Laser and Particle Beams ~2006!, 24, 535540. Printed in the USA.

Copyright 2006 Cambridge University Press 0263-0346006 $16.00 DOI: 10.10170S026303460606071X

Astrophysical radiative shocks: From modeling to laboratory experiments

MATTHIAS GONZLEZ,1 CHANTAL STEHL,2 EDOUARD AUDIT,1 MICHEL BUSQUET,2,6 BEDRICH RUS,3 FRDRIC THAIS,4,1 OUALI ACEF,2 PATRICE BARROSO,5 ABRAHAM BAR-SHALOM,6 DANIEL BAUDUIN,5 MICHAELA KOZLOV,3 THIBAUT LERY,7 ALI MADOURI,8 TOMAS MOCEK,3 and JIRI POLAN 3
DSM0DAPNIA0Service dAstrophysique, CEA-Saclay, Gif-sur-Yvette, France LUTH, UMR811 du CNRS, Observatoire de Paris, Meudon, France 3 Institute of Physics, PALS Center, Prague 4 DSM0DRECAM0SPAM, CEA-Saclay, Gif-sur-Yvette, France 5 GEPI, UMR 8111 du CNRS, Observatoire de Paris, Meudon, France 6 ARTEP Inc., Columbia, Maryland 7 DIAS, Dublin, Ireland 8 LPN UPR20, Marcoussis, France
2 1

~Received 25 July 2006; Accepted 2 August 2006!

Abstract Radiative shock waves are observed around astronomical objects in a wide variety of environments, for example, they herald the birth of stars and sometimes their death. Such shocks can also be created in the laboratory, for example, by using energetic lasers. In the astronomical case, each observation is unique and almost fixed in time, while shocks produced in the laboratory and by numerical simulations can be reproduced, and investigated in greater detail. The combined study of experimental and computational results, as presented here, becomes a unique and powerful probe to understanding radiative shock physics. Here we show the first experiment on radiative shock performed at the PALS laser facility. The shock is driven by a piston made from plastic and gold in a cell filled with xenon at 0.2 bar. During the first 40 ns of the experiment, we have traced the radiative precursor velocity, that is showing a strong decrease at that stage. Three-dimensional ~3D! numerical simulations, including state-of-art opacities, seem to indicate that the slowing down of the precursor is consistent with a radiative loss, induced by a transmission coefficient of about 60% at the walls of the cell. We infer that such 3D radiative effects are governed by the lateral extension of the shock wave, by the value of the opacity, and by the reflection on the walls. Further investigations will be required to quantify the relative importance of each component on the shock properties. Keywords: Laboratory astrophysics; Laser plasmas; Radiative shock waves; Radiative transfer

1. INTRODUCTION Radiative shock waves are highly hypersonic shocks where radiation transport is important in the total energy budget. At high shock velocity, the shocked medium is heated, ionized, and emits radiation, which in turn ionizes and heats the cold unshocked gas, and leads to the creation of a radiative precursor. Since the pioneering work of Zeldovich and Raiser ~1967!, radiative shock waves have been widely theoretically and experimentally studied. The structure of the shock varies with the shock velocity ~Mihalas & Mihalas,
Address correspondence and reprint requests to: Chantal Stehl, Observatoire de Paris-Meudon LUTH, 5 Place Jules Janssen, 92 195 Meudon, France. E-mail: chantal.stehle@obspm.fr

1984!. Radiative shocks are found in different applications: entry of rockets or comets into planetary atmospheres, pulsations of evolved stars, accretion flows in the early stages of stellar evolution, final supernovae stages of stars, and laser inertial fusion. After the pioneering experimental work of Bozier et al. ~1986! on CEA0Limeil laser ~40 J, 0.35 mm!, several experiments dedicated to radiative shocks have been conducted on the LULI laser ~Fleury et al., 2002; Bouquet et al., 2004!. High velocity shocks have been obtained on Xe at 0.2 bar, with velocities of about 70 km0s. The electronic densities in the radiative precursor was about 4 10 19 cm 3, and indication of a shock temperature of 15 to 20 eV was found ~Vinci et al., 2005!. Experiments were also performed on 535

536 the OMEGA facility ~with 10 beams, 0.35 mm, 1 ns, 4 kJ!, mostly on Xe at 1 bar. The shock velocity was measured between 110 and 150 km0s. The numerical description of the coupling between radiation and hydrodynamics in these shocks is a difficult task due to the strong gradients, and the different scale-lengths involved in hydrodynamics and radiative transport. It requires also a good set of equations of state and opacity for a wide range of plasma conditions. The complete detailed study of the structure of the shock, involving the computation of the spectral radiation flux can presently only be overtaken using one-dimensional ~1D! picture. On the other hand, the finite lateral size of the real shocks in the universe or in laboratory experiments may introduce departures from 1D behavior, especially due to the lateral radiative losses. Radiative losses occur in the precursor when the photon mean free path is not small compared to the lateral dimension of the shock. Loss of ionizing photons as well as decrease of the temperature of the shocked gas, when internal energy is given to the surrounding medium will strongly affect the development of the radiative precursor, and lower its velocity ~Leygnac, 2004!. Such radiative losses were suspected in the above mentioned experimental studies in low pressure Xe ~Bouquet et al., 2004!. The low experimental value of the precursor velocity compared to 1D simulations was attributed to radiative losses ~Leygnac, 2004!, and confirmed later with two-dimensional ~2D! simulations ~Boireau, 2005! using the FCI code ~Dautray & Watteau, 1993!. In order to study these radiative losses effects, we conducted an experiment with the PALS laser ~Ullschmied et al., 1999; Jungwirth et al., 2001; Jungwirth, 2005!, increasing the recording time ~50 ns instead of 10 ns!, and the length of the cell. This gave the first experimental proof of the damping of the precursor radiative wave. The experimental results are consistent with the hydro-radiative simulations obtained with the three-dimensional ~3D! radiation-hydrodynamics code HERACLES ~Gonzlez & Audit, 2005; Gonzlez et al., 2006!, using xenon opacity computed with the super transition arrays ~STA! formalism ~Bar-Shalom et al., 1989, 1995!, in the temperature density conditions ~Te 120 eV, and r 10 4 10 3 g0cm 3 !. 2. EXPERIMENTAL SETUP The shock tube is made from a Xe-filled glass cell with a thin gilded plastic foil as a pusher, accelerated by rocket effect from the laser created plasma on its front surface ~cf. Fig. 1!. The iodine 700 J PALS laser with a pulse duration of 0.35 ns is used at his third harmonic ~l 438 nm! and smoothed with a phase zones plate ~PZP!, in order to have a uniform focal spot of approximately 650 mm with a resulting intensity of 11.5 10 14 W0cm 2. The glass cell has a square section of 700 mm and a length of 4 mm. A 10 mm CH foil is glued on top of it. It is filled with 0.2 bar Xe using a 1 cm 3 reservoir to maintain the pressure. The back surface ~toward the Xe! of the foil is overcoated with a 0.5 mm gold

M. Gonzlez et al.

Fig. 1. Picture of the target and its support. The gilded plastic foil is visible on the top.

layer to prevent preheating of the gas by keV X-rays emitted during the laser-plasma interaction. We set an interferometric measurement of the electron density with a Fresnel biprism and a green laser beam transverse to the cell. This probe is obtained with a continuous 5 W Verdi laser operating at 12 W and 532 nm, and an acousto-optical switch giving a flat 150 ns pulse. A visible Hamamatsu TM C5680 streak camera records ~green! light through the cell along its axis during 50 ns. However, for the shot we analyzed in this paper, most of the signal comes from the stray light originating from the scattering of visible emission of the ablated plasma, and going through the cell which is absorbed by the precursor, as illustrated in the experimental setup reported in Figure 2a. It gives also a good time marker, visible as horizontal bright area in the streak camera record in Figure 2b. In these figures, the PALS laser comes from the left and the shock propagates from the left to the right along the x-axis. Some optical defect in the target prevents observing the first half millimeter after the initial position of the plastic foil, and may hide the position of the pusher interface. The propagation of a discontinuity with a roughly parabolic trajectory is clearly seen. We shall show that it corresponds to the position of the precursor front whereas the shock front is not seen in the shot. 3. MODELING The modeling of the experiment was done using two codes: the 1D MULTI hydro-Lagrangian code ~Ramis et al., 1988! was used to determine the pusher velocity with all the appropriate physics, and the 3D Eulerian radiation hydrodynamics code HERACLES ~Gonzlez & Audit, 2005; Gonzlez et al., 2006! was used to determine the multidimensional radiative effects. We perform simulation of the laser ablation on the pusher with the 1D Lagrangian code MULTI using experimental

Radiative shocks: Modeling and experiment

537

Fig. 2. Schematic experimental setup ~a! and raw image of the streak camera recording ~b!. The shock progates from the left to the rigth. Time goes upwards on vertical axis ~scale 50 ns!. The shock direction propagation x is on the horizontal axis ~scale 4 mm!. The horizontal bright spot occurs during the laser pulse, 10 ns after the beginning of the record. The initial position of the plastic foil is on the left.

conditions ~foil thickness, laser intensity versus time, Xe pressure! and detailed physics ~laser absorption, realistic ionization, and EOS. . .!. The Xe has an initial density of 1 10 3 g0cm 3 and a temperature of 0.03 eV. We find a mean pusher rear surface velocity Vp of 60 km0s without any noticeable slow down for more than 50 ns after the laser pulse ~cf. Fig. 3!. Note that in our conditions, the pusher velocity is roughly proportional to the mean laser intensity IL , thus, the precision on the velocity is equal to the precision on IL mentioned in Section 2. We then used the HERACLES code to perform a 3D simulation of the shock propagation, taking into account the radiative loss on the cell walls. In order to save computation time, we used a 2D cylindrical simulations, after making a test run to check that if it gives the same result as a 3D square cell simulation if the ratio surface over volume was preserved. The pusher velocity due to the ablation by the laser can not be computed self-coherently in the HERACLES simulation. We therefore used the pusher velocity given by the MULTI code for the first 0.3 ns ~duration of the laser pulse! and the shock then propagates inertially. We have checked that the pusher kinematics had very little influence on the precursor velocity, at least for the first 30 ns. The wall of the cell are reflective for the hydrodynamics and have an adjustable reflectivity for the radiation. This reflectivity is the only parameter that we modified to fit the experimental results. The Xe opacities used are those of the

Planck mean kP given by the STA code. A typical result of simulation after 50 ns is shown in Figures 4a 4c ~density, temperature, and optical mean free path 10kP r!. The shock front is slightly bent and the compression is higher on the

Fig. 3. Mesh position x in mm versus time obtained with MULTI. Laser comes from the left. Time starts at the maximum of the laser pulse and position at the rear surface of the pusher. The pusher velocity is found equal to 60 km0s.

538

M. Gonzlez et al.

Fig. 4. Maps of density in g0cm 3 ~a!, temperature in eV ~b! and mean free path divided by the width of the canal ~0.7 mm! ~c! 50 ns after the laser pulse computed with HERACLES. In abscissa, the distance in mm from the cell symmetry axis and in ordinate the distance from the initial position of the pusher. The shock propagates from bottom to top. The right boundary is reflective for hydrodynamics and with a 40% reflectivity for the radiation.

Fig. 5. Profile of the density ~a!, temperature ~b! and mean free path ~c! along the symmetry axis ~x 0! corresponding to Figure 4.

side due to the lateral radiative loses. The maximum compression factor is about eight and the temperature in the shock is around 15 eV. In the precursor, the typical temperature values are 510 eV. Profiles of the density, temperature, and mean free path are displayed in Figures 5. Whereas the temperature decreases smoothly along the shock structure, the density displays two peaks, one at the shock ~;1.7 mm!, and the other at the front of the precursor ~;2.5 mm!. The first spike ~;2 mm! in the mean free path is due to the sharp density drop after the shock, and the second ~;2.7 mm! corresponds to the correlation between the strong decrease in the temperature and the density bump ~the minimum of the mean free path corresponds to the maximum of the density!. Once known the pusher velocity, the free parameter in the simulations, which needs to be determined to fit the experimental results is the reflectivity of the lateral walls. We have explored its effect on the shock structure and dynamics by changing the coefficient between 40% and 80%. In all the simulations done, this parameter has a negligible influence on the shock velocity. However, walls reflectivity has sig-

Radiative shocks: Modeling and experiment

539 light!. At the considered wavelength, the main contribution comes from free-free interactions ~corrected by the induced emission!. These opacities are then integrated along the width cell in order to compute the transmission of the cell. Finally, we reconstruct the time and space transmission plot to be compared with the experimental recording ~cf. Fig. 7!. The transmission ratio obtained between the unperturbed medium and the precursor is 6070%. This has to be compared to 70% found in the experiment. On the other hand, we see that the transmission in the shock is nearly zero and that we should see this feature in the experimental diagnostic, which is not the case. This can be due to several reasons. In the simulations performed by HERACLES, the pusher and the ablation are not treated. Second, the piston made of gold and plastic is not considered and behind the shock, plastic will have a strong emission. Finally, as it can be seen in the experimental figure, this region is polluted over a few hundred microns by remaining light. This issue needs further investigations. 4. CONCLUSION In this paper, we presented the first experiment on radiative shock performed at PALS laser facility. The shock is driven by a piston made from plastic and gold in a cell filled with Xe at 0.2 bar. We experimentally recorded the radiative precursor velocity over the first 40 ns and observed a strong decrease of velocity. The simulations, including state of art opacities indicate that the slowing down of the precursor is consistent with a radiative loss induced by a transmission coefficient of about 60% at the walls of the cell. Such 3D radiative effects are governed by the lateral extension of the shock wave, by the value of the opacity, and by the reflec-

Fig. 6. Position of the front of the precursor versus time. Laser comes from the left. The blue points were obtained from the experiment, the green ~resp. red, black! line corresponds to a simulation with a wall transmission of 40% ~respectively, 60%, 80%!.

nificant influence on the precursor velocity, as might have been seen in Figure 6. The figure shows a slow down of the precursor because of lateral losses, which increases with decreasing reflectivity. We then tuned this parameter to fit the experiment curve and found that a 40% ~6 20%! reflectivity can reproduce very well the precursor velocity. This is in reasonable agreement with the albedo of glass heated by a radiative wave at 15 eV. Once the simulation is tuned to reproduce the precursor velocity, we simulate the diagnostic. Every 0.5 ns, we get a map of density and temperature. These maps are then used to deduce the opacity ~in the wavelength range of the stray-

Fig. 7. ~a!: simulated transmission of the stray-light using the HERACLES simulation. ~b!: false color streak camera recording. The laser beam comes from the left. The horizontal bright spot occurs during the laser pulse. Time goes upwards for 40 ns. The black curve is the precursor position obtained in the simulation.

540 tion on the walls. These studies are relevant to analyse the radiative shock structure and are therefore relevant for astrophysical shocks in terms of energetics balance, shock dynamics and structure. ACKNOWLEDGMENTS
We acknowledge financial supports from the Access to Research Infrastructures activity in the Sixth Framework Programme of the EU ~contract RII3-CT-2003-506350, Laserlab Europe!, also by RTN JETSET ~contract MRTN-CT-2004 005592!, and by french CNRS program PNPS.

M. Gonzlez et al.
Fleury, X., Bouquet, S., Stehl, C., Koenig, M., Batani, D., Benuzzi-Mounaix, A., Chize, J.P., Grandjouan N, Grenier, J., Hall, T., Henry, E., Lafon, J.P.J., Leygnac, S., Malka, V., Marchet, B., Merdji, H., Michaut, C. & Thais, F. ~2002!. A laser experiment for studying radiative shocks in astrophysics. Laser Part. Beams 20, 263. Gonzlez, M. & Audit, E. ~2005!. Numerical treatment of radiative transfer. APSS 298, 357. Gonzlez, M., Audit, E. & Huynh, P. ~2006!. HERACLES: A three dimensional radiation hydrodynamics code. A&A, submitted. Jungwirth, K. ~2005!. Recent highlights of the PALS research program. Laser Part. Beams 23, 177182. Jungwirth, K., Cejnarova, A., Juha, L., Kralikova, B., Krasa, J., Krousky, E., Krupickova, P., Laska, L., Masek, K., Prag, A., Renner, O., Rohlena, K., Rus, B., Skala, J., Straka, P. & Ullschmied, J. ~2001!. The Prague asterix laser system. Phys. Plasmas 8, 2495. Leygnac, S. ~2004!. Etude numrique et Exprimentale Dondes de Chocs Surcritiques et Effets Multidimensionnels du Rayonnement, PhD Thesis. Orsay: Universit Paris XI. Mihalas, D. & Mihalas, B.D. ~1984!. Foundation of Radiation Hydrodynamics. Oxford: Oxford University Press. Ramis, R., Schmalz, R. & Meyer-Ter-Vehn, J. ~1988!. MULTI: a computer code for one-dimensional multigroup radiation hydrodynamics. Comp. Phys. Comm. 49, 475. Ullschmied, J., Cejnarova, A., Juha, L., Jungwirth, K., Kolman, B., Kozlova, M., Kralikova, B., Krasa, J., Krousky, E., Krupickova, P., Kubat, P., Laska, L., Masek, K., Mocek, T., Pfeifer, M., Prg, A., Renner, O., Rohlena, K., Rus, B., Skala, J., Straka, P. & Turcicova, H. ~1999!. PALSThe first year of operation. Laser Part. Beams 17, 179. Vinci, T., Koenig, M., Benuzzi-Mounaix, A., Boireau, L., Bouquet, S., Leygnac, S., Michaut, C., Stehle, C., Peyrusse, O. & Batani, D ~2005!. Density and temperature measurements on laser generated radiative shocks. APSS 298, 333. Zeldovich, Y.B. & Raiser, Y.P. ~1967!. Physics of Shock Waves and High Temperature Hydrodynamic Phenomena. New York: Academic Press.

REFERENCES
Bar-Shalom, A., Oreg, J., Goldstein, W.H., Schvarts, D. & Zigler, A. ~1989!. Super-transition-arrays: A model for the spectral analysis of hot, dense plasma. Phys. Rev. A 40, 3183. Bar-Shalom, A., Oreg, J., Seely, J.F., Feldmann, U., Brown, C.M., Hammel, B.A., Lee, R.W. & Back, C.A. ~1995!. Interpretation of hot and dense absorption spectra of a near local thermodynamic equilibrium plasma by the super-transitionarray method. Phys. Rev. E 52, 6686. Boireau, L. ~2005!. Astrophysique de Laboratoire: Modlisation Analytique et Numrique du Choc Radiatif. Expriences au Moyen de Lasers de Puissance, PhD Thesis. Paris: Universit Paris VI. Bouquet, S., Stehl, C., Koenig, M., Chize, J.P., BenuzziMounaix, A., Batani, S., Leygnac, S., Fleury, X., Merdji, H., Michaut, C., Thais, F., Grandjouan, N., Hall, T., Henry, E., Malka, V. & Lafon, J.P.J. ~2004!. Observations of laser driven supercritical radiative shock precursors. Phys. Rev. Lett. 92, 5001. Bozier, J.C., Thiell, G., Lebreton, J.P., Azra, S., Decroisette, M. & Schirmann, D. ~1986!. Experimental-observation of a radiative wave generated in xenon by a laser-driven supercritical shock. Phys. Rev. Lett. 57, 1304. Dautray, R. & Watteau, J.P. ~1993!. La Fusion Thermonucleaire Inertielle par Laser. Paris: Eyrolles.