V
PREPRINT UCRL- 77944 Ur,\>\--m'>te^-2fi Lawrence Livermore Laboratory STATUS OF LARGE NEODYHIUM GLASS LASERS
James A. Glaze, William W. Simmons and Wilhelm F. Hagen March ID, 1976 mm
This Paper was Prepared for Submission to the Society of Photo-Optical Instrumentation Engineers SPIE Meeting, Reston, VA March 22-23, 1976
This is a preprint ot a paper intended for publication in a journal or proceedings. Since changes may be made before publication, this preprint is made available with the understanding that it will not be cited or reproduced without the permission of the author.
i*^P^*S'
««*- -»v«>*^_
DiSTRiKUTiONC; 7602
STATUS OF LARGE NEODYMIUM GLASS LASERS* •ponum] by llw Van* Stu „ o| M[l Blmc*i ng PttritrmtM A« Canes A. Glaje, William H. Simmons and Wlhelm F. Hagen uitKnnmtion. 01 IIKU tiqitiqRt, Laser Fusion Division, Lawrence Livermore Laboratory nmnir.
Master Pulse Pulse Pre Pre— pulse oscillator switch-out shaper amplifiers isolation
Apodizing Rod ASE Disk ASE amplifier amplifier aperture groups isolation groups isolation
I Spatial filter, polarizer-rotator Disk amplifier Spatial filter . V — Typical disk amplifier group Fig. 2. Block diagram of a typical Nd:Glass laser chain as 1t would be staged for laser fusion expenrrents. A disk amplifier group is also shown. Such a group may contain several amplifier heads; these heads would generally have the same clear aperture. The rod amplifier groups are staged in a similar manner. The master oscillator used at LLL is a NdrYAG laser passively mode locked with a flowing bleaehahle dye- Bandwidth is established with an ctalon whose temperature is maintained within + Q.1°C with a closed cycle coolant loop. Uniform pumping of the rod 1s achieved with an afocal double ellipse reflector qeoretry The oscillator output consists of a train of subnanosecond pulses, one of which 1s selected and switched out for injection into the preamplifier stages. Energy of the largest pulse in the train 1s typically 2 mJ. Because of the statistica1 manner in which the dye is bleached by fluorescent noise spikes, this oscillator does not have the pulse-U-pulse temporal stability that is desired for large system applications. (Vari ations of up to 30% are sometimes observed). Attempts are currently being rude to develop oscillators with improved longitudinal mode discrimination. One promising approach.uses a synchronously driven Intracavity loss modulator (fast Pockels' cell) in addition tc the dye cell.('J Transform limited pulses ranging from <50 psec to >.5 nsec have been achieved. This oscillator to date, however, has problems of pulse train Instability thought to arise from Improper impedance matching of the PFN and Pockels1 cell.
The oscillator pulse swltchout incorporates a fast Pockels1 ce'll (1 cm diameter) positioned between crossed polarizers. In early designs half-wave voltage was applied to the cell by switching a high-pressure Ng-fllled spark gap with optical energy derived from the early portion of the pulse train. Since the opti cal energy arrives at the spark gap in the form of pulses whose spacing (typically 10 ns; is long compared to gap break-down times, it is possible to adjust the gas pressure so that the state of ionization required for full avalanche breakdown across the gap is achieved just prior to the arrival at the Pockels' cell of the pulse selected for switchout. This scheme is widely used, but suffers from jitter in the gap break down time, resulting in part fi'om shot-to-shot variability of the pulse train amplitude and In part from electrode degradation.with time. More recently, modern electronic switches have eliminated the troublesome spark gap entirely.**' In the electronic switchout circuit in present use at LIL, a Krytron switch, acti vated by an avalanche transistor stack, discharges a stripllne circuit which 1s accurately impedance matched to the Pockels' cell. This circuit derives Us timing signal from the pulse train through a vacuum photo- diode and voltage-settable tunnel diode comparator. By careful adjustment of time delay and threshold voltage, switched-out pulse energy variation can be minimized. The pulse shaper is a device that transforms the temporally gaussian oscillator pulse to i pulse more suitable for target irradiation. Theoretical calculations show that the desired waveform is a function of target size, composition, and available laser energy.(3) Of all the components that comprise the c*iain, this one, to date, Is least clearly defined In terms of performance requirements and construction. However, a large number of schemes have been proposed; these can be broadly characterized as pulse stackers, shutters and active shapers. The stacker produces a composite pulse derived from a single short pulse that is suit? ably delayed and attenuated. A patented scheme using a double etalon is.presently in use at XMS Fusion.W 5 Another scheme uses a Faraday rotator placed between cross polarizers.I '' Successive reflections from thoe *Registered Trademark name. 2 6* v f \i -t* rotator surfaces experience different rotatfons aod nonce different transmissions through the output pola rizer. This device, while simple to implement, does not provide an exponentially Increasing signal 15 dOuS the KMS stacker. Optical shutters have been in use for some tine. In this scheme a Pockels' or Kerr cell modulator fs used to chop-out a portion of a longer Q-switchcd pulse. Banduidth limitations on the nodu- 'lator presently Unit pulse rise tic* to about iCO nsec. furthermore, all shutters ire United by an .intrinsic dynamic amplitude range of 3U-40 dp. Active shapers rely on the transmission properties of saturable absorbers. One example is that proposed by Elliott and Kissey.W In their scheme a series of dye"cells are optically gated with combinations of exponential and Russian pulses suitably delayed In tine. Their calculations predict that the puts? shape discussed in Reference 3 can be closely matched. »
lasor radiance is limited by both linear and nonlinear wavefront distortions. Static aberrations arising from surface figure error can be as high as 0.4 naves RMS for a laser Such as CYC10PS which has over ISO optical surfaces. Quasistatic aberrations arising fron flash)*:-?-Induced theroal distortion of laser disks can be several waves unless elaborate cooling procedures are adopted. Adjustable corrections of several waves can be nade witn simple refractive elements.*'' Astigratlsn, for example, can be eliminated with an adjustable pair of coaxial cylindrical lenses. Cera, spherical aberration and wivefront tilt can be compensated in a similar manner. Hhilc this technique is United to the Soider aberrations, it 1s cost effective and amenable to intcrirotlon into a large systen, Adaptive optics such as deforaabte mirrors or phased arrays may someday be developed for fusion laser applications. At the present tine such devices seem too costly and complex for use in large systems.
In addition to snail-scale self-focusing both whole-bean self-focusing and diffraction can degrade beam focusabiliiy. To .meliorate these effects a bean shaping transmission filter (apodiclng aperture) Is placed at the beginning of the first power amplifier stage. Since the rate at which the phase 1s retarded by the nonlinear interaction is proportional to the intensity, the bean profflc will have a profound effect on the intermediate and far-field phase and intensity distributions of Che focused bean. For a circular bean with a parabolic intensity profile, the phase retardation of course will be parabolic and will behave as if It has passed through a weak positive lens whose focal length chafes in time with the pulse. Vn«n the beam is brought to a focus, this tensing will cause a time varying shift of the focal plene. For a sys tem such as CYCLOPS, operating at a terawatt and using *r> f*/2.5 focusing lens, this shift can be as large as 100 microns. Other bean profiles give rise to far ro»*e complex focal plane behavior; therefore the para bolic shape is preferred for many applications. Because of diffraction, a simple parabolic profile,will not propagate through the chain without developing near-field Fresnel ripples. To aveM this, a transmission filter *h1ch is parabolic over the central portion gf the aperture but decrease smoothly to Zero It the edge 1s used. In choosing the 5i*e and location of the apodiiing aperture, ox rust be careful that diffraction y ripples, which can be further amplified by self-focusing, are minimized, while at the same tlffc amplifier apertures sre optimally filled by the betn for mixlmum power extraction.
For fusion targets, amplified spontaneous emission (A5F.) trust be attenuated to avoid target damage prior to the arrival of the compressing pulse. Experiments,have shown that energy densities as low as 6 J/cmZ will cause damage to plastic parf.s and glue Joints.W Although It is difficult to pinpoint a pre cise design goal for control of ASE, it ».-, currently thought that the output from each chain must be less than 500 11 J. A number of techniques can be used to attenuate ASE; examples Include Pockels' cells, fast Faraday rotators, dye cells, chopper wheels, and terminations of these elements. Pockets' cells up-to 25 Dn In diameter are in routine use today. Cells up to 50 mm in diameter and having 10 nsec shutter speeds are currently being designed for use on the ARGUS and SHIVA sysion*;. D>» cell isolators, while available In all aperture si2es, arc undesirable because of large insertion losses (J0:-^0i) even at high power densi ties. In addition, they contribute to beam degradation because of their large non-linear Index of refrac tion. Fast Faraday rotators currently look promising. It is believed thai, rotators up to 200 rn in diamcter wlth less than 4 usee shutter speeds are feasible.'''
Damage to laser fusion targets can also be caused by precursor pulses from the oscillator. Measurements on glass microspheres have shown that (00 ps pulses with energy as low „; IG0 nlcrojoules can cause damage
to glue joints.") if. |0I. example, the targe: is to be irradiated with a 100 Joule compression pulse, it would be necessary to provide 60 dS of prepulse isolation. There are scvoal sources of oscillator prepulse that can be readily identified. The first is sirply leakage throuqh the Pockets' eell-onlarlrcr switchout. This leafage can generally be kept below 40 dB with qood multilayer dielectric polarizers providing the contrast rat*t> of the Poefcels' cell is 2. fln dB; such Pockels' cells are available. A second source is Inter- pulse noise; this of cuurse is not attenuated during the open time of the switchout. Fortunately it is typically down in Intensity {relative to the desired pulse) by 30-00 ofl. Prepulse from switchout jitter arises when the shutter admits a portion of the pulse preceding the pulse selected for switchout.' This source has been substantially eliminated, however, with modern electronic swttchouts. In simnary, about 30 dB of additional isolation 1s required. This can be obtained Kith small dye cells following early pre- anpHfler stages.
In staging a laser chain for maximum focusable output power it is convenient to consider what Is called an amplifier group. S'jch a group is shown in the inset of Figure 2 and is comprised of isolation and ampli fier stages positioned between spatial filters. Ihe isolation stage fs composed of a Faraday rotator and polarizer; it protects the chain from laser energy reflected trom the target. The spstla? filters arc used to remove beam filaments and spatial noise exacerbated oy small scale self-focusing. These devices are typically constructed from a pair of aspherlzed f/10 lenses with a 300 micron diameter diamond pinhole .positioned at lens focus. They are pumped to a hard vacuum to minimize breakdown effects, iuch a filter has a spatial bandpass of about 9 cm-'; this is sufficient to pass the whole-t-eam while stripping higher frequencies from it. T!.e staging methvdology for designing a eiyen amplifier group has been given a great deal of attention at LLL and is discussed 1n several a^ticlest'O/C'). 3 The mans awi sum smr»s The AKUS liter Is a tm-am syste* with 20 en final a=iii;f ler apertures, and Is capable of delivering •ore than 3 TH of focuiaMe power to Ui\>ti «lth dleaeters of a fe* hundred ntcrons. This capability Is planned to be upgraded In the near future. Here we describe the physical characteristics of the laser and associated facility, and sow of the Important coeponents incorporated therein. Functionally, the later system consists of a woe-locked Nd:«AG oscillator and pulse selector, followed by three 12 «w aperture rod arpllficrs; bean profiling, colllNtlng, and splitting optics; and in each am, three 2.3 en * 25 e» rod amplifiers, one * <• x 25 ca rod aictllfier, three 6-dlsk 8.S c» C.A. disk ampli fiers, and four J-dfsk 20 en C.A. disk amplifiers. A scheutlc vlex of the bean pririle as It traverses key chain components Is shown in Figure 3. Isolation (elements not shown) against prepulse. orclase and backward propagation are aceonplIshed with a suitable combination of dye cells, Pockets' cells, and faraday Isolators. Also shown is the bea* dlawtcr through successive stages, mwlnel bean power (rated at 100 pseO at the entra.ee to each spatial filter, and bank energy required for each of the esellfter stages shown. Multiple vacuus spatial filters (5 In each ana) are eaoloyed at strategic Intervals along each chain to control the growth of snail scale bean break-up, and to serve as bean expansion telescopes, The spatial filters employ f/10 aspherlc lenses, dlanond wire die pinholes, and ion pun? evacuation lo residual pressures less than 10"b torr. AflGUS LASER SYSTEM, SINGLE ARM, 12/23/75
C-boost«r
1 1 ~1 Beam dia.. mm 36.0 85.0 85,0 195.0 195.0 280.0 Power © 100 psec, Gw 4.1 40.0 120.0 800.0 1,560.0 1.500.0 Bank energy, KJ 45.0 288.0 144.0 864.0 288.0 - Break up integral, nep. 0.5 0.7 1.1 2.0 2.2 2.0 Whole beam phase, rati. .5 1.2 2.3 4.3 6.5 8.5
Fig. 3. Optical profile of * single ARSUS lfiier chain. Representative values *rc given for the ljs«»r energy and power it each ar&llfier sUge, Also shown are th« clear apertures of tach ticpJlflcr, tto elec trical energy supplied to the flashUrps. and typical whole frea» nonlinear phase retardAt.cn 41 It act ululates in each analifler stage. Chain design anC spatial filter placement it guided by the following crltetla: {1} The (ncrcr-enta! sni!l*scale bean breakup Integra), .'S • |^/vld., between successfvc spatial filters shall not exceed an upper bound (typically ?.S-],0); (2) the final arplificr stage shall be driven to the highest rtui allwd consistent with criterion (I), Koctnal values of & and whole-beta rwnlinttf phase retardation a arc also shown in Hsu-re 3. 6/ employing state-of-the-art optical components, which present a tafnirua of glass pith length to l?w beaa, S Is further t>infal2Cf*. the energy storage systca consists of approxlnattly 3 capacitor tanks.. For reduced cost and coss>lc*ity *n4 Increased reliability, individual ftfj nodules (18 fcJ driving one series Unp pair) are charged In 1.2 HJ blocks and switched in 600 M blocks. tarap Ignition is effected by a controlled ovcrvoltagc ring-up during the first few ctcroscconos of the discharge, to detect incipient flashlssp failure prior to a laser shot. an tuxfliiry eharge-and firos^tten provides all disk amplifier lai$s with a tow energy discharge, inter prets df;rharge parameters, and displays to the operator the condition of each Ijsp pair. A photograph of the AftGUS User In partial construction (Flgtirc $J illustrates the general system layout. In the near foreground arc the 2S-rm aperture rod amplifier chassis; along the center table arc located oscillator no. 1, the switchout unit, and as sec fa ted seal) optics; along each tm various disk afrpflficrs, spatial flues, and Isolator units; the rack in the center contains the oscillator control units, spatial '/ ,'•' f\ 9
filter vacuum pump power supplies, PILC display panels, and a variety of cable patch junctions; and alonn the far wall, the U% cooling gas flow control. An optical sehenvitic of a SHIVA laser chain Is shown in Figure 5. The nominal focusablc power from this chain is expected to be about 1.- terawatts. While performing about as well as an ARGUS chain, this chain is considerably rare cost effective. The cost savings arist because the final 20 en aperture & power amplifier group is driven by the snallor IS en y amplifier. This of course requires an additional spatial filter In the driver section. The staging of the MUM chain represents our latest thinking on hw to nfninlzt; both whole-btan and snail-scale self-focusing while rjxlmizir.g trio focusable power per unit cost.
Fig. 4. View of the ARGUS laser systen 1n partial construction. The target chamber and beam turning optics arc located behind the far hall.
The SHIVA system which has Jiecn dsscrlbcd in several reports and publications' ~* ' is comprised of 20 chains. These Mill be taunted in columns of 5 on a support structure (space frame) constructed fron Steel tubing.*'*5' Figure 6 shows a scale drawing of Cnc laser and target space francs as they apveer in the bay of the High Energy laser Facility at Livenr-ore. In this figure an iccsahedral geometry *s shown for bringing the beans onto the target. Figure 7 shows a plan view of tt.e laser chain. Of particular interest 1s the elfctro-aptical beam alignment system. This system can be divided Into five parts whose functions include oscillator alignment, main chain pointing, bean centering on the final focusing lens and pointing and focusing of the beam onto the target. The purpose of the oscillator alignment system is to maintain collinearity of the master oscillator and a cw a'ljnment laser. This is achieved with 5-axis mirrors at the output of each oscillator. These mirrors are conti.-llod by servo-loops between the mirror axi» and pointing and centering sensor-- located in the preamplifier stage. Since ccmon sensors are used to control both mirrors, collinearity is assumed as long as detector responses are linear for both the pulsed and cw mode. It is required that the relative angular accuracy of these two beams be about ± 2 microradians. 5 (HIVA LASER CHAIN
•* J, 7 T 7
I I """1i ™T ""i1 3.1 n.i 37 90 140 0.031 0.118 0.37 0.90 1.4 150 ISO 160 200
F1g. 5, Optical profile of a single SHIVA laser chafn. Representative values are given for the laser energy And power at each amplifier stage. Also shown is the clear aperture of each amplifier and the elec trical energy supplied tu the flashlamps. SHIVA SPACE FRAME
-Pulied QKillatot
Folding nwror Littr »p*» frame
Beam splitter
Target chamber
FtMt turning mirrOf —
Fig. 6. Mechanical support structure for the user and target sections. The laser frame fs capable of supporting sorce 600 individual component assemblies and maintaining a tolerance o7 + 4 urad be tween any two points over a period of 100 s. The target frame supports the target chamber, turn ing mirrors and the PFC sensors.
The purpose of the main-chain pointing system 1s to provide an angular correction at the apodizlng aperture for errors incurred in the beam splitter array. This is accomplished with a 2-axis angle global placed 1n front of the apodizlng aperture and a pointing sensor placed immediately after the aperture (this sensor is not shu*n 1n Figure 6). Pointing accuracies of +. 100 micrcradlans an: required to keep the beam centered In the main chain spatial filters. The angle gimbal Is part of the same mirror assembly that Is used for optical path length adjustment. 7(502
SHIVA LASER WITH BOOSTER STAGE AND ALIGNMENT SYSTEM iZrCZSEHZHOS
J ***rv » 3&**m9 W ***•"•» HZH?^5^1C^>&£3!BI[ZHf2a M4m *t Fig. 7. Plan schematic of the SHIVA laser chains and the alignment sy ;t«ra. riot shown is the path length equalization sy.tem which 1s in an ear!,} stage of development. Centering of the beam en the final focusing lens 1$ accomplished by translating the iodizing aperture 1n a plane perpendicular to the propagation direction. Because of 10:1 magnification of the beam dt the focus lens position with respect to the apodizing aperture, there will be a corresponding 10:1 ratio between displacement of the beam at the lens position and displacement of the aperture. A centerini accuracy of about i 3 cm is curreitly required. To sense the position of the beam center, a reflective screen pi aceJ in fron^, of the focusing lens relays the beam to an optical Imaging system. This system in turn Images the plane of the lens (screen) onto a quadrant detector. Error signals from this detector are used to drive the trans'.ation axis of the apadizing aperture. The imaging system sensor 1r located behind the final turning mirror and along an axis passing through the center of the focusing lens and the center of the tar 7 References 1. B. C. Johnson and H. 0. Fountain, An Active/Passive Mode-locked Laser Oscillator, Technical Digest, p. 322, International Electron Devices Meeting, 1974, Washington, D.'C, 2. D. U. Phlllion, Private Communication. 3. J. H. Nuckolls, J. L. Etmiett and L. L. Wood.Laser-Induced Thermonuclear Fusion. Physics Today, August, 1973. 4. C t. Thomas and L. 0. Siebsrt, Appl. Optics, j5_, 462, (1976). 5. G. E. Somraargrcn, Private Communication. 6. R. A. Elliott and G. A. Masse,. Pulse Shaping with Satui'abl ''•sorters. Technical Digest, p. 314, International Electron Devices Meeting, 19737 Washington, D.C. 7. R. A. Buchreeder and R. B. Hooker. Appl. Optics, 14, 2476, (1975). 8. J. F. Holarichter. D. R. Speck, and J. E. Swain. Meeting of the American Physical Society, Albu querque, HsM Mexico, October 2831, 1974, 9. H. L. Gagnon, Private Coimunicatlon. 10, J. B. Trenholme, Optics in Laser Fusion. Proceedings of the Society of Photo-Optical Instrumenta tion Engineers, Vol. 69, i.aser" Systems, p. lfis, (1975). Vt. J. A. Glaze, High Ereroy G1a>s lasers. Proceedings of tho Society of Photo-Optical Instrumentation Engineers, Vol. 69, Laser Systems, p. 45, (1975). 12. laser Program Annual Report, UCRl-50021-74, pp. 68-123, (1974), Lawrence Uvenmre laboratory. 13. T. J. G1lnQ-tin, R. C- Godwin, J. U. Davis, H. F. Hagen, C. H. Hurley, G. W. leppclacler, G. J. Linford, i. 0. Myall, w. C. O'Neal, and J. S. Trenhotme, 10 Kilojoule SHIVA laser System for Vusion Experiments at L'l, 1975 1EEE/0SA Conference on Laser Engineering and Applications, Hay zd-30, 1975T 14. C. A. Hurley and J. 0. Myall. High Stability Space Frame for a Large rus1on Laser. Sixth Symposium on Engineering Problems of Fusion Research, Nov. 17-21, 1975.