FUSION SCIENCE CENTER FOR EXTREME STATES OF MATTER

Transcription

1 FUSION SCIENCE CENTER FOR EXTREME STATES OF MATTER University of Rochester Ohio State University Massachusetts Institute of Technology Lawrence Livermore National Laboratory University of Texas at Austin University of Nevada at Reno General Atomics University of California at San Diego University of California at Los Angeles Institute for Laser Science Applications ANNUAL REPORT April 2009 April 2010 FSC

2 EXECUTIVE SUMMARY The Fusion Science Center (FSC) for Extreme States of Matter and Fast Ignition (FI) Physics began operation in August 2004 under the sponsorship of the Department of Energy (DOE), the Laboratory for Laser Energetics (LLE), the University of Rochester (UR) and the Institute for Laser Science Applications (ILSA) at Lawrence Livermore National Laboratory. In 2009, $150K in matching funds were provided by the New York State Energy Research and Development Authority (NYSERDA) through LLE, UR contributed with $81K of indirect-cost waivers, and the DOE provided $1.57M for FY10 under cooperative agreement DE-FC02-ER Ten institutions and ten principal investigators participate in the FSC. The FSC is aimed at fostering close collaboration among the Center investigators with the ultimate goal of successfully integrating compression and FI-heating experiments. A major goal of the Center is to bring together researchers from around the country and the international community to build a comprehensive understanding of the physics underlying the creation of extreme states of matter and the exploration of advanced inertial-fusion concepts. Another major function of the Center is to stimulate academic involvement and student interest in the area of High Energy-Density Physics (HEDP). As part of its outreach and academic mission, the FSC organized the 2009 Summer School in High Energy Density Physics. FSC Outreach and Education Mission The FSC held its third summer school in High Energy Density Physics the week of July 26-August 1, 2009, in Los Angeles (CA). The summer school was held on the Campus of the University of California Los Angeles. The FSC provided forty-eight scholarships to graduate and undergraduate students. Ten scientists from the field of High Energy Density Physics lectured on a broad range of subjects including: laser-plasma interactions, laboratory astrophysics, equations of state, plasma-based particle accelerators, inertial confinement fusion, high-energy lasers, Z-pinches, material science under extreme conditions, and HEDP diagnostics. Student and participant posters accompanied each day s thematic focus. Daily poster sessions offered the students a unique opportunity to discuss their work in detail, not only with the lecturers but with the many other students who were in attendance. More detailed information concerning the 2009 HEDP Summer School can be found in Section 4 is available on the FSC web site at Progress in HEDP research and Advanced Ignition Concepts The FSC has promoted close collaboration among several HEDP researchers (some supported by the Center and some by other sources) in the area of laser-plasma interactions, electron transport, hydrodynamics and target fabrication, with the main focus of these disciplines centered on high-gain concepts for inertial confinement fusion. The FSC identified the following fundamental areas critical to Fast Ignition and HEDP: [1] Shock Ignition [2] Fast Ignition [3] Magnetized High Energy-Density Physics 1

3 Highlights on Shock Ignition Laser-plasma instabilities were studied in shock-ignition experiments with peak laser intensities of ~ W/cm 2. Plastic shells were compressed on a low adiabat by 40 of the 60 OMEGA beams. The remaining 20 beams were delayed and tightly focused onto the imploding shell to generate a strong shock. Good coupling of the shock-beam energy was observed leading to a neutron-yield increase of about A significant amount (up to 36%) of laser energy from the high-intensity beams was backscattered. Stimulated Raman Scattering (SRS) dominated the backscattering with some contribution from stimulated Brillouin scattering (SBS) but no significant contribution from the twoplasmon decay instability (TPD). About 10% of the high-intensity beam energy was converted into hot electrons. A hot-electron-energy distribution was generated with a temperature of 42±5 kev, independent of laser intensity. This is beneficial for shock ignition since these electrons are stopped in a thin outer layer of the imploding target, augmenting the strong hydrodynamic shock. More details are described in Section 7 and in a paper submitted for publication in Physical Review Letters. A collaboration on shock ignition has been initiated with the French group (G. Schurtz, X. Rubreye) at CELIA (U. Bordeaux), the Italian groups at the University of Milan (D. Batani) and Rome (S. Atzeni), and J. Perkins at the Lawrence Livermore National Laboratory. The European collaborators will contribute with 1D and 2D simulations of shock ignition experiments proposed for the OMEGA and the NIF laser. Two proposals for shot time allocation on OMEGA and NIF have been submitted. The OMEGA proposal has been approved for three shot days as a part of the Laboratory Basic Science solicitation. The NIF proposal is currently under review. Design work is currently under way to develop a shock-ignition point design target for the National Ignition Facility. The target has been designed by the SIcollaboration above and includes all-dt and CH-DT target designs. The SI point design uses the polar drive configuration of the NIF for a ~600kJ laser pulse. The implosions are simulated in 1D and 2D by the different groups and the results are compared. Preliminary results indicate that the targets survive the detrimental effects of inner surface roughness yielding an energy gain of about 60. The first meeting of the SI collaboration took place on April 27, 2010 at the University of Rochester. Several invited talks have been given by the FSC group on shock ignition (R. Betti, TOFE conference 2008; R. Betti, HEDSA meeting 2008; R. Betti, ICF Symposium at the European Physical Society Meeting, 2010). Highlights on Fast Ignition Integrated experiments on OMEGA. Integrated fast-ignition core-heating experiments have been carried out on the OMEGA Laser. This is the first campaign of fast-ignition integrated experiments since the original experiments of Kodama et al. at the Osaka University. Plastic (CD) shell targets with a reentrant cone are compressed with a ~20 kj UV low-adiabat laser pulse. Fast electrons generated by a 1 kj, 10 ps pulse are injected through the hollow cone into the compressed core demonstrating a significant enhancement of the neutron yield. About additional neutrons (Fig. 1) were generated by fast electron heating in a narrow time window of less than 100 ps at a 2

4 timing that is close to the predicted peak areal density from simulations. Shock breakout measurements have been performed with the integrated targets confirming an intact cone tip at the time of peak neutron production and experimentally validating for the first time that the neutron yield is due to fast electron coupling. The major advancement of these experiments is that they have studied fast ignition core heating in a new regime of high-energydensity plasmas and high target compression. Initial 2D DRACO/LSP simulations were performed to calculate the neutron yield for the experimental conditions. Based on the simulation and a simple model, it is estimated that (10 ± 5)% of the short-pulse laser energy is coupled into the target. LSP Figure 1: Measured neutron yield as a function of the arrival time of the short pulse laser. The dashed gray curve is a fit of a Gaussian profile to the square data. The solid lines mark the yield without a heating beam for a standard CD shell (black) and a CD shell lined with a 1 μm CH layer (green). simulations of the fast electron transport through the cone wall indicate that ~half of the electron energy is lost in the cone for the current laser intensities. Simulations at higher short-pulse energies and higher laser intensities predict an improved coupling efficiency of fastelectron energy into the imploded core. Higher laser intensities will be available in the near future that will allow testing these predictions. Integrated experiments provide fastignition-relevant plasma conditions to study fast-electron transport and are essential for code benchmarking. Details of this work are described in Section 8 and in a paper submitted for publication in Phys. Rev. Lett. An invited talk on this subject is scheduled for the 2010 American Physics Society Meeting. Electron transport experiments on Titan. Due to the limited capability of current laser facilities, cone tip electron-transport properties are mostly studied using cold solid targets, either a solid wire or a planar foil attached to the cone tip. To examine the transport properties through the cold Au into dense plasmas, especially the kinetic effects envisaged to dominate from the particle in cell code simulations, an experiment has been performed recently using the Titan lasers at the Lawrence Livermore National Laboratory, where a high-energy nanosecond laser pulse was used to create a warm dense plasma by shock compression and heating of low density Carbonized Resorcinol Formaldehyde (CRF) foams. A sub-picosecond high-intensity laser was irradiated on the Au foil to produce hot electrons. In the transport study, 4 different timing delay, i.e., 3, 8.5, 10 and 13 ns, were used to examine fast electron transport properties in cold, low density foam (3 ns delay case), between fully shocked and maximum compression ( ns) and after shock breaking out at the Au foil side (13 ns case). Short pulse only shots (without the long pulse driver) were also taken for comparison. Additional comparison 3

5 with cold CH (15 μm thick, solid density), as the transport medium between the Au and Cu layers in a non-refluxing condition, was also performed. In this case, the long pulse (1ns, 100 J, 600 μm spot size, 1ns delay between the short and long pulse) was used to produce a large, long scale-length preformed plasma at the back of the target as the getlost-plasma for fast electrons in order to effectively minimize the refluxing. One distinct consistent feature in these experiments was reproducible observation of a large extended Cu Kα-emission spot from the side-on Kα-imager using the foam package targets at the late time delay. From the width of the Kα-spot and the separation distance, it is estimated that fast electrons have a large angular spread (>90 degree) when traversing through the Au foil into the warm dense plasmas. Such wide spread of fast electrons have been observed in the integrated PIC simulations of electron transport through the Au cone tip into low density plasma targets and it has been attributed to the large magnetic fields generated at the interface. Reduction of the fast electrons transport through the cold low density foam is also observed in the measured vacuum electron spectra as shown in Fig. 5. With the increase in time delay, i.e., from low density foam to dense plasmas with same mass density, more fast electron with energies in a few MeV range can traverse through the foam plasma and are registered on the detector which was 37 degree from the rear target normal direction. More details are provided in Sections 9A, 9B and 9D. Electron generation experiments on Titan. The spatial divergence of hot electrons is a critical factor in the ignition of a compressed core. Years of experiments on flat foils have determined the FWHM divergence to be ~40. Conversion-efficiency measurements in cone geometry use metallic wires attached to cones and extract forward-going electrons for analysis; they collect electrons carrying ~20% of the laser pulse energy, but the cone and wire outside surfaces redirect electrons so that their initial angular divergence cannot be determined. Targets have been developed without these constraints, called buried cones. A wire at the cone tip intercepts just the electrons created in the solid angle of the wire. By replacing the wire with a thin copper foil, likewise buried in aluminum, at some depth from the cone tip, allows measurements of their angular distribution through imaging the Cu-Kα fluorescence caused by passing electrons. A thick conducting block glued behind the Cu prevented electrons from passing through the Cu foil multiple times. Initial results from a Titan campaign suggest that electron divergence is increased relative to flat foils when the conical cavity has a 30µm diameter tip, but is unaffected for tip diameter ~90 µm similar to flat target shots from previous experiments at RAL. More details on these experiments are provided in Section 9H. Electron transport experiments on OMEGA EP. Fast electron transport is studied through a set of experiments carried out on the OMEGA EP laser facility. A copper wire is attached to the tip of a hollow gold cone to investigate the characteristics of the fast electrons through the tip of the cone after they are generated by the OMEGA EP 10-ps pulse. The short pulse (interaction pulse) was focused into the cone/wire target. Here, the wire (40-μm diameter, 1 mm long) was made of copper and attached to the tip of a gold cone with a 20-μm-thick sidewall capped with 6-μm-thick, 30-μm-inner-diam foil. The Cu-Kα x-ray emission from the wire was diagnosed with a HOPG spectrometer at the normal direction to the wire axis. The energy spectra of the fast electrons were measured 4

6 along the wire direction, i.e., on the interaction laser axis. In addition, the second short pulse (backlighter pulse) was used to generate a high-energy proton beam to measure the electrostatic field around the cone/wire target using a proton deflectometry technique. The Cu-Kα signal was observed on the HOPG spectrometer with a signal-to-background contrast of up to 1.4. The Cu-Kα signal was found to be linearly dependent on the interaction pulse energy. We have paid particular attention to the laser contrast to explain the difference of Kα yield efficiency. The energy contained in the prepulse of the Titan laser was up to ~10 mj. The energy in the amplified stimulated emission (ASE) of the interaction pulse in the OMEGA EP experiment was estimated to be up to 300 mj. This can create a large preplasma, which can reduce the fast electron coupling to the wire. We have found that the increase of prepulse energy reduces the electron coupling (i.e. Kα yield) as also shown in a Titan experiment using 0.7 ps interaction pulse. The OMEGA EP results (10 ps pulse) are close to the Titan data when the Kα yields are measured as a function of prepulse energy. Therefore, the lower yield on OMEGA EP seems to be caused by the preplasma due to the ASE. These results and others have been presented in an invited talk at the 2009 American Physical Society meeting. More details on these experiments are provided in Section 9C. Hybrid simulations of fast-electron transport. Three-dimensional (3D) hybrid-pic simulations have been carried out to study fast electron-transport experiments recently performed at the Laboratory for Laser Energetics. In this work, the effects of the target resistivity, including different target ionization levels, have been studied in detail. The simulations confirm that electron transport can be explained by partial collimation of hot electrons by resistive magnetic fields. The initial divergence half-angle of the hot electrons in the target is found to be close to the value predicted by the ponderomotive scaling. The experiments were conducted on the Multi-Terawatt (MTW) Laser. A laser pulse of wavelength λ L =1.053 μm, with an energy of ~5 J and a duration of Δt L ~650 fs, was focused with normal incidence to a 4-μm-radius spot, producing an intensity of ~10 19 W/cm 2. The Al, Cu, Sn, and Au foil targets had transverse dimensions of 500 μm and thicknesses ranging from 5 to 100 μm. A coherent transition radiation (CTR) diagnostic was fielded to acquire images of the rear-side optical emission with a spatial resolution of ~1.4 μm. The three-dimensional hybrid-pic code LSP has been used to model the transport of hot electrons in solid targets. The collisional model in LSP has been recently modified to include relativistic effects for hot electrons. Separate species for hot electrons in different energy ranges are used to insure the correct scattering and slowing-down rates for electrons at different energy levels. The collisional model was tested to reproduce the correct ranges, blooming, and straggling of hot electrons, the correct resistive electric and magnetic fields, and to conserve energy. These simulations use the Thomas-Fermi ionization model and equation of state to calculate changes in the ionization state and specific heat capacity of the background electrons with the target temperature. Figure 2(a) shows cross sections through the azimuthal magnetic field 350 fs after the peak of the laser pulse. Figure 2(b) shows the location of the hot-electrondensity isosurface (red-solid) at 50% of the peak density in each transverse plane at the same moment of time; the blue semi-transparent surface corresponds to the case with the magnetic field artificially suppressed. 5

7 (a) (b) FIG. 2 Cross sections through the azimuthal magnetic field (in units of MG) for the 60-μm target, 350 fs after the laser-pulse peak. (b) The fast-electron density isosurface at 50 % of the peak density in each (x, y) plane (redsolid) and the equivalent isosurface with the magnetic field artificially suppressed (blue-transparent). The hot-electron beam is partially collimated by the resistive magnetic field generated at the outer edge of the electron beam. This field is most intense in the first 20 μm in the 0 target. The magnetic field reduces the initial beam divergence half-angle of θ1/ (averaged within the FWHM of the beam spatial and temporal distribution) to θ1/ 2 16, close to the beam divergence half-angle in the experiment. The variation of the beam density distribution with the propagation distance resembles an expanding annulus which breaks into filaments due to the resistive filamentation instability. These results were presented at the 2009 IFSA conference and described in a paper published in Physical Review Letters in More details are provided in Section 9J. PIC simulations of fast electron transport. It is well known that the system of the beam plus the plasma return current it induces is unstable to Weibel/filament (WF) and twostream (TS) instabilities, and in general, oblique modes. However, there were conflict reports on the long-term evolution of the system. Previous studies showed creation of a single current filament in the laser-generated plasma channel. Other studies did not find the merge into the single current filament. The cause of this difference was not known before our study. We have studied the nonlinear evolution of the beam-plasma system due to the interaction of the WF and TS instabilities. We have found that the presence of TS prevents the current filaments due to WF from eventually merging into a single filament with a cross section much smaller than that of the initial beam. As the beam temperature increases, the dominant mode changes from the Weibel/filament type to the two-stream type. The nonlinear evolution can be understood as competition between the Weibel/filament instability and the two-stream instability as the beam energy is converted to the beam and plasma thermal energy. The comparison of the in- and out-of-plane simulations shows the importance of the two-stream instability, which dominates the energy transfer early on and is stabilized before the Weibel/filament instability. The end state of the system is a beam with a moderately increased angular spread, reduced drift energy and no reduction of the initial cross section. We have also studied how collisions affect the beam transport. We found that collisions can reduce the two-stream modes while enhance the Weibel/filament modes (Hao et al Physical Review E 2009). Combined with the results in Kong et al (Physics of Plasmas 2009), they indicate that in the dense region, the magnetic fields generated from the Webel/filament modes may collimate the beam. This work has been published in Physics of Plasmas and Physical Review E. More details are provided in Section 9N. 6

8 Highlights on Electromagnetic Fields Measurements in High Energy-Density Plasmas In 2009, the FSC carried out several experiments devoted to probing electromagnetic fields in laser-produced plasmas. Several of them are described in the following subsections. Many utilized our recently developed technology for monoenergetic proton radiography, in which a glass shell backlighter capsule filled with D 3 He fuel is imploded with a small subset of OMEGA drive beams, resulting in isotropic emission of 15-MeV D- 3 He protons, 3-MeV DD protons, and 3.6-MeV D- 3 He alpha particles. Others involved the development and deployment of diagnostic instruments at the National Ignition Facility for their preliminary shots in Magnetic fields generated by laser-foil interactions. A pulse of 14.9-MeV protons from the backlighter were used to image two identical, expanding plasma bubbles, formed on opposite sides of a 5-µm-thick plastic (CH) foil by two 1-ns-long laser interaction beams. Both beams had spot diameters of 850 µm and intensities of W/cm 2 ; they were fired simultaneously and incident at 23.5º from the normal to the foil [Fig. 3(a)]. To break the nearly-isotropic proton fluence into beamlets (~1000 protons each) whose deflections could easily be observed and quantified, 150-µm-period nickel meshes were placed on opposite sides of the foil. Figure 3(b) is the resulting radiograph, with laser timing adjusted so the bubbles were recorded 1.36 ns after the onset of the interaction beams. (a) Monoenergetic proton radiography setup (b) Radiograph (c) Deflection & field map (at foil) Backlighter Front CH foil mesh Back bubble B Beam (B2) 1.35 cm Beam (B1) 27 cm Back mesh Protons CR-39 detector 10 cm at detector 0.5 cm at foil B 140 B dl (MG-µm) x 10 4 Proton fluence at detector (cm -2 ) 0 Ө (degrees) FIG. 3. FSC-supported MIT proton radiography setup (a), proton radiograph of two laser-generated plasma bubbles (b), and spatial map of proton beamlet deflection angle as a function of position on the foil (c). It will be seen in Fig. 2(b) that the deflections are associated almost exclusively with a B field near the foil, and this means that (c) can also be viewed as a magnetic field map. Part (c) shows that the two bubbles were actually the same size even though the apparent sizes are different in the radiograph. The orientation of the images is as seen from behind the detector, looking toward the backlighter. The radiograph was acquired during OMEGA shot The top bubble image in Fig. 3(b) is a type of image we have recently begun studying and contrasting to predictions of the 2D radiation- hydrodynamic code LASNEX. The simulations indicated that proton deflections are purely a result of a toroidal B, parallel to the foil, arising from the n magnetic-field source term (where n e and T e are the e T e electron number density and temperature). These results were presented in invited talks at 7

9 the 2009 International Conference of High Energy Density Physics and at 2009 Inertial Fusion Science Conference. More details can be found in Section 10 and in papers published in Physical Review Letters and Physics of Plasmas. Charged-particle probing of x-ray-driven implosions. Measurements of x-ray driven implosions with charged particles have resulted in the quantitative characterization of critical aspects of indirect-drive inertial fusion, as described in [C.K. Li et al., Science (2010)].Three types of spontaneous electric fields differing in strength by two orders of magnitude, the largest being nearly one-tenth of the Bohr field, were discovered with time-gated proton radiographic imaging and spectrally-resolved proton selfemission. The views of the spatial structure and temporal evolution of both the laser drive in a hohlraum and implosion properties provide essential insight into, and modeling validation of, x-ray driven implosions. We performed experiments using monoenergetic proton radiography and charged-particle spectroscopy to study the x-ray drive and capsule implosions in gold (Au) hohlraums. These measurements have allowed a number of important phenomena to be observed. In particular, three types of spontaneous electric (E) fields, differing by two orders of magnitude in strength with the largest approaching the Bohr field (= ea 0-2 ~ V m -1, where a 0 is the Bohr radius), were observed. The experiments also demonstrate the absence of stochastic filamentary pattern and striations generally found in laser-driven implosions. We also observed plasma flow, supersonic jet formation, and self-generated magnetic (B) fields, determined the areal density (ρr) and implosion symmetry; and sampled different implosion phases. The experiments were performed at the OMEGA laser facility. In the backlighting experiment the hohlraum had a 2.4-mm diameter, 3.8-mm length, and 100% laser-entrance holes (LEH) with 25-µm-thick Au walls over-coated on the inside with 0.3 µm parylene (CH) liner. The hohlraum was driven by 30 laser beams with wavelength µm and total laser energy ~ 11 kj in a 1- ns square pulse. The laser beams had full spatial and temporal smoothing (15), and phase plates SG4 were used. The capsule was a 30-µm-thick plastic shell of diameter 550 µm filled with 50 atm H 2 or D 2 gas. The designed radiation temperatures (~ 150 ev) and consequent capsule compression ( 10) were low. The backlighter was a D 3 He-gas-filled, glass-shell capsule with a 420-µm diameter and a 2-µm shell thickness, imploded by 30 laser beams. Two types of fusion protons with discrete birth energies of 14.7 and 3.0 MeV are produced in nuclear fusion reactions (D+ 3 He α+p and D+D T+p) ~ 80 ps. The images were recorded by CR-39 track detectors and the timing of the proton sampling was adjustable. Radiographs made by 15-MeV D 3 He protons covering a typical ICF implosion sequence contain both spatial and energy information because the CR-39 detector records the position and energy of each individual proton. The images show proton fluence versus position, to provide time-dependent information about field distributions, capsule compression, and hohlraum plasma conditions. 8

10 TABLE OF CONTENTS EXECUTIVE SUMMARY FSC MEMBERS AND COLLABORATORS FUSION SCIENCE CENTER MEETING THE FSC EDUCTION MISSION Summer School in High Energy Density Physics REFEREED PUBLICATIONS INVITED PRESENTATIONS AT CONFERENCES SHOCK-IGNITION LASER-PLASMA INTERACTION EXPERIMENTS FAST-IGNITION INTEGRATED EXPERIMENTS ON OMEGA ELECTRON GENERATION AND TRANSPORT IN FAST IGNITION A. Electron transport through warm dense matter B. Effects of the get-lost-plasmas on the observed Kα yield C. Electron transport experiments on OMEGA EP 34 D. Modeling of electron transport and Kα production in multilayer targets E. In-situ laser intensity monitor F. Escaping electrons simulations G. Specular reflection simulations H. Electron beam divergence in cones I. LSP+DRACO simulations of LLE integrated fast ignition experiments J. Role of magnetic collimation in the electron transport in solid targets K. Superthermal and efficient heating modes in cone targets L. Hot electron generation forming a steep gradient M.Transport of MA electron currents in ultra-fast heated metal targets N. PIC simulations of core-heating in fast ignition MAGNETIZED HIGH ENERGY-DENSITY PHYSICS AND NUCLEAR DIAGNOSTICS A. Probing ICF experiments and laser-plasma interactions with protons A1. Study of B-fields generated by laser-foil interactions A2. Charge-particle probing of x-ray driven implosions B. Nuclear diagnostic measurements at the NIF B1. Proton spectra...72 B2. Neutron yield measurements

11 2. THE FUSION SCIENCE CENTER MEMBERS, COLLABORATORS, ADVISORS and ADMINISTRATORS Members -Faculty and Research Scientists F. Beg (UCSD) co-pi R. Betti (UR-LLE) Director D. Correll (LLNL) co-pi T. Ditmire (UT) co-pi R. Freeman (OSU) co-pi M. Key (LLNL) collaborator C. Li (MIT) collaborator D. Meyerhofer (UR-LLE) co-pi W. Mori (UCLA) co-pi R. Petrasso (MIT) Deputy Director P. Patel (LLNL) co-pi C. Ren (UR-LLE) co-pi Y. Sentoku (UNR) co-pi R. Stephens (GA) co-pi L. Van Woerkom (OSU) co-pi Members - Post docs and Res. Assoc. K. Anderson (UR) B. Chrisman (UCSD) E. Chowdhury (OSU) V. Decyk (UCLA) O. Gotchev (UR) M. Hohemberger (UR) P. Nilson (UR) D. Schumacher (OSU) A. Solodov (UR) J. Tonge (UCLA) T. Yabuuchi (UCSD) M. Wei (UCSD) Graduate Students D. Casey (MIT) C. Chen (MIT) P. Chang (UR) F. Fiuza (UCLA) A. Kryger(OSU) F. Leblanc (UNR) G. Li (UR) A. Link (OSU) T. Ma (UCSD) M. Manuel (MIT) J. May (UCLA) R. Nora (UR) V. Ovchinnikov(OSU) B. Pradhkar (UCSD) H. Rinderknecht (MIT) M. Rosenberg (MIT) N. Sinenian (MIT) T. Wang (UCLA) R. Weber (OSU) R. Yan (UR) A. Zylstra (MIT) Administrators M. Kyle (UR-LLE) J. Morris (UR) Collaborators - Research Scientists J. Delettrez (UR-LLE) G. Fiksel (UR-LLE) J. Knauer (UR-LLE) J. Myatt (UR-LLE) C. Sangster (UR-LLE) F. Seguin (MIT) C. Stoeckl (UR-LLE) W. Theobald (UR-LLE) External Advisors J. Kilkenny (GA) J. Sheffield (U. Tennessee) M. Tabak (LLNL) R. Town (LLNL) Member is a Researcher (post doc or scientist) partially or fully supported by the FSC, or a Principal Investigator. Graduate students are fully or partially funded by the FSC, or closely collaborating with the FSC. Collaborators are not currently supported by the FSC.

12 3. FUSION SCIENCE CENTER MEETINGS 7 th Fusion Science Center Meeting, May (2008), Rochester NY 8 th Fusion Science Center Meeting, August 2 (2009), Los Angeles CA (FSC meeting presentations are available on the FSC web site at fsc.lle.rochester.edu) 4. THE FSC EDUCATION MISSION The 3rd FSC Summer School in High Energy Density Physics July 26-August 1, 2009 UCLA Campus, Los Angeles, CA Scheduled lecturers and subjects F. Beg - UCSD Physics of Z-Pinches R. Betti UR, Implosion Hydrodynamics P. Drake - UM Introduction to HEDP D. Hinkel - LLNL Laser-Plasma Intensity M. Key - LLNL Fast Ignition D. Meyerhofer UR, HEDP Diagnostics W. Mori UCLA, Laser-Plasma Interaction C. Ren -, UR The Particle-in-Cell Method M. Rosen - LLNL, Inertial Confinement Fusion L. Van Woerkom OSU, Fast-Electron Transport Panel Discussion Members J. Porter, Sandia National Laboratory J. Kilkenny, General Atomics F. Beg, University of California San Diego J. Workman, Los Alamos National Laboratory A. Wan, Lawrence Livermore National Laboratory R. Stephens General Atomics 6/13/

13 The UC Los Angeles hosted the FSC HEDP Summer School 93 attendees 48 financial aid packages from the FSC 62 graduate students 5 undergraduates 2 post docs 24 research scientists Sponsors: FSC, ILSA and GA 6/13/

14 5. REFEREED PUBLICATIONS (Published or accepted for publication in April ) 1. Multidimensional measurable ignition criterion for inertial confinement fusion Chang PY, Betti R, Spears B et al. PHYSICAL REVIEW LETTERS Volume: 104, Article Number: Published: APR Charged particle probing of X-ray-driven inertial fusion implosions Li CK, Seguin FH, Frenje JA et al. SCIENCE Volume: 327 p Published: MAR Advanced-ignition-concept exploration on OMEGA Theobald W, Anderson KS, Betti R, et al. PLASMA PHYSICS AND CONTROLLED FUSION Volume: 51 Issue: 12 Article Number: Published: DEC Laser driven magnetic-flux compression in high energy density plasmas Gotchev OV, Chang PY, Knauer JP et al PHYSICAL REVIEW LETTERS Volume: 103 Article Number: Published: NOV Radiation hydrodynamic theory of double ablation fronts in direct-drive inertial confinement fusion Sanz J, Betti R, Smalyuk VA, et al. PHYSICS OF PLASMAS Volume: 16 Issue: 8 Article Number: Published: AUG Shock Ignition: A New Approach to High Gain Inertial Confinement Fusion on the National Ignition Facility Perkins LJ, Betti R, La Fortune KN, et al. PHYSICAL REVIEW LETTERS Volume: 103 Issue: 4 Article Number: Published: JUL Lorentz Mapping of Magnetic Fields in Hot Dense Plasmas Petrasso RD, Li CK, Seguin FH et al. PHYSICAL REVIEW LETTERS Volume: 103 Issue: 8 Article Number: Published: AUG High-Current, Relativistic Electron-Beam Transport in Metals and the Role of Magnetic Collimation Storm M, Solodov AA, et al. PHYSICAL REVIEW LETTERS Volume: 102 Issue: 23 Article Number: Published: JUN Pressure-driven, resistive magnetohydrodynamic interchange instabilities in laser-produced high-energy-density plasmas Li CK, Seguin FH, Frenje JA, Petrasso RD, et al. PHYSICAL REVIEW E Volume: 80 Issue: 1 Article Number: Part: Part 2 Published: JUL Integrated simulations of implosion, electron transport, and heating for direct-drive fast-ignition targets Solodov AA, Anderson KS, R. Betti et al. PHYSICS OF PLASMAS Volume: 16 Issue: 5 Article Number: Published: MAY Observations of Electromagnetic Fields and Plasma Flow in Hohlraums with Proton Radiography Li CK, Seguin FH, Frenje JA, Petrasso RD, Seguin FH, et al. PHYSICAL REVIEW LETTERS Volume: 102, Issue: 20 Article Number: 20500, Published: MAY Proton radiography of electric and magnetic fields in laser-produced plasmas Li CK, Seguin FH, Frenje JA et al. PHYSICS OF PLASMAS Volume: 16 Article Number: Published: Hot electron generation forming a steep interface in superintense laser-matter interaction Mishra R, Sentoku Y, Kemp AJ, PHYSICS OF PLASMAS Volume: 16, Article Number: Published: Intense laser-plasma interaction: new frontiers in high energy density physics Norreys PA, Beg FN, Sentoku Y et al, PHYSICS OF PLASMAS Volume: 16, Article Number: Published: /13/

15 15. Study of the transpot of high-intensity laser-generated hot electrons in cone-coupled wire targets King J, Akli KU, Freeman RR et al, PHYSICS OF PLASMAS Volume: 16, Article Number: Published: X ray spectroscopu of buried layer foils irradiated at laser intenisties in excess of 10^20 Wcm^-2 Chen SN et al, PHYSICS OF PLASMAS Volume: 16, Article Number: Published: Transport of energy by ultraintense laser-generated electric fields tudy of the transpot of high-intensity laser-generated hot electrons in cone-coupled wire targets Ma T, Key M, Mason RJ et al, PHYSICS OF PLASMAS Volume: 16, Article Number: Published: Energy injection for fast ignition Stephens R et al, JOURNAL OF PLASMA AND FUSION RESEARCH Volume: 4, p.16, Published: A simulation study of fast ignition with ultrahigh intensity lasers Tonge J, May J, Mori W et al, PHYSICS OF PLASMAS Volume: 16, Article Number: , Published: Evolution of a relativistic electron beam-plasma return current system Kong X, Park J, Ren C et al, PHYSICS OF PLASMAS Volume: 16, Article Number: , Published: INVITED PRESENTATIONS AT CONFERENCES R. Betti The physics of ICF ignition Workshop on Burning Plasmas, Livermore, CA, January 27, 2010 R. Betti ICF Ignition, the Lawson Criterion and Comparison with MCF Ignition American Physical Society Division of Plasma Physics Meeting, Atlanta, GA, November 2-6, 2009 R. Betti A Measurable Lawson Criterion and Hydro-equivalent Ignition for Inertial Confinement Fusion Teller Award Lecture: Inertial Fusion Science and Applications. San Francisco, CA, September 7-11, 2009 J.P. Knauer Compressing magnetic fields with high energy lasers American Physical Society Division of Plasma Physics Meeting, Atlanta, GA, November 2-6, 2009 F. Beg Generation and transport of hot electrons in cone-wire targest American Physical Society Division of Plasma Physics Meeting, Atlanta, GA, November 2-6, 2009 R. Stephens Anomalous divergence of laser-produced hot electrons generated in a cone geometry Inertial Fusion Science and Applications. San Francisco, CA, September 7-11, 2009 P. Patel Design of Fast Ignition Core Heating Experiments with Full Scale Hydrodynamic Fuel Assembly Inertial Fusion Science and Applications. San Francisco, CA, September 7-11, 2009 P. Nilson Bulk heating of solid-density matter using kj-pulses on OMEGA EP Atomic Process in Plasmas (APiP). Monterey Bay, CA, March, 2009 Y. Sentoku Hot electron generation forming a steep interface in super intense laser-matter interaction Inertial Fusion Science and Applications. San Francisco, CA, September 7-11, /13/

16 Y. Sentoku Modeling of high energy density plasmas produced by ultra intense laser light Short-pulse laser-plasma workshop. Pleasenton, CA, September 14, 2009 CK Li Electromagnetic fields and plasma flow in laser-irradiated hohlraums Inertial Fusion Science and Applications. San Francisco, CA, September 7-11, 2009 CK Li Spontaneously-generated, dynamic electric and magnetic field in laser-produced plasmas 8th Pacific Rim Conference on Laser and Electro-Optics, Shanghai, China, August 30- September 3, 2009 CK Li Pressure-driven, resistive MHD interchange instabilities in laser-produced plasmas Anomalous Absorption Conference. Bodega Bay, CA, June 14-19, 2009 R. Petrasso Monoenergetic proton radiography of HED plasmas on OMEGA, OMEGA-EP and the NIF International Conference on High Energy Density Physics. Austin, TX, May 19-22, 2009 R. Betti Compressing magnetic fields with high energy lasers International Conference on High Energy Density Physics. Austin, TX, May 19-22, 6/13/

17 7. SHOCK-IGNITION LASER-PLASMA-INTERACTION EXPERIMENTS (UR) Shock ignition (SI) is a two-step inertial confinement fusion (ICF) concept in which a strong shock wave is launched at the end of the laser pulse to ignite the compressed core of a lowvelocity implosion [1]. Two-step processes separate fuel assembly and ignition, relaxing driver requirements and promising high gains [1,2 5]. Another advanced ignition concept is fast ignition [6], which relies on a high-intensity short-pulse laser generating an energetic beam of particles to trigger ignition. SI relies on highly shaped laser pulses that are within the pulseshaping capabilities of currently operating laser systems like the National Ignition Facility (NIF) [7]. Proof-of-principle experiments can be carried out on the NIF [8]. Recent twodimensional simulations [4] have described SI designs with as low as 250 kj of total laser energy. This shows a promising route to high fusion gains at moderate laser energies. SI is currently considered to be an option in the European direct-drive HiPER project [4]. An intensity spike (~ to ~10 16 W/cm 2 ) at the end of the laser pulse launches the ignitor shock and the final fuel assembly develops a centrally peaked pressure profile lowering the ignition threshold compared to standard isobaric assemblies. Laser E17117aJ1 Critical surface Hot e Hot e Target Hot e Laser Laser ~ W/cm 2 FIG. 1. Schematic of the setup for studying laser plasma interactions and preheating at high laser intensities relevant to shock ignition. Forty of the OMEGA laser beams implode the capsule at low intensities. Twenty delayed beams are tightly focused onto the critical density surface, where plasma instabilities lead to the generation of energetic electrons. Parametric plasma instabilities [9] such as stimulated Brillouin scattering (SBS), stimulated Raman scattering (SRS), and two-plasmon-decay (TPD) instability are of concern in an ignition target design. The instabilities increase the back-reflection of the laser light from the target, degrading the laser energy coupling to the capsule. They increase the fraction of the laser energy transferred to suprathermal electrons, a potential source of preheat that reduces the final core compression. In contrast to conventional hot-spot ignition, low-energy hot electrons generated during the power spike with a pulse duration of several hundred picoseconds may have a positive effect on SI. The areal density increases rapidly during the final stages of the implosion. If the range of the hot electrons generated during the intensity spike is less than the shell thickness, they are stopped in the shell and augment the strength of the shock wave. Significant laser plasma instabilities can be expected during the final high-intensity ignitor spike. Whether or not those are harmful depends on the hot-electron temperature and the amount of laser energy that is transferred into hot electrons at high laser intensities. The effect of hot electrons on a NIF-scale SI target [10] was modeled in 1-D for a marginal igniting target using a multigroup diffusion model [11] for the hot electrons. The ignition (time) window for shocklaunching is considerably wider when the effects of moderate-energy hot electrons (a NIFscale target can efficiently stop up to 150-keV electrons) are included, showing that hot electrons can indeed be beneficial for the SI scheme as long as their range is shorter than the shell s thickness. This work provides the first measurements of parametric instability and preheat for conditions relevant for SI (spherical target, long density scale length, and intensities above 2 6/13/

18 10 15 W/cm 2 ). Important physics issues including the hot-electron energy content, the hotelectron temperature, and laser backscattering for various intensities and time delays between fuel assembly and shock generation are studied. Previous SI experiments [12] on OMEGA [13] studied fuel assembly with 60-beam symmetric implosions. The shock wave was launched by a spike at the end of the pulse on all 60 beams with a maximum intensity of ~ W/cm 2. The intensity was too low for studying laser-plasma instabilities for SI, which requires ~ W/cm 2. Switching from a 60-beam to a 40- plus 20-beam configuration on OMEGA using two separate pulse-shaping systems makes it possible to use tightly focused beams that generate a stronger shock wave than in the previous experiments. The implosion nonuniformity reduced target performance but makes it possible to study high-intensity coupling. In a NIF demonstration, the high-intensity pulse will be generated on all beams, preserving symmetry. OMEGA cannot produce the requisite SI intensity with a compression pulse using 60 beams. The main objective of this work is to study the coupling of laser energy from high-intensity beams into a spherically imploding capsule and the laser plasma instabilities at SI relevant laser intensities. Beam power (TW) Hard x-ray (pc) E17962J (a) (c) Variable delay 1 2 Time (ns) Delay (ns) mm 0.77 mm 0.3 mm Areal density (mg/cm 2 ) FIG. 2. (a) Drive-pulse shape (solid) and high-intensity pulse (dashed), [(b) (d)] measured neutron yield, hard x-ray yield, and neutron-rate averaged areal density, respectively. The different symbols represent different focus positions with respect to the shell center. The solid line in (b) is the average yield for 40- beam implosions and the dashed lines represent the error range. The 40-beam implosions produced no measurable HXR signal and fusion yields were too low N-yield ( 10 9 ) (b) (d) Delay (ns) Delay (ns) 0.37 mm 0.77 mm 0.3 mm 40 beam 0.37 mm 0.77 mm The targets were 36- µ m-thick, 430-µ m-outerradius, deuterated plastic shells coated outside with a 0.1-µm layer of aluminum and filled with D 2 gas with a pressure of ~30 atm. Figure 1 shows a schematic of the experiment. The capsules were imploded using 40 of the 60 OMEGA beams with a low-adiabat ( ~ 1.5) pulse shape [14] with ~13.6 kj of UV laser energy. The adiabat α is defined as the ratio of the plasma pressure to the Fermi pressure of a degenerate electron gas [14]. The solid curve in Fig. 2(a) shows the drive pulse shape comprising an ~100-ps full width at half maximum (FWHM) Gaussian picket pulse preceding a shaped main-drive portion with a total duration of 2.7 ns. The 351-nm-wavelength laser light of the 40 beams was smoothed with polarization smoothing [15] and distributed phase plates [16]. A late shock was driven by the remaining 20 beams that were delayed and tightly focused to achieve intensities at the critical surface ranging from ~ to ~ W/cm 2. The 20 delayed beams (~4.6 kj) used an ~600-ps FWHM square pulse shape (dashed curve) without polarization smoothing or phase plates. 6/13/

19 The experimental observables were the neutron yield [17], the backscattered laser energy [18], the hard x-ray signal [19], and the neutron-rate averaged areal density [20]. The laser light reflected back from the imploded capsule was measured from two beam ports [a shock-beam port (#25) and a drive-beam port (#30)], which were equipped with a full-aperture backscatter station [18]. Time-resolved spectra were recorded by two streaked spectrometers covering the wavelength ranges of 351±3 nm for SBS and 450 to 700 nm for SRS. The total backscattered energy in either of these spectral ranges was measured by calorimeters with an uncertainty of ±10%. The hard x-ray (HXR) signals (with photon energies >20 kev) were measured by the HXR detector with four channels measuring x rays >20, >40, >60, and >80 kev, respectively [19]. Areal densities (ρr) were inferred from secondary proton spectra [20]. The delay time defined by the onset of the high-intensity beams with respect to the start of the drive pulse was varied from 2.3 to 2.9 ns. The effect on neutron and HXR yield is shown in Figs. 2(b) and 2(c) and on ρr in Fig. 2(d). The different symbols represent various focus conditions, where the number refers to the focus position in vacuum with respect to the shell s center. A negative number means that the focus is in front of the target toward the laser. The neutron yield increases by a factor of ~7 from to ~ for the shortest time delay. Two reference implosions with only the 40 drive beams produced neutron yields of and ; the solid line in Fig. 2(b) represents the average of both yields. The HXR yield s dependence on delay time is very similar. Figure 2(c) shows that signals measured by the >40- kev channel increase with shorter delay. The HXR signal provided information on the hot-electron energy and temperature. Based on a calibration [21] of the hard x-ray detector, from 5%±2% to 16%±6% of the shockbeam energy was converted into hot electrons. The conversion efficiency was highest for short delays when there was a partial overlap between the drive and shock pulses [Fig. 2(c)]. The hot-electron temperature was determined by fitting estimated values from the convolution of an exponentially decaying hard x-ray spectrum with the sensitivity of the different channels of the HXR detector to the measured four channels [22]. The inferred temperature was 42±5 kev for all the shots, independent of laser intensity. An unbalanced target illumination caused nonuniform implosions with a dominant l = 2 mode. The power imbalance was ~10.6%, given as the root-mean-square variation of the laser power on target. A typical value for a 60-beam symmetrical illumination on a spherical target is ~2% power imbalance [23]. The nonuniformity of the implosion is clearly seen in x-ray pinhole camera images, which shows a strongly perturbed core. The core distortion was reduced when the 20 delayed, tightly focused beams were coupled into the target. Figures 2(b) and 2(c) show that despite large target illumination nonuniformity, a significant amount of the high-intensity pulse energy was coupled into the capsule, producing up to ~20 more neutrons and a strong HXR signal. The correlation of increasing neutron yield with a higher HXR signal suggests that the increased yield was partially due to hot electrons coupled into the outer regions of the compressing target. The late shock appears to be driven by a combination of the standard ablative and hot-electron drives. The areal density did not change significantly with delay. The measured maximum R is 82±11 mg/cm 2, which is the average of four lines of sight, and the error bar is the standard deviation. This is ~30% lower than the expected ~115 mg/cm 2, which is scaled down from the measured 130±17 mg/cm 2 14 that was obtained with a more-uniform 60-beam implosion with the same fill pressure, the same adiabat, and an energy of 20 kj and using the scaling of ρr 6/13/

20 with the laser energy to the power 1/3. 14 The standard deviation of R for all the shots varies from 15% to 35% of the mean ρr value, showing a large fluctuation of the measured areal density over the different lines of sight. The ρr degradation is most likely due to the strongly nonuniform implosion. The plasma reflectivity and HXR production from hot electrons were measured for various laser intensities. This was achieved through an intensity scan by shifting the focus of the 20 shock-driving beams relative to the shell s center. The nominal laser intensity is quoted for the location of the critical-density plasma surface calculated by a 1-D hydrodynamic simulation [11]. The distance from the critical density to the capsule center was ~0.3 mm at 2.7 ns. For the lens position at 0.3 mm, the 20 beams were tightly focused on the critical-density location. The focus diameter of the 20 shock beams was estimated with ~80 µm, which gave a best-focus intensity up to ~ W/cm 2 for the shock beams in vacuum. The foci of the 20 shock beams did not overlap at the critical density for all lens positions used. No overlappedbeam effects [19] were expected and the HXR signal was dominated by single-beam interaction with the target. Normalized hard x-ray (arbitrary units) E17963J Laser intensity Laser intensity 50 (a) 2.3 ns (b) 2.3 ns 2.5 ns ns 2.7 ns ns 2.9 ns 2.9 ns Lens position (mm) Reflectivity (%)(SRS + SBS) Lens position (mm) FIG. 3. Measured (a) hard x-ray signal normalized to nominal beam focus area and (b) backscattered light versus lens focus position with respect to the shell s center. The laser intensity at the critical density is highest for the negative lens position corresponding to an intensity of ~ W/cm 2. The various symbols represent different time delays. Figure 3(a) shows the measured HXR signal normalized to the estimated laser focus area versus lens position. The x-ray signal and consequently the hotelectron production increased with laser intensity presumably because of a larger growth in laser plasma instabilities such as SRS and TPD, the primary sources of hot electrons [19]. Figure 3(b) shows the measured fraction of laser backscatter energy (SRS + SBS) of one shock beam (#25) versus laser intensity. It increases from ~10% at ~ W/cm 2 to ~36% at ~ W/cm 2. SBS increased moderately from ~7% to 12% with intensity, while the SRS signal increased by almost a factor of ~5 and dominates the backscattering at the highest intensity. SRS backscattered ~4%, ~13%, and ~24% of the beam energy for ~ W/cm 2, ~ W/cm 2, and ~ W/cm 2, respectively. The simultaneously measured back-reflection through a neighboring drive-beam port (#30) remained constant at the level of implosions without the 20 shock beams for all beam delays and lens positions. This shows that the light from the shock beams was scattered back in a narrow cone and did not spill over into adjacent ports. No measurable signal of the 3/2 harmonic of the laser wavelength was measured for all intensities. The half-harmonic signal decreased by more than two orders of magnitude with higher intensities. At the maximum intensity, the half-harmonic signal was below the detection threshold, indicating no significant contribution of TPD to the hot-electron production /13/

21 These experiments measured backscattering levels that are comparable to experiments with gas-filled hohlraums with laser intensities approaching W/cm 2 [24 26]. SRS fractions reaching up to ~20% were reported [25] and measured SBS were in the few percent range and in some cases up to 35%, depending on the gas filling [24]. A recent paper studied the saturated level of SBS in preformed CH plasmas with 527-nm and 352-nm wavelength laser light [27]. A nearly constant SBS reflectivity of ~10% was measured between ~ W/cm 2 and ~ W/cm 2 with 351 nm, while the measured SRS reflectivity was much lower, <0.5%. Here, a similar level of SBS was observed that slightly increased with laser intensity, but significantly higher SRS backscattering was detected that is more comparable to the indirect-drive studies. It has been shown that smoothing the intensity distribution in the focal spot with spatial, temporal, and polarization smoothing schemes can substantially reduce the backscattering [24,26]. This is attributed to a reduction of filamentation. No phase plates and SSD were used in the 20 high-intensity beams, which could explain the high levels of backscattering. The absorbed energy rather than the backscattered light is the key issue. If 36% of the laser light is backscattered and 64% is absorbed, it represents a higher absorption fraction than the prediction of collisional absorption at these intensities (~40% to 50%). Because of the highly nonuniform plasma conditions and nonuniform illumination during the shock spike, the measurement of the scattered light through a few lines of sight cannot be used to infer the total absorbed fraction. In the pessimistic case where the predicted absorbed energy is reduced by the backscattered fraction, this can be remedied by an increase in spike power. In conclusion, shock-ignition laser plasma experiments in spherical geometry have been performed with nominal laser intensities of up to ~ W/cm 2. Forty of the sixty OMEGA beams compressed warm plastic shells filled with D 2 gas and the remaining twenty beams were delayed and tightly focused on the imploding capsule. The 20 high-intensity beams enhanced the neutron yields by up to ~20, indicating a good coupling of the shockbeam energy to the core. A significant amount (up to 36%) of laser energy from the highintensity beams was backscattered. SRS dominated the backscattering with some contribution from SBS but no significant contribution from TPD. About 10% of the high-intensity beam energy was converted into hot electrons. A hot-electron-energy distribution was generated with a temperature of 42±5 kev, independent of laser intensity. This is beneficial for shock ignition since these electrons are stopped in a thin outer layer of the imploding target, augmenting the strong hydrodynamic shock. This work was supported by the U.S. DOE under DE-FC52-08NA28302, DE-FC02-04ER54789, and DE-FG02-05ER /13/

22 REFERENCES 1. R. Betti et al., Phys. Rev. Lett. 98, (2007). 2. R. L. McCrory et al., Phys. Plasmas 15, (2008). 3. R. Betti and C. Zhou, Phys. Plasmas 12, (2005). 4. X. Ribeyre et al., Plasma Phys. Control. Fusion 50, (2008). 5. A. J. Schmitt et al., Fusion Sci. Technol. 56, 377 (2009). 6. M. Tabak et al., Phys. Plasmas 1, 1626 (1994) 7. E. I. Moses, J. Phys., Conf. Ser. 112, (2008). 8. L. J. Perkins et al., Phys. Rev. Lett. 103, (2009). 9. W. L. Kruer, The Physics of Laser Plasma Interactions, Frontiers in Physics, Vol. 73, edited by D. Pines (Addison- Wesley, Redwood City, CA, 1988). 10. R. Betti et al., J. Phys., Conf. Ser. 112, (2008). 11. J. Delettrez et al., Phys. Rev. A 36, 3926 (1987); M. C. Richardson et al., in Laser Interaction and Related Plasma Phenomena, edited by H. Hora and G. H. Miley (Plenum Publishing, New York, 1986), Vol. 7, p W. Theobald et al., Phys. Plasmas 15, (2008). 13. T. R. Boehly et al., Opt. Commun. 133, 495 (1997). 14. C. D. Zhou et al., Phys. Rev. Lett. 98, (2007). 15. T. R. Boehly et al., J. Appl. Phys. 85, 3444 (1999). 16. Y. Lin, T. J. Kessler, and G. N. Lawrence, Opt. Lett. 21, 1703 (1996). 17. V. Y. Glebov et al., Rev. Sci. Instrum. 72, 824 (2001). 18. W. Seka et al., Phys. Plasmas 15, (2008). 19. C. Stoeckl et al., Phys. Rev. Lett. 90, (2003). 20. F. H. Séguin et al., Rev. Sci. Instrum. 74, 975 (2003). 21. V. A. Smalyuk et al., Phys. Rev. Lett. 100, (2008). 22. C. Stoeckl et al., Rev. Sci. Instrum. 72, 1197 (2001). 23. F. J. Marshall et al., Phys. Plasmas 11, 251 (2004). 24. B. J. MacGowan et al., Phys. Plasmas 3, 2029 (1996). 25. J. C. Fernández et al., Phys. Plasmas 4, 1849 (1997). 26. J. D. Moody et al., Phys. Rev. Lett. 86, 2810 (2001). 27. S. Depierreux et al., Phys. Rev. Lett. 103, (2009). 6/13/

23 8. FAST-IGNITION INTEGRATED EXPERIMENTS ON OMEGA (UR) The cone-in-shell concept 1 is of significant current interest for fast ignition (FI) inertial confinement fusion 2,3 due to promising initial integrated experiments. 4,5 In FI the fuel assembly and ignition are separated by using a high-energy laser driver and a high-energy petawatt laser (HEPW) potentially allowing higher target gains than with conventional ICF designs. 2,3 The reentrant cone allows the HEPW beam to propagate as close as possible to the dense core, avoiding the need to channel the laser beam through a large region of plasma material. 2 The HEPW laser creates an intense, supra-thermal electron current at the critical density surface inside the hollow cone directed toward the compressed core. In an ignition design, MeV electrons need to deposit several tens of kj of energy into the compressed core. 6,7 There are many unresolved physics issues in the dynamics of the current propagation from the critical density surface through the high-z cone-wall material, and from the cone tip into the assembled fuel. Along the paths of the MeV electrons, the electron densities vary by more than four orders of magnitude and the temperatures vary from ~1 ev to several kev. 8 Integrated experiments on OMEGA and on NIF access these physics conditions that are relevant for fast ignition and will be used to benchmark code predictions that are important for the development of ignition target designs. FI integrated experiments a with reentrant cone were performed first at Osaka University with 1.2 kj (9 beams) long pulse laser energy and flat-top pulses driving the implosion and a 60 J short pulse laser. 4 A thermonuclear neutron yield of up to was observed and a follow up experiment demonstrated yields up to with ~500 J of short pulse laser energy (0.5 PW, 1 ps). 5 This paper describes initial integrated fast-ignition experiments on the OMEGA/OMEGA EP Laser Facility. 9,10 Compared to the previous experiments, a low adiabat pulse shape with more than an order of magnitude higher drive laser energy assembled a factor of ~16 higher target mass studying the FI scheme in a regime of high target compression. For the first time, shock-breakout measurements were performed for the same targets as used in the integrated experiments. Shock breakout was measured to be ~100 ps after peak neutron production. The experiments demonstrate an intact cone tip at the time when the short pulse laser produces up to additional neutrons, experimentally verifying the fast electron coupling. Simulations with a 2D hydrocode coupled to a 3D hybrid particle-in-cell code were performed to calculate the neutron yield for the experimental conditions. The targets were 40-µm-thick, ~870-µm outer diameter empty deuterated plastic (CD) shells with an inserted hollow gold cone. Details of the experimental setup and the target design are described in Ref.11. The cones had an inner full opening angle of 34, a side wall thickness of 10 µm and a flat tip portion with thicknesses of 10 and 15 µm. The diameter of the flat tip was 10 µm. In some shots, CD shells with a 1 µm inner CH layer were used to study its effect on the neutron yield. A low-adiabat laser pulse shape comprised of a short picket pulse and a shaped ~2.7-ns drive pulse with 351-nm wavelength and ~20 kj of imploded the capsule around the cone. The fuel assembly had been optimized in previous work, 12 with a predicted peak areal densities of ~0.4 g/cm 2 for an empty shell target. Secondary proton measurements with 40-µm plastic shells filled with 25-atm of D 2 gas demonstrated with the same pulse shape a close to 1D fuel assembly and a measured neutron averaged ρr of 0.15 g/cm 2 and a peak ρr of 0.26 g/cm The 1053-nm-wavelength short pulse had an energy up to 1 kj, a 10-ps duration, and was focused to a ~50-µm-diameter spot containing 80% of the laser energy. The 6/13/

24 corresponding peak laser intensity at the center of the cone is estimated to be ~ W/cm 2 assuming a Gaussian shaped pulse in space and time. Shock breakout experiments have been performed by imploding the shell without a short-pulse beam. The experiment was performed with the same target using 40 µm diameter flat tip with thicknesses between 5 and 15 µm. Two diagnostics measured the time when optical emission appeared inside the cone: a velocity interferometer system for any reflector (VISAR)i and a streaked optical pyrometer (SOP). 14 Both instruments measure the shock breakout times with temporal resolutions as low as 10 ps. 14,15 In preparation for each integrated FI experiment, the short-pulse laser is precisely timed with respect to the drive laser and pointed into the cone tip. The accuracy of pointing for the short-pulse laser is ~15 µm measured by x-ray pinhole images. 11 The timing is measured in situ with ±30-ps precision at full energy by measuring the temporally resolved the hard x-ray emission produced by the short-pulse laser interaction. Figure 1: Time-of-flight measurements of 2.45 MeV neutrons from integrated fast-ignition experiment on OMEGA. Signals for different timings of the shortpulse EP laser are shown. It was shown in Ref. 4 that the coupling efficiency (CE) of laser energy into the core through the fast electrons can be inferred from the measurement of 2.45-MeV neutrons from D 2 fusion reactions in the CD shell. The caveat of this method is that the inferred CE depends on the assumed model of the fuel assembly. The number of neutrons can be estimated with 2 ( 14) nd σ v dvdt σ is the reactivity, where n D is the number density of the deuterons, v VT 3 of the D+ D He(0.82 MeV) + n(2.45 MeV) fusion reaction, and the integration is performed over time and the volume of the target. The calculated neutron number is strongly temperature and density dependant for a given energy input by the fast electrons. Since there are no temperature and density measurements of the fuel assembly of this FI-cone target available, the inferred CE is based on simple 1D estimates (as was done in Ref. 5) and on 2D hydrodynamic fuel assembly simulations. Initial 2.45-MeV neutron measurements in these experiments were limited by a strong x-ray background that disabled the neutron diagnostic. 13 The background consists of bremsstrahlung emission generated by fast electrons streaming through the gold cone target that overwhelmed neutron diagnostics deployed with conventional plastic scintillators. Plastic scintillators have a significant afterglow from the prompt bremsstrahlung emission that masks the weak neutron signal even when the unit is appropriately shielded and uses a gated microchannel plate detector. To overcome this, a new detector was developed. The detector consists of an organic liquid scintillator that is saturated with molecular oxygen providing a 6/13/

25 fast light decay and very low afterglow. 15 The liquid scintillator detector completely suppresses the hard x-ray background at the time when the 2.45-MeV neutrons arrive and provides reliable neutron yield data. Figure 1 shows time-of-flight measurements of 2.45 MeV neutrons from the integrated fast-ignition experiment on OMEGA. The measurements are for different timings of the EP short-pulse laser with respect to the compression pulse. The signals at 3.5 and 3.55 ns are very similar to measurements without EP beam, while at 3.60 and 3.65 ns a higher neutron signal was measured. About 15 to 35 neutrons contribute to the signals in the detection solid angle. The signals were integrated from 560 to 635 ns, indicated by the two dashed lines in Figure 1, providing a total yield. The accuracy is limited by the counting statistics. The signals were cross-calibrated against another absolutely calibrated neutron detector 16 using a series of shots without short-pulse laser. Figure 2: Measured neutron yield as a function of the arrival time of the short pulse laser. The dashed gray curve is a fit of a Gaussian profile to the square data. The solid lines mark the yield without a heating beam for a standard CD shell (black) and a CD shell lined with a 1 m CH layer (green). The neutron yield measurements as a function of the arrival time of the OMEGA EP pulse are shown Figure 2. The neutron yield peaks at a delay time of 3.65 ± 0.03 ns. The lines show the yield without the heating beam for a standard CD shell (black) and a CD shell lined with a 1 µm CH layer (green). The implosion neutron yield of the latter target is ~25% lower because the CH inner layer partially prevents mixing of CD material into the center of the imploded capsule. The error bars take the neutron statistics and the cross-calibration error into account. The error bar of the neutron yield without EP pulse is significantly lower because it represents the statistical error over many shots. The figure shows an enhancement in neutron yield by more than a factor of 2 for a properly timed OMEGA EP beam. 2-D hydrodynamic DRACO 17 simulations predict the time of peak areal density at 3.54 ns from 18,19, while a 1D LILAC 20 simulation for a similar spherical target without cone, calculated the time of peak areal density at 3.76 ns. The measured time of peak of neutron production is within the time of peak ρr from both simulations. The experiments measure (1.4 ± 0.6) 10 7 additional neutrons due to heating by the short pulse laser in a narrow time window of less than 100 ps. The 1 µm inner CH target produced the same amount of additional neutron yield from the short-pulse laser. Data points in the peak were taken on several shot days showing their reproducibility. 6/13/

26 Figure 3: Measured breakout time of the shock inside the hollow Au cones. The shock breakout was later for thicker cone tips. The compressing plastic shell rapidly increases the pressure in the center of the target, sending a strong shock wave towards the cone tip. The tip is eventually destroyed and hot plasma from the fuel assembly streams into the hollow cone. To keep the inside of the hollow cone clear of plasma for the short laser pulse, the cone tip has to withstand the plasma pressure up to the time when the fast electrons are generated. The breakout of the shock wave is accompanied by a rapid strong light emission that was measured by imploding the shell without a short-pulse beam. Figure 3 shows the time of shock breakout as a function of the tip thickness. Each data point is the average of the VISAR and SOP measurement. With increasing tip thickness the breakout was delayed. For a thickness of 10 and 15 µm, the shock breakout time was 3.70 ± 0.03 ns and 3.75 ± 0.03 ns, respectively. Shock breakout occurred 50 to 100 ps after the peak of neutron production confirming that the cone tip was intact at the time when the EP beam interacted with the target. 2D DRACO simulations for a slightly different cone design with a hyperbolic shaped inner cone surface predict shock breakout at 3.48 ns. 19 A simple estimate of CE, based on a 1D LILAC simulation (ignoring the cone) was made. The amount of energy that is required to produce the measured neutron yield was calculated. LILAC predicts a neutron yield of from the implosion alone, which is much higher than measured. A yield of (1.3 ± 0.3) 10 8 was measured from a spherical target without cone and (1.1 ± 0.1) 10 7 with cone. While the fuel assembly is robust in these thick shell implosions without cone and the areal densities are close to the 1D predicted areal densities (Ref. 12), the yields are significantly over-predicted indicating a lower temperature in the compressed target. For the CE estimate, the LILAC density and an artificially lowered temperature profiles were used to match the experimental yield. An energy coupling between 6 and 17 % produces an increased neutron yield of , depending on where the energy is deposited in the target. The higher number is obtained when the energy is deposited in the colder, outer part of the shell. DRACO/LSP simulations were performed to refine this estimate. 21 The 2-D radiationhydrodynamics code DRACO simulates the fuel assembly of the cone-target in cylindrical geometry. 18 It has been coupled with the 2-D/3-D implicit hybrid particle-in-cell code LSPii describing the fast electron propagation and energy deposition in the target. 26 Simulations were performed for the laser conditions described above assuming an initial divergence angle of the 6/13/

27 electron beam of 40. A neutron yield increase generated by fast-electron heating of ~ is calculated if it is assumed that ~ 10% of the short pulse energy is coupled into the core through fast electrons. This is consistent with the simple estimate described above. The 2-D DRACO simulations did not take radiation transport and power balance into account leading to an over-prediction of density and temperature. The predicted neutron yield produced by the implosion alone is ~two orders of magnitude higher than that measured in the experiment. The electrons are injected at the predicted time of peak of compression in the simulation. The electron transport through the gold cone was not included in this simulation. The lower conversion than observed in other experiments (Ref. 4,5) may be due to the relatively large laser focal spot and the lower laser intensity. Assuming ponderomotive scaling 23 for the hot-electron energy, the mean hot-electron energy averaged over Gaussianshaped spatial and temporal distributions with full-width at half maximum (FWHM) of ~40 µm and 10 ps, respectively, results in a hot electron temperature of ~250 kev. The hot-electron energy is probably too low for good penetration through the cone tip. Continuous slowing down calculations that do not include blooming and straggling yield a mean range of ~0.124 g/cm 2 or ~65 µm for 250 kev electrons in gold. 24 Scattering is very important in gold and significantly reduces the mean free path. A transport calculation 25 including scattering was performed for the electrons in the Au wall. The calculations show that the range in the direction of the initial electron velocity is decreased by a factor of twenty to ~3 µm much less than the cone wall thickness. The range was defined by the distance a mono-energetic electron travels until the averaged cosine of the scattering angle becomes less than 1/e. The fast electron transport through the Au cone wall into the compressed core was simulated with LSP using a simplified model in 2D planar geometry. A fast-electron beam was generated in the laser s forward direction by promoting background electrons to higher energies at the inside cone wall. The spatially and temporally averaged ponderomotive scaling 1 was assumed with a beam angular spread given byθ 1/2 = tan 2 / ( γ 1), where γ = 1 1 ( u/ c) 2 is the relativistic gamma factor, u is the electron velocity, and c is the 2 vacuum speed of light. Electrons with energy E h = ( γ 1) mc are ejected with the angleθ1/ 2 with respect to the laser beam direction from a focused laser beam from relativistic kinematics. 26 This model has been validated against measurements of fast electron produced coherent transition radiation in solid targets irradiated at W/cm (a) (b) 6/13/

28 (c) (d) Figure 4: LSP simulation for a 10 ps, 1 kj, R 80 = 27 μm EP pulse. (a) Plasma density (in g/cm 3 ), (b) fast electron beam density ( cm -3 ), (c) plasma temperature increase (in kev), and (d) magnetic field (MG) Figure 4 shows the simulation result for 10 ps, 1 kj Gaussian pulse in time and space. 80% of the laser energy is contained in a spot of radius of R 80 = 27μm. The ionization in the gold was taken into account using a Thomas-Fermi model for the equation of state and the Lee & More plasma resistivity model for the cold return current. Figure 4(a) shows the initial plasma density, (b) the hot-electron beam density, (c) the plasma temperature increase, and (d) the self generated magnetic field 10 ps after the onset of the HEPW pulse. ~50% of the fast electrons heat the cone tip and do not pass through the cone wall, while the rest reaches the fuel, see Figures (b) and (c). A small number of electrons reach the high density fuel assembly but the self generated magnetic field (d) is unfavorable so that most of the hot electrons stream sideways and miss the high density region. The target temperature (c) reaches several kev in the cone tip, which is overestimated because the code did not include radiation cooling. Simulations with a smaller focus (R 80 = 15 μm) and higher laser energy (2.6 kj) predict a significantly higher coupling. The laser intensity increased by a factor of ~8 giving a mean hotelectron energy of ~1.4 MeV. The mean free path is larger than the cone wall thickness and a significant portion of the electrons pass into the fuel assembly heating it up to a temperature close to 1 kev. The self generated magnetic field enhances the coupling into the dense core. The expected neutron yield increase is ~ with a factor of ~8 increased fraction of hot electrons depositing energy in target densities >100 g/cm 3. Integrated experiments with higher laser intensity will be performed in the next year. In conclusion, integrated fast ignition experiments with cone-in-shell targets were performed with 1 kj of short-pulse energy and ~20 kj of drive energy. About additional neutrons were generated by fast electron heating in a narrow time window of less than 100 ps at a timing that is close to the predicted peak areal density from simulations. Shock breakout measurements have been performed with the integrated targets confirming an intact cone tip at the time of peak neutron production and experimentally validating for the first time that the neutron yield is due to fast electron coupling. The major advancement of these experiments is that they have studied fast ignition core heating in a new regime of high-energydensity plasmas and high target compression. Initial 2D DRACO/LSP simulations were performed to calculate the neutron yield for the experimental conditions. Based in the simulation and a simple model, it is estimated that (10 ± 5)% of the short-pulse laser energy is coupled into the target. LSP simulations of the fast electron transport through the cone wall indicate that ~half of the electron energy is lost in the cone for the current laser intensities. Simulations at higher short-pulse energies and higher laser intensities predict an improved 6/13/

29 coupling efficiency of fast-electron energy into the imploded core. Higher laser intensities will be available in the near future that will allow testing these predictions. Integrated experiments provide fast-ignition-relevant plasma conditions to study fast-electron transport and are essential for code benchmarking. This work was supported by the U.S. Department of Energy Office of Inertial Confinement Fusion under Cooperative Agreement No. DE-FC52-08NA28302, the OFES Fusion Science Center grant No. DE-FC02-04ER54789, the OFES ACE FI grant No. DE- FG02-05ER54839, the University of Rochester, and the New York State Energy Research and Development Authority. The support of DOE does not constitute an endorsement by DOE of the views expressed in this article. REFERENCES [1] M. Tabak et al., Phys. Plasmas 1, 1626 (1994). [2] N. G. Basov, S. Yu. Gus kov, and L. P. Feokistov, J. Sov. Laser Res. 13, 396 (1992). [3] R. Kodama et al., Nature 412, 798 (2001). [4] R. Kodama et al., Nature 418, 933 (2002). [5] S. Atzeni, Phys. Plasmas 6, 3316 (1999). [6] R. Betti, A. A. Solodov, J. A. Delettrez, and C. Zhou, Phys. Plasmas 13, (2006). [7] R. J. Mason, Phys. Rev. Lett 96, (2006). [8] T. R. Boehly et al., Opt. Commun. 133, 495 (1997). [9] L. J. Waxer et al., Opt. Photonics News 16, 30 (2005). [10] W. Theobald et al. Plas. Phys. Contr. Fusion 51, (2009), W. Theobald et al., Bull. Am. Phys. Soc. 54, 187 (2009) [11] C. D. Zhou et al., Phys. Rev. Lett. 98, (2007). [12] T. R. Boehly et al, Phys. Plasmas 16, (2009) [13] J. E. Miller et al., Rev. Sci. Instrum. 78, (2007). [14] R. Lauck et al., IEEE Transactions on Nuclear Science 56, 989 (2009), V. Yu. Glebov et al., Bull. Am. Phys. Soc. 54, 265 (2009) [15] V. Y. Glebov et al Rev. Sci. Instrum. 75, 3559 (2004). [16] P. B. Radha, T. J. B. Collins, J. A. Delettrez et al., Phys. Plasmas 12, (2005). [17] K. S. Anderson et al., Bull. Am. Phys. Soc. 52, 283 (2007). [18] C. Stoeckl et al., Plasma Phys. Control. Fusion 50, (2008) [19] J. Delettrez, R. Epstein, M. C. Richardson, P. A. Jaanimagi, and B. L. Henke, Phys. Rev. A 36, 3926 (1987) [20] A. A. Solodov et al., Phys. Plasmas 16, (2009) [21] D R Welch et al., Nucl. Instrum. Methods Phys. Res. A 464, 134 (2001). [22] S. C. Wilks and W. L. Kruer, IEEE J. Quantum Electron. 33, 1954 (1997). [23] A. A. Solodov and R. Betti, Phys. Plasmas 15, (2008). [24] M. J. Berger and S. M. Seltzer, Tables of Energy Losses and Ranges of Electrons and Positrons, NASA Document SP (National Aeronautic and Space Administration, Washington, DC, 1964) p. 82. [25] A. A. Solodov and R. Betti, Phys. Plasmas 15, (2008). [26] C.I. Moore, J. P. Knauer, and D. D. Meyerhofer, Phys. Rev. Lett. 74, 2439 (1995). [27] M. Storm at al., Phys. Rev. Lett. 102, (2009). 6/13/

30 9. ELECTRON GENERATION AND TRANPORT IN FAST IGNITION (UCSD/GA/OSU/UR/UNR/UCLA) A. Electron transport through warm dense matter (UCSD/GA). In reentrant cone guided fast ignition, relativistic electrons produced from the interaction of the high intensity heating beam with the Au cone tip plasma must traverse through the thin tip of Au cone into hot dense plasmas before reaching the compressed high density fuel. Due to limited capability of the laser facilities, currently, cone tip transport properties are mostly studied using the cold solid target using either a solid wire or a planar foil attached to the cone tip. To examine the transport properties through the cold Au into dense plasmas, especially the kinetic effects envisaged to dominate from the particle in cell code simulations, an experiment has been performed recently using the Titan lasers at the Lawrence Livermore National Laboratory, where high energy ns laser pulse was used to create a warm dense plasma by shock compression and heating of low density Carbonized Resorcinol Formaldehyde (CRF) foams. A sub-ps high intensity laser was irradiated on the Au foil to produce hot electrons. Figure 1 shows the schematic of the target, layout of the laser and main diagnostics set up for the transport study. The foam package target consists of 3.9 µm thick Au, 139 µm thick RCF foam, 5 µm Cu, 25 µm CH with 0.1 µm Aluminum (Al) coated on the surface. The 3 ns long pulse with energies of 300 J was incident from the Al side with a focal spot size of ~600 µm using a large random phase plate. The high intensity CPA short pulse (0.7 ps) with maximum energy of 150 J (with a nominal peak intensity of W/cm 2 ) was tightly focused (~ 10 µm spot size) onto the Au foil. The main diagnostics were two Cu Kα imagers, one at the back and the other on the side; two highly oriented pyrolytic graphite (HOPG) spectrometers (one at the front and one at the back, both were cross calibrated for absolute Kα yield measurement). One electron spectrometer and one bremsstrahlung spectrometer were also used to monitor vacuum electrons and x-rays from emitted sideway. Fig.1. Schematics of the target and the experimental layout. Before the transport study, one needs to know the shock propagation velocity inside the foam so timing can be determined for electron transport study in cold, partially shocked and fully shocked foams. Shock propagation inside the foam was detected using Ti x-ray backlight in the side-on geometry where short pulse (100 ps) was defocused (100 µm spot size) onto a 10 µm thick Ti foil (250 6/13/

31 µm in diameter). Fig. 2(a) shows radiographs of shock propagation at two different time delay (3ns and 5ns) between the short pulse and the long pulse beams. The average flow velocity at 5 ns can be estimated to be about 9 µm/ns. Also shown (Fig.2 (b)) are the snap shots of the density contour of the simulated shock propagation at the same time from a 2D radiation hydrodynamic simulation using the h2d code. The flow velocity in the simulation at 5 ns was about µm/ns. Good agreement (with 50% error bar) on the shock velocity is obtained. H2d simulations suggest the low density foams achieved maximum compression (10X) at about 10 ns as shown in Fig. 2(c). At its maximum compression, a quite uniform region of warm density matter (between the Au foil and the shocked Cu) is created. Fig. 2. (a) Ti x-ray radiographs of the shock propagation in the foam at 3 and 5 ns; (b) density contour plots at the corresponding time from the h2d simulations; (c) density and temperature contour plot at 10 ns from h2d calculation showing 10X maximum compression and 5-10 ev warm dense plasmas created between the A and shocked Cu. In the transport study, 4 different timing delay, i.e., 3, 8.5, 10 and 13 ns, were used to examine fast electron transport properties in cold, low density foam (3 ns delay case), between fully shocked and maximum compression ( ns) and after shock breaking out at the Au foil side (13 ns case). Short pulse only shots (without the long pulse driver) were also taken for comparison. Additional comparison with cold CH (15 µm thick, solid density) as the transport medium between the Au and Cu layers in a non-refluxing condition was also performed. In this case, the long pulse (1ns, 100 J, 600 µm spot size, 1ns delay between the short and long pulse) was used to produce a large, long scale-length preformed plasma at the back of the target as the get-lost-plasma for fast electrons in order to effectively minimize the refluxing. It should be noted that the cold CH transport medium was not disturbed by the shock at the time the short pulse was fired. One distinct consistent feature in these experiments was reproducible observation of a large extended Cu Kα emission spot from the side-on Kα imager using the foam package targets at the late time delay as shown in Fig. 3. On some shots, a relatively small spot separated from the large Cu Kα spot was also clearly visible. From the width of the Kα spot and the separation distance, it is estimated 6/13/

32 that fast electrons have a large angle spread (>90 degree) when traversing through the Au foil into the warm dense plasmas. Such wide spread of fast electrons have been observed in the integrated PIC simulations of electron transport through the Au cone tip into low density plasma targets and it has been attributed to the large magnetic fields generated at the interface [1]. It should be noted that such a large extended Kα spot was not observed in the 3 ns delay case. Fig. 3. Large Cu Kα emission spot observed in the side-on Kα imager for the late delay cases (8.4 ns, 10 ns and 13 ns) where the foam is fully shocked and the shocked Cu is pushed close to the Au. The right schematic shows the diagnostic view of the target from the side. The total Kα yield obtained from the HOPG spectrometers also shows a smaller Cu yield in the 3 ns case as shown in Fig. 4. At later time, the total Kα yield is almost constant. The total Kα yield at maximum compression case is also comparable to the case where cold CH with similar mass density (ρr=1.5g/cm 2 ) was used as the transport medium as shown as blue triangles on the plot. The high yield observed in the short pulse only case with the foam package targets is well known due to the refluxing of fast electron in the target and they interact with the Cu atoms multiple times. Fig. 4. Cu Kα yields from the foam package targets (red square) at various timing delay. 0 delay means it is a short pulse only shot. Blue triangles show the Ka yield from the comparison targets where cold CH was the transport medium in the nonrefluxing condition. Reduction of the fast electrons transport through the cold low density foam is also observed in the measured vacuum electron spectra as shown in Fig. 5. With the increase in time delay, i.e., from low density foam to dense plasmas with same mass density, more fast electron with energies in a few MeV range can traverse through the foam plasma and are registered on the detector which was 37 degree from the rear target normal direction. Inhibition of fast electron transport in low density plastic foam targets from relativistic laser matter interaction with laser energies of 20 J and intensity of W/cm 2 were reported before [2] 6/13/

33 where such inhibition was attributed to the resistive inhibition due the self-generated electric fields in the low density foams for electrons with temperature in 0.5 MeV range. In our case, laser intensity is 10 times higher and the most affected electrons have energies in a few MeV range. We believe both the self-generated electric and magnetic fields may contribute to the observed reduction. Hybrid PIC modeling is underway to examine the resistive filamentation instability in the low density foam plasma and its effects on the fast electron transport. Fig. 5. Fast electron spectra from foam package target at various timing delay cases showing strong inhibition of fast electrons with energies in a few MeV range in the cold low density foams at 3 ns. Also shown is the electron spectrum in short pulse only case where both the low density foam and sheath potential at the back of the target can significantly affect the escaping electrons. Figure 6 shows the comparison of escaped fast electron spectra in the fully shocked foam case and the cold CH as the transport medium case. A factor of 5 reduction in the number of fast electrons with energies in 1-15 MeV range in the latter case is clearly visible. Given the comparable Kα yield in these two cases as discussed earlier, we believe that the large increase in the fully shocked foam case may be due to the large angular spread of the fast electrons which led to more electrons emitted sideway. While in the cold CH transport medium in the non-refluxing condition, the fast electrons may propagate in a collimated fashion. More work (modeling and experiments with multiple diagnostics at various angles) is required to further understand this interesting behavior. Fig. 6. Comparison of electron spectra from the fully shocked foam package targets and the cold CH (with similar mass density) as the transport medium in the non-refluxing condition. Large reduction in the number of the electrons with energies in the 1-15 MeV range in the latter case is clearly seen. 6/13/

34 B. Effects of the get-lost-plasmas on the observed Kα yield (UCSD/GA). In addition, in this experiment we have also examined the effects of the get-lost-plasmas at the back of the targets on the measured Kα yield by using the same long pulse with a short pulse duration (1 ns) and lower energy (100 J) and same large spot size as used in the transport in shocked foam package experiments. The short pulse is delayed by only 1 ns (irradiated right at the end of the 1 ns long pulse) to produce a sufficiently large preformed plasma as the get-lost medium for fast electrons to prevent from refluxing. The targets used in this case were the multiple planar foils consisting of 3.9 µm Au, 15 µm CH, 5 µm Cu and 20 µm Al. The long pulse was incident onto the target from the Al side. Rad-hydro h2d calculations as shown in Fig. 7 suggest that at the time of short pulse is fired, there is a significantly large preformed plasma created at the back of the targets. The 1/10 critical density surface has moved Fig. 7. Contour and line-outs of the electrons density plots at 1 ns from the h2d calculation using the experimental laser and target parameters showing presence of large preformed plasmas at the back of the target. further back by 600 µm on axis. Short pulse with the same parameters used for the transport in shocked foam package targets was used to produce fast electrons. Figure 8 illustrates the total Cu Kα yields measured from both the front and back HOPG spectrometers with and without the preformed plasma cases. 3X reduction in the Kα yield is measured when there is a large preformed plasma at the back. This important result is relevant to the characterization of energy coupling in the forward direction in the cone guided fast ignition where both in the forward and sideway, there are large plasma created in the implosion. Our measurements suggest that such large plasma is a good medium for fast electrons to dump their energy. To efficiently couple the electrons energy into the compressed core, means have to be found to confine electrons to minimize their energy loss via the side. Fig. 8. Comparison of Cu Kα yields with and without a large preformed plasma cases demonstrated effectiveness of get-lost-plasma for minimizing the refluxing. Ratio of the Kα yields measured from the rear and front HOPG spectrometers with the get-lost-plasma case agrees with the transmission of 8.05 kev x-rays in 20 µm Al and 3.9 µm Au foils. 6/13/

35 C. Electron Transport Experiments on OMEGA EP (UCSD/GA). The goal of the University of California at San Diego s FSC/NLUF project is to investigate the fast-electron transport in hot plasmas and to demonstrate collimation of fast electrons by an external magnetic field. The project consists of three steps: (1) characterization of fast electron source and transport through the cone tip; (2) study of fast-electron transport in hot, dense plasmas; and (3) demonstration of fast-electron collimation with an external magnetic field. As the first step of the project, in experiment described here, a copper wire is attached to the tip of a hollow gold cone to investigate the characteristics of the fast electrons through the tip of the cone after they are generated by the OMEGA EP 10-ps pulse. Figure 9 shows the schematic of the experimental setup of cone/wire shots on the OMEGA EP laser. The short pulse (interaction pulse) was focused into the cone/wire target. Here, the wire (40-µm diameter, 1 mm long) was made of copper and attached to the tip of a gold cone with a 20-µm-thick sidewall capped with 6-µm-thick, 30-µm-inner-diam foil. The detailed target information is shown in the inset of Fig. 9. The Cu Kα x-ray emission from the wire was diagnosed with a HOPG spectrometer at the normal direction to the wire axis. The energy spectra of the fast electrons were measured along the wire direction, i.e., on the interaction laser axis. In addition, the second short pulse (backlighter pulse) was used to generate a high-energy proton beam to measure the electrostatic field around the cone/wire target using a proton deflectometry technique. The backlighter pulse axis was perpendicular to the axis of the interaction pulse. A stack of radiochromic films (RCF s) was positioned on the axis of the backlighter pulse to detect protons at various energies, which can provide the field images at various timings, depending on the proton energies with a magnification of 9. The detectable proton energy range was 5 MeV to 60 MeV. The temporal resolution of the proton deflectometry was 10 ps to 50 ps. In the experiment, the interaction pulse energy was varied from 260 J to 820 J at a 10-ps pulse duration. At the best focus position, 80% of the laser energy was contained within a 45-µm-diam spot. The beampointing stability was monitored with an x-ray pinhole camera that can also monitor the plasma inside the cone. A 0.7-ps pulse duration was used as the backlighter pulse to minimize the proton-generation time window. The energy of the backlighter pulse was up to 300 J. Fig. 9. Experimental setup for cone wire shots with proton deflectometry. The directions of the x-ray and electron spectrometers are indicated in the figure. The detailed cone/wire information is shown in the inset. The Cu Kα signal was observed on the HOPG spectrometer with a signal-to-background contrast of up to 1.4. The Cu Kα x-ray signal was observed to be linearly dependent on the interaction pulse energy as shown in Fig. 10 (a) (solid line). The results indicate that the coupling efficiency from the laser to the Kα photons is quasi-constant in the energy range of 260 J to 820 J; therefore, more electrons pass through the tip of the cone at the higher laser energy. However, the Kα x-ray yield was about 1/4 of the data observed in experiments carried out on Titan laser at LLNL (0.7 ps, 150 J in 6/13/

36 maximum), which is represented with open squares in the figure. Those Titan experiments were performed with 30 cones, therefore we performed another set of experiments on Titan using the identical cone-wire targets (40 cones) that were used on OMEGA EP. Results are shown with open circles in Fig. 10 (a), which are very close to the Titan results with 30 cones. Fig. 10. (a) Laser energy dependence of Kα x-ray signal intensity. (b) Trend of Kα signal reduction with increase of prepulse energy. Solid circles (red) are observed on OMEGA EP experiment. We have paid particular attention to the laser contrast to explain the difference of Kα x-ray yield efficiency. The energy contained in the prepulse of Titan laser was up to ~10 mj. The energy in the amplified stimulated emission (ASE) of the interaction pulse in the OMEGA EP experiment was estimated to be up to 300 mj. This can create a large preplasma, which can reduce the fast electron coupling to the wire. We have found that the increase of prepulse energy reduces the electron coupling (i.e. Kα x-ray yield) as shown in Fig. 10 (b) in a Titan experiment using 0.7 ps interaction pulse. The OMEGA EP results (10 ps pulse) are close to the Titan data when the Kα yields are shown as a function of prepulse energy. Therefore, the lower yield on OMEGA EP seems to be caused by the preplasma due to the ASE. It is worth noting that on the Titan laser when the laser prepulse energy was increased from intrinsic level (8 mj) to 1 Joule level, the Ka yield dropped by a factor of eight. This indicates that the coupling of laser energy into fast electrons is significantly affected by the laser prepulse. The observation of large preplasma effect on coupling is consistent with the x-ray pinhole camera observations, which show a large preplasma existence in the cone target. Figure 11 (a) is an image of x-ray pinhole camera recording the x-ray emission from the inside of the cone. The wireframe shows the cone-wire target in the view. The x-ray emitted in the laser-plasma interaction was shielded by the cone walls, therefore only the direct emission was recorded. It should be noted that the x-rays were emitted from the area of 260 µm in diameter, which was about 6 times larger than the laser spot size. Assuming that the preplasma filled the tip of the cone inside, the longitudinal position of the x-ray source can be estimated from the cone target geometry. The position was estimated to be 360 µm away from the tip of the cone (Fig. 11 (b)). Another indication of preplasma existence was found in the electron energy spectra. The energy spectra of the vacuum electrons observed at 260 J and 820 J shots are shown in Fig. 12. The estimated slope temperatures were 1.6 MeV and 3.1 MeV respectively. These temperatures are much higher than the one estimated with the ponderomotive scaling (0.1 MeV and 0.6 MeV respectively). The higher temperature observation agrees with models of laser and low-density plasma interactions [3]. 6/13/

37 Radiation hydrodynamics code Hydra was used to study creation of preplasma with estimated ASE on OMEGA EP, which shows a large scale preplasma production. Figure 13 shows the density Fig. 12. Electron energy spectra measured on the wire axis in the vacuum. Fig. 13. Preplasma density profile at the main pulse interaction timing estimated with radiation hydrodynamic simulation. profile of the preplasma created by the ASE at the timing of the interaction beam injection in the OMEGA EP experiment. This numerical result is now applied to Hybrid/PIC modeling to investigate the fast electron generation and propagation through the preplasma and the cone tip. Finally, Fig. 14 shows the proton radiographs observed on a shot with 260 J in the interaction pulse. Fig. 14 (a) shows a radiograph by protons with an energy of ~18 MeV taken before the interaction pulse interacts with the cone-wire target. The proton beam deflected by the electrostatic field around the target was observed with low-energy protons (5 MeV) as shown in Fig. 14 (b). Protons were deflected by the electrostatic field only in the vertical direction in the figure because the magnetic field was in the azimuthal direction around the wire and canceled out any deflection. The electrostatic field strength observed in Fig. 14 (b) was estimated at approximately tens of kv/µm using a simple calculation of proton ray tracing. Note that maximum field strength could be higher than this estimation because the field strength can vary within a much shorter time scale than the temporal resolution of the diagnostic setup. D. LSP modeling of fast electron transport and Kα production in multilayer solid targets (UCSD/GA). Fast electron transport is commonly characterized by measuring the K-shell photon production as the result of binary collisions with fast electrons in a fluorescent layer buried in the target. To directly compare the observables, i.e., total Kα yields and spatial profile, from the LSP hybrid PIC modeling and the experiments, a medium model has been implemented in the LSP code to calculate the x-ray photons (including both the bremsstrahlung and K-shell photons) at each time step. The K-shell photon cross section is obtained from the ITS code. Fast electrons can be either injected or promoted from the background fluid plasma electrons with its energy spectrum and a conversion rate determined by various scaling laws using the local laser intensity. 6/13/

38 Figure 14. Cone/wire target proton backlight images observed with (a) 18-MeV and (b) 5-MeV protons at an interaction pulse energy of 260 J. The 5-MeV protons passed through the cone/wire target at about 200 ps after the interaction pulse interacted with the target. We have tested this new LSP capability and performed benchmarking modeling for Kα production in a thin (20 µm) Cu foil target illuminated by a short pulse (0.5 ps) high energy (100 J) laser with peak intensity of ~10 19 W/cm 2. With 10% conversion efficiency, i.e., 10 J in the promoted fast electrons, the calculated total Kα yield normalized to the laser energy is about , in good agreement with the published work [4]. The LSP code with Kα package has been used to model a recent experiment with multilayer planar foil targets (10 µm Al/10µm Cu/Al (0-500 µm)/25 µm Ag/500 µm Al) consisting of two fluorescent layers with Cu at a fixed buried depth near the front and Ag layer at various buried depth. The experiment was performed using the Titan short pulse laser (150 J, 0.7 ps and a 10 µm spot size) at LLNL. The total measured Cu yield is in the range of (2-4) 10-3 J and the total Ag Ka yield shows little dependence on the thickness of the middle Al layer. To downscale the target size for practical simulations without compromising the relevant physics, perfect match layer (PML) boundary condition for EM field and conducting boundary for particles are used to avoid strong field generation at the boundary and the resultant fast electron refluxing. Simulations were performed in 2D cylindrical geometry with R=200 µm and Z in the range of µm for three different middle Al layer thicknesses (0, 25 and 100 µm) as shown in Fig.15 inset. For simplification, the background plasma electrons have a uniform density profile ( cm -3 ) and the initial plasma temperature is 100 ev. Fast electrons with a total energy of 30 J are promoted from the background fluid electrons from the first 1 µm layer of the Al plasmas. The T hot (as shown in Fig. 15) and the conversion rate (ratio of the promoted electron density and the background plasma density) for the promoted electrons are calculated using the measured local laser intensity and power distribution obtained from the in-situ equivalent plane monitor of the laser focal spot. Fig. 16 (a) shows the time evolved Cu and Ag Kα photon yields Fig. 15. Fast electron temperature (T hot ) is calculated from the local laser intensity using ponderomotive scaling and Beg s empirical scaling. Inset shows the target plasma and boundary condition used in the LSP simulations. 6/13/

39 for three different thicknesses of the middle Al layer cases. Cu Kα production has an almost constant yield during the constant (0.7 ps) pulse duration. The integrated Cu Kα yield is about Joules, close to the measured value (see Fig 16(b)). It should be noted that there is good consistency in the calculated Cu Kα photon yields as expected with a given electron beam source, while the experiments showed quite large shot-to-shot variation, which could be caused by the nonlinear laser plasma interaction processes. (a) (b) Fig. 16. Time resolved Cu Kα and Ag Kα yield for three different middle Al layer thicknesses from the LSP simulations (a); and the measured Cu Kα yield from the multilayer targets in the experiment (b). The simulated total Ag Ka yield (time integrated) clearly shows the dependence on the its buried depth in the targets as seen in Fig. 17. At 100 µm distance, LSP simulations suggest a factor of ~2 reduction from J (with no middle Al layer) to J (with 100 µm thick middle Al layer) when fast electron T hot is provided with the ponderomotive scaling [5]. Similar reduction is also found for fast electron T hot calculated using Beg s scaling [6]. To better match the experimentally observed flat Ag Ka yield profile as the function of the buried depth, on-going simulations performing parameters scan by tuning the input electron beam and background plasma parameters including dynamics ionization are underway. Fig. 17. LSP calculated total Ag Kα yield as function of the middle Al layer thickness. 6/13/

40 The simulated total Ag Kα yield (time integrated) clearly shows the dependence on the its buried depth in the targets as seen in Fig. 17. At 100 µm distance, LSP simulations suggest a factor of ~2 reduction from J (with no middle Al layer) to J (with 100 µm thick middle Al layer) when fast electron T hot is provided with the ponderomotive scaling [5]. Similar reduction is also found for fast electron T hot calculated using Beg s scaling [6]. To better match the experimentally observed flat Ag Kα yield profile as the function of the buried depth, on-going simulations performing parameters scan by tuning the input electron beam and background plasma parameters including dynamics ionization are underway. E. In-Situ laser intensity monitor (OSU) OSU is collaborating with LLE to develop an in-situ laser intensity monitor for the MTW laser system. The MTW laser delivers approximately 10 J in a 400 fs pulse that can be focused to 10 µm giving an estimated peak intensity 2x10 19 Wcm -2. Fig. 18. Schematic of physical process considered in simulating the properties of electrons escaping a solid density target. The measurement method involves using time-of-flight detection of laser-ionized atoms in a gas jet. Measuring the distribution of ion charge states coupled with known high intensity laser ionization dynamics of atoms yields a bound on the peak laser intensity and spatio-temporal characteristics of the laser beam. The principles of this method have been experimentally demonstrated up to near Wcm - 2 [7]. Note that this method provides not only the peak focused laser intensity, but also a measure of the entire spatial and temporal profile as well due to the dependence of the atomic ionization signal on integration over both space and time. The MTW laser will interact with a gas jet causing the atoms to be ionized. Both neon and argon gases will be studied, separately, to access more ionization potentials. A micro-channel-plate detector and oscilloscope record the ions, which will be accelerated by 5 kv. One of the main problems for this experiment is noise generated by background species that exist in the chamber. To mitigate this problem, two steps have been taken. The first step is to install a liquid nitrogen cold finger in the chamber to reduce the concentration of background species. The second step is to employ an ion selector to block lower charge states from the detector. The ion selector blocks unwanted ions by coupling a high speed, high voltage pulser (3.5kV in 25 ns) to a deceleration plate. The lowest ionization states of the inert gases as well as all hydrogen, carbon, and oxygen states can be blocked from the flight path by carefully timing the selector gating voltage since their flight times are long compared to the ions of interest. The peak intensity monitor is in the final stages of production. The mechanical design was cleared by LLE engineers for production in November All of the machining has been completed and sent to LLE for welding. LLE has strict requirements for structural welding of vacuum parts and are accordingly supplying a vendor. All equipment has been purchased and most has been delivered; all outstanding items are expected to be delivered by the end of February The experiment is currently planned to take place spring of 2010 with the only administrative roadblock being an operational readiness review two weeks prior to the experiment. F. Escaping electron simulations (OSU). Numerical simulations of the energy spectrum of electrons escaping from a target struck by an ultra-intense laser pulse were performed. The simulated energy spectrum as recorded by an electron spectrometer is found to differ significantly from the spectrum computed within the target. We have performed 2-D LSP calculations on a target in which we 6/13/

41 simulated the insertion of hot electrons from intense laser plasma interaction. The conceptual idea of the physical processes is given in Fig. 18. A typical experimental result is given in Fig 19 along with the results of an LSP calculation where only the process suggested in Fig. 18 is included, that is, the target is considered clean and the ions of the target are fixed. It is clear that while our LSP results do indeed agree with the recorded signal for electron energies above 5 MeV, there is indeed something significant missing in the LSP model for electrons recorded below 5 MeV. What is missing is the inclusion of mobile ions on the target. Our analysis concludes that the emission of electrons and ions must be accounted for simultaneously in order to account for the energy distribution of the electrons. Figure 20 shows the result when the simulation includes ion emission. Analysis of the numerical simulations suggests two major mechanisms are responsible for this phenomenon: (1) The emitted electron energy spectrum is heavily influenced by the self-consistent electric fields generated along the target surface as the electrons escape; and (2) these fields are themselves substantially modified by the simultaneous departure of accelerated surface ions. The total self-consistent potential of the target relative to the detector varies with time during the electrons escape leading to recorded electron spectra that is connected to the actual electron energy distributions within the target, but in a complex fashion. In summary, LSP simulations show that the hot electron temperature measured with a time integrated vacuum electron spectrometer can be related to the time integrated input spectrum by noting the modest cooling in the slope temperature. The simulations show that electrons leave the plasma in two separate phases: 1) an early single pass phase where the hot electrons with kinetic energy greater that the electric potential escape to the detector forming the high energy portion of the vacuum electron spectrum; and 2) a long time evolution phase where the formerly trapped electrons accelerate surface ions to expand into vacuum and escape to the detector. These late electrons are lower in energy and have no direct correlation with the characteristics of the original input Fig. 19. Comparison of typical Titan vacuum electron spectrum (black squares) compared to a fixed ion simulation in LSP (red circles). electron distribution. The modest cooling of the vacuum electron temperature compared to the source, as well as the recorded high energy electrons being biased to the rising edge of the pulse, has been demonstrated for a wide range of model parameters: one order of magnitude in intensity, one order of magnitude in pulse duration, two orders of magnitude in pulse energy and for both narrow and broad angular distributions. G. Specular reflection simulations (OSU). In previous experimental campaigns we had measured the reflection and divergence of an intense pulse off a metal target as a function of intensity, polarization, pre-pulse and other parameters as part of our effort to understand the laser-plasma interaction in general and the propagation of a laser through a cone in particular. In general, we observed large changes in reflectivity and modification of pulse divergence. We have performed a modeling effort based on LSP to understand these effects. Our major findings include: 1) the divergence and reflectivity appear to offer an inexpensive way to monitor pre-plasma, 2) the pulse spatial profile effects the dynamic modification of pulse divergence, and 3) LSP appears to be able to capture the relevant aspects of the LPI insofar as these experiments are concerned. 6/13/

42 We found that reflectivity and output divergence both depend on preplasma scale length. (A number of other groups have looked at reflectivity, although under differing conditions [8,9].) For example, figure 21 shows the modeled reflectivity and reflected beam divergence versus scale length (1/e length for pre-plasma density fall-off) for one set of laser parameters. The reflectivity decreases as the outgoing beam divergence increases. Both changes are dramatic and easy to identity. Pirozhkov, et al. have recently noted that the reflectivity provides a useful indicator of when pre-plasma is present [10]. Our modeling indicates that the reflectivity combined with the reflected pulse beam divergence can provide a quantitative single-shot measure of the pre-plasma scale at minimal cost. Current approaches often involve pump-probe techniques measuring shadowgrams or interference patterns and are difficult to employ and have limited dynamic range [11]. We note that the reflectivity and divergence are relatively easy quantities to measure and we are planning to verify this behavior experimentally. We have also explored the cause of the increased beam divergence and found that, although several factors are at play, the laser modified critical surface plays a dominant role as the pre-plasma scale length increases. Although the measurements mentioned at the beginning of this section had limited diagnostics for the pre-plasma extent, we can still test the utility of these ideas by modeling the experiments and using the scale length as a fitting parameter. As an example, we Fig. 20. Comparison of LSP simulations for fixed (blue circles) and mobile (black squares) ion. The green inset shows an enhanced number of lower energy electrons due to the modification of electron energies in the expanding proton cloud. Fig. 21: Reflectivity (black) and reflected beam divergence (red) versus scale length. Simulated conditions: 110 fs pulses, 1.3 x W/cm 2, 28 o angle of incidence on a copper target, simple preplasma characterized by a single scale length consider two shots taken at Callisto and Titan, respectively. The experimental parameters and results are listed in the table below as well as best fits based on LSP modeling. The Callisto shot is reasonably well described, indicating a 1 µm scale length. The Titan shot is not as well described. A 2 µm scale length matches the divergence, but not the reflectivity. A longer scale length (~3-4 µm) can produce a match for the reflectivity, but the predicted divergence is too large (over 30 o ). Our modeling indicates that pulse spatial profile has a significant effect on these results, but LSP provides only Gaussian profiles whereas Titan produces a flat-top. We have modified LSP to produce more general spot shapes and will include this in our next effort. 6/13/

43 Laser Callisto LSP Titan LSP Incident Angle 45 o 28 o Pulse Width 130 fs 700 fs 350 fs Intensity 3 x W/cm2 2 x W/cm2 Spot size 6 µm 35 µm Spatial Profile Gaussian Flat-Top Gaussian (before focusing) Target and thickness 5 µm copper 25 µm copper Scale Length? 1 µm? 2 µm Reflectivity 28% 27% 1.3% 6.1% Divergence 4.2 o 6.6 o 16 o 16 o Table summarizing two experimental shots under very different conditions and the results of our modeling effort using the scale length as a fitting parameter. The Titan shot was modeled with a shorter pulse than used in the experiment, but we have found these results to be insensitive to pulse duration. H. Electron Beam Divergence in Cones (OSU) The spatial divergence of hot electrons is a critical factor in the ignition of a compressed core. Years of experiments on flat foils [12, 13] have determined the fwhm divergence to be ~40. Conversionefficiency measurements in cone geometry use metallic wires attached to cones extract forward-going electrons for analysis (Fig. 22a); they collect electrons carrying ~20% of the laser pulse energy, but the cone and wire outside surfaces redirect electrons so that their initial angular divergence cannot be determined. Targets have been developed without Fig. 22 a) Typical cone-wire used in conversion efficiency experiments. b) New buried cone target with cone structure formed in a solid block of aluminum. these constraints, called buried cones. A wire at the cone tip (Fig.22a) intercepts just the electrons created in the solid angle of the wire; replacing the wire with a thin copper foil, likewise buried in aluminum, at some depth from the cone tip (Fig. 22b) allows, measurements of their angular distribution through imaging the Cu-Kα fluorescence caused by passing electrons. A thick conducting block glued behind the Cu prevented electrons from passing through the Cu foil multiple times. 6/13/

44 Initial results (Fig. 23) from a Titan campaign suggest that electron divergence is increased relative to flat foils when the conical cavity has 30 m diameter tip, but is unaffected for tip diameter ~90 m relative flat target shots from previous experiments at RAL, UK [12]. REFERENCES [1] B. Chrisman et al., Phys. of Plasmas 15, (2008). [2] D. Batani et al., Phys. Rev. E 65, (2002). [3] A.J. Kemp et al., Phys. Rev. Lett. 101, (2008). [4] J. Myatt et al., Phys. of Plasmas 14, (2007). [5] S.C. Wilks et al., Phys. Rev. Lett. 69, 1383 (1992). [6] F.N. Beg et al., Phys. of Plasmas (1997). [7] E.A. Chowdhury et al., Phys. Rev. A (2001). [8] M. Cerchez et al., Phys. Rev. Lett (2008). [9] Y. Ping et al., Phys. Rev. Lett. 100, (2009). [10] A.S. Pirozhkov et al., Appl. Phys. Lett. 94, (2009). [11] R. Benattar et al., Rev. Sci. Inst (1979). [12] R.B. Stephens et al., Phys. Rev. E 69, (2004). [13] K.L. Lancaster et al., Phys. Rev. Lett (2007). Fig. 23. K α image diameter (fwhm) from a Cu foil as a function of distance in Al from the laser-plasma interaction surface for 30 μm (blue diamonds) and 90 μm (red squares) diameter buried cone tips. The blue and yellow splashes indicate data from previously published [Stephens04] flat foil experiments at RAL, UK. I. LSP+DRACO simulations of integrated fast-ignition experiments at LLE (UR) Integrated experiments on OMEGA using low-adiabat implosions of cone-in-shell plastic targets and petawatt heating pulses have begun at LLE [1]. The targets are 40-μm thick empty CD shells of ~870 μm outer diameter [Fig. 1(a)]. A hollow gold cone with an opening angle of 35 or 70 [Fig. 1(b)] is inserted through a hole in the shell. The shell is compressed using a 351-nm-wavelength, highly shaped pulse of ~3-ns duration with ~20-kJ energy [Fig. 1(c)], designed to achieve high areal densities [2]. Previous implosion experiments using similar targets but without the OMEGA-EP heating beam [3], measured a neutron yield from D-D nuclear reactions of (2-3) The OMEGA-EP petawatt laser delivers laser pulses with a μm wavelength, duration of about 10 ps, and energy of kj (expected to be increased to 2.6 kj). To simulate the integrated fast-ignition experiments on OMEGA [4], the 2D radiationhydrodynamic code DRACO [5] has been coupled with the 2D/3D implicit hybrid particle-incell code LSP [6]. DRACO includes the physics required to simulate implosion, ignition, and burn of direct-drive ICF targets. LSP uses an implicit solution for the electromagnetic fields 6/13/

45 and implicit particle push and hybrid fluid-kinetic description for plasma electrons. An implicit algorithm in LSP provides numerical stability even for very dense plasmas, when the numerical time step greatly exceeds the period of plasma oscillations. The collisional model in LSP was modified to include relativistic and high-density plasma effects, and extensively tested to reproduce the correct ranges, blooming, and straggling of hot electrons [7,8]. In the integrated simulations, DRACO is used to simulate the implosion of the plastic cone-in-shell target. Because of numerical difficulties related to the cone gold opacities, the radiation transport was turned off in the DRACO simulations. LSP is used to simulate the transport of hot electrons in the dense plasmas and heating of the dense fuel. Fig.1 Fuel assembly in the integrated simulations for a cone-in-shell plastic target used in the fastignition experiments at LLE. Schematics of (a) a plastic shell and (b) a cone tip; (c) temporal profile of the laser pulse used for the target implosion. Fig.2 (a) Target-density profile at the time of maximum ρr in the integrated simulations for the fast-ignition experiments at LLE. The dashed lines show the initial position of the cone. Hot electrons are injected in the simulations 70 μm from the target center. (b) Lineout of CD density through the z axis. The density increase at z>60 μm is due to compression by a shock reflected from the cone tip. This provides an additional source term in the temperature equations in DRACO. The plasma profiles in LSP are periodically updated according to the DRACO results. With the onset of the short-pulse laser, LSP starts calculating the fast-electron beam transport and heating in the target. LSP runs for periods of time when the hydrodynamic expansion of the plasma is negligible (~1 ps) and generates a time history of hot-electron deposition that is used as an energy source in the temperature equation in DRACO. DRACO runs for the same time interval and updates the hydrodynamic plasma parameters. DRACO and LSP run together for the duration of the short-pulse interaction, while DRACO alone simulates the hydrodynamic response of the target after the short-pulse laser ceases. The short-pulse laser-plasma interaction and the electron transport through the cone were not simulated in LSP. A simplified model was used to generate a fast-electron beam in the laser s forward direction by promoting background electrons to higher energies using ponderomotive scaling [9] and a constant conversion efficiency of laser energy into fast electrons. 6/13/

46 The simulations are performed for a 50 cone target. Figure 2(a) shows the target density obtained in a DRACO simulation at t=3.54 ns, close to the time of maximum areal density ρr~0.8 g/cm 2 (in the direction opposite to the cone). Figure 2(b) shows the density lineout through the z axis. The density in the compressed shell at this time is around 300 g/cm 3. The initial position of the cone is shown by the white dashed lines. At t=3.54 ns the cone tip is displaced away from the target center by the jet of a high-pressure CD gas escaping through the hole in the compressed shell. Despite the fact that the plastic shell is initially empty, plastic is ablated from the inner shell surface and forms a hot, low-density plasma inside the shell. The cone tip is not only shifted at this time but is also crushed, and a plastic and gold plasma fills the interior of the cone. We are currently working on optimizing the cone-in-shell implosions by varying the cone-tip thickness and distance from the target center, to preserve the integrity of the cone tip at the time of maximum ρr. Here we focus solely on the hot-electron transport in the plastic plasma outside the cone tip. We assume that it is possible to optimize the cone-inshell implosion and that the hot electrons penetrating through the cone tip reach the plastic plasma. These assumptions will be verified in future integrated simulations. In the present simulations hot electrons are injected at the time of maximum ρr in the plastic plasma after the cone tip, 70 μm away from the target center. The plastic plasma is assumed to be fully-ionized. Figure 3 shows the results from a simulation with a shot pulse of 1.3 kj energy and 10 ps duration. The electron beam has a Gaussian radial profile with FWHM of 20 μm and initial divergence half-angle of 30. The electron-distribution function is relativistic- Maxwellian with the mean energy found from the ponderomotive scaling [9] and the energy conversion efficiency to hot electrons is 30%. Figure 3 shows the snapshots of the (a) plasma density, (b) electron beam density, and (c) azimuthal magnetic field 6 ps after the beginning of the hot-electron beam. Figure 3(d) shows the plasma temperature increase caused by the heating by hot electrons. Figure 3(b) shows that the electron beam is well collimated by the self-generated resistive magnetic field. Figure 3(d) shows that the maximum temperature increase due to hot electrons is about 2 kev. It is achieved, however, in the low-density part of the plasma. The maximum temperature increase in the dense core is about 500 ev in the righthand-side of the core. Figure 4 shows the predicted neutron yield produced through fast-electron heating as a function of the short-pulse energy. Simulations were performed for various experimental conditions. A conversion efficiency of 30% from short-pulse energy into fast electrons was assumed for the triangle symbols, while 20% was assumed for the square, the cross, and the blue circle. Conversion efficiencies of 20±10% of laser energy into suprathermal electrons were measured from flat Cu foil targets under similar experimental conditions [10]. A neutron yield of up to is predicted for 2.6 kj with a 20 μm focus and a 10ps pulse duration. Figure 4 shows that the neutron yield increases significantly with laser energy and better focusing. Current experiments are performed using 1 kj, 40-μm spot OMEGA-EP pulses. Simulations show good agreement with the measured neutron yield with 10% laser fast-electron conversion efficiency (red circle in Fig. 4). Reduced conversion efficiency in the experiment can be caused by preplasma filling the interior of the cone, when the laser coupling to hot electrons takes place a hundred microns away from the cone tip or so. Not all the electrons reach the cone tip, which effectively reduces the coupling efficiency to electrons heating the target core. The predicted neutron yield is also optimistic because the transport through the gold cone was not included in 6/13/

47 the simulations. Fast-electron transport through the tip of the gold cone is significantly affected by the material properties and will affect the neutron yield. Integrated simulations including the electron transport through the cone tip are currently underway. (a) (b) Fig.3 (a) Plasma density (in g/cm 3 ), (b) hotelectron density (in cm ), and (c) azimuthal magnetic field (in MG) 6 ps after the beginning of the hot-electron beam. (e) Plasma temperature increase (in kev) in the end of the hot-electron pulse. (c) (d) Fig.4 Neutron yield: simulation versus experiment J. The role of magnetic collimation in the hot electron transport in solid-density targets (UR) Powerful lasers focused onto the surface of a solid, produce large numbers of energetic electrons. Such energetic electrons can be used in the fast ignition approach to inertial confinement fusion (ICF). In fast ignition ICF a laser generated beam of energetic electrons deposits its energy inside a precompressed deuterium-tritium fuel pellet, heating a small hot spot fast enough before the fuel expands hydrodynamically. Fast ignition requires that the energetic electrons remain collimated as they propagate through the target. Early experiments using multiterawatt laser systems observed a collimated fast electron flow in low Z targets [11]. Later experiments using imaging of K α emission from high Z buried layer targets (see Ref. [12] and references therein) measured a divergent electron flow with half-angle divergence increasing with the laser intensity: from θ 1/2 12 at I W/cm 2 to θ 1/2 27 at I W/cm 2 according to Fig. 2 of Ref. [12]. Hot-electron collimation in the experiments of Ref. [11] was explained by the presence of self-generated resistive magnetic fields [13]. Recent two-dimensional hybrid-particle-in-cell (PIC) simulations [4] showed that the effectiveness of magnetic collimation decreases with the laser intensity and that the experiments of Ref. [11] can be explained by a partial (not-complete) collimation of hot electrons. It has been, however, noticed [4] that more complete three-dimensional simulations including the details of the resistivities for the target materials are required for a better quantitative agreement of the electron divergence half-angle in the experiments and simulations. 6/13/

48 Here we report the results from three-dimensional (3D) hybrid-pic simulations of the electrontransport experiments recently performed at the Laboratory for Laser Energetics [14,15]. In this work, the effects of the target resistivity, including different target ionization levels, have been studied in detail. The simulations confirm that electron transport can be explained by partial collimation of hot electrons by resistive magnetic fields. The initial divergence halfangle of the hot electrons in the target is found to be close to the value predicted by the ponderomotive scaling. The experiments were conducted on the Multi-Terawatt (MTW) Laser Facility at the University of Rochester s Laboratory for Laser Energetics. A laser pulse of wavelength λ L =1.053 μm, with an energy of ~5 J and a duration of Δt L ~650 fs, was focused with normal incidence to a 4-μm-radius spot, producing an intensity of ~10 19 W/cm 2. The Al, Cu, Sn, and Au foil targets had transverse dimensions of 500 μm and thicknesses ranging from 5 to 100 μm. A coherent transition radiation (CTR) diagnostic was fielded to acquire images of the rearside optical emission with a spatial resolution of ~1.4 μm. Figure 5 shows three images of the rear-side emission plotted in arbitrary units of intensity. From left to right, the targets are 20-μm-thick Al, 30-μm-thick Al, and 50-μm-thick Cu. The emission contains small-scale structures, with a mean diameter of ~4.0 μm, superimposed on a larger annular feature whose diameter increases with target thickness. Figure 6 shows how the size of the rear-surface emission region grows with target thickness; the values were determined by measuring the horizontal and vertical y (µm) µm Al 30 µm Al 50 µm Cu x (µm) x (µm) x (µm) FIG. 5 Targets s rear-surface emission (in arbitrary units of intensity). x Radius (μm) R = 0.28T θ 1/2 16 ± 1 o Thickness (μm) FIG. 6 Transverse size of the rear-surface emission versus target thickness. dimensions of the emission pattern. No dependence on the target material was observed, and each point represents the radial size averaged over all materials at each thickness. The halfangle electron-beam divergence was inferred to be θ 1/2 16 using a least-squares linear fit. The three-dimensional hybrid-pic code LSP [6] has been used to model the transport of hot electrons in solid targets. The collisional model in LSP has been recently modified to include relativistic effects for hot electrons [4]. Separate species for hot electrons in different energy ranges are used to insure the correct scattering and slowing-down rates for electrons at different energy levels. The collisional model was tested to reproduce the correct ranges, blooming, and straggling of hot electrons, the correct resistive electric and magnetic fields, and to conserve energy. The simulations shown here use the Thomas-Fermi ionization model [16] and equation of state [17] to calculate changes in the ionization state and specific heat capacity 6/13/

49 of the background electrons with the target temperature. The simulations also use the Lee and More resistivity model [18]. Hot electrons are promoted from the background of plasma electrons at the left-hand-side target boundary, having an exponential energy distribution, ~ exp( E / E h ). The mean energy is given by the maximum of the ponderomotive [9] and Beg [19] scalings: E [ ] { [( ) ] ( ) } 1/ 1/ 3 h MeV = max Iλ0 / , 0.1 Iλ0 / 10 and the energy conversion efficiency to hot electrons is 20%, found to be independent of the laser intensity [10]. Here I is the intensity of a Gaussian in space and time laser pulse having the maximum intensity of W/cm 2, full width at half-maximum (FWHM) of 5.54 μm, and duration τ=650 fs. Electrons with energy E h = γ 1 mc are randomly injected in a cone with a half-angle [ 2 /( γ 1) ] ( ) 2 θ tan 1 1/ 2 = α which for α = 1 describes the angle at which electrons are ejected from a focused laser beam by the ponderomotive force [20]. Simulations have been performed for different parameters α and Al targets with thicknesses μm and transversal dimensions of 120 μm. Figure 7 (a) shows cross sections through the azimuthal magnetic field 350 fs after the peak of the laser pulse in the simulation for α = 1 and 60-μm thick target. Figure 8 (b) shows the location of the hot-electron-density isosurface (red-solid) at 50% of the peak density in each transverse plane at the same moment of time; the blue semi-transparent (a) (b) FIG. 7 (a) Cross sections through the azimuthal magnetic field (in units of MG) for the 60-μm target, 350 fs after the laser-pulse peak. (b) The fast-electron density isosurface at 50 % of the peak density in each (x, y) plane (redsolid) and the equivalent isosurface with the magnetic field artificially suppressed (blue-transparent). surface corresponds to the case with the magnetic field artificially suppressed. The hot-electron beam is partially collimated by the resistive magnetic field generated at the outer edge of the electron beam. This field is most intense in the first 20 μm in the target. The magnetic field 0 reduces the initial beam divergence half-angle of θ1/ 2 56 (averaged within the FWHM of 0 the beam spatial and temporal distribution) to θ 1/ 2 16, close to the beam divergence halfangle in the experiment. The variation of the beam density distribution with the propagation distance resembles an expanding annulus which breaks into filaments due to the resistive filamentation instability. Electron beam develops an annular shape because electrons propagating at larger angles to the axis are deflected by the edge magnetic field more than electrons propagating at smaller angles. The electron angular distribution peaks at about 16 to the axis and the density distribution develops a peak at the beam edge. The annular beam distribution is reinforced by magnetic fields generated at the inner and outer sides of the annulus, having opposite signs and collimating electrons into the annular region. 6/13/

50 The collimating effect of magnetic field is significant, as it is seen from comparison of the beam density isosurfaces with and without magnetic field in Fig. 7 (b). The magnetic field, however, is not sufficient to completely collimate the beam because of its high initial divergence. Figures 8 and 9 compare the results of simulations for different α. The divergence half-angle increases (Fig. 8) from 0 (complete collimation) for α 0. 7 to more than 16 for α > 1. Figure 9 shows the hot-electron-density distributions (for electrons with E>0.25 MeV) on the back of a 50-μm thick target, averaged in time using n( r, t) dt 1.065τ, for three values of α : α = 0. 7, 1, and 1.2. The annular density distribution in Fig. 9 (b) for α = 1 is similar to the CTR distributions in Fig. 5. Notice that CTR is thought to be produced by the high-energy part of the hot-electron distribution. We studied the spatial distributions of electrons with energies between 1.5 MeV and 2 MeV and above 2 MeV (representing a relatively small fraction of electrons) and found similar annular distributions as for the bulk of hot electrons in Fig. 9 (b). In summary, 3D hybrid-pic LSP simulations are shown to reproduce the details of electron transport in the experiments on MTW. The hot electrons are partially collimated by the self-generated resistive magnetic field. The electron beam propagates as an expanding annulus that breaks into filaments due to the resistive filamentation instability. The experimentally observed fast-electron divergence half-angle of 16 has been reproduced assuming an initial divergence half-angle in the target of ~56, close to the value expected from the simple ponderomotive acceleration formula. Similar results were btained in the simulations using approximate models for the plasma resistivity and specific heat capacity, when a constant degree of ionization of 8 for Al plasma was used [14]. (a) (b) (c) α=0.7 α=1 α=1.2 FIG. 8 Divergence half-angle versus α. FIG. 9 Rear-surface transverse-density distributions of hot electrons for the 50-μm thick target. REFERENCES [1] W. Theobald et al., Plasma Phys. Control. Fusion. 51, (2009) [2] C. D. Zhou et al., Phys. Rev. Lett. 98, (2007). [3] C. Stoeckl et al., Phys. Plasmas 14, (2007). [4] A. A. Solodov et al., Phys. Plasmas 16, (2009). [5] P. B. Radha et al., Phys. Plasmas 12, (2005). [6] D. R. Welch et al., Phys. Plasmas 13, (2006). [7] A. A. Solodov and R. Betti, Phys. Plasmas 15, (2008). [8] C. K. Li and R. D. Petrasso, Phys. Rev. E 70, (2004); C. K. Li and R. D. Petrasso, Phys. Rev. E 73, (2006). [9] S. C. Wilks and W. L. Kruer, IEEE J. Quantum Electron. 33, 1954 (1997). [10] P. M. Nilson et al., Phys. Rev. E 79, (2009). [11] M. Tatarakis et al., Phys. Rev. Lett. 81, 999 (1998); M. Borghesi et al., Phys. Rev. Lett. 83, 4309 (1999); L. Gremillet et al., Phys. Rev. Lett. 83, 5015 (1999). 6/13/

51 [12] J. S. Green et al., Phys. Rev. Lett. 100, (2008). [13] A. R. Bell and R. J. Kingham, Phys. Rev. Lett. 91, (2003). [14] M. Storm, A. A. Solodov, et al., Phys. Rev. Lett. 102, (2009). [15] A. A. Solodov et al., to appear in Journal of Physics: Conference Series. [16] R. M. More, Adv. At. Mol. Phys. 21, 305 (1985). [17] A. R. Bell, Rutherford Appleton Laboratory Report RL , 1980 (unpublished). [18] Y. T. Lee and R. M. More, Phys. Fluids 27, 1273 (1984). [19] F. N. Beg et al., Phys. Plasmas 4, 447 (1997). [20] C. I. Moore et al., Phys. Rev. Lett. 74, 2439 (1995). K. Superthermal and efficient-heating modes in the interaction of a cone target with ultra-intense laser light (UNR) We studied the interactions between a relativistic-intensity laser pulse and a cone-wire target. We observed energetic electron spectra in relativistic laser interactions with a cone target having a wire attached to the tip. We also observed the heating of the wire by energetic electrons when the laser pulse was focused into the cone, while changing the point of irradiation from the tip center (on axis) to the sidewall (off axis) as shown in Fig. 1. Temperatures of the wire were diagnosed by observing spectra of thermal neutrons produced in the deuterated wire. The experimental results show qualitative consistency with twodimensional collisional particle-in-cell (PIC) simulations with PICLS2D code [1]. These experiments were carried out using a peta watt laser at Osaka University [2]. The energy was 120 J, the wavelength μm, and the pulse duration ps. The pulse was focused by an f/7.8 off axis parabola and had a focal spot size of 50 μm, full width at half maximum. A peak intensity of approximately W/cm 2 was obtained at the target. An amplified spontaneous emission pulse started at 4 ns ahead of the main pulse with energy of more than 20 mj. This amplified spontaneous emission would form a preplasma on the inner wall of the cone. A deuterated-plastic wire target was attached to the tip of a gold hollow cone. The inner diameter of the tip was approximately 30 μm, its thickness 10 μm and the opening angle 30. The wire diameter was 20 μm. When the laser pulse was focused on the sidewall, the length of the lateral shift perpendicular to the axis of the laser pulse was μm and the laser pulse defocused on the side wall, indicating the whole laser pulse interacted with the sidewall outside the tip. The laser pulse was focused with p-polarization in the off axis case. Figure 1. Experimental configuration. The laser pulse was focused into the cone, while changing the point of irradiation from the tip center (on axis) to the sidewall (off axis). The laser pulse defocused on the side wall. The black solid line represents the focusing angle of the laser pulse in the on axis case and the gray dashed line represents the off axis. 6/13/

52 A focusing point of the laser pulse on the cone wall was observed with an x-ray pinhole camera. The x-ray pinhole camera directly observed a time-integrated x-ray image on the inner wall of the cone from the side of laser incidence with smaller angle than the opening angle of the cone. Electron spectra was observed by an electron spectrometer, which was placed along the laser axis. Neutron scintillators were used, which measured the heating temperature of the wire. Figure 2(a) shows the electron energy spectra observed with the electron spectrometer. The black points represent the spectrum in the on axis case while the gray points represent the off axis case. The lower components of the electrons with energies below 1 MeV were an order of magnitude higher in the on axis case in comparison with the off axis case. In the off axis case, higher slope components of temperatures 6 10 MeV appeared on the high energy tail of the spectrum and maximum energy was MeV. The maximum energy was 3 4 times higher in the off axis case and the temperature of the on axis case was 2 3 times smaller than that of the off axis case. Taking into account the ponderomotive acceleration [3], the temperature corresponded to that for a laser intensity of about W/cm 2, which is higher by two orders in comparison to actual intensity, indicating that the electrons were additionally accelerated with a different acceleration mechanism. Figure 2(b) shows dependence of the energetic electron temperature on the nominal length. Nominal length in this context means the length of the electron jet, which corresponds to the distance from the focus point on the side wall to the second point of the reflected light on the tip or on the other side wall as shown in Fig. 1. In the on axis case, significant signals of thermal neutrons were observed. Figure 2(c) shows the spectrum of neutrons generated by a reaction of deuterium-deuterium fusion. The number of the neutrons was about In contrasts, when the laser pulse was irradiated on the side wall, thermal neutrons were not observed. The neutron signal for side wall irradiation was less than the detection limit of the diagnostics, in other words less than The plasma ion temperature is estimated using the energy spread of thermal neutrons. The observed spectral width of 130 kev corresponded to 1.5 kev for the ion temperature in the on axis case, taking into account the energy resolution of the detector. 6/13/

53 Figure 2. (a) Electron energy spectra observed with electron spectrometer. The black points represent the spectrum in the on axis case and the gray points in the off axis case. Slope temperatures were 2.57 and MeV, respectively. (b) Dependence of the energetic electron slope temperature on the nominal length. (c) Spectrum of neutrons generated by a reaction of deuterium-deuterium fusion. The number of the neutrons was about In order to further understand the discrepancy between the energy conversion efficiencies, we carried out 2 dimensional PIC simulations. Figure 3(a) shows the simulation configuration. The size of the target in the simulations was smaller than that in the experiment while the geometrical configuration was very similar. This reduction was assumed for computational purposes and provides qualitative comparison with the experiment. The tip size of the hollow cone is 8 μm, the thickness of the tip 10 μm, the opening angle of the cone 30 deg., and the diameter of the wire 4 μm. 6/13/

54 Figure 3. (a) Simulation configuration. (b) Electron energy spectra calculated at the end of the wire (X= μm). The inset shows the lower energy population of the energy spectra (below 5 MeV). (c) Deposit energy in the wire target in X = μm. The spot size of the laser pulse is 10 μm, and the duration of the representative Gaussian pulse, 500 fs. While the laser intensity does not directly reflect experimental parameters, W/cm 2 was chosen to match the total power input into the cone. The lateral shift-length is 6 μm. The cone density is 1600 n c and the wire density 160 n c, respectively, where n c is the critical density. There is a preplasma layer in front of the cone wall, whose scale length is 0.5 μm. These simulations include binary collisions to simulate the energy deposition in the cone and the plastic wire. Figure 3(b) shows the electron energy spectra observed by counting electrons at the end of the wire [X = μm in Fig. 3(a)]. The electron spectrum in the on tip case has a times higher low-energy (below 5 MeV) population than in the off axis case, as shown in the inset of Fig. 3(b) and a lower maximum energy as shown in Fig. 3(b). The slope temperature of the on axis case is 3 times smaller than that of the off axis case. Figure 3(c) shows the deposited energy in the wire target. These results indicate that 2 times more energy is deposited 6/13/

55 inside the wire in the on axis case than in the off axis case. The results qualitatively demonstrate the differing acceleration mechanisms due to sidewall and cone-tip interactions and their impact upon heating the attached wire. Figures 4(a) and 4(b) show the magnetic field distribution in the on axis and off axis cases, respectively. The magnetic field around the wire is higher in the on axis case when compared with that in the off axis case, indicating the higher energy density electrons propagate in the wire in the on axis case. Figures 4(c) and 4(d) show the electron energy density distribution in on axis and off axis cases, respectively. In Fig. 4(d), the solid lines represent the laser incidence axis and the specular direction, and the dotted lines indicate the initial cone position. Electron jets are observed in the specular direction as shown in Fig. 4(d). In previous experiments of oblique incidence on a foil target, jetlike x-ray emission in the specular direction was also observed and simulations show the specularly reflected laser light is temporally modulated near the critical point in a plasma consisting of a steep density gradient [4]. At the sidewall of the cone, similar phenomena were also expected to be generated. The modulation can be explained by absorption at the reflection point, the stimulated Raman scattering. The modulated laser light excites plasma waves. The energetic electrons are accelerated in the plasma waves by laser wakefield acceleration. Therefore, in the off axis case, energetic electrons accelerations were enhanced by the specularly reflected laser light. While in the on axis case, interface steepening occurs at the tip of the cone due to strong photon pressure of the focused light. As a result, a large number of less energetic (~ MeV) electrons are produced as discussed in Ref. [5]. These MeV electrons heat the wire effectively as seen in Fig. 3(c). On the other hand, without the preplasma on the inner wall of the cone, the electron jets might not have been produced as discussed in Ref. [6] and the electron temperature might not have increased in the off axis case. Therefore, the electron temperature is sensitive to the parameters of the preplasma. In conclusion, we have measured energetic electron spectra in the relativistic intense laser interaction with a cone-wire target. We also measured the fast heating of the wire by the electrons when the laser pulse was focused into the cone, while changing the point of interaction from the tip center to the side wall. Higher slope components appeared in the high energy tail of the electron energy spectra in the off axis case. In contrast, when the laser interacted directly with the tip of the cone, the lower energy electron component increased on the wire, resulting in relatively efficient heating of the wire to more than 1 kev. PIC simulations show the electrons are accelerated to energies above 60 MeV by a modulated laser pulse in the sidewall interaction. Therefore, changing the focusing points on the cone wall creates two different modes of heating and accelerations, each of which can be used for different potential applications. 6/13/

56 Figure 4. Magnetic field distributions in the on axis case (a) and in the off axis case (b) at t = 500 fs. Electron energy density distributions in the on axis case (c) and in the off axis case (d) at t = 500 fs. Dotted lines in (c) and (d) represent initial cone positions. Blue solid lines show the laser axis and specular directions of the laser pulse. L. Hot electron generation forming a steep interface in super intense laser matter interaction (UNR) Current advanced laser technology is capable of producing high contrast ratio laser pulses with relativistic intensity > W/cm 2 and ultra short duration (few tens of fs). Such high contrast short laser pulses form only a small pre-plasma thus can directly interact with steep interface of targets, which is important because of various potential applications e.g. isochoric heating of dense plasmas, inertial confinement fusion, bright hard and soft x-ray sources. Although strong bulk heating beyond 100 ev is expected in this direct heating mode, the underlying heating mechanism is not fully understood. In an intense laser solid interaction, there are basically three different stages in a onedimensional interaction scenario. The first important acceleration is the ponderamotive acceleration near critical density n c, according to Wilks's scaling [3]. Then a long and intense laser pulse can sweep off the preformed plasma, and form a shelf plasma with the density ~ n c in front of the steepened interface [7]. Finally the shelf plasma density is being reduced through the interaction, and drops much lower than n c. Following that, the laser light interacts with dense plasma directly at the steepened interface. In this project we had studied this direct interaction mode. We demonstrate a one-dimensional laser-matter interaction using our PICLS code [1], which incorporates the binary collisions. This simulation focuses on the direct interaction mode. The hot electrons produced in this mode are important for the energy transport inside a solid target. 6/13/

57 Figure 5. Electron phase evolution, x-p x. (a) DCP electrons emerging, arrows show electron flow (b) 1st jet of 2ω extraction (c) 2ω and DCP electrons at later time (d) 2nd jet of 2ω. Arrows point in the direction of rotating blob in phase space. Dashed line shows the position of peak electron density. In simulation, a linearly polarized laser pulse (λ=1μm) with semi infinite envelope and (a=10), rising up in 15τ with a gaussian profile, incidents normally on slab targets of width 20λ. Absorbing boundaries are used for particles and electromagnetic fields. To account plasma density dependence on hot electron generation and their energy scaling we performed a set of simulations with different slab targets of densities ranging between 40 to 400n c. Simulations were carried out for approximately 500fs (150τ). Figure 5 shows the time evolved electron phase plots at the irradiation surface during one laser oscillation period. Electrons are trapped in the electrostatic potential and form a blob structure of thickness 3λ s =3c/ω p at the target surface. Peak longitudinal momentum of this blob depends on electrostatic potential, which is balanced by laser photon pressure [1]. Note here that the total number of electrons trapped in this blob are almost constant during the interaction. Leaving electron jets are balanced by return current which supplies electrons to the blob. In Fig.5(b), (c) and (d) we see the continued oscillation of trapped electrons and 2ω electrons extraction from outer blob region by JxB mechanism in one laser period and then injected back into the target. Except 2ω jets, additional 5-6 electron jets with lower energy are observed emerging from the rear end of blob. We define them DC-ponderomotive (DCP) electrons, as they are accelerated via DC field as discussed later. The energy of the DCP electrons depends on the peak momentum of the blob electrons. Note here that the electrons in the central region are always trapped and never come out from the blob, so the electrons forming 2ω and DCP electrons are originally located deeper in the target and they are introduced to the outer most orbit of the oscillating blob by means of the return current. For 200n c target, we observed a similar blob structure of thickness 3λ s and DCP electrons jets, but we do not see 2ω jet. However the height of blob is reduced by a factor of (1/n e ) 1/2 ~ (40/200) 1/2 =0.44. Again, the maximum number of DCP electrons group is five to six in one laser period, similar to the case of 40n c density. In other density cases such as, 70n c, 100n c, 6/13/

58 300n c and 400n c, except 70n c which has an insignificant 2ω jets, we do not observe 2ω electrons. Electron energy spectra obtained for different target densities are plotted together in Fig.6. This picture gives a clear evidence for disappearance of high energy tail in spectrum due to the 2ω jet suppression. Figure 6. Electron energy spectrum obtained in various density targets. Arrows indicate the energy of oscillating electrons in the trapped region, energies in brackets are the estimated peak energy of DCP electrons by Eq. (1), introduced in the later. A normally incident laser on target exerts a pressure P L, P L =(1+η)I/c where η is reflectivity and I is laser intensity. This pressure pushes electrons more than the ions, hence a charge separation is established and an electrostatic potential φ is excited at the target surface. This potential gap can be estimated by conserving the momentum flux of the mass flow (plasma pressure) with the light pressure, (1+η)I/c = n e eφ, at the reflection surface [7, 8]. If electrons accelerate in this potential well up to energy eφ and gain longitudinal momentum P x, then their energy is, (1) here P y is neglected, m e, c, n e are electron mass, light velocity and electron density, respectively. In simulations we observed that the DCP electrons escaping from the oscillation region have peak energy which is the average energy of the oscillations, then the energy of DCP electrons is given by <ε>=m e c 2 a 2 (1+η)n c /4n e, here, time averaging gives <P x 2 >=P x 2 /2 and using relations I/c=m e c 2 a 2 n c /2 and a=ee L /m e cω. The peak energy of oscillations in the blob for 40n c, 70n c and 300n c density cases are calculated by Eq.(1) and indicated with arrows in Fig.6, which are consistent with the observed spectrum. For example, in 40n c case the arrow clearly shows the transition point between DCP electrons and JxB electrons energy in the spectrum. As we see phase plots in Fig.6, trapped electrons form a blob like structure of thickness ~3λ s at the target surface. Note here that area of the blob in the phase space is conserved, meaning no net accumulation of charge or energy, particle conservation gives us a condition; v d n d v r n e, here v d, n d and v_ r, n e are the speed and density of DCP and return current electrons, respectively. Energy flux conservation gives us the condition I in -I rf = χi in ε d v d n d, here χ is the absorption coefficient and ε d is the energy of the DCP electrons. The energy of electrons in return current, which scales with (n d /n e ) 2, is negligible compared to that of the DCP electrons. Plugging in the DCP electron velocity v d in the energy flux conservation condition calculated from the expression which gives average energy of DCP electrons. 6/13/

59 (2) here γ d = (1+<P_x^2>/m e c 2 ) ~ 1+a 2 (n c /n e ), and (1+η)/2 ~ 1. The factor (1-1/γ d ) in Eq.(2) is very small for a 2 (n c /n e )>>1. To estimate maximum absorption we solve χ/ a = 0 and find χ max 0.6 (n d /n e ) at a 1.27(n e /n c ) 1/2, which indicates that the absorption by DCP electrons becomes efficient for a>10 and n e >100n c. The density of DCP electrons, obtained by a rough estimation in the simulation, is n d ~ n e /20 for 400n c and ~ n e /10 for 100n c. Though this is a small fraction of the input energy but DCP electrons deposit energy effectively inside the target in a small volume, which is important to create kev energy - solid density plasmas. Figure 7. Analytical scaling for longitudinal P x in Eq.(1) (DCP electrons) and JxB with the observed maximum electron momentum in simulation for various densities. In Fig.7 we plot Eq.(1) (DCP electrons), and the JxB (2ω) scaling obtained from the Helmholz boundary solution [9]. Electron momentum observed in simulations which agrees well with the 2ω scaling below 100n c, while with the DCP energy scaling above 100n c. A transition occurs around 100n c, in dominating mechanism to produce the maximum energy electrons. 2ω jets are disappeared in higher density targets (> 100n c ) since they do not have sufficient energy to overcome potential gap so remain trapped, also their suppression reduces the total absorption. Since a circularly polarized laser pulse does not produce 2ω electrons, then the absorption will occur mainly because of DCP electron acceleration. To see the laser polarization effect on DCP electron we performed simulations with circularly polarized laser light while keeping other laser and target parameters similar. Note here that in simulations we have found that the ion contribution to the total absorption is extremely small, such as < 1% for a=10, and it drops to ~ 0.001% for a=1. Hence the absorption is mainly by the electrons. The ion absorption rates are consistent with the Denavit [10]. We measured the total absorption for the period of 150τ(~500fs). In Fig.8, we can see that the absorption is low ~1% and almost constant with a=1, since the DCP electrons are inefficient. Also, the absorption for circularly polarized laser light is identical to the absorption of linearly polarized light, above 100n c. We can conclude here that for density >100n c, DCP remains the only absorption mechanism. 6/13/

60 Figure 8. Total absorption measured in various density targets for linearly (red) and circularly polarized (green) with a=10, and linearly polarized laser (blue) with a=1. We studied electron acceleration at overcritical plasma surface and demonstrated the production of DC-ponderomotive (DCP) electrons by the non-oscillating laser photon pressure. We demonstrated that the injection frequency and energy of DCP electrons are independent of laser polarization. More importantly, the DCP mechanism for electron acceleration dominates in high density targets, such as n e > 100n c. Absorption by DCP electrons does not depend on the laser polarization above n e >100n c. These DCP electrons, with short stopping ranges and relatively high density, ~10% of the target density, have a great potential to heat the target in a limited volume. We derived optimal conditions for the absorption via DCP electrons, which suggest, for example, in targets of density 100 n c or above the contribution from DCP electrons to the total laser absorption will be maximum for a pulse with a~10. M. Transport of MA electron currents in ultra-fast heated metal targets (UNR) Transport of superthermal electrons currents driven by relativistically intense laser pulse (intensity I > W/cm 2 ) is crucial for various potential applications, including fast ignition, and the generation of secondary sources of particle (ions, X-rays, positrons, neutrons). However, the stability and uniformity of those currents transported through targets are an important parameter in order these processes efficient. One method to get information about the hot electron beam is using proton imaging. Protons used for the imaging are produced by irradiating a thin foil with a short-pulse, ultra-high intensity laser and are predominantly accelerated from the rear-surface of the target [11]. The fast electrons propagated through the foil excite a ~TV/m electrostatic sheath field on its rear surface that ionizes the surface atoms and accelerates the resulting ions. The accelerated ion beam is composed mostly of protons originating primarily from contaminant layers of water vapor and hydrocarbons on the target surface. Recently it has been shown that hot electron beam are more uniform in conducting metals than in insulators [12], since in insulators there is lack of background electron to provide the return current that neutralizes the forward electrons. In this work, we show that conducting metal target could have stable or unstable transport owing to the resistivity evolution inside the ultra-fast heating target, see Fig.9. This will be an important aspect to control the hot electron transport to realize the applications. 6/13/

61 Figure 9. Experimental setup: a short-pulse, high energy laser beam hits a solid target. Protons stemming out from the rear target surface are imaged on RCFs. (a) Proton image at 6MeV, 3cm away from 40μm aluminum target, (b) same image from 15μm copper, (c) 40μm copper, and (d) 10μm gold. The main purpose of this project is to understand the different transport features in a several metal targets observed in the experiments by particle-in-cell (PIC) simulations using the same laser and targets conditions with the experiments. Typically, the electron currents generated near the vacuum plasma interface by 100TW laser light exceeds mega-ampere (MA), which is higher than the Alfven critical current. In vacuum, these currents induce selfconsistent B fields bending the electron trajectories backwards and preventing their penetration into the overdense plasma. While in a dense plasma important shielding effects arise and the high energy electron current is neutralized by a cold electron return current. This allows the high energy electrons to propagate unimpeded into the overdense plasma. However, the system is unstable to a relativistic electromagnetic two-stream instability so-called Weibel instability [13]. In a solid plasma, the kinetic instabilities are collisionally damped as discussed in Ref. [1], and the resistive magnetic fields driven by the Ohmic fields E R =ηj, which evolve by the following source terms, (3) plays an important role on the hot electron transport. We found that in low-z targets, such as aluminum, the first term of RHS is a dominant source term of the resistive magnetic fields. While in high-z targets due to a large heat capacity, and dynamic ionization, the second term becomes an important source. The magnetic field is able to be amplified in order of ~ 100MG, and changes its pattern following the resistivity evolution. The experiments were performed using the 100 TW short pulse laser at the Laboratoire pour l'utilisation des Lasers Intenses (LULI). Laser pulses of ~ J of 1μm light (350 fs) were focused at I max ~ W/cm 2 onto the front surface of various conductor targets. We used as diagnostics proton imaging, i.e. the accelerated protons that are detected in multiple layers of radiochromic film (RCF) densitometry media. The spatial distribution of the protons in a given RCF layer gives the angular emission pattern at a specific interval of proton energy. As targets, we used Al, Cu, and Au targets. The targets had on their rear surface a grooved pattern. Those targets, that have a periodic shallow modulation of the target rear-surface, allow 6/13/

62 to imprint regular modulations in the proton beam that thus allow to image the accelerating sheath surface. Figure 9 shows the experimental setup and selected images of 6 MeV protons from various targets. We see strong modulation in the image from 40 μm Al target, Fig.9(a), while 10 μm gold produces more smooth distribution with tightly focused peak at center(d). The 6MeV protons from a thin (15 μm) copper have a doughnut distribution (weak at center) (b), while a thicker (40 μm) copper has a smooth distribution and less doses (c). Note here that the maximum proton energy does not vary sensibly with the target materials, but target thickness. Also an experiment with lower (1/3) laser energy with the same spot size and pulse length showed a smooth profile from 40 μm Al target. Based on these results we infer that the fast proton images are sensitive to the target resistivity, which evolves in time due to dynamic ionization. We analyze the experiments using the two-dimensional Particle-in-Cell (PIC) code PICLS2d [Sentoku08], which features binary collisions among charged particles and ionization processes in gas and solid density plasmas. The target is modeled as a uniform slab with a small preplasma with a few micron scale in front of the target. We prepare the same material targets with the images in Fig.9, 40 μm Al, 15 and 40 μm Cu, and 10 μm Au with its surface normal oriented along a Cartesian Y axis. The target is attached to the transverse boundaries, and we use absorbing boundary condition for particles to represent the large transverse volume of target. The ion density is set to 50n c, here n c =10 21 cm -3 is the critical density for laser wavelength 1 μm. The mass (fully ionized charge) of Al, Cu, and Au ions are 27M p (13), 64M p (29), and 197M p (79), respectively, here M p is the proton mass. Then the mass density of each target becomes ρ Al =2.2, ρ Cu =5.3, and ρ Au =16.9 g/cm 3. These are close to the mass density of each metal at solid density. Initially we set the ion charge state Z=3 for all targets, and electron density is set to neutralize ion charges. Our ionization model is the Thomas-Fermi model in dense plasmas. The electron density increases dynamically during laser irradiation via ionization processes. Initially particles are at rest, with initially zero temperature. Our spatial (temporal) resolution is 1/50 of the wavelength (τ: laser oscillation period) for Al and Cu targets, and 1/100 of the wavelength (τ) for Au target. Since the laser is focused on the target surface with a gaussian profile, the hot electron current has a peak at the center, namely, the first term of Eq. (3) naturally becomes positive at lower side of focal area and negative at upper side. The second term depends on the resistivity evolution during heating and ionization by ultra-fast laser irradiation. 6/13/

63 Figure 10. (a) Electron energy density, and (b) quasi-static magnetic fields at 330 fs of 40μm Al target. (c) Resistivity profile obtained 1μm inside Al target at 80 fs. (d) Electron energy density, and (e) quasi-static magnetic fields at 200 fs of 15μm Cu target. (c) Resistivity profile obtained 1μm inside Cu target at 80 fs. (g) Electron energy density, and (h) quasi-static magnetic fields at 200 fs of 10μm Au target. (c) Resistivity profile obtained 1μm inside Au target at 80 fs. Figure 10 summarizes the simulation results of 40vμm Al, 15vμm Cu, and 10 μm Au targets. In the Al target, the hot electron flows (a) split to twin jets (hollow beam), like observed in Ref. [14]. The resistive magnetic fields (b) are in an order of ± 5MG, and they form twin channels with many small filaments, consistent with the hot electron pattern, due to the resistive instability driven by the hot electron current. We plot the resistivity inside 1 μm from the target surface at the early time (t=80fs) before the resistive fields start to grow in order to determine the dominant source term of the resistive fields. In Fig.10 (c) η has the minimum at the center. This is because the aluminum has less heat capacity and easily ionized almost to full charge state. Therefore the temperature increase is predominant over the increase of average charges, namely, η simply drops during the heating since η~z/t 3/2. Then the transverse gradient of resistivity η/ y has the opposite sign to the current gradient J x / y. Because the magnetic fields at 80 fs before breaking into twin jets (before starting the resistive instability) has the same sign with J x / y, we conclude that the current gradient (1st term of Eq.3) is the dominant source term in the Al target. The propagation speed of magnetic fields is relatively fast, which is about 60% of the speed of light and consistent with the speed of ionization waves (or heat waves) observed in the simulation. 6/13/

64 In the Au target, however, η decreases less than the Al target due to large heat capacity, owing higher electron density as a result of ionization and more energy taken to ionize the Au ions from the thermal energy. As a result, η inside the gold could have a peak at the center because the electron density has a peak at the center, and the bulk temperature stays relatively low during ionization. Fig.10 (i) shows η profile inside the Au target observed at the same time of (c). We see the η had a peak at the center. With this η profile the second source term in Eq. (3) has a positive feedback to the first term, then the resistive magnetic fields are excited more effectively. Also the observed propagation speed of the ionization waves in Au target are about 15% of the speed of light, which is ~ 4 times slower than that in the Al target. The slower propagation could grow the magnetic fields temporally longer, then together with the feedback effect of the resistivity gradient, the resistive magnetic fields is amplified to an order of magnitude stronger than those in the Al target. The resistive magnetic fields in Fig.10(h) forms a single channel with amplitude ± ~100MG, which pinches hot electron current as seen in (g). In the copper, the resistivity profile Fig.10(f) is an intermediate between Al and Au targets. The η had a twin peaks profile due to the balance of competition between heating (T increases) and ionization processes (Z increases). As a result, the resistive magnetic fields, of which amplitude is close to in the Au target, have a hollow channel (e). So that the hot electrons inside the Cu target have the hollow beam structure (d), which is similar with the Al target, however the mechanism of the formation of the hollow beams is quite different. We studied the MA current transport in conductive metal targets by two-dimensional collisional/ionization PICLS simulations. We found that the current term ( η J) is dominant source term in low Z (Al) target of resistive magnetic fields. While the resistivity gradient η plays an important role in high Z targets, such as Cu and Au targets. In high Z target, the resistive magnetic fields become extremely strong 100MG, and structure depends on the evolution of resistivity inside the ultra-fast heated target. In the current experimental conditions, the Cu target has hot electrons in a hollow beam structure, and the Au target has one single channel beam pinched by the magnetic fields. Hot electron currents affected by the strong resistive fields modulate the sheath potential patterns at target rear surface, which are imprinted in MeV protons accelerated from the target rear surface, are recorded in the RCF images. These simulations are consistent with the experimental observations. REFERENCES [1] Y. Sentoku and A. J. Kemp, J. Comput. Phys. 227, 6846 (2008). [2] Y. Kitagawa et al., IEEE J. Quantum Electron 40, 281 (2004). [3] S. C. Wilks, W. L. Kruer, M. Tabak, and A. B. Langdon, Phys. Rev. Lett., 69, 1383 (1992). [4] R. Kodama et al., Phys. Rev. Lett. 84, 674 (2000). [5] B. Chrisman, Y. Sentoku, and A. J. Kemp, Phys. Plasmas 15, (2008). [6] Y. Sentoku et al., Phys. Plasmas 6, 2855 (1999). [7] A. J. Kemp, Y. Sentoku, M. Tabak, Phys. Rev. Lett. 101, (2008). [8] Y. Sentoku, W. Kruer, M. Matsuoka, and A. Pukhov, Fusion Science and Technology 49, 278 (2006). [9] P. Gibbon, Short Pulse Laser Interaction with Matter, Imperial College Press, Chapter-5, section(5.2.2) p.139. [10] J. Denavit, Phys. Rev. Lett., 69, 3052, (1992). [11] J. Fuchs et al., Phys. Rev. Rett. 94, (2005). [12] J. Fuchs et al., Phys, Rev. Rett. 91, (2003). [13] E.S. Weibel, Phys. Rev. Lett. 2, 83 (1959). [14] M. Storm et al., Phys. Rev. Lett. 102, (2009). 6/13/

65 N. PIC simulations of core heating in fast ignition (UCLA) One of the critical issues for fast ignition of fusion targets is to understand and optimize the coupling of the ignition laser to the fast particles, and the transport of the accelerated particles in the mildly dense region of the target. Much of our work on PIC modeling of transport in FI scenarios was done assuming electron transport in target regions with flat density profiles at cm -3 [J. Tonge et. al. Phys. Plasmas , (2009)]. These studies provide important insight regarding the formation of the return current and the beam-plasma instabilities at moderate-low intensities. These studies fail to study and predict the transport of the hotelectrons across a density gradient from moderate to high densities. We have performed a series of two-dimensional PIC simulations in order to examine laser absorption into hot electrons, electron transport and energy absorption into the core of the target using ignition lasers with ultrahigh intensities, up to 5x10 21 W/cm2. We have simulated target densities from the critical density, nc, to 1000 nc, and we have used an absorbing region in order to avoid the unphysical reflux of the inward electrons through the target. This absorbing region represents the higher density region as an energy dependent drag on hot electrons, allowing for measurements of the total energy deposited in the high density region of the fast ignition target. Our results show that at the front of the target, the generated hot electrons drive turbulence in the collisionless plasma background, followed by a region with strong magnetic fields generated by the Weibel instability, due to the return current that builds up. At these ultrahigh intensities, a relativistic collisionless shock is launched, mediated by the Weibel driven magnetic fields, in a similar fashion to relativistic shocks in astrophysics. The dynamics of the Weibel/streaming instabilities leads to the slowing down of the accelerated electrons and to a strong isotropization for the lower energy electrons. The electrons in the hot tail are not affected by the magnetic field. Differences between how more energy is released in a density gradient is shown in Figure 1. FIG.1: Comparison of heat flux in a flat profile target with heat flux in a gradient target. The shock releases more energy in the target with a gradient. 6/13/

66 The energy of electrons generated at the laser-plasma interface by the ignition laser is of great interest in fast ignition. Ponderomotive scaling is usually assumed for the energy spectrum of these electrons, but the mechanism for acceleration is surprisingly not well understood. We use the Particle-In-Cell code OSIRIS to the interaction of high intensity lasers (I>=5x10 19 W/cm 2 ) with a sharp, smooth boundary of an over-critical plasma (n>>n c ) at normal incidence. Although this is an idealized version of a real target, we believe the underlying physics is pertinent to fast-ignition. In 1D and s-polarized 2D cases (that is, in cases where the laser electric field points in an unphysical direction), absorption remains essentially zero unless the plasma is preheated to temperatures T e ~100keV; in p-polarized cases, the laser is able to heat the plasma front, which bootstraps into other absorption mechanisms. Preheated s- and p-polarised cases look very similar, implying that after this bootstrap heating phase, further absorption mechanisms which are not polarization dependent take over. In addition we find that, after the plasma front has heated but before the surface has deformed substantially, none of the commonly proposed absorption mechanisms can account for the observed electron energy spectrum. We propose a new mechanism for how electrons are accelerated. A hint as to what is happening was seen in p2 vs. p1 plots for preheated 1D simulations, where p1 is the momentum in the direction of laser propagation and p2 is the transverse momentum. It is observed that the energetic electrons all have non zero transverse momentum as then move forward in the plasma. Since transverse canonical momentum is strictly conserved in 1D and the vector potential is essentially zero in the plasma then electrons which gain energy all started with a a non zero momentum in the plasma. By tracking the trajectory of the most energetic particles this fact is further confirmed. In order for an electron to gain energy it must interact with the laser electric field. Therefore, it must be able to leave the plasma. However, at the sharp interface where there is reflection there is a standing wave and the magnetic field is at an anti-node at the surface. Electrons which leave normal to the surface feel this magnetic field and are returned to the plasma without making a large excursion into the vacuum. However, an electron that leaves obliquely at the correct phase of the laser oscillation can then be rotated by the magnetic field such that as the magnetic field begins to decrease to zero the electron is moving outward normal to the surface. This electron then moves out a quarter wavelength where there is an anti-node of the electric field and it is then accelerated inward. As it does it is the deflected by the magnetic field such that it has the same transverse momentum it started with. For initially cold plasmas the surface magnetic field prevents electrons from reaching the anti-node of the electric field. Using particle tracking in the OSIRIS code, as well as the results of a simple 1D imposed-field simulation, we confirm that this mechanism results in the energy spectrum of the bulk of the accelerated electrons seen in the PIC simulations. This mechanism also limits the final axial momentum to 2a 0. Results from PIC simulations are shown in Figure 2. These results have been presented at several conferences and are now being written up for publication. 6/13/

67 FIG. 2: The above plots are momentum versus position plots of electrons accelerated by high intensity lasers with a 0 of 3 to 24. In all cases the maximum momentum of the accelerated electrons is given by 2a 0 m e c. Late last year we began work on extending OSIRIS to use the hybrid algorithm initially explored by Bruce Cohen, Andreas Kemp, and Laurent Divol*. This framework is very useful for Fast Ignition simulations. In this algorithm or framework, a full PIC with Monte Carlo collisions model is used in some regions and the hybrid algorithm is used in other regions. In the hybrid algorithm the electric field is generated from Ohm s law, and the return current in the bulk is determined by canceling out the current of the fast electrons. The advantage here is that the hybrid algorithm is much more efficient than full PIC at high density while full PIC can be used at lower densities where the laser plasma interactions take place. Particles are used throughout eliminating the need for ad hoc boundary conditions. The key point is that the collisional PIC algorithm must be completely self-consistent with the resistivity in Ohm s law. We have implemented the hybrid algorithm in OSIRIS and done some limited testing of the code with these modifications. One check on the correctness of the code is by comparing the electric field from the Full PIC with Monte-Carlo Collisions and the hybrid algorithm in 1D If the resistivity in these to algorithms is not the same then the electric fields will not be the same and charge will build up at the transition from the PIC to Hybrid regions. 6/13/

68 On the left a comparison of the Electric field generated by full PIC and the full PIC/Hybrid Algorithm with transition at 24 μm. On the right the electric current is shown. *B. Cohen, A. Kemp, and L. Divol, to appear in J. Comp. Phys. (2010) 6/13/

69 10. MAGNETIZED HIGH ENERGY-DENSITY PHYSICS AND NUCLEAR DIAGNOSTICS (MIT) During 2009 we carried out a wide range of FSC-related experiments, as indicated by the above lists of conference presentations and publications. Several of them are described in the following subsections. Many utilized our recently developed technology for monoenergetic proton radiography, in which a glass shell backlighter capsule filled with D 3 He fuel is imploded with a small subset of OMEGA drive beams, resulting in isotropic emission of 15- MeV D- 3 He protons, 3-MeV DD protons, and 3.6-MeV D- 3 He alpha particles. Others involved the development and deployment of diagnostic instruments at the National Ignition Facility for their preliminary shots in A. Probing ICF experiments and laser-plasma interactions with proton radiography A1. Identification and study of B fields generated by laser-foil interactions. We recently demonstrated [R.D. Petrasso et al., Phys. Rev. Lett. 103, (2009)] how monoenergetic proton radiography, used in combination with Lorentz force mapping, allows for precise measurement of plasma field strengths as well as unequivocal discrimination between electric and magnetic fields. Measurement of electromagnetic fields in a high-energy-density plasma can be made by passing monoenergetic protons through the plasma and observing how their trajectories are deflected by the fields. Any trajectory bending is due to the Lorentz force qv B F = E + c, (1) where q is the proton charge and v is the proton velocity, acting over a path length l characteristic of the fields spatial extent. For true quantitative analysis of data it is critical that v be known accurately. If it is known in advance whether a field is B or E, Eq. 1 can be used directly to relate any observed trajectory bending to field strength. If there is bending observed but no absolute knowledge of which field is present, the individual contributions of E and B can be determined with two independent measurements. This discrimination can be accomplished by three methods, though practical implementation is often challenging. The first method involves measurements on the same plasma made in the same way but with the direction of v reversed; the second utilizes measurements made of the same plasma but with protons of two discrete values of v ; and the third utilizes measurements on two plasmas that are identical except for the reversal of any B field; The experiment reported here utilized the third method to resolve ambiguities of field identity and field strength. The experimental setup is illustrated in Fig. 1(a). A pulse of MeV protons from the backlighter were used to image two identical, expanding plasma bubbles, formed on opposite sides of a 5-µm-thick plastic (CH) foil by two 1-ns-long laser interaction beams. Both beams had spot diameters of 850 µm and intensities of W/cm 2 ; they were fired simultaneously and incident at 23.5º from the normal to the foil. To break the nearly-isotropic proton fluence into beamlets (~1000 protons each) whose deflections could easily be 6/13/

70 (a) Monoenergetic proton radiography setup (b) Radiograph (c) Deflection & field map (at foil) CH foil Backlighter Front mesh Back bubble B 140 Beam (B2) cm cm Protons at at B dl B Beam detector foil (MG-µm) (B1) 1.35 cm 27 cm Back mesh CR-39 detector 0 6 x 10 4 Proton fluence at detector (cm -2 ) 0 Ө (degrees) FIG. 1. FSC-supported MIT proton radiography setup (a), proton radiograph of two laser-generated plasma bubbles (b), and spatial map of proton beamlet deflection angle as a function of position on the foil (c). It will be seen in Fig. 2(b) that the deflections are associated almost exclusively with a B field near the foil, and this means that (c) can also be viewed as a magnetic field map. Part (c) shows that the two bubbles were actually the same size even though the apparent sizes are different in the radiograph. The orientation of the images is as seen from behind the detector, looking toward the backlighter. The radiograph was acquired during OMEGA shot observed and quantified, 150-µm-period nickel meshes were placed on opposite sides of the foil. Figure 1(b) is the resulting radiograph, with laser timing adjusted so the bubbles were recorded 1.36 ns after the onset of the interaction beams. The top bubble image in Fig. 1(b) is a type of image we have recently begun studying and contrasting to predictions of the 2D radiation- hydrodynamic code LASNEX. The simulations indicated that proton deflections are purely a result of a toroidal B, parallel to the foil, arising from the n magnetic-field source term (where n e and T e are the electron e T e number density and temperature). While the data and simulations were qualitatively similar, there was a consistent, quantitative mismatch between them throughout the bubble evolution (predicted apparent bubble sizes were ~25% smaller than observed; predicted values of B dl were larger overall than observed; and field morphology details differed). This discrepancy effectively precluded use of the simulations to justify any a priori assumption that observed proton deflections were caused exclusively by a B field and not by any component E (parallel to the foil) of an E field. To provide direct experimental identification of the field type as well as strength, the current experiment was designed so the 2 nd bubble reverses the sign of any B relative to the first bubble (as seen from the detector) while leaving any E unchanged. If the B reversal had no effect on deflections of the monoenergetic protons used to image the plasma, any deflections would necessarily have been dominated by E. If the reversal resulted in equal but oppositely directed deflections of the monoenergetic protons, that would demonstrate the clear dominance of B. Qualitatively, the later is what we see in the image: the bubble on the front side of the foil (top of image) appears expanded, and the bubble on the back side appears contracted. Figure 1(c) shows the absolute values of the beamlet deflection angles Ө as a function of position at the foil; Ө is calculated from the apparent displacement of a beamlet in an image relative to where it would be without deflection. The peak Ө occur at the foil on two circles of 6/13/

71 the same radius, and the amplitudes are the same for both circles. This is seen quantitatively in Fig. 2(a), which shows Ө as a function of radius measured from each bubble s center. Because of Eq. 1, and the fact that B is reversed between the bubbles while E is not, it follows that we can decompose the total deflections Ө top (r) and Ө bottom (r) for the top and bottom bubbles into parts due only to B and E by assuming the two bubbles are otherwise equivalent. Then from which it follows that Ө top (r) = Ө E (r) + Ө B,top (r), (2) Ө bottom (r) = Ө E (r) - Ө B,top (r), (3) Ө E (r) = [Ө top (r) + Ө bottom (r)]/2, (4) Ө B (r) = [Ө top (r) - Ө bottom (r)]/2. (5) The results are shown in Fig. 2(b) after converting Ө B (r) and Ө E (r) to B dl and E dl using Eq. 1. The vertical display scales for E and B were selected so the relative amplitudes of the curves indicate the relative amounts of proton deflection. The effect of B greatly dominates the effect of E, whose measured amplitude is smaller than measurement uncertainties. Figure 1(c) reveals a toroidal topology for the B field. An estimate of the maximum local B for a toroidal height of 400 µm (assuming a height of order the shell thickness) is then 100 MG-µm / 400 µm ~ 0.3 MG. For this field, the Hall parameter ω ce τ (where ω ce is the electron gyrofrequency and τ is the electron-ion collision time) is of order 1. Since thermal conductivity goes as 1/[1 + (ω ce τ ) 2 ], it follows that field-induced inhibition of thermal transport across the plasma bubble boundary will occur Ө (r) (degrees) (a) Measured beamlet deflections Ө top (r) Typical error Ө bottom (r) Bubble radius r at foil (µm) 0 (b) Inferred radial profiles of B and E B dl 0 (MG µm) -50 Top bubble Bottom bubble Typical error Bubble radius r at foil (µm) 0 E dl (10 9 V/m µm) FIG. 2. Measured beamlet deflection angles Ө as a function of radius r in the top and bottom bubbles of Fig. 6(b) (positive is away from the bubble center), and inferred radial profiles of B dl and E dl in the two bubbles. In (b), the vector B dl is plotted as a positive number for a toroidal B field in the clockwise direction of Fig. 1(c), while E dl is plotted as positive for an E field pointing away from the bubble center. B has opposite directions in the two bubbles, while E has the same direction. Note that the absence of information about Ө bottom for r < ~500 µm reflects the overlap of beamlets in the center of the bottom bubble image in Fig. 1(b), which prevented beamlet deflection measurements in that region. Essential to the successful implementation of the technique of field discrimination and quantification are the isotropic and monoenergetic characteristics of the protons (the velocity 6/13/

72 uncertainty was < 1% over the imaged plasma). Other recent important methods of ion generation from intense laser-plasma interactions, while useful in different radiographic settings, would be compromised in the present context because of the energy spread and anisotropy of the ion fluences. And other techniques of single-point field measurement at extremely high laser intensities (~ W/cm 2 ) do not generate global field maps that show the entire laser-plasma morphology, a prerequisite to understanding plasma dynamics. A2. Charged-Particle Probing of X-Ray-Driven Inertial-Fusion Implosions. Measurements of x-ray driven implosions with charged particles have resulted in the quantitative characterization of critical aspects of indirect-drive inertial fusion, as described in [C.K. Li et al., Science (2010)] Three types of spontaneous electric fields differing in strength by two orders of magnitude, the largest being nearly one-tenth of the Bohr field, were discovered with time-gated proton radiographic imaging and spectrally-resolved proton self-emission. The views of the spatial structure and temporal evolution of both the laser drive in a hohlraum and implosion properties provide essential insight into, and modeling validation of, x-ray driven implosions. We performed experiments using monoenergetic proton radiography and charged-particle spectroscopy to study the x-ray drive and capsule implosions in gold (Au) hohlraums. These measurements have allowed a number of important phenomena to be observed. In particular, three types of spontaneous electric (E) fields, differing by two orders of magnitude in strength with the largest approaching the Bohr field (= ea -2 0 ~ V m -1, where a 0 is the Bohr radius), were observed. The experiments also demonstrate the absence of stochastic filamentary pattern and striations generally found in laser-driven implosions. We also observed plasma flow, supersonic jet formation, and self-generated magnetic (B) fields, determined the areal density (ρr) and implosion symmetry; and sampled different implosion phases. The experiments were performed at the OMEGA laser facility. In the backlighting experiment (Fig. 3) the hohlraum had a 2.4-mm diameter, 3.8-mm length, and 100% laserentrance holes (LEH) with 25-µm-thick Au walls over-coated on the inside with 0.3 µm parylene (CH) liner. The hohlraum was driven by 30 laser beams with wavelength µm and total laser energy ~ 11 kj in a 1-ns square pulse. The laser beams had full spatial and temporal smoothing (15), and phase plates SG4 were used. The capsule was a 30-µm-thick plastic shell of diameter 550 µm filled with 50 atm H 2 or D 2 gas. The designed radiation temperatures (~ 150 ev) and consequent capsule compression ( 10) were low. The backlighter was a D 3 He-gas-filled, glass-shell capsule with a 420-µm diameter and a 2-µm shell thickness, imploded by 30 laser beams. Two types of fusion protons with discrete birth energies of 14.7 and 3.0 MeV are produced in nuclear fusion reactions (D+ 3 He α+p and D+D T+p) ~ 80 ps. The images were recorded by CR-39 track detectors and the timing of the proton sampling was adjustable. Radiographs made by 15-MeV D 3 He protons covering a typical ICF implosion sequence (Fig. 3) contain both spatial and energy information because the CR-39 detector records the position and energy of each individual proton. The images show proton fluence versus position (Fig. 3B), to provide time-dependent information about field distributions, capsule compression, and hohlraum plasma conditions. 6/13/

73 D + 3 He α (3.6 MeV) + p (14.7 MeV) Proton radiographs 0.9 ns 1.6 ns 2.2 ns 2.8 ns Fig 3. Data from MIT s FSC-related work on radiographly of laser-driven hohlraums and their capsules. (A) Schematic of the experimental setup, with proton backlighter, hohlraum-driven implosion, CR-39 imaging detector, and laser drive beams. Fifteen laser beams entered each end of the hohlraum: 5 with incident angle 42 and 10 with angle The colors shown on the hohlraum wall indicate the laser intensity distribution [modeled by VISRAD]. The proton backlighter was driven by 30 laser beams with total laser energy ~ 11 kj in a 1-ns square pulse. The 15 MeV D 3 He backlighting protons passed through the laser-driven hohlraum, sampling plasma conditions and capsule implosions at different times. Images in (B) show proton fluence (within each image, darker means higher fluence). B. Proton and neutron diagnostic measurements at the NIF From October through December of 2009, the National Ignition Facility performed its first, preliminary ICF experiments. The fuel used was either D 2 or D 3 He, mixed in either case with a dominant amount of 4 He (~ 90% to 95% by number) to make fusion yields very low (since the goals of the experiments were study and optimization of drive conditions rather than fusion). During a number of these experiments. B1. Proton spectra. During the fall 2009 campaign at the NIF, MIT fielded wedge-range-filter (WRF) proton spectrometers. The setup is shown in Fig. 4a, which shows a WRF module mounted on a polar DIM alongside the gated x-ray diagnostic and viewing the capsule in a hohlraum through a laser entrance hole. Figure 4b shows a spectrum of D 3 He protons from the shot on Novermber 23, and the spectrum clearly shows the protons generated during the shock bang and the compression bang. Such spectra are being used to measure areal densities, evaluate simulations and understand implosion dynamics. 6/13/

74 (a) 2 WRF spectrometers Yield / MeV [ 10 8 ] (b) NIF shot N Compression Shock MeV Birth energy Fig. 4. MIT s FSC-related experimental setup (a) for WRF proton spectrometers at the NIF, and data (b) acquired on November 23. B2. Neutron yield measurements with a compact, passive detector. A new design was developed for a simple neutron detector based on CR-39, intended for use in measuring yields of 2.45-MeV neutrons generated in fusion reactions in inertial-confinement-fusion experiments utilizing deuterium as a fuel [F.H. Séguin et al., to be submitted]. A polyethylene filter is placed in front of half the area of a piece of CR-39, and incident neutrons elastically scatter protons into the CR-39 where they generate tracks that can be counted. By subtracting the number of tracks per unit area in the area with no polyethylene from the number behind the polyethylene, a number proportional to the neutron fluence and independent of noise in the CR-39 is obtained. This is a passive design that is insensitive to EMP problems, x or γ rays, 1/3-µm or unconverted light. It is very compact (~ 6 cm) and easily mounted, so that multiple modules can be used at different positions simultaneously. It is also has a directional response, and therefore rejects a large part of any isotropic field of neutrons scattered from an experiment chamber. The dynamic range is from about neutrons / cm 2. Calibration measurements were made on a modified CW accelerator are shown, and data were acquired during the early operation of the National Ignition Facility (NIF) in Figure 5 shows the hardware implementation we ve adopted. It includes three round pieces of CR-39, 5 cm in diameter. In front of each piece is an aluminum disk. The front disk is 3 mm thick, for protection, while the second and third Al disks are 100-mm thick to prevent protons generated in any CR-39 layer from passing to the next CR-39 layer. Directly in front of half of each piece of CR-39 is a layer of C 2 H 4 that is 100 µm thick (approximately the range of a MeV proton, to give the maximum number of protons at the CR-39 for 2.45-MeV neutrons. The components are contained in a very compact package consisting of a cylindrical sleeve and a backplate with an assembly designed to couple with mounting hardware used for 6/13/

75 experiments at the OMEGA laser. It provides the option of using frontside and/or backside analysis, and has redundancy with three separate pieces of CR-39 in case anything should happen to one of them or in case it is desired to average results for increased accuracy and improved statistics. 9.4 cm 5.3 cm FIG. 5. Drawing of the packaging used for MIT s FSC-related neutron detector (above), and an exploded view of the package contents. During the fall experiments, CR39-based neutron detector packages were fielded at one or two of the three positions indicated in Fig. 6, which were 27 or 127 cm from target-chamber center (TCC) in the equatorial plane, and 52 cm from TCC in the polar direction (toward the top in the figure). Two detector packages were mounted simultaneously at the polar position, where they shared DIM 0-0 with a gated x-ray detector (GXD), looking at the imploding capsule through the hohlraum s laser entrance hole (LEH). Figure7 shows yields measured for these experiments with ntof and with the CR-39 method. For each experiment there are three different ntof measurements and a measurement for each separate piece of CR-39 exposed in the detector packs; the CR39-based measurements are labeled according to the measurement location. Overall, the CR-39 measurements track the ntof measurements well, though they tend to be slightly lower. Where CR-39 measurements are available at different distances from TCC, as they are for shots and 12-04, those that are farther away are slightly higher. For shot 11-23, the ratio of the apparent yields at 52 cm and 27 cm is ~1.12 ± 0.07, while for shot the ratio of the apparent yields at 127 cm and 52 cm is 1.30 ± 0.24 (the uncertainty in these measured ratios is independent of any systematic errors in the calibration). The deviations of these ratios from 1.0 can be interpreted as a consequence of the detection of neutrons scattered from the hohlraum, from other diagnostics, or from the target chamber wall. These measurements were used to check and verify the yield measurements made with ntof and with activation methods. 6/13/

76 Pole direction Neutron detectors on DIM 0-0 GXD Hohlraum at TCC 52 cm 27 cm Neutron detectors on DIM cm FIG. 6. Positions of CR-39 neutron detectors relative to the capsule-in-hohlraum located at the center of the 500-cm-radius NIF target chamber. One or two neutron detectors share a port with the gated x-ray detector (GXD), which is adjacent to their lines of sight to TCC and could scatter some neutrons into the neutron detectors. One neutron detector was generally placed either 27 cm or 127 cm from TCC in the equatorial plane. This figure is approximately to scale Black = ntof Blue = Backside CR39 Red = Frontside CR Yn (10 9 ) cm [DIM 90-45] 52 cm [DIM 0-0] 127 cm [DIM 90-45] NIF 4.7 Shot FIG. 7 Neutron yields measured as part of MIT s FSC activity on NIF shots in 2009 on the dates shown on the horizontal axis. Measurements were made using ntof and using CR-39 detectors. 6/13/