Direct observation and partial-width measurement of γγ decay of charmonium states

  • Published on
    02-Sep-2016

  • View
    212

  • Download
    0

Embed Size (px)

Transcript

  • Volume 187, number 1,2 PHYSICS LETTERS B 19 March 1987

    DIRECT OBSERVATION AND PARTIAL-WIDTH MEASUREMENT OF 73' DECAY OF CHARMONIUM STATES

    Annecy (LAPP) -CERN-Genoa-Lyon ( IPN) -Os lo -Rome-Strasbourg-Tur in Collaboration

    C. BAGLIN a, S. BAIRD b, G. BASSOMPIERRE a, G. BORREANI c, J.-C. BRIENT a, C. BROLL a,l, J.-M. BROM d, L. BUGGE e, T. BURAN ~, J.-P. BURQ f A. BUSSI~RE a, A. BUZZO g, R. CESTER , M. CHEMARIN f, M. CHEVALLIER f B. ESCOUBES a, j. FAY f, S. FERRONI s, V. GRACCO s, J.-P. GUILLAUD b, E. KHAN-ARONSEN e, B. ILLE f K. KIRSEBOM , M. LAMBERT f, L. LEISTAM b, A. LUNDBY b, M. MACRI b, F. MARCHETTO c, E. MENICHETT I c, Ch. M6rch ~, B. MOUELL IC b, N. PASTRONE c, L. PETRILLO h, M.G. PIA ~, J. POOLE b, M. POULET ~, G. R INAUDO ~, A. SANTRONI b, M. SEVERI h, G. SKJEVLING ~ and B. STUGU ~ Lapp, B P 909, F-74019 Annecy le Vwux Cedex, France

    b CERN, CH-1211 Geneva 23, Swttzerland Umverstty of Turm, 1-10125 Turm, Italy

    d CRN, F-67037 Strasbourg Cedex, France Unlverstty of Oslo, N-1000 Oslo 1, Norway

    f IPN Lyon, F-69622 Vtlleurbanne Cedex, France g Umverstty and INFN, 1-16126 Genoa, Italy h Umverstty of Rome, 1-00185 Rome, Italy

    Received 19 December 1986

    As part of the charmonlum formation experiment at the CERN Intersecting Storage Rings, we stud~ed the reactmn plY--ce--.7"r m an antiproton momentum scan through the ~lc, X ~, and X2 regions. We report events observed m the n and X2 regions, whilst no event was observed in the X z region, as expected for a spin-1 state.

    1. Int roduct ion

    The interest in studying the 77 decays ofcharmon- lUre states is well known. Potential models [ l ], as well as QCD sum rules [ 2], can be tested by meas- uring these decays. In addition, since the ratio F(R- ,7 ' / ) /F (R - -~hadrons) (where R is the reso- nant state) is directly related to the running coupling constant as (Q 2) of QCD, the measurement of this ratio allows the validity of perturbative QCD calcu- lations to be checked in this range of Q2.

    When this experiment began to take data, only upper limits for this decay mode of the ~c and X2 states were available [ 3]. Meanwhile, new experi- mental results have been published: in pho- ton-photon interactions at PETRA, the PLUTO

    Deceased, 21 July 1984.

    0370-2693/87/$ 03.50 Elsevier Science Publishers B.V. (North-HoUand Physics Publishing Division )

    Collaboration [4] has observed a bump in the KOK_+x :F mass spectrum, with a mass about 3 GeV. A natural interpretation of this bump is given by the reaction 7Y ~]c-~KK+-x ~:. Using the 1984 Particle Data Group branching ratio BR(~coKK-+Tt :~ ), they obtained the value of the partial width F(TI~' /7) - - (14_+I2) keV. This gives an experi- mental indication of a coupling of the I]~ to 77, but the partial width is still quite uncertain, later, the TASSO Collaboration [ 5 ] observed a signal in the same process, giving F0]~77) =(50_+40) keV. More recently, at PEP, the MARK II Collaboration [ 6 ] observed four events above background in the same channel. The value derived for F(Tl-~77) is (8.0_+ 5.0-+ 2.0) keV. The Crystal Ball Collaboration at SPEAR [ 7 ] have observed the '/7 decays of the go and X2 states, and give the partial width

    191

  • Volume 187, number 1,2 PHYSICS LETTERS B 19 March 1987

    F(X2--,~'t) = (2.8 + 2.0) keV. We report the results obtained by the R704 Collab-

    oration at the CERN Intersecting Storage Rings (ISR) for the reaction

    Pf)--)VT (1)

    studied in the mass region of the ~c, )C ~, and )C2 states of the (cO) system.

    2. Experimental set-up A detailed description of the experimental tech-

    nique and of the apparatus has been given in pre- vious publications [8]. The proton-antiproton

    annihilations were produced, in ring 2 of the ISR, by intersecting a cooled coasting antiproton beam with a molecular hydrogen jet.

    The detector (fig. 1 ) wa a non-magnetic spectrom- eter with two detection arms set symmetrically rela- tive to the beam axis. Each arm had a charged-particle tracking section, including sets of scintillation hodo- scopes, a threshold ~erenkov counter for tagging electrons, and a transversally and longitudinally seg- mented electromagnetic calorimeter. For 77 detec- tion, the calorimeter is the essential part of the apparatus and has to provide a high 7:/7 rejection factor in order to be able to isolate the ~7 signal over

    / / v - X--sONT ,o s s, _ . .

    $

    Fig. 1. Schematic view of the experimental apparatus.

    192

  • Volume 187, number 1,2 PHYSICS LETTERS B 19 March 1987

    the large multineutral hadronic background. The cal- orimeter was set up to optimize

    (i) the efficiency of detection of the low-energy 33'S,

    (ii) the 71/spatial resolution, (iii) the energy resolution. Each calorimeter had four parts: (1) The 1/ converter, or precalorimeter. This was

    made of lead/scintillator sandwiches amounting to a total length of 4.7Xo. The transverse segmentation consisted of 29 vertical strips and 14 horizontal wedge-shaped strips.

    (2) The analog chambers. These consisted of four identical (X, Y) chambers with analog readout of the the cathodic strips (10 mm pitch).

    (3) The shower hodoscope. This consisted of 14 vertical scintillator strips (NEI 10, 1 cm thick) used for triggering.

    (4) The lead-glass wall was made of 66 blocks of F2 glass with the dimensions 15 15 30.5 cm 3.

    The length of the blocks (30.5 cm or 10Xo) ensured full shower containment.

    A system of guard counters completely sur-

    200 o= 27 MeV

    160

    120

    ~0

    I I i i 100 200 300 t00 500

    M('('KI IMeVl

    Fig. 2. Effective mass spectrum for two photons m the same arm.

    rounded the two arms, covering most of the remain- ing solid angle. This guard system was composed of scintillation counters followed by segmented lead /scintillator sandwich counters (4.7Xo). Its aim was to detect charged particles and/or "/-rays outside the detection arms and therefore to set multiparticle or multi-33 veto conditions.

    The fast trigger (200 ns) for the 33y final state was implemented using the guard system and shower hodoscope information.

    A slow trigger (450 ~ts) made use of a fast proces- sor, which read the calorimeter fast ADCs, calcu- lated the energy deposited in the precalorimeter and the lead glass, and performed a cut on the total energy deposited. The trigger rate was a few events per sec- ond. Triggering for the ~ono final state was also obtained at the same time. The ability of the detector to reconstruct n's is shown in fig. 2, which displays the reconstructed mass for the 11 pairs.

    3. Data taking The energy scan was done around the qc, X z, and

    X2 regions. The energies and luminosities are given in table 1. The X~ scan was performed to constrain the background level in the X2 sample.

    Owing to the presence of the multi-y veto condi- tion, the largest background (from hadronic multi- neutrals) is coming from the xx channel, which simulates an exclusive 33y final state from symmetric n decays inside the angular resolution or from very asymmetric n decays where the low-energy y escapes detection. In order to evaluate the multi-33 back- ground to reaction (1), we have also studied the reaction

    pp~nx (2)

    in the same regions.

    Table 1 Candidate events, mass range, and lumlnoszty

    Mass range No. of events Luminosity (MeV) (nb - j )

    2963-2969 0 96 2975-2987 12 511 3023-3025 1 98 3509-3516 0 494 3552-3564 6 821

    193

  • Volume 187, number 1,2 PHYSICS LETTERS B 19 March 1987

    4. Data analysis The selection of T7 final-state candidates in reac-

    tion (1) was based on the single y-shower topology, the energy deposit, and the two-body kinematics.

    As afirst step we used the following criteria in order to obtain an intermediate sample of a reasonable size suitable for a detailed study of the final analysis cuts: - no charged particle entering the detector; - total energy deposited in the calorimeters, > 0.8Etot; - only one electromagnetic shower per arm, with E> 150 MeV; - an effective mass of these two showers > 75% of the centre-of mass energy (i.e. > 2.2 GeV for the q); - extra energy deposited in the detector, < 50 MeV.

    This first step of the analysis led to a total sample of 1173 TT candidate events out of 2.6 million trig- gers (all data taken at the 11c, Z ~, and ~2 states). Their centre-of-mass cos 0* distributions are shown in fig. 3a (s-~9 GeV 2) and fig. 3b (s - ~ 12 GeV2). The cor- responding CM angular distributions of =o~o events (4 y converted in the calorimeters) are shown in figs. 3c and 3d. A comparison of figs. 3a and 3c indicates

    250

    ~, 200

    150 o

    100

    50

    al

    ~ l , I , "--] 02 01, 06

    COS e e

    1000

    000

    600

    Z~O0

    200

    _7

    , L Ot , I 02 06

    cos O*

    that in the q~ region there is an important ~o~o back- ground in the 77 sample, since we expect an iso- tropic angular distribution for the T1~77 decay. In order to reduce this background, we disregarded for the TI~ state the region cos 0"> 0.35, where the ~o~o. induced background is the highest. For the Z i and Z2 states, where the ~o~o angular distribution flattens and there is no definite expectation for the ~2---~/~/ angular distribution, we have considered data up to cos 0*=0.5 so as to avoid edge effects at the limit of the acceptance.

    Thefinal step of the data analysis was based on (i) restricted kinematical constraints using the

    shower position measurements given by the analog chambers for which angular resolution was obtained using the exclusive JA sample [8,9] (pp-~J/~t --,e+e-);

    (ii) longitudinal and transverse shower develop- ment characteristics; these were derived also by the analysis of the exclusive J/V sample;

    (iii) no possible 7:o reconstruction using the main shower correlated with a very low energy deposit.

    The efficiency of this analysis is about 60%. This has been derived by reanalysing the exclusive J/~ sample, disregarding the charged-particle tracking section and ~erenkov information.

    The final results of the analysis chain are reported in table 1.

    5. Results 5.1 The vc sample. Fig. 4 displays the CM angular

    distribution for the 7)' candidates in the qc region. The hatched event (corresponding to a mass of

    g t.o

    '~ 30

    20 E

    02 Or, 06 COS B e

    30

    2O

    10 I

    0 02 OZ, 06 COS e e

    Fig. 3. The cos 0* angular dsstributlons for the first-step 77 sam- ple (a) at s-~9 GeV 2 (qc sample), (b) at s= 12 GeV 2 (Z2 sam- ple), and cos 0* angular distribution for nn, (c) at s - 9 GeV 2, (d) a ts= 12 GeV 2.

    4

    3

    2 I 1 i

    [ ] r/c -~ "V)" events

    [ ] Background sample

    I L [ 01 02

    COS 8

    , I O3

    cos O angular d]stnbut]on for the final "qc sample. Fig. 4. The *

    194

  • Volume 187, number 1,2 PHYSICS LETTERS B 19 March 1987

    12 '1 " "l I . . . . . l " " I 1

    II ~oa g if)

    ~3o4

    02

    2 97 2 98 2 99 ,3. 3 01 3 02 3 03 Moss (GeV)

    Fig. 5. Excltatmn curve m the r h mass region. The curve is the result of the maximum likehhood procedure descrabed m the text The dashed line is the background contribution parametnzed by the expressmn (rb,~k= ab [ 1--O.0024(pbe~,,--3635)] pb obtained from the study of the n n sample ( Ob ~S the parameter used m the maximum hkehhood fit;/h,~,~ xs m MeV).

    3023.5 MeV) , far f rom the central mass value o f the resonance, is considered as a background event. The excitat ion funct ion in this ~1~ mass region is shown in fig. 5. The curve is the result o f a fit obta ined with a max imum l ikel ihood method, descr ibed in ref. [ 10]. The total width Ftot, the mass, the branching ratios, and the background level were free parameters :1. An energy-dependent background was used. The best-f it parameters are

    m~(2982.6 +2 ~) MeV, +7 5 _ F to t= (7.0_7 o) MeV,

    BR (r)~ ~p13) X BR(q~--)~/y) +o 42 = (0.68_o,3z) 10 -6 ,

    :t If Fro, is fixed al (11 +_4) MeV [ 11 ], the best-fit parameters are: mno=(2982.4_+2 6) MeV, BR(q--,pl))BR(rI:~Ty ) =(0.57_+0.26) -6 +14o X 10 , ab,~k= (90-7o) pb, glwng the follow- mg branching ratm: BR('q:-.77) = (0.5 +_0.2 +-0.3) X 10 -3 and partial width T'( rlc --. t,'}' ) = ( 5.7 +- 2.6 -%_ 3.7 ) keV.

    Table 2 Resonance parameters from R704

    +2OO aback = ( I00_ loo)Pb

    This cross section takes into account the eff ic iency and acceptance correct ions (~=0.05) and corre- sponds to the value o f the background at the B~ mass. I f we use the publ ished value o f BR0q:- )p lb) = (I . I + 0.6) 10- 3 [ 12 ], we obtain the fol lowing branching ratio:

    _ 10 ~+o.4 + 0 .4) X 10 -3 BR(rh-- 'T~') - , " " -o 3 -

    and partial width

    t 4 a+3 4 +2.4) keV F (q-*~' ) = ~ .~-3 7 -

    The first errors come f rom the max imum likeli- hood fit. The last errors correspond to the uncertain- ues in the other parameters. These uncertaint ies are composed quadratical ly.

    We have checked the stabil ity o f our results with respect to the dif ferent cuts used in the analysis. The

    Resonant Mass Total width BR(R--,pl~) BR(R--,~) F(R--, yT) BR(R--,~) state [MeV] [MeV] ( 106 ) [keV] ( x 103)

    29 ~':''~+2~ 7 a75 0.68 -o3~ -,.-,-~ ~ - ~. 06-o~ +-0.4 a) ~c v~.v - -23 "v - -70 +042 A " / -3 -~-9 '4 +04 X2 3556.9+0.4_.+05 b~ 2 6-+] 4 b) +0046 0 099-oos5 2.9_+I 3 + +os o-1.7 1.1_o4-+0.4~)

    a) From ref. [12]' BR(n:-,plb) = (1.1 +__0.6) 10 -3. b) From ref. [10] c) From ref [ 10]: BR (Z2--,plb) = (0,9- + ] ) 10 -4.

    195

  • Volume 187, number 1,2

    Table 3 Summary of partml widths (m keY).

    PHYSICS LETTERS B 19 March 1987

    Resonant Parhal width [keV] state

    R704 PLUTO [4] TASSO [5] MARK II [6] CRYSTAL BALL [7]

    4 a+34"4''~ A 14+12 50+40 80--+5.0+2.0 ~c .~37 .2-'/~...t X 2 9-+I 03+ 1.7 2.8-+2.0

    results are stable for slight variations of these cuts ~;2. Table 2 summarizes our results. The value of the

    partial width is in agreement with the theoretical expectations [ 2,13 ] and is consistent with the upper limit given by the Crystal Ball [ 7 ] and with the val- ues given by PLUTO [4], TASSO [5], and MARK II [6 ]. Table 3 shows our results and the present published values for this partial width.

    5.2 Thex~ andg2 samples. The Z~ and Z2 samples are given in table 1. No candidate event is found in the Z ~ region

    A likelihood maximization has been performed using the two samples. In this case, the mass, the total width of the Z2, and the BR(z2--,p~) are as deter- mined in ref. [ 10]. So we have only two free param- eters: the branching ratio BR(z2~T) ' ) and the background cross section. The ~2--~/'/ CM angular distribution is strongly dependent on the helicity amplitudes of the initial and final states and is not likely to be isotropic. However, in the absence of any experimental result for this angular distribution, we consider it as isotropic and give results in this frame- work. These results are

    BR (X 2--,pp ) X BR (Z2--)TT) +0 46 = (0.99_0 35) 10 -7 ,

    0"back < 70 pb.

    This upper limit (90% confidence level) for the background also takes into account the efficiency and acceptance corrections (~ = 0.07). Using

    ~2 In particular, the maximum hkehhood method applied to the reduced sample corresponding to cos 0"< 0.3 leads to the fol- lowing results. No. of events in the tie mass intervals of t...