VOLUME 1$, NUMBER 4 P H Y SI CAL RE V I E%' LETTERS 27 JULY 1964

EVIDENCE FOR THE 2rr DECAY OF THE Km MESON*1

J. H. Christenson, J. W. Cronin, V. L. Fitch, and R. Turlay~Princeton University, Princeton, New Jersey

(Received 10 July 1964)

PLAN VIEW

I root

VFEEEPEEEEPz

57 Ft. to =internal target

Cerenkov

FIG. 1. Plan view of the detector arrangement.

This Letter reports the results of experimentalstudies designed to search for the 2m decay of theK, meson. Several previous experiments haveserved"~ to set an upper limit of 1/300 for thefraction of K2 's which decay into two charged pi-ons. The present experiment, using spark cham-ber techniques, proposed to extend this limit.

In this measurement, K,' mesons were pro-duced at the Brookhaven AGS in an internal Betarget bombarded by 30-BeV protons. A neutralbeam was defined at 30 degrees relative to the

1 1circulating protons by a 1&-in. x 12-in. x 48-in.collimator at an average distance of 14.5 ft. fromthe internal target. This collimator was followedby a sweeping magnet of 512 kG-in. at -20 ft. .

and a 6-in. x 6-in. x 48-in. collimator at 55 ft. A1~-in. thickness of Pb was placed in front of thefirst collimator to attenuate the gamma rays inthe beam.

The experimental layout is shown in relation tothe beam in Fig. 1. The detector for the decayproducts consisted of two spectrometers eachcomposed of two spark chambers for track delin-eation separated by a magnetic field of 178 kG-in.The axis of each spectrometer was in the hori-zontal plane and each subtended an average solidangle of 0.7&& 10 steradians. The squark cham-bers were triggered on a coincidence betweenwater Cherenkov and scintillation counters posi-tioned immediately behind the spectrometers.When coherent K,' regeneration in solid materialswas being studied, an anticoincidence counter wasplaced immediately behind the regenerator. Tominimize interactions K2' decays were observedfrom a volume of He gas at nearly STP.

Water

The analysis program computed the vector mo-mentum of each charged particle observed in thedecay and the invariant mass, m*, assumingeach charged particle had the mass of thecharged pion. In this detector the Ke3 decayleads to a distribution in m* ranging from 280MeV to -536 MeV; the K&3, from 280 to -516; andthe K&3, from 280 to 363 MeV. We emphasizethat m* equal to the E' mass is not a preferredresult when the three-body decays are analyzedin this way. In addition, the vector sum of thetwo momenta and the angle, |9, between it and thedirection of the K,' beam were determined. Thisangle should be zero for two-body decay and is,in general, different from zero for three-bodydecays.

An important calibration of the apparatus anddata reduction system was afforded by observingthe decays of K,' mesons produced by coherentregeneration in 43 gm/cm' of tungsten. Since theK,' mesons produced by coherent regenerationhave the same momentum and direction as theK,' beam, the K,' decay simulates the direct de-cay of the K,' into two pions. The regeneratorwas successively placed at intervals of 11 in.along the region of the beam sensed by the detec-tor to approximate the spatial distribution of theK,"s. The K,' vector momenta peaked about theforward direction with a standard deviation of3.4+0.3 milliradians. The mass distribution ofthese events was fitted to a Gaussian with an av-erage mass 498.1+0.4 MeV and standard devia-tion of 3.6+ 0.2 MeV. The mean momentum ofthe K,o decays was found to be 1100 MeV/c. Atthis momentum the beam region sensed by thedetector was 300 K,' decay lengths from the tar-get.

For the K,' decays in He gas, the experimentaldistribution in m is shown in Fig. 2(a). It iscompared in the figure with the results of aMonte Carlo calculation which takes into accountthe nature of the interaction and the form factorsinvolved in the decay, coupled with the detectionefficiency of the apparatus. The computed curveshown in Fig. 2(a) is for a vector interaction,form-factor ratio f /f+= 0.5, and relative abun-dance 0.47, 0.37, and 0.16 for the Ke3, K&3, andEg3 respectively. The scalar interaction hasbeen computed as well as the vector interaction

138

VOLUME 1),NUMBER 4 PHYSICAL REVIEW LETTERS 27 JULY 1964

QATA: 52ll EVENTS~ ---- MONTE-CARLO CALCULATION

VECTOR —' -0.5 - 6oo

484&m" & 494 -. IO

- 500

I0IIl~II

I

- 400

- 500

- 200

- IQO494& m~& 504

30

CA

-- 201 LU

LLI

- IO

i iJ500 550

I I I

400 450 500 550 600 MeVX

p iZ

(b)504&m" & 5 l4 -- IO

"---"MONTE- CARLO CALCULATION -~ )20VE CTOR — & 0.5 "-

I IOf+ - IOP90

~,- 80ewe $0

r-- ~ spr-i ' ~ 5py

II 40

- 50. 20IO

10.998 0,999cos g

FIG. 2. (a) Experimental distribution in rn~ com-pared with Monte Carlo calculation. The calculateddistribution is normalized to the total number of ob-served events. (b) Angular distribution of those eventsin the range 490 &m*&510 MeV. The calculated curveis normalized to the number of events in the completesample.

with a form-factor ratio f /f+ =-6.6. The dataare not sensitive to the choice of form factorsbut do discriminate against the scalar interac-tion.

Figure 2(b) shows the distribution in cos8 forthose events which fall in the mass range from490 to 510 MeV together with the correspondingresult from the Monte Carlo calculation. Thoseevents within a restricted angular range (cos8&0.9995) were remeasured on a somewhat moreprecise measuring machine and recomputed usingan independent computer program. The results ofthese two analyses are the same within the re-spective resolutions. Figure 3 shows the re-

00.9996 0.9997 0.9998 0.9999 I.OOOO

cos 8

FIG. 3. Angular distribution in three mass rangesfor events with cos0 & 0.9995.

suits from the more accurate measuring machine.The angular distribution from three mass rangesare shown; one above, one below, and one encom-passing the mass of the neutral K meson.

The average of the distribution of masses ofthose events in Fig. 3 with cos8 &0.99999 isfound to be 499.1 + 0.8 MeV. A correspondingcalculation has been made for the tungsten dataresulting in a mean mass of 498.1 + 0.4. The dif-ference is 1.0+0.9 MeV. Alternately we maytake the mass of the E' to be known and computethe mass of the secondaries for two-body decay.Again restricting our attention to those eventswith cos0&0.99999 and assuming one of the sec-ondaries to be a pion, the mass of the other par-ticle is determined to be 137.4+ 1.8. Fitted to aGaussian shape the forward peak in Fig. 3 has astandard deviation of 4.0 + 0.7 milliradians to becompared with 3.4+ 0.3 milliradians for the tung-sten. The events from the He gas appear identi-cal with those from the coherent regeneration intungsten in both mass and angular spread.

The relative efficiency for detection of thethree-body E, decays compared to that for decayto two pions is 0.23. %e obtain 45+ 9 events in

139

VOLUME 1$, NUMBER 4 PHYSICAL REVIEW LETTERS 27 JuLY 1964

the forward peak after subtraction of backgroundout of a total corrected sample of 22 700 K,' de-cays.

Data taken with a hydrogen target in the beamalso show evidence of a forward peak in the cos0distribution. After subtraction of background,45+ 10 events are observed in the forward peakat the K' mass. We estimate that -10 events canbe expected from coherent regeneration. Thenumber of events remaining (35) is entirely con-sistent with the decay data when the relative tar-get volumes and integrated beam intensities aretaken into account. This number is substantiallysmaller (by more than a factor of 15) than onewould expect on the basis of the data of Adairet al. '

We have examined many possibilities whichmight lead to a pronounced forward peak in theangular distribution at the K' mass. These in-clude the following:

(i) K,' coherent regeneration. In the He gas itis computed to be too small by a factor of -10' toaccount for the effect observed, assuming reasonable scattering amplitudes. Anomalously largescattering amplitudes would presumably lead toexaggerated effects in liquid H, which are notobserved. The walls of the He bag are outsidethe sensitive volume of the detector. The spatialdistribution of the forward events is the same asthat for the regular K,' decays which eliminatesthe possibility of regeneration having occurredin the collimator.

(ii) K&3 or Ke3 decay. A spectrum can be

constructed to reproduce the observed data. Itrequires the preferential emission of the neutrinowithin a narrow band of energy, +4 MeV, cen-tered at 17+ 2 MeV (K&3) or 39+ 2 MeV (Ke3).This must be coupled with an appropriate angularcorrelation to produce the forward peak. Thereappears to be no reasonable mechanism whichcan produce such a spectrum.

(iii) Decay into w+7t y. To produce the highly

singular behavior shown in Fig. 3 it would benecessary for the y ray to have an average ener-gy of less than 1 MeV with the available energyext nding to 209 MeV. We know of no physicalprocess which would accomplish this.

We would conclude therefore that K2 decays totwo pions with a branching ratio R = (K2- w++ w )/(K,'- all charged modes) = (2.0+ 0.4) && 10 wherethe error is the standard deviation. As empha-sized above, any alternate explanation of the ef-fect requires highly nonphysical behavior of thethree-body decays of the K,'. The presence of atwo-pion decay mode implies that the K,' mesonis not a pure eigenstate of CI'. Expressed asK,0=2 "'[(K,-KO)+e(KO+KJ] then I&I'= R&T—IT2where 7, and T, are the K, and K,' mean livesand RZ is the branching ratio including decay totwo r'. Using RT = &R and the branching ratioquoted above, l et =—2.3x 10

We are grateful for the full cooperation of thestaff of the Brookhaven National Laboratory. Wewish to thank Alan Clark for one of the computeranalysis programs. R. Turlay wishes to thankthe Elementary Particles Laboratory at Prince-ton University for its hospitality.

*Work supported by the U. S. Office of Naval Re-search.

This work made use of computer facilities sup-ported in part by National Science Foundation grant.

~A. P. Sloan Foundation Fellow.~On leave from Laboratoire de Physique Corpusculaire

h Haute Energie, Centre d'Etudes Nucldaires, Saclay,France.

M. Bardon, K. Lande, L. M. Lederman, andW. Chinowsky, Ann. Phys. (N. Y. ) 5, 156 (1958).

D. Neagu, E. O. Okonov, N. I. Petrov, A. M.Rosanova, and V. A. Rusakov, Phys. Rev. Letters6, 552 (1961).

3D. Luers, I. S. Mittra, W. J. Willis, and S. S.Yamamoto, Phys. Rev. 133, B1276 (1964).

R. Adair, W. Chinowsky, R. Crittenden, L. Leipun-er, B. Musgrave, and F. Shively, Phys. Rev. 132,2285 (1963).

140