PlutoniumA Wartime Nightmarebut a Metallurgist’s Dream

by Richard D. Baker, Siegfried S. Hecker, and Delbert R. Harbur

In 1942 the theoretical outline for anatomic bomb was clear: compressenough fissionable material longenough to properly ignite a chain

reaction. Construction of an actual weapon,however, required translation of “fissionablematerial” into real pieces of plutonium oruranium metal. These metals had to be freeof impurities that would adversely affect theneutron flux during the chain reaction andyet be fabricable enough that precise shapescould be formed. Whether this would even bepossible with plutonium was not then known,however, because plutonium was a new,manmade element and the metal had notbeen produced.

Accounts of the Manhattan Project haveneglected (for security reasons, initially) theimportant metallurgical work that precededfabrication of these materials into integralparts of real weapons. For example, theSmyth Report* devotes one short paragraphto the wartime work of the entire Chemistryand Metallurgy Division at Los Alamos—adivision that in 1945 numbered 400 scientistsand technicians. Our article will attempt tofill part of this gap for plutonium by high-lighting key developments of the wartimeresearch and will continue with some of theexciting research that has occurred since thewar.

Research from 1943 to 1946

The Los Alamos work on plutonium andenriched uranium, the so-called special nu-clear materials, was extensive, covering avariety of research problems ranging frompurification of material received from reac-tors to the prevention of oxidation of thefinal product. Further, because many chemi-cal processes and physics experiments re-quired very pure materials, such as gold,beryllium oxide, graphite, and many plastics,considerable general materials research wasalso carried out.

Much of the chemistry and metallurgy ofuranium was already known from the pro-duction of uranium metal for the uranium-graphite reactor pile at Chicago in 1942. Thework remaining on enriched uranium in-cluded preparation of high-purity metal, fab-rication of components, and recycling ofresidues. However, the most challenging re-search and development was carried out onthe new element plutonium.

Table I gives the important dates in theearly history of plutonium and shows theshort time—four years—that elapsed be-tween its discovery and its use in the firstatomic device at Trinity. The discovery oc-curred, as predicted by nuclear theory, whenuranium was bombarded with 16-million-

electron-volt deuterons in the cyclotron atBerkeley. Within about a month it wasshown that plutonium-239 fissioned whenbombarded with slow neutrons, and a de-cision was made to build large reactors atHanford for the production of pluto-nium—this before the uranium-graphite pileat Chicago had demonstrated that a sus-tained and controlled chain reaction waseven possible! That demonstration soon fol-lowed, proving that large quantities of pluto-nium could be produced, although no pluto-nium was extracted from the Chicago reac-tor.

At this point only microgram amounts ofplutonium had been separated from thetargets used in the cyclotrons. Remarkably,the basic chemistry of plutonium wasworked out at Berkeley and Chicago on thismicrogram scale, and it formed the basis forthe scale-up—by a factor of a billion—needed for plants that would eventuallyseparate plutonium from spent reactor fuel.At the same time the first micrograms of themetal were produced at Chicago by the

*Henry DeWolf Smyth, Atomic Energy for Mili-tary Purposes: The Official Report on the Devel-opment of the Atomic Bomb under the Auspicesof the United States Government, 1940-1945(Princeton University Press, Princeton, 1945),pp. 221-222.

142 Winter/Spring 1983 LOS ALAMOS SCIENCE

reduction of fluorides, and preliminarymetallurgical properties were determined.However, the influence of impurities on suchtiny samples distorted many of the results;for example, the melting point of plutoniumwas first thought to be about 1800 degreesCelsius, considerably above the true meltingpoint of 641 degrees. Ultimately, theproperties of plutonium were found to beincredibly sensitive to impurities.

It had been agreed that Los Alamos wouldnot work on batches of plutonium of lessthan about 1 gram, and the microgram-scalework continued at Chicago. Finally, in early1944 Los Alamos received plutonium nitratesamples containing half-gram amounts of theelement from the “Clinton” reactor and pilotextraction plant at Oak Ridge. Later, largeramounts were received from the productionfacility at Hanford.

The plutonium nitrate arrived in relativelyimpure form, and techniques and equipmenthad to be developed for a number ofprocesses, including purification, preparationof plutonium tetrafluoride and other com-pounds, reduction to metal, and metal fabri-cation. Also, because plutonium was in veryshort supply, it was imperative to developprocesses to recycle all residues.

Initially, the purity requirements for themetal were very stringent because someelements, if present, would emit neutronsupon absorbing alpha particles from theradioactive plutonium, These extra neutronswere undesirable in the gun-type plutoniumweapon then envisioned: they would initiatea chain reaction before the material hadproperly assembled into its supercritical con-figuration, and this “pre-initiation” woulddecrease the explosive force of the weapon.The purity requirement for certain elementswas a few parts per million and for some,less than one part per million. As a result, allthe materials used in the preparation of theplutonium metal, everything from the proc-ess chemicals to the containers, had to be ofvery high purity. This necessitated develop-ment work on many materials, including an

LOS ALAMOS SCIENCE Winter/Spring 1983

TABLE IEARLY HISTORY OF PLUTONIUM

Plutonium discovered February 23,1941Neutron-induced fission of plutonium-239, proved March 25, 1941Decision reached for large, full-scale plutonium production December 6, 1941First controlled fission chain reaction achieved, December 2, 1942

proving method for full-scale production of plutoniumPreparation of plutonium metal from microgram November, 1943

quantities produced with cyclotronGram quantities of plutonium nitrate from March, 1944

experimental reactor received at Los AlamosPlutonium nitrate from production reactor received mid 1944

at Los AlamosPlutonium weapon demonstrated with Trinity test July 15, 1945

extensive effort to obtain pure and nonreac-tive refractories to contain molten phuto-nium. The high purity requirements alsonecessitated the development of new meth-ods for analysis of all materials, includingplutonium.

The potential health problem associatedwith the handling of plutonium had beenrecognized at Chicago, and work on thesubject began with receipt of the first smallamounts of plutonium. A Health Group wasformed to monitor plutonium work areas,and, within the Chemistry and Metallurgy

Division itself, committees were establishedto design suitable radiation detectors andapparatus for handling plutonium and toformulate safe handling procedures. Becausealpha counters then lacked either sensitivityor portability and were in short supply, oiledfilter paper was swiped over surfaces to pick

up possible stray bits of plutonium and thenmeasured at stationary counters. Similarprocedures were used to detect suspectedcontamination of hands and nostrils. The air-conditioning system in the plutonium labora-tory (D Building), which was installed in-itially to help maintain high purity by filter-ing out dust, ultimately served the moreimportant function of confining the pluto-nium. The building was equipped with hoodswith minimal ventilation and with the fore-runner of the modern glove box—plywood“dry boxes.” The successful handling oflarge quantities of plutonium without seriousproblems was at that time an outstandingachievement.

Two early discs of plutonium metal afterreduction from the tetrafluoride. Pluto-nium generally arrived at the Labora-tory from the Hanford reactors in theform of a relatively impure nitrate solu-tion. Techniques were developed at LosAlamos for purification, preparation ofvarious compounds, reduction to themetal, and metal fabrication.

143

At first plutonium metal was prepared atLos Alamos either by lithium reduction ofplutonium tetrafluoride in a centrifuge, themetal settling out as the closed reactionvessel rotated, or by the electrolysis of fusedsalts containing plutonium. Soon, however,calcium reduction of the tetrafluoride wasperfected. The vessel used to contain thisreaction was called a stationary “bomb”because the reaction was highly exothermicand the metal product settled out in theclosed, nonrotating vessel simply by gravity.This technique became the preferred methodand was used to prepare plutonium foralmost all metallurgical studies and for thenuclear devices.

The microgram metallurgy at Chicagohad provided values for the density of themetal that clustered about either 16 or 20grams per cubic centimeter. This bimodalspread, due surely in part to impurities,nevertheless pointed toward interestingmetallurgy by hinting that the element hadmore than one phase. Working with largeramounts, Los Alamos refined these measure-ments and by the middle of 1944 haddiscovered that plutonium was a nightmare:no less than five allotropic phases existedbetween room temperature and the meltingpoint. Unfortunately, the room-temperaturealpha (a) phase was brittle, and the metalexperienced a large volume change whenheated and then cracked upon cooling. Theseproperties made fabrication very difficult,and there was not enough time for detailedfabrication development on the a-phase ma-terial. It was thought likely that anotherphase would be malleable and easily shaped;the problem was how to stabilize such aphase at room temperature. It was thendiscovered that alloying plutonium withsmall amounts of aluminum stabilized thedelta (8) phase, which was, in fact, malleable.However, aluminum was one of those ele-ments that emitted neutrons upon absorbingalpha particles and so would exacerbate thepre-initiation problem. Beneath aluminum inthe periodic table was gallium, which did not

144Winter/Spring 1983 LOS ALAMOS SCIENCE

Plutonium

The hydrofluorination of plutonium. The upper photograph shows the chemical hoodsin D Building used for this process, which converted the oxide to the tetrafluoride.Four furnace controllers are at the top of the panel with one controller open showingthe temperature program cut into its rotating disc. Note the bucket of calcium oxide tobe used for treatment of hydrogen fluoride burns. The lower photograph shows one ofthe hydrofluorination furnaces inside the hood.

LOS ALAMOS SCIENCE Winter/Spring 1983

undergo this type of nuclear reaction. Pluto-nium-gallium alloys were found to be stable

the required hemispheres. Thus the problemof fabrication was solved. To avoid oxidationof the metal and to contain the radioactivity,the pieces were ultimately coated with nickel.

In July 1944 it was discovered that theplutonium-239 generated in the high-neu-tron-flux production reactors at Hanfordcontained too much plutonium-240. Pluto-nium-240 was undesirable because it had amuch higher spontaneous fission rate thanplutonium-239 and emitted far too manyunwanted neutrons. As a result, pre-initiationin the gun weapon could not be avoidedwithout the difficult task of separating theseisotopes. Instead an intense effort wasmounted to develop an implosion weapon inwhich pre-initiation could be avoided be-cause of its higher assembly velocities. Thisturn of events allowed the purity require-ments for the metal to be somewhat relaxed,simplifying many of the process operations.The necessary pieces of plutonium were thenfabricated in time to construct the Trinityand Nagasaki devices.

The extreme press of time during the warallowed for the immediate problems of fabri-cation, stability, and oxidation protection tobe solved only empirically. A comprehensiveprogram of basic research on this mostfascinating element had to wait until after thewar. In Table II we summarize the propertiesof plutonium metal known in 1945.

Postwar Research and Development

As the war ended, construction began atDP West site on a new, more permanentfacility for the plutonium effort. This activityreflected the government’s decision to in-crease production of nuclear warheads and,thus. to scale up all processes associatedwith the fabrication of plutonium metalparts.

Because the plywood dry boxes of old DBuilding posed a tire hazard, they were

145

TABLE II

PROPERTIES OF PLUTONIUM METAL KNOWN IN 1945’

Phase Temperature Range Crystal Structure Density Average Electrical Temperature Coefficientof Stability (’W) (g/cm’) Linear Expansion Resistivity of Resistivity

Coefficient (@cm) (@cm per “C)(per ‘C)

Alpha Below 117 Orthorhombic (?) 19.8 55 x 10-6 150 at 25°C -29.7 X 1 0-4

Beta 117 to 200 Unknown (complex) 17.8 35 x 10-6 110 at 200°C - oGamma 200 to 300 Unknown (complex) Unknown 36 X 1 0-6 110 at 300°C ˜ 0

Delta 300 to 475 Face-centeredcubic 16.0 -21 x 10-6 102 at 400°C +1.5 x 10-4

Epsilon 415 to 637 Body-centeredcubic 16.4 4 x 10-6 120 at 500°C - 0

Liquid Above 637±5

aFrom Cyril Stanley Smith, Journal of Nuclear Materials 100,3-10 (1981).

replaced with stainless-steel glove boxes. Tobetter contain the plutonium, the glove boxeswere equipped with elaborate ventilation-filtration systems devised to keep the at-mosphere within each glove box at a lowerpressure than the surrounding air so that anyleak in the system would not release pluto-nium to the room. In addition, the breathingair in the laboratories was filtered andchanged several times each hour.

Since all of the processes for purification,preparation, and fabrication of the metal andfor recycling of the residues of plutoniumand enriched uranium were developed at LosAlamos during the war, there was no otherplace for the production of nuclear war-heads. It was decided that Los Alamosshould not continue in production but shouldconcentrate on research and development.The transfer of all the special processes to beused in the new production plants was amajor postwar undertaking. Plutoniumprocesses were transferred to Hanford,Savannah River, and Rocky Flats. The en-riched uranium processes were transferred toOak Ridge.

The work at DP West thus settled into aprogram of basic research and development,and major advances were made in the fledg-ling plutonium technologies of vacuum cast-ing, metal working, machining, electrorefin-ing, and aqueous processing of scrap. Sev-eral plutonium reactor fuels, both metallicand ceramic, and the plutonium-238 heatsources for thermoelectric generators forspace and other missions had their begin-nings at DP West.

In 1978 the plutonium activities at DPWest were moved to the newly completedLos Alamos Plutonium Facility, the mostmodern and complete plutonium research

146

Plutonium

and development center in the country. Itincorporates state-of-the-art designs andequipment for the safe containment of pluto-nium and the protection of workers duringall credible accidents or natural disasters,including earthquakes and tornado-forcewinds.

After the war the continued improvementin process chemistry and applied metallurgyof plutonium came about through a betterunderstanding of its basic properties.Aqueous processes were developed for

separating plutonium from virtually everyelement in the periodic table. The wartime“bomb” reduction process was augmentedwith other pyrochemical processes such asdirect reduction of the oxide and electrorefin-ing. These processes not only yielded a purerproduct but also minimized the amount ofplutonium-bearing residues and the as-

sociated radiation exposure of personnel.Plutonium casting was first carried out

with ceramic crucibles and molds becausethey were known to be compatible withmolten plutonium. The discovery thatslightly oxidized tantalum was quite unreac-tive with molten plutonium led to the devel-opment of reusable foundry hardware. Also,the development of several ceramic coatingprocesses, based on either calcium fluorideor the stable oxides of zirconium or yttrium,permitted the use of easily machined graphitemolds. It was discovered that microcracksresulting from the multiple phase changesthat occur as the metal cools and freezescould almost be eliminated by casting thepure metal into chilled aluminum molds, aprocess that virtually by-passes most of theintermediate phase transformations.

The development of new plutonium alloysfor both reactor and weapons use proceededhand in hand with the determination of theequilibrium phase diagrams of plutoniumwith most other elements, and the associatedcomplex crystal structures of phases andcompounds. Early on we realized that wewere dealing with alloys that were metastablein many environments. Thus, it became

147

Fig. 1. The actinides and the configuration of their outermost in the electrons near the Fermi level. Thorium is thereforeelectrons. This series of elements is characterized by the filling regarded as the first actinide.) The properties of the 5 fof 5 f electron orbitals. (The 5 f orbitals of thorium lie above the electrons, particularly their participation in atomic bonding,Fermi level but are sufficiently close to induce some f character are the key to the unusual properties of these elements.

imperative to understand microsegregationof alloying elements and phase stability dur-ing all processing steps. Complex heat treat-ments were developed to homogenize thealloys or to further stabilize the properphases.

In 1954 the purity of plutonium wasincreased sufficiently that a new phase, delta

this new phase proved to be inconsequentialto applied plutonium technology, its dis-covery certainly showed the necessity ofusing high-purity, well-characterized pluto-nium in basic research. A seventh allotrope

covered many years later in 1970 duringcareful studies of the equilibrium pressure-temperature phase diagram of plutonium.This phase exists only at high temperatureover a limited pressure range and has such acomplex crystal structure that it still today

148

has not been positively identified.

Current Understanding of Plutonium

During the postwar period major activitiesin plutonium metallurgy in the United Stateswere centered at Argonne National Labora-tory, Hanford, Lawrence Livermore Labora-tory, Rocky Flats, and of course, Los Ala-mos. Important contributions were alsomade at the Atomic Weapons ResearchEstablishment’s facility at Aldermaston inthe United Kingdom and at Centre d’EtudesNucleaires de Fontenay-aux-Roses inFrance. Except at Los Alamos much of theresearch activity was terminated or severelycurtailed in the 1970s. However, we arecurrently seeing a revival of plutonium re-search at several locations.

In spite of many years of concentratedresearch and great strides in the practicalaspects of plutonium metallurgy, this field is

still in its infancy. A comparison with steelsupports this perspective. The metallurgy ofsteel has been studied intensely in manycountries for more than 100 years, yetimportant discoveries are still commonplace.Metallurgists have learned to manipulate thethree allotropic phases of iron to tailor theproperties of steel to specific applications.The six allotropic phases of plutonium andits much wider range of crystal structures

and atomic volumes provide many morepossibilities—and pitfalls.

The focus of the postwar research was tostudy all aspects of the behavior of this newelement and, thus, be prepared for all of itspeculiarities. In contrast, the focus of thepast decade has been to exploit the complex-ities of plutonium. Much of the effort hasbeen devoted to alloy development and thedetermination of structural properties andhas resulted in several new alloys with inter-esting properties. Many of these results re-

Winter/Spring 1983 LOS ALAMOS SCIENCE

Plutonium

Fig. 2. A connected phase diagram of binary alloys of the actinides (prepared by E. A.

Kmetko and J.L. Smith) shows the preponderance of low-symmetry crystal structures,

the large number of phase changes, and the depression @melting points in the vicinity

of plutonium. The crystal structures are body-centered cubic (bcc),face-centered cubic

(fcc), tetragonal (tetr), orthorhombic (orth), exotic cubic (exe), monoclinic (mono), and

double hexagonal close-packed (dhcp).

main classified.In the past decade we have also turned our

attention toward a more fundamental under-standing of plutonium on the atomic level.This effort has opened a most fascinatingchapter in solid-state physics—the electronicproperties of the actinides, the seventh periodin the periodic table. Interest in the actinideshad stemmed primarily from their specialnuclear properties. Yet, it is the properties ofthe electrons (not the nuclei) that govern allchemical and structural behavior. The ac-tinides are characterized, as shown in Fig. 1,by the progressive filling of 5f electronorbitals. It is the participation of these 5 f

electrons in atomic bonding that leads to thepeculiar and complex behavior of actinidemetals and alloys.

Although details of the 5f bonding in theactinides are still being contested, it is gener-

LOS ALAMOS SCIENCE Winter/Spring 1983

ally agreed that the 5f electrons in the earlyactinides (through plutonium) are not fullylocalized and thus participate in bonding.The 5f bonding increases to a maximum atplutonium and vanishes as the electronsbecome localized near americium. The ef-fects of 5f bonding on the behavior of thelighter actinides are dramatic, and mostdramatic for plutonium. We will highlighthere only three of the most important effects.These are demonstrated in Fig. 2, a con-nected binary alloy phase diagram of theelements in the actinide series. First, as onemoves from actinium to plutonium, a changefrom highly symmetric cubic to low-sym-metry crystal structures occurs. Second, thenumber of allotropic phases increases. Fi-nally, the melting points decrease dramati-cally.

How can these effects be explained? To

begin, the wave functions off electrons (likethose of p electrons and unlike those ofs andd electrons) have odd symmetry. This prop-erty is not compatible with symmetric cubiccrystal structures but rather favors the low-symmetry crystal structures and, thus, thestability of monoclinic and orthorhombicphases. Only with increased temperature andlattice vibrations is the f character suffi-ciently overcome in plutonium to permitcubic crystal structures. Beyond plutoniumthe localization (nonbonding) of the f elec-trons leads to a return of more typicalmetallic behavior. Also, because the f elec-trons are just on the verge of becominglocalized and magnetic, small changes intemperature, pressure, or alloying have dra-matic effects on phase stability andproperties. Hence, allotropy is promoted.Finally, the f electrons bond quite easily inthe liquid phase because its less rigid struc-ture increases rotational freedom. This easeof bonding promotes the stability of theliquid (or, equivalently, limits the stabilityrange of the solid) and lowers the meltingpoints.

We see that the f electrons are the cause ofmany of plutonium’s peculiarities andcomplexities, which have important practicalconsequences. Its low melting point andlimited solid stability are particularly impor-tant because, as a liquid, plutonium is ex-tremely reactive and corrosive and hencedifficult to contain. Liquid plutonium alsohas the greatest known surface tension andviscosity among metals because off bonding.A less obvious consequence arises from thefact that most rate processes in solids dependupon homologous temperature, that is, tem-perature relative to the absolute meltingpoint. Hence, diffusion and other thermally

activated processes are quite rapid at roomand slightly elevated temperatures.

The most significant consequence of plu-tonium’s large number of phases is thermalinstability of the solid. This property is bestillustrated by a plot of length change duringheating. Figure 3 compares the behavior of

149

plutonium with that of iron. Most phasetransitions in plutonium are accompanied bylarge length and thus volume changes. Suchvolume changes are difficult to accommo-

date in solids at relatively low temperatureswithout loss of physical integrity. In addi-

very large thermal expansion coefficients.For example, the thermal expansion coeffi-cient of the a phase is about five timesgreater than that of iron. Therefore specialcompatibility problems arise wherever pluto-nium is in contact with other metals. Fig-ure 3 illustrates two exceptionally peculiarproperties of plutonium: the negative thermal

and the contraction upon melting, whichresults from increased f- electron bonding inthe liquid phase.

The crystal structures and the correspond-ing densities are also listed in Fig. 3. Notethat the three structures that are stable at I I

Plutonium

Iron

/temperatures closest to room temperatureare of low symmetry. The cubic structuresthat are typical of most metals appear onlyat high temperatures where the 5 f -electronbonding is overwhelmed. The low-symmetrystructures (especially the a phase) exhibitvery directional bonding. The a-phase mono-clinic structure is essentially covalentlybonded. Its unit cell contains 16 atoms with8 different bond lengths ranging from 2.57 to3.71 angstroms. Consequently, most of itsphysical properties are also very directional.In addition, the a phase is a poor conductorand is highly compressible.

The low symmetry and nearly covalentnature of bonding in the a phase greatlyaffect its mechanical properties, which morenearly resemble those of covalently bondedminerals than those of metals. The a phase isstrong and brittle because the low symmetrycontrols the nature and motion of defects.

hand. behaves much like a normal metal. InFig. 3, A plot of percentage length change as a function of temperature illustrates the

malleability of aluminum. One must remem- dramatic changes that occur with each of plutonium’s phase changes. The more sedatebehavior of iron is shown for comparison.

150 Winter/Spring 1983 LOS ALAMOS SCIENCE

Fig. 4. The plutonium-gallium phase diagram serves as an example of the complexitiesthat occur when plutonium is alloyed. The region of concern during the war was the

gallium concentrations below 10 per cent.

the 5f bonding is essentially gone.

temperature by alloying. As we pointed outearlier, this fact was already recognizedduring the war and led to the use of galliumto stabilize this phase. It is now recognizedthat most trivalent solutes, such as gallium,aluminum, cerium, americium, iridium, and

shows the plutonium-gallium equilibriumphase diagram as determined at Los Alamosin the postwar era. Note the expanded field

complexities that result from alloying pluto-

behaves much like a normal metal and hasseveral advantages over the a phase, includ-ing excellent ductility (fabricability), a much

Further Reading

O. J. Wick, Ed., Plutonium Handbook: A Guide to Technology (Gordon and Breach Science Publishers,New York, 1967),

A. S. Coffinberry and W. N. Miner, Eds., The Metal Plutonium (University of Chicago Press, Chicago,1961).

G. T. Seaborg, The Transuranium Elements (Addison-Wesley Publishing Co., Reading, Massachusetts,1958).

larger range of thermal stability, and a lowerthermal expansion coefficient (nearly zerofor most alloys).

So far we have not mentioned the effectsof pressure. As one might expect, hydrostaticpressure tends to collapse the low-densitycrystal phases. Hence, in pure plutonium the

kilobar. Here is where the seventh allotrope

pressures. Only moderate pressures are re-

higher density phases. When dealing withalloys at high pressures, we are faced withthe problem of what happens to the soluteatoms, since they are generally insoluble in

question of the response of alloys undernonequilibrium cooling conditions typify thefascinating world of nonequilibrium phasetransformations in plutonium, which isbeyond the scope of this article.

Plutonium is without question the mostcomplex and interesting of all metals. Moreso than in any other metal, a fundamentalunderstanding of its metallurgical behaviormust be rooted in an understanding of elec-tronic structure. We have highlighted thepeculiarity and complexity of plutonium re-sulting from the 5f electrons. The complex-ity, hidden until after the war, makes theaccomplishments of the metallurgists andchemists during the Manhattan Project evenmore remarkable. ■

W. N. Miner, Ed., Plutonium 1970 and Other Actinides (American Institute of Mining, Metallurgical, andPetroleum Engineers. Inc., New York, 1970).

LOS ALAMOS SCIENCE Winter/Spring 1983 151