DEFENSE PROGRAMS

The Nova laser fusion research Jacility, currently under construc­ The Nova tion at LINL, will provide researchers with powerful new tools Jor the study oj nuclear weapons physics and inertial confinement fusion. laser fusion The Nova laser system consists oj 20 large (74-cm diam) beams, facility Jocused and aligned precisely so that their combined energy is brought to bear Jor a fraction oj a second on a tiny target containing thermonuclear fuel (deuterium and tritium). The ultimate goal oj the LINL inertial confinement fusion program is to produce fusion microexplosions that release several hundred times the energy that the laser delivers to the target. Such an achievement would make in­ ertial confinement fusion attractive Jor military and civilian applica­ tions. The Department oj Energy (DOE) has approved construction oj 10 Nova laser beams and the associated laboratory buildings (Phase 1). For further information contact Phase 2, which includes the remaining 10 beams and harmonic­ William W. Simmons (422-0681). conversion crystal arrays, is presently under consideration by DOE. By the mid- to late 1980s, Nova should demonstrate that we can produce the extremes oj heat and pressure required to achieve igni­ tion oj the thermonuclear fuel. Additional developments in the area oj high-efficiency drivers and reactor systems may make inertial con­ finement fusion attractive Jor commercial power production.

One technical approach to the times liquid 0-T density), ignition problem of controlling thermo­ of thermonuclear bum, scientific nuclear fusion reactions is to bring breakeven (in which the energy in­ small deuterium-tritium (0-T) fuel cident on the target equals the ther­ pellets to very high temperatures monuclear energy released), and and densities in such a short time significant energy gain in an ICF ex­ that the thermonuclear fuel will periment. These achievements are ignite and bum before the com­ necessary precursors to the suc­ pressed core disassembles. This ap­ cessful realization of either the proach, known as inertial confine­ military or the energy objectives of ment, relies upon a driver (e.g., a the program. From a technical laser) to deliver the extremely high­ point of view, the ignition milestone power, short-duration burst of is the most important, and its at­ energy required. tainment will provide a strong LLNL has been involved in the signal to move forward national inertial confinement fusion aggressively. (ICF) program since the early We chose solid-state 1960s. The objectives of the neodymium-glass lasers for our ex­ program have been twofold: in the perimental driver systems because short term to develop the military applications of ICF, and in the long term to evaluate ICF as an energy source. 1,2 Our immediate scientific objectives are the demonstration of high compression (100 to 1000

1 that particular technology was most review by DOE, an additional 10 we can double or triple the fre­ advanced at the time the decision Nova beams will be installed in the quency of the basic 1.05-/-Lm light was made. In separate programs, Shiva building. The full Nova com­ from high-power Nd-glass lasers we have continued to develop ad­ plement of 20 beams will be with efficiencies exceeding 70%. vanced laser concepts with the goal brought to an integrated target Since shorter wavelengths are of reducing system costs and chamber in two opposed clusters of much more favorable for ICF laser providing high-efficiency, high­ 10 beams each. Other target con­ target physics, we plan to imple­ repetition-rate drivers. Over the figurations are possible if required. ment this option in the Nova past several years, we have built Like its predecessors, Nova will facility. (This enhancement of the and operated a series of in­ be a neodymium-glass system Nova facility is presently under creasingly powerful and energetic with a working wavelength near review by DOE and will be deter­ laser systems to study the physics of 1.05 /-Lm. However, in recent ex­ mined shortly.) We would thus be ICF targets and laser-plasma in­ periments on the Argus laser able to focus approximately 200 kJ teractions. Nova, the latest in this system, we have demonstrated that of green (O.53/-Lm) or blue series, is the successor to the Argus by using the nonlinear optical (O.35/-Lm) light onto laser fusion and Shiva lasers. The Nova laser properties of potassium dihy­ targets. The total cost of the Nova will consist of 20 beams, capable of drogen phosphate (KDP) crystals, project will be $195 million if we concentrating 200 to 300 kJ of energy (in 3 ns) and 200 to 3001W of power (in 100 ps) on experimental targets by the mid- 1980s. Its purpose is to demon­ strate the ignition of thermonuclear burn. The first phase of the Nova proj­ ect involves construction of a 10 684-m 2 ( 115,000-ft2) laboratory building and installation of a 10-beam neodymium-glass laser system adjacent to the existing Shiva facility (see Fig. 1). With Phase 2 funding, presently under

lFD(Qlo II The Nova laserfusion research facility, currently underconstruc- c:2) tion at LLNL, will provide researchers with powerful new toolsfor the study of nuclear weapons physics and inertial confinement fusion. Its 20 large (74-cm diam) beams, focused, timed, and aligned precisely, will concen­ trate 200 to 300 kJ of energy for a few billionths of a second on a tiny pellet con­ taining thermonuclear fuel. Nova is being built to demonstrate the ignition of advanced thermonuclear-fuel targets in the mid-1980s.

2 DEFENSE PROGRAMS

are limited to the 1.05-,um the fuel to twice the no-bum tem­ into 20 beams. After traversing an wavelength and close to perature. If our experiments are adjustable optical delay path (used $238 million if we develop the successful, it is also possible that to synchronize the arrival of the green or blue light capability. the Nova facility will achieve scien­ various beams at the target), the The light from Nova, particularly tific breakeven (in which the energy pulse enters the amplifier chain, if it is of the green or blue type, released by the thermonuclear where (1) disk amplifiers increase should be capable of producing the reaction equals the laser energy the pulse power and energy, (2) extremes of temperature and delivered to the target). spatial filters maintain the spatial pressure required to ignite a small smoothness of the beam profile pellet of thermonuclear fuel. That Laser design and while expanding its diameter, and is, it should create conditions such performance (3) isolators prevent the entire laser that the energy released from alpha The Nova laser system has from breaking spontaneously into particles in the central core of the master-oscillator-power-amplifier oscillations that could drain its compressed fuel (accounting for (MOPA) architecture. As shown in stored energy and damage the 3.5 MeV of the 14.6-MeV total Fig. 2, a laser pulse of requisite target prematurely. (For a discus­ energy released by the 0-T reac­ temporal shape is generated by the sion of spatial filtering and isolation tion) is trapped and used to heat oscillator, preamplified, and split processes, see the box on pg. 4.)

Output sensor . .. Spatial filters package • Amplifiers +.L Isolation D Space for added amplifiers 46.0-cm diam '\. / Turning mirrors

cm diam equal ization 15.0-cm diam

20.8- cm diam

[pO 0 ~ Schematic diagram showing the major op- in cm) is expanded in increments by spatial filters judiciously lJ@o tical components through which each of 20 spaced in the chain. The large-output sections of the chain Nova laser beams will pass on its way from the oscillator to (31.5, 46.0, and 74.0- cm in diameter) will require significant the target. The overall Nova system architecture is similar extensions of (and some innovation from) Shiva optical to that of the Argus and Shiva solid-state lasers (currently technology. The plasma shutter, which isolates the laser operational at LLNL). A collimated laser pulse traverses from potentially damaging reflections from the target, will progressively larger amplifier sections between the os­ be used for the first time. cillator and the target. The beam diameter (expressed above

3 Spatial filtering and isolation

A spatial filter consists of a pair of lenses plasma). The plasma then expands to fill the hole (positioned so that the distance between them is through which the pulse is passing. We refer to this equal to the sum of their focal lengths) and a small phenomenon as "self-closure." No closure occurs if pinhole at their common focus. Spatial filters are the laser pulse is short compared to the pinhole used in the laser chain to increase the diameter of closure time. However, for long pulses, the intensity the beam and to remove intensity variations (Le., of the beam focal spot at the pinhole edge must be lighter or darker regions of the beam produced by less than the threshold for plasma formation, or the optical imperfections and self-focusing effects). The pinhole must be much larger than the product of the input lens of a spatial filter focuses the main beam pulse duration and plasma velocity. For Nova, f/20 (all parallel light) to a tiny spot that can pass through spatial filters satiSfy all imaging, filtering, and pinhole a correctly positioned pinhole without being clipped self-closure criteria. (The f number describes the on the edges. The nonparallel light, which is respon­ ratio of the focal length to the beam diameter.) sible for intensity variations, hits the edges of the Isolators act as optical diodes in the laser chain, pinhole and is stopped. permitting light to pass in one direction, but not in A spatial filter projects a real image of the input the other. They are used to prevent back-reflected beam and minimizes the undesirable effects of dif­ laser light from reentering the amplifier stages. This fraction rings. The primary purpose of this imaging is goal is achieved either by the Faraday effect to maintain a high filling factor (defined as the ratio (wherein the polarization plane of transmitted light is of the integral of the beam energy density over its rotated by an amount proportional to the strength of spatial profile to the energy in a flat beam) through a magnetic field applied along the direction of the progressively larger amplifier sections. propagation), or by the Pockels effect (wherein the The optimum length for a spatial filter is deter­ polarization is "rotated" by an amount proportional mined by its imaging properties and is independent to the applied voltage). of the diameter of the filtering pinhole. Optimum The Faraday effect is particularly attractive use of amplifier apertures is achieved if the input because of its scalability to large diameters. It can be lens of one filter is imaged onto the input lens of the used to rotate forward-propagating, linearly next filter in the chain. Simple formulas for a two­ polarized light 45° in a right-hand sense with respect lens system tell us that the length of the filter is equal to the direction of the magnetic field. Thus, the to LN/M, where LNis the length of the optical­ plane of polarization of any back-reflected light will component train between the filter in question and be perpendicular to that of the initial beam and can the one follOwing it, and M is the magnification (Le., be rejected by a polarizer. The Faraday effect occurs the ratio of the diameter of the beam exiting the filter in almost all glasses and is particularly strong in to the diameter of the beam entering it). Thus, the paramagnetic glasses. Nova will use terbium-doped optimum length of a spatial filter is uniquely deter­ silicate glass manufactured by Hoya Glass Works. mined by the diameter of the components on either The Pockels effect occurs in non-centrosymmetric side of it and the length of the optical train between crystals such as potaSSium dihydrogen phosphate it and the next filter. (KDP). One achieves isolation by rotating the beam A second factor that must be considered in deter­ polarization (by applying the appropriate voltage) for mining the spatial-filter length is plasma formation in the requisite time period (approximately 10 ns), thus the pinhole. Pinholes intercept some beam energy allOwing the pulse through the crossed polarizers. while performing their filtering function. If the Light arriving at the Pockels crystal before and after amount of energy is too large, material around the this lOons interval from either direction is rejected by edges of the pinhole is ablated (Le., converted to the polarizers.

4 DEFENSE PROGRAMS

The beam is collimated between located between the final laser am­ plified in the process), and it might spatial filters in the laser chain. plifier and the optical turning destroy some of the optical compo­ Thus, each of the components in a mirrors. nents. particular section has the same When the pulse exits from the In each section, the beam is am­ diameter. In the 4.0-cm section final beam-expanding spatial filter, plified to the damage threshold of (see Fig. 2), the amplifier is a single it has been amplified to an energy the lenses at a maximum energy glass rod, and the isolator is an level of 10 to 15 kJ, and its output for a specified pulse dura­ electro-optic (Pockels) cell crystal diameter is 74 cm. Turning mirrors tion. This "isofluence" design max­ placed between crossed polarizers. direct the beam to the target cham­ imizes the energy output per unit This cell operates as a fast (10 ns) ber, where focusing lenses concen­ cost, while keeping the chain as a optical gate, preventing interchain trate it on the target. The first of the whole below the component oscillations and at the same time turning mirrors is partially trans­ damage limit. The f1uence (energy reducing to tolerable levels un­ parent, allOwing approximately 2% per unit area) at which optical com­ wanted amplified spontaneous of the pulse to enter the output ponents suffer damage has been emission (i.e., radiation at the laser sensor package. This unit senses the subject of intense investigation wavelength, amplified by passage and reports on the alignment throughout the country. through the chain, which can strike status, energy and power, spatial On the basis of extensive mea­ and damage the target before the quality, and other characteristics of surements, we have established laser pulse arrives). In all larger the beam. The plasma shutter, numerical damage thresholds for diameter sections, the amplifiers located at the focal position of the components representative of consist of face-pumped disks set at final spatial filter, protects the laser current surface-preparation Brewster's angle to the passing by preventing light reflected from technology (Table 1). It is apparent beam. Polarization-rotating the target from reaching the laser that uncoated surfaces will tolerate isolators (relying upon the Faraday amplifiers. In the absence of this higher fluences than antireflection­ effect) assure interchain isolation. protection, such light would travel coated (AR-coated) surfaces. Thus, Substantial savings will be realized back down the chain (being am- spatial filter input lenses, which are by reusing most Shiva laser compo­ nents in Nova. 1 (Nova components Surface damage threshold for anti- up to 21 cm in diameter are vir­ 1f~@il® 11 reflection-coated and uncoated optical tually identical to Shiva compo­ surfaces (Nova performance at 12 kJ). nents.) Optimum spatial filter design Damage threshold, J /cm2 provides entrance-Iens-to-entrance­ 0.1 os 1.0 os 3.0 os lens imaging. Thus, smooth beam intensity (through the cross­ Coated surfaces 2.5 8.0 8.5 sectional area) is projected along Bare polished surfaces 6.0 19 33 the chain, and energy extraction by the laser pulse is maximized. The beam has a filling factor of more than 70% (defined as the ratio of the integral of the beam energy density over its spatial profile to the energy in a flat beam). Pinholes (located at the spatial filter focal planes) are large enough to avoid self-closure (due to the ablation of material around the pinhole edge) during the passage of the pulse. The longest spatial filter in the Nova chain is 23 m (75 ft) long. It is

5 subject to the highest peak (0.1 to 3 ns) for the baseline Nova curve, which represents an upper fluences, are left uncoated. chain. Results are summarized in limit on performance of the laser at Figure 3 gives an example of ex­ Fig. 4. The upper curve represents any pulse duration. In the case of pected performance at the damage hypothetical single-chain perfor­ long pulses, nonlinear effects limits shown in Table 1 (for a mance at the first-component-to­ become less important. Therefore, nominal I-ns pulse). The location damage limit if a perfectly smooth the ratio of input to output peak of each spatial filter lens in the beam with a 70% filling factor were fluences at the spatial filters ap­ chain is shown at the top of the to pass through it. However, real proaches the geometric expansion figure. The solid line represents the beams exhibit spatial modulation as ratio. Since this ratio is less than the peak fluence at each lens, and the a result of imperfections encoun­ ratio of coated to uncoated data points represent the average tered upon passage through optical damage thresholds (Table 1), the fluence at each lens. Both peak components. Therefore, realistic AR-coated output lens tends to be and average fluences increase as performance levels must be set on the limiting element for long pulse the pulse passes through the the basis of the peak/average ratio performance. In the case of very various amplifier sec.tions. The of the beam at the location of the short pulses, nonlinear growth of average fluence grows as a result of threatened component. Computer spatial peaks along the propaga­ amplification of the pulse energy. estimates thus lead to the lower tion path from the output of the last However, the peak fluence grows faster because it is also affected by nonlinear self-focusing and spatial noise sources (discussed below). Beam-expanding spatial filters 20 reduce both the average fluence and the peak/average ratio. Thus, the spatial-filter output lenses and N the target lens are AR-coated. The 515 design laser can focus up to 9.4 kJ -----, • Average fluence I - Peak fluence on the target in a I-ns pulse before Q) U any lens is threatened. t: Calculations such as those ~ 10 LL described above have been per­ formed over the temporal range for coated exit lenses 5 •

50 100 150 Distance along laser chain - m

~g C[j) Interchain peak fluence as a function of distance along the laser U l.l@o cQ) chain for a 9A-kJ, 1-ns laser pulse. The location of spatial-filter lenses is indicated along the top of the figure. These lens surfaces will suffer damage if the peak fluence exceeds the indicated limits. The laser can be operated without damage at or below 9.4 kJ. The dots indicate the average fluence at various pOints in the chain.

6 DEFENSE PROGRAMS

spatial filter could damage the final compares code simulations with ac­ plane of each spatial-filter entrance focusing lens. tual chain performance. lens. We expect that significant MAlAPROP sequentially The full two-dimensional developments in surface tech­ calculates the evolution of a 512 X 512 mesh version of nology stemming from ongoing spatially profiled pulse as it passes MAlAPROP is used to assess non­ research and development will lead through the various elements of a linear spatial noise growth and to higher component-damage laser chain. Spatial filtering and pinhole filtering realistically for thresholds. Graded-index surfaces free space propagation between beam diameters as large as those look very promising. 3 Also, elements are modeled by means of required for Nova. A longer run­ researchers have demonstrated a two-dimensional fast Fourier ning version of MAlAPROP is significant increase in the damage transform algorithms. Nonlinear available to model the effect of am­ threshold of bare surfaces "bur­ media (e.g., glass) are represented plifier gain saturation on a local nished" with a directed C02 laser as phase transformations. In the scale throughout the laser chain. beam. 4 Finally, there is evidence model, all glass components are (This effect is important in the case that damage thresholds are higher treated as "thin." Pinholes in spatial of temporally long pulses.) An ex­ for AR coatings applied to "super­ filters are modeled by aperture ample of this more detailed calcula­ polished" surfaces (i.e., those truncation in the Fourier transform tion is shown in Fig. 6. Experimental whose substrate has been finished via a special process). Our greatest short-term hopes for optical com­ / / ponents suitable for damage-free / / / / laser operation at and above the 0.1 ns / / / 1 ns fluence levels indicated in Fig. 4 are s: / / Peak/ based on these three develop­ l- / / I / / ments. c: / <0 / / We have allocated space in the .s= / 3 ns ~ 10 31.5- and 46.0-cm sections for ad­ Q) a. / ditional amplifiers. We are planning ... / / Q) / that advances in optical-surface o~ a... / / technologies such as those men­ / / Damage tioned above will make it possible / / threshold to install amplifiers there that will / / / / Performance increase overall Nova performance / / at very modest cost. The computations illustrated in 10 100 Figs. 3 and 4 were performed using Energy per chain - kJ the fast Fourier transform code ~g®o 'i1l Performance limits over the full temporal MAlAPROP, which (a) accu~ately J.J ll(Q) ~ design range of Nova (0.1 to 3 ns). If the simulates detailed beam propaga­ beam were perfectly smooth spatially, and if it filled 70% of the available amplifier aperture, the laser could be tion through spatially filtered am­ operated at points indicated along the upper curve. plifier chains, (b) accounts for the However, small component imperfections introduce performance of the Argus and modulation on the beam. Thus, the laser can only be Shiva systems, and (c) confirms the operated safely at or below the performance curve, which spatial filtering strategy for the is determined by dividing the upper curve by the larger Nova system. The predictive peak/ average ratio. success of this comprehensive code is primarily a result of its spatial noise model, which correlates well with independent statistical obser­ vations of noise sources. Figure 5

7 results with longer pulses {and Quantitative knowledge of these All major laser systems are ul­ split beams} have been amply sup­ parameters is crucial as we try to timately limited by small-scale ported by such calculations. judge the relative merits of different spatial fluctuations, which originate In practice, the code calculates spatial filter staging strategies. In from optical imperfections in the {on the basis of fixed laser chain addition, pinhole diameters are components or on their surfaces. staging and input power} the chosen such that beam intensities We have introduced a spatial­ peak intensity at each lens in the on their perimeters are not greater noise-source model, which ac­ system, the intensity at the perim­ than 1011W/cm2 {which is below curately mimics the effects of eter of the spatial filter pinholes, the threshold for plasma forma­ small scattering sites {approx­ and the intensity distribution in the tion} . Focal spot blurring is also imately 200 J.Lm in diameter} final or target lens focal plane. estimated. located primarily on disk surfaces.

(b) Simulated profile

20

16

N E u ~ l:J 12 I >- (c) Experimental profile '...v; t:

4

4 8 12 16 20 24 Distance - em . IFD@o fID Nova laser design is based upon experimen- diameter section through this contour plot. The modulation (§) tally verified computer codes. A computer- depth and the peak-to-average ratio of (b) are both in ex­ generated intensity contour plot (a) representative of a cellent agreement with the experimental profile (c). The lat­ typical Argus near-field beam (at about 21W) is shown at ter was generated from a microdensitometer scan of an ac­ the left. The simulated beam profile (b) represents a tual beam photograph taken at this power level.

8 DEFENSE PROGRAMS

We set points (or point clusters) tion count statistics (based upon plifier assembly, maintenance, and randomly on the MALAPROP comparison between MAlAPROP operation will be required to computational grid to zero, thus results and Argus and Shiva ex­ achieve this degree of perfection. representing opaque particles or perimental data) is the fractional "obscurations." From careful obscured area per disk surface Components and subsystems laboratory examination of disk am­ (typically 5 X 10-5). We have used Disk amplifiers and other com­ plifiers and other sources, we this numerical parameter in our ponents with apertures of 21 cm or characterize the number and size simulations of Nova performance less are typical of Shiva-based distribution of these obscurations. and as a design criterion for the technology. 1 The larger Nova am­ The parameter that best correlates surface quality of the Nova disks. plifiers will feature a new rec­ system performance with obscura- Scrupulous attention to clean am- tangular internal geometry (as op­ posed to the current cylindrical geometry) that will permit flashlamps to pump the laser disks more efficiently. A side view of a partially assembled 46-cm rec­ tangular amplifier is given in Fig.7a, and an internal view of a 34-cm-aperture amplifier is featured on the cover of this issue. Flashlamps will run lengthwise along two opposing sides of the rectangular case. Each flash lamp will be backed by a silver-plated crenulated reflector, which will reflect light into the disk faces while minimizing absorption by neighbor­ ing flash lamps. We will use an elec­ troform process to manufacture these reflectors for Nova. Flat, silver-plated walls will form the remaining two sides. The cavity is Time - ns very reflective and will provide tight optical coupling of light from .....t::::::.===-====-_.-/ 2.50 flashlamps to disks. Careful design -40 -20 o 20 40 has made possible efficiency Beam diameter - em

IFn(i)l @ Afull computer simulation of a Nova beam profile as it appears 2&0 0 at the target focusing lens. This beam is carrying 12.8 kJ in a pulse of nominal3-ns duration. The profiles shown toward the back of the figure arrive at the target earliest. Peak intensity occurs at the diffraction edge of the beam slot, which itself is caused by the split disks used in the 46-cm Nova am­ plifier stage.

9 improvements of a factor of two portant. For this reason, flat quality of the beam by keeping over the cylindrical amplifiers used transparent glass shields will be thermal disturbances formed in the in the Shiva laser system. used to isolate them from the atmosphere around the f1ashlamps Maintenance of the high surface f1ashlamps. The shields also pre­ from penetrating the optical beam quality of the disks is extremely im- vent serious degradation of the path. We will employ rectangular

Major new compo­ nents for Nova. (a) Partially assembled 46-cm rec­ tangular laser amplifier. Half disks will be inserted in the elliptical holders. Transverse flashlamps and their reflectors appear at the back of the amplifier head. The interior metal parts of the amplifier are in­ sulated to 50 kV to prevent arc-over. Most of the metal parts shown are electroformed nickel; optical cavity interior parts are silver-plated for high reflectivity. (b) Examples of split disks. The split reduces parasitic amplification within the disk, allowing more energy to be stored in a large-aperture amplifier. The (blue) monolithic edge cladding also enhances disk energy storage by reducing internal losses from amplified spontaneous emission. ----;., (c) Turning-mirror blanks made of 1- - __ - borosilicate glass by Schott Optical Company. Mirrors made from these 1.2-m-diameter blanks will be used to direct the Nova beams to the target chamber.

10 DEFENSE PROGRAMS disk amplifiers in the final three am­ At the largest Nova amplifier Naturally, a split disk produces a plifier sections of Nova. Design diameter (46 cm), drastic measures split beam, and although diffraction criteria, most of which have been must be taken to suppress this effects originating at the split can be met in component tests, are sum­ drain. This is the reason why we diminished by careful beam shap­ marized in Table 2. split the disks along their minor ing (apodization) techniques, Phosphate-based glass features diameters. Much higher energy residual edge modulation due to very high intrinsic gain, as well as storage and gain can be realized diffraction still remains (Fig. 6). sufficient energy storage capacity from a 46-cm-diameter disk when it Nevertheless, at beam apertures for the realization of Nova laser is split as shown in Fig. 7b. For greater than about 40 cm, the cost performance goals. Furthermore, it Nova, each disk half will be com­ and performance benefits of split has proven to be manufacturable in pletely surrounded by a disks far outweigh the diffraction large sizes to Nova specifications "monolithic" edge cladding (shown problems. relating to optical quality and in blue) manufactured with glass HistOrically, we have used Fara­ resistance to damage. Fluoro­ that is thermally and mechanically day rotator-polarizer combinations phosphate glass has a lower, more compatible with phosphate-based as isolators to protect the laser from desirable nonlinear refractive index. laser glass. This edge cladding reflected target light. However, For that reason, it is capable of per­ serves two purposes. First, it has the these become very expensive as formance superior to that of same index of refraction as the their size increases (because of the phosphate for pulses less than 1 ns laser glass, so that reflections of the cost of the terbium-based glass and in duration. Unfortunately, internal amplified spontaneous energy storage required). Faraday fluorophosphate has not proven to emission that could reenter the disk isolation at the final amplifier aper­ be manufacturable in large pieces. and undergo further paraSitic am­ ture of Nova would cost approx­ We have therefore selected plification are minimized. Second, imately $800 000 per beam. We phosphate as the Nova laser am­ it strongly absorbs energy at the examined alternative methods, plifier material. laser wavelength (1.05 Jlm) , serving finally settling upon the plasma With disks of larger diameter, the as a "sink" for unwanted energy. shutter. gain path for internally generated These claddings represent a signifi­ In concept, the plasma shutter amplification of spontaneous emis­ cant departure from and an im­ consists of a wire (or foil) metallic sion (ASE) becomes longer. Inter­ provement upon conventional sample closing an electric circuit. nal ASE represents a parasitic drain claddings employed on Shiva and This circuit stands ready until the on the energy stored in each disk. Argus laser disks. optical pulse has passed the

1r(~Jl~n® ~ Design criteria for rectangular Nova disk amplifiers.

Amplifier aperture, em 20.8 31.5 46

Glass type Phosphate Phosphate Phosphate Number of lamps/amplifier head 16 20 80 (transverse) Number of disks/amplifier head 3 2 2 (split) St ored energy (nominal), kJ 300 375 600 Small signal gain/head (nominal) 2.3 1.78 1.98

11 pinhole at the final spatial filter. At pinhole. To be useful for system ap­ than the equivalent Faraday that instant, an electrical surge plications, it must work reliably for rotator-polarizer combination and large enough to sublimate the foil is many shots. This has been accom­ is also optically lossless. Perfor­ applied. This creates a plasma jet, plished conceptually with a novel mance parameters appear in which is directed transverse to the foil-chip changer, which replaces Table 3. beam path near the pinhole. The the used foil with a fresh one after Beam-handling components for driving current pulse must be very every shot. We are currently prov­ the Nova system will be larger and rapid to create the plasma, which ing this concept in practice. Third, more numerous than those blocks light reflected from the the plasma formed by the l-MA associated with previous systems. target. (Such light reappears at the electrical pulse must be directed so Knowing that optical coating 2 plasma within about 400 ns.) Con­ that no debris accumulates at the damage occurs at 8 J/cm , that the sequently, advanced rail-gap spatial filter lenses. This is done by peak f1uence is about twice the technology has been used to carefully shaping the foil (and the average f1uence, and that 20 beams minimize electrical circuit induc- . changeable chip in which it resides) are planned, we can use simple tance. Tests have confirmed that prior to the application of the elec­ arithmetic to determine that beam the 3-cm/f.lS plasma velocity that is trical pulse. All debris from the ex­ diameters greater than 70 cm will created with an energy store of 6 kJ ploding foil is captured at a plasma be required to achieve the project is sufficient to ensure closure. dump located transverse to the goal of 300 kJ. Nova's final focus­ Three criteria must be sa,tisfied beam propagation direction. The ing lenses will in fact accommodate by the plasma shutter driver circuit. entire unit, including its control 74-cm-diameter beams. The First, the plasma jet must be suf­ electronics, pulse-forming network, aggregate area of optical compo­ ficiently dense to block all of the and power supplies, will be located nents in the final beam-handling light returning from the target. To at the final spatial filter itself. Elec­ stages will approach that of compo­ ensure this, we generate a plasma tromagnetic interference with nents in some of the world's larger of greater than critical density (for system electronics in the Nova test telescopes. (The primary of Mt. 1.05-f.lm light). Second, the plasma facility is currently being evaluated. Palomar's 200-in. telescope has a shutter must operate in the vacuum We anticipate no major problems. 200 000-cm 2 aperture; Nova's 20 surrounding the spatial filter This unit is significantly cheaper beams have a combined aperture of 88 000 cm 2. ) To direct and point these beams, large mirrors are UChl@il® ~ Plasma shutter physical parameters. required. The blanks from which the mirrors will be fabricated are Round trip beam time of flight shown in Fig. 7c. (Nova will require (shutter to target) 400 ns 40 to 50 mirrors.) Eastman Kodak Peak current 0.7 MA Incorporated, Tinsley Laboratories, Nominal plasma jet velocity 3 cm/ /ls and Perkin-Elmer Corporation Pla~ma temperature 30 eV Plasma density (singly ionized) 1.5 X l

12 DEFENSE PROGRAMS

Figure 8 is an artist's conception target. The cone angle itself can be struments. In fact, our target­ 0 of the Nova target chamber. The varied (from 80 to 1000) to irradiation experiments will be west beams are equally spaced in provide flexibility in dealing with diagnosed and analyzed more angle upon the surface of a cone various experimental target thoroughly than the underground whose vertex is at the target. These designs. Figure 8 also shows some nuclear tests conducted by LLNL beams are opposed by the east of the ancillary target event in Nevada. The line drawing in beams so that east and west beams diagnostic systems. This configura­ Fig. 8 illustrates design criteria for can radiate into each other through tion will allow for a full complement the overlap of Nova beams at a a coordinate system centered at the of experimental and diagnostic in- representative target. In the

Target positioner 150 (design pm limit on Soft-x-ray overlap streak camera ~ ~ diam) t

Cluster of f / 3 beams Target al ignment optics

F iIter-fluorescer x-ray spectrometer

"----Vacuum system

o (0) Artist's conception of the Nova target cham­ beams in each cluster are positioned on the surface of a 90° IFD@o (Q) ber, showing the typical positions of some of cone. While a 100° cone angle represents our optimal choice the major experimental diagnostics. Each of the 74-cm at the present time, provisions have been made to permit lenses focuses a beam onto the target through a nominal variation of the cone angle between 80° and 100°, depending focal length of 2.25 m. The beams are arranged in two op­ on the nature of the target experiment. posing clusters of 10 beams each. As shown on the left, the

13 (common) focal plane, the beam transfer rates (10 Mbit/ s) , low Present status overlap spot is not expected to ex­ overhead through direct access to Nova building construction is ceed 150 ,um in diameter, including computer memory, and program­ well under way. Concrete walls allowances for alignment, position­ mable network connections. To 1.8 m (6 ft) thick surround the ing, and verification tolerances. facilitate block data transfer, which target room, and walls of lesser This criterion applies for a nominal is especially useful for image data thickness surround the "switch­ focal length of 2.25 m. processing, the sophisticated yard," which will house primary turn­ Complex systems like Nova, re­ NOVANET interconnection system ing mirrors and output sensor quiring literally hundreds of elec­ has been implemented using in­ packages. Structural steel has been tronic and electromechanical con­ telligent "Novalink" controllers, erected for the new laser bay. trol functions for a single laser­ which can communicate both to Design engineering is being done target experiment, must rely upon remote LSI-11/ 23 computers and by Kaiser Engineers and Bechtel, an extensive, sophisticated to remote memories of stored Inc., in collaboration with the LLNL computer-control network. The video images through extensive engineering staff. control system architecture is fiber optics communication chan­ We have negotiated and designed to handle 'multiple tasks nels. finalized contracts for all of the ma­ (such as laser alignment, target One example of the system's jor optical components except the alignment, capacitor-bank charge­ remote, closed-loop alignment laser glass, and procurement action and-fire sequencing, and laser and capability is its ability to align each on it will occur this year. Most of the target diagnostic data processing) of the 140-odd spatial filter major additional components are from a centralized location. Com­ pinholes scattered throughout the close to the end of their respective mon hardware and software laser bay. An alignment beam is developmental/prototyping routines allow functional redun­ detected with a CCD array (the phases, and detailed design work dancy. For example, two (of the solid-state eqUivalent of a television on many of them has begun. Active three) top level VAX-ll/ 780 com­ vidicon camera) located in the out­ controls-subsystem development is puters (Digital Equipment Cor­ put sensor package (Fig. 2). This now under way. We anticipate the poration's top real-time computer array represents the state of the art timely completion of Nova in 1983. system) are capable of operating in image sensors and offers greater the entire system through a task­ than a factor of 10 improvement in Key words: glass lasers; inertial confinement fu· sharing network. The same is true dynamic range. The sensor­ sion; laser ampli fiers; Nova; spatial fi lters; ther· monuclear reactions. of the operator touch-panel display package optics image the plane of consoles, which are interchange­ the pinhole upon the CCD array. able. The image is automatically Notes and references The control system com­ processed by the control system, 1. See the August 1977 Energy and municates with many distributed which then commands stepper Technology Review (UCRL·52000-77-8). 2. J. Emmett, J. Nuckoll s, and L. Wood, "Fu· devices through an extensive fiber motors to position the pinhole to a sion Power by Laser Implosion," Scientific optic network, featuring high data preset location (as judged by the American 230, 24' (June 1974). pinhole image location on the 3. M. J. Minot, "Single· layer, Gradient Refrac­ CCD). The alignment control tive Index Antirefl ection Films Effective from 0.35 to 2.5 Ii ," J. Opt. Soc. Am. 66, 515 subsystem is capable of aligning (1976); M. J. Minot, 'T he Angu lar Reflec· all of the Nova pinholes within tance of Single- layer Gradient Refracti ve­ thirty minutes. index Films," J. Opt. Soc. Am. 67, 1046 (1977) 4. P. A. Temple, D. Milam, and W. H. Lowder· milk, "COz-laser Polis hing of Fused Sil ica Surfaces for Increased Laser Damage Resistance at 1.06 lim," in Laser Induced Damage in Optical Materials: 1979, Eds. H. E. Bennett, A. J. Glass, A. H. Guenther, and B. E. Newnam, NBS Special Publica· tion 568 (1980), pp.229-236.

14