Vouuse 19, Nustou 21 PHYSICAL REVIEW LETTERS 20 Novena 1967 obtaining the expression (11) the mass difference ‘between the charged and neutral has been ignored. 42yt, Ademollo and R. Gatto, Nuovo Cimento 444, 262 (4966); see also J. Pasupathy and R. B. Marshak, Phys, Rev. Letters 17, 888 (1966). ‘The predicted raito leq. (12)] from the current alge- bra Is slightly Iarger than that (0.288) obtained from the p-dominanee model of Ref. 2. This seems to be ‘true also in the other case of the ratio P= x*=7y)/ ly) calculated in Refs. 12 and 14, SL, M. Brown and P. Singer, Phys. Rev. Letters 460 (1962), A MODEL OF LEPTONS* Steven Weinbergt Laboratory for Nuclear Seience and Physies Department, Massachusetts Institute of Technology, Cambridge, Massachusetts (Received 17 October 1967) Leptons interact only with photons, and with the intermediate bosons that presumably me- diate weak interactions. What could be more natural than to unite? these spin-one bosons into a multiplet of gauge fields? Standing in the way of this synthesis are the obvious dif- ferences in the masses of the photon and inter- mediate meson, and in their couplings. We might hope to understand these differences bby imagining that the symmetries relating the weak and electromagnetic interactions are ex- act symmetries of the Lagrangian but are bro- kken by the vacuum. However, this raises the specter of unwanted massless Goldstone bosons.* ‘This note will describe a model in which the symmetry between the electromagnetic and weak interactions is spontaneously broken, but in which the Goldstone bosons are avoided by introducing the photon and the intermediate boson fields as gauge fields.” The model may be renormalizable. We will restrict our attention to symmetry groups that connect the observed electron-type leptons only with each other, i.e., not with muon-type leptons or other unobserved leptons or hadrons. ‘The symmetries then act on a left- handed doublet avoil'®) ® 2 and on a right-handed singlet Re[b0-y)le. 2) ‘The largest group that leaves invariant the kine- matic terms ~IyH0,,L-RyH8,R of the Lagrang- ian consists of the electronic isospin T acting on L, plus the numbers Nz, Np of left- and right-handed electron-type leptons. As far a8 we know, two of these symmetries are en- tirely unbroken: the charge Q=73-Np-4Nz, and the electron number =p +z. Bat the gauge field corresponding to an unbroken sym- metry will have zero mass, and there is no massless particle coupled to N,° 60 we must form our gauge group out of the electronic iso- spin F and the electronic hyperchange Y= Np +N. Therefore, we shall construct our Lagrang- inn out of Land R, plus gauge flelds Ay, and By, coupled to F and ¥, plus a spin-2er0 dow- biet a) ® ‘whose vacuum expectation value will break T and Y and give the electron its mass. ‘The on- ly renormalizable Lagrangian which is invar- jant under T and Y gauge transformations is ee 2 oA seh xR Pe 2-0 3, -0,K, +88, xR,)-40,8,-9,8, Ry! a ae RM (0, ~ig'B \R-Ly! (0 igt+K ibe" YL y VM, ~ie'B RLY! it, 18a, t oriole? A. Tosi 2G Gor+Ro't)-u2 10 ,9-igh,, To +ike’B, of -C Lor +Ro'L)-M,"@ We have chosen the phase of the R field to make Gp real, and can also adjust the phase of the Zand @ fields to make the vacuum expectation value A=(¢") real. The “physical” g fields are then @~ 1264Vouuste 19, Nustoen 21 PHYSICAL REVIEW LETTERS 20 Novimner 1967 and 2 (0 +9 -aAF @, 6 We see immediately that the electron mass is Gz. ‘The charged spin-1 field is ‘The condition that ¢, have zero vacaum expec Ww a2 sia) ® tation value to all orders of perturbation the- ¥ ae ony tells us that X*=4,*/2h, and therefore the field 9, has mass M, while ¢, and g~ have mass and has mass zero. But we can easily see that the Goldstone bosons represented by 2 and g~ have no phys Mya Be. ® ical coupling. The Lagrangian is gauge invar~ sant, #0 we can perform a combined isospin ‘The neutral spin-1 fields of definite mass are and hypercharge gauge transformation which tliminates o~ and 9, everywhere! without chang Zi-letse*)MM(eA *46°B), (10) ing anything else, We will see that Gp is very # we small, and in any case M, might be very large,” aymusygea * 50 the Y, couplings will also be disregarded Aya (ere -W'A,*468,). (2) in the following. ‘The effect of allthis is just to replace @ ev- They masses are exywhere by its vacuum expectation value @ © My = alate, a2) ‘The first four terms in £ remain intact, while the rest of the Lagrangian becomes as) pel oF AF “ 80 Ay, is to be identified as the photon field. “H(g4,246°B,, 96,20, (1), The interaction between leptons and spin-1 fe mesons 1s Beet, tes eanne, HE eV (Be) He oy ocala gy, yy fo (GB oreo y se oma syg]z,, (4) We see thatthe rationalized electric charge ||| f-——————___________ is en ge' (gt +g) (1s) and, assuming that W,, couples as usual to had- rons and muons, the usual coupling constant of weak interactions is given by Gy B= /8M y? = 1/20. (as) Note that then the e-y coupling constant is 07X10" The coupling of y, to muons is stronger by factor My,/Me, but still very weak. Note al- so that (14) gives g and g’ larger than e, 50 (16) tells us that My>40 BeV, while (12) gives ‘Mz >My and Mz >80 BeV. ‘The only unequivocal new predictions made by this model have to do with the couplings of the neutral intermediate meson Z, . If Zy does not couple to hadrons then the best place to look for effects of Z,, is in electron-neutron scattering. Applying a Fierz transformation to the W-exchange terms, the total effective e-v interaction is, G, { Gee") Ww, He story, TE Oe ae gayY © eee Ie g>e then g>8", and this is just the usual e-v scattering matrix element times an extra factor 3. If ge theng«g’, and the vector {interaction is multiplied by a factor -} rath- er than 3. Of course our model has too many arbitrary features for these predictions to be 1265Vouume 19, Num 2 PHYSICAL REVIEW LETTERS 20Novemnre 1967 taken very seriously, but it is worth keeping in mind that the standard calculation’ of the eleetron-neutrino eross section may well be wrong 1s this model renormalizable? We usually do not expect non-Abelian gauge theories to be renormalizable if the vector-meson mass is not zero, but our Z, and Wy, mesons get, thelr mass from the spontaneous breaking of the symmetry, nt from a mass term put in at the beginning. Indeed, the model Lagrang- Jan we start from is probably renormalizable, so the question is whether this renormalizabil- ity Is ost in the reordering of the perturbation theory implied by our redefinition of the fields And if this model is renormalizable, then what happens when we extend itt include the coup lings of Ay and 2, to the hadrons ? Tam grateful to the Physics Department of MIT for thelr hospitality, and to K. A. Johnson for a valuable discussion. ‘This work is supported in part through funds pro- vided by the U. S. Atomic Energy Commission under Contract No. AT(90-1)2098), 0n leave from the University of California, Berke- ley, California, ithe history of attempts to unify weak and electro- ‘magnetic interactions is very long, and will not be re- viewed here, Possibly the earliest reference Is E. Fer- nai, 2. Physik 88, 161 (1934). A model similar to ours ‘was discussed by'S. Glashow, Nucl. Phys. 22, 579 {0961); the chlet difference is that Glashow introduces symmetry-breaking ferms Into the Lagrangian, and therefore gts less definite prediction 8]. Goldstone, Nuovo Cimento 19, 154 (1961); J. Gold~ stone, A, Silam, and 8, Weinberg, Phys. Rev. 127, 85 (1962). 3p, W. Higgs, Phys. Letters 12, 132 (1904), Phys. Rev, Letters 13, 508 (1964), and Phys, Rev. 143, 1158 (0960); F. Eaglert and R. Brout, Phys. Rev. Letters 413, 521 (1964); G. 8. Guraltky C. Re Hagen, and T. W. B, Kibble, Phys. Rev, Letters 13, 585 (1904). ‘seo particularly T. W. B. Kibble, Phys. Rev. 155, 1554 (1967). A similar phenomenon oceurs inthe strong interactions; the p-meson mass in zeroth-order Perturbation theory is Just the bure mass, while the ‘A; meson picks up an extra contribution from the spon taneous breaking of chiral symmetry. See 8. Weinberg, Phys, Rev. Letters 18, 507 (1967), especialy footnote Ted. Schwinger, Phys, Letters 24B, 473 (967) 5. Giashow, Hl. Schntzer, and 8. Weinberg, Phys. Rev. Letters 19, 139 (1967), Bq. (13) ef soa. Sn, D. Lee and C. N. Yang, Phys. Rev. 98, 101 (1855). This 'is the aame aortof transformation as that ‘which eliminates the nonderivative F eouplings in the © model; ace S, Welnberg, Phys. Rev, Letters 18, 188 (0967), ‘The ¥ reappears with derivative coupling be- cause the strong-interastion Lagrangian is no invari~ ‘nt under chiral gauge transformation. ‘por a similar argument applied tthe # meson, see Weinberg, Ref. 6. i. Feynman and M. Gell-Mann, Phys, Rev. 100, 193 (951). SPECTRAL-FUNCTION SUM RULES, »-y MIXING, AND LEPTON-PAIR DECAYS OF VECTOR MESONS* R. J. Oakest ‘Brookhaven National Laboratory, Upton, New York and 3. J. Sakurai ‘The Enrico Fermi Institute for Nuclear Studies and the Department of Physics, ‘The University of Chicago, Chicago, Hlinois (Received 18 Octaber 1967) Within the framework of vector-meson dominance, the eurrent-mixing model is shown to be the only theory of w-o mixing consistent with Weinberg’s first sum rule as applied to the vector-carrent spectral functions. Relations among the leptonlo decay rates of 6, ©, and ¢ are derived, and other related processes are discussed. We begin by considering Weinberg’s first sum rule! extended to the (1+8) vector currents of the eightfold way*s Sam?{n= 2 gg MO) + ag onl 1266 ap *S"Paq%gor a)