LA-UR-76-899

I

TITLE: EXPLOSIVELY PRODUCED MEGAGAUSS FIELDS AND APPLICATIONS

AUTHOR(S): C. M. Fowler, R. S. Caird, W. B. Gam andD. J. Erickson

SUBMI’1”1’EDTO: Invited paper, Joint MMM-lntennag Conference,

Pittsburg, PA, June 15-18, 1976. To be publishedin IEEE Transactions on Magnetics

By acceptance of thin artlc:e for publlmtlon. the’publisber recomtlmmthe Cavemmmit’m(Ilcenm) rlght~in aay eapyright and theGovernment and itaauthorizedrepreaentatlveehave unrestricted right to reoraduce inwhole or In part raid ●rticle under ●ny copyrightescured by the pubhoher.

The I.ae Alamoe Sclentlflc LAoratory requeetethatthepubllcher Identify thin mrtlcle●s work performed underthe ●uspkweof the USI!MDA.

sckmtific laboratory

of tho University of California10S ALAMOS, NEW MEXICO 07s4s

/\#R Affirmative Ac!ion/Equal Oppa-!unitT Employe~

7{”,:, . ,f 1“’:1 1)

i ,J.i :,, ,, ,, II!’ ,,.’

hrm No. RIOS4BNUw!

UNITED FITATFNENERGY RliSEAlict[ AND

DIWKt,OiWfItNT AI)hllNIS’l~A’rIONcomw,x W-7WG.HNG, 3rJ

EXPLOSIVELY pRdDU~.ED MEGAGA(JSS FIELDS AND AppLICATIONS*

C, M. Fowler, R. s. cai~d, w, B, Garn and ~. J, EricksonLOS A]amos Scientific Laboratory, university of California, Los A~=s. NH ~ls~s

ABSTRACT

We describe various .axpl~sive magnetic flux cow

pression devices that produce p~]sed megagauss fields,and a rlumber of applications in ~hic~ they have beenused. hong the systems described are relatively sim-ple ones that generate fields lip to 250 T in largefixed volumes, and cylindrical implosion systems thatproduce fields in excess of 1000 T. Small fixed volumeSystems are described that may be used in the labora-tory. They require only small amounts of explosi~~e andcan produce 100 T fields in coils 25 mm long and 10 mndiameter. We discuss measurements made on variousmaterials in mqagauss fields, often at cryogenic temp-eratures, including magnetoresi stance, magnetic suscep-tibility, optical absorption, Faraday rotation, andZeeman splittings. We also discuss experiments [nwhich large magnetic pressures have been used to com-press solid deuterium isentropically. In flux com-pression devices part of the energy of the explosivesis converted to electromagnetic energy. This has ledto their use as compact single-shot high power energysources. At times, it is necessar.~ to transformercouple loads to the device outputs. We describe suc-cessful operation of transformers in 165 T fie!ds, andsuggest that t+ey can operate in much higher fields.

INTRODUCTION

Exploslve flux compression devices are generallydesigned to produce very large magnetic fields or togenerate large pulses of electromagnetic energy. Thepresent discussion, as requested, surveys the highfield program at the Los Alamos Scientific Laboratorypand therefore is limited largely to those de’~ices thatgenerate large magnetic fields in large working volumesand to the experiments that utillze these fields. WefirsL describe several cmmnonly used systems and pre-sent characteristic magnetic field versus time plots.Next described are high field diagnostics and experi-ments in which information is obtained electrically.This is followed by a description of scwcral magneto-optical experiments. We then outline a series of ex-periments in which the large pressures accompanyinghigh magnetic fields are used to compress soliddcuterium iscntropically to state conditions otherwisedifficult to obtain, We conclude this discussion by

describing experiments in w+ich transformers have beenused successfully in pulsed high field envirmmentsgreater than 100 T. This resl.llt is not entirely ex-

pected in view of the extreme effects of pressure andeddy current heating on conductors.

HIGH FIELD GENERATORS

Explosive fiux compression devices con:;ist basi-cally of a cavity defined by surrounding metal con-ductors, some or all of which may be overlaid withexplosives. After an initial magnetic flux is intro-duced in the cavity, the explosives are detonated.The subsequc.nt motion of the conductors leads to com-pression of the flux into regions of smaller crosssection or of lowered inductance. in analogy withengineering convention, such devices are called genera-tors with their explosively driven components termedarmatures. In sane of these devices, magnetic flux isapproximately conserved. Under this condition, bothenergy and current increase as the system inductanceis forccably reduced by the moving conductors.

Schematic drawings of three flux compression gen-erators are shown in Fig. 1. Other types of generatorsarc described in references 1-4 and the referencescited therein. Fig. la shows a cross sectional view ofa strip gene~mator. initial flux is developed in thetriangular generator cavity, bounded by copper sl-rect,and the cylindrical load coil, drilied out of a brassblock. The flux is obtained from capacitor dr;vencurrent supplied through the opening at the top left.Strips of sheet explosive, mounted m both plates ofthe generator, are detonated near p~ak system flux.Motion of the top strip first closes the current inputslot, thus trapping the fiux in the generator volumeand load coil. As detonation proceeds along the stripsthe generator volume is reduced and the curre,lt andmagnetic field increase in the system. The configura-tion of the copper plates during compression is shownby dashed lines at a time when the detonation frontshave progressed to the points Iahel led D. Peak fieldsdevel(,pecl depend on the load coii di.’lmeter and rangefrom 100 T at 19 mm diam to 150 T for smaller dia-meters. Coil length valiationsp from 70 to 100 mm,have little effect on the peak fields developed. Themagnetic field generated in a 16 mm diam load coil isshown in Fig. 2.

The amount of explosive used in this system variesfrom 0.75 to 1.50 kg. We arc presently developing aminiaturized version of this qenerator for use in thelaboratory. To date, fields of lifl T have been qene-rated in load coils 25 mm long and 10 mm diem usingless than 90 g of explosive. The generator body and

load coil ar= formed from a single sheet of copper 0.52rrrn thick. A slotted piece of brass pipe is placed overthe load coil to limit coil expansicn from magneticforces. The devices are not substantial enough fordirect currer!t feed. They are therefore placed withinsimply fabricated capacitor driu?n external CCIIS thatinduce the required Iilitial flux.

Fig. lb shows a two-stage system. The first stageconsists of a pair of strip ~enerators. The second-stage cavity is the triailgular wedge cut out of a brassblock also containing the cylindrical load coil. Thetop plates of the strip generators and the plate overthe second-stage cavity are formed from a single sheetof copper. The first-stage generators are driven bysheet explosive while the second-stage plate IS drivenby a block ~f pressed explosive initiated simultaneous-ly over its rurface. In:tial flux is supplieu to thecomplete system by capacitor driven current through theinput opening at the upper left. At peak initial flux,the first-stage strip generator explosives are initia-ted. As detonation proceeds, flux is concentrated inthe second-stage cavity. The second-stage initiationis timed to start compressing the flux in the triangu-lar cavity at burnout of the strip generators. Peakfields obtained in the cylindrical load coils aga~ndepend upon the coil diameter. The higher field curveIn Fig. 2 shows a field vs time plct obtained in a 16m diam coil. Peak fields of !75 T are typica! for19 mm diam c ils, and about 250 T, for 10 nwn dlamcoils.

Fig. Ic shows two axial views of a cylindrical im-plosion system. Initially, a thin walled metal cylin-(Ier or liner is centered within an armJlar charge of

high expiosive. Initial maqnetic field BO is inducedwithin the Iiner by energizing a pair of coils, ane oneach side of the liner, or a single coil large nnoughto house the entire assembly. A detonation system,such as the rw of detonators shwn , cylindricallyInitiates the explosive to start radial liner cnrn-pression at peak initial flux. The right hand view inthe figure shows the liner at a later stage of com-pression. Conservation of fIux implies a magneticfield buildup inversely proportional co the square ofthe radius. However, field diffusion into the llner,lack of perfect Implosion synrnetry, and the large mag-netic pressures developed limit the peak fields ob-tainable. Still, peak fields in the range 1000-I5OO Twere reported in the first fl~qx compression experimentsdescribed in the literature.~

ELECTRICAL MCASIIREMENTS

Electrical measurements in these systems posespecial problems because of the rapid rates of magnetic

field rise. From Fig. 2, fields increase as fast as107 T/s fOr the Sin9~e-Stage ~ystems and still fasterfor the two-stage systems. In cylindrical implosionsystem,, rates larger than 109 T/s are normal duringsome phases of the implosion. I’lagnetic fields areusually determined by integrating the voltage signalsobtained frcmr small pickup coils of known area that areplaced in the high field regions. It i% especiallynecessary to CJLIard against high voltage breakdown forimplosion system measurements. Here probes “with areasof only 10 rrrn2 will develop potentials in excess of 10kV . Care must also be taken to minimize stray areasthat would contribute additional voltages to the prohesignal. With use of twisted leads up to the activeprobe areat external shielding, and alignment of anynecessary untwisted signal wires parallel to the fieldaxis, our probes generally yield field measurementsaccurate to within 2% for the slow systems and 4% forthe implosion systems. These measurements have beenconfirmed by measuring the Zeeman splittings ofselected transitions and comparing the magnetic fieldsnJseded to ca!culate the obs rved splittings with fieldsdetermined from the probes. %

Great care in terms of technique is also necessarywhen accurate sample voltages are required to determinea physical quantity as a function of field. For exam-ple, we show In Fig. 3 the transverse magnetoresistar,cemeasured for Bi. The high field data was obtainedusing a single-stage strip generator. The sample was4N polycrystal line Bi wire mounted transverse to themagnetic field. To permit longer sample lengths, along rectangular load coil was machined into a brassblock body, with about the same cross sectional area asa 19 mmdiam cylindrical coil. The field pulse wastherefore similar to that shown in Fig. 2, with a peakfield of about 100 T. C~rrent through the sample wasmonitored outside the high field volume and itsmeasurement presented no difficulty. in contra~.p the

true voltage across the sample was obtained only afterthe measured voltage was ~orrected for pickup from thelead configuration. From Fig. 3, the magnetoresistanceincreases linearly to about 25 T with saturation be-ginning to appear at higher fields. In order to mini-mize Joule heating effects ~ current was passed throughthe sample for only 450 ps, the last 50 PS or so repre-senting the time during which the shots were fired.Calculations showed that no appreciable rise in sampletemperature took place uncer these conditions.

Another technique requiring special care is themeasurement of the differential magnet-c susceptibility

kin a rapidly changing magnetic field7~ . The basictechnique employs two pickup loops of nearly equalareas with their planes normal to the field. One is

wound around the sample and gives a signal proportionalto dH/dt + d?l/dt. The second is located s~e distancefran the sample and gives a signal proportional todH/dt. [Ipon subtraction, the result is proportional todM/dt. Dividing this by the signal ~roportional todH/dt, one obtains the differential susceptlbi]itydMi’dH. The usefulness of the technique depends uponthe accuracy in subtracting the tw signals. The indi-vidual loop voltages are USUal~y quite large and mustbe divided before monitoring. By making small adjust-ments in the division ratiO for one of the loop signalssmall differences in cOii areas can be c~pensated toyield an essentially null difference signal. Thesecompensation tests are performed nondestructively onthe assembiy before the high field shot, using onlycapacitor bank discharges to supply the magneticfields. Wi i this procedure we have achieved a differ-ence signal of less than 1% of the signal in eitherloop in sample free impiosion experiments. This pre-cision is sufficient to observe a first or second ordermagnetic transition in many materials. We plan to usethis method to complement optical absorption techniquesto determine the critical field for the spin flop toparamagnetic transition in MnF2 as a function of temp-erature and orientation. We have previously presentedoptical data9~10 that confirm theoretical predictionsthat this transition should occur at Iow temperatureat about ItlG or twice the exchange fieId.

OPTICAL MEASUREMENTS

Our high field optical measurements have beencontined primarily to the detection of visible lighttransmitted through various samples utilizing a highspeed, rotating mirror spectrograph. This spectro-graphic camera covers, in a singIe shot, the cc-npletevisible spectrum over the high magnetic field I“ange.Fig. 4 shows a schematic drawin~ of the system asadapted for a low temperature experiment. intenselight sources, which are reasonably white over mostof the visible spectrum, are produced by shockingtubesof argon gas. The light is transmitted throughthe sample under investigation, passes through agrating that provides spectral resolution, impingeson a rotating mirror that provides time resolution,and finally is recorded on film. Magnetic fields aremeasured simultaneously by a pickup probe mounted inthe annular space between the high field coil and thesample housing assembly. A bridgewire, mounted in thefie!d of view of :he spectrograph, Is electrically ex-pIoded late in time of the experiment to produce aknown spectrum on the film for wavelength calibration.With a standardized helium flow system, the samples

achieve temperatures of 6.5 ~ 0.5 K at the time theexperiments :Ire initiated. Sample diameters are re-stricted to 6-7 m for 200 T devices when the cryo-ge~ic housing is require.i,

We discuss here several representative experimentsthat are based on o tical phenomena.

!described elsewhere-12. Forexample,p~~~~O~s~~~~ “the transmission spectrum for thin platelets of CaSeobtained with the geometry shown in Fig. 4. The c axisof the crystal was aligned parallel to the magneticfield. The first several exc!ton absorption bands canbe readily distinguished from the diffraction fringesvisible at the longer wavelengths. From this recordand several others derived from 200 T systems togetherwith lower field data previously reportedly, we havetentatively adopted the level scheme shown in Fig. 6for this material. Data obtained up to 17.5 T, shown

~S and by H*lpernin the rectan le at the lower ~~ft, were obtained byAoyagi et al. In Fig. 6, we have——used the level notation of Aoyag; et al., \lith the ex-ception of the very weak line previously unreported,labelled n = 1’. The other dashed lines indicate theexistence of more than one absorption level, althoughowing to spectral resolution limitations, the locationsare not very precise.

By nmunting thin polaroid discs on both sides ofthe sample in Fig. 4, Faraday rotation experiments havebeen conducted. Fig. 7 shows a Faraday rotation recordfor a 1.02 mm long crystal of cubic phase ZnSe. Thecrystal was cut In such a way that the magnetic fieldwas perpendicular to the {110’ planes. The unusualcharacter of the record can be understood by followinga particular wavelength along the magnetic field. Asthe field increases, the ptllarization angle of thelight increases. The light and dark regions thencorrespond to 90° rotation increments. The band rever-sals occur at peak magnetic field. Photodensitometermeasurements across the record yield accurate positionsof the pea!:s and valleys as functions of the field.Plots of the total angle of rotation vs magnetic fieldwere straight lines fur all wavelengths between 450 and63o nm even though the clearly seen absorption brndedge is around 440 nm. In cont~ast, similar experi-ments with CdS samples showed a marked curvature of theangles vs field for wavelengths near the absorptionedge., particularly at the higher magnetic fields101120The resulting Verdet constants for ZnSe calclllated fromFig. 7 are shown in Fig. 8.

We have also made high field Faraday rotationmeasurements with discrete wavelength lines, such asthe green and blue linc~ of Hg Photomultiplier tubeswere used as the detectors , while pulsed mercury lampswith appropriate filters served as the sources. Al-

though this approach wa$ abandoned in favor of thebroadband method, we Plan tO initiate similar experi-ments with long wavelength laser sources. BesidesFaraday rotation experiments, we intend to make infra-red cyclotron resonance measurements similar to th seof Herlach et al. .1?.15 and Iluira, Kido and Chikazuml

As a l~t~xample, we describe briefly a seriesof Zeeman effect experiments. The spectra were dis-played photographically as Shown in Fig. 4. The lightsources, however, were electrically exploded bridge-wires placed in the high field region. In some cases,the bridgewires were coated IIyhtly with othermaterials whose known spectra were desired. Al thoughhigh field splittings were observed for many spectrallines, our primary motivation was to use the splittingsof well understood lines as a means to calculate magne-tic fields, thus furnishing a means of checking fieldvalues obtained from pickup probes as discussed above.The largest field measured in this manner was 510 Tcalculated from an observed 16.5 nm splitting of theNa D lines6.

FIP.GNETIC FIELDS AS PRESSURE SOLiRCES

Fig. 9 is a schematic of an experiment used to ob-tain high pressure equation uf state data for soliddeuterium. The experiments are being done jointly withmembers of our laboratory cryogenics group, who de-signed the sample holders and furnist, the deuteriumloaded samples at shot time. The outer vacuum wall ofthe sample holder is nonconducting, while the innerwall is copper. The entire sample assembly fits intoa 16 m diam load coil. The high field system employedis the two-stage system of Fig. lb. As the fieldbuilds up in the annular space between the field coiland the deuterium confining copp~r sleeve, the magneticpressure compresses the sleeve, and consequently, thedeuterium sample. The primary diagnostics consist of afieid probe, which monitors the field during the entirecompression stage, and a 600 kV flash x-ray pulse,aligned to radiograph the cross section of the deuter-ium sample. Limited to a single x-ray exposure duringa shot, the pulse is timed to correspond to a particu-lar magnetic f!eid durini~ the compression stage.

Fig, 10 shows an x-ray exposure of the setup con-figuration, and ~n exposui”e taken durim the implosionphase wt’en the field was 128 T. From the right ordi-nate of F!y. 2, the drive pressure for th;s field is6.5 GPa (6s ‘<h). With use of image qua;~tizing equip-ment, ti’e arsa! comp~cssilm cf the d~uterium deter-mined from iha radiographs was 3.55 ~ O.LOF leadingto a tel~tative state (P,P) point of (6.5 Gpa, 0.71 tO.10 g/cm3).

The interpretation and validity of the data restheavily on calcul~tiona! reSults. A number of onedimensional hydrodynamic calculations were made to in-vestigate the uniformity of the deuteriurn sample overthe cross sectior,, using hypothetical de:~teriumequations of state and copper sleeve drive pressurescorresponding to the field curve of Fig. 2. The cal-culations showed that for the last several microsecondsof compression, the pressures and densities over thedeuterium cross section were quite uniform, and fur-ther, that the deuterium pressure was the same as thatof the magnetic dr;ve pressure to within a few percent.A second result from these calculations was that thecompression was essentially isentropic. This contrastssharply with shock compression to the same pressurewhere, for compressible materials, significant in-creases in entropy and far greater temperatures are ob-tained with substantially less compression. A few twodimensional hydrodynamic calculations were done, againusing hypothetical models for the deuterium equation ofstate. Results from ttese calculations showed thatthere was also negligible axial variation of densityand pressure in the central region of the sample. Thisis the region of greatest compression and defines thesampie cross section: shown in the radiographs.

Tww Iowr pressure data points have also been ob-tained, the lowest of which agrees within our experi-mental error with that of Stewarti7. All data pointsare assumed to belong to the isentrope passing throughthe initial conditions of the experiment (p. = 0.200g/cm3, To = IOK).

So far, we have been iimited in pressure to about6.5 GPa, because of x-ray beam attenuation at thehigher deuterium densities. A more energetic 2 tleVmachine has recently become available, which should per-mit an extension of the data to 15 GPa, probably withbetter volume resolution, and therefore greater densityaccuracy.

Aithough our present data are not very precise,the method seems particularly suited for high pressllreequation of state investigations of soft materials atiow temperatures. As noted above, the use of fixedvolume ioad coils aliows measurement of both pressureand volume, aiong an isentropic thermodynamic path.tluch higher pressures can be obtained with implosionsystems, such as shown in Fig. lc. The pressures, howe-

ver, cannot be monitored during the finai stages ofcompression, because the field measur!ng probes aredestroyed. For this reason, when impiosion systemshave been used in compression experiments, the pres-sures have had to be inferred, such as in our earlyimploding liner plasma compression exPeriments18B or inthe more recent work by Hawke and comrkers’9 in theirattempts to make metallic hydrogen. More accurate high

pressure data have been obtained by shock techniques,but since the thermodynamic paths are such that thedata points are far removed from those obtained isen-tropically, the tm techniques are complementary. Lowtemperature isothermal data can also be obtained inhydrostatic presses, but accurate (p,P) data for deu-terium have been limited to only a few GPa17#20.

HIGH FIELD ENVIRONMENTS

The current densities required to produce largefields lead to destructive effects upon metal conduc-tors. Furth, Waniek and Levine21, in a classic paper,have considered both pressure and temperature effects.Using a simplified analysis, they concluded that thesurface temperature of several different metals placedin a magnetic field B rose approximately as AT(K) = 0.3B2 . They then characterized these metals by a criticalfield above which melting wuld begin. The fieldsgeneralli ranged from about 50 T for copper to about100 T for tungsten. The analysis showed that the co-efficient 0.3 was pulse shape dependent, but not timedependent. Later, more detailed work22 has shown thatthe interaction of a large field with a metal is farmore complicated and is time dependent. Generally,larger magnetic fields can be produced when pulse timesare shorter, such as the ?GOO T and greater fields pro-duced in cylindrical imploslon systems. Here, it isalnmst certain that the inner surface of the liner hasmelted and has perhaps even achieved a plasma state.

Knowing that the actively driven liners were ableto function under these extreme conditioris, it was byno means certain that a passive element, such as atransformer, could survive in a high field environment.To investigate this point, a number of experiments wereperformed in which small, multi turn transformers wereplaced in high field load coils. The first attemptsresulted in failure usually caused by turn-to-turn coilbreakdown before peak fields were reached. It. shouldbe notsd that a shorting of z turn ir: these p~;sed sys-tems effcctiwely exclw’~,s ~urther flux from per,etra?ingthe ::!rdings,, thus terminating the transformer action.The b;eakdown problem vias eliminateri by winding thecoils cm a soft substrate, such as polyethylene.Apparently, softer materials are able to adapt a~ thetransformer windings distort such that turn-to-turn in-sulation is preserved, With primary c~ils driver bys!ngle-stage flux cr.mpression systems (Fig. la),, trans-for-:,ers were show~~ to function properly in peak fieldsas !arge as 110 T. Fig. 2 shows that the field rise

time is abou”: 7 us (from 0.5 max to i~~x). With a two-stage sy’tem (Fig. Ib), successful ‘Li&~sforme~ opera-

tion was achieved in fields of 165 T with field rise

times of 4-5 microseconds. Based upon these results,we now have some confidence that transformers shollldfunction in implosion produced fields of 500-1000 T,where the rise times are less than one microsecond.

The importance of these transformer studies isassociated with the use of transformers :ri flux com-pression devices that are used as power so~rces. Suchdevices are inherently low impedance sources, and arebest suited for driving low inductance loads. Thetransformers serve as Impedance matching transfer sys-tems enabling us to power widely differing loads inclu-ding resistive elements, large inductcrs, and ca~aci-tors. Often the primary load coils can be made largeenough that the resulting fields are sufficiently lowto prevent serious damage to the tr~nsformers. How-ever, there remain sitl’ations where small coils may berequired for compactness or for very fast energeticpulses that can be obtained only by the ultra highfields produced by implosim systems.

We are indebted to many Los Alamos people who havecontributed to various parts of this program, in parti-cular to D. B. Thomson who wrked with us for a numberof years. Our collaborators on the deuterium compres-sion experiments are R. Mills and F. Edeskuty from thelaboratory cryogenics group, and we have received cal-culational and theoretical support on this program fromJ. N. Fritz, G. 1. Kerley, and S. R. Orr. We ars alsoindebted to b’. G. Wittcman who grew the GaSe crystalsused in the optical transmission experiments.

REFERENCES

2.

39

4.

5.

6.

7.

Work done under the auspices of the United StatesEnerw Research and Development Administration.Proc;edinqs of the Conference on Megagauss Maqne-tic Field Generation by Explosives and RelatedExperiments, H. Knoepfel and F. Herlach, Eds.(Euratom, Brussels, 1966).F. Herlach, Rept. prog. phys. 31, 341 (1968)=H. Knoepfel, Pulsed High hlagne~c Fields (North-Holland. Amsterdam, 1970).C. M. F&wler, R. S; Caird, and W. B. Garn, LosAlamos Scientific Laboratory report LA-5890-MS(1975).c. M. Fowler, W. B. Garn, and R. S. Caird, J. APP1.Phys. 31, 588 (1960).W. B. ~rn, R. S. Caird, D. B. Thomson, and C. M.Fowler, Pev. Sci. Instr. 37, 762 (1966).S. Foner and S.-IL. Hou, J~Appl. Phys. 33 (5uPPI .),1289 (1962).

—

8.

9.

10.Il.

12.

13.

14.

15.

16.

1“1.10.

I?.

20.

21.

22.

K. U. Blazer and H. Rohrer~ Phys. Rev. 173, 574(1968) .

—-

R. S. Caird, W. B. Garn, C. H. Fowler, and D. B.Thomson, J. Appl. Phys. ~, 1651 (1971).C. M. Fowler, Science 18o, 261 !1973).W. B. Garn, R. S. Cair~C. M. Fowler, and D. B.Thomson, Rev. $ci. Instr. 39, 1313 (1968).R. S. Caird, C. H. Fowler, J. N. Fritz, W, B. Garn,and D. B. Thomson, Los Alamos Scienti fic Laboratoryreport LA-5065-MS (1972).K. Aoyagi, A. ?4isu, G.. Kuwabara, Y. Nishina, S.Kurita, T. Fukurai, O. Akimotc, H. Hasegawa, H.Shinada, and S. Sugano, J. t“?ys. Sot. Jap. ~(SUPPI .), 174 (1966).J. Halpern, J. Phys. SOC. Jap. 21 (Suppl.), 180( 1966).

—

F. Herlach, J. Davis, R. Sctmnidt, and H. Spector,Phys. Rev. B1O, 682 (1974).N. Miura, G.~ido, and S. Ch’kazumi, Univ. ofTokyo ISSP report No. 723-Ser. A (1975).J. U. Stewart, J. Phys. Chem. Solids ~, 146 (19s6).“Atomic Energy Research in the Life and PhysicalSciences” (report of the U. S. Atomic EnergyCommission, available from th~ Superintendent ofDocuments, Government Printing Office, WashingtonD. ~., 1960), p. 104; see also D. B. Thomson, R.S. i~aird, W. B. Garn, and C, H. fo~ler, in Ref. 1,

p. ;91.R. tiawke, D. Duerre, J. Iiuebel, R. Keeler and H.K; ‘pper, Phys. Earth Planet. Inter. ~, 44 (1972).M. S. Anderson and C. A. Swenson, Phys. Rev. BE,5184 (1974).H. P. Furth, M. A. Levine, and R. W. Waniek, Rev.Sci. Instr. ~, 949 (1957).R. E. Kidder, in Ref. 1, P. 37.

Shoot Qwloaive-Row of detonators

(o) (c)

(b)

Fig. 1. Cross-sectional views of three high field fluxcompression devices, (a) single-stage strip system, ty-pical peak field is I1O T in 16 nrn diam load coil; (b)two-stage system, typical peak field is 200 T in 16 mm

diam load coil; (c) cylindrical implosion system, peakFields exceed 1000 T in diameter of a few rran.

rTwo- stage

d

/

.

[:

.

0“-// ‘Sing’’-s+og

JLJLbLL~24 20 16 IQ e 4 0 -2

17.5

1s,0

12.5

.-

1.0

0

Time before f~eld peek (ps)

Fig. 2. Magnetic field and corrcs~onding magneticpressure developed In 16 nrn aiam load COIIS vs time forthe single-stage and two-stage systems.

ii

40

30

K“

2

Zo

10

I

‘T~ [ I Im L

0- Laboratory data on same Somple.

.- Increasing field

X-Field decoy

~~~-o 10 20 30 40 90 60 ?0 ao 90 100 110

B Field ( Tesk )

Fig. 3. The norms I I zed transverse magnetores i stance ofBi at, about 300 K. Ope~i circles represent data ob-

tained nondestructively in the laboratory. The datarepre~ented by X’s were obtained during tbe decay ofthe field after peak was reached. All other data were

tdken whiie the field was increasing.

AB

c

Fig. 4.eraturc,

K!!!!!?Rod j

c $- IllSch@matic of ttraak spoctrogrophie camera—. \*.~1~

* - \\*\ ●

Grating N.-

Rotating mirror.

B

Schematic diagram of a setup used for low te.mp-high fie.lcl, optical transmission experiments.

-s95,7

-63e,8

I I 111I 5.7 ;3 63 154 l@6173

Flcld (~)

Fig. 5. Light transmis~ ion record obtained for GaSeat about 7 K in a two-stage system load coi 1. Abovethe diffraction fringes at the bottom of the record areseen several of the exciton absorption bands. Theknown spectrum of an exploding bridgewire appears atthe right of the record.

20

-

TE9s 19

18

17

0 100 200

Magnetic Field (T)

6. Tentative exclton Ievcl scheme for GaSe at 7 K.nr:w Ievc I n = 1’ is weak. The other dashed Iincs

cotc poorly resolved band structure.

-405.0

‘424. S

-4S14i

E-479.3 ~

-583.7

-595.7

Fig. 7. Faraday rotation record obtained for ZnSe atabout 7 K in a single-stage strip system load coil.The absorption edge at abouth~[~ nm is clearly evident,

VQrdct Con8tont

ZnSe, -7° K

I

wawlafqj?h (rim)

Fig. 8. The Verdct coefficients for ZnSe obtained fromthe record shown in Fig. 7.

\

High field COII

\

Fieldprobe

\

Fig. 9. Schematic drawing of a system used to compresssolid deuterium isentropically.

effective deuteriumrings define spacessleeves and between

sample area. The dar..( annularbetween the copper and micartathe micarta sleeve and field coi!.