IL NUOVO CIMENTO

VOL. 107A, N. 11

Novembre 1994

Results from ~p ~ ~ (*). J E T S E T COLLABORATION L. BERTOLOTTO(5), A. BUZZO(5), P. T. DEBEVEC (6), D. DRIJARD(4), S. EASO(5) R. A. EISENSTEIN(6), W. EYRICH(2), M. FERRO-LUZZI(4), J. FRANZ(3), R. GEYER(2) N. HAMANN(4) (¢), P. HARRIS(6), D. W. HERTZOG(6), S. A. HUGHES(6), T. JOHANSSON(9) R. JONES (4), K. KILIAN (7), K. KIRSEBOM (5), A. KLETT(3), H. KORSMO(8) M. LO VETERE (5), M. MACRI(5), M. MARINELLI(5), B. MOUELLIC (4), W. OELERT(7) S. OHLSSON(4), A. PALANO(I), S. PASSAGGIO(5), J. M. PERREAU(4), M. G. PIA(5), S. POMP (2), p. E. REIMER (6), j. RITTER (6), R. ROBUTTI (5), K. R()HRICH (7), M. ROOK(7), E. ROSSLE(a), A. SANTRONI(5), H. SCHMITT(3), O. STEINKAMp (7) (**), F. STINZING(2), B. STUGU(7), M. TSCHEULIN(3), H. WIRTH(3) and H. ZIPSE (3) presented by O. STEINKAMP (1) INFN, Sezione di Bari e Universit5 di Bari (2) Physikalisches Institut, Universitdt Erlangen (3) Fakultdt fi~r Physik, Universitdt Freiburg (4) CERN, Genevo~ Switzerland (5) INFN e Universit5 di Genova (6) University of Illinois at Urbana-Champaign (7) Institut fi~r Kernphysik) Forschungszentrum Ji~lich (s) Universitet Oslo, Norway (9) Universitet Uppsala, Sweden (ricevuto il 14Aprile 1994)

Summary. - - The reaction ~p ~ ~¢ is investigated in the JETSET (PS202) experiment using an internal target in LEAR. Data have been analysed at a variety of beam momenta from 1.2GeV/c to 2.0GeV/c, corresponding to centre-of-mass energies between 2.15 GeV and 2.43 GeV. A strong ~-signal was observed over the whole energy range, preliminary results giving ~-yields of (1-2)~b. No evidence for narrow resonances was found within the accumulated statistics. The capability of the experiment to measure ~--~ K ÷ K- decay angular distributions and thus put constraints on the spin-parity quantum numbers of a possible resonant intermediate state is demonstrated. PACS 14.40 - Mesons and meson resonances. PACS 01.30.Cc - Conference proceedings.

(*) Paper presented at the conference HADRON '93, Como, June 21-25, 1993. (**) Now at European Organisation for Nuclear Research (CERN), Geneva. (t) Deceased. 2329

2330 1. -

o. STEINKAMP Introduction.

The J E T S E T (PS202) experiment [1] is being performed with an internal target at the Low Energy Antiproton Ring (LEAR) at CERN. The aim of the experiment is to search for hadronic resonances in the mass range above 2GeV, using in-flight antiproton-proton annihilations. Special emphasis has been put on the investigation of the gluon-rich reaction ~p --. ¢¢, where the ~-mesons are detected via their decay into two charged kaons. The accessible energy range is from the reaction threshold at 2.04 GeV to 2.43 GeV, the upper limit being given by the maximum ~-momentum of 2.0GeV/c. The excellent momentum resolution of the LEAR ~-beam ( 5 p / p ~ 10 -3) translates into an invariant-mass resolution of O(1 MeV) and makes this experiment especially suited for the study of narrow resonances. Figure 1 shows a naive quark-line diagram for the reaction ~ p - - ~ , which demonstrates why it is regarded to be a particularly promising hunting ground for ,,gluonic, states of hadronic matter. In this simple model, all quark-antiquark pairs of the initial ~p-state annihilate into a purely gluonic intermediate state from which two s~ quark-antiquark pairs are created. This process is completely quarkqine disconnected and should be strongly suppressed according to the empirical 0kubo-Zweig-Iizuka (OZI)[2] rule. Such suppression may be overcome if a resonant intermediate gluonic state (,,glueball,0 is formed. Recent lattice-QCD calculations [3] indicate that glueballs with spin-parity of jR__ O- and jR.= 2 + fall into the mass range just above 2 GeV. However, other production mechanisms are possible. Quarkline-connected processes involve intermediate states such as K* K* or ~ . The final state ~ could be produced via co-~ mixing from an initially produced (oco pair, or s§ pairs might be directly produced from the quark sea of the proton or the antiproton. Only two experiments have so far published results on the reaction ~p --~ ~ . In an early bubble chamber experiment Davidson et al. [4] isolated six events of the type ~ p - . K ÷ K- K ÷ K- at different ~-momenta between 1.6GeV/c and 2.2GeV/c. The cross-section for this reaction was calculated to be (3.8 +- 1.7)~b. One of the events was compatible with a ~¢ intermediate state, which led to an estimate of 600 nb for the resonant reaction ~p ---) ¢~ ---) K + K- K ÷ K - . At a centre-of-mass energy of 3 GeV the experiment R704 at the Intersection Storage Ring (ISR) at C E R N [5] measured the cross-section for the reaction ~p--, ¢~ to be 25 rib. Several experiments have investigated ~¢ production in radiative J/V-decays and in hadronic reactions. Strong claims for the observation of glueballs in the region above the ~ threshold came from an experiment at the BNL-AGS [6]. In the process n - p - - . ~¢n, an excess of ¢~-events was found compared to uncorrelated ~p--~ K ÷ K- K ÷ K- production. The ~ mass spectrum showed a broad structure extending from threshold up to 2.5 GeV. In a detailed partial-wave analysis, this structure was u

p ud pu

a

'

~

.

9

*' ~ ' ~ T ~ Y / ~ f

*.

~s O

.

. "

s¢ ~-

'

~

Fig. 1. - Quark-line diagram for the reaction ~p---) ~ .

RESULTS FROM

~p~

2331

resolved into three broad jPC = 2 ÷ ÷ resonances which were named gT, g~, and g~ (for tensor glueball) by the authors, but are termed officially '~f2" by the Particle Data Group (PDG). The experiment WA67 at the CERN Omega-spectrometer investigated the reaction = - B e - - . ~ + X[7] and observed an enhancement above threshold in the ~ mass spectrum, which could be fitted with two broad jPC = 2 + + resonances. In apparent contradiction to these results, a jR __ 0- structure near threshold in the ~ mass spectrum was found in the reaction J/~--. ~,~ both by the MarkIII experiment at SLAC-SPEAR [8] and by the DM2 experiment at Orsay-DCI [9]. They did not find evidence for a jR = 2 + contribution. The Jetset experiment examines this disputed energy region with a very good mass resolution.

2. - The experiment. The central feature of the Jetset apparatus is the hydrogen-cluster jet target that is installed in one of the straight sections of LEAR and intersects the coasting ~-beam at right angles. The use of an internal target has been made possible by the LEAR stochastic cooling system which compensates for the perturbation of the coasting ~-beam. Barring technical problems, beam lifetimes of a couple of days were achieved, making very economical use of the antiproton resources. With a typical target density of a few 1012p/cm 2, beam intensities of up to 5.101° ~, and a revolution frequency of the antiprotons of 3.24MHz (at a momentum of 1.5GeV/c), peak luminosities of several 1029cm -2 s -1 have been achieved. An additional feature of this set-up is the relatively small interaction region which extends over < 1 cm3. The main advantage is, however, that the combination of a relatively thin target with an intense antiproton beam fully exploits the small momentum spread of the LEAR machine, leading to a centre-of-mass energy resolution of 0(1 MeV). Figure 2 shows a top view of the apparatus. The interaction region, defined by the intersection of the LEAR beam (in the figure from left to right) and the hydrogen jet (from bottom to top), is surrounded by a compact detector which has been primarily designed to observe events of the type ~ p - - ~ K ÷ K - K ÷ K - . An array of 60 scintillation counters (3) surrounds the beampipe just downstream of the target. They are segmented, like the rest of the detector, into -forward- and ,,barrel, units. The trigger required a total multiplicity of four hits in these counters, at ]east three of which had to fall into the forward region. The tracking detector (4) consists of a total of 2500 cylindrical aluminium drift tubes ( ,,straws , ). It allowed the reconstruction of the direction of charged tracks with a resolution of about 1°. An array of two planes of silicon pad detectors (5) in the forward region measures the energy loss (dE/dx) and was used for particle identification in the offiine analysis. Segmented threshold Cherenkov counters (7) filled with either water (fit = 0.75) or C6F14(/~t=0.80) provide a fast pion/kaon separation in the first-level trigger. Depending on the incident antiproton momentum, a maximum of one or two hits in these counters was allowed. The measured signal was also used for particle identification in the offiine analysis. An outer three-layer scintillator hodoscope (8) forms a multipurpose detector. Due to its high granularity, it gives a fast information on the directions of charged particles which was employed in the trigger and in the track-finding algorithm. The pulse-height information in the scintillators gives

2332

O. STEINKAMP

Fig. 2. - Top view of the JETSET apparatus. additional particle identification and the multiplicity measurement in the three layers was used in the first-level trigger. A high-resolution electromagnetic calorimeter (9) in the forward region and a photon detector (10) in the barrel complete the set-up. They were used to identify and reject events with neutral final-state particles, but also give the possibility to measure complementary physics channels.

3. - A n a l y s i s .

J E T S E T took the first physics data in three run periods during the year 1991. A total of 16 beam momenta, representing two energy scans, have been analysed. One scan, in coarse steps of 100MeV/c over a wide momentum region from 1.2GeV/c, was intended to give a first overview of the accessible momentum range and to establish the performance of the detector. The second scan, in fine steps of 15 M e V / c over the region from 1.4GeV/c to 1.5GeV/c, was motivated by the search for the ~/f4 (2220) resonance [10]. Simple cuts on multiplicities in the detector components were applied to eliminate events which were incompatible with four charged and no neutral particles seen in the detector. Based on the measurements in the drift chambers, events with four

RESULTS FROM

~p-->~

2333

1.4 GeV/c

1400 ~ 1200

50

.....

1000

4o

800

30

600

20 10 •~

'V. 400

o .95 1.25 1 . 1 5 ~ ~ ~ 1.05 " 1.05 0 cP5 1 15

200 0 0.95

2

1.00

1.05

1.10

1.15

1.20

2

M (KI, K2)~"M (K3, K4)(GeV/c )

Min v (K s, K 4) (GeV/c )

Fig. 3. - Plot of invariant masses of 2-kaon pairs (left), and one-dimensional projection (right) Mft = 1019 MeV/c 2,/'flit 15.58 MeV/c 2). Shown is the invariant mass of one kaon pair if the other pair gives a mass in the region of the ~-meson. =

tracks from a common production vertex inside the interaction region were selected. The information from the photon counters was employed to reject events with neutral particles in the final state. Making use of the precise knowledge of the initial state and the measured directions of the outgoing particles, the events were kinematically reconstructed under the hypothesis of a 4-kaon final state. This zero-constraint kinematic reconstruction may give two solutions. The compatibility of the reconstructed momenta with the measurements in the associated thresholdCherenkov and silicon-dE/dx detectors was used to eliminate non-4-kaon background and to define a unique solution for each event. In the search for a ~ intermediate state, the four final-state particles were grouped into pairs of two and the invariant masses of these pairs were calculated. Since the charge of the outgoing particles is not measured, each event gives rise to three combinatorial solutions. In fig. 3 the two invariant masses of the kaon-pairs are plotted against each other. A clear signal at the ~ mass is visible above a broad background, the structure near the diagonal of the plot being a ,,reflection- of wrong combinatorial solutions from events in the ~-peak. Also shown is a one-dimensional projection of this distribution, where the invariant mass of one pair is plotted if the second pair gives an invariant mass within 24 MeV of the ~. In order to obtain the best possible statistics, all analysed data have been merged into this plot. The fit of a Breit-Wigner resonance curve above a parametrised background distribution gave the following values for the mass and width of the ~-signal: m+ : ( 1019.30 _+ 0.16) MeV F~ = (15.58 _+ 0.52) MeV

[(1019.413 _+ 0.008)MeV / '[

(4.43 -+ 0.06) MeV

']"

The reconstructed mass is in excellent agreement with the PDG value [11], quoted in brackets, whereas the width is determined by the resolution of the track reconstruction. Different methods have been used to extrapolate and statistically subtract the background under the ~-signal. They lead to compatible results. In order to estimate the trigger acceptance and the reconstruction efficiency, a

2334

o. STEINKAMP

full Monte Carlo simulation based on the C E R N Geant package, has been developed. A total of 100000 ~p--~¢~ events at each momentum setting were generated assuming isotropic production and decay of the ~-mesons. These events were tracked through the apparatus, taking into account material effects as well as the resolution and the efficiencies of the detector components. The number of Monte Carlo events found by the same analysis chain as used for data defines the reconstruction efficiencies. The luminosity determination relied on the measurement of the elastic crosssection at large angles (G.m. = 90°). For this purpose elastic events were accumulated with a dedicated trigger together with the 4-kaon data. 4. - R e s u l t s .

The analysis [12] reveals at all measured energies a strong ¢~-signal, which clearly dominates over the competing physical processes ~p--* K ÷ K - K ÷ K- and ~p--* --* CK ÷ K - . A signal of the latter process is observed at the highest energies, whereas at low energies it is submerged in the combinatorial background. The determination of total cross-sections for the reaction ~p --* ¢$ suffers from the vanishing acceptance of the detector in the forward region, which is caused by the large size of the LEAR beampipe traversing the apparatus. Figure 4 shows the reconstructed cOS~c.m, angular distribution of the ~ p - - * ~ production and the corresponding acceptance curve from phase-space-generated Monte Carlo. In order to obtain good statistics, all analysed points from 1.405GeV/c to 1.5GeV/c beam momentum have been merged in these plots. The acceptance decreases with increasing cOS~c.m,and vanishes for cosG.m. > 0.8. For this reason we call our measured quantity a ,,yield, rather than a cross-section. In the following, the integrated ~-yield is calculated by assuming a constant differential cross-section over the whole cos ~e.m. range, including the unseen forward region. Under this assumption it is 1

N~

1

Y¢¢-fLdt%¢ BR2(¢ - - * K + K - ) ' where I L dt is the integrated luminosity, N¢¢ is the number of reconstructed events

150 : ~

~40~- i.~0~GeV) C~ 4000 arY-3000.~

"~ 10G

2000 O

o 5C

I000~

,xl

~

0

0.2 0.4 0.6

0.8

1.0

COS 0e.m.

Fig. 4. - The angular distribution of the ~p--* ~}~}production (cos ~c.m.)" The data points are the raw distribution. The solid curve gives the corresponding acceptance from phase-space--generated Monte Carlo.

2335

RESULTS FROM ~p --> ~ beam momentum (GeV/c) 1.21.31.41.51.61.71.81.92.0 3.0

,

2.5

[

J,,,,,,,

,

,

Tf~

,

,

'

15

preliminary ~J

10

~. 2.0 ~•

~J × ~J

1.5

i

-e-

~

5

1.0

oJ

0.5 0.O_

• 20

. ~ . . . . I. . . . . . .t ~ 2.1

2.5

0 , .... I......... ] ......... I ......... 2.2 2.3 2.4 L5 c.m. energy (GeV)

beam momentum (GeV/c) 1.405 1.435 1.390 ~ 1.420 ~ 1.4501.4651.480 ~ I1 J J I I t

T

¢J

~-

~ threshold

3.0

.... *.... * " * .... *.... *

+

1.500 l

preliminary

I0'

8

10 i

4~ 1.0 0.5

- ¢ ~ = - ~

.....

~" .... -¢--'--

~-

......

~. . . . . .

* ........

=-~:

2

0.0_.21' . . . . . . . 2.221 . . . . ,/ . . . .2.23' . . . . . . . . . 2.241 . . . . . . . . . 2.251. . . . . . . . 2.26 0 ~/f4(2220) c.m. energy (GeV) Fig. 5. - The ~p--* ~ excitation function (open circles), calculated under the assumptions described in the text. Diamonds indicate the corresponding Monte Carlo efficiencies, to be read with the right-hand scale (the dashed line through these points is plotted to guide the eye). All analysed data are included in the upper plot, whereas the lower plots shows enlarged the energy region between 2.21 GeV and 2.26 GeV. of the type ~p --~ ~ --. K + K - K ÷ K - , s** is the average ~p --. ~ reconstruction efficiency from Monte Carlo, and BR ( ~ - ~ K ÷ K - ) is the branching ratio for the ~-meson decay into two charged kaons. F o r this branching ratio the PDG value of 49.1 _+ 0.8% was used. Figure 5 shows preliminary ~ - y i e l d s for all energies measured in 1991 and, enlarged, for the region of the ,,fine, scan between 2.21 GeV and 2.26 GeV. The e r r o r s

2336

o. STEINKAMP *~ 250 ~ O

" ~

~ ' ~

'

4000

200 150

3000 2000

50 !preliminary 0

3000~ 1.405- 1.500 GeV/c 2500 preliminary 2000

1 4

I00

~

15o

,~ 5000

'

"

0.2 "

O:i' (.6' o'.8' COS

0K

1000 0 1,0

1500 o

~

50

I000 500

.1=

~

~

o

o

20

40 60 Z (degrees)

80

0~

Fig. 6. - ~ -* K ÷ K- angular distributions, on the left for the polar angle of the kaon in its parent meson's rest frame (cos~g), on the right the angle between the two ~ decay planes (~). Shown are the raw distributions, the solid line indicating the acceptance from phase-space-generated Monte Carlo. include the statistical errors on the number of reconstructed events, the background subtraction, the efficiency correction, and the determination of the integrated luminosities. Data at 1.5GeV/c beam momentum (2.254GeV centre-of-mass energy) were taken in each of the three separate running periods to give a relative normalisation for the measurements. The results agree within the errors, indicating that the statistical errors dominate over systematic effects in the point-to-point determination of ~¢-yields. The plots also show the corresponding reconstruction efficiencies, which rise smoothly with increasing energy. The reconstructed ~¢-yields of (1-2) ~b are almost two orders of magnitude higher than the value seen at 3.0 GeV by the R704 Collaboration [5], but in good agreement with the estimate made by Davidson et al. from the one single reconstructed event [4]. Within the statistics the excitation function does not exhibit evidence for narrow resonant structures. Constraints can be put on the spin-parity quantum numbers of a possible intermediate state by investigating ~--~K+K decay angular distributions, in particular, the polar angle of the kaon in the rest frame of its parent ~-meson (cos ~ K) and the angle ~ between the two ~-meson decay planes. Figure 6 shows preliminary distributions of the angles, again summed over all data taken between 1.405GeV/c and 1.5GeV/c beam momentum, and the corresponding acceptance curves from phase space Monte Carlo. Within the statistics, the data show no deviation from the phase space distribution, as would be expected for a j R __ 2 +, but not for a pure j R = = 0- intermediate state. Each ~ p - ~ event is, however, described by six independent parameters, and the correct treatment of these distributions involves a very careful acceptance correction performed simultaneously in all six parameters. This work is still in progress and no final conclusions should be drawn from the shown one-dimensional projections.

5. -

Conclusions.

J E T S E T has measured the reaction ~p--) ~ in an interesting region above threshold. A strong ~¢-signal, which clearly dominates over uncorrelated ~p--~ -* K ÷ K- K ÷ K- production, has been observed over the whole measured energy range.

RESULTS FROM~p --> ~¢

2337

Preliminary ~-yields have been determined to be of the order of (1-2) ~b. Within the statistics that have been analysed so far, no evidence is seen for narrow structures in the excitation function. In spite of acceptance limitations in the ~-production angular distribution, the analysis of decay angular distributions will allow us to put constraints on spin-parity quantum numbers of possible intermediate states. Work is in progress on this topic. More data have been accumulated during the past two years attempting to fill gaps in the ~p --) ¢~ excitation function and to statistics, especially in the region of the ~/f4(2220). The analysis of these data is in progress. With a significantly improved luminosity, which is due to machine and jet target improvements, J E T S E T will be able to complete the study of the reaction ~p---) +~ in a reasonable time. The installation of a forward RICH detector has greatly increased the particle identification capabilities of the apparatus and thereby opens new physics possibilities.

We want to acknowledge the important contribution to the success of this experiment from our colleague and friend Niko Hamann, whose sudden and untimely death has left a great void in our community.

REFERENCES [1] JETSET COLLABORATION(C. EVANGELISTAet al.): The ~p--> ~ reaction near threshold~ CERN/SPSLC 92-42, SPSLC M501. [2] S. OKUBO:Phys. Lett., 5, 165 (1963). [3] UKQCD COLLABORATION(G. S. BALI et al.): A comprehensive study of SU(3) glueballs, Liverpool Preprint LTH 303. [4] J. DAVIDSON, J. W. CHAPMAN, R. W. GREEN, J. LYS and B. P. ROE: Phys. Rev. D, 9, 77 (1974). [5] R704 COLLABORATION(C. BAGLINet al.): Phys. Lett. B, 231, 557 (1989). [6] A. ETKIN, K. J. FOLEY, R. W. HACKENBURG,R. S. LONGACRE,W. A. LOVE, T. W. MORRIS, E. D. PLATNER,A. C. SAULYS,S. J. LINDENBAUM,C. CHAN, M. A. KRAMER:Phys. Lett. B, 201, 568 (1988). [7] WA67 COLLABORATION(P. S. L. BOOTHet al.): Nucl. Phys. B, 273, 677, 689 (1986). [8] MARKIII COLLABORATION(Z. BAI et al.): Phys. Rev. Lett., 65, 1309 (1990). [9] DM2 COLLABORATION(D. BISELLOet al.): Phys. Lett. B, 241, 617 (1990). [10] For an overview of the status of the ~/f4(2220), see PS185 COLLABORATION(P. D. BARNESet al.): Measurement of the reaction ~p ~ K~ K ° in the region near V~ ~ 2230 MeV, accepted for publication in Phys. Lett. B, and references therein. [11] PARTICLE DATA GROUP: Review of Particle Properties, Phys. Rev. D, 45, VII22 (1992). [12] O. STEINKAMP:Messung der Reaktion ~p - , ~ am Experiment JETSET, PhD thesis, to be published.