Observation of the decay η → π0γγ

  • Published on
    22-Aug-2016

  • View
    212

  • Download
    0

Embed Size (px)

Transcript

  • IL NUOV0 CIMENTO VOL. 71 A, N. 4 21 0ttobro 1982

    Observation of the Decay ~--~ ~yy.

    ( IHEP (*)-IISN (**)-LAPP (***) COLLABORATION)

    F. BINON (**), C. BRICMAN (**), V .A . DAVYDOV (*), S .V . DONSKOV (*) J . ])UFOURNAUD (***), ~P. DUTEIL (***), M. GOUAN]~RE (***)~ A .V . INYAKIN (*) V .A . KACHANOV (*), D .B . I~AKAURIDZE (*), G .V . KHAUSTOV (*) A .V . KULIK (*), J . P . LAGNAUX (**), A .A . LEDNEV (*), YU. V. MIKHAILOV (*) J . P . PEIGNEUX (***), YU. D. PROKOSHKIN (*), YU. V. I:[ODNOV (*) S .A . SADOVSKY (*), A .V . SINGOVSKY (*), A .V . STARZEV (*) J . P . STReeT (**) and V .P . SIYGO~YAEV (*)

    Jo int Exper iment o] IHEP . Serpukhov, USSR CERN - Geneva, Switzerland

    (ricevuto il 21 Giugno 1982)

    In memory of A.V. STARZEV who passed away unexpectedly on June 14, 1981

    Summary . - The rare radiative decay ~-~~ w has been effectively observed and studied at the 70 GeV IHEP accelerator with the hodo- scope spectrometer GAMS-2000. The branching ratio for this decay is found to be BR(~ --~rOyy) = (9.5 :~ 2.3)- 10 -4.

    1. - In t roduct ion .

    A previous exper iment at IHEP has shown (1) that the branch ing rat io for the rad iat ive decay of the ~-meson into n o and a photon pai r

    (1) . .q __~ ~Oyy

    (*) Institute for High Energy Physics, Serpukhov, USSR. (**) Institute Interuniversitaire des Sciences Nucl6aires, Brussels, Belgium. (***) Laboratoire d'Annecy de Physique des Particules, France. (***) CERN, Geneva, Switzerland. (1) F. BINON, C. BRICMAN, V. A. DAVYDOV, S. V. DONSKOV, J. DUFOURNAUD, P. DUTEIL,

    497

  • 498 IHEP-IISN-LAPP COLLABORATION

    is smaller than 3.10 -3, bringing to a~n end a long-lasting controversy (2) between current theoretical values lower than 10 -~ (8,4) and exper imental results up to 10 -~ which cast some doubts on schemes like PCAC, current ~lgebra, etc.

    Measurements that have been pursued at the IHEP accelerator h~ve al- lowed us to observe the decay mode (1) for the first time.

    A contamination mainly due to the intense decay channel

    (2) ~ -~ 3n ~

    might explain earlier claims for the observation of decay (1). The present work is devoted to these new results.

    2. - Exper imental procedure.

    The measurements have been performed with ~-mesons produced by 30 GeV/c negative pions in the charge exchange reaction (5)

    (3) T.-p -~ ~n.

    The hodoseope ~erenkov spectrometer GAMS-2000 (for details concerning the set-up see (1,6,7)) has been used to detect the photons emitted by the

    M. GOUAN~RE, .A.V. INYAKIN, V..A. KACHANOV, D. B. KAKAURIDZE, G. V. KHAUSTOV, A. V. KULIK, J. 1 ~. LAGNAUX, _A_. A. LEDNEV, YU. V. MIKHAILOV, J. i ~. PEIGNEUX, YU. D. PROKOSHKIN, YU. V. RODNOV, S. A. SADOVSKY, A. V. SINGOVSKY, A. V. STARZEV, J. P. STROOT and V. P. SIIGONYAEV: :Fad. iz., 33, 1534 (1981); Lett. Nuovo Cimento, 32, 45 (1981); see also the l iterature quoted therein. (~) PARTICLE DATA GROUP: .Rev. Mod. Phys., 52, 70 (1980). (s) G. OePo and S. ONEDA: Phys. Bey., 160, 1397 (1967). (4) T. P. CHENG: Phys. l~ev., 162, 1734 (1967). (5) W.D. APEL, K. H. AUGENSTEIN, E. BERTOLUCCI, YU. B. BUSHNIN, S. V. DONSKOV, R. FRANCESCHINI, . GIVOLETTI, :M. I. GRACHEV, A. V. INYAKIN, R. JOHNSON, V. A. KACHANOV, W. KITTENBERGER, R. N. KRASNOKUTSKY, M. KRUGER, G. LEDER, A. A. LEDNEV, I. MANNELLI, YU. V. MIKHAILOV, H. MOLLER, R. OBERPARLEITER, M. PER- NICKA, G. ~ . PIERAZZINI, YU. D. PROKOSHKIN, 1V[. QUAGLIA, D. SCHINZEL, H. SCHNEIDER, A. SCRIBANO, F. SERGAMPIETRI, R. S. SHUVALOV, G. SIGURDSSON, H. M. STAUDENMAIER, M. STEUER, A. THYS, A. N. TOROPIN and M. L. VINCELLI: ~'a~. ~iz., ").9, 1519 (1979); Nucl. Phys. B, 152, 1 (1979). (6) F. BINON, C. BRICMAN, V. . DAV2gDOV, S. V. DONSKOV, J. DUFOURNAUD, P. DUTEIL, M. GOUAN:ERE, .A.V. INYAKIN, V. A. KACHANOV, D. B. KAKAURIDZE, G. V. KHAUSTOV, A. V. KULIK, J. P. LAGNAUX, A. A. LEDNEV, YU. V. iV[IKHAILOV, J. P. ])EIGNEUX, YU. D. I~ROKHOSHKIN, YU. V. RODNOV, R. ]{OOSEN, S. ~k. SADOVSKY, D. SILLOU, A. V. STARZEV, J. P. STROOT and V. P. SIIGONYAEV: Nuel. Instrum. Methods, 188, 507 (1981). (7) F. BINON, V. A. DAVYDOV, S. V. DONSKOV, J. DUFOURNAUD, P. DUTEIL, IV[. GOUA- NgRE, A. V. INYAKIN, V. A. KACHANOV, D. B. I{AKAURIDZE, G. V. KHAUSTOV, YU. S.

  • OBSERVATION OF THE DECAY "~--~7~OT~ 4~9

    decaying ~'s. More details about the spectrometer will be given in a forth- coming paper. Only the modif ications which have been brought to the experi- menta l set-up used in the previous work (~) are ment ioned here.

    The distance L between the l iquid-hydrogen target (7) und GA~IS has been increased f rom 9 m to 12 m and the momentum of the r:- beam bus been lowered by 30 %. The distance between T-shower impacts on the spectro- meter is, as a consequence~ increased by 70 ~o. As a result, spurious processes (like (2)), which could be confused with decay (1) when showers overlup~ are better identified. Simultuneously this has brought an improvement of the mass resolution of GAMS for decaying particles (aM/M _~ 1 ~ ) (fig. ] ) and easier working conditions for the shower reconstruct ion programs.

    2000

    iroo i i

    490 520 550 580 ~: :(I,I eV/c 2)

    3000

    N

    Fig. 1. - Mass resolution of the spectrometer GAMS-2000 for MO-particles produced in reaction (4) decaying into 3re ~ The peak corresponds to ~-~ 37: ~ the arrow points to the value of the ~ mass given in the tables. The mass resolution of GAMS is charac- terized by a M = 5.1 MeV/c ~, ~M/M = 0.9%. The curve is a Gaussian distribution. I t practically coincides with the distribution of events simulated with real showers (see below).

    The ~- beam intensity has been increased up to 1.5.30 T per burst for a proton spill on the internal target of 1.5 s, at the cost of a more intense ir- radiat ion of GAMS. In order to avoid a gradual b lackening of the lead glass

    KHODIREV, A. V. KULIK, J . ~). LAGNAUX, .A .A . LEDNEV, YU. V. ~/~IKHAILOV, J . 1 :). PEIGNEUX, YU. D. PROKOSHKIN, YU. D. I~ODNOV, 1:~. ROOSEN, S, A. SADOVSKY, A. V. STARZEV, J. P. STROO~, V. P. STSGONYAEV and A. E. YAXVTIN: Fad. ~iz., 33, 1244 (1981); Z. Phys. C, 9, 109 (1981).

  • 500 IHEP-IISN-LAPP COLLABORATION

    cells which are nearest to the beam, the central radiator cells of GA~r have been replaced by other ones made of a special glass 80 times more radiation resistant (8).

    In these conditions the number of events

    (4) ~-p -~ M~ ]---> my

    detected by the spectrometer, after selection by the trigger system, amounted to ~ 300 per beam burst. Fast processors have been added to the data acqui- sition system. They allow us to compute the total energy release ~ ----- ~ ~Aj and the first radial moment ~r ~-- ~ ~Ajrj/L of the showers in GAMS (At is the magnitude of the signal in the j-th counter, rj is its distance to the beam axis and ~j is its normalization coefficient). The ~ values are determined ex- perimental ly during the calibration of GAMS in a 25 GeV/e electron beam (~,7) and are stored in the memory of the processors.

    _~1~]~ is proport ional to to M ~ -~ p~, where M and PT are the mass and the transverse momentum of the decaying particle, respectively. The selection of events having radial momenta ~r above ~ 350 MeV eliminates most of pion charge exchange processes =-p-~7:~ (~), without affecting the detection efficiency for ~-mesons. The number of events stored into the computer HP 2] 00 A is thus brought down to ~ 100 events/burst only.

    During the four days of data taking, 3.1011~ - have reached the liquid- hydrogen target leading to the production of 600000 ~-mesons in the u-p --~ ~u reaction. This number is 30 times larger than the statistics in our previous experiment and in all other works which were devoted to the study of decay (1), altogether. The background due to ~ dec~ying into 37:0 and other processes, which may simulate decay (1), has been lowered at the same t ime by more than one order of magnitude leading to a level sensitivity

  • OBSERVATION OF T ITEDECAY ~__~:Oyy 501

    The new reconstruction program looks for showers and evaluates the co- ordinates and energy of each Y by making use of the information gathered on shower shape and fluctuations during the calibration of the spectrometer with electrons of 10 GeV and 25 GeV. Events are sorted out according to the multi- plicity m of 7's emitted in reaction (4). I t has been much improved with regard to the version used in our previous work (1), giving a better separation of near- lying showers. Classification ambiguities are much reduced. Less ~han a few per cent of events need to be transferred between adjacent classes m and re:k1 in the course of the analysis. Technical improvements of the spectro- meter GAMS have brought the min imum energy threshold for y detection down to Ew~ . ----0.15 GeV, resulting in better constraints on decay (2) and a large reduction of the main background.

    Events classified in the class m ~ 4 have been retained for further kinema- tical analysis. Those which satisfy the kinematics of reactions

    (5) rr-p ~ r:~

    (6) :z-p -~ K~A

    or : : -p -~~ (lo), with Z2< 12 (99% confidence level), have been discarded. The compatibi l i ty of the remaining events with the hypotheses M ~ -~v.~ and M ~ -~ ~]yy has been checked. I f the value of g s is smallest and not larger than 7 (97 % confidence level) when the first hypothesis is retained, events are further considered as possible candidates for reaction

    (7) ~-p -~ r~~

    Then the region of small M ~ transverse momenta PT < 0.3 GeV/c has been excluded in order to reduce the background. At the cost of a 20 % loss of good events (3) (5), this allowed us to suppress the contribution of reactions like (5), which, going through one-pion exchange, are concentrated at small PT (11), and the contribution of fortuitous coincidences.

    Contamination due to reactions like (5) and (6), giving ~o pairs in the final state, is further suppressed by selecting events in which the mass corresponding to the pair of noncorrelated y's in (7) is larger than 180 MeV/e ~ only. Of these

    (lO) W. D. APEL, K. H. AUOENSTEIN, E. BERTOLUCCI, S. V. DONSKOV, A. V. INYAKIN, V. A. KACHANOV, R. ABDEL KttALEK, R. : i . KRASNOKUTSKY, ~. KR~.TGER, U. LEDER, A. A. LEDNEV, I. MANNELLI, YU. V. MIKHAILOV, I-I. ~r G. M. PIERAZZINI, YU. D. PROKOSIIKIN, ~V~. QUAGLIA, H. SCHNEIDER, .A. SCRIBANO, F. SERGIAMPIETRI, ]~. S. SItUVALOV, G. SIGURDSSON and M. L. u Natl. Phys. B, 193, 269 (1981). (11) V. N. BOLOTOV, V. V. ISAKOV, D. B. KAKAURIDZE, YU. D. PROKOSHKIN, S. A. SAD0VSKI and G. V. KRAUST0V: Yad. Eiz., 24, 106 (1976).

  • 502 ] I I~P- I ISN-LAPP COLLABORATION

    reactions, the most tr icky source of background is certainly (6) (in the case in which the K ~ decays at a few metres from the target). The detection of the decay products of the recoil hyperons by the counters of the guard system, which surrounds the Larget in the set-up, allowed us to reduce the contribu- tion of reaction (6) to the level of a few percents of the main process (3).

    The min imum value of energy threshold Eth ~. has only been used in the y reconstruction program in order to reduce as much as possible the main background which comes from decay (2), when any two y's are being missed. One drawback of such u low threshold is that the program sometimes finds spurious low-energy y's generated by the noise of the counters and of the elec- tronics of the spectrometer. At the stage of the kinematical analysis, a quite higher-energy threshold for the y's, Eta,, has been used. By increasing Eth , the above-mentioned instrumental background decreases fast.

    4. - Mass spectra.

    The effective mass spectrum of the ~oyy system, constructed from the events which have passed through the selection procedure described above, is shown in fig. 2a). A narrow peak, with a statistical significance exceeding 7 standard deviations, shows up in the region of the ~ mass.

    This peak has been fitted with a Gaussian distribution F(M- -M, a~), the parameters _M and ~ and the normalization being left free. The back- ground part of the spectrum, which is bell-shaped with a max imum near 500 l~eV, has been fitted with a polynomial function Fb(M). This procedure provides a statistically significant description of the measured spectrum as :~2/nDgl (e.g. fig. 2a) and b)).

    The evaluated ~]I value is found to be independent of the threshold energy E~ value between 0.5 GeV and 3.5 GeV. ~t is equal to (549~=2)MeV/c 2, a value which agrees within 0.2 ~o with the tabulated ~) mass.

    The region of the mass spectrum around i GeV/c 2 has been studied to check the l inearity of the spectrometer and the accuracy of its calibration. Here too a peak appears in the mass spectrum of the 7:~ system. I t is due to the rare decay ~'-+o)$-+~:~ (BR ~ 2.10 -3) (2) of ~'(958) mesons produced in reaction (4). The final-state topology is the same as th,~t of decay (]). The position of the ~' peak agrees to 0.4 ~o with the value given in the tables for the ~' mass. The measured value of the Ks ~ mass in reaction (6) is also in good agreement (0.3 ~o) with the tables, which demonstrates the accuracy of the mass scale of GAMS.

    The width of the ~ peak shows also no dependence on the y energy threshold Eth. The value aM--~ (7.2=[:1.5)MeV/e 2 obtained in the fit ugrees with the intrinsic resolution of the apparatus (8.5 h[eV/e~).

  • OBSERVATION OF THE D~CAY ~--~r@yT 503

    30

    f/

    20

    b)

    510 i

    55O 590

    40

    N

    20

    a)

    0 I i &00 500 600 70O

    M~ol (M eV/c z)

    Fig. 2. - a) Mass spectrum of the n0yy system measured in reaction (7). The arrow points to the ~] mass. The dashed curve is a polynomial function ~b which describes the background part of the spectrum, b) Mass spectrum in the region of the ~-meson obtained from a) after subtraction of the background distribution ~b (dashed curve). The dotted curve is the Gaussian distribution 17,(M--_~r, a~) obtained with the best estimates of the parameters, after fitting 2~ = 549MeV/e ~, a M = 7.2 MeV/c% The measured spectrum has been fitted by a superposition of /7 and /~b with z~/n, : 0.7.

    5. - S imulat ion procedure .

    A s imulat ion procedure based on a bank of measured showers has been developed to determine the reg is t ra t ion efficiency for decay (1) and the value of BR(~ 1 _ .~oyy) : F(~ --* ~~ --* all).

    The 7's f rom the decay B --* 27, which are detected in the spect rometer together with those produced in decay (]) , are well separated spat ia l ly . They

  • 504 II-IEP- I ISN-LAPP COLLABORATIO I~

    have been used as a source of real individual showers. The set of amplitudes A; in the (~ fired ~) cells of showers produced by 150000 y's of various energies and co-ordinates have been stored in a spectrometer data bank.

    In a first stage, events produced in reaction (3) have been generated by the method of Monte Carlo by using the known t-dependence of the differential cross-section (5). The phase-space distribution has been assumed to be uniform. Further, the information stored in the bank of showers has been used for each y entering GA~S. After, the...