Data are presented on the reaction p̄p → ωωπ at rest from the Crystal Barrel detector. These data identify a strong signal due to f2(1565) → ωω. The relative production from initial p̄p states P2, P1 and S0 is well determined from ωω decay angular correlations; P-state annihilation dominates strongly. A combined fit is made with data on p̄p → 3π at rest, where f2(1565) → ππ is observed. A Flatté formula is fitted to the f2(1565), including the s-dependence of decay widths to ωω and ρρ. The data then determine the K-matrix mass, M = 1598± 11(stat)± 9(syst) MeV. The decay width to 2π is very small, of order 2% of the total width. We investigate the f2(1565) through its decays to ωω in the process p̄p → ωωπ at rest. Data on p̄p → 3π at rest are also used to estimate relative widths to ππ and ωω. The first clear identification of the f2(1565) was by May et al. in the Asterix experiment on p̄p → πππ at rest [1,2]. Subsequently, it has been identified by the Obelix collaboration in n̄p → πππ [3], and in Crystal Barrel data on p̄p → 3π at rest [4,5]. However, it has become clear that the f2(1565) also couples strongly to ωω (and therefore probably to ρρ). Abele et al. [6] found a strong cusp in the ππ D-wave at the ωω threshold in the 3π data. This led to the identification of f2(1565) as the same 2 + resonance as observed by GAMS [7] and VES [8] in ωω just above threshold. Their results are listed under f2(1640) by the Particle Data Group (PDG) [9], who also list other sightings there and under f2(1565). The properties of this resonance have remained elusive. Here, we show that it is important to fit with a Flatté formula which includes the s-dependence of decays to ππ, ρρ and ωω; there is a sharp cusp at 1564 MeV due to the opening of the ωω channel. In outline, the elements in the analysis are as follows. Firstly, ωω correlations in the Crystal Barrel data on the ωωπ channel determine well the relative amounts of annihilation from initial states S0, P1 and P2. Secondly, earlier Crystal Barrel data on the 3π 0 channel determine well the lower edge of the resonance. Both of these sets of data are readily fitted with a Breit-Wigner amplitude which includes the s-dependence of widths for decays to ρρ, ωω and ππ. However, Now at IHEP, Beijing 100039, China

Upper limits for the production of new light gauge bosons in 0 , and 0 decays have been obtained with the Crystal Barrel detector. The kinematically well-constrained reactions pp ! 0 0 P have been studied, where P 0 ; ; 0 decays through the emission of a single photon recoiling against a missing state X. X can be a long-lived weakly interacting particle or X !. This paper gives full detail of the analysis of our previous work and extends the search to 0 decay. No signal has been observed and branching ratio upper limits (90 % C:L:) of 6 10 ?5 have been obtained for masses of the gauge boson lying between 65 MeV and 125 MeV (0 decay), 6 10 ?5 , for X masses between 200 MeV and 525 MeV (decay), and 4 10 ?5 , for X masses between 50 MeV and 925 MeV (0 decay). The 0-decay limit represents a factor of 4 to 8 (depending on the X-mass) improvement when compared to the existing limit, whereas the and 0 decay limits have been measured for the rst time, thereby extending the M X range from 125 MeV up to 925 MeV.

We have determined branching ratios for antiproton annihilations at rest on protons or neutrons in liquid deuterium which we compare to frequencies of isospin-related processes in antiproton-proton annihilations. Using the annihilation rates into 0 0 and ? 0 where the annihilation took place on the proton or neutron, respectively, we discuss the fraction of S-wave and P-wave annihilation in liquid D2. The frequencies for ? ! and ? 0 , and ? and ? 0 and the corresponding frequencies for pp annihilations are used to determine isoscalar and isovector contributions to the protonium wave function. The isospin decomposition of the annihilating pp system in the 3 S1 or 1 S0 state is consistent with both, pure pp initial wave function and with the predictions of NN potential models. For the 3 P0 state of the pp atom we nd consistency with a pure pp system at annihilation while NN potential models predict large nn contributions. We observe-! interference in pp! + ? and + ? 0 annihilation which we compare to-! interference in e + e ? annihilation. The interference patterns show striking similarities due to similar phase relations; the interference magnitude depends on the !-production ratio. The similarity of the phase in all 3 data sets demonstrates that isovector and isoscalar parts of the protonium (pp atomic) wave function are relatively real, again in connict with NN potential models. The annihilation rate for pd!K ? K 0 p connrms the dominance of the isovector contribution to NN! KK annihilations. No complications due to initial state interactions are required by the data. Furthermore, we searched for narrow quasinuclear bound states close to the NN threshold, also predicted by NN potential models, but with negative outcome. We conclude that NN potential models are not suited to provide insight into the dynamics of the annihilation process.

Antiproton-proton annihilation into 0 has been studied with the Crystal Barrel spectrometer at CERN at an incident beam momentum of 1.94 GeV/c. The data were taken with a trigger requiring neutral nal states. A new isovector state, the a2(1660) decaying to 0 , is observed. In the invariant mass region around 2.1 GeV=c 2 , strong production of a heavy resonance is required, but our analysis does not distinguish between J P = 0 + , 2 + and 4 +. The production of the f0(1500) in reactions in ight is also observed.

The annihilation channels (p) over bar p-->omega pi(o), omega eta, omega eta' were studied with the CRYSTAL BARREL detector at LEAR at (p) over bar-momenta of 600, 1200, and 1940 MeV/c. In most cases angular distributions were measured which allowed a complete J(P)-analysis using the helicity formalism. The contribution of all relevant initial states could be determined. The maximal contributing angular moments are dependent on the (p) over bar-momentum and range up to J = 5.

A high-statistics data sample on (p) over bar p --> pi(+)pi(-)pi(0) is studied to deduce the mass, width, and Line shape of charged and neutral rho mesons. We find that the line shape is well described by a relativistic Breit-Wigner function. We find mass degeneracy of the rho(+/-) and rho(0) masses when mixing with the omega in its pi(+)pi(-) decay mode is taken into account. A fit with one mass and width for the three rho charge states yields m(rho) = (763.0 +/- 1.3) MeV and Gamma(rho) = (149.5 +/- 1.3) MeV. The mass difference between charged and neutral rho's is determined to be (1.6 +/- 0.6 +/- 1.7) MeV. A possible influence on the rho mass due to the isobar approximation is discussed. (C) 1999 Published by Elsevier Science B.V. All rights reserved.

.   Antiproton–proton annihilation into has been studied with the Crystal Barrel spectrometer at CERN at an incident beam momentum of 1.94 GeV/c. The data were taken with a trigger requiring neutral final states. A new isovector state, the decaying to , is observed. In the invariant mass region around 2.1 GeV/, strong production of a heavy resonance is required, but our analysis does not distinguish between , and . The production of the in reactions in flight is also observed.

Data are presented on the reaction p̄p→ωωπ0 at rest from the Crystal Barrel detector. These data identify a strong signal due to f2(1565)→ωω. The relative production from initial p̄p states 3P2, 3P1 and 1S0 is well determined from ωω decay angular correlations; P-state annihilation dominates strongly. A combined fit is made with data on p̄p→3π0 at rest, where f2(1565)→π0π0 is observed. A Flatté formula is fitted to the f2(1565), including the s-dependence of decay widths to ωω and ρρ. The data then determine the K-matrix mass, M=1598±11(stat)±9(syst) MeV. The decay width to 2π is very small, of order 2% of the total width.

We report the first observation of antiproton annihilation on the deuteron into two-body channels with open strangeness. These reactions are of the Pontecorvo type, designating annihilations kinematically not possible on a free nucleon. With antiprotons stopped in a liquid deuterium target, a KS and Λ enriched sample of events, corresponding to about 109 annihilations, was recorded with the Crystal Barrel experiment, using a Silicon microstrip detector for triggering on secondary decays. The final states ΛKS and Σ KS, with Σ 0 →Λγ, Λ→pπ−, KS→π +π−, and the equivalent final states ΛKL and Σ KL, with noninteracting K 0 L, were identified. The pd branching fraction into ΛK0 and Σ0K0 are (2.35± 0.45)× 10−6 and (2.15± 0.45) × 10−6, respectively. This similarity disagrees strongly with two-step model predictions and supports the statistical (fireball) model. Annihilation reactions of antinucleons on atomic nuclei that demand the participation of more than one nucleon were envisaged by Pontecorvo [1] shortly after the discovery of the antiproton. By energy and momentum conservation, these so-called Pontecorvo reactions cannot occur on a free nucleon. Antiproton annihilation on the deuteron into a single meson and a baryon is the prime example, elaborated in a variety of theoretical papers (see, e.g.[2-7]). Experimentally, only a few pd Pontecorvo channels have been clearly identified: pπ− [8,9], nπ, nη, nω [10] and ∆(1232)π [11]. The branching ratios of pd (at rest) into these channels, in the range (0.3-3.0) ×10, did not allow a distinctive evaluation of the different theoretical approaches. The models fall into two classes using either the two-step or the fireball picture, which are extremes with respect to the involved degrees of freedom. The former [2-4] assumes annihilation of the antiproton on one of the nucleons in the deuteron into a pair of mesons and subsequent absorption of one of them on the other nucleon. The intermediate meson is highly virtual. The latter model 1 Now at University of Mainz, Mainz, Germany 2 Now at University of Freiburg, Freiburg,Germany 3 University of Ljubljana, Ljubljana, Slovenia 4 University of Silesia, Katowice, Poland 5 Now at SLAC, Stanford, USA 6 Now at University of Bonn, Bonn, Germany 7 Now at Uppsala University, Uppsala, Sweden 8 This work is part of the PhD thesis of C. Regenfus at the University of Munich

A study of antiproton annihilation in liquid deuterium into pi(+)pi(-)pi(-) and a spectator proton is presented. For a long time this reaction resisted a description by final state interactions which is surprising (and disturbing) given the success of the final state interaction model in other annihilation reactions. It is shown that the introduction of p(1450) is essential to Set a reasonable description of the measured Dalitz plot. This resonance was never tried in previous attempts to understand this data. A possible isospin-2-pi-pi S-wave contribution was tested, but no evidence was found for such a contribution. (C) 1999 Elsevier Science B.V. All rights reserved.

We report the first observation of antiproton annihilation on the deuteron into two-body channels with open strangeness. These reactions are of the Pontecorvo type, designating annihilations kinematically not possible on a free nucleon. With antiprotons stopped in a liquid deuterium target, a K0S and Λenriched sample of events, corresponding to about 109 annihilations, was recorded with the Crystal Barrel experiment, using a Silicon microstrip detector for triggering on secondary decays. The final states ΛK0S and Σ0K0S, with Σ0→Λγ, Λ→pπ−, K0S→π+π−, and the equivalent final states ΛK0L and Σ0K0L, with noninteracting K0L, were identified. The pd branching fraction into ΛK0 and Σ0K0 are (2.35±0.45)×10−6 and (2.15±0.45)×10−6, respectively. This similarity disagrees strongly with two-step model predictions and supports the statistical (fireball) model.

A partial wave analysis is presented of two high-statistics data samples of protonium annihilation into pi(0)pi(0)eta in liquid and 12 arm gaseous hydrogen. The contributions from the S-1(0), P-3(1) and P-3(2) initial atomic fine structure states to the two data sits are different. The change of their fractional contributions when going from liquid to gaseous H-2 as calculated in a cascade model is imposed in fitting the data. Thus the uncertainty in the fraction of S-state and P-state capture is minimized. Both data sets allow a description with a common set of resonances and resonance parameters. The inclusion of a pi eta P-wave in the fit gives supportive evidence for the <(rho)over cap>(1405). with parameters compatible with previous findings. (C) 1999 Elsevier Science B.V. All rights reserved.

The rho mass, width and line shape in p(over)-bar-p annihilation at rest into pi(+)pi(-)pi(0)