The discovery of fissionOtto R. FrischJohn A. Wheeler

Citation: Physics Today 20, 11, 43 (1967); doi: 10.1063/1.3034021View online: http://dx.doi.org/10.1063/1.3034021View Table of Contents: http://physicstoday.scitation.org/toc/pto/20/11Published by the American Institute of Physics

Articles you may be interested inOn the belated discovery of fissionPhysics Today 68, (2015); 10.1063/PT.3.2817

The Discovery of Nuclear FissionPhysics Today 42, (2008); 10.1063/1.881174

The Discovery of FissionInitial formulations of nuclear fission are coloredwith the successes, failures and just plain bad luck of several scientistsfrom different nations. The winning combination of goodfortune and careful thought made this exciting concept a reality.

by Otto R. Frisch and John A. Wheeler

Otto R. Frisch, professor of naturalphilosophy (physics) at CambridgeUniversity, England, did research in

Berlin (1927-30), Hamburg (1930-33), London (1933-34), Copen-

hagen (1934-39) and Birmingham(1939-40). During the war he

worked on the A-bomb at LosAlamos. He was first to observe

energy liberated in the fission of asingle uranium nuoleus.

John A. Wheeler, one of the firstAmerican scientists to concentrateon nuclear fission, worked at theU. of Copenhagen in 1934 as aNational Research Fellow withNiels Bohr. Wheeler received hisPhD in physics at the JohnsHopkins University prior to hisresearch in Copenhagen. In 1938he joined Princeton's physics de-partment, where he remains active.

How It All Beganby Otto R. Frisch

THE NEUTRON was discovered in 1932.Why, then, did it take seven years be-fore nuclear fission was found? Fissionis obviously a striking phenomenon; itresults in a large amount of radioactiv-ity of all kinds and produces fragmentsthat have more than ten times the totalionization of anything previouslyknown. So why did it take so long?The question might be answered bestby reviewing the situation in Europefrom an experimentalist's point ofview.

Research in Europe

In Europe there were few laboratoriesin which nuclear-physics research wasconducted, and I think the word"team" had not yet been introducedinto scientific jargon. Science wasstill pursued by individual scientistswho worked with only one or two stu-dents and assistants.

Paris harbored some of the most ac-tive research laboratories in Europe.It is the city in which radioactivityhad been discovered and where Ma-dame Curie was working until herdeath in 1934. She still dominatedthe situation: Techniques were quitesimilar to those used at the turn of thecentury; that is, ionization chambersand electrometers. This state of af-fairs is good enough for performing ac-curate measurements on natural ra-dioactive elements, but it is not reallyadequate for much of the work on nu-clear disintegration. Madame Curie

had little respect for theory. Once,when one of her students suggested anexperiment, adding that the theoreti-cal physicists next door thought ithopeful, she replied, "Well, we mighttry it all the same." Their disregardof theory may have cost them the dis-covery of the neutron.

Cambridge is the second place wor-thy of discussion. Ernest Rutherford,whose towering personality dominatedCambridge research, had split atomicnuclei in 1919; since 1909 he had, infact, been keenly concerned with theobservation and counting of individualnuclear particles. He first introducedthe scintillation method and stuckfirmly to it. His great preference wasfor simple, unsophisticated methods,and he possessed a strong distrust ofany complicated instrumentation.Even in 1932, when John Cockcroftand Ernest Walton first disintegratednuclei by artificially-accelerated pro-tons, they used scintillations to detectthe process. By that time Rutherfordhad realized that electronic methodsof particle counting must be devel-oped. The reason was that the scin-tillation method clearly had its short-comings. It did not work for very lowor high counting rates and was notreally reliable. This deficiency washighlighted by the results that camefrom the third laboratory I want tomention—Vienna.

Vienna is where I began my careerand it was in those days a sort of en-

PHYSICS TODAY • NOVEMBER 1967 • 43

fant terrible of nuclear physics. Sev-eral physicists were claiming that notonly nitrogen and one or two others ofthe light nuclei could be disintegratedby alpha particles but that practicallyall of them could and did give manymore protons than anybody else couldobserve. I still do not know how theyfound these wrong results. Apparent-ly they employed students to do thecounting without telling them what toexpect. On the face of it, that opera-tion appears to be a very objectivemethod because the student wouldhave no bias; yet the students quicklydeveloped a bias towards high num-bers because they felt that they wouldbe given approval if they found lots ofparticles. Quite likely this situationcaused the wrong results along with agenerally uncritical attitude and con-siderable enthusiasm over beating theEnglish at their own game.

I still remember when I left Viennaat just about that time (after havingescaped the duty of counting scintilla-tions). My supervisor, Karl Przibram,told me with sadness in his voice, "Youwill tell the people in Berlin, won'tyou, that we are not quite as bad asthey think?" I failed to persuadethem.

Germany had nuclear-physics re-search in several places. The team ofOtto Hahn and Lise Meitner, whichhad been one of the first groups tostudy radioactive elements, had at thattime separated to carry out indepen-

dent research. Hahn was working onvarious applications of radioactivityfor the study of chemical reactions,structures of precipitates and similarsubjects, whereas Lise Meitner wasusing radioactive materials chiefly toelucidate the processes of beta andgamma emission and the interaction ofgamma rays with matter.

In addition, Hans Geiger was inGermany. He had been with Ruther-ford from 1909 onwards, in the earlydays before the nucleus was discov-ered. Rutherford felt uncertain aboutthe scintillation method and askedGeiger to develop an electric counterto check on it. But as soon as Ruther-ford saw that the two gave the sameresults, Rutherford returned to thescintillation method, which appearedto be simpler and more reliable whenused with proper precaution. Geigerwent back to Germany and perfectedhis electric counters, and in 1928, to-gether with a student namedW. Miiller, he developed an improvedcounter that could count beta rays.Earlier counters were inadequate forthis purpose, and scintillation methodswere also incapable of detecting betarays. However the new counters werestill very slow because the dischargebetween the central wire and the cy-lindrical envelope was quenched by alarge resistor of many megohms placedin the circuit; consequently the count-ing rate was limited to numbers notmuch greater than with the scintilla-

( I )

(2)

(3)

92U+

92U +

92U+

n->

(92U -f- n)

(92U + n)

(92U + n)

ioSek.

—?->59 Min.

2,5 Std.

P40 Sek.

57Std.

23 Min.

93EkaRe

95EkaIr

97EkaAu

93EkaRe

95EkaIr ?

93EkaRe

> 9 4EkaOs^^^^2,2 Min.

66 Std! 9G J H l i

—•-• 9 4 E k a O s ^16 Min. 9 4

?

COMPUTATIONS, indicating chains of radioactive elements, were published in a1938 Die Naturwissenschaften article by Hahn, Meitner and Strassmann. —FIG. 1

tion method. Even at a few hundredparticles a minute there were quitelarge corrections to be applied.

Walther Bothe was the first to usethe coincidence method, both in an at-tempt to do something about cosmicrays and also for measuring the energyof gamma rays by the range of the sec-ondary electrons they produced. Thiswas really the first reliable method formeasuring the energy of weak gammaradiations.

Until 1932, the only source of par-ticles for doing atomic nuclear disinte-gration was natural alpha particles:either polonium, which was difficultto come by (in fact one practicallyhad to go to Paris) or sources of one ofthe short-lived decay products of radi-um, which were very clean but wereshort-lived and usually had lots ofgamma radiation.

The year of discovery

But in 1932, that annus mirabilis, notonly the neutron was discovered buttwo other developments took place.In the US Ernest O. Lawrence madethe first cyclotron that showed prom-ise of being useful, and in EnglandCockcroft and Walton built the firstaccelerator for protons capable of pro-ducing nuclear disintegrations. Ineed not state that this was the begin-ning of an enormous development;most of nuclear physics as we know itwould have never come about withoutat least one of those two instruments.But the interesting thing is that theyplayed practically no role in that nar-row thread that led to the discovery ofnuclear fission.

I do not want to dwell on the dis-covery of the neutron very much be-cause it was discussed in several inter-esting lectures in 1962 at the Historyof Science Congress held in Ithaca,New York. The published proceed-ings contained interesting contribu-tions by Norman Feather and SirJames Chadwick, who showed that theneutron was discovered in Cambridge,not simply by chance with everybodyelse having done the groundwork, butbecause a search for the neutron hadbeen going on in Cambridge (admit-tedly with wrong ideas). The peopleat Cambridge were keyed up for thisdiscovery. They had made one obser-vation that was important and thattends to be overlooked: H. C. Web-

44 • NOVEMBER 1967 • PHYSICS TODAY

ster showed that those queer pene-trating rays that beryllium emittedwhen alpha particles fell on it weremore intense in the forward directionthan in the backward direction. Thisresult was quite incomprehensible ifthe radiation were gamma rays as ev-erybody believed. Even the Frenchphysicists Curie and Joliot shared thatbelief in the teeth of all theoreticalpredictions. Then Chadwick's experi-ment showed clearly that the mysteri-ous radiation consisted of particles hav-ing approximately the mass of the pro-ton. There was a bit of confusion atthe time because the word "neutron"had been used by Enrico Fermi andWolfgang Pauli to indicate the particlethat later came to be called the "neu-trino."

After the neutron was discovered,there was of course a certain rush ofactivity, but nobody knew quite whatto do. Neutrons were rather few innumber. They were, after all, secon-dary products of nuclear disintegra-tion. With only natural alpha sourcesavailable at first, neutron productionwas low.

Moreover the main instrument fordetection was essentially the cloudchamber. With cloud chambers onlya limited number of tracks due to neu-trons could be found. And it wasslow work to make any sense out ofthe few detected tracks of recoil nu-clei. Leo Szilard once joked that if aman suddenly does something unex-pected there is usually a woman be-hind it, but if an atomic nucleus sud-denly does something unexpected,there is probably a neutron behind it.

Electronic counting methods hadonly just been developed; largely as areaction to the wrong results comingout of Vienna that nobody else couldconfirm, it had been decided that itreally was necessary to build electron-ic amplifiers and counters. Actuallythe Viennese themselves started thatkind of work but were not very suc-cessful. The work was also started inSwitzerland with some success byHermann Greinacher. Yet I think themain thread that led to the develop-ment of decent counters took place inEngland, where Charles Wynn-Wil-liams used proper screening and tubeswith low noise level etc. to produceelectronic counters. Neverthelessthose counters, although Chadwick

GREAT AND GOOD FRIENDS. Lord and Lady Rutherford (left) withNiels and Margrethe Bohr in Rutherford's garden. The photograph wastaken about 1930.

had used them with good effect to pindown the neutron, were still too noisyto be of much use.

Artificial radioactivityThings really got moving when, in1934, artificial radioactivity was foundby Curie and Joliot. I think they musthave been very happy to have madeup for their failure to spot the neutrontwo years previously. Almost to theday two years previously both discov-eries came out in the middle of Janu-ary. They had known for manymonths before that aluminum bom-barded with alpha particles emits posi-trons, but it had never occurred tothem that this might be a delayedprocess. They had only observed thepositron during bombardment. Law-rence and his cyclotron people in Cali-fornia had made the same mistake. Infact they had noticed that the countersmisbehaved after the cyclotron wasswitched off. I am told that theybuilt in special gadgetry so that thecounters were automatically switchedoff together with the cyclotron! Oth-erwise they would have found artifi-

cial radioactivity before the French.It is astonishing that nobody ap-

pears to have thought beforehand thatthe result of a nuclear disintegrationmight be an unstable nucleus althoughthe existence of unstable nuclei had, ofcourse, been known for thirty years ormore. I have been told that, after thediscovery, Rutherford wrote to Joliotand congratulated him on his discov-ery saying that he himself had thoughtthat some of the resulting nuclei mightbe unstable, but had always looked foralpha particles only because he wasnot really interested in beta particles.

As soon as this work became knownin January 1934 a lot of people rushedto repeat and extend the experiment.But most of them rushed in a straightline indicated by Curie-Joliot, bom-barding other elements with alphaparticles. (So did I in Blackett's labo-ratory in London.)

But in Rome Fermi at that time hadalready decided that nuclear physicswas an important and interesting line,and he had started to set up some in-strumentation. So when this discoverycame along, he began working quite

PHYSICS TODAY • NOVEMBER 1967 • 45

fast to see whether neutrons wouldform radioactive nuclei.

I remember that my reaction andprobably that of many others was thatFermi's was a silly experiment becauseneutrons were much fewer than alphaparticles. What that simple argumentoverlooked of course was that they arevery much more effective. Neutronsare not slowed down by electrons, andthey are not repelled by the Coulombfield of nuclei. Indeed, within aboutfour weeks of the discovery by Curieand Joliot, Fermi published the firstresults proving that various elementsdid become radioactive when bom-barded with neutrons. Only anothermonth later he announced that bom-barding uranium produced some newradioactivity that he felt must be dueto transuranic elements. Becauseboth on theoretical grounds (Coulombbarrier and all that) and as far as theexperiments confirmed it, all heavierelements were known to absorb neu-trons without splitting anything off.And so it was felt that must also be thecase with uranium.

This work was of course considera-bly interesting to radiochemists. Sev-eral took it up, but once again, oddlyenough, one false result started thingsreally moving—a note by Aristid vonGrosse, a German-born chemist work-ing in the US, who thought one ofthese elements behaved like protactin-ium. He had done some of the earlywork with Hahn on protactinium soonafter it was discovered in 1917; so hissuggestion put Hahn and Meitner ontheir mettle. They felt protactiniumwas their own baby and they weregoing to check it. Lise Meitner per-suaded Hahn to join forces again.They soon showed that von Grossewas wrong: It was not protactinium.On the other hand there were so manyodd things there that they were cap-tured by this phenomenon and had togo on. The results were most pecu-liar.

Figure 1 shows one of the tabula-tions indicating the chains of radioac-tive elements that Hahn and Meitnerhad thought identified them. Theydid not give new names to the trans-uranic elements that they thoughtthey had identified, but they used theprefix "eka" to indicate that they werehigher homologues of rhenium, os-mium, etc. up to ekagold. Obvious-

LINKS IN THE CHAIN. Cockcroft(top) and Walton contributed to thenew ideas when they disintegratednuclei by artificially-accelerated pro-tons.

ly, Hahn was excited to have a wholenew lot of chemical elements to playwith and to study their properties.Today, of course, these elements afteruranium are known as neptunium, plu-tonium, americium etc., and areknown to be chemically quite differentfrom those that Hahn was studying.

Parallel chains

The results were astonishing for tworeasons. In the first place, it ap-peared that there were three parallelseries. And from the yields obtained

they must all derive from uranium 238or possibly one of them from 235(which is already much rarer). So itlooked as if there were at least twoparallel chains of isomeric elements.This isomeric property had to be prop-agated all along the chain of betadisintegrations.

Nuclear isomerism was still fairlynew in 1938, and its interpretationwas not altogether clear. It had beensuggested (as we now accept) that itwas due to high angular momentum,but there were also proposals that itmight be due to the existence of rigidstructures inside nuclei. One couldimagine that such a rigid structuremight survive a beta decay and mightinfluence the half-life of the subse-quent product.

But then there was still the mysteryof the great length of those chains.Uranium, after all, was not beta un-stable itself. The other elements inthat region never had more than twobeta decays in succession; yet herefour or five had been found. So Hahnthe chemist was delighted by so manynew elements, but Hahn the radio-physicist or radiochemist was ratherworried about the mechanism thatcould account for them.

All this work was made difficult bythe political situation in Germany.Hitler was in power and the institutehad to play a delicate game of politicsto prevent racial persecution from re-moving some of its personnel. In1938, when Austria was occupied bythe Nazis, Lise Meitner felt very inse-cure; rumors began to float aroundthat she might lose her post and beprevented thereafter from leavingGermany because of her knowhow. Acertain amount of panic resulted.Dutch colleagues offered to smuggleher to Holland without a visa. Thusshe left Germany in the early summerof 1938, went from Holland for a briefstay in Denmark, and was offered hos-pitality by Manne Siegbahn at theNobel Institute in Stockholm.

Near misses

After that, the team that had alreadybrought Strassmann in with Hahn as asecond chemist had to carry on with-out her. In the meantime some workhad been started in Paris. It is inter-esting that they had a different angle.They were at first not so interested in

46 • NOVEMBER 1967 • PHYSICS TODAY

the transuranic elements; but theyrealized that if thorium is bombardedwith neutrons, one ought to find thebeginning of the new and missing ra-dioactive chain with the atomicweight 4n + 1. One realizes that theothers, 4n, 4n + 2, 4n + 3, are allrepresented by the natural radioactiveseries. But the 4n + 1 was missing,and so Irene Curie, the daughter ofMadame Curie, together with Hansvon Halban, an Austrian, and PeterPreiswerk, a Swiss, set out to searchfor that series and published somework on it.

Later that team broke up becauseHalban came to Copenhagen and, fora time, worked with me on the studyof slow neutrons. Irene Curie found anew collaborator in Pavel Savitch, aYugoslav. They tried to disentanglethe transuranic elements. Havingrealized that there was a great varietyof different materials, Irene Curie hadthe good idea of selecting one of themsimply by the high penetration of itsbeta rays. They covered theirsamples with a fairly thick sheet ofbrass and only studied the substancewhose radiation penetrated. They didnot realize that even that methodmight not select a single substance al-though the substance appeared tohave a reasonably unique lifetime of3.5 hours. From the chemical behav-ior they first thought it looked like tho-rium.

This work was checked by Hahn,who concluded that it was not thoriumand wrote so to Paris. Curie and Sav-itch continued the work and in a laterpaper in the summer of 1938 acknowl-edged that the 3.5-hour substance wasnot thorium but behaved a bit morelike actinium and even more like lan-thanum. She had come very close in-deed to the concept of nuclear fissionbut unfortunately did not state itclearly. She said that it was definitelynot actinium and that it was quite sim-ilar to lanthanum, "from which it couldbe separated only by fractionation."But she did think it could be sepa-rated. The reason was probably thatshe still had a mixture of two substan-ces; in that case of course one does ef-fect a partial separation. Then thiswork was in turn checked by Hahnand Strassmann who discovered ra-dioactive products that behaved partlylike actinium, partly a bit like radium.

RUTHERFORD was the first to use scintillation methods to detect particles.

There was another near miss atabout the same time: Gottfried vonDroste, a physicist working with LiseMeitner, looked for long-range alpharays from uranium during neutronbombardment. If he had supressedthe ordinary alpha rays by applying abias to the amplifier, he would nothave failed to find fission. Unfortu-nately instead of using a bias he used afoil, and that foil was thick enough tostop not only uranium alpha rays butalso the fission fragments; nor did hefind any long-range alpha rays, whichhad to be there if radium or actiniumisotopes were formed.

Then Hahn and Strassmann checkedthe chemical properties of this "ra-dium" with care and found that theywere identical with those of barium.

A propitious visitThis is where I came in because LiseMeitner was lonely in Sweden and, asher faithful nephew, I went to visit herat Christmas. There, in a small hotelin Kungalv near Goteborg I foundher at breakfast brooding over a letterfrom Hahn. I was skeptical about thecontents—that barium was formedfrom uranium by neutrons—but shekept on with it. We walked up anddown in the snow, I on skis and sheon foot (she said and proved that shecould get along just as fast that way),and gradually the idea took shape thatthis was no chipping or cracking ofthe nucleus but rather a process to beexplained by Bohr's idea that the nu-cleus was like a liquid drop; such adrop might elongate and divide it-self. Then I worked out the way theelectric charge of the nucleus would

diminish the surface tension and foundthat it would be down to zero aroundZ = 100 and probably quite small foruranium. Lise Meitner worked outthe energies that would be availablefrom the mass defect in such a break-up. She had the mass defect curvepretty well in her head. It turned outthat the electric repulsion of the frag-ments would give them about 200MeV of energy and that the mass de-fect would indeed deliver that energyso the process could take place on apurely classical basis without havingto invoke the crossing of a potentialbarrier, which of course could neverhave worked.

We only spent two or three days to-gether that Christmas. Then I wentback to Copenhagen and just managedto tell Bohr about the idea as he wascatching his boat to the US. I remem-ber how he struck his head after I hadbarely started to speak and said:"Oh, what fools we have been! Weought to have seen that before." Buthe had not—nobody had.

Lise Meitner and I composed apaper over the long-distance tele-phone between Copenhagen andStockholm. I told the whole story toGeorge Placzek, who was in Copenha-gen, before it even occurred to me todo an experiment. At first Placzekdid not believe the story that theseheavy nuclei, already known to sufferfrom alpha instability, should also besuffering from this extra affliction."It sounds a bit," he said,. "like theman who is run over by a motor carand whose autopsy shows that he hada fatal tumor and would have diedwithin a few days anyway." Then he

PHYSICS TODAY • NOVEMBER 1967 • 47

THE JOLIOT-CURIESartificial radioactivity.

discovered CENTRAL FIGURES in the discovery were Otto Hahn and Lise Meitner,here shown in front of the institute that bears their names

said, "Why don't you use a cloudchamber to test it?" I did not have acloud chamber handy and thought itwould be difficult anyway. But Iused an ionization chamber and it wasa very easy experiment to observe thelarge pulses caused by ion fragments.

I do not think chronology meansvery much and certainly cannot claimany particular intelligence or original-ity. I was just lucky to be with LiseMeitner when she received advancenotice of Hahn's and Strassmann's dis-covery. Then I had to be nudged be-fore I did the crucial experiment on 13January. By that time our joint paperwas nearly written. I held it back foranother three days to write up theother paper, and then they were bothsent to Nature on 16 January but pub-lished a week apart. In the first paperI used the word "fission" suggested tome by the American biologist, WilliamA. Arnold, whom I asked what onecalls the phenomenon of cell division.

The second paper also contained asuggestion from Lise Meitner that fis-sion fragments emerging from a bom-barded uranium layer could be collect-ed on a surface and their activitymeasured. The same thought inde-pendently occurred to Joliot, and hesuccessfully did this experiment on 26January. About that same time thenews reached the US; what happenedthen is discussed by Wheeler.

Serendipitous searches

To come back to my initial question:Why did it take so long before fissionwas recognized? Indeed, why wasn'tthe neutron found earlier? Ruther-ford thought about it and foretoldsome of its properties as early as hisBakerian lecture in 1920; but Joliotdid not read it, expecting a public lec-ture to contain nothing new! WhenCurie and Joliot found that the "beryl-lium radiation" ejected protons fromparaffin, they put it down to a kind ofCompton effect of a very hard gammaradiation (some 50 MeV), ignoringthe objections of theoretical physicists.The neutron was finally observed inCambridge, where such a particle wasexpected and had been sought.

At the time the neutron was foundin 1932 pulse amplifiers and ionizationchambers were available for a faciledetection of fission pulses. But thatwould have been too big a jump to ex-pect. The liquid-drop model of thenucleus was born late; the compound-nucleus idea was conceived by Bohronly late in 1936. It would have beena stroke of genius to think of fissionthen, and nobody did.

The discovery of artificial radioac-tivity in 1934 was again a chance dis-covery; no one had looked for it ex-cept Rutherford, who looked in vainfor alpha decay. And indeed the

Berkeley team turned a blind eyewhen their counters "misbehaved."After the discovery there was asheep-like rush to repeat the experi-ment with only the most obvious vari-ation (I was one of the sheep). OnlyFermi had the intelligence to strikeout in a different and tremendouslyfruitful direction.

But then Fermi got on the wrongtrack: He felt sure that uranium, likeother heavy nuclei, would obedientlyswallow any slow neutron that fell onit. He did make sure that the radioac-tive substances that were formed fromit were different from any of theknown elements near uranium. IdaNoddack, a German chemist, quiterightly pointed out that they might belighter elements; but her comments(published in a journal not much readby chemists and hardly at all by physi-cists) were regarded as mere pedant-ry. She did not indicate how suchlight elements could be formed; herpaper had probably no effect whatev-er on later work.

In the end it was good solid chemis-try that got things on the right track.Irene Curie and Pavel Savitch camevery close to it; only the presence oftwo substances with maliciously simi-lar properties prevented them from es-tablishing uranium fission beforeHahn and Strassman finally accom-plished it. D

48 • NOVEMBER 1967 • PHYSICS TODAY

Mechanism of Fissionby John A. Wheeler

IN EARLY JANUARY 1939 the Swedish-

American liner, MS Drottningholmcarried a short message across thestormy sea from Copenhagen to NewYork. This message symbolized thesteady transfer of nuclear discoveriesfrom Europe to the US that had beengoing on during the Hitler years.

Although these transfers were fate-ful for the US and the rest of theworld, the act of relaying this particu-lar message was simple: a few wordsspoken by Otto Frisch to Niels Bohron the pier in Copenhagen and a fewwords spoken to Enrico Fermi andme by Bohr on the pier in New York.As a junior participator in the eventsthat occurred then and in subsequentmonths, I shall relate the activities thatled to the publication of a Physical Re-view paper by Bohr and me. In thispaper we summarized the thoughtsexpressed in the message: the liquid-drop model that Frisch had applied tothe mechanism of fission and the de-terminations of packing fraction thatLise Meitner considered when arrivingat the first estimate of energy release infission.

No one looking at such a novelprocess at that time could fail to callon everything he knew about nuclearphysics to seek an interpretation. For-tunately the key ideas for unravelingthe puzzle had already been de-veloped. It may be appropriate torecall what had been learned aboutnuclear physics in the preceding halfa dozen years.

Clues to the answer1933 was a fruitful year for someonelike me, who was just earning his doc-tor's degree. It was the year of thediscovery of the neutron and WernerHeisenberg's great paper on the struc-ture of nuclei built out of neutrons andprotons. These discoveries made onefeel that he might soon know as muchabout the nucleus as he already knewabout the atom.

Encouraged by the vision that in-spired so many young men, me in-cluded, at that time, I spent 1933-34

working with Gregory Breit, to whoseinsights I owe so much. He and thegroup of which I soon found myself amember accepted almost unconsciouslythe model of the nucleus of that day:neutrons and protons moving in a com-mon self-consistent potential, closelyanalogous to the electric potential ofthe atom. "Unconscious" our accept-ance of the model was, yes; but alsoshadowy. None of us took it too liter-ally, especially not Breit, With his cau-tion and insight. Thus he was alwayswilling to consider alpha particles inthe nucleus as well as neutrons andprotons when that point of view madesense in considering a particular reac-tion. Breit also directed especial at-tention to areas of investigation asnearly free as possible of model-de-pendent issues. Thus much work wasdone on the penetration of chargedparticles into nuceli and how the crosssection for a nuclear reaction dependson energy. The analysis of scatteringprocesses in terms of phase shifts alsoreceived much attention.

With Breit's warm endorsement Ispent the following year at Niels Bohr'sinstitute in Copenhagen. Here I wasinitiated into the study of many newideas, but nothing was more impressivein nuclear physics than the messagethat M0ller brought back during thespring of 1935 from a short Easter visitto Rome: It told of Fermi's slow-neu-tron experiments and the astonishingresonances that he had discovered.Every estimate ever made before thenindicated that a particle passingthrough a nucleus would have an ex-tremely small probability of losing itsenergy by radiation and undergoingcapture if the current nuclear modelwas credible. Yet, directly in opposi-tion to the predictions of this model,Fermi's experiments displayed hugecross sections and resonances that werequite beyond explanation.

Of course a number of weeks wentby before the most significant resultsof this discovery could be sorted out.Everyone was actively concerned, butno one more so than Bohr, who paced

up and down in the colloquium andtook a central part in discussions.

Liquid dropsThe story of the development of theliquid-drop model and the compound-nucleus picture is a familiar one.What is not so clear and was certainlynot evident at the time is the distinc-tion between these ideas: (1) Thecompound-nucleus model shows, inessence, that the fate of a nucleus isindependent of the mechanism bywhich it has been formed, and (2) theliquid-drop model is, so to speak, aspecial case of the compound-nucleusmodel, a particular way of makingsuch a model of nuclear structure rea-sonable. Bohr proposed that the meanfree path of nucleon is short in rela-tion to nuclear dimensions instead ofbeing long, as assumed in all previousestimates. This new idea made some-thing like a liquid-drop model exceed-ingly attractive.

No one looking back on the situationfrom today's vantage point can fail tobe amazed at "the great accident ofnuclear physics"—the circumstance thatthe mean free path of particles in thenucleus is neither extremely short com-pared with nuclear dimensions (as as-sumed in the liquid-drop picture) norextremely long (as assumed in theearlier model) but of an intermediatevalue. Moreover, all the marvelousdetail of nuclear physics turns out todepend in such a critical way on thevalue of this parameter. As Aage Bohrand Ben Mottelson have taught us inrecent years, no one could have pre-dicted the precise one among manyalternative regimes in which thephenomenology would actually liefrom any advance estimate of themean free path. Only observationcould suffice! Knowing as little asone did in 1935 about the value ofthis decise parameter, still less aboutits cirticality, one had no option butto explore with all vigor the idea thatthe mean free path is very short.

The development of the liquid-dropmodel, which was applied to a varietyof processes, took place in the handsof Fritz Kalckar and Niels Bohr in1935-37. They applied it to a varietyof processes. At the center of everysuch application stood the idealizationof the compound nucleus, that is, theconcept that a nuclear reaction occurs

PHYSICS TODAY • NOVEMBER 1967 • 49

STROLLING THINKERS, Fermi (left) and Bohr, are well known fortheir important applications and expansions of early ideas of nuclear fission.

in two well separated stages: First,the particle arrives in the nucleus andimparts an excitation; then in someway the nucleus uses that energy forradiation, neutron or alpha-particleemission or any other competing pro-cess.

Bohr brings the news

The message that Frisch gave Bohras Bohr left Copenhagen opened up anew domain of application for thisconcept of the compound nucleus. Bythe time Bohr had arrived in New Yorkhe had already recognized that fissionis one more process in competition withneutron reemission and gamma-rayemission. Four days after his arrivalhe and Rosenfeld finished a paper sum-

marizing this general picture of fissionin terms of formation and breakup ofthe compound nucleus.

Rosenfeld had originally accom-panied Bohr to Princeton for severalmonths of work on the problem ofmeasurement in quantum electrody-namics. During Rosenfeld's Princetonsojourn Bohr gave less than half adozen lectures on that issue. Never-theless, that and many other questionsconspired to take much of his time.No one could go into his office withoutseeing the long list of duties and peoplehe had to give time to. That list madeit easy to appreciate the pleasure withwhich he came into my office to discussthe work that we had under way. Wewere trying to understand in detail the

mechanism of fission and, not least,analyze the barrier against fission andthe considerations that determine itsheight.

First of all, of course, we had toformulate the very idea of a thresholdor barrier. How can there even beany barrier according to the liquid-drop picture? Is not an ideal fluidinfinitely subdivisible? And thereforecannot the activation energy requiredto go from the original configuration toa pair of fragments be made as smallas one pleases? We obtained guidanceon this question out of the theory ofthe calculus of variations in the large,maxima and minima, and criticalpoints. This subject we absorbed byosmosis from our environment, sothoroughly charged over the years bythe ideas and results of Marston Morse.It became clear that we could find aconfiguration space to describe the de-formation of the nucleus. In this de-formation space we could find avariety of paths leading from the nor-mal, nearly spherical configuration overa barrier to a separated configuration.On each path the energy of deforma-tion reaches a highest value. Thispeak value differs from one path toanother. Among all these maxima theminimum measures the height of thesaddle point or fission threshold oractivation energy for fission.

While we were estimating barrierheights and the energy release in vari-ous modes of fission, the time came forthe fifth annual theoretical physics con-ference held in Washington on 26 Jan.Bohr felt a responsibility toward Frischand Meitner and thought that word oftheir work-in-progress and their con-cepts should not be released until theyhad the proper opportunity to publish,as is the custom throughout science.Even though this was the situation, atthe outset Rosenfeld did not appreciateall the complications and demands ofBohr's position. On the day of Bohr'sarrival in the US Rosenfeld went downto Princeton on the train. (Bohr hadan appointment later that day in NewYork.) Rosenfeld reported the newdiscovery at the journal club—the regu-lar Monday night journal club—and ofcourse everybody was very excited.Isidor I. Rabi, who was at the journalclub, carried the news back to Colum-bia, where John Dunning started toplan an experiment.

50 • NOVEMBER 1967 • PHYSICS TODAY

Nevertheless, even on 26 Jan., Bohrwas reluctant to speak about Frisch'sand Meitner's findings until he re-ceived word that they had actuallybeen published. Fortunately thatafternoon an issue of Die Naturwissen-schaften, which contained work byHarm and Fritz Strassmann, washanded to him; thus he could tell aboutit. Of course everybody started hisexperiments. The first direct physicalproof that fission takes place appearedin the newspapers of the twenty-ninth.

Shaping the theory

The analysis of fission led to the theoryof a liquid drop and this in turn ledback to a favorite love of Bohr, who,for his first student research work, ex-perimented on the instability of a jetof water against breakup into smallerdrops. He was quite familiar with thework of John W. Strutt, the third LordRayleigh. This work furnished a start-ing point for our analysis. However,we had to go to terms of higher orderthan Rayleigh's favorite second-ordercalculations to pass beyond the purelyparabolic part of the nuclear potential,that is, the part of the potential thatincreases quadratically with deforma-tion. We determined the third-orderterms to see the turning down of thepotential. They enabled us to evalu-ate the height of the barrier, or at leastthe height of the barrier for a nucleuswhose charge was sufficiently close tothe critical limit for immediate break-up.

Here we found that we could reducethe whole problem to finding a func-tion / of a single dimensionless variablex. This "fissility parameter" measuresthe ratio of the square of the charge tothe nuclear mass. This parameter hasthe value 1 for a nucleus that is alreadyunstable against fission in its sphericalform. For values of x close to 1, by thepower-series development mentionedabove one could estimate the height ofthe barrier and actually give quite adetailed calculation of the first twoterms in the power series for barrierheight, or /, in powers of (1 — *)•The opposite limiting case also lent it-self to analysis. In this limit the nu-cleus has such a small charge that thebarrier is governed almost entirely bysurface tension. The Coulomb forcesgive almost negligible assistance inpushing the material apart.

ROSENFELD, with Bohr, summarizedthe idea of fission.

Between this case (the power seriesabout x = 0) and the other case (thepower series about x = 1) there wasan enormous gap. We saw that itwould take a great amount of work tocalculate the properties of the fissionbarrier at points in between. Conse-quently we limited ourselves to inter-polation between these points. In the28 years since that time many workershave done an enormous amount ofcomputation on the topography of thedeformation energy as depicted overconfiguration space as a "base" for thetopographic plot. We are still far fromcompleting the analysis. Beautifulwork by Wladyslaw J. Swiatecki andhis collaborators at Berkeley has taughtus much more than we ever knew be-fore about the structure of this fissionbarrier and has revealed many unsus-pected features for values of x thatare remote from the two simple, origi-nal limits.

From fission barrier we turned tofission rate. All of us have alwaysrecognized that nuclear physics con-sists of two parts: (a) the energy of aprocess and (b) the rate at which theprocess will go on. The compound-nucleus model told us that the rateshould be measured by the partialwidth of the nuclear state in questionfor breakup by the specified process.

Toward a simpler theoryHow could we estimate this width?Happily, in earlier days, several per-sons in the Princeton community—among them Henry Eyring and EugeneWigner—had been occupied by thetheory of the rates of chemical reac-

tions. Also we derived some usefulinformation from cosmic-ray physics.Who does not recall the many detailedcalculations St0rmer and his associatesmade on the orbits of cosmic-ray par-ticles in the earth's magnetic field?Fortunately Manuel Sandoval Vallartaand later workers were able to sparethemselves almost all of these details.They had only to employ Liouville'stheorem. It said that the density ofsystems in phase space remains con-stant in time.

The same considerations of phasespace were equally useful for evaluat-ing the rate of fission. It turned outthat we could express the probabilityof going over the barrier as the ratio oftwo numbers. One of these numbers isrelated to the amount of phase spaceavailable in the transition-state con-figuration as the nucleus goes over thetop of the barrier. We were forced tothink of all the degrees of freedom ofthe nucleus other than the particularone leading to fission. All these otherdegrees of freedom are summarizedin effect in the internal excitations ofthe nucleus as it passes over the fissionbarrier. In classical terms this con-cept is well defined. It is a volumein phase space completely determinedby the amount of energy.

The other quantity, appearing in thedenominator of the rate-of-fission ex-pression, is linked with the volume ofphase space accessible to the com-pound system. In all the complexmotion short of actual passage over thebarrier the ensemble of systems underconsideration remains confined to thenarrow band of energies, AE, definedby the energy of the incident neutron.What counts is this energy intervalmultiplied with the rate of change ofvolume in phase space with energyfor the undissociated nucleus. Thebeauty of this derivation is the fact thatthese classical ideas lend themselvesto direct transcription into quantum-mechanical terms. Thus the Went-zel-Kramers-Brillouin approximationtaught us that volume in phase spacedetermines the number of energylevels. So we concluded that thewidth—the desired width measuringthe probability for fission—is given bya ratio in which the numerator is thenumber of states accessible to thetransition-state nucleus as it is goingover the barrier, that is, the number of

PHYSICS TODAY • NOVEMBER 1967 • 51

PLACZEK was helpful in formulatingtheories of fission.

states of excitation other than motionin the direction of fission. In the de-nominator appears the spacing be-tween nuclear energy levels, dividedby 2?r. Thus we had attained the mostdirect tie with experimentally interest-ing quantities. The formula that wasobtained in this way for the reactionrate, or the level width, applied to awide class of reactions as well as tofission, and was more general thanany that had previously been availablein reaction-rate theory. The newformula gave considerable insight intothe rate of passage over the fissionbarrier.

At this particular point it is inter-esting to note the caution with whichBohr adopted the formula. He wouldcome in every other day or so, and wewould go at it for perhaps a half a day,trying out first this approach and thenthat approach. But his supreme cau-tion was most evident when we wantedto interpret the number of levels ac-cessible in the transition state. Todaythat number is called "the number ofchannels," and we use it as a formulato describe the channel-analysis theoryof fission rate. Also we apply similarchannel-analysis considerations toother nuclear reactions. But at thattime the idea that each one of theseindividual channels has in principle adefinite experimentally observable sig-nificance was, for us, of dubious cer-tainty. Still less did we appreciate,until the later work of Aage Bohr, thepossibility that each individual chan-nel would have its individual angular

distribution from which one could de-termine the K values of that channel.The cautious phrase that was used inreference to that channel number ap-pears in the following quotation: "Itshould be remarked that the specificquantum-mechanical effects which setin at and below the critical fissionenergy may even show their influenceto a certain extent above this energyand produce slight oscillations in thebeginning of the yield curve, allowing,possibly, a direct determination of thenumber of channels." Of course weknow how later on in the 1950's thesevariations were observed by Lamphereand Green and others and how theyled to direct measurement of the chan-nel number.

Bohr's epiphany

The most important part of this Prince-ton period happened when I was notin direct touch with Bohr. One snowymorning he was walking from theNassau Club to his office in Fine Hall.As a consequence of a breakfast dis-cussion with George Placzek, who wasdeeply skeptical of these fission ideas,Bohr began struggling with the prob-lem of explaining the remarkable de-pendence of fission cross section onneutron energy. In the course of thewalk he concluded that slow-neutronfission is caused by U235 and fast-neutron fission by U238. By the timehe had arrived at Fine Hall and he andI had gathered together with Placzekand Rosenfeld, he was ready to sketchout the whole idea on the blackboard.There he displayed the concept thatU238 is not susceptible to division byneutrons of thermal energy, nor is itsusceptible to neutrons of intermediateenergy but only to neutrons with ener-gies of a million electron volts or more.Further, the fission observed at lowerenergies occurs because U235 is pres-ent and has a 1/v cross section forcapture. We already knew experi-mentally that neutrons of intermediateenergy undergo resonance capture.And, with the help of simple con-siderations, we could show that theresonance reaction of neutrons withuranium could not be due to U23r\ Weconcluded this because we knew thatthe resonance cross section would ex-ceed the theoretical limit given by thesquare of the wavelength if U235 wereresponsible for the resonance effect.

So the resonance had to be due toU238, and the very fact that the reso-nance neutrons did not bring aboutfission proved that U238 was not sus-ceptible to fission by neutrons of suchlow energy. Thus if it was not sus-septible at that energy, it would cer-tainly not be susceptible at lowerenergies; consequently low-energy fis-sion must be due to U235.

A few days later, on 16 April,Placzek, Wigner, Rosenfeld, Bohr, my-self and others discussed whether onecould ever hope to make a nuclearexplosive. It was so preposterous thento think of separating U235 that I can-not forget the words that Bohr usedin speaking about it: "It would takethe entire efforts of a country to makea bomb." He did not foresee that, intruth, the efforts of thousands ofworkers drawn from three countrieswould be needed to achieve that goal.

The theory of fission made it pos-sible to predict in general terms howthe cross section for fission would de-pend upon energy. In Palmer PhysicalLaboratory Rudolf Ladenberg, JamesKanner, Heinz H. Barschall and VanVoorhies, just at the time we wereworking on the theory, actually mea-sured the cross section of uranium inthe region from two million to threemillion volts—and also the cross sectionfor thorium, all of which fitted in withpredictions. The same considerationsof course made it possible to predictthat plutonium 239 would be fissile.For this application of the theory weare especially indebted to Louis A.Turner. One started on the way thatultimately led to the giant plutoniumproject having only this theoreticalestimate to light and encourage the firststeps.

Spontaneous fission offered a mostattractive application of these ideas inconjunction with the concept of barrierpenetration. Another application dealtwith the difference between promptneutrons and delayed neutrons. Inconclusion, nuclear fission brought us aprocess distinguished from all theother processes with which we everdealt before in nuclear physics, in thatwe have for the first time in fission anuclear transformation inescapablycollective in character. In this sensefission opened the door to the develop-ment of the collective model of thenucleus in the postwar years. 0

52 • NOVEMBER 1967 • PHYSICS TODAY