129

OBSERVATION OF THE T AND T '

IN ANNIHILATIONS

C.W. Darden, H. Hasemann, A. Krol zig, W. Schmidt-Parzefal l, H. Schroder, H.D. Schulz, F. Selonke, E. Steinmann , R. Wurth , Deutsches Elektronen-Synchrotron DESY, Hamburg

W. Hofmann , D. Wegener, Institut fLlr Physik, Uni versitat Dortmund

H. Albrecht, K.R. Schubert, Insti tut fUr Hochenergiephysik der Universitat Hei delberg

P. Bockmann, L. Jonsson Insti tute of Physics, Uni versity of Lund

presented by W. Schmi dt-Parzefall

Abstract : •

The T and T ' have been observed using the DASP detector at the DORIS storage ring. A fi rst study of the decays of the T meson has been performed .

Resume : Le et le T ' ont ete observes en uti l isant le detecteur DASP aupres de l'anneT au de stockage DORIS. Nous presentons des premieres etudes sur les desi ntegrations du meson T. 130

A Columbia-FNAL-Stony Brook col l aborati on in studying the reaction + - l) p + N + µ µ + X discovered peaks in the pair invariant mass distri bution near 10 GeV. They identified these peaks wi th three new , which they called T(9.4), T' (10.0) and T"(l0.4) . Since the observed width was equal to the experimental resol ution, these new mesons were interpreted in analogy wi th the J/¢ mesons as a bb of a new heavy b with a new flavour. + The best way to investigate such a state is its production in e e- annihilation. Thus can be found out whether it is real ly a state as narrow as expected for a bound bb system , and a detailed study of its decay properties can be perfonned .

The upgrading of the e+e- storage ring DORIS to reach the energy region of the T required its conversion into a single-ring , single-bunch machine and the instal lation of additional RF cavi ties. This operation, which started in fall of 1977 was done in several steps ending in Apri l 1978, when DORIS started to 30 -2 -1 operate in the T region wi th a peak luminosi ty of about 10 cm s . A maximum energy of 10.2 GeV has been reached.

In the two interaction regions of DORIS, initially measurements were perfonned with the detectors PLUTO and DASP . The Double Arm Spectrometer DASP, however, was run by a new col l aboration. Later on PLUTO moved to PETRA and was replaced by the Na I -Lead-Gl ass detector of the DESY-Heidel berg group, again operated by 2l a new col l aboration .

Thus each measurement has been performed twi ce, and the agreement between the results has been generally very close.

We report here on results obtained with the Double• Arm Spectrometer DASP The 3) . apparatus consists of a non-magnetic inner detector, which covers a sol id angle of about 50% of 4rr, and two magnetic outer detectors , which cover appro ximately 5% of 4n . The properties of the detector have been descri bed before .

The storage ring luminosity wa s continuously monitored by a counter system of four telescopes positi oned at scatteri ng angl es of 8° to record smal l an le g Bhabha scattering. Large angle Bhabha scattering was recorded wi th the inner detector independently. The two methods agreed to within ± 15%.

The cross section of a for the reaction e+e- + has been obtained in h 4 the fo owing way : Abo1Jt 10 triggers lead to one good event . To reject inter­ 11 acti ons with the res idual gas in the storage ring and cosmic rays , a set of 131 selection criteria was applied :

Particle tracks were reconstructed with the five layer scintillator hodoscopes of the inner detector. At least three tracks were required , which did not al l lie in the same hemisphere. The total energy deposited in the shower counters' of the detector was required to be more than MeV wi th no single track 450 depositing more than GeV. Events whi ch passed thi s tes t were visually .5 scanned for tracks in3 the proportional tube counters . Again at least three tracks were required which did not all lie in the same hemisphere . They had to originate from a common vertex point close to the interaction region. Finally the accepted events were computer reconstructed .

The resulting visible cross secti on versus energy is shown in Fig. 1. The curve 1 is the best fi t of a gaussian wi th radiati ve corrections 4 ) and a ;s back­ ground. A clear resonance signal is observed at a mass of ± 0.010) GeV . (9.457 The error refl ects the present precision of the energy cal i bration of DORIS. The measured r.m.s. wi dth of the resonance is 7.6 MeV , in perfect agreement wi th the mass resolution of the storage ring, and exactly what would be expect­ ed for a narrow resonance. The corresponding resul ts of the PLUTO col l abor­ ation 5) is shown in Fig. 2.

+ The partial width for the decay T + e e- can be obtai ned from the ree integrated cross section 'hdE for production, using the Breit-Wigner formula f

M dE 2 h 6 2 1f I (J

Wi th the assumption 5 )

h R : (J ± 0. at GeV I crµµ 4.5 5 9.4 we determine the acceptance of our detector and selection criteria for hadron events off resonance to. be n Since the average multipl icity that we c = 4 %. measure increases slightly for events2 on resonance, we estimate that the acceptance of our detector increases to on resonance . Including radiati ve 47% corrections we get an integrated hadronic cross section of

o dE (370 ± 100) MeV nb. J h = 132

nb OASP Z 8

4

t t t t 0 9.35 9.40 945 9.5 0 GeV

Visibl e cross section of the reaction e+e- +hadrons . The solid curve is a single resonance with Gaussian energy resolution and radiative corrections . One standard deviation of the energy resolution is 7.6 MeV .

20 cr(nb) PLUTO 15

10

0 9.1 5 9.25 9.35 9.45 9.55 is(GeV) Fig. 2 Cross section for the reaction e+e- +hadrons versus CM energy as measured by PLUTO. 133

The error contains the estimated systematic uncertainties .

+ - To determine the branching ratio B for the decay T(9.4) + µ µ we used the DASP inner detector with its largeµµ sol id angle. The selection criteria were : exactly two tracks, no showers and the tracks should be col linear. To reject cosmic ray , the time of fl ight to the fifth layer of the scintillator hodoscope had to be equal within 1 ns. For cosmic ray muons the difference is typical ly 6 ns.

The detector blocks above and below the beam tube could not be used since there the flight path was too short for suffi cient discrimination against cosmi c rays . Finally, the events had to occur in coincidence with the bunch-crossing. The two muon track determined the z-coordinate - along the beam - of the inter­ action . The z-distribution is shown in Fig. 3. After subtraction of the remain­ ing cosmic ray background , we obtained 86 ± 13 events in the T resonance and 44 ± 10 events outside.

The visible cross section for muon pair production in the continuum outside the resonance is c a . (0.30 ± 0.08) nb , µµVlS which determi nes the acceptance of the detector, 0.31 ± 0.08 . '\iµ = The visible cross section in the peak of the resonance is T . (o.41 ± 0.07) nb . a µµVlS + - Combining the resul ts, we find the integrated cross section of e e + r(9.4) + + - µ µ Ja dE (10 ± 8) nb. µµ =

We assume the three leptonic widths r ' and r of the (9.46) to be ee r TT equal and estimate that about half of the µµ decays are included in our measured hadronic cross section. This allows us to obtain values for the partial width r and the branching ratio B r /r : ee ee ee

r ee 0.4) keV. (1.5 ± B B (2.5 ± 2.1) %. ee = µµ = 134

N 25

20

15

10

5

o+-�...-.�...... � ...... ���� ...... -150 -100 -50 0 50 100 150 (mm) �

Vertex distribution for the reaction T(9.46) µµ showi ng the � si gnal and a background ori ginating from cosmi� c rays .

KeV re ele� in KeV 20

w Y(eQ= -1/3) J/¢ 10 i· � t

0 2 4 6 B 10 M(GeV) Fi g. 4 The ratio of the electronic width r over the square of the ee quark charge e2 for radial ground state vector mesons as Q a function of mass. 135 The quoted errors are one standard deviation, including estimated systematic errors. The total width r of the T(9.46) has a best estimate of 60 keV and a two-standard deviations lower limit of 20 keV; the statistics obtained do not yet permit us �o give a reasonable upper limit. Non-relativistic potential model s of a bound bb system make specific predictions for ree· It is a direct measure of the square of the charge eb of the b quark.

6) where ¢(0) is the wave function at the origin. Eichten and Gottfried find 1 ree = 0.7 keV for e b = ;3 e0• It is important to note that the tatio r /e is roughly the same for the vect0r mesons ¢?)and J/¢. As can be ee p, � 4, �· seen from Fig. the measured ree suggests lebl = 13 e0 3 According to QCD, the predominant decay mode of a heavy quark antiquark s1 state is b6 3 3 jets. + +

Assuming that this decay is the only hadronic decay mode of the T (g .46) and taking the quark charge as 1/3 e , Kol ler and \fol sh predicted - 0 S) ee : r 1 27, which is compatible with the result of this experiment, r = B = (2.5 2 1 %. Our results therefore favor the assignment of charge . ) ee ± 1/3 e0 to the new fifth quark and indicate that the T(9.46) is a bb state.

The inclusive production spectra have been studied with the magnetic spectrometer of the DASP detector. Figs. Sa and b show the invariant production cross section for the inclusive reactions

as a function of the particle energy on the resonance and off the resonance. Above the momentum cutoff of 200 MeV/c, the show an exponential spectrum E d3 o/dp3 exp(-E/t with � 0) E0 (on resonance) (260 ± 25) M:eV, E0(off resonance) (240 ± 40) MeV. 136

Within the stati stical accuracy, there is no difference between the values on and off resonance . The on and off resonance and the on reson­ ance fol low roughly the wel l-known law that they are produced as frequently as pions with the same particle energy. The relative abundance of kaons is about the same on resonance as it is off resonance . Hence no abnormal coupling of the T(9.46) to the different flavors has been observed .

According to QCD the hadronic decay of the T(9.4) should proceed via the emission of 3 gluons . The experimental observation of this decay would be a strong indication for the validity of QCD. However, so far no fi rm theoretical predictions exist, how gluons should fragment into the eventually detected hadrons .

One of the early assumptions 8) was, that a gluon first decays into a quark pair and the subsequently fragment into hadrons . This picture would predict an increase of a factor 1.5 to 2 of the average mul tiplicity of hadrons produced on resonance as compared to off resonance.

+ We have studied the average multiplicity of events of the reaction e e- + 3) hadrons . All events contained in the visible cross section of Fig. 1, (� wi th three or more visible tracks includi ng converted were ful ly re­ constructed in the inner detector. The mul tipl icities have not been corrected for the acceptance of the detector.

Fig. 6 shows the average visibl e mul tiplicity of charged tracks and converted photons observed in the selected events as a function of the centre of ma�s energy . As can be seen, there is a small but clear rise of the average mul ti ­ plicity by ± 2) % in the resonance region . This result is not suggesti ng (8.4 that gluons decay mainly into (symmetric) quark pairs, where a larger ri se was expected .

It is however consistent with a Monte Carl o simulation in which gluons were allowed to directly fragment into pions as though they had been quarks , using 9l the fragmentation algorithm of Field and Feynman for the simulation of gluon jets .

A more detai led search for the gluon decay of the T meson has been performed 3 by studyi ng the topology of the hadron events . 137

On Resofllnce · Off Resonance 100 100 2 OASP 2 DASP jo E� E d dp' • dP' t +" 10 10 t ft } t r t K t t\ t t t ' t It t t 01 t 01 -t- 2t 2 Eh !GeVI Eh!GeVI Fig. 5 Invariant production cross secti on for the inclusi ve reacti ons + - ± ± ± e e + rr X, K X, p X on and off the T(9.46 ) resonance.

DASP 2 (nh+ny)6.0

5.5 r1l 5.0

45 .L-....._���---'��-+-��-1---l 9.4 0 9.45 9.35t �so ev G

Fi g. 6 Average visible mul tiplicity of charged and converted photons in the inner detector. The values are not corrected for acceptance. 138

+ The existence of quark jets in e e- annihilation has been successful ly demonstrated by the use of the topology variable sphericity Four topology lO). va ri ables have been proposed 11) to search for 3-gl uon-jet . events : Spheri ci ty, . spherocity, thrust and acoplanarity. The first three essentially measure the particl es' transverse momenta with respect to a preferred axis, and the fourth one measures the momentum components perpendicul ar to a preferred plane . Two­ jet events have small spheri (o)city and smal l acopl anari ty , whereas 3-jet­ events are supposed to show bigger val ues of the spheri( o)city but comparably small val ues of the acoplanarity.

Si nce we cannot determine particle momenta with the DASP inner detector, we study the behaviour of four "pseudo " topological variables and determine them fo r each event : 2 Pseudosphericity Mi n < sin o > , i 2 Pseudospherocity (� Min ) 'IT Pseudo thrust Max , 2 Pseudoacoplanarity 4 (Min ) •

where is the angle of each track wi th respect to the preferred axis, or, 8 for pseudoacoplanarity, wi th respect to the normal to the preferred plane . Our fo ur defi nitions would coincide with those of reference 11, if all particles had equal momenta.

Figs. 7a - 7d show the mean values of the distri buti ons of the four quantities as a function of the centre-of-mass energy; and Fig. 8 shows two distributions of the pseudospherocity averaged over all centre-of-mass energies on and off resonance. We note a definite change in all four mean values at the centre-of­ mass energy of the resonance and a definite change of the distributions with Fig. 8 as one exampl e When going from off-resonance to on-resonance, the two­ jet character of the events decreases , the events become more isotropic. This increase of isotropy takes pl ace in three (and not on ly in two ) dimensions, since the acoplanari ty is also increasing. The changes in event topology are so pronounced that they are even seen in geometrical quantities wi thout knowl edge of the parti cle momenta . Are these observations in agreement wi th a possible decay of the 7(9.46) into three-gluon jets ?

Si nce there are no theoreti cal predictions for our geometrical q11anti ties and since th e detector acceptance has an appreciable infl uence on their val ues , 139

Pseudo- Pseudo- spheroc1ty thrust

045 085 ft

040 080 + I + + + T + t+ I . + +ti t t t + 035 ms

Pseudo- Pseudo- spher1c1ty acoplanarity

055 02 ft H

05 t 015 jt + + t lt + + + +I 045 01 {s g35+ 94 Vs \GeV) 945 tgs m 94 (GeVl 91'..S +g5

Fig. 7 Mean val ues of the to pology variables as a function of centre­ of-mass energy .

DASP 2 100

N 80

60

40

20

02 0.4 06 0.8 PseLdo-spheroc it y Distribution of the pseudospherocity 'nea r the resonance (open poi nts) and for the direct resonance decay (ful l points ). The direct decay distribu tion is obtained by subtracting the resonance back­ ground and decays through the one channel . 140

we have tri ed a simp le Monte-Carlo simulation of 2-jet and 3-jet events . Two­ jet events are generated according to an angular distri bution

2 do(jet)/dQ 1 + cos 8(jet) �

9 ) and to a "standard " fragmentation into pions wi th

320 MeV/c . t = Three-jet events are generated according to the jet-momentum distribution descri bed in reference 6 and assumi ng that gl uons fragment into pions exactly th e way quarks do . Al l events are passed through a rough model of the detector acceptance and are treated afterwards exactly like real events . The resul ts are summarized in Tabl e 1.

The experimental col umn shows the relati ve change of the values from off-reson­ ance to on-resonance energies , and the Monte Carl o col umn incl udes non-resonant background and resonance decays through the one-photon channel . The agreement between the two col umns is surprisingly good.

We draw the fol lowi ng concl usions from th is agreement : If the T (9.4) decays into hadrons by way of three gluons , then these gl uons must fragment into hadrons in a way simi lar to quark fragmentation, with nearly the same p,, - distribution and nearly the same P - distri bution . Direct evi dence for the t decay of the T(9.46) into three gl uons remains to be found .

The T-meson shoul d have exited states if it is a b6 state. Indications of such an exited state have already been observed by the Col umbi a-FNAL-Stony Brook collaboration in studying the reaction p + N X. The energy di fference + rr+rr- + between the ground state T and the first exited state was fi tted to be r' (610 ± 42) MeV .

The mass splitti ng between the J/� and the �· Mesons is 586 MeV . According to QCD the force between a quark and its antiquark should be almost independent of the quark mass. With this assumption the mass spl itting is given by the 12) shape of the potential between the quarks . Quigg and Rosner have shown that for a logarithmical ly ri sing potential the exitation levels are independ­ 6) ent of the quark mass . For the potential used by Ei chten and Gottfried Coulomb pl us a linear confining term, the level spacing decreases with ri sing quark mass.

+ - A scan was performed wi th DORIS in order to search for the Meson in e e - + - r' annihilations. The reaction e e hadrons was studied wi th the DASP detector + 141

using the same analysis criteria as described before.

Data were taken at 15 different centre-of-mass energies between 9.98 and -1 10.04 GeV, the total lumi nosity accumulated was about 120 nb . The resulting cross sections are given in Fig. 9.

We observe evidence for a narrow resonance at (10.012 O.J20) GeV; the error ± refl ects mainly the uncertainty of the energy cal i bration of DORIS. The measured rms wi dth of 9 MeV is consi stent with the mass resolution of the machine.

As part of the error in the energy cali bration cancels for mass differences , the mass splitting between the and the new resonance which we identify as T the T' can be estimated wi th higher accuracy

(555 11) MeV . ±

Thi s value is slightly smaller than that obtained for

586 MeV

in accordance wi th non-relativistic potential models of a bound quark-antiquark state.

The partial decay width ree is eval uated from the integrated cross section a dE for hadron production assuming R (4.5 ± 0.5) and imposing a constant h = 4) accf eptance of 42 4% and applying radiative corrections ±

(0.35 0.14) keV . ±

The error contains systematic uncertainties .

Fig. 10 shows the measured mass spl itting and the exitation levels for a 13) logarithmi cal ly rising potential . Rosner et al . have computed the particle decay with r of the T and the T ' meson for various simple potentials. The ee shaded area of Fig. 11 corresponds to a quark charge of 1/3.

For a quark charge of 2/3 the r of both, the and the T ' mesons should be ee T fo ur times as big. As can be seen from the figure, the two measured r ee val ues define a region in the shaded area, consistent with non rel ativistic potential models and again strongly indicating that the new heavy quark has charge 1/3.

A summary of the resonance parameters as measured so far is given in Table 2. 142

O VIS ( nb) 7 DASP 2 (1978)

900 9.50 1000 10.40 rs ( GeV J + Fig. 9 Visible cross section for the reaction e e- hadrons shewing the � and the ' meson . T T

3S

.Q 0.8 2S :§ ·0 >< \IJ' •y' W 0.4

y IS 0 \I} 5 I 2. 10 2.0 mo/me = [ V ( r) ln �o Fi g. The exi tation energies for a logari thmical ly r1 s1ng potential 10 adjusted to the charmonium states . Al so shown is the measured exitation energy of the meson. r' 143

I I I I I I I I I � 2.0 I I I I "' I "' I t ---�! ------f-< I h 1,0 I I I I I I

0.5 1.0 1.5 f(T'-e·e-), keV

The electronic wi dth of the versus the electronic width of the T' meson. The shT aded area is computed for vari ous potenti als and a quark charge 1/3 . The data fi t well wi th this assumption. 144

TABLE

CHANGE ON/OFF EXP [%] MC [%]

Pseudoacoplanarity < > 18.8 ± 4.0 23.0 Pseudosphericity < > 9.1 ' 1.8 9.8 Pseudospherocity < > 9.5 ± 2.0 14 .0 Pseudothrust < > 2.8 ± 0.6 3.1 Mu1 tiplicity < > 8.4 ± 2.0 7 .o !I

TABLE I I

Results onl(9.46l M(Yl !YI 8µµ ( eVI Exp.Width(MeVI r,.(k eVI ('I.I r.01(k eVI 9.46G !0.01 18!2 1.h0.4 2.7!2.0 >20(2s d.I DAPLUTOSPil 9.46!0.01 18!2 1.5!0.4 2.5t2.1 20( d > s D-H I 9.46! 0.01 17!2 1.04t0.28 10. •3.4 12 l -1.0 >15 2s.d I Mean Values r = (12 0.21 keV 8µµ= 12.6 !1t .41 '/, ot 25 keV (95'1.c .LI (Bestft value liot =50keV)

Results on "((10.02) M lrl-M!Yl l;.(Y') r..!Yl (GeV)rn (MeV) (keV) lee(Y') DASPI 10.012!0.020M 555!11 0.35!0.14 4.3!1.5 D-H 10.02!0.02 560! 10 0.32!0.13 3.3 !0.9 Il Mean 10.016t0.021 558! 10 0.33!0.10 3.6!0.8 9 M(Y'l-M(Y) M(tjJ') -M(J/tjJ) < 145

Acknowledgement

We express our thanks to Dr. Degele and the DORIS group for their excellent work in upgrading and operating the storage ring.

References :

1) S.W. Herb et al ., Phys . Rev. Lett. 39 (1977) 252 2) J.K. Bienlein et al ., Phys . Lett. 78 (1978) 360 3) C.W. Darden et al ., Phys . Lett. 76B (1978) 246; ]SB (1978) 364; SOB (1979) 419 4) J.D. Jackson and D.L. Scharre , Nucl . Instr. Meth ._gtl_ (1975) 13 5) Ch . Berger et al ., Phys . Lett. 76B (1978) 243 6) Eichten and Gottfried, Phys . Lett. 66B (1977) 286 7) See f.e. T.F. Wa lsh, DESY 76/13 (Apri l 1976) 8) K. Koller and T.F. Wa lsh, Phys . Lett. � (1977) 277, Erratum ]3B (1978)504 9) R.D. Field and R.P. Feynman, Nucl .Phys . Bl36 (1978) 1 10) R.F. Schwitters et al ., Phys . Rev. Letters (1975) 1230 ;!2 G. Hanson et al ., Phys . Rev. Letters ;!2 (1975) 1609 11) A. de Rujula et al ., QCD Predictions for Hadronic Final States in - e+e Annihilations , Prepri nt TH . 2455 - CERN 12) J.L. Rosner et al ., Phys . Lett. ]4B (1978) 350 13) C. Quigg and J.L. Rosner, Fermilab-Pub-77/82-THY (August 1977) .