los alamos sciencenumber 24 1996 - jean-pierre petit · 50 los alamos sciencenumber 24 1996 a time...

48

Los Alamos Science Number 24 1996

In the midst of a cold Russian

winter, these Russian and

American experimentalists

attempted to produce a plasma

that was one hundred times

hotter than the surface

of the sun.

The first international conferenceon Megagauss Magnetic FieldGeneration and Related Topics

was held in 1965 in Frascati, Italy. Bythen, Max Fowler,Wray Garn, and BobCaird had alreadyspent the better partof eight years pro-ducing megagaussmagnetic fields. Thesmall group of LosAlamos scientists hadpioneered a techniquecalled magnetic-fluxcompression, whichtakes the energystored in the chemi-cal bonds of high ex-plosives and convertsit to magnetic fieldenergy. The energyis then delivered toan experiment as apulse of either ex-tremely strong mag-netic field or ex-tremely largeelectrical current.Although the LosAlamos magnetic-flux compression ef-fort was relatively modest, Fowler andhis team had achieved considerable suc-cess at building flux compression gen-erators and had already produced mag-

netic fields above 10 megagauss (mega= 106). By comparison, the Earth’smagnetic field is about 0.5 gauss, andthat of an ordinary refrigerator magnet

about 10 gauss. One of Fowler’s motivations for

building these devices was to use theenormous field to contain or compress aplasma. This compression could be a

means of achieving thermonuclear fu-sion (the process by which the sun pro-duces energy), which might make avail-able to the world an unlimited energy

source. Even withoutthat exceptionallypractical goal, Fowlerand his team recog-nized that ultrahighmagnetic fields andintense electrical cur-rents could find ap-plication in the studyof phenomena rang-ing from materialproperties to x-raygeneration.

While thumbingthrough the abstractssubmitted to that1965 conference,Fowler, to his sur-prise, noticed thatsome were from theSoviet Union. Nine-teen scientists wererepresented in eightabstracts, and the So-viets were going todiscuss the generation

of megagauss fields by the technique ofmagnetic-flux compression.

“That was the first time I had seenanything of their work.” said Fowler.“We had certainly never met any of

Number 24 1996 Los Alamos Science 49

Lab-to-LabScientific Collaborations Between Los Alamos and Arzamas-16

Using Explosive-Driven Flux Compression Generators

Stephen Younger, Irvin Lindemuth, Robert Reinovsky, C. Maxwell Fowler, James Goforth, and Carl Ekdahl

A smiling Steve Younger congratulates Russian delegation leader Alexander

Bykov after the success of the first, collaborative, nonweapons-related scientific

experiment to be carried out on U.S. soil. The experiment, performed by Russian

nuclear-weapon scientists and their Los Alamos counterparts, occurred in De-

cember 1993. On the left is Russian translator Elena Gerdova.

them. It was strange, because theirwork seemed to be of the same scopeas ours, and they were alluding to thesame problems and the same solutions.”

But the papers referenced by thoseabstracts were never submitted to theconference. No Soviet scientists at-tended, and the international communi-ty was left with only a tantalizingglimpse of the Soviet research program.

The Russian Magnetic-FluxCompression Program

We now know that the Soviet workhad begun as early as 1951 when An-drei Sakharov, one of the premier sci-entists of the Soviet nuclear weaponsprogram and winner of the 1975 NobelPrize for peace, had sketched out anidea for compressing magnetic flux andgenerating high fields or currents. LikeFowler, Sakharov was seeking a meansto achieve thermonuclear fusion, and hehelped identify several schemes inwhich high magnetic fields could poten-tially help the fusion process. Some ofthe schemes were purely for researchpurposes, whereas others could poten-tially be used for weapons work.

Sakharov’s ideas initiated a programinvolving some of the best Sovietweapons scientists, and an intense ef-fort was devoted to the development ofthe high-field and high-current genera-tors required to implement those ideas.The work was performed at Arzamas-16, the secret city that harbored the All-

Russian (formerly the All-Union) Sci-entific Research Institute of Experimen-tal Physics (VNIIEF), the SovietUnion’s first nuclear weapons laborato-ry. Initially, much of the experimentalflux compression work was carried outby Robert Lyudaev, who in 1952 suc-ceeded in producing a magnetic pulseof approximately 1.5 megagauss. (Inthese explosive-driven flux compressionschemes, the entire experiment is overin less than a millisecond. The field orcurrent pulse rises “slowly,” thenquickly reaches peak value in the fewmicroseconds before the generator isdestroyed. In general, this research isreferred to as high-explosive, pulsed-power research.)

Lyudaev’s work was extended andadvanced by scores of skilled Russianscientists, including AlexanderIvanovich Pavlovskii and VladimirKonstantinovich Chernyshev, scientistswho more than three decades laterwould play pivotal roles in establishingscientific collaborations between theRussian Federation and the UnitedStates. Pavlovskii eventually refined agenerator, called the MC-1, to the pointthat it could reliably and predictablyproduce magnetic fields in excess of 10megagauss. This was about the samefield magnitude produced by Fowler’sgenerators, but it was established in alarger and therefore more useful vol-ume. Chernyshev’s team developed aflux compression generator, called theDEMG, that could produce currents ex-ceeding 200 megamperes. The Russian

investigation into magnetic-flux com-pression continues to this day.

The Russian-American Pulsed-Power Collaborations

The independent development of theLos Alamos and Soviet pulsed-powerprograms represented something of ananomaly within the framework of mod-ern science. Basic research is difficultand success often elusive, and the freeexchange of ideas is vital. Yet herewere two groups that were unable tocommunicate, much less exchangeideas. Despite the fact that flux com-pression generators were primarily usedfor pure scientific research, these de-vices could potentially aid in weaponsdevelopment.† In the suspicion-chargedatmosphere of the cold war, potentialthreats to national security supersededthe desire for scientific exchange.

But times and situations change, andwhen the second Megagauss conferencewas held in Washington, D.C. in 1979,some fourteen years after the first con-ference, Soviet research papers were ac-tually presented. However, neitherPavlovskii nor Chernyshev nor theirteam members were allowed to attend.Instead, a close colleague of theirs read

Lab-to-Lab Scientific Interactions

50 Los Alamos Science Number 24 1996

A time sequence of an explosive-driven, magnetic-flux compression experiment per-

formed at the Ancho Canyon site in Los Alamos. The time elapsed between the first

and last photos is on the order of fifty milliseconds.

†A ten-megagauss magnetic field can exert anenormous pressure on a conducting material, onethat is exceeded only by the pressures achievedin a nuclear explosion. The generators cantherefore be used to study weapons materialsand evaluate diagnostics without detonating anuclear device.

a number of papers that were of interestthe non-Soviet scientists in attendance.

Communication and interactions be-tween the Los Alamos and Arzamas-16pulsed-power groups gradually in-creased during informal meetings atsubsequent conferences. Fowler firstmet Pavlovskii in 1982 at the thirdMegagauss conference in Novosibirsk,U.S.S.R. The two scientists had beenindirectly influencing each other’s workfor more than a decade, but now a per-sonal relationship developed betweenthe two men. With Fowler’s assistance,Pavlovskii visited both the UnitedStates and Los Alamos for the first timein 1989.

Megagauss-V was held later in 1989in Novosibirsk. Pavlovskii, who wasnot in attendance due to health prob-lems, had a letter delivered to Fowlerthat raised the issue of a joint researchprogram for producing fields in the 20to 30 megagauss range. The sugges-tion, though informal, was a recognitionof the obvious. Faster progress wouldbe achieved by both groups through acollaborative effort, and both groupswould benefit.

Megagauss-V was also where BobReinovsky and Irv Lindemuth of LosAlamos met Vladimir Chernyshev forthe first time. The Los Alamos and So-viet teams were by then well acquaint-ed with each other’s publications, andthe meeting led to several speculativediscussions about the possibility of fu-ture collaborations. The talk becamemore serious at the 1991 InternationalPulsed Power Conference, held in SanDiego, and culminated in September ofthat same year when Chernyshev andVladislav N. Mokhov met with Linde-muth in Moscow and presented a writ-ten proposal for a formal collaborationon thermonuclear fusion research usingflux compression generators.

The Soviet proposal called for a gen-erator to create a large magnetic fieldthat would be used to implode a liner,which is a hollow metal cylinder. Theliner would surround a dense, hot, plas-ma that would be created in a secondmagnetic field. This method of prepar-

ing a “magnetized” plasma was notakin to any method then being pursuedin the United States. Imploding theliner would potentially compress theplasma to the very high densities andtemperatures needed to initiate ther-monuclear fusion. This speculative fu-sion scheme is known as MAGO in theSoviet Union. The collaboration pro-posal was signed by VNIIEF DirectorVladimir Belugin and, evidently, hadthe support of Yuli Khariton—the “So-viet Oppenheimer”—as well as high-ranking officials from the Soviet Min-istry of Atomic Energy. However,although the Soviets were willing toshare with the Americans the results oftheir pulsed-power program, includingtheir MAGO thermonuclear fusion re-search, the global political climate waschanging so abruptly in the latter partof 1991 that the formal proposal wentunanswered by the United States gov-ernment.

In fact, the political climate turnedsevere with the collapse of the SovietUnion in December of 1991 and theRussian Federation’s subsequent rapiddecline towards economic chaos. With-in the nuclear cities, the formerly elitenuclear weapons scientists were sud-denly facing food-distribution problemsand shortages of medical supplies. Itwas perceived by many in the West thatthe situation was becoming unstableand could potentially result in break-downs in the security that safeguardednuclear weapons and materials. Manyfeared that weapons of mass destructionor fissile materials could be stolen orsold to rogue nations or terrorists.President Bush himself was deeply con-cerned about the possibility of the so-called “brain drain,” wherein nuclearweapons scientists would migrate toand work for other countries.

Los Alamos Laboratory Director SigHecker, aware of the various overturesextended to Los Alamos scientists bythe Arzamas-16 scientists, pointed outto then Secretary of Energy AdmiralWatkins that perhaps the Russian labo-ratory leaders themselves knew the bestway to keep their scientists at home.

That simple acknowledgment, andWatkins quick approval, led directly tothe Laboratory Directors’ exchange vis-its in February of 1992.

The Directors’ exchanges would formthe beginnings of the “lab-to-lab” collab-orations between the United States andRussian nuclear laboratories. Scientifi-cally, this program was for the purposeof conducting pure research, and was notdirected towards the development of anyweapon, fusion or otherwise. TheAmericans, and presumably the Rus-sians, came to recognize that the techni-cal advances that could emerge from theresearch would have a minimal and re-mote risk of being applied to weaponsthat posed a threat to either country.

Instead, the collaborations wouldhave the positive effect of infusing asmall amount of money into the Russ-ian complex. This would help stabilizethe financial situation and help keep theRussians scientists working. The Unit-ed States would also reap the benefitsof scientific exchange with world-classresearch institutions. It is interesting tonote, however, that although the Direc-tors’ exchange formally cut the ribbon,the bridge that spanned the East-Westpolitical gulf had been built by scien-tists reaching out to one other. Afriendly handshake between MaxFowler and Alexander Pavlovskii wastransformed into a tangible link be-tween Russian and American scientists.

Irv Lindemuth, Bob Reinovsky, MaxFowler, and Stephen Younger visitedArzamas-16 in June of 1992. Duringthat visit, Younger, then the ProgramDirector for Above-Ground Experi-ments, suddenly found himself elevatedto the role of negotiations point man.Younger succeeded in forging an agree-ment that laid out the rules for the lab-to-lab program. The Russians wouldprovide manpower, expertise, andequipment for joint experiments. LosAlamos would finance part of the ex-periments and would complement theArzamas-16 devices with its significantexpertise in fast diagnostics, recordinginstrumentation, and supercomputermodeling.



It was agreed that two experimentalcampaigns would initially be conduct-ed, the first to take place at Arzamas-16. That experiment would testChernyshev’s DEMG high-current fluxcompression generator. The Russianscientists would then come to LosAlamos and help conduct an experi-mental series in superconductivity usingPavlovskii’s MC-1 generators that hadbeen purchased by Los Alamos. Thecontract establishing the lab-to-lab col-laborations was signed at Los Alamosin November 1992.

That initial contract and the diversecollaborations that developed from it (in-cluding an on-going exploration of theMAGO fusion scheme) signified a mani-fest thawing of Cold War relations and atrue shift in the respective roles of thelabs. But another, more personal thaw-ing took place as well. After more thanforty years of mutual distrust and enmi-ty, Russian and American weapon scien-tists were going to work together as col-laborators and “side-by-side as equals.”

The remainder of this article de-scribes some of the experiments thatwere performed between 1993 and1995. All of those experiments neededmegagauss magnetic fields or megam-pere electrical currents to achieve theirobjectives. There will be a briefoverview of the principle of magnetic-flux compression that is the basis forultrahigh magnetic field or current generation, followed by a cursory de-scription of several types of flux com-pression generators. The article willthen proceed to describe five differentseries of experiments that used thosegenerators.

The Principles of Magnetic-Flux Compression

Early in the nineteenth century,through the work of Oersted, Ampere,and others, it was recognized that anelectrical current always generated amagnetic field. The size of the currentdetermined the field strength, and thefield always pointed in a direction that

was at right angles to the direction ofcurrent flow (Figure 1).

Although many physicists during the1820s were aware that currents were the

source of magnetic fields, it wasn’t until1831 that Michael Faraday showed theconverse to be true; a changing magnet-ic field generates an electric field that



Fig. 2/magnetic flux4/19/96

Magnetic field strength = B

Surface area = A

Flux = Φ ~ B × A

Top view shows that the area of theloop encompasses nine field lines.Uniform magnetic field

Definedsurface

Wireloop

Field line

Figure 2. Magnetic Flux In general, magnetic flux is calculated by integrating the perpendicular component of a

magnetic field passing through a surface over the area of that surface. For the uniform

magnetic field shown in the figure, the calculation is greatly simplified. The surface is

the inside of the circular loop of wire, and the flux is simply the field strength times the

area of the loop. Because the field strength is represented by the density of magnetic

field lines, the flux is represented by the number of field lines. (Flux = number of lines

per unit area × area = number of lines.)

Figure 1. Magnetic Fields and Electrical CurrentsCurrent (red) flowing through a straight wire creates circular magnetic field lines (blue).

The field lines are drawn such that the field strength is indicated by the density of the

lines (number of lines per unit area). Thus, the magnetic field strength decreases with

distance from the wire. The direction of the magnetic field can be found by the “right

hand rule.” If the thumb of your right hand points in the direction of current flow, the

magnetic field lines will point in the direction that your fingers curl. The magnetic field

created by a current-carrying solenoid exits from one end of the coil and circles

around to enter the other end. The field on the inside of the solenoid is relatively

strong and uniform (equally spaced, dense field lines), whereas the field decreases in

strength and is nonuniform outside of the coil.

Current

Magneticfield lines

Magneticfield direction

Currentdirection

Fig. 1/mag field & elec curr4/19/96

+ -Current

SolenoidMagneticfield lines

causes a current to circulate in a con-ductor. Faraday summarized his obser-vations by stating that a change in the“magnetic flux” that threaded a loop ofwire would generate an electromotiveforce, that is, a voltage, which would in-duce current to flow.

Figure 2 illustrates the concept ofmagnetic flux. Although the flux can bedefined and calculated for any arbitraryconfiguration of field and conductors, asimple case is shown in the figure.There, a uniform magnetic field passesstraight through a circular loop of wire.The flux in this case is simply the fieldstrength times the area of the loop.

As described at the start of this sec-tion, a current is the source of a mag-netic field, so that if the flux thatthreads a loop changes, and the changeinduces a current to flow in the wire, anew magnetic field is also induced.Faraday demonstrated that the directionof that new field counteracts the changein the flux (a phenomenon that hadbeen described, but not quantified, byLenz’s law). In other words, attemptsto change the flux through a conductingloop are counteracted by the inductionof currents and fields. The inducedfield points in a direction that negatesthe flux change.

Suppose our loop is made from a per-fectly conducting material, meaning thatcurrents can circulate around that loopwithout losing energy. For a perfectlyconducting loop, a change in the fluxwill induce a current that will be of suf-ficient strength to exactly counteract thechange. As illustrated in Figure 3, theflux before and after will be the same,and the flux is said to be conserved.

Most materials are not perfect con-ductors but have some resistance. Cur-rent flowing through a copper or alu-minum wire loses energy, which isdissipated as heat. An induced currentwill continuously decay at some charac-teristic rate (which depends on both theresistivity of the material and the “in-ductance” of the loop), and therefore,the induced magnetic field also decays.It becomes unable to counteract the fluxchange. A loop made of one-millimeter

thick copper wire at room temperatureand a few centimeters in diameter willmaintain a constant flux for less than amillisecond. On the time scale of anexplosion, however, which may lastonly a few microseconds, that loopmaintains flux quite well. Thus, onshort time scales, shorter than the char-acteristic decay time, even normal ma-terials approximate perfect conductors,and flux is approximately conserved.

Suppose that instead of changing the

magnetic field through the loop, theloop itself is changed and shrinks insize. The flux, which is proportional toboth the field and the area, should de-crease, but again, currents are generatedin the conducting loop that create a newmagnetic field. The induced fieldpoints in the same direction as the orig-inal field to counter the flux change,and the total strength of the fieldthreading the loop increases.

This is the way ultrahigh fields and



Fig. Flux conservation3/21/96

Uniform magnetic field,strength |B|

Reduced magnetic field,

strength |B|2

Surface area A

Current

New field lines

New magnetic field configuration

Figure 3. Faraday’s Law and FluxConservation

An external magnetic field (blue lines)

threads a closed, perfectly conducting

loop. Nine field lines, which represent

the flux, thread the loop.

The external magnetic field is reduced to

half its value, such that only five external

field lines pass through the loop and con-

tribute to the flux. This change in flux in-

duces a current in the loop, which gener-

ates a new magnetic field (green lines).

The current flows in such a direction that

the induced magnetic field adds to the

external field. The induced field negates

the flux change, and the total flux

through the loop is maintained (four

green field lines plus five blue equals

nine field lines).

Summing the external field and the in-

duced field gives the final field configura-

tion. The distribution of the magnetic

field through the loop has changed, but

the total amount of flux is conserved.

ultrahigh currents are created. A fluxcompression generator may use a hol-low metal pipe instead of a loop, and aportion of an external field will godown the center of the pipe. High ex-plosives, arranged symmetricallyaround the pipe, are detonated, and thepipe is rapidly compressed by the pres-sure of the explosion. The pipe wallcollapses towards the axis. On theshort time scale of the explosion, theflux is approximately conserved and re-mains relatively constant as the pipecross section shrinks (Figure 4). Theflux is “compressed” because the sameamount of flux now occupies a signifi-cantly smaller area. To maintain thetotal flux, the magnetic field strengthgets greatly enhanced, and that increas-ing magnetic field, in turn, generates alarge current in the collapsing wall.

The high explosive plays a dual rolein this scheme. First, it collapses theconductor so quickly that flux conser-vation is approximately true. Second,it is a source of energy. Energy isstored in a magnetic field and theamount of energy is proportional to thesquare of the field magnitude (B2). Be-cause the field magnitude increases, theenergy content must also grow. Thatenergy comes from the chemical ener-gy stored in the molecular bonds thatmake up the explosive material. Whenthe explosives are detonated, energy isreleased and does work on the conduct-ing surface, so that it collapses. Theconductor, in turn, does work on thefield by compressing the flux, and theultimate repository for the releasedchemical energy is the magnetic fielditself.

Regions of high energy density wantto expand and equilibrate with regionsof lower energy density. A magneticfield of high energy density will, there-fore, exert a physical pressure againstany barrier that is trying to contain orexclude that field. The magnetic pres-sure also scales as B2, and for the hugefields created by these flux compressiongenerators, that pressure is enormous.A 1-megagauss field exerts a pressureof about 40,000 bar (a bar is about



��

Fig. 4/exp-driv flux comp4/19/96

R0

R1

��

Φ0 ≅ B0 πR02

Φ1 ≅ B1 πR12

Φ1 = Φ0

B1 = B0 R0

2

R1

High explosiveHigh explosive

Highexplosive

SolenoidElectriccurrent

Slit

ClosedSwitch

Magneticfield

Copper cylinder

Fig. 5/Sakharov/Lyudaev4/19/96

��

��

+_ Remote

capacitorbank

Figure 4. Explosive-Driven Flux CompressionA magnetic field is established within the interior of a metal pipe. The boundary for

the flux is the pipe wall, and the surface that defines the flux is the cross-sectional

area of the pipe. On short time scales, magnetic flux is conserved so that rapidly im-

ploding the pipe and reducing the interior area compresses the flux (the density of

field lines increases). Thus, although (ideally) the flux stays the same, the total mag-

netic field strength increases.

Figure 5. An Early Flux Compression GeneratorThe central copper cylinder is cut by a long slit, so that it is not initially a closed con-

ducting surface and currents cannot circulate around its circumference. Flux cannot

be conserved. When the remote capacitor bank is discharged and current runs

through the solenoid, an initial magnetic field is easily established inside of the cylin-

der. Detonating the high explosives compresses the cylinder, and the slit closes. It is

now a closed surface that conserves the flux. As described in the text, the magnitude

of magnetic field inside the cylinder increases rapidly.

14.7 pounds per square inch), whichwill easily cause metals to buckle anddeform. Between 1 and 2 megagauss,the pressure will cause the surface of aconductor to liquefy and vaporize.Above 2 megagauss, the vaporizationoccurs so rapidly and violently that thesurface of a conductor is blasted offand shock waves penetrate into the ma-terial. A 10-megagauss magnetic fieldexerts on a conducting surface a pres-sure of 4 megabars, or 60 millionpounds per square inch! This is largerthan the pressure values existing in thecenter of the Earth (3.7 megabars).

Figure 5 shows the type of flux com-pression generator built by Robert Lyu-daev. This device is very similar to adesign published by Fowler and his LosAlamos team in the proceedings of a1961 conference on high magneticfields (see Further Readings, page 66,third reference). The device used a so-lenoid to establish an initial magnetic

field inside a copper cylinder, and thecylinder was then imploded. The fluxwas compressed inside the metal cylin-der, and the initial field was amplifiedby a factor of 10 or more. The peakvalue of the resulting transient mag-netic field was estimated to be about1.5 megagauss.

Fowler’s and Lyudaev’s early gener-ators, as well as Pavlovskii’s MC-1 gen-erator, were intended to use the highmagnetic field directly on an experimentthat was placed within the central cylin-der of the device. But as previouslymentioned, the high magnetic field in-duces a large current in the collapsingconductor, and that current can be theintended output of the generator. Ingeneral, the design of a generator willdiffer depending on whether it is to de-liver a high magnetic field or high cur-rent to the experiment. A helical gener-ator, shown in Figure 6, is designed todeliver high current to a load located

outside of the explosive region of thedevice. Often, helical generators areused as the first stage in a multistageflux compression scheme. The highoutput current is used to establish anew, very high initial magnetic field ina second generator.

Before leaving this section to discussthe various experiments, there is onefinal point to be made. These experi-ments are true one-shots deals. Thegenerators work because high explo-sives are detonated, and therefore, theentire experiment must be completed insubstantially less than a millisecond,after which time the generator and mostof the experimental apparatus is com-pletely destroyed. This places stringentconditions not only on the type of phenomena that can be investigated, butalso on the reliability and predictabilityof the generator and experiment diag-nostics. One does not have a secondchance.



��

Fig. helicalgenerator3/21/96

+−

+−

+−

Metal Armature

Solenoid

Load

Load switch

CapacitorBank

HighExplosive

Figure 6. Helical generatorA helical generator has a long metal ar-

mature that is packed with high explosive

and placed within a solenoid. As the ca-

pacitor bank discharges, the current gen-

erates a magnetic field in the space be-

tween the solenoid and the armature.

The load switch is initially in the closed

position, preventing the current from

flowing through the load.

The explosive is detonated at one end,

and the armature expands—like inflating

a long balloon. The volume between the

solenoid and the armature decreases in

both the radial and longitudinal direc-

tions. This causes the magnetic flux to

be compressed. Flux conservation re-

sults in an enhanced magnetic field,

which induces a large current in the re-

maining loops of the solenoid.

At peak flux compression, the load

switch is opened, and a greatly enhanced

current is delivered to the load.

The DEMG

The first scientific experiment con-ducted jointly by the nuclear-weaponslaboratories of the United States andthe Russian Federation occurred at ahigh-explosive facility at Arzamas-16on September 22, 1993—the day afterPresident Yeltsin sent tanks to surroundthe Russian White House. (The LosAlamos contingent, consisting of all theauthors except Max Fowler, plus LynnVeeser, Pat Rodriguez, and Jim King,

tried to ignore the growing political cri-sis as they completed the final prepara-tions for the experiment.) The objec-tive of the experiment was to verify theperformance of the unique high-currentgenerator, the Disk-Explosive MagneticGenerator (DEMG) developed byChernyshev, that could potentially beused for the MAGO plasma compres-sion experiments, as well as other high-energy-density physics experiments.

The DEMG has no counterpart inthe United States, and its properties and

operation were unknown. Althoughsmall models of the DEMG had beenbriefly described at the Megagauss-IIIconference (1983), it was not untilMegagauss-V (1989) that the full powerof the DEMG was revealed.

The device, shown in Figure 7, hascylindrical symmetry and consists of aseries of concave conducting disks thatare stacked together in pairs, like op-posing pie pans. Magnetic flux istrapped in the space between two disks.Detonating the DEMG collapses the



Figure 7. The Disk Explosive Mag-netic Generator (DEMG)The DEMG consists of pairs of concave

conducting disks that are stacked togeth-

er. A device of 15 disks is shown. It has

cylindrical symmetry about the labeled

axis. Current flows as indicated by the

red line, and an azimuthal magnetic field

is established within each toroidal disk

cavity. When the DEMG is detonated, the

explosion begins on axis and proceeds

radially outward. As the disk cavity col-

lapses, the magnetic flux within it is

compressed and pushed into the

thin region at the outer circum-

ference of the device. That

region is bounded by

conducting surfaces, so

when the flux density within

that space rapidly increases, a huge

current is induced to flow. When a fuse

opening switch is used, the current caus-

es the fuse to melt and open. At the

same time, the load switch is forced shut.

The current is then delivered to the load,

which is often a liner (see below).

Magneticfield

Current

Magneticpressure: B2

8π

Fig. implosion package3/21/96

Thin metalliner Imploded

liner

Figure 8. Implosion of a LinerA liner is a hollow cylinder made of metal.

Initially, there is no magnetic flux inside

the cylinder. When an intense current

pulse from a generator (represented by

the single red line) passes down the walls

of the liner, a large magnetic field is creat-

ed. The inside of the liner remains at

zero field due to flux conservation and

field exists only on the outside. The mag-

netic pressure drives the liner inward.

Fig. 7/The DEMG4/18/96

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

Loadswitch

Disk cavity

Input fromhelicalgenerator

Input fromhelicalgenerator

Current inCurrent out

Implosionpackage(liner)

��

�

��

�

��

�

��

�

��

�

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

�

��

�

��

�

��

�

��

��

��

��

��

��

��

��

��Disk

cavityDisk

cavity

20 cmFuse-openingswitch

Magneticfield

Current

Thin metalliner

Magneticpressure:|B|2

8πImplodedliner

Axis ofsymmetry

Axis ofsymmetry 15-element disk generator

Copper

High explosive

Detonator

Initial current path

Final current path

��

Explosive-driven loadswitch

disks, and the magnetic flux is com-pressed into a thin region bounded by aconducting cylinder. The enormouslycompressed magnetic flux generates ahuge current in that conducting surface,and this current can be delivered direct-ly to the experiment or “stored” forsubsequent, rapid delivery to the exper-iment using a fast-opening switch.

For the 1993 DEMG test, a capacitorbank provided the initial current to cre-ate a magnetic field in a helical genera-tor. The helical generator amplified thecapacitor’s output current of approxi-mately 20 kiloamperes to the 6-megam-pere current required to power the mainDEMG. That device had fifteen disksof 0.2 meter radius. It was to generatesome 60 megamperes and deliver asmuch as 35 megamperes to a cylindri-cal aluminum liner, 2 centimeters longand 6 centimeters in diameter.

A high current pulse sent down theliner creates a large magnetic field that,for a short time only, exists on the out-side of the liner wall (Figure 8). Thelarge magnetic pressure drives the linerinward at huge velocities (up to hun-dreds of kilometers per second for verylight liners). Diagnostics placed insidethe liner at different azimuthal anglesor different axial positions can detectthe liner’s arrival, and hence, measurethe symmetry of the implosion. Theliner can be in a solid, liquid, or plasmastate as it implodes, depending on theamount of heat generated by the currentand field. Shock wave phenomena, hy-drodynamics, and material propertiescan all be studied with this type ofelectrical load. For this experiment, theliner was simply a well-understood andconvenient diagnostic.

To improve the timing of the currentdelivery, a thin metal fuse was addedthat initially allowed the DEMG outputcurrent to be diverted away from theliner. When the current reached a criti-cal value, the fuse melted. The highcurrent was then delivered to the linerin less than 1 microsecond.

The rate of change of the currentand pulse shape were measured at vari-ous points along the DEMG using

VNIIEF-built probes (mostly tiny pick-up coils called B-dots, which measurethe rate of change of the magnetic fluxproduced by the current). Los Alamosfielded two current probes (Faraday ro-tation probes, described in the followingsection) that allowed a more precisemeasurement of the DEMG’s perfor-mance than had been previouslyachieved. The result of the experiment,shown in Figure 9, agrees with modelpredictions calculated using Los Alamoscodes and parameters provided by theRussian scientists. But a probe locatednear the liner indicated that there was apartial failure in a transmission line, sothat only 20 megamperes of the DEMGoutput was delivered to the load.

Still, the disk generator worked as theRussians had described in the literature,and this first collaborative experimenthelped allay many lingering suspicionsthat existed within both camps. Whatremained was an atmosphere of enthusi-asm, for it was clear that after years of

parallel but separate research, scientistswith similar backgrounds, interests, andgoals were working together.

Measurement of the CriticalField of YBCO Superconductor

At the end of 1993 and two monthsafter the DEMG experiment was per-formed at Arzamas-16, a group ofeight Russians came to Los Alamos,bringing with them five MC-1 genera-tors that had been purchased by LosAlamos as part of the November 1992agreement. The MC-1s (Figure 10)were used in a series of experiments tomeasure a key parameter of high-tem-perature superconductors. Unfortunate-ly, the principal developer of the MC-1,Alexander Pavlovskii, had died in Feb-ruary of 1993 and did not live to seecome to fruition the collaboration forwhich he had worked so hard.

A superconductor is a material that



10

8

6

4

2

0

-2

-4

-60 5 10 15 20 25

Time (microseconds)

dI/d

t (m

egam

pere

s/m

icro

seco

nd)

Model

Experiment

Fig. 9/DEMG output4/19/96

Figure 9. Output from the DEMGThe rate of change of current was measured by a probe that was located near the

transmission line that led to the load. Although there are some discrepancies in the

late time behavior, clearly the DEMG worked as predicted. The theoretical curve is

from a Los Alamos computer model developed by Bob Reinovsky, which was run

using input parameters provided by the Russian pulsed-power group.

when cooled below a certain criticaltemperature, Tc, experiences a suddendrop in its electrical resistance to im-measurably low values, and a directcurrent moving through a superconduc-tor flows with no energy dissipation.How and why superconductivity occurswas described by Bardeen, Cooper, andSchrieffer in 1957 when they publisheda detailed microscopic theory of super-conductivity.

The cornerstone of the BCS theory isthat, below Tc, electrons with equal butopposite momentum and opposite spinform what is called a Cooper pair. Byforming a pair, the two electrons lowerthe sum of their total energy, and thus,pair formation is energetically favorable.

Below Tc, a macroscopic number ofelectrons condense into paired stateswith total spin zero. This means thatthe pairs obey Bose statistics, and theentire ensemble of Cooper pairs can oc-cupy the same quantum state and ex-hibit collective behavior. It is the col-lective behavior of the Cooper pairsthat leads to resistanceless current flow,often called supercurrent flow.

To understand the supercurrent, firstconsider the normal current flow due tounpaired electrons moving through amaterial’s crystal lattice. The electronswill scatter from atomic defects in thelattice and lose energy. An analogy is toconsider the defects as bumps in an oth-erwise smooth road, and to consider thefree electrons that make up the normalcurrent as cars driving down the road.Each time a car encounters a bump, itslows down or changes direction. Thecars encounter “resistance” to theirmovement.

In the collective state, the cars areall jammed together, front-to-back andside-to-side, forming a pack. Withinthe pack, cars are linked together as“Cooper pairs” (although the cars form-ing the pairs are not necessarily rightnext to each other). The entire packspeeds down the road, each car movingwith the exact same velocity as all theothers. Small bumps cannot affect themomentum of this single, collective“state,” and the cars move down the

road without resistance. Analogies not withstanding, the col-

lective state can be broken. The attrac-tive interaction binding Cooper pairs to-gether is very weak, and above thetemperature of absolute zero, thermalenergy is often sufficient to cause pairsto break. As the temperature of the ma-terial rises, the number of Cooper pairsdecreases, until above Tc, all Cooperpairs are broken and a normal currentflows through a resistive material.

A magnetic field can also destroy thesuperconducting state. Above a fewhundred gauss, magnetic fields willpenetrate most superconductors in theform of quantized vortices, which arecircular tubes of circulating supercur-rent. At the core of the vortex, super-conductivity is suppressed over a radiustermed the “coherence length,” which isroughly equal to the size of the Cooperpair. As the applied magnetic field in-creases, the density of vortices increas-es proportionally.

At an external field value referred to

as Hc2, the cores overlap and the super-

conducting state is destroyed throughoutthe entire sample. Thus, Hc2

establishesthe highest field in which a supercon-ducting device can be operated withoutreverting to the “normal” resistive state.From an engineering standpoint, estab-lishing the magnetic field dependence ofa superconductor is extremely impor-tant. From a research standpoint, Hc2 isrelated to the size of the vortex core orthe coherence length, and knowing itsvalue and temperature dependence,Hc2

(T), is of great theoretical interest.Prior to 1986, all of the conventional

superconductors had to be operated ator near liquid helium temperature (4.2 kelvins), and that required expen-sive refrigeration technology. Thehighest Tc that had been observed inany superconductor was 23 kelvins forthe compound Nb3Ge, which has anHc2

of 0.4 megagauss.In 1986, a new class of supercon-

ductors, the “cuprates,” was discoveredthat were based on a layered structure



Figure 10. The MC-1 Flux Compression Generator Max Fowler peers into the central region of an MC-1. The white ring is a mock-up of

the high explosive that surrounds the central solenoid. Current is carried to and from

the solenoid along numerous cables, although only a single cable is shown.

of copper-oxide sheets separated bynon-superconducting layers. Thecuprates exhibit Tc’s extending farabove liquid nitrogen temperature (77 kelvins), a much easier temperatureto maintain. The present record Tc of135 kelvins is held by a mercury-cuprate compound. The potential forapplication of these high temperaturesuperconductors in motors, generators,and high-field solenoids that can oper-ate more economically at liquid nitro-gen temperature is enormous.

Naturally, there is a great interest inmeasuring the critical field for thesenew cuprate superconductors. Howev-er, for compounds with critical temper-atures above 90 kelvins, critical mag-netic fields have been observed toexceed 0.3 megagauss at temperaturesnear 70 kelvins. That magnetic field isapproximately at the limit of presentlyavailable direct current magnet technol-ogy. Since Hc2

only increases as thetemperature plunges towards absolutezero, a measurement of the critical fieldat lower temperature values has notbeen possible.

Thus, the value of Hc2(T) was more

than just idle curiosity. The model out-lined above for how a magnetic field

destroys superconductivity is quite gen-eral and has been experimentally veri-fied in detail for the conventional su-perconductors and, in many respects,for the new cuprates. However, there isyet no established theory for the micro-scopic mechanism of superconductivityin the cuprates, and there is growingevidence to support the idea that thereare fundamental differences with lowtemperature superconductivity. Recentexperiments indicate, for instance, thatCooper pairs in the cuprates may havenonzero orbital angular momentum, incontrast to the BCS model and to theestablished behavior of conventionalcompounds. This difference could af-fect the detailed functional form ofHc2

(T) at high fields. In addition, therehave been predictions of novel magnet-ic structures developing at high fieldsthat differ from the usual vortex latticestructure. It is clear that a determina-tion of Hc2

(T) over the range from Tc tolow temperatures and in fields of sever-al megagauss will be important in an-swering these questions.

The Los Alamos-Arzamas-16 collab-oration was interested in directly mea-suring Hc2

(T) for a YBCO (Yttrium-Bar-ium-Copper-Oxygen) high-temperature

superconductor as a function of tempera-ture, data that previously could not bemeasured because of the high criticalfield value. A sample of the YBCO ma-terial was placed along the axis of theMC-1 generator. A flow-through cryo-genic system maintained the sample at apredetermined temperature between 4and 80 kelvins. For a given fixed tem-perature, the state of the material wouldbe monitored while the magnetic fieldstrength was continuously measured as itincreased. At the critical field, the su-perconducting sample went normal.

The transition to a normal state washeralded by the appearance of a mil-limeter-wave signal at a receiver.When superconducting, the ceramicYBCO sample reflects electromagneticradiation at millimeter wavelengths, butthe radiation passes straight through thematerial when it is normal. As seen inFigure 11, the sample was sandwichedbetween two plastic dielectric wave-guides. The probe waveguide broughta 4-millimeter wavelength (75 GHz)signal to the 0.15-micron thick YBCOsample. When the sample went nor-mal, the radiation passed through thematerial, entered the detection wave-guide, and was detected by a receiver.

Magnetic field values were measuredwith both B-dot pickup coils and withoptical probes. An optical probe makesuse of the Faraday effect, in which theplane of polarization of polarized light isrotated as it passes through an optical el-ement situated in a magnetic field. Theamount of rotation is proportional to thefield strength. A polarized laser beamwas transported to and from a cylinderof flint glass (the optical element) byfiber optic cables, and a comparison ofthe plane of polarization between theoutgoing and the incoming laser beamsmeasured the magnetic field.

To complement the high field, lowtemperature measurements, two addi-tional experiments were performed athigher temperatures using low fieldgenerators built by Los Alamos.Figure 12 shows the four data pointsthat were generated. At the lowesttemperature, about 4 kelvins, the criti-



��

Styrofoam cascadeYBCO sample

Detection waveguide

Liquid helium or nitrogen

Styrofoam cascadeYBCO sample

Cryogenic chamberCryogenic chamber

Magentic field direction

Magentic field direction

Detection waveguide

Liquid helium or nitrogen

Microwavesignal

Microwavesignal

Fig. 83/18/96

Optical fiberOptical fiber

Figure 11. Schematic of the Critical-Field Experimental SetupA styrofoam cylinder, placed within the center of the MC-1, held the entire experiment.

A channel cut into the styrofoam formed a conduit for the delivery of cryogenic fluids

that cooled the sample. The temperature was adjusted by changing the particular

cryogen. The plastic waveguides directed a millimeter-wave signal to the YBCO sam-

ple for the detection of the superconducting to normal phase transition. The field was

measured by both optical (Faraday rotation) and inductive (B-dot) probes.

cal field was over three megagauss,more than six times the peak fieldachievable in prior laboratory experi-ments. The seven collaborative experi-ments mapped out the curve of the crit-ical field over the full temperaturerange. The data provides valuable in-formation for theorists and experimen-talists studying this material.

Fowler and Bruce Freeman of LosAlamos led the American team of morethan two dozen scientists in these chal-lenging experiments. This effort wasthe first time that Russians—let aloneRussians from a nuclear weapons insti-tute—had worked “behind the fence” atLos Alamos. Although most of thegenerators were Russian (Pavlovskii’sMC-1 generator), the high explosivesthat powered them were American, andLos Alamos explosives engineers hadto learn how to load the special “Russ-ian initiator blocks” that served to deto-nate uniformly the exterior of the mainexplosive charge.

Hot Magnetized Plasmas

The third series of experiments,which were initiated at Arzamas-16 inApril 1994, was the start of our collab-oration on the MAGO thermonuclearfusion scheme. This was the topic thatwas originally proposed by Chernyshevand Mokhov in September of 1991.The goal of this series was to investi-gate the first step of the MAGOscheme, that is, the production of a hot,magnetized plasma that could potential-ly be imploded to thermonuclear fusionignition conditions.

Fusion is the process by which twolight atomic nuclei combine to form aheavier nucleus. But fusion does notnormally occur under the conditionsfound here on Earth. All nuclei arepositively charged, and as the familiarmaxim states, like charges repel. Eachnucleus is surrounded by a Coulombbarrier that normally prevents the nucleifrom coming too close to each other.

But in the same way that a speedybullet can pass right through a thickwall, nuclei moving at extreme speedshave sufficient energy to penetratethrough the Coulomb barrier. A colli-sion between intensely energetic nucleiwill bring them so close that they feelthe strong attractive nuclear force. Thetwo nuclei will come together, fuse,and form a heavier composite nucleus.

As illustrated in Figure 13, a deu-terium (D) nucleus and a tritium nucle-us (T), two of the lightest nuclei avail-able, will fuse to form an isotope ofhelium (5He). That composite nucleusquickly decays into a neutron and analpha particle (a 4He nucleus). There isa large net energy release from the re-action, and both the alpha particle andthe neutron fly off with a considerableamount of kinetic energy.

Because energy is released, scientistshave long recognized the potential offusion to be the basis for a commercialenergy source. But realizing that po-tential has proven to be remarkably dif-ficult. For decades, scientists havebeen frustrated in their attempts to ad-vance beyond even the first critical step

in energy production, which is achiev-ing a self-sustaining, thermonuclear fu-sion reaction.

In thermonuclear fusion, the “fuel”for the reaction is a plasma (a state ofmatter consisting almost entirely of ionsand electrons) that is heated to millionsof degrees. That plasma temperature isa measure of the average kinetic energyof the ions and electrons. Because theparticle energies are distributed accord-ing to a Maxwell-Boltzman distribution,a tiny fraction of the ions have energiesthat are much higher than the averageenergy. For all present day thermonu-clear fusion schemes, the initial plasmatemperature is such that only those fewnuclei at the extreme high energy tail ofthe thermal distribution are sufficientlyenergetic to overcome the Coulomb bar-rier and fuse.

Energy is released by those early fu-sion events in the form of fast movingparticles. If those particles are captured



P N

N

NP

N

P N

N PN

PN

PN

Deuteron

Triton

Fusion reaction

Energetic neutron (14.1 MeV)

Energetic helium nucleus (3.5 MeV)

Fig. DT fusion 3/18/96

Figure 13. Thermonuclear FusionEinstein’s famous equation, E = mc2, re-

lates mass to energy. The sum of the

deuteron and triton rest masses at the

start of the fusion reaction is actually

more than the sum of the alpha particle

and neutron rest masses after the reaction

has finished. As a result of fusion, some

mass is converted into energy, and that

energy is imparted to the reaction prod-

ucts. Both the alpha particle and the neu-

tron emerge from the fusion event with a

significant amount of kinetic energy.

50Temperature (kelvin)

1

2

3

4

00 100

Crit

ical

fiel

d (m

egag

auss

)

SuperconductingphaseSuperconductingphase

Normalphase

MC-1

U.S.

Fig. 11/YBCO4/19/96

Figure 12. Critical Field of theYBCO SuperconductorThe critical field,

Hc2, for the YBCO su-

perconductor is plotted versus tempera-

ture. The critical temperature, Tc, for this

material is about 90 kelvins. The border

of the shaded region was drawn by hand

to help guide the eye and is not a fit to

the data. The temperature dependence

roughly follows that of metallic, low tem-

perature superconductors: Hc2(T) = Hc2

(0)

[1-(T/Tc)2)], where Hc2(0) is the critical

field at absolute zero.

and become part of the plasma, the ener-gy released by early fusion events willgo into increasing the plasma tempera-ture. The number of energetic nucleiwill increase, and the probability thattwo nuclei fuse will go up. The fusionreaction can become self-sustaining.

Unfortunately, there are always ener-gy losses that cool the plasma and killthe fusion reaction. Plasma particlesare in constant motion, and each timean electron scatters and gets acceleratedby an ion, energy is radiated away (ascontinuum radiation, also known asbremsstrahlung). The plasma cools.To maintain the temperature, enoughenergy must be pumped into the plas-ma, either by initial fusion events orexternally, to counteract those losses.

Because the energy gained by fusionand the energy lost throughbremsstrahlung both have a temperaturedependence, equating the two allowscalculation of an “ignition” tempera-ture, above which the plasma tempera-ture is maintained and the fusion reac-tion becomes self-sustaining. For theDT reaction, the ignition temperature isabout 4000 electron volts, or about 45million degrees (one electron volt cor-responds to about 11,600 kelvins).

Other loss mechanisms cool theplasma, but they are more amenable toexperimental control. One is the loss

of ions or electrons from the hot plas-ma. These carry energy away and theplasma cools. A second loss mecha-nism involves contaminants of “heavy”impurity ions, such as aluminum oriron, that increase the rate ofbremsstrahlung, and again the plasmacools. If enough impurities are present,one can never win in the energy bal-ance equation, and ignition can neverbe reached. Because impurities arenearly always present due to the out-gassing of walls and insulator materialsthat comprise the plasma chamber, min-imizing impurities has been a majorchallenge to all fusion schemes.

Even in this simplified picture ofthermonuclear fusion, it is clear thatconstructing a system that is designedfor getting useful power from fusion isa difficult undertaking. One wants asystem that sustains a high particle col-lision rate for a long a period of time.But in any real system, these are oftenconflicting demands. For any giventemperature, the collision rate can be in-creased by increasing the plasma densi-ty. But a high-temperature, high-densityplasma exerts an outward pressure, andthe higher the density, the more difficultit is to keep the plasma confined.

By making general assumptionsabout how much energy will be pro-duced by a plasma and how much ener-

gy will be lost by that plasma, one canarrive at minimum conditions forachieving useful power. The product ofthe density, n, and the plasma confine-ment time, τ, that is, nτ, is the relevantparameter, and the Lawson criterionstates that a minimum value for nτ beapproximately 1014 sec-cm-3. There islittle hope of achieving power from fu-sion unless the criterion is satisfied.

In the United States, fusion researchhas proceeded mostly along two paths.The first approach involves using atoroidal, or donut-shaped, reaction ves-sel, called a tokamak, to confine a lowdensity (n ~ 1014 cm-3) plasma. Highcurrents are sustained within the plasmathat heat it to ignition temperatures. Asshown in Figure 14, a charged particlewill spiral around a magnetic field line.Within the tokamak, magnetic fields arecreated that twist around the interior ofthe torus. The field lines form closedsurfaces, which the plasma particles areconstrained to follow. In principle, theplasma is confined forever. Dynamicalinstabilities actually limit the confine-ment time τ to 0.1 to 1 second, but thisis sufficiently long to balance the lowparticle density and bring nτ to withinthe range of the Lawson criterion. Gen-erally, the tokamak is considered to bethe most promising method for achiev-ing fusion, and worldwide, billions of



Force

B-field

Velocity

x x

x

x

xx

x x

x x xx x

x x

x x

x

x

xx

x x

x x

The magnetic force on a charged particle is directed at right angles to both the magnetic field and the particle's velocity.

A charged particle with a velocity that is entirely in the plane perpendicular to a uniform magnetic field moves in a circle.

A particle with a velocity component that is parallel to a uniform magnetic field spirals along the field line, which actsas a guiding center for the particle motion.

VF

FV

B-fieldB-field directed into page

Figure 14. Motion of a Charged Particle in a Magnetic FieldWhen a plasma consisting of bare atomic nuclei and electrons is subjected to a magnetic field, the individual particles will spiral

about the field lines. Stronger fields exert more force and the motion is a tighter spiral. Because the particles follow the field lines,

magnetic fields can be used to contain a plasma and increase the particle confinement time.

dollars have been invested in building,understanding, and developing theselarge and highly complex reactors.

The other mainline approach to ther-monuclear fusion, vigorously pursuedin the United States, is inertial confine-ment fusion (ICF). In an ICF scheme,a sphere of solid deuterium and tritiumis subjected on all sides to an implod-ing force that drives the DT fuel in-ward. The severe compression createsa hot, high-density plasma and resultsin fusion reactions. However, there isno way to confine the plasma once it iscreated, and the heat of the initial fu-sion events tend to expand the sphereand cool the plasma before ignitiontemperature is reached. It is only be-cause the implosion occurs so quickly(in billionths of a second) that the iner-tia of the inwardly moving fuel is ableto hold the sphere together and main-tain the temperature. The confinementtime, τ, is on the order of only 10-11 seconds, which is balanced by the

very high particle density (n ~ 1024 to1025 cm-3).

So far, the most successful implod-ing force has been created by usinglaser pulses generated by the hugeNOVA laser located at Lawrence Liver-more Laboratory, or by the OMEGAlaser located at the University ofRochester. However, an even morepowerful implosion is needed to bringthe plasma to ignition. It is hoped thatthe next-generation laser, to be built atthe National Ignition Facility, will pro-duce the required power.

An alternative approach to thermonu-clear fusion, one that used elements ofboth the tokamak and the ICF approach-es, was proposed by Andrei Sakharov(who incidentally helped elucidate theprinciples of the tokamak). He consid-ered creating a high-temperature, DTplasma in a strong magnetic field so thatthe charged ions and electrons were“stuck” to magnetic field lines, as in atokamak. The field would prevent ener-

getic electrons from leaving the plasmaand thus help reduce themal losses.

The hot, “magnetized” plasma wouldthen be imploded by an external forceas in an ICF scheme (Figure 15). Theimplosion would heat and compress therelatively dense plasma, and the strongfield would help capture the energeticalpha particles produced during the fu-sion events. The approach could poten-tially simplify the apparatus required tobring about ignition.

The Russian scientists call this fu-sion concept MAGnitnoye Obzhatiye,or magnetic compression (MAGO),whereas the U.S. researchers refer to itas Magnetized Target Fusion (MTF).To implement the scheme, VNIIEF in-vented a novel, two-section chamberthat produced a hot magnetized plasmaby means of hypersonic flow (Figure16). A gas mixture of DT is introducedinto both sections of the chamber. Twocurrent pulses sent through the chambercause a portion of the DT gas in onesection to become ionized and then pro-pelled through a nozzle so that it entersthe second section at a very high veloc-ity. The effect of the abrupt collisionbetween this plasma, moving at hyper-sonic speeds, and the relatively staticgas in front of it is to raise the tempera-ture of the gas rapidly to several thou-sand electron volts. This newlyformed, extremely hot plasma quicklyequilibrates to a temperature of severalhundred electron volts, at which point itis a large volume, relatively dense, hotplasma, referred to as the target plasmain Figure 16.

In a full MAGO fusion scheme, thetarget plasma would be surrounded by athin liner. Another current pulse, sentdown the walls of the liner, would cre-ate a magnetic field that implodes theliner. This action would compress theplasma and potentially bring it to igni-tion conditions. (Figure 16 shows thechamber that was used for the plasmaformation tests. In compression experi-ments, the chamber would be modifiedby replacing the thick, stationary outerwall with a thin liner.)

Producing the target plasma is the



Implode a linerthat surrounds

the plasma.

Fusion reactions take place in the hot, highly compressed plasma.

A high temperature plasma in the presence of a magnetic field. Charged nuclei spiral around the field lines.

1 2

Figure 15. The MAGO Two Step ProcessIn the first step of MAGO, a DT gas inside of a thin liner is heated and ionized to a

plasma in the presence of a magnetic field. The plasma particles are constrained to

follow the magnetic field lines. In the second step, the liner surrounding this “magne-

tized” plasma is imploded, and the plasma gets compressed. The higher particle den-

sity results in an increased collision rate, which leads to more fusion events. The

magnetic field reduces thermal energy losses and potentially helps capture the 3-MeV

alpha particles that are released from D-T fusion events. If other thermal losses can be

minimized, the plasma temperature may increase and reach ignition.

intriguing aspect of the MAGO scheme,and the Arzamas-16 scientists presentedsome neutron data as evidence that theplasma had been created. The initialplasma temperature of several thousandelectron volts is sufficient to initiate aburst of thermonuclear reactions, sothat even without further compression,a small fraction of the plasma producedon the order of 1013 neutrons. Al-though those neutrons were simply aby-product of the plasma formationmethod, ironically, this neutron produc-tion was comparable to the highest everachieved in the United States in pulsed-power or ICF experiments.

The objective of the first MAGO ex-periment, held in April of 1994, was toproduce and diagnose the hot, magne-tized plasma. The Chernyshev teamprovided a unique two-pulse helicalgenerator to power the plasma chamber,and Los Alamos brought to Arzamas-16more than a ton of advanced diagnos-tics equipment, which included spec-trometers, plasma interferometers, andprecision current probes. Excellentdata were obtained with the U.S. instru-ments, and the experiment greatly im-proved our understanding of plasmaflow through the nozzle as well as thefinal temperature and density distribu-tion of the hot, dense plasma.

Still, the effectiveness of a magneticfield in reducing electron losses couldnot be deduced from that initial experi-ment. Thus, four more experimentswere done by a team of Russian andAmerican scientists at Los Alamos inOctober 1994. VNIIEF sent two oftheir two-pulse helical generators andtwo test armatures to Los Alamos. Thefirst two experiments tested the perfor-mance of American explosives in dri-ving the armature of the complex Russ-ian generator. The third was a fullMAGO plasma formation shot usingthe same Russian generator, but puredeuterium was used in the chamber in-stead of a deuterium-tritium mix. Thepurpose of that shot was to confirm theelectrical performance of the deviceusing Los Alamos explosives and ourcapacitor bank. The experiment served

also to verify the operation of new di-agnostics that would be used on thefourth shot.

Fourteen VNIIEF scientists andmore than fifty Americans participatedin the final experiment. Chernyshevand Mokhov led the Russians, andReinovsky and Goforth were the LosAlamos shot coordinators. The experi-ment again used a Russian helical gen-erator along with as complete an arrayof diagnostics as Los Alamos couldprovide. Two major neutron diagnos-tics were fielded. One, based on mea-surements of the time of flight of theneutrons to the detectors, attempted toobtain an indication of the plasma tem-

perature. The second, based on neutronimaging, attempted to define the preciseregion from which the neutrons wereproduced. An array of optical and x-ray spectrometers were designed to pro-vide critical information on the time de-pendence of plasma temperature as wellas the presence of heavy ion impuritiesin the plasma.

The results of the experiment werevery encouraging. The data analysissuggested that a hot, dense plasma hadindeed been produced. Significantly,there were also indications that impuri-ties generated in the first plasma cham-ber were delayed by several microsec-onds before arriving in the second



��

��

Targetplasma

Nozzle Region

Coppercenter

electrode

Initial D-Tbreakdown

InsulatorsInsulators

Ionizing shock wave

Cou

ple

to G

ener

ator

Deuterium and tritium gas mixture

Current arc

Initial D-Tbreakdown

region

Section I Section II

Figure 16. MAGO Two-Section Chamber and Target Plasma FormationA cross section of the cylindrically symmetric, two-section MAGO chamber. The two

sections are joined by a narrow opening that acts as a nozzle. Initially, a DT gas fills

both sections. A current pulse of about 2 megamperes sent through the electrode cre-

ates a complex magnetic field pattern throughout the entire chamber. A second current

pulse, reaching 6 to 8 megamperes, arcs through both section I and the nozzle region

and creates a weak plasma. Due to the Lorentz force, this plasma is propelled through

the nozzle. When the high-velocity plasma collides with the relatively static gas filling

section II, shock waves are produced. These shock waves ionize the bulk of the gas

and create a large volume, relatively dense plasma at a temperature of 100 to 300 elec-

tron volts. Such a plasma could possibly be compressed to thermonuclear ignition con-

ditions in future experiments. (Figure courtesy of N. Shea, Defense Science)

chamber. This meant that the DT plas-ma in the second chamber would re-main relatively free of harmful impuri-ties and was likely to remainsufficiently hot for the 5 to 10 mi-croseconds required to compress it toignition conditions.

Another series of experiments atArzamas-16 are planned to testMAGO/MTF concept. Ultimately, oncea plasma has been judged suitable interms of temperature, density (n ~ 1018

cm-3), and purity, the experiments willattempt an implosion using the sametype of plasma formation chamber asbefore and a DEMG to provide theroughly 65 megajoules of energy esti-mated to bring about ignition. A jointexperiment at Arzamas-16, planned forthe summer of 1996, will be the firstdevelopmental test of the “high-energy”liner that will implode the hot plasma.

Isentropic Compression

The behavior of matter under ex-treme compression is of interest interms of understanding phenomena asdiverse as the atmospheres of gaseousplanets and the structural mechanics ofrock deep within the Earth. For exam-ple, the properties of materials underextreme pressures is important to geo-physicists studying the origin and dy-namics of earthquakes. Because manyearthquakes occur deep beneath the sur-face, knowing the shear strength ofrock at conditions found there could beimportant for developing predictivemodels of earthquakes.

One of the most successful tech-niques for compressing materials tohigh pressures is to use a diamond anvilpress, which can currently achieve pres-sures up to about 2 megabars. Abovethat, a standard technique is to use highexplosives to drive shock waves direct-ly through the material. Although ul-trahigh densities can be achieved viathis technique, the shock waves abrupt-ly jar the material and generate heat asthey propagate. Strong gradients andtransient effects often complicate the

interpretation of data obtained by thismethod.

An alternative technique for achiev-ing pressures above 2 megabars is to use magnetic pressure to implode aconducting surface that surrounds the sample of interest. The implosion can subject the sample to even higher pres-sures than are possible with shock wavemethods. Because a flux compressiongenerator produces a magnetic field that builds slowly and reaches its peak value after a few microseconds,the pressure increases in a relativelysmooth and steady fashion. Thus,shock wave production and sampleheating are minimized, and materialscan be compressed with a minimumchange of entropy (isentropic compres-sion). This simplifies not only the datainterpretation, it also opens up the pos-sibility of studying the low-temperaturebehavior of materials.

Our Russian colleagues at Arzamas-16 had employed isentropic compres-sion to study many different materialsat pressures of many megabars. Hydro-gen was of particular interest in theearly Russian work. At very high pres-sures, this gaseous element was predict-ed to undergo a transition to an atomic,metallic phase. It proved to be verydifficult to identify unambiguously theatomic phase, because under extremepressure, hydrogen can form many dif-ferent molecular phases that tend to ob-scure the interpretation of the data.

In 1994, we began discussions withthe Russians to perform an isentropiccompression experiment. Eventually, itwas decided that we would attempt tomeasure the electrical conductivity ofsolid argon as it was compressed undera peak pressure of over 6 megabars.

Argon solidifies at liquid nitrogentemperatures. Because it is a closed-shell atom, argon is insulating undernormal conditions, and even when so-lidified, the atoms of the crystal retaintheir monatomic character. Under ex-treme pressure, however, the atomic or-bitals of adjacent atoms are predicted tooverlap, which would allow electronsgreater mobility, effectively increasing

the electrical conductivity. The solidargon is predicted to undergo a transi-tion to a conducting state at about 5megabars. Any change in the electricalproperties of the sample could be attrib-uted to quasi-molecular or many-bodybehavior.

A preliminary attempt to measureelectrical conductivity of the samplefailed, however, due to the prematuredestruction of the current probes. Asecond experiment, conducted in Au-gust 1995, used a simpler current-probedesign and very clearly demonstrated aconducting state for argon at pressuresbetween 5 and 6 megabars.

This experiment was the firstdemonstration of the transition of argonfrom an insulator to a conductor at highpressure, and it held some surprises.The conductivity was remarkably low,indicating that rather than creating aconduction band of current carryingfree electrons, the electrons were tend-ing to “hop” from one atomic site toanother. This behavior was unexpect-ed, and thus the experiment has gener-ated some theoretical interest. Futureexperiments will attempt to achieveeven higher pressures, so that thecrossover to the metallic phase shouldbe more apparent.

Soft X Rays

Another topic of mutual interest toArzamas-16 and Los Alamos is the cre-ation of a soft x-ray source. Mostpulsed-power sources of x rays arebased on the fast implosion of a cylin-drical liner. As described earlier, avery light liner driven inward by mag-netic pressures can reach fantasticspeeds of hundreds of kilometers persecond. The interaction with the mag-netic field heats the imploding liner andturns it into a moving wall of plasma.When this cylindrical wall of plasmareaches the implosion axis, it collideswith itself, stops moving, and convertsits kinetic energy into internal heat en-ergy. That hot, stagnated plasma radi-ates x rays as it cools.



For the above concept to work, theliner must reach a very high velocity.Otherwise, the total energy in the sys-tem is below what is necessary to cre-ate an intense thermal x-ray source. Inaddition, the implosion must proceedwith a high degree of symmetry. Ifsome section of the liner is movingfaster than the rest, it will prematurelyarrive at the implosion axis. Stagnationwill occur somewhere off-axis, and thehot plasma will be distributed over abroad, indeterminate region.

Although many ideas have beentried, almost all of them have fallenshort of the two criteria mentionedabove. More often than not, the limit-ing factor is the growth of dynamicalinstabilities that cause the liner to breakapart prematurely, so that the implosionis severely asymmetric. But obtaininga very rapidly rising current pulse isalso problematic. The current sourcemust deliver all of its energy in thetenths of microseconds before therapidly moving plasma shell reaches the

implosion axis. Designing a fast switchrepresents a significant challenge forany pulsed-power system.

The Chernyshev-Mokhov team con-ceived a novel approach to solve theseproblems. Rather than accelerating alow-mass liner, a magnetic field im-plodes a large-radius (19 centimeters),“heavy” (0.5-millimeter thick) alu-minum liner. The acceleration occursduring the several tens of microsecondsthat the generator is powering up.When the generator has reached peakcurrent, the liner, now in a liquid state,is cut by a knife-like protrusion calleda “clipper.” In a manner similar torunning a wire through a film of soapywater, the break in the liquid linercauses a “bubble” to form between theclipper and the remaining liner, asshown in the Figure 17.

The bubble is really a section of theliner that is “thinned” to the point thatthe magnetic forces can ionize it andturn it into a plasma. The magneticfield that was driving the heavy liner

now rapidly accelerates this plasmabubble so that it converges upon theaxis of the device. The hot plasmastagnates and produces x rays.

The advantage of this scheme is thatwhile the generator is powering up, theheavy aluminum liner is moving rela-tively slowly, so the opportunity for thegrowth of instabilities is greatly re-duced. After the bubble is formed, itslow mass can be accelerated rapidly bythe peak field. There are no switchesinvolved. In addition, the surface densi-ty of the bubble is much lower than thatof the liner, which also helps in the sup-pression of hydrodynamic instabilities.

After a detailed analysis of the Russ-ian’s two-dimensional calculations, wedefined a set of Los Alamos diagnosticsthat would test the key elements of theconcept. A microwave interferometerwas designed to measure the initial mo-tion of the heavy liner. A set of fiber-optic and magnetic probes measured theprogress of the plasma bubble duringthe fast phase of the implosion. A



��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

Fig. Liner3/21/96

"Slow"-moving liner

Plasma arc forms

Rapidlyexpandingplasmabubble

Plasma

��

��

��

��

��

��

��

��

��

Liner

Magneticpressure

Clipper

Axis ofsymmetry

Current (65–70 MA)

(a)

(b) (c) (d)

Figure 17. Generation of a Plasma Bubble The figure shows one half of a cross section through a cylindrically symmetric cham-

ber. a) Current begins to run through the thick aluminum liner. The current generates a

magnetic field and the magnetic pressure accelerates the liner inward toward the clip-

per. b) At peak current, the liner moves past the clipper. The break in electrical conti-

nuity causes a “spark,” or an arc of plasma, to form between the liner and the wall. c)

The magnetic pressure expands the plasma arc into a “bubble.” Due to its low mass,

the bubble rapidly accelerates towards the implosion axis (the axis of symmetry). d)

The remainder of the slow-moving heavy liner stays behind while the bubble races in-

ward. Upon reaching the implosion axis, the plasma collides with plasma coming from

other sides of the chamber, stagnates, and emits x rays as it cools.

DEMG was used to provide the currentto drive the heavy liner. This ambi-tious experiment was conducted in Feb-ruary 1995 at the same firing pointwhere the previous DEMG and magne-tized plasma experiments had been conducted.

The results of the experiment weremixed. Los Alamos and VNIIEFanalyses suggest that a bubble was in-deed formed, although some significantasymmetries appear to have occurredduring its implosion. The implosionaxis was shifted approximately one cen-timeter off-center of the DEMG sym-metry axis, probably because of a sig-nificant azimuthal asymmetry in thedensity of the plasma bubble thatformed. The reason for the densityasymmetry is not clear. One possibleexplanation is that the heavy liner mayhave had a nonuniform electrical con-nection to the current source, resultingin nonuniform acceleration. In anycase, unless the unpredictable shift canbe controlled, the scheme in its presentconfiguration is unusable as an x-raysource because the x rays would begenerated from an unknown location.

This experiment highlights the diffi-cult nature of explosive-driven pulsed

power research. The results of monthsof effort culminated in one irrepro-ducible experiment that lasted but a fewmicroseconds. The outcome was notall that had been hoped for, althoughanalyses showed that the implodingplasma may well have had more implo-sion kinetic energy than presently avail-able in any other concept. Ways of im-proving the technique and removing theasymmetries may therefore be exploredin the future.

The Future

The unprecedented collaboration be-tween the nuclear weapons laboratoriesat Arzamas-16 and Los Alamos reflectsthe changes that have occurred in thepost-Cold War period. Scientists whowere previously intense competitors inthe design of weapons of mass destruc-tion are now working together to applytheir skills to problems of general sci-entific interest. In just over two years,Los Alamos and VNIIEF have per-formed experiments on ultrahigh cur-rent generation, the properties of high-temperature superconductors, theproperties of magnetized plasmas, the

compression of materials undermegabar pressures, and the creation ofa soft x-ray source. These experimentswere conducted at the very sites previ-ously used for weapons development.

Both sides are enthusiastic aboutcontinuing and expanding the collabora-tion. There is much to be learned aboutthe promising MAGO/MTF fusionscheme first suggested by AndreiSakharov. In forthcoming experiments,we hope to compress helium to thesame conditions found in the gas-giantplanets and thereby gain a better under-standing of these remarkable bodies. ALos Alamos proposal that involves fly-ing an explosive generator on a high-al-titude balloon to stimulate lightning ar-tificially has been accepted by theRussians. Several experiments to ex-plore quantum field effects at highmagnetic fields using the MC-1 genera-tor have already been performed at LosAlamos (see “The Dirac Series—ANew International Pulsed-Power Col-laboration” on page 68). A DEMG ex-periment to drive the most energeticsolid liner ever will be conducted thissummer. In short, there seems to be noend to the possibilities for collabora-tions on scientific endeavors.

■



Further Reading

F. Bitter. 1965. Ultrastrong magnetic fields.Scientific American. 213 (July: no.1): 65-73.

C. M. Fowler, W. B. Garn, and R. S. Caird.1960. Production of very high magnetic fields byimplosion. Journal of Applied Physics. 31(March: no. 3): 588-594.

H. Kolm, B. Lax, F. Bitter, and R. Mills, eds.1962. High Magnetic Fields, Proceedings of theInternational Conference on High MagneticFields, November 1-4, 1961. New York: TheM.I.T. Press and John Wiley and Sons, Inc.

H. Knoepfel and F. Herlach (L.G.I.) eds. 1966.Conference on Megagauss Magnetic Field Gener-ation by Explosives and Related Experiments:Proceedings of a Symposium sponsored by theItalian Physical Society and organized by TheLaboratorio Gas Ionizzati. Brussels: EuropeanAtomic Energy Community (Euraton).

H. Knoepfel. 1970. Pulsed High MagneticFields. American Elsevier Publishing Company,Inc.

I. R. Lindemuth, et. al. 1995. Target plasma for-mation for magnetic compression/magnetized tar-get fusion. Physical Review Letters. 75 (Septem-ber: no. 10): 1953-1956.

P. Sheehey. 1995. Magnetized target fusion.The World and I. May: 192-199.

S. M. Younger. 1993. AGEX II: the high-ener-gy-density regime of weapons physics. LosAlamos Science. 21: 63-69.

Carl Ekdahl earned his Ph.D. in Physics at theUniversity of California, San Diego, in 1971. Hefirst joined the Laboratory in 1975 to carry out ex-periments in controlled thermonuclear fusion, thenagain in 1982 tolead experiments toheat a high-densityplasma with elec-tron beams and tolaunch a high-power microwave-source developmentprogram. In 1983,he joined SandiaNational Laborato-ries to continuewith beam-propaga-tion experimentsand became Supervisor of the High-Energy BeamPhysics Division. He rejoined the Laboratory in1986 to design, execute, and analyze experimentsusing the radiation from underground nuclear-weapon tests. As leader of a nuclear-test diagnos-tics group, he directed their transition into above-ground experimental activities, including the firstlab-to-lab experiments with VNIIEF. He is cur-rently Program Manager for high-energy-densityphysics in the Nuclear Weapon Technology Directorate. Prior to joining the laboratory, Ekdahl held positions with Scripps Institute ofOceanography, the Laboratory of Plasma Physicsat Cornell University, and Mission ResearchCorporation.

C. M. (Max) Fowler joined the Laboratory permanently in 1957 with the responsibility of assembling a team to develop and apply explo-sive-driven magnetic-flux-compression devices.

The early work ofthis team influ-enced subsequentmegagauss solidstate research,liner implosion ofplasmas, and wasinstrumental instarting the“Megagauss” Con-ferences. Throughhis career, Maxand his colleagueshave used the ex-

plosive-driven magnetic-flux-compression tech-nique to generate energy sources to power a number of plasma-producing devices, lasers, imploding foils, electron-beam accelerators, andrailguns. This energy source was also used topower high-magnetic-field generators to studymaterials in megagauss fields, including high-temperature superconductors. Fowler received hisB.S. in chemical engineering from the Universityof Illinois and his Ph.D. in physics from the Uni-versity of Michigan. He was recently awarded anHonorary Doctorate from Novosibirsk State Uni-versity for his work in high-energy-densityphysics and for furthering scientific relations between the United Sates and Russia. Fowler isactive as a Los Alamos Laboratory Fellow and aFellow of the American Physical Society.

James H. Goforth received his M.S. in physicsfrom New Mexico State University in 1973.During a tour of duty at the Air Force WeaponsLaboratories in Kirkland, NM, he directed theoperation of a state of the art, 250-kilojoule capacitor bank for driving plasma z-pinch experi-ments. Also at the Air Force Weapons Laborato-ries, he participated in z-pinch and fuse openingswitch experiments powered by flux compressiongenerators at Los Alamos National Laboratory.He joined the Laboratory in August, 1976, ashead of the detonator exploratory developmentunit. In 1981, he joined the Shock-Wave PhysicsGroup, where he continues to do explosivepulsed-power research and development . Hismajor contributionis the developmentof the explosivelyformed fuse open-ing switch that isin current use asthe primary pulsecompression stageof the Procyon explosive pulsed-power system. Hehas also played asubstantial part inthe development of all explosive pulsed powersystems for the High Energy Density PhysicsProgram. Goforth is currently project leader forthe development of driver systems to be used forhigh-energy liner experiments.

Irvin R. Lindemuth received his B.S. in electri-cal engineering from Lehigh University in 1965,and his M.S. and Ph.D. in engineering-appliedscience in 1967 and 1971, respectively, from theUniversity of California, Davis/Livermore. Priorto joining the Laboratory in 1978, Lindemuth wasa technical staff member at the Lawrence Liver-more National Laboratory. His areas of researchinclude thermonuclear fusion, advanced numericalmethods for computer simulation of fusion plas-mas, and related pulsed power-technology. Hecurrently is Project Leader for the InternationalCollaboration in Pulsed Power Applications atLANL and has responsibility to provide technicalleadership for the pulsed-power/magnetized-target

fusion collabora-tion between LosAlamos and VNIIEF at Arza-mas-16. Linde-muth is creditedwith establishing aSister City rela-tionship betweenthe two nuclearcities and activelycontinues his par-ticipation and support of the

program. In 1992, he was the recipient of a Distinguished Performance Award for his workin the formative stages of the LANL/VNIIEFcollaboration.

Robert E. Reinovsky received his M.E. andPh.D. degrees in physics from Rensselaer Poly-technic Institute in 1971 and 1973, respectively.From 1974 through 1986, he worked at the AirForce WeaponsLaboratory (nowAir Force PhillipsLaboratory) in plas-ma and pulse-powerphysics. His princi-ple interests werehigh-density plasmaimplosions, radia-tion processes, plas-ma diagnostics, andpulse-powerphysics. Reinovskywas responsible fordeveloping and building four generations of theworld-class SHIVA family of high-current, low-impedance pulse-power systems. Techniques inultrahigh-current, high-explosive pulse power thatwere developed in Los Alamos in the1950scaught Reinovsky’s interest. He joined the Labo-ratory in 1986 to continue work applying thesetechniques to ultrahigh-current plasma systemsfor applications to high-energy-density physics.Reinovsky led the explosive pulse group from1990 to 1993 and later joined the program inhigh-energy-density physics as project leader forthe Athena pulse-power project. He is currentlyChief Scientist for that program.

Stephen M. Younger received his Ph.D. in theo-retical physics from the University of Marylandin 1978. Prior to employment with the Laborato-ry, Younger worked at the National Bureau ofStandards in Washington D. C., on related topics

in theoreticalatomic physicsand was a memberof the Nuclear De-sign Departmentat Lawrence Liv-ermore Laborato-ry. His researchat Livermore in-cluded advancednuclear weaponsdesigns and super-vising designgroups for the nu-

clear-driven x-ray laser and other nuclear-explo-sive concepts. He came to Los Alamos in 1989and has directed programs in inertial-confinementfusion and above-ground experiments. In 1994,he was named Deputy Program Director for Nu-clear Weapons Technology and was responsiblefor the physics associated with nuclear weapons.Younger currently is the Director of the Centerfor International Security Affairs at the Laborato-ry. His responsibilities include oversight for in-teractions involving Los Alamos and the Newly Independent States.





This April, scientists from seven laboratories under four flags gathered at LosAlamos to conduct a campaign of pioneering experiments using ultrahighmagnetic fields. This collaboration among Americans, Russians, Aus-

tralians, and Japanese is without precedent. This series of experiments was namedafter the great physicist P.A.M. Diracbecause his monumental contributions toquantum theory touch on all aspects ofthe physics and chemistry we intend toexplore. We are sure Dirac would haveappreciated the unification of world sci-entific efforts represented by this collab-oration, as the world appreciated the uni-fication he brought to science.

Some of the participants in this col-laboration are Florida State University,the University of New South Wales,Louisiana State University, the Univer-sity of Tokyo, the National Institute ofMaterials and Chemical Research(Tsukuba, Japan), Bechtel Nevada, andthe All-Russian Institute of Experimen-tal Physics (Arzamas-16). The LosAlamos contingent consists of programmanager Johndale Solem, shot coordina-

tor Jeff Goettee, and local staff members Max Fowler, Will Lewis, Dwight Rickel,Murry Sheppard, and Bill Zerwekh.

The Dirac series included four 1.5-megagauss experiments, using an explosive-driven generator designed at Los Alamos, and three 10-megagauss experiments,using the MC-1 explosive-driven generator designed at Arzamas-16. A brief outlineof the goals of each experiment is given.

The Quantum Hall Effect at High Electron Density. The Hall effect describesthe development of a transverse electric field in a current-carrying conductorplaced in a magnetic field, and it was discovered nearly a century ago by EdwinHall. The quantum Hall effect was discovered in 1980 by Klaus von Klitzingusing the two-dimensional electron gas formed in a metal-oxide, silicon, field-ef-fect transistor. At low temperatures, the degenerate electron ground state breaksup into energy levels called, “Landau levels.” As von Klitzing adjusted the gatevoltage to raise the Fermi energy level, he observed a quantized sequence ofplateaus in the Hall conductivity at integral multiples of e2/h, suggesting a funda-mental unit of electrical conductivity. These plateaus were accompanied by near-vanishing resistivity in the electric-field direction. Von Klitzing won the NobelPrize for his discovery of this “integer quantum Hall effect.”

But the story was far from over. Using much higher fields and lower tempera-tures, researchers in 1982 reported a fractional quantum Hall effect; plateaus oc-curred in fractions of e2/h. At first, only odd denominators were reported (1/3,2/5, 3/5, 2/3, and so forth). These were quickly attributed to the interaction be-tween electrons, that is, collective effects or quasiparticles. Sensible theories werepropounded as to why the denominators were all odd, but in 1993 many re-

The Dirac Series

The International Group of Scien-tists and Technicians that CarriedOut the Dirac Series. More than

seven universities and institutes, repre-

senting four countries, participated in the

experiments that were conducted in

Ancho Canyon at Los Alamos. The large

white tubing seen in front was a vacuum

line that was eventually connected to a

cryostat located inside an explosive-dri-

ven flux compression generator.



searchers reported even denominators. At present, many theorists believe the frac-tional quantum Hall states are actually integral quantum Hall states of compositeFermions. For example, the 2/5 state has 5 flux quanta for every 2 electrons (thatis, 2 filled Landau levels of composite electrons).

Although many experiments have been performed, precisionexperiments on the quantum Hall effect are often limited by im-perfections in the sample. Fortunately, samples with higher elec-tron densities are less sensitive to imperfections, and higher mag-netic fields allow observation of the quantum Hall effect insamples with large electron densities. Ultrahigh magnetic fieldsare required to observe the effect. The object of this experimentis to explore integer and fractional quantum Hall effects in a highelectron density, two-dimensional electron gas in a semiconductorheterostructure device. Clean data from this experiment will sup-ply a stronger experimental basis for building a complete under-standing of magneto-quantum electronic effects in solid statephysics.

Quantum Hall Effect and Quantum Limit Phenomena in Two-Dimensional Organic Metals. Two-dimensional metals may beseveral orders of magnitude more conducting in the x and y direc-tions than in the z direction. Their anisotropic conductivity sug-gests that these metals should behave somewhat like a compositeof two-dimensional electron gases. The integer quantum Hall ef-fect has been observed in preliminary laboratory experiments up to about 5 mega-gauss. At extremely high fields, the magnetic and Fermi energies are comparable,and we enter the realm called the quantum limit.

What happens to the two-dimensional metals in the quantum limit is simply un-known. If they retain their Fermi-liquid character, we expect something akin tothe fractional quantum Hall effect, although we may see entirely new collectiveelectronic configurations. On the other hand, the field may localize the conductionmechanisms and cause the material to behave more like a semiconductor or an in-sulator. The results will certainly lead to a deeper understanding of these very in-teresting materials as well as conduction mechanisms in general. Curiously, thesetwo-dimensional metals have many aspects in common with biological materials,so the implications may transcend the domain of solid state physics.

Magnetic-Field Induced Superconductivity. Superconductivity derives from anet attractive interaction between electrons in the neighborhood of the Fermi sur-face. In conventional superconductors the interaction is the sum of a repulsiondue to the Coulomb force and an attraction due to ionic overscreening.

As described in the main article, a magnetic field can break the superconduct-ing state, although how it does so depends on the type of superconductor. For-mally, there are two types of superconductors. Type I superconductors exhibitperfect diamagnetism: the magnetic field is abruptly expelled at the superconduct-ing transition, and once above a critical magnetic field, the entire specimen re-verts to the normal state. In a Type II superconductor, there is no flux penetra-tion below a first critical field, but there is partial flux penetration in the form ofevenly spaced thin filaments below a second critical field. In both Type I and

A New International Pulsed-Power Collaboration

Waiting for Dirac. Program manager

Johndale Solem and Max Fowler (fore-

ground) have done their jobs. On the day

of the shot, responsibility for the experi-

ment falls to the technicians and the shot

coordinator, and to the individual re-

searchers. In the background are Andy

Maverick from Louisiana State University

and Hiroyuki Yokoi from the National In-

stitute of Materials and Chemical Re-

search, Tsukuba, Japan.



Type II superconductors, the critical field is a function of temperature.Theoretical work at Los Alamos and elsewhere has suggested that in the quan-

tum limit (the lowest Landau level), the temperature for a transition to the super-conducting state can actually increase with field. The electron-electron repulsionis screened by the Debye length, and it can be shown that above some ultrahighmagnetic field values, the Debye length increases with field. The electron-electronrepulsion can be reduced until attraction dominates.

This new kind of superconductivity has never been observed, and in principle,it can be observed only at ultrahigh fields. Besides leading to a deeper under-standing of superconductivity, this research could result in a new kind of super-conductor that thrives, rather than quenches, in a magnetic field.

Zeeman-Driven Bond Breaking in Re2Cl--8 . Quadruplybonded metal complexes are a relatively new discoveryin physical chemistry. Four bonds are formed betweentwo metal atoms, and that two-atom core is free to inter-act with a variety of ligands. These complexes are ofconsiderable interest, and they enjoy symmetry propertiesthat make them simple to describe.

The lowest excited state of the rhenium-chloridecomplex consists of a singlet state (no spin) and tripletstate (one unit of spin). The singlet is readily accessibleby photoexcitation, and hence its energy level has beenmeasured and is well-known. Little is known about thetriplet other than it has an electron in an antibonding or-bital. Thus, two rhenium atoms can form only threebonds when excited to the triplet state.

In this experiment, a new type of chemical manipula-tion will be attempted. The Zeeman effect, which is ashift of the energy level of an atomic or molecular statedue to the presence of a magnetic field, will be used toreduce the energy level of one component of the tripletuntil it lies below the ground state. This level “crossing”will break the fourth bond, an event that will be visible inthe material’s spectroscopy. The experiment is intendedto give a measurement of the energy level of the tripletstate, which has been heretofore inaccessible. This tech-nique may usher in a new way of doing chemistry.

High-Field Exciton Spectrum of Mercury Iodide. Ex-citons are electron-hole pairs that act like loosely boundatoms within a solid host. Excitons in tetragonal crystals

of mercury iodide have been studied by absorption and photoluminescence at lowtemperatures. In a direct-gap semiconductor, the hole and electron combine fromthe lowest energy state with the same crystal momentum. Direct-gap semiconduc-tors produce light easily and are the basis of many of the light-emitting devices inuse today. In an indirect-gap semiconductor, the hole and electron combine fromthe lowest energy state with a different crystal momentum and, consequently, pro-duce light rather poorly.

Mercury iodide is somewhere in between. The crystal possesses a secondarylocal minimum in energy at different crystal momentum. A magnetic field breaksthe symmetry and makes it possible to see which emissions in the near-band excitonphotoemission spectrum are due to direct or indirect processes. Observing the spec-trum at very high fields will enhance our understanding of these solid state devices.

Preparing for the experiment.Mikhail Dolotenko of Arzamas-16 (kneel-

ing), oversees the installation of his sam-

ples and diagnostics into the bore of the

MC-1 flux compression generator (orient-

ed vertically for this experiment). Lying

down are Los Alamos technicians Tommy

Herrera (facing) and Dave Torres.



Ultrahigh Magnetic-Field Calibration Standard. In some materials, a magneticfield along the direction of propagation will cause two circularly polarized compo-nents of an electromagnetic wave to propagate at different velocities. Thus, a lin-early polarized wave will rotate as it travels through the material. This is calledthe “Faraday effect.” The strength of the Faraday effect in a material is usuallycharacterized by the “Verdet coefficient,” which measures the rotation per unitfield per unit length.

Materials were fabricated with either samarium or europium embedded in aplastic matrix. These rare-earth elements have ground states and excited statesthat are split by the spin-orbit interaction into numerous levels. Due to the Zee-man effect, an applied magnetic field will cause some excited states levels andground state levels to interact and cross.After each crossing, the Verdet coefficientchanges, and steps appear in a plot of theFaraday effect versus magnetic field.These steps are a function of only the in-teratomic state and are not influenced bythe surrounding matrix. The specific mag-netic-field value at which each crossing oc-curs can be calculated using well-definedatomic constants, and thus observation ofthe crossing can be used to calibrate theexternal field. In the sample with the eu-ropium impurity, the first crossing shouldbe observed around 10 megagauss, the sec-ond around 12 megagauss, with periodiccrossings up to 50 megagauss. In the sam-ple with samarium impurity, the first cross-ing may be observed about 3 to 5 mega-gauss, with periodic crossings also up to 50megagauss. These samples may prove tobe the only probes capable of measuring magnetic fields up to 50 megagauss.

Faraday Rotation in Cd1-xMnxTe. Cd1-xMnxTe is a member of a group of mate-rials, called “diluted magnetic semiconductors,” that contain magnetic ions (Mn++

in this case) that can undergo a spin-exchange interaction with band electrons.This spin-exchange produces an enormous spin splitting of the energy bands and,consequently, a giant Faraday effect. At low magnetic fields and room tempera-ture, the Verdet coefficient is directly proportional to the field. At high fields,however, the Verdet coefficient is expected to reach a saturation level and evendecrease slightly. At low temperature and high field, steps appear in the Verdetcoefficient that are attributed to the coupling of pairs of the magnetic ions andmore complex (3, 4, 5, and so forth) clusters of magnetic ions. In the linearregime, Cd1-xMnxTe is of great practical importance as an optical sensor of mag-netic fields. Extension of the data for this material to ultrahigh fields will lead toa more complete understanding of the effect of magnetic clusters in diluted mag-netic semiconductors.

Conclusion. The Dirac series of experiments will explore fundamental physics inthe ultrahigh magnetic field regime of several different disciplines. These are ex-tremely difficult experiments, and new measurement techniques are already beingdeveloped in the course of designing and performing these investigations. This in-ternational effort is a fitting extension to the Russian-American pulsed-power col-laboration initiated under the lab-to-lab program. ■

An International Exchange. Members of the Russian delegation (from

left to right: Elena Gerdova, Vadim

Platonov, and chief scientist Olga Tat-

senko) discuss physics with Noboru

Miura from the University of Tokyo.

los alamos sciencenumber 24 1996 - jean-pierre petit · 50 los alamos sciencenumber 24 1996 a time...

Documents