Fu ion with Green and Blue Light

We will use a frequency-conversion subsystem to produce green and blue light with the Laboratory's Novette and Tova to improve substantially the performance of Fo r further informatIO n ontac Da\ id fimer' ( <1S) 422--505 inertial-confinement fusion targets.

18 INERTIAL FUSION

In inertial confinement fusion, tiny shorter wavelengths (0 .53 and 0.35 ~m) but powerful thermonuclear explosions both increase the maximum absorption are produced by focusing a laser beam and cause more of the incident light to on microscopic (0.1 to 5.0 mm) be absorbed at high intensities. Analy­ deuterium-tritium (D-T) targets. Be­ sis of these results indicates that the cause the laser beam is very intense absorption of laser light is improved 7 2 (l05 to 10 GW/cm ), it generates a because inverse bremsstrahlung is more dense plama where it impinges on the efficient at shorter wavelengths. target material. This interacts Shorter wavelengths also affect pre­ strongly with the laser beam. Recent heating of the fuel in these laser-driven experiments on laser-plasma interac­ targets. The Pd V work necessary to tions have confirmed theoretical predic­ compress a gas such as a D-T mixture tions that target performance improves increases with the gas temperature. at laser wavelengths shorter than Therefore, any heating of the fuel be­ 1 ~m , the wavelength of most target­ fore compression increases the work irradiation studies at LLNL,I This result (i.e., the laser energy) required to com­ has encouraged us to develop short­ press it. High-velocity suprathermal wavelength target-irradiation facilities. electrons produced by collective laser­ The desired wavelengths can be most plasma interactions may preheat the efficiently achieved by converting the target before compression by penetrat­ fundamental wavelength produced by ing to the fuel, where they deposit their current -glass lasers. The kinetic energy. Recent experiments have Laboratory's laser, now under shown that at shorter wavelengths, the construction, is a neodymium-glass sys­ number of suprathermal electrons is re­ tem that will produce a beam of 80 to duced significantly. Thus, by reducing 120 kJ.2 A frequency-conversion subsys­ preheating, shorter wavelengths make tem between the output beam and the another contribution to reducing the target will enable Nova to operate at laser energy required for ignition. shorter wavelengths. We plan to con­ duct target-irradiation studies with Generating Shorter Nova to measure quantitative improve­ Wavelengths ments in target performance at The foregoing experimental results wavelengths of 1.052, 0.53, and 0.35 ~m. indicate that target performance should We review, here, the physics of laser­ improve as the wavelength of the laser plasma interaction at shorter wave­ drive is reduced below 1 ~m . There are, lengths and describe the frequency­ however, several fundamentallimita­ conversion system to be used with tions on the range of operating Nova. wavelengths for technically feasible laser systems.3 These include the avail­ Laser-Plasma Interaction ability of a lasing medium, the effects Studies of light propagation in optically non­ Two effects lead us to believe that linear materials, and the damaging ef­ the performance of directly driven fects of intense, short-wavelength light fusion targets will improve at shorter pulses on optical materials. . wavelengths of incident laser light: the Because of these constraints, the target absorbs more light, and the dele­ shortest wavelength generally consid­ terious preheating of the fuel by fast ered practicable for laser drivers is electrons is reduced. These effects de­ about 0.25 ~m. In contrast, the most pend on the highly nonlinear physics powerful existing target irradiation fa­ of plasma instabilities and their excita­ cilities use lasers that operate either tion by an intense electromagnetic near 10 or near 1 ~m . As a practical wave. matter, then, we decided to reduce the Recent experiments have studied the wavelength of the driver up to a factor absorption of light in various materials of four. By interposing a frequency­ as a function of intensity and wave­ conversion subsystem between the length.I The experiments revealed that output ports of the laser arms and the

19 target, we can shorten the wavelength media, which requires no auxiliary in­ of the laser output. This approach has puts and thus has the virtue of simplic­ proven to be most cost effective with ity and economy. This process is the neodymium-glass lasers and is the one best understood frequency-conversion we have adopted. technique and also provides the most control over the beam quality of the converted light.

Harmonic Generation Recent experiments on laser-plasma Harmonic-conversion techniques are interactions have confirmed theoretical well understood and are routinely im­ plemented in small-aperture commer­ predictions that target performance cial lasers.4 Very intense laser light inci­ dent on a transparent medium interacts improves at laser wavelengths shorter with the material's atomic structure to generate electromagnetic radiation with than 1 ~m ... frequencies that are multiples of the fundamental frequency of the incident light, w (see box on p. 22). These are known as harmonics of the funda­ Frequency Conversion mental frequency. The strongest of In principle, any physical phenome­ these is at 2w, and the next strongest is non that changes the frequency of light at 3w. For high conversion efficiency, is a candidate for frequency conversion. the phase velocities of the waves must In practice, however, other require­ be commensurate; the process that ments limit our choice. These include: achieves this is called phase matching. • A focusable output (less than 20 Phase-matched conditions generate times the diffraction-limited spot size). high-power harmonics of the funda­ • A high conversion efficiency mental wave that have a focal spot ap­ (greater than 70%). proximately the same size of that of the • A high damage threshold (above fundamental wave. 2 5 J/cm ). In negative uniaxial crystals such as • Simplicity. potassium dihydrogen phosphate • Availability of materials. (KDP), we can phase match in one of With regard to conversion efficiency, two ways (called type I and type II), for example, there would be no cost ad­ depending on the orientation of the vantage to a technique that requires a polarization of the waves relative to the larger neodymium-glass laser facility to crystal axis and to the direction of reach ignition with frequency conver­ propagation (see box on p. 22). Phase sion than without it. matching thus requires precise orienta­ To satisfy these conditions, frequency tion of the crystals. Our design for a conversion must be based on a physical KDP frequency-conversion subsystem phenomenon that is coherent, satu­ uses type-II phase matching and a rable, and virtually lossless. Among novel arrangement of the crystals to such phenomena are stimulated provide maximum flexibility for the multiphoton Raman scattering, the gen­ laser system. eration of a sum frequency by two-, three-, or four-wave mixing, and the KDP as a Conversion generation of harmonics in optically Medium nonlinear media. Each of these has its Many materials exhibit the properties own impact on laser system design. needed for frequency conversion. The Where the input for the conversion most commonly used are the crystalline process requires waves of different fre­ solids potassium dihydrogen phosphate quency, the effect on the laser system is (KDP), ammonium dihydrogen phos­ considerable. We have decided to use phate (ADP), and their deuterated harmonic conversion in nonlinear isomorphs. These materials also have

20 INERTIAL FUSION

the properties, listed above, essential to harmonic is required, another attractive the fusion application. (We considered candidate material is CDA (cesium other materials, such as atomic vapors, dihydrogen arsenate). Phase matching vapors of asymmetric molecules, and is much easier with CDA than with various solid and liquid media; none, KDP, so long as the crystal temperature however, proved as readily adaptable is well controlled. We are continuing to to the neodymium-glass laser.) examine other candidates for future KDP was the first material found to laser systems to increase system flex­ exhibit phase-matched generation of ibility and to decrease the cost of the the second harmonic.s Although the frequency-conversion subsystem. number of such materials has since climbed into the hundreds, KDP re­ Frequency Conversion in mains a primary choice. During the N ovette and Nova 1970s, the development of KDP for use The Laboratory first applied KDP in Pockels cells (a type of optical crystals to harmonic conversion with its switch) led to the production of crystals . With both the upgraded of excellent optical quality with a clear Argus (Novette) and the high-power aperture approaching 10 cm in diame­ Nova neodymium-glass lasers, we will ter. Major crystal growers are continu­ use KDP to generate second and third ing their efforts to increase aperture harmonics for studies of the irradiation diameter, now emphasizing the undeu­ of fusion targets. Our objective is to terated form. Undeuterated KDP crys­ convert the (1.053-.um) light tals have an optical quality superior to produced by these lasers to visible the deuterated form, do not require as green (0.53-.um) or blue (0.35-.um) light. expensive a growth facility, and have This conversion entails five major de­ provided most of the growers' sign requirements: experience. • High conversion efficiency from Compared with other candidate ma­ the fundamental frequency to the sec­ terials, KDP is resistant to laser-induced ond and third harmonics. damage, has an adequately high • with a large harmonic-generation coefficient and a (74-cm) aperture. very low self-focusing index, and trans­ • Multiwavelength flexibility at mits ultraviolet frequencies well. It can minimum cost. be grown from solution to unusually • Integration with the target align­ large sizes and can be phase matched ment and diagnostics systems. to generate 2w, 3w, and 4w harmonics • Minimization of nonlinear propa­ from 1.064- or 1.053-.um light (actual gation effects. fundamental wavelength is a function of the specific glass used in the Conversion Efficiency neodymium-glass laser). These proper­ Conventional methods of generating ties outweigh KDP's relative disad­ the second and third harmonics tend to vantages, which include undesirable work efficiently over a limited intensity absorption (0.05S/cm) at 1.064.um, its range. Applied to Nova, such methods softness and consequent polishing diffi­ would require three interchangeable culty, significant hygroscopicity, suscep­ crystal arrays, each with a different tibility to fracture, and slowness of crystal thickness, to achieve the desired crystal growth (about 1 mm a day). performance over a wide range of input Several of the candidate materials are intensities. We have developed a sys­ superior to KDP in at least one cate­ tem capable of generating the second gory. At present, however, KDP re­ and third harmonics over the required mains the only material that satisfies all intensity range using only a single ar­ the fundamental requirements of har­ ray containing two crystals. Our design monic generation at 1.064 and 1.053.um is called the quadrature-green/cascade­ and can be grown in crystals of an ade­ blue scheme.6 quate size in time for installation in the To generate the third harmonic, we Nova laser. If only the second use two crystals already in the basic

21 Harmonic Generation site, we obtain both odd and even multiples. By far the strongest of these in KDP is 2w and, in order of rapidly decreasing When a low-intensity light wave strength, 3w, 4w, and higher propagates through a transparent frequencies. (There is also a zero­ medium, it drives the electron cloud of frequency, or dc, polarization term). the constituent atoms into forced os­ Each of the oscillatory motions gives cillation (a). This oscillating electric rise to a radiated wavelet (a harmonic) charge constitutes a polarization wave at multiples of the input frequency. that gives rise to a radiated wave. The In generating second harmonics, for radiated wave represents the usual example, each wavelet at the second "linear" response of the medium to the harmonic frequency, 2w, radiates incident light. Because the oscillatory outward at the phase velocity of the motion of the electrons follows the medium for that frequency, which is vibratory motion of the driving field, the vacuum speed of light divided by the radiated wave is of the same the at 2w (v2w = c/n2w)' frequency as the incident wave. The The fundamental wave, however, wave thus emerges from the propagates with a different phase transparent medium with its frequency velocity (vw = c/nw)' Because the unchanged but with a phase delay that fundamental wave imposes its phase depends on the material's refractive on each of the electron oscillators, and index. This applies to all low-intensity thus on the radiated 2w wavelets, the light waves passing through net radiated wave at 2w is the sum transparent media. of many out-of-phase wavelets from Laser pulses typically have a spectral the host of oscillators throughout the brightness (power density per material. We describe this situation by steradian-hertz) many orders of saying that the process is not phase magnitude greater than that of matched (d). The result is a very weak incoherent sources. It is this property conversion of the laser energy from that enables a laser pulse to produce a the fundamental to the desired more vigorous oscillation of the frequency, 2w. electron cloud. The electrons may be To achieve the phase-matched displaced from their equilibrium condition, in which the 2w wavelets position by a distance many times will add coherently, we must arrange larger than the truly minute excursion for the phase velocity of the induced at low intensity. As a result, fundamental wave to equal that of the the electrons come into greater contact second-harmonic wave. This can be with neighboring atoms, probing their done by choosing a material with repulsive potential with each oscillatory anisotropic characteristics that cycle. This process enables the structure compensate for the variation in phase of the medium to influence the velocity with wavelength. The wavelets frequency of light radiated by the from all the oscillators in such a oscillating electrons. If the strength of material then will reinforce one the repulsion experienced by an another (d), resulting in a high electron driven by intense laser light is conversion efficiency. KDP crystals asymmetric (different in opposite display this property. directions), its motion will be The phase velocity of light in a biased (b). transparent medium is inversely If we decompose an electron's proportional to its refractive index, motion into its frequency components, which, in tum, depends on the spatial they appear at multiples of the input orientation of the material and on the frequency, w (c). For a centrosymmetric wavelength of the light. It is possible to site, the components appear at odd orient a crystal of KDP so that the multiples (mostly 3w and, to a vastly propagation velocity of incident light smaller extent, Sw) . For an asymmetric waves produces exact phase matching.

22 INERTIAL FUSION

(a) (c)

Symmetric electron motion w ~ r'\ ~( ~.,.p::: -L :::c ,;,. ;;[) A A AQ3W ~ . '-../ ~ \ ~ . '-../ \.../"" ~ 5w/V\JVV\I\N Incident wave ...... t 1/ Transmitted wave

r Electric field amplitude Asymmetric electron motion w ~ , .... ,/ ~ , .... -" , ,- ~ .....--...... ~ Q 2w ~ , \ ~ ~ !/ ~~~. Time \IV 3w ~ ' ... --' '-~ '-,' 4wf\../\.f\.r\f\. Electron motion 5wrv\M.MJ\I\J

(b) (d) Symmetric repulsive potential ( centric site)

Not phase matched

Input wave at w '\ , / ..... _"

Asymmetric repulsive potential (acentric site) Phase matched

The optical axis of KDP is such that light polarized per­ pendicular to the axis experiences an angle-independent (ordinary) refractive index, no' whereas light of the other polarization experiences an angle-dependent (extraordi­ (e) nary) refractive index, ne (e). The extraordinary index ne(O) at the second harmonic is equal to the ordinary index at the first harmonic at a particular angle 0, the phase­ matching angle of incident light. This condition, called type-I phase matching, produces phase matching for the second harmonic. c:: .2 Phase matching may also occur when the second har­ ti monic is matched with two fundamental waves polarized ~ in orthogonal directions.4 This is called type-II phase ~ '0 matching. In this way, a KDP crystal provides a high con­ )( Ql version efficiency in one particular orientation but a low "C efficiency in any other orientation. We have demonstrated c:: r------~~~~------~,------_J experimentally this conversion technique, called angle­ tuned harmonic generation, with the Laboratory's Argus --- and other laser facilities. This technique requires precise orientation of the KDP crystal, typically to within 100.urad. Wavelength

23 orientation for generating blue aperture crystals have a high aspect ra­ 7 (0.355-)lm) light. ,8 This is the standard tio (aperture divided by thickness) and cascade-blue configuration. For high a long fabrication time. This limitation conversion efficiency, we optimize the can be overcome, however, by fabricat­ crystal lengths for efficient performance ing an array of small-aperture crystals over the full operating range of input to provide a large total aperture. We intensity. To generate the second­ have used this technique in designing harmonic green (0.532-)lm) light, we the frequency-conversion subsystem for have developed a new approach, the Nova. In our two-crystal subsystem, quadrature scheme (see box on p. 26). each 74-cm aperture is fabricated of In this scheme, two crystals are nine KDP segments (Fig. 2). Each array arranged almost exactly as in the is sandwiched between two windows cascade-blue configuration. To change coated with graded-index antireflection the system from generating blue to layers. 9 Losses from Fresnel reflections generating green light, we need only at surfaces inside the assembly are re­ rotate the two-crystal array about the duced with three thin layers of index­ beam axis by 10 deg and angle-tune the matching fluid. The composite arrays second crystal (about only one axis of are assembled by precisely machining the assembly) from the blue phase­ the KDP crystals so that the phase­ matching angle to the green phase­ matching direction (the beam direction) matching angle-a rotation of just a is accurately aligned with the surface few milliradians. In contrast to conven­ normal. This approach was made tional approaches, there is no need for feasible by the use of diamond-turning interchangeable arrays. technology to machine the crystals and Our scheme produces green light ef­ by precise measurement of the phase­ ficiently over a large operating range of matching angle (± 30 )lrad). input intensity, much larger than the Exacting requirements for the optical system requirement (Fig. 1). It performs quality of the beam severely con­ best at the low-intensity end of Nova's strained the mechanical design of the pulse-width range (2 to 5 ns). This assembly. We developed and tested design is consistent with other system two designs using a I5-cm, clear­ constraints such as minimizing aperture prototype. In the "egg crate" nonlinear propagation effects. design, the crystals are embedded in a stainless-steel lattice that supports the Large-Aperture Arrays windows against an internal vacuum Nova requires frequency-conversion load. We have also tested a "close arrays with a large (74-cm) aperture. packed" design, in which the individual The feasible aperture size of single KDP crystal segments are placed directly crystals is limited, at present, as large- against each other, and the whole array Fig. 1 is supported by the windows and per­ 1.0 r------, Calculated conversion efficiency of haps some internal post supports. (Ex­ the quadrature-green / cascade-blue perimental results from a prototype arrangement, assuming a Nova whole­ 0.8 close-packed array are shown in Fig. 3.) beam divergence of 200l'rad and the >­ <.) We are evaluating, both experimentally optimum crystal thickness of 1.4 cm c: Ql for both crystal arrays. (a) The conver­ and theoretically, the relative merits of ~ 0 .6 sion efficiency (1)2 w) of the Q; these two designs. fundamental harmonic (w) to green c: The intensity of high-power lasers is o light (the second harmonic) is high .(jj limited by self-focusing, a phenomenon over the full operating range (0.5 to 4i 0.4 > Third harmonic 3 GW/ cm2); (b) the efficiency of con­ c: whereby small-amplitude, small-scale o version (1)3) to blue light, (the third U intensity ripples in the laser beam am­ harmonic) declines at higher intensi­ 0.2 plify to very high intensities that can ties of the fundamental frequency. damage optical components.3,IO,1l This Efficient conversion to blue light is not effect, caused by nonlinearities in the necessary beyond 2.5 GW/ cm 2 , as the intensity of the incident laser 1.0 2.0 3.0 optical materials, occurs to some extent beam is limited by self-focusing in the Incident intensity in all optical components, including the target-focusing optics. (fundamental wavelength), GW/ cm 2 target-focusing optics. The severity of

24 INERTIAL FUSION the effect increases with the optical path of the laser beam through glass is path length and with the laser intensity minimized so as to maximize the high­ and decreases with wavelength. Seg­ est power at which the laser operates mented crystal arrays induce ripples in without undergoing self-focusing.6 the beam when it is diffracted by the The spatially filtered infrared laser gaps between segments, subjecting it to beam produced by each arm of Nova is self-focusing in the target-focusing op­ imaged onto the target-focusing optics tics, especially at the third harmonic through a system of mirrors that are (0.35 ~m) . In our design, we minimize rippling by using apodization to control diffraction from the gaps (apodization is a method of softening hard edges Index-matching Fig. 2 fluid « 100 I'm thick) that clip the beam). Each 74-cm frequency-conversion ap­ erture is fabricated of nine KDP seg­ Design of the Frequency­ ments. The three-by-three array is Conversion Subsystem sandwiched between two anti­ reflecting windows. Losses from The conversion scheme produces two Fresnel reflections at surfaces inside wavelengths by using the same arrays the assembly are reduced with three that generate the second harmonic to thin layers of index-matching fluid. generate the third harmonic; the color The KDP crystals are machined so that is changed merely by adjusting the ori­ the phase-matching direction (the beam direction) is accurately aligned entation of the arrays. The rest of the with the surface normal. frequency-conversion subsystem is de­ signed to be similarly color flexible. ~WI'dOW'~ Components of high-power lasers typically perform best (in terms of damage threshold, reflectivity, and wavefront distortion) when they are optimized for monochromatic opera­ tion. It is not economically feasible, however, to use, say, three sets of 74-cm-aperture focusing optics, one for each operating wavelength. Therefore, we designed the frequency-conversion and target-focusing subsystems to oper­ ate at multiple wavelengths, minimiz­ ing the number of multiwavelength components. Furthermore, the total

Fig. 3 Using a prototype close-packed KDP crystal array (left), we were able to generate the second-harmonic green light (right). Data were taken with the Argus laser facility: laser output at 2 1.064 I'm and 0.7 ns was 2 GW/ cm ; crystal width was 5 cm; crystal thick­ ness was 11.8 mm. The measured conversion efficiency from the funda­ mental infrared wavelength to the sec­ ond harmonic (green) was 75%.

25 (a) Quadrature Conversion Scheme The quadrature conversion scheme is a method of generating the second harmonic of a fundamental wave. The scheme, which uses two crystals in series, has several advantages over single-crystal or other two-crystal schemes:

(b)

~

~ k z, o

e

~ Beam direction, k

o

(c) (d)

1.0

0.8

0.6

0.4 'tl

'"E 0.2 !i Cl c: 0 '"Cl cc: .a - 0.2 Q) c - 0.4

- 0.6

- 0.8

- 1.0 0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16 Incident intensity, GW/ cm2

26 INERTIAL FUSION

• High conversion efficiency over a The conversion properties of a single large dynamic range of drive intensity crystal of KDP 1.4 cm thick are plotted and de tuning angle. as contours of constant conversion • Easier crystal alignment. efficiency in (c). Consider a large­ • The same crystal configuration as aperture laser beam incident on the the two-crystal methods of generating crystal. Because of the unavoidable the third harmonic, making the slight aberrations present in any laser transition from a 2w to a 3w output beam, the intensity and local direction much simpler. of the beam (angular detuning) will Consider a pair of KDP crystals cut vary over its face. As the crystal acts for type-II phase matching [in (a) we locally in converting the laser light to show the first crystal of the pair]. In the the second harmonic, the conversion quadrature scheme, the optical axes of efficiency varies from point to point the crystals are arranged so that the over the face of the beam. The relevant planes containing the direction of the conversion efficiencies are defined, laser beam and their optical axes (the therefore, by the range of intensities kz planes) are mutually perpendicular and angular detunings present in the (b). This arrangement has two impor­ beam. This range defines an area on tant properties. First, in type-II phase the detuning-intensity plot. For efficient matching, the incident wave is polar­ conversion, the efficiency within that ized at 45 deg to the kz plane of the region must be high. Clearly, any crystal. Thus, in the quadrature scheme, energy in the laser beam that meets the if the incident wave is correctly polar­ crystal with an intensity and detuning ized for efficient conversion in the first for which the conversion efficiency is crystal, it is also correctly polarized for low will remain unconverted. For efficient conversion in the second efficient operation, the crystals thus crystal (b). Both crystals can therefore must convert efficiently over a suffi­ convert efficiently. ciently large area in the intensity­ Second, the second harmonic waves detuning plane. The required size of generated in the two crystals are this area depends on the quality of the orthogonally polarized (hence the term incident laser beam. As shown in (c), a quadrature). The output from the first single l.4-cm crystal converts efficiently crystal thus has the wrong polarization in a roughly rectangular region (defined to experience in the second crystal. by the 90% contour) 200 to 400 JLrad 2 If the conversion efficiency is 771 in the wide from 4 up to about 20 GW/cm • first crystal and 772 in the second, the Any laser light falling in this region output from the first crystal is 7711 will be converted to the second (where I is the incident intensity of the harmonic with an efficiency greater fundamental wave) and the input than 90%. intensity to the second crystal is Figure (d) shows the same plot for (1 - 771)1. Because the second crystal two l.4-cm crystals arranged in quadra­ does not change the output of the first, ture (see b). In this case, the the overall output is simply the sum of 90% contour occupies a much larger the output from each crystal. The region- about 1 mrad wide at overall conversion efficiency is then 6 GW/cm2 and falling to 0.5 mrad wide 2 at 15 GW/cm • The quadrature system is clearly much more forgiving of beam aberrations than is a single crystal. For the beam quality expected from Nova, The advantage of quadrature conver­ angular detunings are not expected to sion is that it enables us to arrange for exceed 0.2 mrad. In this case, the quad­ the second crystal to convert whatever rature system gives a conversion effi­ light is left unconverted from the first ciency greater than 80% for intensities crystal. above 2 GW/cm2.

27 highly reflective at the beam's funda­ fl13 aplanatic lens with the last surface mental wavelength (Fig. 4). In our de­ uncoated. Combined, these two lenses sign, the two frequency-conversion provide an fl4 focusing system with a arrays, operating in series, are placed 4 % reflection of the harmonic between the last turning mirror and the wavelengths for diagnostic purposes. focusing optics in each arm. This allows The reflection is focused and returned us to miminize the number of compo­ to a diagnostic sensor via the same nents required to perform under high­ turning-mirror system that transports power incident light at more than one the beam to the arrays. By changing the wavelength. spacing between the two lens elements, The multiwavelength portion of the we can control the convergence on the system includes the focusing optics, diagnostic reflection to accommodate debris shields, a vacuum barrier, and arm-to-arm variations in the distance beam dumps. Of these, only the beam from the lens to the sensor. This spac­ dumps are wavelength unique: these ing change also helps compensate for absorb the undesired residual funda­ the focal shift in the target plane due to mental (1.052-/J-m) or harmonic dispersion in the glass. (0.53-/J-m) wavelengths. Changing the wavelength of the laser output thus re­ Integration of Diagnostic and quires only a slight reorientation of the Alignment Systems Fig. 4 crystal arrays, a minor repositioning of The harmonic wavelengths partially Frequency-conversion subsystem of the target-focusing optics, and replace­ reflected from the focusing-lens system Nova, showing KDP crystal positioning ment of the beam dumps. Because no are transported by the turning mirrors and target-focusing optics. Legends major components need be replaced, to a diagnostic sensor (about 30 m show the operations necessary to set the system is thus highly color flexible. away). The sensor is located behind the the subsystem for (a) red, (b) green, or fl6 (c) blue light. The conversion wave­ The first element of the aspheric, first turning mirror in an area shielded length can be changed without replac­ focusing lens serves as a vacuum bar­ from electromagnetic interference and ing major components. rier (Fig. 4). The second element is an from produced by fusion reac­ tions. As the iritensity of the reflected beam is low (less than 4% of the har­ monic intensity) but overly energetic for diagnostic purposes, the mirrors Aw = 1.052 /Lm, Em". = 15 kJ / 3 ns need be neither highly damage resis­ • Detune crystal array: tant nor highly reflective at the second Ewalong crystal axis. and third harmonics. The reflected sig­ • Remove beam dump. nal is used for calorimetry and for • Position doublet lenses. spatial and temporal measurements. The alignment procedure can be con­ (b) sidered in two parts. The first is align­ A2w = 0.53 /Lm, Em". = 12 kJ/ 3 ns ment of the turning mirrors and focus­

• Tune crystal array for 2w: ing optics to enable the high-power rotate 45 deg, tilt to match phases. harmonic pulse to properly illuminate • Insert beam dump. the target. The second is alignment of • Position doublet lenses. the crystal array in the phase-matching direction for efficient harmonic conver­ (c) sion of the incident fundamental pulse. A3w = 0.35 /Lm, Em". = 10 kJ / 3 ns To align the optical elements of the system at the target, we need to know • Tune crystal array for 3w: rotate 10 deg, tilt 0.25 deg. the beam direction. This is provided by • Change beam dump. low-power reference beams that are • Position doublet lenses. coliri.early aligned with the main (high­ power) laser beam. For experiments f / 13 aplanatic element! splitter using infrared (1.052-/J-m) light, we will f / 6 aspheric lens/ vacuum window use the system alignment laser, an in­ frared reference beam designed into Nova. To align the optics for

28 INERTIAL FUSION experiments with green (0.53-,um) or The payoff of such a conversion sys­ blue (0.35-,um) light, we will use an an­ tem in improved target performance cillary low-power, continuous-wave promises to be considerable, and KDP laser operating at the same wavelength crystal array technology is currently as the desired harmonic. The ancillary sufficiently advanced to meet Nova laser is colin early aligned with the specifications. Accordingly, plans are system-alignment laser, thereby provid­ for the (in late 1982) and ing a low-power reference source of the the Nova laser (in 1984 or 1985) to pro­ desired wavelength. The turning mir­ vide high-energy, short-wavelength rors and focusing optics are then posi­ target irradiations. tioned using standard target-alignment techniques. Key Words: Argus; frequency conversion, laser; harmonic generation; inertial confinement fusion To align the crystal arrays, we use (lCF); laser-plasma interaction; Nova; Novette; the master oscillator pulse train (operat­ short-wavelength lasers; potassium dihydrogen ing at the fundamental frequency), am­ phosphate (KDP). plified by small-aperture rod amplifiers with a high repetition rate (0.1 Hz). We Notes and References then orient an array so that it converts 1. Recent work at LLNL was described in the Laser Prog ~am A I1IUlal Report, Lawrence Liv­ the second harmonic with maximum ermore National Laboratory, Rept. UCRL- efficiency as monitored in the target 50021-81 (1981), Sec. 6. chamber. This technique, however, will 2. The design and operating principles of the not produce a measurable third­ Nova laser were described in Energy and harmonic signal. We align the array for Technology Review (UCRL-S2000-80-12), December 1980, p. 1. the third harmonic by an open-loop 3. J. R. Holzrichter, D. Eimerl, E. V. George, correction of its position at the second J. B. Trenholme, W. W. Simmons, and J. T. harmonic. This amounts to a tilt of Hunt, High Po wer Lasers for Fu sion, Lawrence 4.4 mrad about one axis of the array. Livermore National Laboratory, Rept. UCRL-86161 (1980). 4. F. Zernike and J. E. Midwinter, Applied Non­ Summary and Conclusions linear Optics (Wiley lnterscience Publications, Both theory and experimental results New York, 1973). indicate that the performance of 5. P. A. Franken, A. E. Hill, C. W. Peters, and inertial-confinement fusion targets can G. Weinreich, "Generation of Optical Har­ be substantially improved by operating monics," Phys. Rev. Let t., 7, 118 (1961). 6. M. A. Summers, L. G. Seppala, F. Reinecker, the driving laser at wavelengths shorter D. Eimerl, and B. C. Johnson, "A Two-Color than 1 ,urn. The most efficient way of Frequency Conversion System for High­ attaining wavelengths that are three to Power Lasers," presented at CLEO '8I, four times shorter than the current op­ Washington, D.C., June 10-12, 1981. erating wavelength is by converting the 7. W. Seka, S. D. Jacobs, J. E. Rizzo, R. Boni, . and R. S. Craxton, "Demonstration of High­ laser's fundamental frequency. We have Efficiency Third Harmonic Conversion of designed a frequency-conversion sub­ High-Power Nd-Glass Laser Radiation," Opt. system using multicrystal arrays of Co mmul1., 34, 469 (1980). KDP. When installed with the Nova 8. R. S. Craxton, "Theory of High-Efficiency and Novette lasers, the system will op­ Third Harmonic Generation of High-Power Nd-Glass Laser Radiation," Opt. Commun ., erate at the second (0.53-,um) or third 34, 474 (1980). (0.35-,um) harmonic of the fundamental 9. The principles and fabrication of graded­ frequency (1.052 ,urn). The operating index antireflection surfaces were described wavelength can be easily changed. The the March 1982 issue of En ergy and Technol­ system should achieve high conversion ogy Review (UCRL-S2000-82-3), p. 9. 10. E. S. Bliss, J. T. Hunt, P. A. Renard, G. E. efficiency and will accommodate non­ Sommargren, and H. J. Weaver, "Effects of linear propagation effects over the in­ Nonlinear Propagation on Laser Focusing tensity range of interest. Requirements Properties," IEEE J. Quantum El ectron ., QE-12 for aligning the system and for di­ (7), 71 (1976). agnosing target performance can be 11. J. T. Hunt, W. W. Simmons, R. Speck, W. Warren, and D. Eimer!, Beam Propagation met at all three wavelengths without in the Frequency Converted Subsystems of Nova compromising the multiwavelength and Novelle, Lawrence Livermore National flexibility. Laboratory, Rept. UCID-19086 (1981).

29