Energy and Technology Review Lawrence Livermore National Laboratory April 1985 Novette Target-Ph ysics Experiments

We have used the Novette to explore physical effects vital to the attainment of ICF goals: those governing the transfer of energy from a laser beam to a target capsule and the peak compression reached in the fuel.

The objectives of the LLNL laser to produce the targets for laser For further information contact fusion program are to understand and irradiation, and develop our modeling R. Paul Drake (415) 422-6706. develop the inertial-confinement fusion capability so that we can accurately (IeF) approach to controlled specify both the laser facility and the thermonuclear power for both military targets we will need to achieve our goal and civilian (i.e., power-production) of producing high- implosions. applications. In IeF, an intense energy Target-physics experiments serve source, such as a laser, is used to two purposes. First, they determine the implode a capsule containing constraints on and targets intended thermonuclear fuel (e.g., a mixture of to produce high-gain implosions. Second, deuterium and tritium) to high densities they test our understanding of the low­ and temperatures. Weapon-physics gain implosions we can now produce. experiments have already been To do these novel experiments, we have conducted with the , and had to develop many new diagnostics more are planned with the laser and experimental techniques.} Figure 1 facility. However, full realization of the illustrates the variety of diagnostic military and civilian IeF applications instruments arrayed around a single reF requires, first, the attainment of well­ experiment. Our experiments with diagnosed, high-gain performance of Novette2 produced more information thermonuclear capsules. per laser shot than any previous reF To achieve such high-gain implosions experiments, helping us to make rapid (in which the energy released by progress in several areas of target implosion and fusion of the fuel is many physics. times the laser input energy), we must Our target-physics studies can be advance the state of the art in laser roughly categorized either as laser­ construction, target fabrication, computer interaction experiments or as modeling, and target-physics implosion-physics experiments. In the experiments. We must learn to build laser-plasma interaction experiments, we increasingly powerful lasers at an explore the phenomena that occur as the affordable cost, develop new materials laser light interacts with the target and fabrication and assembly techniques plasma (e.g., absorption and hot-electron

35 production). This information, in turn, targets with one or two beams of green is used to define laser and target (0.53-.um) or ultraviolet (O .26-.um) light, specifications. For example, a poor choice frequency-converted from the laser's of laser and target can produce too many fundamental (1.0S-.um) hot electrons, which prevent high-gain wavelength. At the end of the laser implosions. Laser-plasma interaction amplifier chain, each 74-cm-diameter experiments in which we study hot­ beam was focused by a large lens to electron production enable us to avert strike a spot on the target less than a few this problem by better target design. millimetres in diameter. The concentrated In implosion-physics experiments, we laser light had an intensity greater than focus on understanding the physics of 10 14 W/cm2 (more than a million billion imploding targets. The detailed modeling times the intensity of sunlight at the required in these experiments is done earth's surface). with the the LASNEX computer code.3 The high temperatures produced by Well-diagnosed implosion experiments such a beam cause the surface of the enable us to compare predicted and target to ablate, and the electrons in the measured performance. The models can heated material are torn from their atoms then be improved against the deficiencies to form a fully ionized gas called a exposed and further tested. These two plasma. The laser light absorbed by the types of experiments-laser-plasma plasma heats it to so high a temperature interaction and implosion physics-thus that it radiates soft x rays. The soft x rays provide a bootstrap process, in which strike the capsule and vaporize its progress in one area supports and surface. The vaporized material blows off Fig. 1 facilitates progress in other areas. equally in all directions at high velocity. The array of diagnostic instruments To understand why such experiments This violent outward flow of material used in Novette target-physics experi­ are necessary, we must look at the produces an equal and opposite reaction, ments. These instruments imaged sequence of events that produce a fusion causing a rocket-like inward thrust that the source and measured the timing reaction. In meaningful laser-fusion compresses the fuel to high density, and spectral composition of the light and x rays from the target, as well experiments, the laser pulse must strike simultaneously heating at least part of it as the yield and the neutron an appropriate target at the right to the ignition temperature of the fusion activation. location. Our Novette laser irradiated reaction. If the inertia of the burning fuel confines it long enough (hence the term Radiochemistry ~ inertial confinement fusion), there will be D a net production of energy. o Arrays of photodiodes There are several villains in this piece, Neutron one of which is the high-energy yield electrons generated in the plasma by <' Raman scattering and other processes. o These hot electrons have enough energy to penetrate deep into the fuel, depositing energy and heating it. This is exactly what we want to avoid. Hot fuel is harder to compress than cold fuel, so ~ t ======:J prematurely heated targets never reach Ta ..etl t:J the densities they would otherwise theoretically attain. It has become clear that one of the major keys to the ) Light collectors development of high-gain implosions is ¢:) for optical to suppress these high-energy electrons. ~ We used the Novette laser to B investigate several critical aspects of this o o ~\~OPtical/x_ray complex process: the interaction of the streak camera Dante soft-x-ray j ~ laser light with the plasma, the efficiency spectrometers J X-ray microscope of soft-x-ray production, and the compression of the D-T fuel to high Cassegrain Leake. "y,tal X-

36 INERTIAL FUSION

ser-Plasma Interaction Raman scattering was insignificant Experiments because the scattered light wave and the UIn laser-plasma interaction plasma wave could not become very experiments, we try to determine what intense before escaping the resonant happens when a high-intensity laser region. Because the Novette laser was so beam penetrates and interacts with a much more powerful, we could spread plasma. The mission of such experiments its beam over a larger target area and for ICF is to identify the processes that still maintain the same intensity as in the affect high-gain targets and to discover Argus experiments. This produced larger, how they vary as laser and plasma more planar plasmas that gave these parameters change. waves room to grow more intense before This mission has three parts. First, escaping. The plasmas in fu ture, high­ we must produce and diagnose plasma gain experiments will be even larger. conditions that are relevant both to current target-physics experiments and Raman Scattering in Large to future high-gain targets. Second, we Plasmas must identify both those processes that To simulate these future, larger can scatter the laser light and keep it plasmas, we used thin foil targets made from being absorbed in the target and of plastic or gold. These targets exploded those that convert the laser light into hot early in the laser pulse and expanded electrons. Finally, we must determine the both toward and away from the laser to severity of adverse effects and, if produce a very large plasma. Figure 2 is necessary, mitigate them. The study of a time-resolved image of the ultraviolet Raman scattering illustrates all three of bremsstrahlung emission from such a these. target, color-coded by computer The most desired process for ICF by processing to show the variation in which laser light is absorbed in a plasma intensity of the emissions as a function is collisional absorption. In this process, of time (horizontal axis) and distance the incident light wave causes the from the target plane (vertical axis). electrons in the plasma to oscillate. The Figure 3 compares data from a oscillating electrons then collide with the large number of such shots with the ions in the plasma, absorbing the energy predictions of the LASNEX computer of the light. code. Figure 3a shows the density along In contrast, Raman scattering converts the axis of the plasma, at the center of the incident light into two waves. One the I-ns laser pulse. The gray area shows of these is an electron plasma wave, the behavior predicted by LASNEX, and in which undulations in the electron the solid circles show the density profiles density travel through the plasma. The energy in an electron plasma wave is often converted into (unwanted) hot electrons. The second kind of wave is a scattered light wave, which escapes and carries its energy out of the plasma. Raman scattering, thus, prevents the laser light from being absorbed gradually through electron collisional absorption Fig. 2 and also produces harmful hot electrons.4 A thin plastic film , irradiated with green Raman scattering, like several other (O.S3-jlm) light from the Novette laser, explodes to produce a large, uniform interaction processes, occurs only at a plasma. This streak-camera image was specific location in the plasma (a made with the ultraviolet light emitted resonant region) where the laser light by the plasma. False color was added wave, the scattered light wave, and the by computer processing to indicate the plasma wave can properly couple to one range of intensities. The resulting im­ age shows the plasma expanding on another. In past experiments using the both sides of the target plane (vertical , which produced small, axis) as the laser pulse continues spherically expanding plasmas, the (horizontal axis).

37 obtained from the bremsstrahlung x-ray spectrum. Figure 5 shows that the emission. inferred hot-electron fraction for thick, Figure 3b shows the decrease of the flat target experiments with Novette foil density throughout the laser pulse, as correlates very well with the Raman determined by measuring cutoff times of light fraction, and hence with the optical emissions from the plasma at corresponding energy in the electron different wavelengths with a streak plasma waves. These observations camera. The emission at a given suggest that no other process is a wavelength stops when the plasma significant source of hot electrons under density drops below a certain value. The these conditions. gray area again shows the LASNEX prediction, which agrees well enough Collisional Plasmas Reduce with the data to be used to design future Raman Scattering experiments of this type. Both theory and computer simulation Because the resonant region was suggest that Raman scattering can be much larger, we expected the plasmas reduced by collisions in the plasma, produced using exploding foils to exhibit which become more frequent when the increased Raman scattering. Figure 4 plasma is denser, cooler, or contains shows the fraction of the incident laser more highly charged ions. To create energy scattered as Raman light, plotted such a collisional plasma, we used the as a function of the plasma scalelength Novette laser to irradiate exploding gold (which corresponds roughly to the size foils with ultraviolet (0.26-.um) light, of the plasma). Experiments with the which is absorbed at higher densities Argus laser using thick, flat targets and produces lower temperatures than produced comparatively small plasmas green light of the same intensity.5 and scattered only 0.01 % of the incident Furthermore, the resulting gold ions are laser energy as Raman light. However, more highly charged than the carbon the exploding foils used in the Novette and hydrogen ions that come from experiments scattered more than 10%. plastic foils. Bremsstrahlung emission The thick, flat targets irradiated by measurements showed that these Novette were in between. plasmas were as large as the other When one of these thick targets is hit exploding-foil plasmas, so that plasma by the hot electrons produced by Raman size was not a factor in limiting the scattering, it emits high-energy x rays. Raman scattering. We can measure these x rays with an Figure 6 shows the results of this x-ray spectrometer and infer the number experiment. In all of the other and temperature of the hot electrons experiments, more than 1% of the laser from the magnitude and slope of the light was dissipated by Raman scattering.

Fig. 3 21 10 r------~ 1022 ,.-----,.... ------, Experiments with thin, exploding-foil (a) (b) targets test the predictive ability of our computer modeling. (a) The density LASNEX (spatial profile) of the plasma produced by irradiating a 2-JlI1l-thick plastic film with 3.6 kJ of green light, measured '", half way through the 1-ns pulse. Red lines are optical-pyrometer measure­ ments, and the laser beam arrived from • the right. (b) Dependence of the peak density of the same plasma on time, measured by the extinction of light of Laser direction different wavelengths (solid circles) ,. and the shape of the Raman spectrum (red lines). The predictions of the LASNEX computer code (gray shaded 1019L-----~------~------~----~ 1020L-____~ ______~ ______~ ____~ areas) are in reasonable agreement - 1.0 - 0.5 0 0.5 1.0 - 1.0 - 0.5 o 0.5 1.0 with the data in both cases. Axial distance, mm Time from center of pulse, ns

38 INERTIAL FUSION

However, in this collisional plasma, the 100r------~ Fig. 4 Raman fraction was less than 0.001 %. The proportion of Raman scattering ob­ This demonstrates that we can eliminate served as a function of plasma size. In Raman scattering if we make the plasma Novette foils • a small plasma, such as those pro­ 10 collisional enough. duced on the previous Argus laser

;!!.0 (red), the Raman light escapes the c plasma before growing to large ampli­ Brillouin Scattering 0 :e tudes. Hundreds of times as much laser Another quantitatively important energy was dissipated by Raman scat­ ~ Novette disks process we studied in these same c:: tering in the exploding foil experiments til experiments is Brillouin scattering, which E on Novette (green). til reflects up to about 30% of the incident a: green laser light from the plasma. 0.1 Although this is a significant loss, the important result is that we observed about the same level of Brillouin scattering with the small plasmas produced by Argus and with the large Density scalelength, LfA exploding-foil plasmas produced by Novette. This suggests that Brillouin scattering may become no worse in the 10 - 1 r------~ very large plasmas needed for high-gain Fig. 5 implosions. The observed hot-electron fraction, determined from the x rays produced when the electrons stop in the target, oft X-Ray Production 10 - 2 plotted as a function of the fraction of In the indirect-drive targets with c:: the laser energy scattered as Raman which we expect to achieve high­ .2 S ti light. The line represents the fraction of gain implosions, laser energy will be ~ the laser energy converted into Raman c:: plasma waves, which correlates well transferred to the fuel capsule by a 10- 3 g with the hot-electron fraction. This cor­ strong flux of soft x rays. Thus, it is 0 CI) relation indicates that in experiments Qj important for the designer of a high-gain using thick, flat targets, the dominant ~ system to know, in detail, how the :z: source of hot electrons is Raman conversion efficiency varies as the laser 10 - 4 scattering. wavelength or intensity is changed. In Novette experiments using gold • disk targets, we found that the conversion efficiency from laser energy 10 - 5 to soft-x-ray energy depends on laser 10 - 4 10 - 3 10 - 2 10 - 1 intensity, laser wavelegth, and target Raman scattered light fraction materia1. 6 In these experiments, we placed a soft-x-ray spectrometer at 30 deg from the plane of the target to measure Fig. 6 the flux of soft x rays. From this 10 - 1 The effect of collisional damping on ~ .! Raman scattering in large plasmas pro­ measured flux, we inferred the total soft­ c:: x-ray energy based on plausible angular 0 duced by irradiating a thin foil of either ti ! gold or plastic. (The red point is for distributions of such emissions and on ~ 10 - 2 ~ Shiva experiments with 1.0S-pm light, more detailed data from Argus 1: the green circles are for Novette ex­ experiments. ~ periments with 0.53-pm light, and the "C black circles are for Novette experi­ Figure 7 shows the resulting soft-x-ray I!! 3 ! 10- ~ ments with 0.2S-pm light. Bars on the conversion efficiency from gold targets, til . 0.2Spm 0 data points indicate the range of values III (Novette) as a function of laser intensity, and C observed on several shots.) Ultraviolet includes data obtained with the Novette til • 0.53 pm (0.2S-pm) light irradiating a gold foil E 10 - 4 ~ and Argus lasers. We are still a:til (Novette) produces a cooler, more dense, and investigating the reasons for the • 1.0S pm more highly ionized plasma than the differences between the Novette data (Shiva) other experiments. As a result, colli­ 10 - 5 I f sions in the plasma were frequent and the Argus data for irradiations with Plastic Gold enough to render the Raman scattering green (0.53-.um) light. Atomic number of target undetectable.

39 In general, it appears that we can measure the number of enhance the production of soft x rays by produced by the fusion reaction and using light of shorter wavelengths and radiochemistry techniques to measure the lower intensities. In particular, in the amount of radioactivity induced in the Novette experiments, as much as 80% of fuel-capsule shell by the neutrons from the 0.26-.um light was converted into soft the target. From the radiochemistry x rays, a result that may find uses in measurement, we infer the compression other research in addition to ICF. of the shell. Figure 8 presents the results Fortunately, the same trends that in terms of compression (the ratio of increase the soft-x-ray production­ final to initial pusher mass per unit area) shorter laser wavelength and lower as a function of neutron yield. intensity- also act to reduce the plasma As Fig. 8 shows, it is possible to obtain instabilities such as Raman scattering. a high neutron yield with very little compression. This occurs because, at the mplosion Experiments subkilovolt temperatures achieved in Computer modeling indicates that to these experiments, the neutron yield is I obtain high gain in energy by ICF, proportional to roughly the sixth power the capsule implosion must compress the (the cube of the square) of the ion fuel to densities of about 200 g/cm3 temperature, but only to the two-thirds (more than 1000 times the density of power of the density. The fuel in many Fig. 7 liquid hydrogen). Previous experiments of the Shiva capsules became hot early The efficiency of conversion from laser with the , comparable to those in the implosion and produced many light to soft x rays as a function of the performed with Novette, used infrared neutrons, but these high temperatures intensity of the laser light on target. (1.06-.um) light. The Shiva experiments prevented it from being compressed to We represent the various wavelengths achieved fuel densities of about 10 g/cm3. high densities. When we modified the used in these experiments by red for infrared (1 .06 JIm), green for green A major goal of our Novette implosion Shiva experiment design to reduce (0.53 JIm), blue for near-ultraviolet experiments was to achieve higher preheat, we managed to obtain more (0.35 jim), and black for far-ultraviolet densities. The key to this process is to compression but at a lower neutron (0.26 jim). Argus experiments are rep­ produce a gentle, symmetrical ablative yield. resented with open symbols, Novette implosion without hot-electron The first experiments with the with closed symbols. We obtained very large conversion efficiencies in the preheating. Novette laser produced nearly as many Novette experiments using 0.26-Jlm To monitor the performance of these neutrons and significantly higher light. targets, we used neutron counters to compression than the final comparable experiments with Shiva. The fuel density 100.------~------~ inferred from the Novette data is about 100 times the density of liquid deuterium, or about twice the density inferred from these Shiva experiments. 80 We attribute the difference in

0~ compression to the negligible target ::.:; Co) preheat by fast electrons achieved in c: CIl the Novette experiments. Ti 60 The Shiva results were obtained in the CIl ==c: course of several years of experiments to 0 'iii optimize implosion conditions. The lack iii >c: of optimization in the Novette 0 40 Co) experiments is indicated by the neutron >- l!! yield, which corresponds to an ion X Argus Novette temperature of only 500 to 600 eV (about 1.06 jim 20 0 5 million degrees). Computer modeling, 0.53j1m 0 • based on the measured driving 0.35j1m 0 conditions and the assumption that the 0.26 jim • implosion was stable and symmetrical,

OL------~------~------~ indicates that the temperature (and the 1013 1014 1015 1016 neutron yield) should have been larger. Laser intensity, W/cm2 Follow-up experiments with the Nova

40 INERTIAL FUSION

laser system will allow us to investigate 120 Fig. 8 the causes of this discrepancy. 0.53 pm The results of radiochemistry experi­ 100 ~ (Novette) ments that measure the compression onclusion ~ _ 1.06 pm of the fuel-capsule shell, plotted as a c:: The target-physics ICF 0 80 (Shiva) function of the observed neutron yield ·iii for a number of similar experiments experiments with the Novette fJI C ~ with Shiva and Novette. The targets ir­ laser have advanced our knowledge in c. E radiated with green light by Novette 0 60 several areas and bode well for the 0 (green circles) produced almost as future. We have produced large plasmas, 4) many neutrons as the Shiva experi­ ~ relevant to future experiments even fJI 40 ments (red circles), and they com­ :;Q) pressed the shell (and the fuel) to much beyond Nova, and have observed that fJI C. higher densities. under such conditions Raman scattering III u 20 can, like Brillouin scattering, interfere significantly with the efficient absorption I 0 .- of the laser light. In addition, we have 6 7 - 10 10 continued to see larger soft-x-ray Neutron yield conversion efficiencies at short wavelengths and low intensities, even in larger plasmas, which augurs well for further implosion experiments. In future short-wavelength laser systems. addition, we will address several new We have traced the production of hot areas of target physics, including the electrons in large plasmas on flat targets growth of hydrodynamic instabilities and to the Raman scattering process, and we the development of higher ablation have shown that this scattering can be pressures with pulse shaping. ill: dramatically reduced by making the plasma more collisional. Thus, we have at least one way to reduce Raman scattering in high-gain systems. Finally, the first implosion experiments with Key Words: ablation pressure; Brillouin scattering; Novette produced greater fuel collisional plasma; computer code-LASNEX; hot­ compression than the final experiments electron preheat; inertial confinement fu sion (I C F); made with the longer-wavelength Shiva laser-Argus, Janus, Nova, Novette, Shiva; Raman scattering; soft x rays. laser. Thus, the Novette laser continued to demonstrate that targets irradiated with short-wavelength (0.26- to 0.53-.um) lasers produce significantly better results Notes and References in many respects than those irradiated 1. H. C. Ahlstrom, "Diagnostics of Experiments with longer wavelengths (1 to 10 .urn). In on Laser Fusion Targets at LLN L," in Physics particular, the hot-electron production is of Laser Fusion, Vol. 2, Lawrence Li vermore smaller, the energy absorption is higher, National Laboratory, Rept. UCRL-53106 the x-ray conversion efficiency is higher, (1982). and greater fuel compression can be 2. The Novette facility was described in Energy and Technology Review (UCRL-52000-85-1), obtained. Current projections to high­ January 1985, p. 10. gain systems indicate that we may even 3. G. B. Zimmerman and W. L. Kruer, Comments be able to use a wavelength as long as Plasma Phys. Controlled Fusion 2, 51 (1 975). 0.53.um, rather than requiring an even 4. R. P. Drake et al., Ph ys. Rev. Lett. 53, 1739 shorter wavelength of 0.35 or 0.26.um. (1984). 5. R. E. Turner et al. , Phys. Re v. Lett. 54, 189 The Nova laser, a more powerful (1985). and versatile experimental facility just 6. R. L. Ka uffman et al., "X-Ray Conversion coming into operation, will help answer Effi ciency," Laser Program Annual Report, this and other questions. We will use Sec. 5.2.8, Lawrence Li vermore National Nova to continue our studies of laser­ Laboratory, Rept. UCRL-50021-84 (1984). 7. F. Ze et aI. , Wavelength Scaling of Laser­ plasma interactions and their Driven Implosions, Lawrence Li vermore consequences and of soft-x-ray Nati onal Laboratory, Rept. UCRL-91087 production and transport, and to perform (1984).

41