwith a ~150-ps x-ray laser pulse from anexploded yttrium x-ray laser target. The x-raylaser operates at 155 Å and has an output energyof ~7 mJ.

In Fig. VIII-19 we show the x-ray laser backlitimage from an accelerated foil experiment asrecorded on an x-ray CCD detector. The foil isconstructed of a 10-µm plastic layer and a 3-µmaluminum layer, irradiated, with the beamsmoothed using random phase plates. The spatialresolution is estimated to be better than 2 µm.The 5- to 6-µm filament structure observed on therear surface is likely due to the speckle patternproduced by the random phase plates.

This result clearly shows the merits of usingx-ray lasers to image high-density plasmas.Although direct imaging using an x-ray laserdoes not require several of the attributes of thex-ray laser (specifically, small source size, narrow


0 100 200 300 400 500 600 700

0

50

100

150

200Y

(µm

)

300 350 400 450 500 550 600 650100

110

120

130

140

150

160

170

X(µm)

Laser intensity 1014 w/cm2

Figure VIII-19. X-ray-laser-backlit image of accelerated foil, recorded on an x-ray CCD detector. Foilconsists of a 10-µm plastic layer and a 3-µm aluminum layer, irradiated, with beam smoothed usingrandom phase plates. Upper image is enlarged view of central region of foil, showing non-uniformities (filament structure) at the rear surface with dimensions characteristic of the specklepattern produced by the random phase plates.

divergence, and narrow bandwidth), its highbrightness makes it possible to neglect self-emission. Although not required for directimaging, the short duration of the x-ray laser isuseful in that it allows us to use an x-ray CCD asthe detector. The x-ray CCD has better spatialresolution (27-µm pixel) and higher dynamicrange (>1000) than existing gated microchannelplate systems.

Using a similar experimental setup, Moirédeflectometry experiments have been performedto measure electron density profile in a laser-irradiated plastic plasma. The only differencebetween the imaging and Moiré experiments isthe addition of two Ronchi rulings directlyinfront of the detector on the Moiré experiments.The rulings have a period of 10 µm and a relativerotation of 4 degrees to give a fringe spacing of

142 µm in the image plane—i.e., 14.2 µm in theobject plane. The rulings were separated by19.35 mm to give a sensitivity of 5.2 Mrad/fringe-shift at the object plane.

Figure VIII-20 shows the measureddeflectogram obtained of a plastic targetirradiated at 2.5 × 1013

W/cm2. The laser pulsewas 1 ns square at a wavelength of 0.53 µm, andwas focused to a 3-mm-diameter spot. The imageshows the expected unperturbed Moiré patternfar away from the surface, but shows cleardeflections closer to the irradiated side of thetarget. Analysis of the fringe pattern indicatespeak electron densities of 2 × 1021

cm–3 near thesurface. This corresponds to a peak line-integrated density of 6 × 1020

cm–2 , a factor of 30higher than is accessible using conventional UVinterferometry at λ ~ 2650 Å.


Y(µ

m)

0 100 200 300 400 500 600 700 800 9000

100

200

300

400

500

600

700

800

900

X(µm)

Figure VIII-20. Moiré deflectogram of laser-irradiated plastic target. Laser beam is incident from theright. The image shows the expected unperturbed Moiré pattern far away from the surface, but showsclear deflections closer to the irradiated side of the target. Analysis of the fringe pattern indicatespeak electron densities of 2 × 1021

cm–3 near the surface. This corresponds to a peak line-integrateddensity of 6 × 1020

cm–2 , a factor of 30 higher than is accessible using conventional UV interferometryat λ ~ 2650 Å.

The second application employs the x-raylaser in an imaging microscope. Figure VIII-21shows a schematic of the x-ray imagingmicroscope. It uses x-rays from a tantalumnickel-like collisionally pumped x-ray laseroperating at 44.83 Å, which are collected andfocused onto a specimen. The x-ray laser isgenerated by irradiating a 3.5-cm long, 2000-Åthick plastic foil coated with 900 Å of tantalumusing two cylindrically focused laser beamsoperating at a wavelength of 0.53 µm.12 The twobeams heat the foil, which explodes toform a high-temperature plasma with lowdensity gradients.

For this experiment, two optical beams fromthe Nova laser were used to generate a combinedintensity on the target foil of 3.0 × 1014 W/cm2 fora duration of 500 ps. The x-ray laser pulse(~200 ps FWHM) is characteristic of explodingfoil amplifiers. The x-ray laser originates from a~100-µm diameter gain region at the center of theplasma and has a beam divergence of 10 MradFWHM. The output energy is ~10 µJ, which gives

a brightness of 1021 photons/(s-Mrad2-mm2-0.01% BW).13

Figure VIII-22 shows an image of a resolutiontest pattern consisting of radial gold bars thattaper down at the center to ~350 Å. The gold barsare 1000 Å thick and are on a 1000-Å siliconnitride substrate. Features near the diffraction-limited resolution of ~550 Å are clearly observed.The nonuniform illumination pattern is due tothe finite source size of the x-ray laser, which isdemagnified onto the test pattern by themultilayer mirror collecting optic. (The effectivesource size of the x-ray laser is estimated to be50 µm by 80 µm.) The grainy nature of the imageis due to the microchannel plate detector and thelow signal levels.

Using this x-ray microscope, images havebeen made of three rat sperm nuclei that haveundergone different preparation techniques. Thegoal was to investigate the advantages of goldlabeling and to access the effects of the variouspreparation techniques.


Ta foil target

Laser beams

Filter

X-ray laser beam

Multilayer mirror condenser

Test specimenon pinhole

Zone plate objective

Filter

MCP detector

Photographic film

Figure VIII-21. Schematic of x-ray microscope, showing main components. The light source is atantalum nickel-like collisionally pumped x-ray laser operating at 44.83 Å. The x-rays it produces arecollected and focused onto a specimen using a spherical multilayer mirror.

20

16

12

8

4

00 4 8 12 16 20

X (µm)

Y (

µm)

Figure VIII-22. Image of resolution test patternobtained with x-ray laser microscope. The testpattern consists of 1000-Å-thick radial gold barson a 1000-Å silicon nitride substrate. The barstaper down at the center to ~350 Å. Thenonuniform illumination pattern is due to thefinite source size of the x-ray laser, which isdemagnified onto the test pattern by themultilayer mirror collecting optic. The grainynature of the image is due to the microchannelplate detector and the low signal levels.

The nuclei were first prepared by treatingsperm isolated from rat epididymides to exposethe DNA-protamine complex that comprisessperm chromatin. This was done by treating thesperm with a disulfide reducing agent anddetergent to dissolve the tails, acrosome, andnuclear membranes. A droplet containing theamembraneous nuclei was deposited onto asilicon nitride window with dimensions300 × 300 µm by 1000 Å thick, and the liquidcontaining unbound nuclei was removed after30 seconds.

Figure VIII-23 shows three x-ray microscopeimages of rat sperm nuclei. In (A) the rat spermnucleus is unstained, in (B) the nucleus is“stained” with mouse antiprotamine 1 andtagged with 400-Å diameter gold, and in (C) thenucleus is stained with antiprotamine 2antibodies and goat anti-mouse antibodiestagged with 400-Å diameter gold.

The images show distinct differences. Thestained images show high concentrations of goldalong the edge of the sperm and some evidenceof individual 400-Å gold particles on the surface.The clumping of the gold labels is particularlyevident in the image of the nuclei stained with


A) B) C)

10 µm

Figure VIII-23. X-ray microscope images of rat sperm nuclei prepared using different techniques:A) unstained, with no gold labeling; B) stained with antiprotamine 1 and gold-labeled, and C) stainedwith antiprotamine 2 and gold-labeled. The stained images show high concentrations (clumping) ofgold along the edge of the sperm and some evidence of individual 400-Å gold particles on the surface.The clumping is particularly evident in (B). The frame structures observed in the antiprotamine-1-stained nuclei (B) are a consequence of the preparation technique that unmasked internal sites forantibody binding. These images clearly show the value of gold labeling to enhance contrast, but alsoshow the need for avoiding unnecessary preparation techniques that can lead to artifacts.

the antiprotamine-1 antibody (B). The framestructures observed in the antiprotamine-1-stained nuclei are a consequence of thepreparation technique that unmasked internalsites for antibody binding. These images clearlyshow the value of gold labeling to enhancecontrast, but also show the need for avoidingunnecessary preparation techniques that canlead to artifacts.

XUV Interferometry

Optical interferometry has an importanthistory in the study and characterization ofplasmas. However, probing of high-densityand/or large plasmas has been difficult due tothe high absorption of optical probes, the effectsof refraction, and the impossibility of probingbeyond critical densities. Recently, a neon-likeyttrium x-ray laser has been used to develop thetechnology that will allow interferometry to beperformed at wavelengths of 155 Å, which has acritical density ~1025 cm–3.

In Fig. VIII-24, we show the experimentalsetup used to test a Mach-Zehnder interferometersuitable for probing high-density, long-scale-length plasmas. The output comes from astandard 3-cm-long yttrium x-ray laser with anoutput energy of 1 mJ in a 150-ps FWHM pulse.The output is collimated by a multilayer mirrorand injected into a Mach-Zehnder interferometer.The interferometer uses molybdenum/siliconbeamsplitters consisting of 8 layer pairs on1000 Å of silicon nitride. An imaging optic at theoutput of the interferometer images a planewithin the interferometer where a secondaryplasma is produced.

Figure VIII-25 shows a recorded interfero-gram of a plasma formed by laser irradiationof a mylar target. Excellent fringe visibility isobserved, clearly indicating the feasibility ofthis technique.

The limits of the current system are alsoapparent. In the figure we see that the blow-off ofthe plasma in the central region of the image is


Laser beam

Multilayer imaging mirror

Multilayer collimating mirror

X-ray laser target

Detector

Sample

target

Laser beam

Figure VIII-24. Experimental setup for x-ray laser interferometry suitable for probing high-density,long-scale-length plasmas. The output comes from a standard 3-cm-long yttrium x-ray laser with anoutput energy of 1 mJ in a 150-ps FWHM pulse. The output is collimated by a multilayer mirror andinjected into a Mach-Zehnder interferometer. The interferometer uses molybdenum/siliconbeamsplitters consisting of 8 layer pairs on 1000 Å of silicon nitride. An imaging optic at the output ofthe interferometer images a plane within the interferometer where a secondary plasma is produced.

0 400 800 1200 1600 20000

200

400

600

800

1000

1200

1400

Y(µ

m)

X(µm)

Figure VIII-25. Interferogram obtained using x-ray laser interferometer to image a plasma made byirradiating the surface of a mylar plastic with a second beam. The plasma-induced fringes can bereadily seen, clearly indicating the feasibility of this technique. The thick solid band runninghorizontally is the plastic sample, while the bright spot above the plastic sample is the self-emissionof the plastic.


sufficient to completely obscure the laser. This isdue to the existence of an oxygen transition nearthe laser wavelength at 155 Å, causing refraction.Further, the edges of the image show that thelimit of the fringe count is caused by attenuationas the plasma becomes dense near the surface.

Thus, although the x-ray laser interferometerhas been proven to work, there is still much workto be performed. This technology will beimportant in the characterization of NIF-scaleplasmas. In addition, with the NIF we can pushto shorter x-ray laser wavelengths, making this aneven more important diagnostic tool.


Intense radiation at wavelengths extendingfrom the far-infrared to the x-ray region of thespectrum can be generated by high-energy, laser-matter interaction. The characteristics of thisradiation depend in detail on parameters suchas the magnitude and mechanism of laserabsorption, relevant atomic physics, the equationof state of the heated material, and hydro-dynamics. In order to be able to tailor suchradiation sources to the needs of specificexperiments, considerable effort has gone intoboth detailed measurements of emitted spectraand extensive modeling work for a variety ofmaterials and laser conditions. As a consequenceof this work, scientists have developed sophisti-cated simulation tools that allow extrapolation ofthe capabilities that lie ahead using even higher-energy laser drivers, such as the NIF.

In the sections that follow we will discuss thetypes and capabilities of radiation sourcespossible using a high-energy laser driver, anddiscuss possible applications of these sources.First we discuss incoherent x-ray radiationsources; this includes both broadband sourcesand line emitters. Next we describe coherentsources; this includes an extrapolation of existingcollisionally pumped x-ray lasers, new inner-shellpumped x-ray laser schemes, and high-orderharmonic generation. The latter two of thesecoherent sources will take advantage ofultrashort-pulse capabilities on the NIF. Finally,we briefly discuss the potential of the NIF as ahigh-energy particle source.

Incoherent X-ray Sources

Broadband SourcesLaser-produced plasmas have long been

recognized as efficient sources of x-rays. Novaexperiments have already demonstratedconversion efficiencies as high as 50% from goldplasmas (see Fig. VIII-8 for the absolutelymeasured spectra of laser-irradiated targets ofdifferent elements, shown as a function of targetZ). For the NIF this implies that megajoules ofx-rays can be generated from a millimeter-sizesource. The effective blackbody temperature ofthis radiating plasma can be estimated from theStefan-Boltzmann relationship to be given byTBB(eV) ~ (10-5 I laser)0.25 . Here TBB is the plasmatemperature in eV and Ilaser is the laser intensityin W/cm2.

Radiation sources driven on flat targets canbe configured as point sources, with increasedspatial coherence, or as large-area radiators forefficient radiative energy transfer in a close-coupled geometry. The most intense radiationenvironment created using a high-energy laser isfound inside a radiation case, called a hohlraum.

In a hohlraum, laser energy is focused on theinterior walls of a hollow cylinder that has laserentrance holes on the ends. The interior walls ofthe hohlraum are vaporized and heated to a hightemperature. (This temperature has beenmeasured at Nova; typical data for its temporaldependence is shown in Fig. VIII-10.) Section III,Experimental Capabilities, under “HohlraumRadiation Sources,” has information about thetypes of hohlraum radiation environments nowavailable. With this information and a simplemodel (see Fig. III-3), simple extrapolations canbe made to the NIF. At the NIF, by the simplestscaling, the radiation temperature in hohlraumswill be increased by more than a factor of two,with energy outputs approaching two orders ofmagnitude larger than currently available.

Continuum x-ray sources, convenientlydescribed by effective blackbody temperatures inregions of the spectrum determined by thespecific ionic species contained in the plasma, caneasily be spectrally filtered and tailored for awide variety of applications, including opticalpumping of absorbing targets in materials


research and astrophysics. For examples refer toSection IV, Astrophysics and Space Physics;Section VI, Material Properties; and Section IX,Radiative Properties.

Laser-driven sources can be configured toefficiently yield hard x-rays, such as those usedin imaging and opacity experiments. At thehighest energies, implosion experiments on theNIF will achieve capsule temperatures of5−10 keV, which will produce extremely brightbursts of continuum, hard x-rays.

Short-pulse capability on the NIF will allowthe generation of intense, hard x-ray sources inthe ultrashort temporal regime. This capabilitywill allow the study of the dynamics of materialsundergoing rapid phase transformations on thetime scale of individual atomic motion.

Such studies in the past have been limitedbecause of insufficient photon flux and becauseof the lack of hard x-rays. The hard x-rays areneeded not only to heat but also to probematerials to sufficient depths to eliminate theeffects caused by surface interactions and theeffects of thermal wave propagation (shocks).(These effects are, of course, interesting in theirown way—see Section VI, Material Properties).However, here we have the capability to studyintrinsic material time scales. One intriguing pos-sibility is the potential of rapidly heated materialsto form metastable or short-lived phases; the typeof pump-probe experiment possible with the NIFwould allow observation of such materialthrough characteristic x-ray scattering.

Narrow or Line SourcesLine sources can yield narrow-bandwidth

radiation with increased temporal coherence.They represent an important diagnostic tool forhard x-ray imaging and opacity measurements.Conversion efficiency into a single spectral linehas been measured to vary from 1% at x-rayenergies below 3–4 keV to 0.1% beyond 4 keV.

The character of the emission spectra isstrongly affected by the dominant shell. Forinstance, K-shell emitters are dominated byisolated strong hydrogen-like and helium-likeline emission. In contrast, high-Z emitters such asgold are dominated by large bands of unresolvedtransitions. Reference 1 contains a compilation of

the experimental results for a variety of elementsand laser conditions.

A NIF laser will allow higher ionizationstates to be accessed, extending the region ofmonochromatic hard x-ray sources to the hardx-ray region. The application of narrow band-width sources for the probing of polycrystallinestructures is discussed in Section VI, MaterialProperties.

Coherent X-ray Sources

X-ray LasersPlasma-driven lasers now operate over a

wavelength range extending from the ultravioletdown to the soft x-ray region near 35 Å.14 Suchsystems utilize single-pulse optical lasers to pro-duce a hot and uniform plasma suitable for laseramplification and propagation. Figure VIII-26shows a graph of both calculated and measuredwavelengths of the dominant x-ray laser lines incollisionally pumped systems. Wavelengths areshown as a function of atomic number. With theNIF it would be straightforward to extend colli-sionally pumped neon-like 3p–3s and nickel-like4d–4p systems to short wavelengths and highoutput energies.

Previous Nova x-ray laser experiments havefound that the total irradiance required in a500-ps pulse of 0.53-µm laser irradiation scaleswith x-ray laser wavelength λ is

I (W/cm2) ~ 2 × 1017 (λ /20)–3.5

for neon-like systems andI (W/cm2) ~ 4.3 × 1015 (λ /20)–3.5

for nickel-like systems. Here λ is the wavelengthin Ångstroms of the short wavelength J = 0–1 innickel-like and the brightest J = 2–1 line inneon-like systems.

Extending neon-like x-ray lasers belowwavelengths of 20 Å will be difficult given thehigh power requirements. Nevertheless, the highgain of these systems will make it possible toproduce saturated x-ray lasers with plasmalengths of 2–3 cm.

Extrapolating the nickel-like system touranium with a J = 0–1 transition at 21.46 Å, weestimate that an irradiance of 4 × 1015 W/cm2

would be required to make an x-ray laser at this


0 50 100 150 200 250 300 35020

30

Ato

mic

num

ber

40

50

60

70

3p–3sNe-like

4d–4pNi-like

600 300 200 60100

Energy (eV)

Wavelength (Å)

80

90

100

Figure VIII-26. Collisionally pumped x-ray laser wavelengths as a function of atomic number. Thegraph shows calculated wavelengths (open symbols) and measured wavelengths (solid symbols). TheNIF would allow the probing of short wavelengths and high output energies.

wavelength. If we assume a 500-µm-wide linefocus, a 6-cm-long laser, and a 500-ps pulse, theenergy required to saturate a nickel-like uraniumlaser is approximately 500 kJ, well within thedesign limits of the NIF. The output x-ray laserenergy from a saturated amplifier is calculated tobe ~0.2 J. Using x-ray optics to focus the x-raylaser, we will be able to achieve peak intensitiesof 1 × 1017 W/cm2.

Numerous candidates have been proposedfor photopumped x-ray lasers, yet to date nosuccessful demonstration has been reported. Thedifficulty in these experiments remainsproducing bright x-ray pump sources that canefficiently couple to the candidate plasma. A NIF-size facility will allow large-volume pumpplasmas to be produced that can efficiently pumpcandidate plasmas and allow testing of thispromising concept.

The short-pulse capabilities of the NIF willallow us to extend x-ray lasers to the hard x-rayregime and to investigate a variety of new x-raylaser schemes, including inner-shell pumped andoptical-field ionized lasers.

Inner-shell pumped x-ray lasers rely onbright x-ray sources to produce inner-shellvacancies and direct population inversion. Thefast auger-decay rates of these systems make itnecessary to use extremely short x-ray pulses.The main advantage of inner-shell pumpedsystems is the prospect of generating short-wavelength lasers with low-Z atomic systems; forexample, 10-Å lasers from neon. Kapteyn,15 andmore recently Strobel et al.,16 have investigatedthe feasibility of these systems and have shownthat high gains are possible with bright 100-fsx-ray pump sources.

Recombination x-ray lasers, which rely onfast three-body recombination in a cold plasma toproduce an inversion, have long been viewed asan alternative to collisionally pumped systems,offering the potential for higher conversionefficiencies. To date, however, it has beendifficult to achieve large gain length products (oralternatively, saturated output) in these systems.A primary reason has been the difficulty ofproducing long uniform plasmas suitable for


inversion and x-ray propagation. A NIF-sizefacility will allow recombination systems to beadequately tested because of the large-volumeplasmas available.

Optical-field ionized x-ray lasers, which are avariant of standard collisional and/orrecombination lasers, rely on a short laser pulseto multiphoton-ionize a neutral gas. The shortoptical pulse reduces collisional heating—or itcan, by design, enhance it. This x-ray laserscheme has the potential for a short output pulse,on the order of 100 fs.

High-Order Harmonic GenerationThe generation of odd harmonics, up to the

eleventh, from the interaction of high-intensitylasers with dense gases had long beenobserved.17 However, researchers have nowmeasured harmonics extending to the 109th of811-nm radiation and the 143rd of 1053-nmradiation, by shortening pulse duration to avoidionization and by increasing intensity.18,19

The harmonics in these experiments resultfrom the periodic oscillation of quasi-freeelectrons across the atomic potential. As theelectron passes the nucleus, it suffers animpulsive distortion in its trajectory, whichgenerates a broad spectrum of radiation. Becausethe harmonics are produced only by the drivingfield of the laser, their pulse width is equivalentto that of the driving optical pulse.

Figure VIII-27 shows an example of the typeof data obtained from harmonic generationexperiments. In these experiments, high-orderharmonics are used to generate coherent XUVradiation. In rare gases this can produce aradiation source that could be useful in the 200-Åregion. The experiment is quite straightforward,as the laser is focused into a gas cell containing,for example, helium. The laser is an 800-fs-duration, 1.053-µm-wavelength system with upto 8 J. The diagnostic is a streak camera coupledto a XUV grating spectrometer that produces aflat field.

1600

1400

1200

1000

800

600

400

200

0150 200 250 300

17th

29th

Wavelength (Å)

Rel

ativ

e in

tens

ity

Wavelength (Å)145 330

0

23

Tim

e (n

s)

Figure VIII-27. Streak image of coherent XUV radiation generated by high-order harmonics. Thestreak image covers from 145 to 330 Å in the abscissa and from 0 to 23 ns in the ordinate. The long-wavelength, long-lived transitions are the Lyman series of hydrogen-like helium. The short-livedfeatures that are evenly spaced in wavelength are the odd harmonics of the laser. On the right is anintensity trace through the harmonics taken at a time indicated by the arrow on the streak image. Thetrace indicates the existence of the 17th through the 31st harmonic, occurring at 170 Å. The use of longfocal-length lenses will make it possible to use the NIF to generate coherent tunable XUV radiation.


On the left of the figure is a typical streakimage, covering from 145 to 330 Å in the abscissaand from 0 to 23 ns in the ordinate. The long-wavelength, long-lived transitions are the Lymanseries of hydrogen-like helium. On the right is anintensity trace through the harmonics thatindicates the existence of the 17th through the31st harmonic, occurring at 170 Å.

Coherence and overall conversion efficiencyare limited by phase matching between theharmonic field and the incident laser field. Recentexperiments have produced XUV harmonicradiation near 20 nm with conversion efficienciesas high as 10–7.20 Longer wavelength radiation,greater than 50 nm, can be produced with aconversion efficiency exceeding 10–6. The use oflong focal-length lenses will make it possible touse this technique on the NIF to generatemicrojoules of coherent tunable XUV radiation.

Applications of X-ray LasersBiological Imaging. The high brightness of

x-ray lasers makes them well-suited for a varietyof applications ranging from biological imagingand microscopy to nonlinear optics andinterferometry. X-ray microscopy offers a way tostudy structure in wet, thick, 2–10-µm, biologicalspecimens. Demonstrated resolution isapproximately five times better than that ofconventional and confocal optical microscopesfor samples that are thicker than the electronmicroscopes’ limit of 0.4 µm.21 X-ray microscopy,then, allows whole cells or organelles to bestudied without sectioning.

When dealing with wet specimens, theradiation dose imparted to the specimen byelectron microscopy leads to mass loss anddecomposition of specimens. Resolution is,therefore, limited by radiation damage andexposure time.

X-ray lasers, by virtue of their highbrightness, have the potential for producinghigh-resolution biological images beforesignificant damage occurs to the specimen. Forinstance, the tantalum nickel-like x-ray laser hasbeen used to image dry biological gold-stainedspecimens. However, developing full three-dimensional imaging of biological specimens willrequire a significant increase in the output energy

of the tantalum x-ray laser. A NIF-size facilitywill make high-contrast imaging possible, byallowing us to produce saturated x-ray laseroutput at laser wavelengths above and below thecarbon absorption edge at 44 Å.

Plasma Diagnostics. X-ray lasers arecurrently finding important applications asdiagnostics tools for high-density plasmas. Todate most of this work has used the neon-likeyttrium x-ray laser operating at 155 Å, because ofits high output power of 30 MW.22 X-ray laserdiagnostic techniques have significantadvantages over conventional optical diagnostictechniques because they reduce refraction effectsand eliminate problems with critical densitylayers. However, the high absorption of relevantplasmas at 155 Å creates a real need to push thework on x-ray laser diagnostics to shorterwavelengths in order to allow accurate densitymeasurements of cool plasmas at solid densities.

Nonlinear optics at soft x-ray wavelengths isan area largely ignored because of the difficultyin generating sufficient x-rays to makemeasurement viable. Numerous experimentshave demonstrated multiphoton processes in thenear UV by focusing optical lasers throughneutral gases. To extend these experiments to thesoft x-ray region, a plasma is necessary both toreduce absorption and to allow for resonantenhancement of the nonlinear susceptibility. Onepossibility that is particularly interesting is four-wave mixing, in which an optical laser is mixedwith an x-ray laser to produce a tunable soft x-raysource. Aside from its clear applications, theseexperiments would allow measurements ofnonlinear susceptibility to be made in a regimethat is largely unexplored.

Particle Sources

A NIF laser will permit the generation of awide variety of particle sources ranging fromhigh-energy electrons and ions to neutrons andother fusion products. The generation of high-energy electrons through parametric instabilitiesin plasmas is discussed in Section VII, PlasmaPhysics. Pair production using high-intensitylasers has long been proposed but is outside therealm of existing facilities. The demonstration of


pair production will be marginally possible witha NIF-size facility, as discussed in Section IV,Astrophysics and Space Physics.

At a high-energy facility like the NIF,implosion and subsequent burn will produce avariety of fusion products. Initial calculationsindicate that neutron production from implosionand burn could exceed 1019 in a short, 100-psburst. The use of these neutrons for producinguniform high-density plasmas exists, but thisapplication is seriously limited by the lowtemperatures of current facilities. The high fluxlevels of higher temperatures on the NIF can beuseful, however, for studying material propertiesin a regime where the generation and growth ofdislocations become nonlinear.

E. References

1. K. Eidmann, Phys. Rev. A 41, 3270 (1990).2. D. Kania, H. Kornblum, B. Hammel, J. Seely,

C. Brown, U. Feldman, L. Da Silva,B. MacGowan, D. Montgomery, C. Back,R. Doyas, J. Edwards and R. Lee, Phys. Rev. A46, 7853 (1992).

3. D. Matthews, E. Campbell, N. Ceglio,G. Hermes, R. Kauffman, L. Koppel, R. Lee,K. Manes, V. Rupert, V. Slivinsky, R. Turner,F. Ze, J. Appl. Phys. 54, 4260 (1983).

4. P. Carroll and G. O’Sullivan, Phys. Rev. A 25,275 (1982).

5. K. Eidmann and T. Kishimoto, Appl. Phys.Lett. 49, 377 (1986).

6. H. Nishimura, H. Takabe, K. Kondo, T. Endo,H. Shiraga, K. Sugimoto, T. Nishikawa,Y. Kato, and S. Nakai, Phys. Rev. A 43, 3073(1991).

7. D. Matthews, E. Campbell, N. Ceglio,G. Hermes, R. Kauffman, L. Koppel, R. Lee,

K. Manes, V. Rupert, V. Slivinsky, R. Turner,F. Ze, J. Appl. Phys. 54, 4260 (1983).

8. J. N. Nilsen et al., Phys. B 26, L243 (1993).9. J. N. Nilsen et al., Phys. Rev. A 48, 4682 (1993).10. J. N. Nilsen et al., Phys. Rev. Lett. 70, 3713

(1993).11. D. L. Matthews et al., Phys. Rev. Lett. 45, 110

(1985).12. B. J. MacGowan et al., Phys. Rev. Lett. 65, 420

(1990).13. L. Da Silva et al., Science 258, 269 (1992).14. B. J. MacGowan et al., Phys. Fluids 4, 2326

(1992).15. H. Kapteyn, Appl. Opt. 31, 4931 (1992).16. G. L. Strobel, D. C. Eder, R. A. London, M. D.

Rosen, R. W. Falcone, and S. P. Gordon, in“Inner-shell photo-ionized x-ray laserschemes” (SPIE Proceedings, Los Angeles,1993), Vol. 1860, pp. 157.

17. J. F. Reintjes, Nonlinear Optical ParametricProcesses in Liquids and Gases (AcademicPress, Orlando, Fl, 1984).

18. J. J. Macklin, J. D. Kmetec, C. L. Gordon, Phys.Rev. Lett. 70, 766 (1993).

19. J. K. Crane, H. Nguyen, M. D. Perry(submitted for publication).


21. L. B. Da Silva, J. E. Trebes, R. Balhorn,S. Mrowka, E. Anderson, D. T. Attwood,T. W. Barbee, J. Brase, M. Corzett, J. Gray,J. A. Koch, C. Lee, D. Kern, R. A. London,B. J. MacGowan, D. L. Matthews, andG. Stone, Science 258, 269 (1992).

22. L. B. Da Silva, B. J. MacGowan, S. Mrowka,J. A. Koch, R. A. London, D. L. Matthews,and J. H. Underwood, Opt. Lett. 18, 1174(1993).

Radiative Properties 133 Section IX

Section IX

Radiative Properties

The study of radiative properties using high-energy lasers has a long history. The irradiationof a surface with a high-energy laser providescopious x-ray emission that is easily accessibleand inexpensive.

Over the years the development of x-rayspectroscopy has led to further inquiries into theorigin of the radiation, kinetics processes ofimportance, and radiative transfer in theseplasmas. These are all grouped here fordiscussion. Some of the topics summarized beloware quite complex, while others are dealt with ina more schematic manner. This difference intreatment is due to the varying levels ofsophistication we now have in the various areas.

There are various constraints onexperimental aspects of the problem of radiativeproperties. In order to cover these constraints in acoherent way, the background material onradiative properties has been grouped so thateach topic builds on the previous discussions.Thus, the topics are grouped into fivesubsections:

A. Atomic Physics and Isolated EmitterSpectroscopy

B. Plasma-Emitter Radiative PropertiesC. Dynamics PropertiesD. Plasma SpectroscopyE. Radiative TransferThe various ideas for the study of radiative

properties in hot, dense matter or high-energy-density conditions fall naturally into these fiveareas. The areas can be traced to the variousdegrees of complexity that the surroundingmatter imposes on the radiator. For example, atthe first level (Subsection A, Atomic Physics and

Isolated-Emitter Spectroscopy) we are onlyinterested in the isolated ionic emitter, and thefact that there are interactions is significant onlyin that the interactions serve to create the highlyionized or specially populated ionic radiator. Thespectroscopy of highly stripped ions falls into thiscategory.

Subsection B, Radiative Properties of PlasmaEmitters, goes into the effect of the plasmaenvironment on the emitter.

Subsection C, Dynamics Properties, coversthe dynamics, or evolution, of the radiator in theplasma.

Subsection D, Plasma Spectroscopy, thendiscusses measuring quantities relevant to thesex-ray- or laser-produced plasmas.

And finally, Subsection E, Radiative Transfer,covers the transport of radiation, first forconditions of local thermodynamic equilibrium(LTE), and then for plasmas that do not satisfythe conditions for LTE, which is a more complexsituation.

A. Atomic Physics andIsolated-Emitter Spectroscopy

The production of highly stripped ions is nota straightforward extrapolation of laser energy.For example, in small 100-joule 1-ns-type lasers,it is possible to reach ion stages of K-shelltitanium. Further, the 24-beam Omega laser,which delivered 50 joules per beam, was used toprobe transitions between the 4n-type groundconfiguration and any excited configuration overthe entire periodic system.

Section IX 134 Radiative Properties

The limits for some other transitions are:• 2s2S1/2–2p2P1/2,3/2 and 2p2P1/2,3/2–

3d2D3/2,5/2 transitions in lithium-likesystems.Current knowledge: Ge29+ (Z = 32)Source for highest degree of ionization:OmegaDegree of ionization reached with theOmega laser: Ge29+

• 2s22p5 2P3/2,1/2–2s p6 2S1/2 transitions influorine-like systems.Current knowledge: Sn41+ (Z = 50)Source for highest degree of ionization:Nova 2-beamDegree of ionization reached with theOmega laser: Sn41+

• 3s 2S1/2–3p 2P1/2,3/2 and 3p 2P1/2,3/2–3d2D3/2,5/2 transitions in sodium-likesystems.Current knowledge: Gd53+ (Z = 64)Source for highest degree of ionization:Nova 2-beamDegree of ionization reached with theOmega laser: Sn39+

However, it is clear that using ever-higherlaser energies in a straightforward application ofdirect target irradiation will not necessarilypermit access to yet higher ionization stages. Thedifficulty is that much of the additional energywill go into kinetic directed energy and thecoupling efficiency will be relatively small. Onesolution for this is the use of the rather newlyperfected “gas-bag” targets. These are simpletargets made of a thin plastic bag, currentlycontaining large molecular structures of low-Zatoms. This allows the ambient pressure of thebag to be low (≥1 atm), while the fully ionizedtargets can have electron densities on the order of1021 cm–3. These are described in more detail inSection VII, Plasma Physics.

The study of the atomic physics of highlyionized atoms came into the modern era withthe availability in the late 1960s of reasonablyhigh-energy lasers. The production mechanismfor obtaining ionized material on a high-energylaser is straightforward: irradiate a solid surfaceof the element of interest with a laser at

intensities greater than the surface meltingpoint. This technique has produced spectra andline identifications of numerous highlyionized species.

The method has been improved with the useof a localized dot of the material of interest in amatrix of a lower-Z material, which will notinterfere spectroscopically. This helps isolate theelement of interest to a finite column that willmove outward along the laser axis from the solidsurface. The dot spectroscopy method isschematically illustrated in Fig. IX-1.1

The dot spectroscopy technique has beenwidely used to improve spectral resolution(which can be compromised by broadening dueto the finite source size), and to permit spectralinformation from spatially resolving instrumentsto isolate regions of the plasma in which there aresmall temperature and density variations.

Along this line we mention two results. First,we have the work of Seely and others performingclassical line identification on higher-Z elements.In this work a survey soft x-ray spectrometerwith wide spectral coverage is used to look in theXUV region. Results of this technique have beenused to obtain, for example, the configurationinteraction and level crossing data relevant to the3d104l4l ' levels in highly ionized zinc-like atomswith atomic numbers from 40 to 80.2 Theseexperiments are used to provide difficult-to-obtain data.

Second, the identification of highly strippedions in laser-produced plasma can lead tointeresting diagnostic possibilities. For example,the density-sensitive emission line ratio fromholmium XXXIX has been inferred from anexperiment where the XUV spectrum can be bothspatially and spectrally resolved.3 Fig. IX-2 showsan intensity-vs-wavelength plot of the lines ofholmium XXXIX at one spatial position. When atemperature of 1780 eV and an optically thinplasma are assumed, various ratios of theseemission lines form a reasonable diagnostic.This is illustrated in Fig. IX-3, which shows theratios of various line emission intensities vselectron density.

The final comment on spectroscopy of laser-produced plasma is to point out the interest in


Front view Top view

Laser spot

Dot ofmaterial of interest

Spectrometer

Low-Z substrate

Figure IX-1. Schematic of dot spectroscopy technique. The dot of the material of interest is placed on,or slightly beneath, a target that will not interfere spectroscopically. When the surface is irradiated, thedot is confined, reducing the spatial variations possible.

4s1/2 – 4p3/2

4p3/2 – 4d5/2

Inte

nsity

4d5/2 – 4f7/2

87 88 89 90

Wavelength (Å)

Figure IX-2. Example of the spectrum of lines ofholmium XXXIX n = 4–4 lines at one spatialposition. The lines used in Fig. IX-3 areindicated here.

u

u u u

u

† ††

†

††

232221201918

10

1

0.1

0.0117

Log (Ne)

Line

inte

nsity

rat

io

Figure IX-3. Ratios of various line emissionintensities:

u = 2P3/2–2D5/2 to 2S1/2–2P3/2† = 2D5/2–2F7/2 to 2S1/2–2P3/2o = 2P1/2–2D3/2 to 2S1/2–2P3/2

• = 2D5/2–2F7/2 to 2P3/2–2D5/2


the structures called unresolved transition arrays(UTAs), which are spectral features composed ofmyriad line transitions from complex ions. Thesestructures provide a wealth of information onthe plasma formation and atomic physics ofcomplex atoms.4

B. Plasma-Emitter RadiativeProperties

To study the effect of the plasma environ-ment on the emitter, you must create a plasmaand the ionic emitter in a condition that is well-characterized. A special plasma-productionmechanism is required to study the intrinsicprofile of a line transition, the shape of theresonances in a line series leading up to theionization potential, the interaction of the discreteresonances with the underlying continua thatoccur above the first ionization potential of anion, or the absolute shape and intensity ofthe continua.

In the past, the production of the ionicemitter in a hot and dense plasma and verifyingits existence was of sufficient scientific interest topermit the use of relatively simpler plasma-creation strategies; for example, the irradiation ofa solid target. However, when the goal is todisentangle the temporal and spatial variations ofthe plasma in order to determine thefundamental radiative properties of an ionimmersed in a plasma, a method must be foundto determine the plasma conditions—temperatures, densities, etc.

Determining the conditions of a plasmarequires systems that are sufficiently steady-stateand, ideally, have no gradients. The steady stateis required to ensure that the radiative property isnot a function of the time history of the system,but of the instantaneous conditions. The spatialconstraint (i.e., no gradients) comes from the factthat critical determination of the effects of interestrequires that the contribution from widelyvarying plasma conditions be minimized. As asimple example, in order to determine theexistence (or non-existence) of a transparencywindow between the last existing line in a seriesand the associated bound–free continuum, it

must be possible to ensure that there is no fillingof the proposed window due to the existence ofplasma at slightly different conditions.

Generating this type of single-density, single-temperature plasma is highly energy intensive.First, direct irradiation of a sample will producelarge gradients. Second, the problem withgradients has not to date been remedied by theuse of microdots of the element of interestembedded in a solid material. Direct irradiationwould therefore seem to be a poor method forobtaining the test-bed plasma for studyingplasma-emitter radiative properties.

However, there are experiments that mayindicate a way to create the ideal system. A greatdeal of work has been performed in thecharacterization of the opacity of elements withlow- and middle-range Zs. This work indicatesthat the entire capability of Nova is required toachieve tens of electron volts at 0.001 ofsolid density.

The reasons why the generation of single-density, single-temperature plasmas requires theentire capability of Nova are clear. First, thematerial from which the plasma is created mustbe volumetrically heated. Second, this volumetricheating is most easily achieved by creating anx-ray flux. And third, the sample must behydrodynamically isolated. (Hydrodynamicisolation keeps the sample from suffering shocks,etc., from laser deposition or indirectly fromcollision with laser and x-ray ablated material.)

The two processes of up-conversion of laserlight to x-rays and the volumetric deposition ofthe x-rays reduce the amount of energy reachingthe sample, resulting in less than 1% of the initiallaser energy being coupled into the sample. Theneed for hydrodynamic isolation, implyingseparation of the sample and x-ray source, causesthe coupling to drop further by an additionalmultiplier that will be less than 0.1.

These factors together result in less than 10–3

coupling of the initial laser energy into thematerial of interest. Experience in the use ofhohlraums and x-ray-irradiated samples,therefore, leads to a simple scaling for theproduction of characterized samples—that we arelimited to temperatures well below 100 eV, andprobably 50 eV is the maximum currently. It is


therefore implicit in all the discussions of theeffects that we wish to study that the greatlyincreased energy of the NIF will allow us toextend our range of temperatures and densitiestoward the higher-Z materials and regimesof importance.

C. Dynamics Properties

The next level of detail is the study of thedynamics, or evolution, of the radiator in theplasma. At one level, the effects discussed aboveof the plasma on the imbedded emitter are a formof kinetic process. By dynamics we mean here theactual time dependence of the detailed ionic statepopulations (i.e., the result of the rate equationsolution for the populations). The plasma-affected radiative properties (e.g., line shapes)discussed above require a plasma averaging thatcan, by the ergodic hypothesis, be replaced by atime average. The time average then indicates adynamic process; however, it is one where theaveraging process replaces the actual temporalhistory. In looking at the dynamics of theradiator, we are interested in the details of thetime histories, population densities, andpopulation mechanisms.

In studies attempting to isolate the dynamicalproperties of a hot, dense plasma, numerousprocesses play a role. The formation of atomicmodels can require exhaustive detail, and this initself has become a topic of research over the pastdecade. That is, some states of the atom may haverelatively little observable contribution to thespectrum, but play an important part in the flowof population. This makes the task of kineticsmodeling difficult, insofar as the observables aredark shadows providing only a vague sense ofthe processes required to create the populations.

In studies of dynamical properties the openquestions include difficulties with the truncationof high-principal-quantum-number states, theinclusion of states that have auto-ionizationchannels, and the level of detail for the states. Themost obvious example where theseconsiderations are critical is for the prediction ofinversion and laser gain in rapidly varying

systems. However, the evolution of theradiators in hot, dense matter in general is ofgreat importance.

The additional constraint on the experimentalsetup, beyond those discussed above in referenceto radiative properties, comes from the ability toprobe, in time, the level populations that maytake part in the formation of the populations.This potential creates an interest in developingthe ability to probe states that may not be seen inemission. The development that will be requiredis the ability to probe states and transitions bymethods that are similar to those that we requirefor radiative properties (discussed above) andthose required for radiation transfer (as discussedbelow), but with the additional fact that theplasma characterization must be performed in atime-dependent manner. Determination ofradiative properties from a single-temperatureand -density plasma may provide definitiveinformation; however, this is only a part of thepuzzle when it concerns the flow of populationdensity through the various ion species.

A critically important advance would be theaddition of a probe that could be used to generatefluorescence from excited states. Developing sucha probe is difficult. With low-energy-densityplasma, the equivalent to this experimentrequires lasers as the pumping source. In thehigh-energy-density regime, the production ofappropriate short-wavelength lasers is not aneasy task, and to date no line coincidences havebeen located for use of the existing soft x-raylasers as pump sources.

Thus, we await, in the current generation, anappropriate source. With the advent of the NIF,photon densities from quasi-continuous pseudo-Planckian sources could be spectrally filtered toprovide the necessary pumping for fluorescenceand pump-probe experiments. Until such time asa source of the magnitude of the NIF becomesavailable, we will have to rely on the search forcoincidences with existing bright laser-plasmasources, or the use of high-emissivity lineradiation from spatially separate plasmas, toprovide a photon pump of the same transitions inanother plasma.


D. Plasma Spectroscopy

The plasma produced by directly irradiatingtargets gives rise to states that are well outside ofthe defined limits of the condition of LTE in theplasma. The time dependence of the heatingsource, together with the fact that the plasmaexpands rapidly during heating by the laserpulse, indicates that these plasmas must bestudied using non-LTE (NLTE) techniques. This,in turn, means that atomic models coupled toNLTE rate equations must be employed topredict the populations in the plasma, and thusthese models are necessary for interpretation ofthe spectral character of the plasma emission ingeneral. The further complication that theplasmas are spatially varying indicates that thehydrodynamics of the plasma must also bemodeled for a fuller understanding of theplasma.

To understand laser-produced plasmas,much effort has been expended to develop adiagnostic complement that will provide data

independent of plasma simulations. Thesediagnostic methods include development of theinstruments (which is discussed in Section III,Experimental Capabilities), as well as of plasma-spectroscopic diagnostics. Many of thesediagnostics are complementary to thosedeveloped for astrophysical, atmospheric,or other laboratory plasmas.

One of the methods that has found greatsuccess is the previously mentioned dot-spectroscopy technique. This technique has beenperfected so that the temporal, spatial, andspectral information is obtained on a single shot,and it can also provide independent measuresof, for example, electron density. A schematic ofsuch a setup is shown in Fig. IX-4.5 Figure IX-5shows the results for the electron temperaturemeasurement that uses the slope of the bound–free emission continuum of the hydrogenicspecies. This experiment represents the type ofdetailed spectroscopy that can be performed onthese plasmas even though they are rapidlyevolving during the experiment.

4-frame UV probe beam

Laser spot overfills microdot target

Spatial imaging slit

Photocathode entrance slit

X-ray streaked crystal spectrograph

Time-integrated spatially resolved

crystal spectrograph

Framing crystal X-ray spectrometer

Plasma column of Z of interest

4-frame hologram

Figure IX-4. Schematic of a spectroscopy experiment that provides time, space, and spectralinformation. In this experiment a holographic interferometer is used to also provide an independentmeasure of the plasma electron density.


Distance (µm)

0

200

400

600

800

1000

1200

0 50 100 150 200 250

780 ps1050 ps1250 ps

Ele

ctro

n te

mpe

ratu

re (

eV)

Figure IX-5. Temperature measurements from aspectroscopy experiment of the type shownschematically in Fig. IX-1. The temperature fromthe hydrogenic bound–free recombinationcontinuum is measured for several frames on aspatially resolving spectrometer.

This brings us to an important point con-cerning the laser irradiation of a surface. Theplasma that is formed during the laser–matterinteraction will be dense and warm at the surface,becoming tenuous and hot as the matter movesaway from the target. Therefore, defining thetemporal and spatial history of the material is achallenging feature of these experiments. This, asone might expect, gives rise to a field of study ofits own.

One rather elegant advance was to use aseries of concentric dots of differing elements,bull’s-eye fashion (to isolate the spatial informa-tion in the direction perpendicular to the laseraxis), and then to use a gated spectrometer thatprovided spatial resolution along the laser axis.In this form of the experiment the NLTE charac-ter of the plasma can be both analyzed and at thesame time used for simple line identification.

The methods developed for use on laser-produced plasmas include the standard lineintensity ratio techniques, the use of spectral linebroadening, and most recently, a novel butsimple extension of the line ratios technique tothe use of two different elements.6

This new idea is illustrated in Fig. IX-6,which shows the ratios of the helium-like 1s2–1s3p transitions from titanium to chromium inthe same plasma conditions. Figure IX-7 shows

the results of an experiment on a long-scale-length plasma created by irradiating a bag of gasthat is underdense. That is, when fully ionized,the electron density (Ne) of the gas is less thanthe critical density (Nc) for the laser frequencyused. The laser heats the gas to a few kilovolts,and the chromium/titanium line ratios are agood diagnostic of this temperature. The lineemissions are obtained from chromium- andtitanium- coated fibers introduced into a bag ofn-pentane having an Ne of 1021 cm–3. Thechromium and titanium line ratios in Fig. IX-7indicate that the Te is ~3 keV.7

In terms of novel diagnostics, measurementtechniques have been developed that employ lineshapes,8 inner-shell absorption techniques,9 andquasi-steady-state models,10 as well as x-raylaser probes.11 Some these will be discussedbelow in Subsection F, Future NIF Experiments,under “Plasma Spectroscopic Topics.”

E. Radiative Transfer

One of the unique capabilities of high-energylasers is that they can reach sufficient energydensity, albeit for short times, to permit the studyof radiation transfer as a topic in its own right.Since the transport of radiation plays an integralpart in stellar interiors, as well as in inertialconfinement fusion (ICF) implosions, this hasbecome an area of some interest.

There are clearly two general regimes. In thefirst regime, when the transfer of radiation occursin material that satisfies the conditions for localthermodynamic equilibrium, the transfer isstudied as a flow of radiation through thematerial. In these experiments we are interestedin measuring flow characteristics as these arerepresented by quantities such as the Rosselandopacities and described by the radiation diffusionapproximation.

The second regime deals with plasmas thatdo not satisfy the conditions for LTE, and thedemands become somewhat more complex.Whereas in LTE we can use the Saha-Boltzmannequation or some related statistical method toobtain populations, in NLTE cases we requireatomic models of some sophistication that


2.0

1.5

1.0

0.5

2423222120Log(Ne)

5000 eV 4500 eV

4000 eV

3500 eV

3000 eV

2500 eV

2000 eV

1000 eV

500 eV

1500 eV

Rat

io o

f Cr

to T

i He-

like

1s2-1

s3p

Figure IX-6. Ratio of the chromium to titanium helium-like 1s2–1s3p intensities for a range oftemperatures and densities. Note that each curve represents a single temperature.

0.8

0.6

0.4

0.2

0.0

6.05.85.65.45.25.04.84.64.4keV

Ti XXI 1s2 - 1s2p

Cr XXIII 1s2 - 1s2p

1.0

Inte

nsity

Figure IX-7. A spectrum of the chromium and titanium produced by irradiating a large bag of gas witha single 3-ns long pulse of 3600-J energy. The chromium and titanium are on fibers introduced into abag of n-pentane, yielding an Ne of 1021 cm–3. The line ratios indicate the Te is ~3 keV. Because theelectron density of the gas is relatively undisturbed, due to the small amount of hydrodynamicmotion, temperature can be determined.


demand rates for all the processes of importance.In addition, the radiation diffusion equation isno longer useful, and we must solve the detailedradiative transfer equation that, in general, mustbe solved self-consistently with the rateequations.

LTE Radiation Flow

Currently there is a series of experiments tostudy LTE radiation flow—the transfer of heat byx-rays through targets of varying types. The basicexperimental geometry is shown in Fig. IX-8. Thex-ray drive for these experiments is created byfocusing eight of the ten laser beams at Nova intoa gold hohlraum. The x-rays from the hohlraumare incident on one end of a millimeter-scale tubecontaining one of several possible experimentalpackages (see Fig. IX-9). From the opposite end ofthe tube the backlight passing through the

package is observed with an x-ray crystalcoupled to an x-ray streak camera.

Backlighting illumination is created using theother two Nova beams. Behind the radiation flowtarget is a tiny fiber of samarium that is heated bylaser light. This produces a broad-band x-raybacklighting spectrum at photon energies fromaround 1 to 2 keV.

Figure IX-9 shows the various possibletargets used in the radiation flow experiments.(Hohlraums represent a feature of these, andmany other, experiments. For more informationabout the hohlraum, see Section III, Fig. III-1 andFig. III-2.)

The tube contains a series of thin fiducial foilsthat register the transfer of heat by allowing us toobserve the absorption spectra of Kα transitionsin the foils. As the fiducial foils heat, they gothrough several different ionization stages. Each

Streaked SSC using apoint backlighter or

Dante, SXRFC, SXI, SOP

Scale-1 hohlraumx-ray drive

Backlighter

Ly-α fiducialx-ray foil (Z1)

Laser beam

Radiativeflow

target

Dante, SXRFC, SXI,SOP

Figure IX-8. Schematic of radiative flow experiment. X-rays from a Nova hohlraum are incident on oneend of a millimeter-scale tube containing an experimental package. Eight of Nova’s ten beams arefocused on the hohlraum; the remaining two are used for the backlighter. (See Section III forinstrument names.)


Hole ingold washer

Z2

Doped or undoped foam

Occluding gold obstructions

Observationholes

Empty or foam-filled tube

Vacuum between Z1 and Z2

Z2

Figure IX-9. Possible experimental packages forradiation flow experiment shown schematicallyin Fig. IX-8.

ion stage absorbs different frequencies from thex-ray backlighting spectrum. The streak camerarecords the ionization history of each of thefiducial foils in the tube. From this, the timehistory of the temperature of each foil can beinferred.12 Fig. IX-10 presents an example of astreak-camera data record for radiation flowthrough a tube.

Both one-dimensional burn-through foils andmore complicated two-dimensional geometriesare being studied. The initial x-ray driving fluxhas been well characterized in terms of its time-,angle-, and frequency-dependence. Heat-transferexperiments are also being performed using atwo-dimensional x-ray framing camera tostudy the longitudinal distribution of heat inboth unobstructed and baffled straightcylindrical geometries.

These initial experiments have also beenextended using smaller hohlraums that createhotter radiation sources capable of driving longerand more complex two-dimensional geometries.The radiative drives from these hotter sourceshave also been well-characterized.

Non-LTE Radiative Transfer

Interest in the generation of x-ray lasers ledto a series of relatively recent advances in theoryand experiment. These advances have brought usto a position where detailed radiationhydrodynamics simulations are available and

Time

C-likeF-like

He-like

C-like

F-like

Si i

ons

Al i

ons

Figure IX-10. Streak-camera record of backlightpassing through a radiation-flow package in theexperimental setup shown in Fig. IX-8. Timeruns horizontally and the spectrum is vertical.At early time the silicon Kα lines are seen. Asthe radiation flows up the tube, at late time, thealuminum Kα lines rise up.

detailed kinetics models can be produced, withsome effort. (It is important to note the centralrole of the astrophysicist in the development ofthese simulation capabilities. Although thelargest part of the motivation for the experimentsshown here was derived from interest inlaboratory plasma sources, astrophysicists wereproducing the methodology for doing NLTEradiation transfer “correctly.”)

Thus, there are simulation codes that nowprovide a radiative-hydrodynamics capability forthe analysis of NLTE experiments. A descriptionof the basics of these codes is not necessary here,but the fact remains that these codes exist andthat they have been developed at great expense.Rigorous testing of the codes has essentially beenconfined to benchmarking numerical problemswhere analytic solutions exist, and no rigoroustests have been performed against controlledhigh-energy-density plasma experiments.

The first goal of the effort on NLTE radiativetransfer will be to develop a method to test thesimulation codes by making it possible tocompare the results of radiative hydrodynamics


simulations with experimental results. Thismethod must go as far as possible towardsallowing detailed comparisons to be performed.We must be able to provide the initial conditionsto the code more accurately than has beenpreviously attempted, and then provide theresults of the experiment. The hope is to provideexperimental results in the form of absoluteintensities so that the details of simulations canbe tested for the first time.

More specifically, there are three areas inwhich it is felt simulation capabilities should betested. First, the accuracy of the atomic kineticsmodels will be measured insofar as we canprovide spectral information of high enoughquality to check model details.

Second, the suitability of the kinetics modelsin simulating plasma dynamics can be tested.This testing will be made possible by the abilityto obtain experimental information on theevolution of the radiatively driven system.

Third, and most difficult, the level ofinformation that will be available on a NIF-scale facility may be sufficient to constrain theradiation transfer algorithms in these time-dependent plasmas with multilevelatomic systems.

Note that the testing of radiative transferwould require the generation of spectra thatcontain line transitions formed in a radiativelydominated regime. This is a very difficult regimeto attain in the laboratory, as can be seen from thesimple rule of thumb that for radiation todominate the collisions in the line formationprocess, the following must be true:

• The ε parameter (defined as the ratio ofthe collision decay to the radiative decayof a transition) must be small comparedto unity

• The optical depth τ of the transition mustbe large

• The source function S should reach theblackbody limit, B .

The ε parameter can be evaluated asε ≈ 10–4 Ne λ3

where λ is the transition wavelength in cm andNe is the electron density in cm–3. (Here weassume that the Gaunt factor is 0.2 and theplasma temperature is 100 eV.)

Next, the optical depth, τ , of a Dopplertransition can be approximated by assuming thatthe oscillator strength is ≈0.5, the temperature is≈100 eV and the atomic number is ≈20. Then

τ ≈ 6.3 × 10–10 Ng λ lHere l is the plasma column length in cm and Ngis the ground-state number density.

If we want the line formation to work in aregime where the source function reaches theblackbody limit (so that the medium is effectivelythick), we can relate ε to τ by the equation

Smax / B ≈ ετ ln τ( )for Doppler line profiles. Analysis indicates that,for example, a column length of 10 µm of plasmawith an electron density of 1020 cm–3 and 1019

cm–3 ground-state ions will yield (for 100-Åradiation) an ε ≈ 0.01, a τ ≈ 6.40, and a maximumsource function near the blackbody limit.

These conditions are difficult to attain, butare possible in the type of experiments that arecurrently being performed. The main object ofthis work to obtain a test bed for radiativetransfer experiments, therefore, would be todevelop the techniques to perform measurementsin such a plasma. Detailed studies of parameterregimes where ε is much smaller and/or τ ismuch larger are important to the verification ofthe radiation transfer schemes, but the goal of thecurrent experiments is to create an experimentalenvironment in which radiation dominates thecollisions in the line formation process so thatdetailed studies can eventually be performed.But, first things first.

The progress so far on the currentexperiments is as follows. First, based on severalconsiderations, an experimental configuration isnow being tested. The considerations are:

• The use of radiation enclosures (i.e.,hohlraums) is not necessary for theproduction of NLTE experiments. Theenclosure provides nothing of interest tothe experimental design, and in factcomplicates the measurement process tothe point of rendering the experimentalplan all but unattainable.

• The use of direct-drive experiments,although of great interest to thecommunity of experimentalists as afertile source of ideas and developments,


produces a plasma that is heated bymechanisms not relevant to the testing ofthe NLTE radiative hydrodynamicssimulations.

• The dimensionality of the experimentalsystem must be reduced to the minimumso that we can provide, as nearly aspossible, an ideal NLTE system. This isachieved by using a low-Z tamping layerto contain the material of interest.

• Because the temperatures that can bereached with current facilities are wellbelow the kilovolt range, spectralmeasurements should be made in thesub-kilovolt range. This will require aneffort to develop XUV instruments.

Given these considerations, the experimentcan be defined. A schematic of the experiment tostudy NLTE phenomena as it is being fielded isshown in Fig. IX-11.

The experimental plan is now going forwardon three levels. First, we have developed a fullunderstanding of the radiation heating sourceused in these experiments. The source, whicharises from a gold burn-through foil, has beenmeasured in time, spectrum, and angle (for moreinformation see Section VIII, RadiationSources).13

Backlight beam

Heater beam

2500-Å Au foil

2500-Å Au foil

~ 1000-Å Low Z~ 200-Å Mid Z~ 1000-Å Low Z

Spectrometer

Tamped sample:

Figure IX-11. Schematic of setup to study NLTEphenomena. Heating is achieved by using theflux from the rear side of a gold burn-throughfoil. The sample in this case is tamped, and thebacklight can be delayed in time forthe heating.

Second, we have developed methods toperform the absorption spectroscopy of theheated samples. Figure IX-12 shows an exampleof the spectrum from an untamped boron nitridetarget. The figure shows a comparison of thetheoretical absorption spectrum with theexperimental results for this case. The boronnitride sample is untamped, so effects due to thevacuum/matter interface will be present. Notethat this sample should not come too far out ofLTE, and comparisons have been performed thatindicate that this is true.

The third step is spectroscopy of the tampedsample and the relevant characterization of thematerial with small gradients. In Fig. IX-13 wesee the experimental results from first attempts atmeasuring a Stark-broadened absorption lineprofile to ascertain Ne as a function of time. Theresults are for absorption of a gold backlightthrough a plastic-tamped Teflon sample. Thefigure shows both the measured lithium-likefluorine and the line shape predicted by theory,and includes the densities inferred from themeasured results. Density can be determined as afunction of time by using a long-durationbacklight (i.e., absorption source) with an XUVstreak camera. Then the time-dependent width ofthe absorption line yields a time-dependentdensity. The development of in-situ temperaturediagnostics and further density diagnostics isbeing evaluated.

F. Future NIF Experiments

Spectroscopy of High-Z Elements

To provide a background to assist inunderstanding the importance of experimentsusing high-energy lasers, we will describe thecurrent state of knowledge of highly ionizedatoms as it relates to some key isoelectronicsequences. In particular, this description will helpin extrapolating those experiments to the NIF.

The current state of experimental knowledgeon isoelectronic sequences that lie between thosepresented below (i.e., K-shell, lithium-like,sodium-like, and copper-like ions) is even lesscomplete. A typical example of a current


1.0

0.8

0.6

0.4

0.2

0.0

Tra

nsm

issi

on

600500400300200

Energy (eV)

Experiment Radiative transfer simulation

Figure IX-12. Spectrum from absorption of untamped boron nitride. The experiment is the thick solidline; two predictions of the spectra are also shown. The thin line is the prediction of the NLTEradiative transfer code ALTAIR. Note that this sample should not come too far out of LTE, andcomparisons have been performed that indicate that this is true.

experimental setup is the dot spectroscopyschematic shown in Fig. IX-1. It is believed thatusing a source as powerful as the NIF laser withspecial foam target designs, or using othermechanisms for producing long-scale-lengthplasmas, can achieve coronal temperatures of6 keV, with eventual production of 10-keVplasmas. Such temperatures will make it possibleto probe most elements of the periodic table toany desired degree of ionization.

Observing transitions in higher-Z ions isimportant for determining the contributions tothe transition energies that become significantonly at high Z. These include QED andrelativistic contributions to the transitionenergies. In addition, electron correlation effectsare important at high Z. Comparison of observedand calculated transition energies motivates and

guides the improvement of the atomic theory ofhighly charged ions. For example, a great deal ofwork is presently being done on non-hydrogenicQED theory. Further, the atomic models areinput to the codes that simulate opacity andatomic kinetics.

The simplest spectra to interpret, and theeasiest transition energies to calculate, are forions with one weakly bound electron outside aclosed shell. These are in effect “one-electronspectra,” and they occur for lithium-like ionswith 3 electrons, sodium-like ions with 11electrons, and copper-like ions with 29 electrons.To date, the copper-like sequence has beenstudied to the end of the sequence (i.e., up tocopper-like uranium). The NIF would permit thestudy of other sequences, such as the sodium-likesequence up to uranium.


1.2

1.0

0.8

0.6

128.0127.9127.8127.7127.6127.5

Wavelength (Å)

1.9 ns;

3 ×10 20

1.8 ns;

4 ×10 20

1.6 ns;

5 ×10 20

1.5 ns;

7 ×10 20

Figure IX-13. Experimental results forabsorption of a gold backlight through a plastic-tamped Teflon sample. The measurementshows the lithium-like fluorine measured alongwith the theoretical prediction for the lineshape. The densities inferred from themeasured lithium-like fluorine line shape arealso shown.

Isoelectronic sequences that are near the one-electron sequences are also of interest. Examplesare the fluorine sequence with 9 electrons and themagnesium sequence with 12 electrons. Of thesetwo, the fluorine sequence has one electron lessthan the closed-shell neon-like sequence, andonly two transitions are intense in current laser-produced plasmas. Sequences such as thefluorine sequence are used to study the atomicphysics of “nearly closed shells.” The magnesiumsequence has two electrons outside closed shells,and electron correlation effects require detailedcalculation. This effect has become an importanttopic in atomic physics.

The work planned on high-energy lasersmust be referenced to the possibilities on othersources of high-Z ions such as the EBIT andTokamak devices. With regard to EBIT, only two

types of transitions are typically intense. Theseare the resonance transitions that are collisionallyexcited from the ground state of the ion, and thetransitions from the auto-ionizing levels that areresonantly excited by dielectronic recombination.With the EBIT it is not possible to efficientlyexcite the transitions between excited states. Thesame limitation applies to Tokamak spectra—only the ground-state transitions are intense.

In addition, although heavy ion acceleratorscan produce highly charged ions, Doppler shiftscomplicate the analysis. Also, synchrotronradiation can be used only for the study of coldmaterial, not highly charged ions. Thus, the NIFis necessary for the in-situ and complete study ofentire spectra of the highly charged ions thatcannot be produced by currently available lasers.

The Transparency Window

The possible existence of a transparencywindow in the spectral region between thebound–bound transitions and the formation ofthe bound–free continuum has been discussed formany years, and the issue is still unresolved.14

This possible effect is important because such adecrease in absorption probability can haveseveral serious effects. First, reduction of opacityin this spectral region can lead to overestimatesof the opacity of the models used. Second, therecombination rate of ions into the high-lyingstates could be seriously reduced by such atransparency window, changing the kinetics ofthe plasma. Finally, modification of the density ofstates will have an effect on the thermodynamicproperties of the plasma. This last effect is due tothe direct connection between the density ofstates and the partition functions, from which allthe thermodynamics properties are developed.

To measure radiative properties of this type(that is, those measurements that go beyondspectral identification), the phenomenon ofinterest needs to be isolated by creating highlyconstrained plasma conditions. Figure IX-14shows a schematic of a method that generates asingle-density single-temperature plasma forthese types of studies. To ensure uniform heating,the experiment shown uses volumetricirradiation with x-rays. The sample is tamped tokeep its density and temperature uniform.


40 NIFbeams X-ray flux

Mid-point

1/4 point

Edge

0.2 µmAl sample

CH tampers(3 cases: 0 Å, 50 Å, 10 µm)

X-ray flux

X-ray flux

X-ray flux

High-Zburn-through foil

40 NIFbeams

40 NIFbeams

40 NIFbeams

Figure IX-14. Schematic of proposed experiment to obtain uniform temperature and density. Toensure uniform heating, volumetric irradiation using x-rays is employed. The tampers are used to keepthe density and temperature of the sample at uniform levels.

The requirements for a thin sample and x-rayheating imply that little absorption of the x-rayenergy occurs. The low absorption, coupled withthe fact that the sample x-rays are produced bylaser-plasma generation, indicate that the laserenergy must be large to yield substantial plasmaheating. Combined with this is the requirementthat the sample must be hydrodynamicallyisolated, which means that laser-plasma creationmust occur at substantial distances. The distancefurther diminishes the coupling and increases thelaser energy requirements. Simple calculationsusing approximately 160 of the NIF beams ongold burn-through foils yield x-ray flux that willprovide researchers in the area of radiativeproperties an ideal plasma insofar as it hasthe correct conditions in which to performmeasurements.

Figures IX-15 and IX-16 show plots oftemperature and density as a function of time forsamples handled two different ways, one over-tamped and one undertamped. Figure IX-15shows a plot of temperature and density for a2000-Å aluminum sample that is overtamped toprovide a high degree of uniformity. The figureshows information for the three locations in thesample, the midpoint, the quarter-point (halfwaybetween the edge and the midpoint), and the

edge. The overtamping minimizes density andtemperature gradients in the sample. Over-tamping produces some extreme states of matterfor study—for example, note that at 1 ns thetemperature is over 65 eV and the density is 0.25g/cm3. This single temperature and density pointis, of course, attainable in many plasmas;however, the study of radiative properties ofsamples at this single condition will be uniqueand requires the capability of the NIF.

Figure IX-16 shows an untamped sample ofaluminum for the same thickness, with thegradients in temperature and density indicated.The figure shows information for the threelocations in the sample: the midpoint, thequarter-point (halfway between the edge and themidpoint), and the edge. The fact that the sampleis untamped leads to an exponential decay ofdensity and the formation of gradients. The outeredge of the sample blows down first, and there isa factor-of-three density gradient between theedge and the middle. The temperature shows thatthe lower-density outer region is hotter. This isconsistent with the conservation of energydensity of the sample. Clearly, tamping of thesample (as shown in the previous figure) willassist in the performance of single-temperatureand single-density experiments.


3.0

2.5

2.0

1.5

1.0

0.5

Den

sity

(g/

cm3

)

2.01.51.00.50.0

Time (ns)

100

80

60

40

20

0

Tem

pera

ture

(eV

)

mid-point quarter-point edge

Figure IX-15. Temperature and density vs time for an overtamped sample of 2000-Å aluminum. Theplots associated with the left-hand axis show density for the three locations in the sample. Themidpoint is shown as a thick solid line; the quarter-point (the position halfway between the edge ofthe sample and the midpoint) as a thinner line; and the edge as a dotted line. The plots associated withthe right-hand axis indicate the temperature. Again, the midpoint is shown as a thick solid line, thequarter point as a thinner line, and the edge as a dotted line. (Note that the quarter-point temperatureis not visible as a separate line, as it overlaps the mid-point temperature.) The density and temperatureenhancement at 0.4 ns is due to a shock generated in the rather thick 10-µm plastic tamper. Theovertamping minimizes density and temperature gradients in the sample. Overtamping producessome extreme states of matter for study—for example, note that at 1 ns the temperature is over 65 eVand the density is 0.25 g/cm3.

Transparency (and the other residual effects)are predicted to arise when two conditions exist.The first requirement is a strongly coupledplasma, or a state when the plasma-ion radiatorpotential energy is greater than the thermalenergy, so that the required approximation forthe plasma-ion interaction becomes complex. Thesecond is to have an ion described by a non-spherical potential, because for sphericalpotentials it can be shown that transparency andall the other residual effects do not occur. Thus,in strongly coupled plasmas with complex ions,the transparency effect may be important.

Figure IX-17 shows a calculation illustratingthe transparency window position and its effect.The logarithm of opacity is plotted againstenergy for an aluminum plasma at a temperatureof 450 eV and an electron density of 1019 cm–3.

The calculations are performed for two cases, onewith only Doppler broadening and one with bothDoppler and Stark broadening. The region wherethe ionization potential of the helium-like speciesoccurs is also shown. The Doppler profilespectrum in this region shows a substantialdecrease—this is the region of thetransparency window.

The existence of a transparency window is inquestion. The subtlety of the effect indicates thathigh-quality experiments are required to make aquantitative assessment of the effect, and this willrequire a NIF-scale facility.

The difficulty here is that to observe a changein the kinetics of the plasma or its thermo-dynamic properties requires a level of detailedinformation that is not possible, or thinkable, atthe present time. However, there is the possibility


0.001

0.01

0.1

1

Den

sity

( g

/cm

3 )

2.01.51.00.50.0

Time (ns)

100

80

60

40

20

0

Tem

pera

ture

(eV

)


Figure IX-16. Temperature and density vs time for an untamped sample of 2000-Å aluminum. Theplots associated with the left-hand axis show the logarithm of density for the three locations in thesample. The midpoint is shown as a thick solid line; the quarter-point (the position half-way betweenthe edge of the sample and the midpoint) as a thinner line; and the edge as a light line. The plotsassociated with the right-hand axis indicate the temperature. Again, the midpoint is shown as a thicksolid line, the quarter point as a thinner line, and the edge as a light line. The fact that the sample isuntamped leads to an exponential decay of density and the formation of gradients. The outer edge ofthe sample blows down first, and there is a factor-of-three density gradient between the edge and themiddle. The temperature shows that the lower-density outer region is hotter. This is consistent withthe conservation of energy density of the sample. (Figures IX-15 and IX-21 show an overtampedsample and an undertamped sample, respectively.)

that we can probe the region in absorptionbetween bound–bound and bound–freetransitions in a specially prepared plasma. First,the effect we are looking for is a subtle change inabsorption in a region that has bound–bound andbound–free transitions. Second, the very natureof the strongly coupled plasma problem indicatesthat the bound–bound transitions will berelatively broad and will form a quasi-continuum. The ability to differentiate theseeffects and obtain absolute absorption crosssections requires the ability to minimize thegradients of the system. Methods for generatinghot, dense plasmas of this type were discussed inSubsection E and here. Importantly, there hasbeen much effort spent on developing thetechniques to perform precision absorptionspectroscopy on these kinds of plasmas.

The proposed experimental setup to studythe transparency window might look similar tothe experimental setup shown in Fig. IX-14. Withthis configuration, small differences in opticaldepth (on the order of 0.1 or less) can bemeasured accurately. However, it is difficultto generate a benign environment that canbe probed.

Strongly Correlated Effects

It is important to make the distinctionbetween the concepts of strongly coupled plasmaeffects and strongly correlated plasma effects.Strongly coupled plasma effects concern only therelative contribution made by the plasmainterparticle Coulombic potentials to the plasmastatistical mechanics. The Coulombic potentials


Log

(rel

ativ

e in

tens

ity)

24002200200018001600Energy (eV)

Al K-shell at Te = 450 ev, Ne = 1019 cm-3

Stark Doppler

Figure IX-17. Transparency window (calculated) in the spectral region between the bound–bound andbound–free transitions. The logarithm of opacity is plotted against energy for an aluminum plasma ata temperature of 450 eV and an electron density of 1019 cm–3. The spectral features arise from thehelium-like 1s2–1snp 1P series and hydrogen-like Lyman series 1s–np, which overlap. The calculationsare performed with only Doppler broadening (indicated by the grayed line) and Doppler and Starkbroadening (indicated by the solid line). The inset shows the region where the ionization potential ofthe helium-like species occurs. The Doppler profile spectrum in this region shows a substantialdecrease—this is the region of the transparency window. Note that for the Stark-broadenedtransitions, transparency is substantially reduced.

are given by V~ Z1Z2e2/r12, where the Zs are thecharges on particles and r12 is the meaninterparticle spacing.

Strongly correlated plasma effects can beschematically defined as the regime in whichmodification of the interatomic potentials of theions by other ions in turn substantially modifieseffects on radiative properties.

The regions in temperature and densityparameter space where strong correlations occurare shown in Figs. IX-18 and IX-19. Figure IX-18indicates these regions with reference to physicalparameter configurations for a pure hydrogenplasma. The figure shows the regions of currentexperiments. The radiatively heated experimentsare shown in the left-hand rectangle, the plasmas

arising from spherical implosion in the upperright rectangle, and the planar shock experimentsin the lower rectangle.

To indicate the regions that are stronglycoupled, the contour where the strong couplingparameter Γ is unity is shown with a thicker line;the region below this line is strongly coupled.The strong coupling parameter Γ is defined as theratio of the ion-ion potential energy (Z2e2/r0) tothe thermal energy (kT). Z is the mean ion charge(here taken to be 1 for hydrogen), and r0 is themean distance between ions [i.e., equal to(3Z/4πNe)1/3].

The region where the electrons in a hydrogenplasma become degenerate (that is, the conditionwhere the chemical potential is zero) is indicated


104

103

102

10

1

10–3 10–2 10–1 1 10 102 103

Density (g/cm3)

Tem

pera

ture

(eV

)Γ = 1

µ = 0ρ−T track in Sun

Sphericalcompressions

Planar shockcompression

Uniformradiativelyheated foils

Figure IX-18. Temperature and density parameter configurations for various laser experiments. Theregions of current experiments are the radiatively heated experiments (left-hand rectangle), theplasmas arising from spherical implosion (upper right rectangle), and the planar shock experiments(lower right rectangle). To indicate the regions that are strongly coupled, the contour where the strongcoupling parameter Γ is unity is shown (thick line); the region below this line is strongly coupled. Theregion where the electrons in a hydrogen plasma become degenerate (that is, the condition where thechemical potential is zero) is indicated by the contour labeled µ = 0 (thin line). All regions below thisline require treatment of partial or completely degenerate plasmas. Finally, the regions that will beavailable with the NIF are indicated by the large light-gray area toward the top of the graph.

by the contour labeled µ = 0 (the thinner line). Allregions below this line require treatment ofpartial or completely degenerate plasmas.

Finally, the regions that will be available withthe NIF are indicated by the large light-gray areatoward the top of the graph. It is easy to see thatextension of these experiments to conditions ofinterest in the study of the radiative properties ofstrongly coupled matter will be greatly enhancedby the advent of the NIF.

Figure IX-19 isolates the temperature anddensity case for aluminum, showing the effects ofionization in the study of strongly coupledplasmas. The figure defines three regions. Theregion where classical plasma physics isappropriate is shown in white. In this regionDebye-Hückel theory can be used, because thereare many electrons in a Debye sphere.

In the dense plasma region (the gray area inthe middle), the Debye length is becomingcomparable to the interparticle spacing, and

high-lying states—the Rydberg states—of one ioncan interact with other ions to form band-likestates. This yields, among other things, loweringof the ionization potential.

In the high-density-matter region (the darkarea to the right of the dotted line), beyond thecontour where degeneracy becomes important,three things happen: the plasma now becomesclosely associated to a liquid metal; inner-shellstates form energy bands; and the ions aredensely packed, so distortion from neighboringions is substantial. The Γ = 1 contour is irregular,which is due to the fact that detailed effects ofionization of the aluminum create changes in theZ and thus in the average potential.

We can use as a measure of strong correlationthe amount by which the ionization potential ofthe ion is perturbed. This is a reasonablemeasure, in that it points out that strongcorrelations occur for some principal quantumnumber in all plasmas. But it also shows that


104

103

102

1

10

10–4 10–21 102 104

Classical plasma

Dense plasma

Tem

pera

ture

(eV

)

Density (g/cm3)

Γ = 1

Deg

ener

acy

High-densitymatter

Γ = 100

Γ = 10

Figure IX-19. Temperature and density diagramfor a specific element (aluminum). The diagramshows the effects of ionization on the region ofinterest to the study of strongly coupledplasmas. Three regions are defined. The regionwhere classical plasma physics is appropriate isshown in white. In this region Debye-Hückeltheory can be used, because there are manyelectrons in a Debye sphere. In the denseplasma region (the gray area in the middle), theDebye length is becoming comparable to theinterparticle spacing, and high-lying states—theRydberg states—of one ion can interact withother ions to form band-like states. This yields,among other things, lowering of the ionizationpotential. In the high-density-matter region (thedark area to the right of the dotted line), beyondthe contour where degeneracy becomesimportant, three things happen: the plasma nowbecomes closely associated to a liquid metal;inner-shell states form energy bands; and theions are densely packed, so distortion fromneighboring ions is substantial. The Γ = 1contour is irregular, which is due to the fact thatdetailed effects of ionization of the aluminumcreate changes in the Z and thus in the averagepotential.

modification of the radiative properties may notbe substantive. This means that, for example, thedepression of ionization potential, which isdiscussed below, can be well approximated by

simple means, and is not formally associatedwith strong correlations.

On the other hand, the formation of spectralfeatures due solely to the correlations (e.g., quasi-molecular formation) is considered an effect ofstrong correlation. Further, we note that it issimpler to get a strongly correlated plasma thanto get a strongly coupled plasma. For example,the plasmas in stellar interiors are very dense,and while their plasma properties are calculablewith weak coupling, or Debye, theories, strongcorrelations will exist because the overlap of theinteratomic potentials is large.

An example of these strongly correlatedeffects is the transient quasi-molecular effectgiving rise to satellite lines. There are severalsuggestions in the literature concerning thesesatellite lines. First, there is the prediction of amolecular resonance on the far wings of theLyman α transition of hydrogen in a plasma, dueto strong correlation with other hydrogenatoms.15 The upper limit of density for sucheffects can be estimated by considering that whenthe orbital of interest is at the mean ion sphereradius, quasi-molecular formation will occur. Forthe n = 2 state of a hydrogenic species of atomicnumber Z , the orbital radius is r = 6/Z a0, so thatthe limiting electron density would be Z4 × 7.46× 1021 cm–3. The quasi-molecules would form atdensities below this, but there would have to be aprobability distribution associated with thedynamics of the system; at higher density theplasma would form band structures.

Second, this strong correlation effect onradiative property has been numerically studiedfor a pure hydrogen plasma.16 This study doesshow that it is possible to get groupings of atomsin which the charge distribution is highlydistorted, resembling molecular clusters.However, these clusters are not necessarilybound. Static calculations performed on argon ina hydrogen medium yield configurations ofatomic positions from a molecular dynamicscalculation. Subsequent electronic structurecalculations over many configurations show thatbands and line broadening form in a non-perturbative manner, and the averaged resultsform one composite transition. Preliminary workfor these static calculations indicate that the n = 3


manifold of argon is heavily influenced by thehydrogen band.

Third, there have been claims of satellites,assumed to be quasi-molecular in origin, thatform in laser-produced plasmas. The effect wastentatively observed in the spectra from laserplasma produced by irradiating a lithiumfluoride planar target. This was done byanalyzing the Lyman β transition in fluorine withan emphasis on the short-wavelength wing. Theresults provided small features that may beattributable to quasi-molecular resonances.17

These observations are not sufficient to prove theexistence of the satellite lines and do show thedifficulty with current experimental techniques.

The performance of experiments on plasma-emitter radiative properties will require a moreidealized setup than employed in the direct-laserplanar target irradiation experiment. The effectssought are quite subtle, so working as close to agradientless case as possible will improve thechances of success. To have at the same time botha uniform plasma and a plasma in an extremestate requires the expenditure of a large amountof energy.

The highest level of success would beachieved by using a well-structured shock-basedsystem. For example, an adiabatic or isentropicshock could be produced using a tailored x-raydriving source. This may allow the possibility ofcompressing material to 10 times denser thanliquid density. The difficulty will be inmaintaining a low temperature in the shock-compressed region and devising a diagnosticcomplement that can observe the effect. There arefurther discussions of shock compression inSection VI, Material Properties, Subsection C.

Plasma Spectroscopic Topics

Spectral Line Shifts, Level Shifts, andContinuum Lowering

The measurement of spectral line shapes,including spectral line shifts, has been a topic ofcontinuing interest because of the diagnosticcapability provided by this non-interfering probeof plasma conditions. This has meant that foreach new plasma source there is a new andchallenging series of theoretical developments

that must follow to provide the diagnosticinformation.

Thus, on the one hand, experimentaldevelopment has fueled the need for line shapemodeling. On the other hand, the line shape,width, and shift, and the formation of the lineseries leading to the bound–free limit containsome of the most fertile ground for exploring thegeneral area of radiator-plasma interactions. Inthis way the underlying plasma-electron densityand temperature are important, but so are themore subtle facets of the plasma, such asfluctuation levels, composition of the plasma, andnon-equilibrium velocity distributions.

All of these facets can provide importantsignatures in the line spectrum and therebyprovide a method of testing theoretical statisticalmechanics and kinetics formulations, whichmeans the line spectrum has both diagnostic andintrinsic interest. Figure IX-20 shows an exampleof the competing effects of line shapes, linemerging, and continuum lowering for a K-shellaluminum sample at several densities and asingle temperature (450 eV).

The spectral features arise from the helium-like 1s2–1snp 1P series and hydrogen-like Lymanseries 1s–np, which overlap. In the lowest densitycase (at 1019, lower plot), the line spectrum hasmany states leading to the bound–freecontinuum. In the intermediate density case (at1022, middle plot), the lines are substantiallybroadened and there is a merging of levels, dueto the broadening near the continuum edge. Atthe highest density (at 1025, top plot), thespectrum shows the ionization potential to bedepressed by 250 eV, and thus only the stateswith principal quantum numbers equal to 2still exist.

It can be seen from these results that single-density, single-temperature samples will becritical to evaluating the complex interaction ofthe related mechanisms of line broadening,ionization potential depression, and line merging.The NIF holds out the promise of attaining theseplasma conditions in the ideal experimentalconditions.

One aspect of the line spectrum that isopened to study is the evaluation of line shifts.Line shifts are relatively small compared to line


Log

(rel

ativ

e in

tens

ity)

24002200200018001600

Energy (eV)

Ne = 1025

Ne = 1022

Ne = 1019

Al K-shell for Te = 450 eV

Figure IX-20. Effects of line shapes, line merging, and continuum lowering. The logarithm of therelative intensity of the aluminum K-shell transitions is plotted against energy for three plasmaconditions—three different electron densities at an electron temperature of 450 eV. The spectralfeatures arise from the helium-like 1s2–1snp 1P series and hydrogen-like Lyman series 1s–np, whichoverlap. In the lowest density case (lower plot, Ne = 1019 cm–3), the line spectrum has many statesleading to the bound–free continuum. In the intermediate density case (middle plot, Ne = 1022 cm–3),the lines are substantially broadened and there is a merging of levels, due to the broadening near thecontinuum edge. Increased plasma density has depressed the continuum edge by 25 eV. At the highestdensity (top plot, Ne = 1025 cm–3) , the spectrum shows the ionization potential to be depressed by 250eV, and thus only the states with principal quantum numbers equal to 2 still exist.

width, and are therefore difficult to measure andextremely difficult to calculate. No significantadvance, in terms of a theory, has been madeon the calculations of the spectral line shifts sincethe first formulations of the standard theorywere developed in the late 1950s by Barangerand Griem.18

The shapes of many lines have still not beenquantified, because there have been very fewhigh-energy-density experiments wheresupporting data on the temperature and densityof the plasma has been measured. To date thereare, for example, no experiments where thedensity and temperature have been measuredwith sufficient accuracy in any plasma with adensity higher than those probable by Thomsonscattering. Thus, although we have for twodecades diagnosed laser-plasma densities using

line widths, these have never been independentlyverified. The difficulty in obtaining single-density, single-temperature plasma is discussedabove. The main point is that hot, dense plasmais highly transitory and usually has largegradients. These hot, dense plasmas do not easilylend themselves to the production of high-precision measurements.

The situation becomes even more complex aswe attempt to study the entire line series with itsassociated continuum. The various processes thatarise as the series limit is approached are linebroadening, line merging, and ionizationpotential lowering. To distinguish these is verydifficult, but experiments that allow for acontrolled plasma environment may be able tosolve some of theses problems. As indicated inSubsection E, Radiative Transfer, under Non-LTE


Radiative Transport, single-temperature, single-density gradient data may allow us to separatethe effects of various mechanisms.

We note further that the various processesare in fact all due to the same perturbativemechanisms—the plasma particle fields cause thebroadening and shifting of a line, while at thesame time they change certain orbitals frombound to free. It is the interplay among thesevarious physical manifestations that forms thenexus for our understanding of radiator-plasmainteraction.

We note the fact that there is no completetheory describing the broadening of a bound–bound transition as it goes over to a bound–freetransition. From a purely phenomenologicalview, what happens is that the bound–boundresonance must continuously broaden in energyspace while the change to a bound–freeresonance is made. A schematic approach to thisbehavior is discussed by Liberman.19

To measure an effect of this type requires anexperiment that would have observably largeionization-potential depression and substantialline broadening. These two effects are largelycomplementary in that these both increasedirectly with electron density. Further, to be ableto isolate the line broadening and the transitioninto the continuum from the effects of linemerging (i.e., the Inglis-Teller limiting case), wewould like to go to high density and high Z inorder to depress the ionization potential to a lowenough principal quantum number to allow thebroadened spectral lines to be distinguishable(i.e., not merged).

Such experiments have been possible in thecore emission of gas-filled microspheresimploded by both direct and indirect laserirradiation.20 However, these experimentsindicate that the technique of using implosioncores is not sufficient for the task. Although theexperiments provide indications of the variouscritically interesting effects and an interestingset of possible explanations, the data is notof sufficient quality to validate anyparticular theory.

We note that the development of betterdiagnostics and smoother implosion-coredensities may be the solution, but these would

require the increased energy of the NIF. Theimplosion dynamics that are now available willnot allow adequate constraining of the problem.This is simply because the system is driven topeak densities for very short times, and thegradients thus produced are large. This isunacceptable for spectroscopy of the details ofthe line shape, line merging, and ionizationpotential depression processes. On the otherhand, using somewhat lower-Z elements thanargon as fill gases may provide someamelioration of the steep density effect bypermitting the study of the interplay of thevarious line shape effects at a lower density.

Ion DynamicsThe role of the dynamics of ions in the

formation of spectral lines in plasma has been atopic of intense investigation for the last twentyyears. Much effort has been expended since theexperimental verification that the Lyman αtransition of neutral hydrogen was seriouslyaffected by the dynamics of the ions, forcingmodification of the standard theory that assumesquasi-static ions. The fact that this discoverycame at a time when ionic emitter line profilesstarted to become ever more available from laser-produced plasma experiments led to much workon the ion dynamics of spectral lines that arisefrom ionic emitters.

However, to date there are no experimentsthat can be used to provide a serious test of thetheories for hot, dense plasmas. Verification ofthe ion dynamics contribution is of criticalimportance for two complementary reasons.First, the use of spectral line broadening may beseriously compromised if the ion dynamicsseriously change the widths of the spectral.Secondly, even in those cases where the iondynamics are negligible, it is only possible toknow this by access to a verified model.

To study ion dynamics requires the sametype of uniform plasma conditions that arerequired by the study of plasma-emitter radiativeproperties, and the discussion on that subjectshould be consulted. Additional considerationsfor these experiments require observation of lineprofiles from the same ionic radiator emitting atthe same plasma conditions but from plasmaswith different ion perturber masses. The use of


implosions is attractive from the point of viewthat hot, dense conditions can be reached.However, both the gradients in these implosionsand the difficulty in obtaining temperature anddensity histories with different atomic species inthe fill are unsolved problems. On the otherhand, this area is ripe for investigation.Alternatively, the tamped-sample-type plasmamay be the solution, but reaching the correctconditions will definitely require NIF-sizeexperiments.

Continuum MeasurementsThe problem of providing continuum

measurements is similar to the difficulty ofmeasuring the absolute values of the continuumin any given experiment. The spectral localizationof bound–bound or even bound–free edgesprovides a great simplification in separating outsignal from background compared to thedetermination of continua. This is true in allplasmas, but when we add the hot, dense and/orstrongly coupled (and correlated) effects we arefaced with, a daunting problem arises in how toproceed with measurement. While it is clear thatinteresting processes, such as strongly correlatedscattering cross sections, can come into play inthe formation of continua, what is not clear ishow to proceed.

The experimental setups discussed above(e.g., Fig. IX-14) are not sufficient in themselves toprovide good continuum measurement. Theadditional constraint is that the materialsurrounding the material of interest, whether itbe a tamped sample or an implosion, must havecontinuum contributions that are either small orcompletely quantified. It is easy to envisionkeeping the contributions small. In such cases thedensity of the material of interest would have tobe increased relative to, for example, a sampleprepared for the line shape measurements, wherethe optical depth in the lines would be keptsmall. However, with increased thickness comeincreased plasma gradients, and the problems ofvolumetric heating are exacerbated. This makesthe energy-rich environment of the NIF evenmore desirable.

In this context it would be appropriate tonote that the important aspects of bound–freeedges and continuum processes are at some level

the same, with the formation of free–freecontinua being an additional part of the puzzle.Further, the study of auto-ionization processesand how these behave in dense plasmas is also ofgreat interest, not only in the study of plasma-emitter radiative properties, but also in the areaof kinetics and radiative transfer.

Population Kinetics

The study of the dynamics, or kinetics, of anatom immersed in a plasma is at the center of theability to model time-dependent behavior ofatoms in plasma. For the present discussion weassume that radiative transfer is not an integralpart of those kinetics. (Situations where theradiative transfer is important are discussed inthe next section, Radiative Transfer and LineFormation.)

Even when the radiative properties havebeen understood, the kinetics of the plasma arestill in question. The reason is quite simple; inorder to model population dynamics, the rates(or more accurately, the cross sections) for all theprocesses involved must be well known. Further,in those cases where non-Maxwell-Boltzmanndistributions are important, the kinetics behaviorwould have to be accompanied by ameasurement of the velocity distributions in lieuof temperature measurements.

Knowing the rates and measuring velocitydistributions is an extremely difficult task and isnot made simpler by choosing to study thebehavior of the atom in hot, dense media.Although the densities may be high and give riseto plasmas that rapidly thermalize the velocitydistributions, the rates for ionization andrecombination are also rapid. Thus the dynamicionization balance can be dictated by non-thermal distributions. This will be particularlyimportant for experiments with short-pulse,intense laser–solid matter interactions.

In the laser intensity regimes of 1015 W/cm2

and above, which will be realized for sub-picosecond pulses, the electron distribution thatis produced in the laser–matter interaction ishighly non-thermal. This non-thermal electrondistribution is created through non-localtransport of electrons from the laser interactionregion to a region of higher density, where the


electron temperature (as measured by theelectron subsystem energy content) is predictedto be up to ten times higher than the local iontemperature. Ionization, therefore, is of greatinterest. However, the more mundane plasmas(at least by these standards) created byvolumetric x-ray heating or implosion of amicrosphere have the same interesting problems.

Further, we note that at these high intensitiesthe laser electric field equals or exceeds thebinding field in many atoms. In such cases, thelaser field must be included non-perturbatively inthe calculation of atomic and molecularproperties. The laser field also affects scatteringprocesses. For example, Feshbach-like resonancescan be induced even in such a simple system aselectron + proton, which without the field canhave no resonances. Of course, these studies ofintense field effects are in their infancy, but theywill have to be addressed at some stage in futureplasma models and experiments.

To map the kinetics of atoms in a plasmarequires a rich set of diagnostics that can providespatial, temporal, and spectral coverage of thesame event. This requires not only a well-characterized plasma, but a resource base that isbeyond the level that can be brought to bear byan independent effort without the support of afacility of the scale of the NIF. Further, primaryinformation on the plasma character must beavailable to corroborate the results. Again, theability to perform temperature and densitymeasurements is critical to the success of theseexperiments.

We note here that the numerous studies todate on the kinetics of plasma formed by directillumination (for example) have left manyunresolved issues. This is not for lack of desire,but mostly for the lack of both resources andplasma-generating devices that would allowcritical investigations. The NIF would providesuch a research environment.

Below we will discuss one type of experimentthat is motivated by astrophysical considerations.However, the range of areas in which kinetics isof importance is much broader. The mostimportant factor is that with the plasma-production capability and the diagnosticcomplement that the NIF will offer, the complex

mix of processes relevant to the kinetics problemcan be attacked in a coherent manner.

Radiation Transfer and Line Formation

The study of the formation of spectra fortime-dependent hot, dense media is the culmina-tion of all the topics that have been discussedabove. First, the radiative properties are requiredas initial input, and second, the kinetics modelmust be verified by experiment so that one canisolate the radiative transfer effects. With thesetwo in place, it is possible to study the non-localeffects that arise in the formation of the spectrum.

For these effects one must solve the radiativetransfer equation in tandem with the equationsfor the population dynamics. Because each affectsthe other, the problem is a compounding ofprevious investigations. Here, the regimes wherethe radiation transfer will be severely stressedwill also be challenging for experiments. Part ofthe problem is the need to quantify the sources ofthe radiation field and radiation matter coupling.The rest of the problem is to create high-optical-depth, radiatively dominated transitions. Theseare extremely difficult in the present generation,and the NIF will certainly be required toinvestigate the regime thoroughly.

Fluorescence has been used in the study ofkinetics to probe the movement of populationaround a network of levels, thus allowing thedirect study of the redistribution of population.However, the difficulty in generating efficientphotopumps is dependent on the radiativetransfer process. For hot, dense matter, pump-probe experiments are beyond the realm ofpresent-day laser systems. There may be thechance line coincidence between a soft x-ray laserand a transition of interest, but other than this bitof serendipity, pump-probe experiments willhave to wait for the increased photon densitiesavailable with the NIF laser. Then it will bepossible to use filtered backlights of sufficientbrightness to pump line transitions in hot,dense matter.

A schematic experiment would look roughlyas follows. We take a large low-Z tamped sampleof mid-Z element. The sample is irradiated withan x-ray source that is filtered to ensure that thereis no surface deposition due to critical density


effects. Next, with the tamping controlling thegradient scale-length in the problem, we performabsorption and emission studies, in differentspectral regimes. It should be possible to obtaintemperatures of >100 eV for centimeter-sizesamples. The limiting factor here is thathydrodynamic isolation is absolutely required toensure that blast wave and debris do notcompromise the results.

Next, a photopump of measured temporal,angular, and spectral content is used to pump theplasma. The technique can now be used toobserve line formation in the chosen transitions.

It would then be possible to explore thecontrol of velocity gradients, and more generallythe interaction of the radiative transfer and thehydrodynamics of the plasma. We note again thatthis is a regime that is completely beyond currentcapability. Measurement of the velocity fields inaddition to the other parameters required wouldbe impossible in the current experiments.

Figure IX-21 shows a calculation of thebehavior of an undertamped aluminum samplewith a modest gradient. Undertamping leads tocontrollable density and temperature gradients inthe sample. Obtaining these conditions in acontrolled manner requires the energy of a NIF.This undertamped case, together with theovertamped and untamped cases shown inFigs. IX-15 and IX-16, indicate the degree ofcontrol that the x-ray heating flux will provide tothe users of the NIF.

We note that with hot, dense matter, for thevelocities to be of concern, they must be largerthan the line width, which we assume is given bythe Doppler width of the transitions of interest.Thus, for example, for a 100-eV plasma withZ ~ 16, velocities in excess of 5 × 106 cm/s wouldbe required. Such velocities are easily obtainedfrom a free expanding surface and can beassociated with untamped surface expansion.

0.0001

0.001

0.01

0.1

1

Den

sity

( g

/cm

3 )

2.01.51.00.50.0

Time (ns)

100

80

60

40

20

0

Tem

pera

ture

(eV

)


Figure IX-21. Temperature and density vs time for an undertamped sample of 2000-Å aluminum. Thisrepresents the time evolution of a plasma that can be used to study radiative transfer and lineformation in a controlled system. The plots associated with the left-hand axis show density for thethree locations in the sample. The midpoint is shown as a heavy solid line; the quarter-point (theposition half-way between the edge of the sample and the midpoint) as a thinner line; and the edge asa dotted line. The plots associated with the right-hand axis indicate the temperature. Again, themidpoint is shown as a thick solid line, the quarter point as a thinner line, and the edge as a dottedline. Undertamping leads to controllable density and temperature gradients in the sample. (Figs. IX-15and IX-16 show an overtamped sample and an untamped sample, respectively.)


Strong Magnetic Field Effects

It is straightforward to deduce fromAmpere’s law that hot, dense plasmas with smallscale lengths might contain strong magneticfields, with B > 106 G, and indeed this is the case.Flux compression yields megagauss fields as amatter of course in explosive pinch experiments,and suprathermal electron currents arising fromlaser plasmas can produce fields exceeding 100megagauss.21,22

The behavior of matter in such strongmagnetic fields continues to pose difficultquestions in atomic/plasma physics, withimportance for astrophysical as well as laboratoryapplications. For instance, some white dwarfstars have megagauss surface fields, andgigagauss fields are usually ascribed to neutronstars. Strong, primordial fields may also exist inthe interiors of giant planets and in the cores ofnormal stars, including the sun, but theirpresence cannot be deduced with our primitiveunderstanding of how equations of state andtransport coefficients are modified by largeB-values.

Recent years have witnessed considerableprogress in understanding certain effects of thehuge 1012-G fields expected in neutron starcrusts.23 However, the fact that Coulombinteractions are relatively unimportant thereactually makes it easier to develop goodtheoretical models for this regime than for theintermediate 107- to 109-G regime, wheremagnetic and Coulomb effects are comparable24

and perturbation methods are reliable. The suiteof diagnostics expected to be available forstudying high-energy-density plasmas with theNIF will provide a unique opportunity to probespecific issues related to intensely magnetizedplasmas, such as:

• How do the strong B-fields modifyplasma ionization balance?

• How important are the anisotropicpressure effects that arise when B2/8π iscomparable to the gas pressure?

• What is the effect on transportcoefficients of non-spherical atomiccharge distributions that occur whenB > 108 G?

• Is it possible to use Stark broadening bymotional (i.e., v × B) electric fields as areliable diagnostic of the strong B-fieldsin plasmas?

G. References

1. P. G. Burkhalter et al., Phys. Fluids 26, 3650(1983).

2. J. F. Seely et al., “Zn-like 4l 4l',” submitted toPhys Rev. Lett.

3. J. F. Seely et al., “Ho XXXIX density sensitivelines,” accepted to Phys. Rev. E .

4. See, for example, J. Bauche, C. BaucheArnoult, and M. Klapisch in Advances inAtomic Physics , D. R. Bates, Ed. (AcademicPress, New York, 1988).

5. B. K. Young et al., Phys. Rev. Lett. 61, 2851(1988).

6. R. Majoribanks et al., Phys Rev. A 46, R1747(1992).

7. W. Hsing, B. MacGowan et al., privatecommunication.

8. R. W. Lee, AIP Conf. Proc., E. Marmar andJ. Terry, Eds. (American Institute of Physics,New York, NY, 1992), Vol. 257.

9. T. S. Perry et al., Phys. Rev. Lett. 67, 3784(1991).

10. B. K. Young et al., Phys. Rev. Lett. 62, 1266(1989).

11. L. Da Silva et al., SPIE Proceedings (SPIE) Vol.2012, page 158.

12. T. S. Perry, Phys. Rev. Lett. 67, 3784 (1991).13. D. R. Kania et al., Phys. Rev A 46, 7853 (1992)

and C. A. Back et al., JQSRT 51 (1994).14. Some theoretical discussions of the

transparency window include: L. G.Dyachkov et al., JQSRT 44, 123 (1990) andB. G. Berkovsky et al., J. Phys. B 26, 2475(1993). Some experimental work includes:V. E. Fortov et al. in Strongly Coupled Plasmas,S. Ichimaru, Ed. (Elsevier, 1990), p. 571; andPopovic & Dodevic, in Strongly CoupledPlasmas, H. Van Horn & S. Ichimaru Eds.(University of Rochester Press, 1993), p. 273.

15. S. J. Rose, Journal de Physique Coll. Ser. C 8(1982); using methods of D. R. Bates, MNRAS112, 40 (1952); D. Salzmann and J. Stein, Phys.Rev. A 45, 3943 (1992).


16. I. Kwon, L. A. Collins, J. D. Kress,N. Troullier, and D. L. Lynch., Phys. Rev. E49, R4771 (1994).

17. E. LeBoucher-Dalimier, A. Poquérusse,P. Angelo, I. Gharbi, and H. Derfoul, JQSRT51, 187 (1994).

18. See the references in H. R. Griem, SpectralLine Shapes from Plasmas (Academic Press,New York, 1972).

19. D. Liberman and J. R. Albritton, JQSRT 51,197 (1994) and the references therein.

20. B. A. Hammel, C. J. Keane, M. D. Cable, D. R.Kania, J. D. Kilkenny, R. W. Lee, andR. Pasha, Phys. Rev. Lett. 70, 1263 (1993); andC. F. Hooper, D. P. Kilcrease, R. C. Mancini,

L. A. Woltz, D. K. Bradley, P. A. Jaanimagi,and M. C. Richardson, Phys. Rev. Lett. 63, 267(1989).

21. N. Miura and F. Herlach, in Strong andUltrastrong Magnetic Fields and TheirApplications, F. Herlach, Ed. (Springer, Berlin,1985).


23. P. Meszaros, High-Energy Radiation fromMagnetized Neutron Stars (U. Chicago Press,Chicago, 1992).

24. E. P. Lief and J. C. Weisheit, Contrib. PlasmaPhys. 33, 471 (1993).

Techn

ical In

formation

D

epartmen

t • L

awrence L

ivermore N

ational Laboratory

University of C

alifornia • L

ivermore, C

alifornia 94551

Documents

Science on High-Energy Lasers: From Today to the NIF Page... · 2011. 3. 22. · June 1995 UCRL-ID-119170 LAWRENCE LIVERMORE NATIONAL LABORATORY University of California • Livermore,