The Uranium Bomb^ the Calutron,and the Space-Charge ProblemA participating scientist relates the story of a World War IIproject dedicated to electromagnetically separatinguranium-235 from uranium-238.

William E. Parkins

U ranium! Uranium! Uranium!" A voice sbouted out intotbe night from the second floor of a dormitory in Oak

Ridge, Tennessee. It was 6 August 1945. That day, Presi-dent Harry S Truman bad announced to the world that theUS had dropped a new weapon, a uranium bomb, on thecity of Hiroshima, Japan. For years, tbose of us on thebomb project were cautioned not to say the word uranium,but now it was okay. There were code words and code let-ters for tbe things we worked with, and eacb of our newdesigns received a new name. The teletype messages tbatwent back and forth between tbe radiation laboratory inBerkeley, California, and the Y-12 plant in Oak Ridge weretotal gihherish. The purpose of our effort was to separate"P," or ^ ''U from "Q," or ^ U. Tbose were easy to remem-ber hecause P stood for precious and Q stood for qrap.

Afew days later, another word burst on the scene withthe issuance of the Smyth Report, the official governmentaccount of the history of the bomh project.' That word was"calutron." Now that the device had achieved its objective,Ernest Lawrence wished to give recognition to the Uni-versity of California by using the name calutron for the ap-paratus developed to separate P from Q. He had made anarrangement with report author Henry Smyth that thename he included, but never divulged the deal until thewar was won.

The calutron's separation method was based on elec-tromagnetic mass spectrometry. (See box 1 on page 46 for atutorial.) All critical material was transported in beams ofpositive ions on wbich electric and magnetic fields could act.The needed quantity of material demanded very intensebeams witb a bigh density of electric charge. But the posi-tive beam itself should create so-called space-charge fieldswhose repulsive forces would alter ion trajectories and pre-vent the desired isotopic separation by mass from occurring.At least in 1940, any thinking physicist knew that.

Cornell UniversityThe calutron story starts around 1940 at Cornell Univer-sity. Fellow student A. Theodore Forrester and I were fin-ishing our graduate work under the direction of Lloyd P.Smith, who had obtained a contract to separate a quantityof lithium-6 for use in an experimental study of a new can-cer therapy. Knowing about the space-charge problem and

Bill Parkins retired from Rockwell International, where he wasdirector of research and technology for the energy systemsgroup. He now lives in Woodlarid Hills, California.

© 2005 American Inslitule ot Physics. S-0031-9228-0505-020-9

realizing he had to deal with only twofractions, " Li and ' Li, Smith suggestedtrying a geometry, illustrated in figure 1,similar to that of the electron mag-netron. An arc ion source would bealong the centerline parallel to themagnetic field, and ions would be ac-celerated radially outward. They

would describe circular paths, and the fields would be ad-justed so that the heavy fraction would collect near the180° focus on a peripheral cylinder while the light fractionreturned to be collected near the center. Smith reasonedtbat the symmetry would eliminate space-charge fields inthe 0 direction (in the usual cylindrical coordinates).Space-charge fields would have r and z components, andSmith had calculated how much ion current one would ex-pect before the radial fields ruined the resolution in massseparation.

The experimental apparatus was readied in the sum-mer of 1941. Almost immediately, we observed clean reso-lution with higher currents than should have been possi-ble. Shortly thereafter, we realized that we had stumbledonto a process wherein the ion beam automatically neu-tralizes itself by ionizing residual gas in the vacuum cham-ber. The positive ion beam presents a potential well to elec-trons. They are trapped while the ions they leave areimmediately swept out along magnetic-field lines in the zdirection. And tbe process is fast. Even when the acceler-ating voltage is swept at 60 Hz, the neutralization followsthe beam location perfectly.

Furthermore, the process is self-limiting. Electronsaccumulate while oscillating up and down along magnetic-field lines only until tbe potential well is filled. They movelaterally with small cycloidal paths because of any resid-ual space-charge fields perpendicular to the magneticfield. Collisions of the electrons with gas molecules causethe electrons to start new cycloids—fortunately alwayscloser to the center of the beam. {See box 2 on page 50 formore on heam neutralization.)

We at Cornell didn't know tbat at this very time, theUranium Committee, an arm of the Office of Scientific Re-search and Development (OSRD) under Vannevar Bush,was negotiating with Smyth and Lawrence to start proj-ects at Princeton University and tbe University of Cali-fornia. These would investigate whether space-charge ef-fects might be overcome sufficiently to permit use of anelectromagnetic method for quantity separation of ^ ^U.Lawrence, who was in the process of building a giant 184-inch magnet at UC Berkeley, volunteered the use of it andthe existing 37-inch cyclotron magnet. He would try theclassic Dempster mass-spectrometer arrangement'^ andsee how well he could do. Smyth proposed a time-of-flightmethod he called the isotron, invented by Robert R. Wil-son. It used no magnetic fields, but employed a broad two-dimensional ion source to increase currents. Both of thoseprojects were started in late 1941.

May 2005 Physics Today 45

Word reached Lawrence of our workat Cornell. He contacted Smith and in-vited the three of us to join his proj-ect at Berkeley. Pearl Harhor hadrecently been attacked, and thecountry was at war. We felt aduty to go, although Smithwould have to return withina few months. In mid-February we boarded atrain in Ithaca and leftthe ice and snow forsunny California. From arailroad station on theway, we mailed a manu-script to Physical Reviewwith the request that, be-cause it should now be re-garded as secret, its publi-cation he postponed until thewar was over. We wished toget credit for having discoveredand explained the automatic self-neutralization of intense ion beamswhere there are no applied electricfields. Our manuscript was received on18 February 1942 and published on 1 Decem-ber 1947 after declassification.^

The Berkeley radiation laboratoryWhen we arrived at Berkeley, experiments were alreadyunder way in a small vacuum-chamber tank placed be-tween the poie pieces of the 37-inch magnet. Also, tberewas arc-ion-source development using a smaller magnetfrom the cosmic-ray program. Getting sufficient ion cur-rent from the source was the greatest problem. By mid-March the ion currents were up and, for the first time, ex-ceeded those possible without some space-chargeneutralization.

The current above which resolution in mass separa-tion is lost may readily he calculated from tbe divergenceof the ions at the heam's boundary, which is caused by thespace-charge field there. In the Dempster spectrometer,the beam is accelerated at the ion source from a narrowslit that is long in the direction of tbe magnetic field. Tbebeam takes tbe shape of a douhle-bladed wedge bent intoa semicircle, and comes close to focusing at tbe 180" point.If the space-charge field widens the focus even further,

Figure 1. The radial magnetic separator atCornell University was used to separate

lithium isotopes. This cross sectionshows the relationship of its ion tra-

jectories with those of the con-ventional Dempster mass spec-

trometer (blue area). Alsoshown are a circular arc (redarea) struck trom the cathode(A), short alternating sec-tions of metal tubing andscreen at ground potential(B), and matching sectionsof metal tubing and openmesh grid (C) at the accel-erating potential. The loca-tion of the collector pocket

for the lighter isotope is indi-cated (D), as is the focus

where the heavier isotope col-lected (E). For ease of viewing,

the elements A, B, and C havebeen somewhat enlarged.

until the additional width is equal to theseparation distance of the two isotopes being sep-

arated, then tbe useful beam current limit has beenreached. The current / of the desired isotope in milliampsper centimeter of height of beam in the magnetic-field di-rection is^

/=8 .55x VH,

where rj is tbe fractional abundance of the desired isotope,Aj and A are the atomic weights of the two isotopes beingseparated. Vis tbe accelerating voltage, and H is the mag-netic field in gauss.

For separating ' ' ' U from ' ''U, the maximum / of ^"Uis 4 X 10" niA/cm for a voltage of 35 kV and a magneticfield of 3500 gauss. With that current, it would take moretban 5 years to accumulate 1 kg of' •''U with 1000 separa-tor units having beams 60 cm in height and operating atfull capacity without interruption. That was consideredunachievable, so our challenge was to see how much beamcurrent might be increased above the space-cbarge limit.

A significant change took place in the beginning of

Box 1. Electromagnetic Mass Spectrometry

Despite their varied geometries and field combinations,mass spectrometers incorporate two steps, each of

which filters ions of similar energy, momentum, or velocity,tn each step, the trajectories are completely determined bythe physical parameters of the apparatus along with themass, charge, and velocity of the ion. For example, an en-ergy filter can be a simple acceleration of the ion withcharge e and mass m through a potential difference V. Byequating the energy gained through that acceleration withthe ion's kinetic energy, one obtains mv'/2 eV. where v isthe ion's velocity.

A magnetic field H perpendicular to the path of the ionprovides a momentum filter. The ion subjected to such a fieldwill describe a circular path of radius r. After equating cen-tripetal and magnetic forces, one obtains the momentum ex-pression mv - Her.

A velocity filter can use crossed electric and magneticfields. In a simplest case, the electric field E and magnetic

field Hare perpendicular and the ion moves in a straight lineorthogonal to both fields. Equating the electric and magneticforces determines that v - E/H.

From any two of the three equations, velocity can be elim-inated as a variable. Thus, using two mass-spectrometer stepsallows a discrete solution for mass, which effectively resultsin spatial separation of different isotopes.

For some mass spectrometers, the electric field varies intime. Such time-of-flight spectrometers typically use a fixed-potential accelerating electric field as an energy filter, fol-lowed by a velocity t'llter that uses an RF electric field. Nomagnetic field is necessary.

A Dempster-type mass spectrometer' uses the fixed-potential energy f ilter, followed by a semicircular path in auniform magnetic field as a momentum filter. The calutrondeveloped during World War II modified that basic arrange-ment to increase ion currents while still retaining adequateresolution in mass separation.

46 May 2005 Physics Today http://www.physicstoday.org

Figure 2. The 184-inch magnet at the University of California,Berkeley. The gap between the 184-inch diameter pole pieces

is 72 inches. The photo, courtesy of the Lawrence BerkeleyNational Laboratory Image Library, is from

June 1942: The giant 184-inchmagnet on the hill hehind thecampus became ready for oper-ation (see figure 2). Smith hadreturned to Cornell. More andmore of Lawrence's former stu-dents were arriving to join theeffort. The center of activitiesmoved to the 184-inch magnethuilding, although work contin-ued at the old radiation lahhuilding and at others on thecampus. Rohert Oppenheimer'stheoretical group, which was as-sisting us, operated from thephysics huilding, Le Conte Hall.

Big-time physicsThe large circular huilding thathoused the giant magnet wasideal for our purposes. Aroundits wall on the inside were nu-merous shops and hatches ofheavy electrical equipment. An upper level included officesand conference rooms. In the center was the hig magnetwith a hastily erected platform at the level of the lowerpole-piece surface. Two large vacuum-chamher tanks withslightly over 2 feet of inside clearance in the magnetic-fielddirection were stacked in the 72-inch gap. They had accessfaces on opposite sides for mounting ion sources and col-lectors. Control rooms for each tank were nearhy at plat-form level.

Crews for each tank worked around the clock in threeshifts. Much of the work was done inside the magnet gapwith the magnet turned on. We all knew not to wear awatch or carry keys. The nails in our shoes were no proh-lem, hut they made walking seem as though we were work-ing in a muddy field. We had nonmagnetic tools made froma beryllium-copper alloy. Occasionally, a nail or other fer-rous object would get loose and come flying like a bullet tothe lip of the pole piece. After a couple of accidents, welearned that the liquid-nitrogen-containing Dewar flasks,which were in large metal canisters on casters, needed tohe chained to the railing at the edge of the platform. A sofawas placed on top of the magnet where anyone detainedfor an extended run could catch a nap. Also, it was thewarmest location in the huilding on a foggy night.

One immediate need was for experts in high-power-circuit design. Lawrence contacted movie studios in Hol-lywood. They were looking for work and immediately senta very competent team that stayed through our entire proj-ect. Marcus Oliphant, who had come to the US from Eng-land, arranged for a group of superb physicists to comefrom that war-torn nation. Both Oliphant and HarrieMassey, who also came, were subsequently knighted by thequeen. The General Electric Research IJahs sent a goodgroup headed hy Kenneth Kingdon. Westinghouse sent ateam led hy William Shoup, who was joined later hy Ed-ward U. Condon. Lots was happening; everybody was co-operating and one could feel the excitement. We were likea swarm of bees in a huilding that even looked like a hive!

But there was no question as to who had the role ofqueen bee. It was, as we all called him, E.O.L. But to his

face, it was always Ernest. Lawrence, seen in figure 3, wasa big man with strong hands, a healthy boyish complexion,a ready smile, and a hig shock of hair. But most impressivewere those penetrating hluish eyes. Nobody worked harderor had more enthusiasm than Lawrence, and his approachwas perfect for the kind of development heing done. He be-lieved in thinking a little and experimenting a lot. There wereso many variables and we lefl none uncovered. I can't beginto explain all of the interesting avenues our work explored.But I must mention one that had critical consequences.

On the matter of neutralization of the space charge,no means of introducing electrons into the beam workedbetter than simply depending on the beam to ionize theresidual gas in the tank. But the pressure had to he ahout2 X 10" torr. Half of that and the beam became unstable,especially with sparking. Unfortunately, we were trying toseparate a lighter isotope from a much more abundantheavier one, - "U. Scattering of the beam by residual gascaused the heavier ions to reduce energy and radius, andto enter the collector intended for the -'" U, therehy reduc-ing the enrichment obtained. We knew this, because ex-periments decelerating the collected ions to near zero en-ergy improved the enrichment. But such a collectorreduced the final currents too much.

A compromise had to be struck, and reluctantly thedecision was made to go to a two-stage process to achieveenrichments necessary for the bomh design. The stageswould be called the alpha and beta. In the beta stage, therewould be the new problem of chemical recovery of all ura-nium used because beta-stage feed material would be sovaluable.

The atomic homb project was now being taken veryseriously. The Manhattan District of the US Army Corpsof Engineers, under the command of Major General LeslieR. Groves, had been brought in to take charge. It had al-ready set up the Los Alamos weapons lab in New Mexico.The isotron project at Princeton was shut down for lackof any positive results. Key personnel from there, includ-ing Wilson and Richard Feynman, went to Los Alamos.The project at Berkeley was transferred from OSRD to

http://www.physicstoday.org May 2005 Physics Today 47

Figure 3. Ernest Lawrence as he poses on the bighind the University of California, Berkeley, campus. The

184-inch magnet, housed In the building visible in thebackground, became operational in June 1942, about

the lime this picture was taken. {Courtesy of the EmilioSegre Visual Archives, PHYSICS TODAY Collection

the US Army on 1 May 1943. A short timelater came the surprising tiews that a plantfor its process would be built in eastern Ten-nessee, where there was access to powerfrom the Tennessee Valley Authority.

The calutronThe electromagnetic isotope separator designdeveloped at Berkeley was a variation of theDempster mass spectrometer. Did it reallydeserve the name calutron, which advertisedthe university? I believe it did, hecause of fourimportant new features that contributed toincreased throughput and resolution. Theseare features other than the use of extremelyhigh accelerating voltages.

The first new feature was the use of mag-netic shims. As designed by Oppenheimer'stheoretical group, the shims employed twoshaped iron sheets approximately 3 feet wideand extending all the wayacross the tank, asviewed from the ionsource. They were boltedto the top and bottom sur-faces in the tank's centralregion. Their purposewas to slightly increasethe magnetic field. That preferentially, if onlyvery slightly, decreased the radii of the tra-jectories of ions exiting the ion source with small diver-gence, and brought those ions to a focus at the collector to-gether with the ions of wider divergence. In effect, thealtered magnetic field produced better resolution withwider angular divergence from the source, which improvedboth throughput and resolution. A disadvantage was thatthe focus at the collector was an odd-shaped nonplanarcurve instead of a straight line.

Lawrence wanted to alleviate the problem with so-called fish shims that would be anchored at midplane likea flat fish. To Lawrence's great disappointment. Grovessaid "No," and that was that. Groves felt the urgency andwisely foresaw new problems. But he had the greatest ad-miration for Lawrence, whom be considered a nationaltreasure. He even refused to allow Lawrence to fly, and re-quired him only to travel by train.

The second calutron innovation was tbe use of multi-ple beams. Several arc ion sources were located a fewinches apart on a line perpendicular to the initial directionof tbe accelerated beam. Of course, multiple collectingpockets bad to be provided at the 180" focus. In traversingf"rom source to collector, the beams had to cross, and at firsttheir mutual interference caused great trouble. HansBethe, who was visiting at the time, quipped, "Lawrenceexpected multiplication, but he got division." With furtherexperimentation, we eventually learned the conditions forstable operation with multiple beams. Production designsincorporated eitber two or four beams per tank.

Other featuresUp until tbat time, the ion source had been operated atground potential and the accelerating slit at high negativevoltage. That required the collectors to be at higb negativevoltage and to bave a metallic tank liner at the high neg-ative. The liner was a constant problem, and it reduced tbeusable beam height. The next major innovation was toeliminate the high-voltage liner, put the collectors atground potential, and operate the source at high positivevoltage. But that created a monstrous problem: The tank

was at ground potential, so the region around the sourcehad exactly tbe conditions for the classic Phillips dis-charge—magnetic field extending between negative endplates and an electrode maintaining a positive region inbetween.

The region immediately surrounding the source pre-sented a positive-potential well to electrons, and trappedthem much as the ion beam did. But in this case, the ap-plied field was one tbe electrons could not neutralize. Andnow the electrons could have many thousands of volts ofenergy, depending on where they were created by ioniza-tion. At the residual pressure around the source, the elec-tron current oscillating up and down the magnetic-fieldlines could multiply exponentially into an avalanche thatwould destroy any positive electrode surface that finallycollected it. Holes were melted through quarter-inch-thicktungsten plate.

The solution, while hard to descrihe, was truly ingen-ious. A grounded and fitted shield, with plenty of perfo-rated holes to allow for vacuum pumping, was installed toclosely surround the big source block. The shield had ob-long bulges called blisters that were each several incheslong and that enclosed fins—short lengths of copper-platestrips brazed to the source block. Two overlapping se-quences of fins and blisters just above and below tbe mid-plane perpendicular to the magnetic field reduced the dis-tance in which the trapped electrons could oscillate fromabout 2 feet witbin tbe tank to about 2 inches within theblister. Furthermore, as the electrons traveled laterallywithin the blister, they executed cycloids in tbe directionof E X H. Tben, as tbey went around tbe end of a fin, tbeywould see, along magnetic-field lines, tbe overlapping finin the otber plane and be collected. That arrangement pre-vented wasteful electron drain currents from building up,and it worked!

The fourth major feature of the calutron was the so-called accel-decel. The positive source now made it possi-ble to interpose a high negative accelerating slit betweenthe ion source slit and the final slit at ground potential. The

48 May 2005 Physics Today http://www.physicstoday.org

ions would first be accelerated by, for example, 70 kV, andthen decelerated by 35 kV to follow trajectories at 35 kV intbe grounded tank. The very high initial voltage allowed usto extract higher ion currents. But because of tbe deceler-ation step, one could retain the same radius of trajectory intbe tank without increasing the magnetic field.

The large amount of electrical equipment at high volt-age presented our greatest hazard. We took care to use cir-cuit interlocks on gates and cages. Those were essentiallyswitches that, wben opened, prevented tbe bigh voltagefrom being turned on. Fortunately no one was electro-cuted, altbough in the frenzy of work, there were a fewclose calls. I was one of those, and I owe my life since tbatday to a scream by a quick-thinking Frank Oppenheimer.

Tbe ion sources, except for tbeir large scale, were moreconventional. Each was a full-length arc struck between areplaceable, large-gauge tantalum wire filament and agrapbite box enclosing the arc. Tbe heavy source block,supported on porcelain bushings, bad beaters thatwarmed a large reservoir of uranium tetracbloride. Care-ful temperature regulation ensured optimum vapor pres-sure in tbe boxes. Willard Libhy, later of carbon-14 fame,came from the gaseous diffusion project at Columbia Uni-versity to acquaint us with tbe tricks of using uraniumbexafluoride. That compound had the advantage of beinga gas at room temperature and would have been easier touse. In tbe end, UF,. was rejected because of a mucb re-duced lifetime for the arc filament and its colhmating slot.

In late 1943, the production design of what was calledAlpba-I bad to he frozen before development work on tbehigb-positive source and accel-dece! bad been completed.Tbose features would later be incorporated into tbe Alpba-II design. Tbe plant in Oak Ridge was coming together.Tbe first team to go tbere bad tbe assignment of mappingtbe magnetic fieids. Others for plant start-up would soonbe following.

The Y-12 plantThe US Army Corps of Engineers had created a town intbe mud of the Tennessee hills west of Knoxville. In adja-cent valleys, army contractors were building plants for the

Figure 4. Tbe Y-12 plant in OakRidge, Tennessee, was a part of (beManhattan Project. The site includednine production buildings and wasdedicated to electromagneticallyseparating uranium-235 fromuranium-238. (Courtesy of Oak

' National Laboratory.)

Manhattan Project. Ourswas the Y-12 plant (seefigure 4), which eventu-ally included nine largeproduction buildings.Each building containedone or two racetrack-sbaped assemblies sucbas tbe Alpba-I tracksbown in figure 5. Theracetracks alternatedtanks, set on edge, withmagnet excitation wind-ings. Copper was scarce,and the US Treasury De-partment loaned 15 000tons of silver for thewindings. As many as 96

gaps for tanks were designed into a single racetrack. It wasstaggering!

Big industry had been called in. The engineer-constructor was Stone and Webster of Boston; the race-tracks and their large generators were built by AllisChalmers; Westingbouse made the internals, sources, andcollectors; General Electric handled tbe electrical cubicles;and tbe operating company was set up as a division ofEastman Kodak called Tennessee Eastman.

Tbe first track to go into operation, in early 1944, wasthe Alpha-I. It had shims, two beams per tank, a groundedsource, and no accel-decel. Its performance was not nearlyas good as that of tbe Alpha-II, which followed. Alpba-IIhad sbims, four beams, a high-positive source, andaccel-decel. The Beta tracks tbat followed next were likeAlpha-II, except parameters were halved—two beams,2-foot-radius paths instead of 4-foot, and a beam height of7.5 inches instead of 16. The magnetic field for tbe Betatracks bad to be doubled to between 6000 and 7000 G. Andtbe Beta tanks required a water-cooled stainless steel linerfor the cbemical recovery of all uranium tbat did not reacbtbe collector pockets.

Tbe design of all of tbis equipment was hased on re-sults from experimental apparatus at Berkeley. The rushto get into production omitted any pilot-plant phase, andmany serious start-up problems resulted.^ Tbe magnetcoils in tbe first Alpha-I track bad shorts to ground fromrust and sediments in the cooling oil. A four-montb delayensued wbile all of tbe silver coils from tbat track weresbipped back to the Allis Chalmers plant in Milwaukee,Wisconsin, for cleaning and otber corrective measures.

Alpba-II suffered devastating failures of the large in-sulating busbings supporting tbe source block. Operationswere hampered for months until we obtained bushingsmade from improved porcelain. Even tbe cbemical recov-ery of enricbed uranium was initially so poor tbat it tbreat-ened the viability of the overall project. These and otherdifficulties were overcome one by one, as Lawrencesteadily maintained his unflagging optimism and drive.

One of tbe best beam diagnostic techniques was sim-ply to look. Through a window with protective shutter built

http://www.physicstoday.org May 2005 Physics Today 49

Box 2. Space-Charge Neutralization

Since the time of Charles Darwin, discovery of simple,effective, and beneficial biological processes has not

been a surprise—they all have had the slow guiding hand ofevoiution. But it is extremely rare to discover such a processin the physical world. Instead, society undertakes to bendnature to serve its technological needs. Automatic self-neutralization of ion beams is a wonderful solution to theproblem of space-charge repulsion, and it occurs totallywithout human intervention.

With some form of Dempster mass spectrometer, all thatis required is a low pressure of residual gas in the vacuumchamber to permit ionization by the ion beam itself. Scatter-ing of ions by that gas reduces the enrichment of separatedisotopes, but a broad window of operating conditions makespossible essentially 100% neutralization and reasonably highenrichment factors.

With the Alpha-ll calutron operation, about 10% of thebeam ions traversing from source to collector cause ioniza-

tion. Given their transit time, the beam would be neutralizedIn a few milliseconds. There is then continuing production oftrapped electrons; those having a little more momentum inihe direclion of the magnetic field are constantly replaced byelectrons of lower energy. With any interruption of beam cur-rent, the electrons would be swept out in microseconds.

Discovered' in 1941, beam neutralization• is established in all beam regions at a rate independent ofthe beam current in the region,• can follow fluctuations in beam location and intensity atfrequencies up to aboul 1 kHz,• appears to be 100% effective with steady beams, and• places no limit on total beam current.The process is a gift from Nature and, in contrast to those inthe iiiological world, is one of a small number of beneficialnatural physical processes. It has made possible the quantityseparation of isotopes of elements throughout the periodictable for use in science, technoiogy, and medicine.

into the tank wall, one could easily see the beam floatinglike a pale blue 3D ghost. Depending on conditions, ion-ized chlorine or Cl , ionized combinations of U and Cl, andeven doubly ionized U might be visible. And in the Betatank, one could actually see the beam of ^ ''U, although itwas approximately lC/f of the total U current.

The calutron was a temperamental piece of equip-ment. Each unit was operated from a control panel at thefront of a cubicle containing the electrical equipment. Theoperators were mostly young women scoured from the backcountry of Tennessee. Many were uncomfortable having towear shoes, and their drawl was often hard to understand.But when it came to patience, no one was their superior.And the calutron required patience. With no scientific un-derstanding, those women became much better operatorsthan our lab personnel from Berkeley.

As many as 22 000 nonscientist employees worked atY-12.' What did they think wewere doing? The plant had noreceiving dock and no loadingdock. Nothing went in or out,but everybody was very busy.For a time, they were given acock-and-bull story aboutbroadcasting radio signals thatjammed the communications ofour enemies in Europe and thePacific. But it was wartime, andpeople accepted the fact that ourwork, whatever its purpose,must be important.

Late in 1944, a fortunatechange boosted the throughputand final ^ ''U enrichment fromY-12: Low-enrichment feed ma-terial from Oak Ridge's thermaldiffusion plant became avail-

Figure 5. This Alpha-I racetrack was partof the Y-12 plant in Oak Ridge, Ten-

nessee. In it, process tanks with calutronunits alternate with magnetic windings

made of silver. Electrical-equipment cubi-cles with control panels are in a separate

bay. Vacuum pumps are on a lower level.

50 May 2005 Physics Today

able. Its output, obtained independently of the Y-12 efTort,was introduced at a time when Y-12 was at its full capac-ity of 1152 tanks. A big push was mounted around the be-ginning of 1945 that led, within a few months, to the ac-cumulation of the necessary enriched uranium, and toTruman's announcement of 6 August.

Just three years earlier, our challenge had been to seehow much we could exceed the limit set by the space-chargeeffect on ion beam currents. Now we knew. A good run onan Alpha-II unit would last 7 to 10 days and was usuallyterminated when the graphite facing on the ^ U collectorpocket had been sputtered away. During that time, the totalcharge measured to the • 'U collector would be in the neigh-borhood of 15 A-h per beam. If averaged over a one-week pe-riod, that amounts to a beam current approximately 400times the calculated space-charge limit!

http://www. physicstoday.org

The legacyThe electromagnetic separation of ^ U could not havetaken place without our having overcome the space-chargelimitation. We at Cornell and, later, Lawrence at the Uni-versity of California independently suggested using aDempster-type instrument—-the only basic electromag-netic method of isotope separation that offers a solution.Its elements are an accelerating-potential-based energyfilter in which space-charge fields are unimportant, fol-lowed hy a momentum filter consisting of a drift space ina magnetic field, where no applied electric fields exist andwhere space charge can he neutralized by trapped elec-trons. No other combination of steady electric and mag-netic fields has that capability.

During World War II, the Germans gave up any at-tempt to separate '•'"'U electromagnetically because theydid not consider the space-charge problem soluble.^ TheJapanese also considered the electromagnetic method, hutgave up on it for the same reason.''"

We at Cornell were the first to observe and explain theion beam neutralization process. Our article, withheld frompubhcation for more than five years, was titled "On the Sep-aration of Isotopes in Quantity hy Electromagnetic Means."That, of course, was the purpose of the calutron project thatcame later and produced the Hiroshima bomh material. Butthe bomb has not been the project's most important legacy.Other methods were being used to make bomh material. Aplutonium bomb was dropped over Nagasaki, Japan, threedays afler Hiroshima was hombed, and the gaseous diffu-sion plant was coming on line. It would produce enricheduranium at a much faster rate and lower cost.

The most important legacy of the project has been thecontribution to science, technology, and medicine madepossible through the use of separated isotopes of nearly allthe elements of the periodic table. Hundreds of kilogramsbave been prepared for research and diagnostics inphysics, chemistry. Earth sciences, biology, and medicine.This service bas heen provided at cost for almost 60 yearsthrough the use of calutrons in the pilot units and Betatracks at Y-12, all operated by Oak Ridge National Labo-ratory. Nationally and internationally, thousands of cus-tomers and millions of medical patients bave benefited.^

The development and use of the calutron to produceenriched uranium for the first atomic bomb that was ex-ploded in warfare, and then to produce tbe full spectrumof separated isotopes for uses in peacetime, is tbe greatestexample of heating swords into plowshares in tbe historyof humankind. For its contribution in both wartime andpeacetime, tbe physics profession can be proud.

References1. H. D. Smyth, Atomic Energy for Military Purposes, Princeton

U. Press, Princeton, NJ (1948).2. A. J. Dempster, Phyn. Rev. 11, 316 (1918).3. L. P. Smith, W. E. Parkins, A. T. Forrester, Phys. Rev. 72, 989

(1947).4. R. G. Hewlett, 0. E. Anderson, A History of the United States

Atomic Energy Commission, vol. 1, Pennsylvania State U.Press, University Park {1962).

5. C. W. Johnson, C. O. Jackson, City Behind a Fence: OakRidge, Tennessee, 1942-1946, U. of Tennessee Press,Knoxville(1981).

6. D. Irving, The German Atomic Bomb, Simon and Schuster,New York (1967).

7. F. H. Schmidt, Science 199, 1286 (1978).8. W. E. Parkins, Science 200, 255 (1978).9. L. O. Love, Science 182, 343 (1973). •

May 2005 Physics Today 51

Step Into the3D World of AFMwith the MFP-3D°

AFM 3D rendering (top), anaglyph (below) of

quantum dot structure created on a GaAs

wafer using oxidation lithography, 2.5 pm scan

Image courtesy D. Graf and R. Shieser,

Ensslin Group, ETH Zurich

Email us for your own pairof 3D glasses

YLUMwww.AsylumResearch.com805-696-6466sales@AsylumResearch.com

wfww.pt.ims.ca/5988-25 or Circle #25