The idea of the neutrinoLaurie M. Brown Citation: Phys. Today 31(9), 23 (1978); doi: 10.1063/1.2995181 View online: http://dx.doi.org/10.1063/1.2995181 View Table of Contents: http://www.physicstoday.org/resource/1/PHTOAD/v31/i9 Published by the American Institute of Physics. Additional resources for Physics TodayHomepage: http://www.physicstoday.org/ Information: http://www.physicstoday.org/about_us Daily Edition: http://www.physicstoday.org/daily_edition

Downloaded 21 Sep 2012 to 141.108.40.129. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms

The idea of the neutrinoTo avoid anomalies of spin and statistics Pauli suggested in 1930

that a neutral particle of small mass might accompany the electron in nuclearbeta decay, calling it (until Chadwick's discovery) the neutron.

Laurie M. Brown

During the 1920's physicists came to ac-cept the view that matter is built of onlytwo kinds of elementary particles, elec-trons and protons, which they oftencalled1 "negative and positive electrons."A neutral atom of mass number A andatomic number Z was supposed to containA protons, all in the nucleus, and A neg-ative electrons, A — Z in the nucleus andthe rest making up the external electronshells of the atom. Their belief that bothprotons and negative electrons were to befound in the nucleus arose from the ob-servations that protons could be knockedout of light elements by alpha-particlebombardment, while electrons emergedspontaneously (mostly from very heavynuclei) in radioactive beta decay. Anyother elementary constituent of the atomwould have been considered superfluous,and to imagine that another might existwas abhorrent to the prevailing naturalphilosophy.

Nevertheless, in December 1930Wolfgang Pauli suggested a new elemen-tary particle that he called a neutron,with characteristics partly like that of thenucleon we now call by that name, andpartly those of the lepton that we now callneutrino (more precisely the electronantineutrino, but this distinction is notneeded here). Pauli's neutron-neutrinoidea became well-known to physicistseven before his first publication of it,which is in the discussion section fol-lowing Heisenberg's report on nuclearstructure at the Seventh Solvay Confer-ence,2 held in Brussels in October 1933.

Shortly after attending this conference,Enrico Fermi published his theory of betadecay, which assumes that a neutrino al-ways accompanies the beta-decay elec-

Laurie M. Brown is a professor in the Departmentof Physics and Astronomy, Northwestern Uni-versity, Evanston, Illinois.

tron, and that both are created at theirmoment of emission. Perhaps because ofthe rapid acceptance of Fermi's theoryand the tendency to rethink history "as itshould have happened," the true natureof Pauli's proposal has been partly over-looked and its radical character insuffi-ciently emphasized. Contrary to theimpression given by most accounts, Pau-li's "neutron" has some properties incommon with the neutron James Chad-wick discovered in 1932 as well as withFermi's neutrino.

Flaws in the modelBy the end of 1930, when our story be-

gins, quantum mechanics had triumphednot only in atomic, molecular and crystalphysics, but also in its treatment of somenuclear processes, such as alpha-particleradioactivity and scattering of alphaparticles from nuclei (including the caseof helium, in which quantum-mechanicalinterference effects are so important).However, the situation regarding elec-trons in the nucleus was felt to be critical.The main difficulties of the electron-proton model of the nucleus were:• The symmetry character of the nuclearwave function depends upon A, not Z aspredicted by the model; when A — Z isodd the spin and statistics of the nucleusare given incorrectly. For example, ni-trogen (Z = 7, A = 14) was known fromthe molecular band spectrum of N2 tohave spin 1 and Bose-Einstein statis-tics.• No potential well is deep enough andnarrow enough to confine a particle aslight as an electron to a region the size ofthe nucleus (the argument for this isbased on the uncertainty principle andrelativistic electron theory).• It is hard to see how to "suppress" thevery large (on the nuclear scale) magneticmoments of the electrons in the nucleus,

which conflict with data on the hyperfinestructure of atomic spectra.• Although both alpha and gamma decayshow the existence of narrow nuclear en-ergy levels, the electrons from a givenbeta-decay transition emerge with a broadcontinuous spectrum of energy.

The strong contrast between the suc-cesses and the failures of quantum me-chanics applied to the nucleus are no-where more evident than in a book byGeorge Gamow.3 In it, all the passagesconcerning electrons in the nucleus are setoff in warning symbols (skull and cross-bones in the original manuscript).

Some physicists (among them NielsBohr and Werner Heisenberg4) took thesedifficulties to indicate that a new dy-namics, possibly even a new type ofspace-time description, might be appro-priate on the scale of nuclear distancesand energies, just as quantum mechanicsbegins to be important on the atomicscale. These physicists were impressedby the similarity of the nuclear radius tothe value e'z/mc2, the classical electronradius of H. A. Lorentz. At this distanceit had been anticipated that electrody-namics would probably fail (and maybe,with it, the special theory of relativity).Bohr was willing to relinquish the con-servation of energy, except as a statisticallaw, in parallel with the second law ofthermodynamics. At the same timeHeisenberg was considering the intro-duction of a new fundamental length intothe theory. It seemed that anythingmight be considered acceptable as a wayout of the dilemma—or perhaps anythingexcept a new elementary particle.

Pauli's proposal

It was in this context of ideas that Paulidared to suggest the existence of a newneutral particle. His proposal, intendedto rescue the quantum theory of the nu-

0031-9228/78/3109-023$0.50 © 1978 American Institute of Physics PHYSICS TODAY / SEPTEMBER 1978 2 3

Downloaded 21 Sep 2012 to 141.108.40.129. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms

GOUDSMIT COLLECTION, AIP NIELS BOHR LIBRARY

PAULI ON THE WAY TO PASADENA, 1931

cleus from its contradictions, was pre-sented in good humor as a "desperateremedy," although it was a serious one.(The Viennese version would have been,according to the old joke: desperate, butnot serious.) During the next three yearshe lectured on what he called the "neu-tron" at several physics meetings and hediscussed it privately with colleagues.

Pauli's first proposal was put forwardonly tentatively, as he recalled in a lecturehe delivered in Zurich in 1957, after re-ceiving news of the experiments con-firming parity violation in beta decay.5

Invited to a physics meeting in Tubingen,Germany, which he was unable to attend(because of a ball to be held in Zurich, atwhich he declared he was "indispens-able"), he sent a message with a colleagueas an "open letter," although it was in-tended mainly for Hans Geiger and LiseMeitner. An English translation of thisletter is given in the Box on page 27.Pauli was anxious for their expert adviceas to whether his proposal was compatiblewith the known facts of beta decay.

In the 1957 lecture Pauli also tells howhe became convinced of a crisis associatedwith beta decay. During the decade thatfollowed the discovery by Chadwick in1914 of beta rays with a continuous energyspectrum, it became established thatthese were the true "disintegration elec-trons," rather than those making up dis-crete electron line spectra, which werelater shown to arise from such causes asphotoelectric effects of nuclear gammarays, internal conversion and Auger pro-cesses. Because a continuous spectrumseemed to disagree with the presence of

discrete quantum states of the nucleus (asindicated by alpha and gamma emission),some workers, including Meitner, thoughtthat the beta rays were radiating some oftheir energy as they emerged through thestrong electric field of the nucleus.5'67

This led C. D. Ellis and WilliamWooster at the Cavendish Laboratory inCambridge, England, who did not believein the radiation theory, to perform a ca-lorimetric experiment with radium E(bismuth) as a source. Their result, laterconfirmed in an improved experiment byMeitner and W. Orthmann,8 was that theenergy per beta decay absorbed in athick-walled calorimeter was equal to themean of the electron energy spectrum,and not to its maximum (endpoint).Furthermore, Meitner showed that nogamma rays were involved. According toPauli (in 1957), this allowed but two pos-sible theoretical interpretations:• The conservation of energy is valid onlystatistically for the interaction that givesrise to beta radioactivity.• The energy theorem holds strictly ineach individual primary process, but atthe same time there is emitted with theelectron another very penetrating radia-tion, consisting of new neutral particles.To the above, Pauli adds, "The first pos-sibility was advocated by Bohr, the secondby me." 5

But although the conservation of en-ergy, and possibly other conservation lawsin beta decay were very much in Pauli'smind at this time, this was not his onlyreason for proposing the neutrino. Hemakes this point (already obvious fromhis Tubingen letter) quite explicit in his

1957 Zurich lecture. After pointing outone of the major difficulties with the nu-clear model containing only protons andelectrons (the symmetry argument men-tioned above), Pauli says:

"I tried to connect this problem of thespin and statistics of the nucleus with theother of the continuous beta spectrum,without giving up the energy theorem,through the idea of a new neutral parti-cle."

Neutrinos—ejected or created?

It is often overlooked in discussing thehistory of the neutrino idea that Paulisuggested his particle as a constituent ofthe nucleus, with a small but not zeromass, together with the protons and theelectrons. (Chien-Shiung Wu, for ex-ample, emphasizes the non-conservationof statistics that would occur in beta decaywithout the neutrino.67'9 However, Paulirefers rather to the spin and statistics ofstable nuclei such as lithium 6 and nitro-gen 14.) This point is of some signifi-cance; had Pauli proposed in 1930 thatneutrinos were created (like photons) intransitions between nuclear states, andthat they were otherwise not present inthe nucleus, he would have anticipated bythree years an important feature of Fer-mi's theory of beta decay. Pauli did notclaim to have had this idea when he wrotethe Tubingen letter, but he did say (in hisZurich lecture) that by the time he wasready to speak openly of his new particle,at a meeting of The American PhysicalSociety in Pasadena, held in June of 1931,he no longer considered his neutrons to benuclear constituents. It is for this reason,he says, that he no longer referred to themas "neutrons"; indeed, that he made useof no special name for them. However,there is evidence, as we shall see, thatPauli's recollections are incorrect; that atPasadena the particles were called neu-trons and were regarded as constituentsof the nucleus.

I have not been able to obtain a copy ofPauli's Pasadena talk or scientific noteson it; he said later that he was unsure ofthe matter and thus did not allow hislecture to be printed. The press, how-ever, took notice. For example, a shortnote in Time, 29 June 1931, headed"Neutrons?", says that Pauli wants to adda fourth to the "three unresolvable basicunits of the universe" (proton, electronand photon); adding, "He calls it theneutron."

Upon examining the program of thePasadena Meeting, I discovered thatSamuel Goudsmit spoke at the same ses-sion as Pauli (and even upon the sameannounced subject—hyperfine structure).I wrote to Goudsmit and received a mostinteresting reply, from which I should liketo quote:

"Pauli accompanied my former wifeand me on the train trip across theUS. I forgot whether we started inAnn Arbor or arranged to meet in Chi-

24 PHYSICS TODAY / SEPTEMBER 1978

Downloaded 21 Sep 2012 to 141.108.40.129. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms

cago. We talked little physics, moreabout physicists. Pauli's main topicat the time was that he could imitateP. S. Epstein and he insisted that Itake pictures of him while doing that.We spent a couple of days in SanFrancisco, where we almost lost him inChinatown. He'd suddenly rushahead and around a corner while wewere window shopping . . . He mayhave talked about the "neutron" onthat trip, but I am not at all certain

Goudsmit does not now recall exactlywhat Pauli said at Pasadena, except thathe mentioned the "neutron"; however, hesent me a copy of his report at the RomeCongress on what Pauli had said fourmonths earlier in Pasadena. To continue,then, with Goudsmit's letter:

"Fermi was arranging what was prob-ably the first nuclear physics meeting.It was held in Rome in October 1931. . . It was the best organized meeting Iever attended, because there was verymuch time available for informal dis-cussions and get-togethers . . . Fermihad arranged marvelous leisurelysightseeing trips for the group. Therewere about 40 guests and 10 Italians.

"Fermi ordered the then 'young'participants, namely [Nevill] Mott,[Bruno] Rossi, [George] Gamow (whocould not leave Russia but sent a man-uscript) and myself, to prepare sum-mary papers for discussion . . . As youknow, I don't use and don't keepnotes. But I have a clear picture ofPauli lecturing [at Pasadena] and hismention of the 'neutron'. . . Pauli wassupposed to attend the Rome meet-ing, but he arrived a day or so late. Infact, he entered the lecture hall thevery moment that I mentioned hisname! Like magic! I remarkedabout it and got a big laugh from theaudience."

Goudsmit's Rome report

At Fermi's request, then, Goudsmitreported at the Rome Conference onPauli's talk in Pasadena. Here is what hesaid:10

"At a meeting in Pasadena in June1931, Pauli expressed the idea thatthere might exist a third type of ele-mentary particles besides protons andelectrons, namely 'neutrons.' Theseneutrons should have an angular mo-mentum V2 h/2-K and also a magneticmoment, but no charge. They arekept in the nucleus by magnetic forcesand are emitted together with beta-rays in radioactive disintegration.This, according to Pauli, might re-move present difficulties in nuclearstructure and at the same time in theexplanation of the beta-ray spectrum,in which it seems that the law of con-servation of energy is not fulfilled. Ifone would find experimentally thatthere is also no conservation of mo-

GOUDSMIT COLLECTION. AIP NIELS BOHR LIBRARY

GOUDSMIT (MIDDLE) AND FERMI (RIGHT) WITH UNIDENTIFIED MAN

mentum, it would make it very proba-ble that another particle is emitted atthe same time with the beta-particle.The mass of these neutrons has to bevery much smaller than that of theproton, otherwise one would have de-tected the change in atomic weightafter beta-emission."

Goudsmit added that Pauli believed"neutrons may throw some light on thenature of cosmic rays."

It does appear clear from this passage(to which Pauli evidently made no objec-tion at the time) that at Pasadena theneutron was intended to be a particle thatcould be bound in the nucleus by mag-netic forces. In his letter to me Goudsmitalso said, "It was Maurice Goldhaber whosome time ago pointed out that I was thefirst to put Pauli's idea on paper and inprint."

After leaving Pasadena Pauli remainedin the United States until the fall, whenhe went to Rome. He gave a seminar atthe Summer Session of the University ofMichigan at Ann Arbor (probably at oneof their Symposia on Theoretical Physics,where Fermi had, the previous summer,given his famous lectures on the quantumtheory of radiation). At the seminarPauli spoke, according to the Berkeleytheorists J. F. Carlson and J. Robert Op-penheimer,11 about "the elements of thetheory of the neutron, its functions and itsproperties."

Tracks in the cloud chamber

Carlson and Oppenheimer wonderedwhether Pauli's "neutrons" could be usedto solve yet another puzzle: the appear-

ance of certain lightly ionizing cloud-chamber tracks from cosmic rays that hadbeen reported.

The complex problem of the energy lossof relativistic charged particles was crucialto the interpretation of the various com-ponents of the cosmic rays observed in theatmosphere, and had attracted the at-tention of many theorists. Carlson andOppenheimer were unable to account forcloud-chamber tracks that appearedthinner than those of an "ordinary ra-dioactive" beta particle. Their calcula-tions of energy loss (which agreed in ageneral way with independent calcula-tions by Heisenberg and Hans Bethe, andwith an older classical estimate by Bohr)showed that charged particles should havea relativistic increase of ionization withenergy. The particles leaving light trackswere very penetrating (and thus probablyrelativistic) and it was concluded thatthey could not be electrons or protons.(These quarklike tracks have not, to myknowledge, been explained. Perhapsthey were examples of old and "faded"tracks, which often plagued cloud cham-bers of the untriggered variety.)

Carlson and Oppenheimer decidedtherefore to make a theoretical investi-gation, as they said,11 of the "ionizingpower of the neutrons which were sug-gested by Pauli to salvage the theory ofthe nucleus. These neutrons, it will beremembered, are particles of finite propermass, carrying no charge, but having asmall magnetic moment. . ."

Could thin tracks, like those in thecosmic rays, be seen from beta decays?

"If they were found, we should be cer-

PHYSICS TODAY / SEPTEMBER 1978 25Downloaded 21 Sep 2012 to 141.108.40.129. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms

AIP NIELS BOHR LIBRARY

The participants in the Seventh Solvay Conference, where Pauli pre-sented his neutrino idea, included, in the first row, E. Schrodinger, I. Joliot,N. Bohr, A. Joffe, M. Curie, 0. W. Richardson, P. Langevin. Lord Ruther-ford, T. De Donder, M. de Broglie, L. de Broglie, L. Meitner, J. Chadwick,and in the second row, E. Henriot, F. Perrin, F. Joliot, W. Heisenberg, H.

A. Kramers, E. Stahel, E. Fermi, E. T. S. Walton, P. A. M. Dirac, P. Debye,N. F. Mott, B. Cabrera, G. Gamow, W. Bothe, P. M. S. Blackett, M. S.Rosenblum, J. Errera, E. Bauer, W. Pauli, J. E. Verschaffelt, M. Cosyns(in back), E. Herzen, J. D. Cockcroft, C. D. Ellis, R. Peierls, A. Piccard,E. 0. Lawrence, L. Rosenfeld. The photograph is by Benjamin Couprie.

tain that the neutrons not only playeda part in the building of nuclei, butthat they also formed the cosmicrays."The calculations of Carlson and Op-

penheimer were published12 almost a yearlater, in September 1932; by that timethey no longer believed that "neutrons"might leave observable cloud-chambertracks. In addition, the situation in nu-clear physics had changed profoundly, asit also was about to in cosmic-ray physics:Chadwick's study of "the penetratingradiation produced in the artificial dis-integration of beryllium" had revealed theexistence of the neutron, announced theprevious February; Anderson's discoveryof the positron in cosmic rays was an-nounced in August.13 Certainly onecould no longer speak of the proton assynonymous with positive electricity, andone might suppose that now a new parti-cle like Pauli's would be acceptable; butthis was not the case:

For one thing, the positron was thoughtto be only the absence of a negative elec-tron of negative energy, a hole in thevacuum. For another, the neutron ofChadwick, the heavy neutron, was gen-erally regarded as a composite object (itwas not thought to be unstable when free),a kind of tightly bound hydrogen atom orneutral nucleus made of a proton and anelectron, like other nuclei. It was perhapsthought to be elementary only by EttoreMajorana in Rome who (according toEmilio Segre) called it the neutral pro-ton.

For our purposes, the Carlson-Op-penheimer article is significant in what ittells us about the view held by Pauli, inthat summer of 1931, about his neutralparticle, which, following the Berkeleyauthors, we will now call the magnetic

neutron to distinguish it from Chadwick'sneutron and Fermi's neutrino. Carlsonand Oppenheimer state that the neutralparticle of spin lli, satisfying the exclusionprinciple, was introduced by Pauli notonly to resolve the difficulties in nucleartheory, but "on the further ground thatsuch a particle could be described by awave function which satisfies all the re-quirements of quantum mechanics andrelativity . . . The experimental evidenceon the penetrating beryllium radiationsuggests that neutrons of nearly protonicmass do exist; and since our calculationsmay be carried through without specify-ing the mass or magnetic moment of theneutron, we shall consider the most gen-eral particle which satisfies the waveequation proposed by Pauli. It is im-portant to observe that there may verywell be other types of neutral particles,which are not elementary, and to whichour calculations do not apply . . . "

Thus we find, surprisingly, that therewere thought to be also purely theoreticalgrounds for considering a neutral particlewith a magnetic moment; it is one of thefew simple types of elementary particlesthat are allowed by relativistic quantumtheory. In the wake of Chadwick's neu-tron discovery, Carlson and Oppenheimerin 1932 redefined Pauli's particle to be onewhose wave function obeys a certain rel-ativistic wave equation. We should not,however, assume that the Berkeley theo-rists were soft on new particles. On thecontrary, the final paragraph of theirlengthy article reads, "We believe thatthese computations show that there is noexperimental evidence for the existenceof a particle like the magnetic neutron."

Pauli's wave equation for the neutralparticle, given at Ann Arbor, is a variantof the linear Dirac equation for the elec-

tron, containing an additional term (Zu-satzglied) called the "Pauli anomalousmagnetic moment" term.14 This equa-tion describes a spin-1^ particle that maybe either charged or neutral; the extraterm makes a contribution to thecharge-current four-vector, which neednot vanish for a neutral particle.

Fermi is positive

Carlson and Oppenheimer derived ageneral formula for the collision crosssection of magnetic neutrons and exam-ined the result for small velocities. (Theywere well aware of the perils involved inpushing this highly singular interaction toexcessive energies.) For the collision ofa neutron against a particle of equal mass,they found a large probability, nearly in-dependent of velocity and proportional tothe square of the magnetic moment. Theaverage energy loss per collision was rel-atively large, and they deduced that sucha particle "will never produce ion tracesin a cloud chamber, since it tends to losean appreciable fraction of its energy, andsuffer an appreciable deflection at everyimpact." For targets much lighter orheavier than the neutron, smaller energylosses occur; cloud-chamber tracks mightresult in this case, but the collisionprobabilities are small unless the mag-netic moment of the neutron is assumedto be improbably large. The concluded(correctly) that there is no evidence formagnetic neutrons. (The heavy neutron,with a magnetic moment only one thou-sandth of a Bohr magneton, leaves notracks.) At the Seventh Solvay Confer-ence in 1933, Pauli no longer felt themagnetic neutron to be "well-founded."

Let us return now to the Rome Con-gress of 1931, which Pauli consideredimportant in the development of the

26 PHYSICS TODAY / SEPTEMBER 1978Downloaded 21 Sep 2012 to 141.108.40.129. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms

neutrino concept, for there he had theopportunity to discuss it with Bohr andespecially with Fermi, with whom he hada number of private conversations. WhileFermi's attitude toward the neutrino wasvery positive, Bohr was totally opposed toit, preferring to think that within nucleardistances the conservation laws werebreaking down.15

"From the empirical point of view,"said Pauli, "it appeared to me decisivewhether the beta spectrum of the elec-trons showed a sharp upper limit" or, in-stead, an infinitely falling statistical dis-tribution. Pauli felt that if the limit weresharp, then his idea was correct, andBohr's was wrong.

In mid-1933, Ellis and Mott suggestedthat the beta-ray spectrum has indeed asharp upper limit, corresponding to aunique energy difference between parentand daughter nucleus.16 Furthermore,they added,

"According to our assumption the j3-particle may be expelled with less en-ergy than the difference of the ener-gies . . . of the two nuclei, but not withmore energy. We do not wish in thispaper to dwell on what happens to theexcess energy in those disintegrationsin which the electron is emitted withless than the maximum energy. Wemay, however, point out that if the en-ergy merely disappears, implying abreakdown of the principle of energyconservation, then in a 0-ray decayenergy is not even statistically con-served. Our hypothesis is, of course,also consistent with the suggestion ofPauli that the excess energy is carriedoff by particles of great penetratingpower such as neutrons of electronicmass."The question of the upper limit of the

beta spectrum, although not easily re-solved, is of some importance, for theshape of the upper end of the spectrum issensitive to the neutrino mass. This wasdiscussed again by Ellis at an interna-tional conference in London, held in thefall of 1934, where he referred to accuratemagnetic spectrograph measurements ofW. J. Henderson that strongly suggesteda neutrino of zero mass.17 Fermi's theoryof beta decay had already been pub-lished,18 and Ellis assumed it in his anal-ysis, but an energy-nonconserving theory,that of Guido Beck and Kurt Sitte, sharedequal time with Fermi's at the confer-ence.

Fermi spoke at the London Conference,but his subject was the neutron-activationwork of the Rome experimental nuclearphysics group. He also had attended theSeventh Solvay Conference, held in Oc-tober, 1933, where he heard Pauli presenthis first suggestion for publication of theexistence of a neutrino. The completeSolvay remarks of Pauli are given in En-glish translation in the Box on page 28; weleave it to the reader to decide whetherPauli still thought that the neutrino or the

Pauli proposes a particleThe letter in which Pauli proposed the neutrino, translated from the German of reference 5,reads as follows:

Zurich, 4 December 1930Gloriastr.

Physical Institute of theFederal Institute of Technology (ETH)

ZurichDear radioactive ladies and gentlemen,

As the bearer of these lines, to whom I askyou to listen graciously, will explain moreexactly, considering the "false" statistics ofN-14 and Li-6 nuclei, as well as the contin-uous /3-spectrum, I have hit upon a desperateremedy to save the "exchange theorem" *of statistics and the energy theorem.Namely [there is] the possibility that therecould exist in the nuclei electrically neutralparticles that I wish to call neutrons, whichhave spin 1/2 and obey the exclusion princi-ple, and additionally differ from light quantain that they do not travel with the velocity oflight: The mass of the neutron must be ofthe same order of magnitude as the electronmass and, in any case, not larger than 0.01proton mass.—The continuous /3-spectrumwould then become understandable by theassumption that in /3 decay a neutron isemitted together with the electron, in sucha way that the sum of the energies of neutronand electron is constant.

Now the next question is what forces actupon the neutrons. The most likely modelfor the neutron seems to me to be, on wavemechanical grounds (more details are knownby the bearer of these lines), that the neutron

at rest is a magnetic dipole of a certain mo-ment fi. Experiment probably requires thatthe ionizing effect of such a neutron shouldnot be larger than that of a 7 ray, and thus /ushould probably not be larger than e.10~13

cm.But I don't feel secure enough to publish

anything about this idea, so I first turn con-fidently to you, dear radioactives, with thequestion as to the situation concerning ex-perimental proof of such a neutron, if it hassomething like about 10 times the pene-trating capacity of a 7 ray.

I admit that my remedy may appear tohave a small a priori probability becauseneutrons, if they exist, would probably havelong ago been seen. However, only thosewho wager can win, and the seriousness ofthe situation of the continuous /3-spectrumcan be made clear by the saying of my hon-ored predecessor in office, Mr. Debye, whotold me a short while ago in Brussels, "Onedoes best not to think about that at all, likethe new taxes." Thus one should earnestlydiscuss every way of salvation.—So, dearradioactives, put it to the test and set itright.—Unfortunately I cannot personallyappear in Tubingen, since I am indispensablehere on account of a ball taking place inZurich in the night from 6 to 7 of Decem-ber.—With many greetings to you, also toMr. Back, your devoted servant,

W. Pauli

* In the 1957 lecture, Pauli explains, "This reads: exclusion principle (Fermi statistics)and half-integer spin for an odd number of particles; Bose statistics and integer spin for aneven number of particles."

CHADWICK

PHYSICS TODAY / SEPTEMBER 1978 27Downloaded 21 Sep 2012 to 141.108.40.129. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms

electron were constituents of the nucleus.(That a massless neutrino could becreated at the moment of its emissionwith the electron was clearly proposedthat year19 by Francis Perrin, who alsoattended the Seventh Solvay Conference.)There was, in any case, no doubt that alight or massless neutral particle of spinV2 has to be emitted with the beta-decayelectron in order to save the conservationlaws, and that is surely the idea of neu-trino!

Fermi's theory of beta decay is in manyways still the standard theory. Called byVictor Weisskopf "the first example ofmodern field theory,"20 it eventuallycaused Bohr to withdraw21 his doubtsconcerning "the strict validity of theconservation laws." A radical generali-zation of quantum theory was not re-quired, though new particles and new in-teractions were. Within a few months ofFermi's theory, positron beta decay wasseen (the first example of artificial ra-dioactivity); and beta decay was to be theprototype of a larger class of weak inter-actions.

The neutrino can be regarded as one ofthe first (if not the first) of the new par-ticles that made the new physics of the1930's, even though it took two moredecades to observe the first neutrino-capture event. The weak interactionshave been notorious for their capacity toflout the expectations of physicists withregard to symmetries and conservationlaws. Although Bohr was too willing, inhis 1931 Faraday Lecture,15 "to renouncethe very idea of energy balance," theconclusion of that lecture is probably stillappropriate today: " . . . notwithstandingall the recent progress, we must still beprepared for new surprises."

This work was supported in part by a grantfrom the National Science Foundation. Iwould like to express my sincere appreciationto Arthur L. Norberg of The Bancroft Library,University of California, Berkeley, and toJudith Goodstein of the Robert A. MillikanMemorial Library of the California Instituteof Technology. I am much obliged to SamuelGoudsmit for his letter and for his kind per-mission to quote from it.

References

1. R. A. Millikan, in Encyclopedia Britan-nica, 14th edition, volume 8, page 340(1929).

2. Rapports du Septieme Conseil de Phy-sique Solvay, 1933, Gauthier-Villars, Paris(1934), page 324. Pauli's remarks are inFrench.

3. G. Gamow, Constitution of Atomic Nucleiand Radioactivity, Oxford U.P. (1931).

4. J. Bromberg, Hist. Stud. Phys. Sci. 3, 307(1971).

5. W. Pauli, Aufsdtze und Vortrdge u'berPhysik und Erkenntnistheorie, Braun-schweig (1961); Collected Scientific Pa-pers, volume 2, Interscience (1964), page1313.

Pauli becomes bolderThe discussion comments in which Pauli presented the idea of the neutrino at the SeventhSolvay Conference, ref. 2. The text is based on the translation from the French originalby Chien-Shiung Wu, ref. 9, with corrections by Laurie Brown noted in brackets.

The difficulty coming from the existence ofthe continuous spectrum of the 0-rays con-sists, as one knows, in that the mean life-times of nuclei emitting these rays, as thatof the resulting radioactive bodies, possesswell-determined values. One concludesnecessarily from this that the state as well asthe energy and the mass, of the nucleuswhich remains after the expulsion of the /3particle, are also well-determined. I will notpersist in efforts by which one could try toescape from this conclusion for I believe, inagreement with Bohr, that one alwaysstumbles upon insurmountable difficulties inexplaining the experimental facts.

In this connection, two interpretations ofthe experiment present themselves. Theinterpretation supported by Bohr admits thatthe laws of conservation of energy and mo-mentum do not hold when one deals with anuclear process where light particles play anessential part. This hypothesis does notseem to me either satisfying or even plau-sible. In the first place the electric chargeis conserved in the process, and I don't seewhy conservation of charge would be morefundamental than conservation of energy andmomentum. Moreover, it is precisely theenergy relations which govern severalcharacteristic properties of beta spectra(existence of an upper limit and relation withgamma spectra, Heisenberg stability crite-rion). If the conservation laws were notvalid, one would have to conclude from theserelations that a beta disintegration occursalways with a loss of energy and never again; this conclusion implies an irreversibilityof these processes with respect to time,which doesn't seem to me at all accept-able.

In June 1931, during a conference inPasadena, I proposed the following inter-pretation: the conservation laws hold, theemission of beta particles occurring togetherwith the emission of a very penetrating ra-

diation of neutral particles, which has notbeen observed yet. The sum of the energiesof the beta particle and the neutral particle(or the neutral particles, since one doesn'tknow whether there be one or many) emittedby the nucleus in one process, will be equalto the energy which corresponds to the upperlimit of the beta spectrum. It is obvious thatwe assume not only energy conservation butalso the conservation of linear momentum,of angular momentum and of the character-istics of the statistics in all elementary pro-cesses.

With.regard to the properties of theseneutral particles, we first learn from atomicweights [of radioactive elements] that theirmass cannot be much larger than that of theelectron. In order to distinguish them fromthe heavy neutrons, E. Fermi proposed thename "neutrino." It is possible that theneutrino proper mass be equal to zero, sothat it would have to propagate with the ve-locity of light, like photons. Nevertheless,their penetrating power would be far greaterthan that of photons with the same energy.It seems to me admissible that neutrinospossess a spin 1/2 and that they obey Fermistatistics, in spite of the fact that experimentsdo not provide us with any direct proof of thishypothesis. We don't know anything aboutthe interaction of neutrinos with other ma-terial particles and with photons: the hy-pothesis that they possess a magnetic mo-ment, as I had proposed once (Dirac's theoryinduces us to predict the possibility of neutralmagnetic particles) doesn't seem to me at allwell founded.

In this connection, the experimental studyof the momentum difference [read balance]in beta disintegrations constitutes an ex-tremely important problem; one can predictthat the difficulties will be quite insur-mountable [read very great] because of thesmallness of the energy of the recoil nucle-

6. C. S. Wu, S. A. Moszkowki, Beta Decay,Interscience (1966).

7. C. S. Wu, in Trends in Atomic Physics (0.R. Frisch et al, eds.), Interscience (1959),page 45; C. S. Wu, in Five Decades of WeakInteractions (N. P. Chang, ed.), New YorkAcad. Sciences, New York (1977), page 37;A. Pais, Rev. Mod. Phys. 49, 925 (1977).

8. C. D. Ellis, W. A. Wooster, Proc. Roy. Soc.(London) A 117,109 (1927); L. Meitner, W.Orthmann, Zeit. f. Phys. 60, 413 (1930).

9. C. S. Wu, in Theoretical Physics in theTwentieth Century (H. Fierz, V. F. Weis-skopf, eds.), Interscience (1960), page249.

10. S. A. Goudsmit, in Convegno di FisicaNucleare, Reale Accademia d' Italia, Atti,Rome (1932), page 41.

11. J. F. Carlson, J. R. Oppenheimer, Phys.Rev. 38, 1737 (1931).

12. J. F. Carlson, J. R. Oppenheimer, Phys.Rev. 41,763(1932).

13. C. Weiner, PHYSICS TODAY, May 1972,page 40.

14. W. Pauli, in Handbuch der Physik, Band24/1 (1933), page 233; ref. 12, page 778.

15. N. Bohr, in ref. 10, page 119; J. Chem. Soc.(London), page 349 (1932).

16. C. D. Ellis, N. F. Mott, Proc. Roy. Soc.(London), A 141, 502 (1933).

17. C. D. Ellis, in International Conference onPhysics, London, 1934, Vol. I, NuclearPhysics, Cambridge (1935); W. J. Hen-derson, Proc. Roy. Soc. (London) A 147,572 (1934).

18. E. Fermi, Z. f. Phys. 88,161 (1934); Englishtranslation: F. L. Wilson, Amer. J. Phys.36,1150(1968).

19. F.Perrin.Compt. Rend. 197,1625 (1933).

20. V. F. Weisskopf, in Exploring the Historyof Nuclear Physics (C. Weiner, ed.), Amer.Inst. of Physics, N.Y. (1972), page 17.

21. N. Bohr, Nature 138, 25 (1936). •

28 PHYSICS TODAY / SEPTEMBER 1978Downloaded 21 Sep 2012 to 141.108.40.129. Redistribution subject to AIP license or copyright; see http://www.physicstoday.org/about_us/terms