Section 2

2 Research with Antiprotons Hadron Spectroscopy and Hadronic MatterOverviewThis section addresses the many unexplored properties of the strong interaction in the subnuclear regime. It describes a proposed physics program to study the structure of hadrons and their interaction with the nuclear medium, using cooled antiproton beams of unprecedented quality and intensity in the momentum range from 1.5 15 GeV/c stored in the High Energy Storage Ring (HESR). The accepted theory of the strong interaction is Quantum Chromodynamics (QCD). This theory is extremely successful in describing phenomena at high energies, where the interaction among quarks by gluon exchange can be treated in a highly accurate approximation by perturbation theory in close analogy to Quantum Electrodynamics. In the energy regime of interest here, hadrons bound systems of quarks and gluons become the relevant degrees of freedom. The complementarity between QCD and hadronic descriptions provides an ideal paradigm for one of the central issues of present day physics: the description of complex systems. In nearly all cases, different concepts are introduced for different levels of complexity and the challenge is to understand their relationship. For a number of processes both formalisms are applicable, allowing for a systematic study of how quark-gluon and hadronic degrees of freedom are related. At low energies, the perturbative treatment of QCD can no longer be applied. We enter the regime of strong coupling where astonishing new phenomena arise which represent open problems to date Free quarks have not been observed, they are confined within hadrons. The mass of hadrons is much larger than the mass of their quark constituents. Because of the characteristic self-interaction among gluons, glueballs and hybrids should exist, i.e. states which consist predominantly of gluons or of glue and a quark-antiquark pair.

All these phenomena are long-standing puzzles; they have their origin in the specific properties of the strong interaction, and represent a major intellectual challenge in our attempt to understand the nature of this fundamental force and the structure of multi-particle systems bound by this interaction. Experimentally, studies of hadron structure can be performed with different probes such as electron or cooled antiproton beams, each of which have their specific advantages. Electrons are pointlike objects and interact with hadrons via the well known but weak electromagnetic force. In antiproton-proton annihilation, particles with gluonic degrees of freedom as well as particle-antiparticle pairs are copiously produced, allowing spectroscopic studies with unprecedented statistics and precision. Antiprotons of 1 15 GeV/c will therefore be an excellent tool to address the open 193

Section 2 problems mentioned above. Vivid discussions within the hadron physics community in Europe and overseas have been instrumental in shaping a physics program which encompasses the following research directions: Charmonium ( cc ) spectroscopy: precision measurements of mass, width, decay branches of all charmonium states, especially for extracting information on the quark confinement. Firm establishment of the QCD-predicted gluonic excitations (charmed hybrids, glueballs) in the charmonium mass range (3 5 GeV/c2). Search for modifications of meson properties in the nuclear medium, and their possible relationship to the partial restoration of chiral symmetry for light quarks. Particular emphasis is placed on mesons with open and hidden charm in order to learn more about the origin of hadron masses. Precision -ray spectroscopy of single and double hypernuclei for extracting information on their structure and on the hyperon-nucleon and hyperon-hyperon interaction.

With increasing luminosity of the HESR facility further possibilities will emerge: D-meson decay spectroscopy (rare leptonic and hadronic decays). Search for CP-violation in the charm and strangeness sector (D meson decays, system). Extraction of generalized parton distributions from p p annihilation. Fundamental physics with stopped antiprotons.

With the quest for the origin of hadron masses and the study of hypernuclei, this physics program is intellectually interrelated and complementary to other parts of the proposal which address nucleus-nucleus collisions and nuclear structure studies. To prepare an antiproton beam with small energy spread is a technological challenge as it requires electron cooling at electron energies of several MeV. Currently, a collaborative effort with Fermilab and Novosibirsk is being prepared. The instrumentation for the research with antiprotons comprises a multi-component general purpose detector (Figure 2.1), characterized by high resolution charged particle tracking, neutral and charged particle identification over a broad momentum range, excellent mass-, momentum- and energy resolution, high rate capability and a sophisticated, fast trigger scheme. The costs for the detector are estimated to be 28 M, including 13 M for the most costly component, the electromagnetic calorimeter. The project of a high energy antiproton storage ring is strongly supported by groups who have been active in this field at CERN (LEAR) and Fermilab (E831). Because of the strong overlap with the ongoing hadron and nuclear structure program at GSI the community interested in this project has grown and is estimated to include 300 194

Section 2 scientists, mainly from Europe with particularly strong participation from Italy and Germany.

Figure 2.1. Artist's view of the universal detector system for experiments at the internal target of the antiproton storage ring. It allows the detection and identification of neutral and charged particles generated over the relevant angular and energy range. This task will be shared by the combination of a central and a forward spectrometer of modular design which are optimized for the specific kinematics of the antiproton annihilation process.

In comparison with other facilities, the physics opportunities outlined in the present proposal go far beyond the earlier SUPER-LEAR concept and are complementary to the physics program at the planned Japanese High Intensity Hadron Facility which is focused on kaon- and neutrino beams. The search for gluonic excitations at HESR is complementary to the corresponding program at the proposed 12 GeV upgrade at Jefferson lab which due to the accelerator energy is limited to light quark hybrids. There is partial overlap with research at BES and the D meson physics program proposed at Cornell where - being e+e- colliders - studies of in-medium meson properties can, however, not be performed. With the realization of the HESR-project, GSI will thus play a pioneering and unique role in the experimental exploration of long distance (non-perturbative) QCD and the structure of hadronic matter. GSI has a distinguished history of many important contributions to the physics of the strong interaction. This program will enable GSI to play an equally significant role in the future. 195

Section 2

2.1 Science case2.1.1 Introduction Until the end of the 19th century atoms were considered to be the smallest building blocks of matter. The physics of the 20th century has revealed that atoms are not elementary as their name suggests ( = indivisible) but rather consist of an atomic nucleus surrounded by a cloud of electrons. The nucleus itself is built of nucleons, positively charged protons and neutrons which are electrically neutral. Since the late 1950s we know that nucleons also have a substructure. They themselves are composed of elementary particles later named quarks. Together with the family of leptons - to which the electron belongs - these particles are considered today to be the elementary building blocks of matter. Matter thus manifests itself in a hierarchy of composite systems. The stability of each system is determined by the interaction among the respective constituents: Atoms are held together by the electromagnetic force between electrons and the protons in the atomic nucleus; nuclei are bound due to the attractive strong interaction among the nucleons. In both cases, the mass of the composite system is governed by the masses of the constituents with some small correction due to the binding energy. In case of the nucleon, two completely new aspects are encountered: the bare masses of the constituents (quarks) contribute only a few percent to the nucleon mass. The latter mass is almost entirely determined by the strong interactions of quarks and gluons in the nucleon. Secondly, quarks have never been observed as isolated free particles; they are confined inside hadrons. Both phenomena, the origin of mass of strongly interacting, composite systems and the confinement of quarks and gluons are long-standing puzzles and represent the key challenge in our quest to understand the nature of the strong interaction. This chapter describes theoretical concepts and experimental approaches to address these fundamental questions, making use of antiproton beams of unprecedented intensity and quality which will be provided by the new GSI facility. In high energy physics, one studies the question how the masses of the elementary quarks and leptons are generated through the so called Higgs-mechanism: particles acquire an effective mass by an interaction with the Higgs field which may manifest itself in the existence of the Higgs-particle. This particle is and will be intensively searched for at high energy accelerators like LEP, LHC (CERN) and TEVATRON (Fermilab). In hadron physics, the physics of strongly interacting composite systems, the masses of composite particles like nucleons are attributed to the strong interactions among their quark constituents through gluon fields (see Figure 2.2). The underlying theory is Quantum Chromodynamics (QCD). QCD is conceptually elegant and deceptively simple. It generates an enormous richness and complexity of phenomena and forms of matter. They range from all kinds of hadronic and nuclear species, systems like neutron stars and black holes to the quark-gluon plasma, a state of matter in the early universe. The physics of strong interactions is undoubtedly one of the most challenging and fascinating areas of modern science. 196

Section 2

Figure 2.2. Schematic view into the nucleon, illustrating its complex structure. The three valence quarks of the nucleon are held together by exchanging gluons (symbolized by springs) which for tiny moments split into pairs of so called virtual quarks and antiquarks.

QCD is well understood only at short distance scales, i.e. much shorter than the size of a nucleon (10-15m). Here the basic quark-gluon and qq interaction is characterized by a small coupling strength parameter s 0.1. In this case, perturbation theory is applicable, a well established calculational technique, yielding results of high precision and predictive power. Indeed, for high energy processes, perturbative QCD provides a quantitatively successful description of strong interactions, and it is here that one finds up- and down quark masses of only 1% or less of the nucleon mass. Similar conditions are encountered in the interior of hadrons at distances much smaller than the size of the nucleon. At distance scales comparable to the nucleon radius, however, a new physics regime is entered and other theoretical approaches are required; at these scales the interaction between quarks and gluons becomes so strong that further separation of single quarks from their partners is inhibited. This unusual behavior is related to the fact that gluons do not only interact with quarks but also among themselves, forming a kind of spring which holds the quarks together like a rubber band. This self-interaction among gluons appears to be the origin of quark confinement. It is one of the major challenges to understand this phenomenon not only qualitatively but quantitatively in the framework of strong interaction theory. At these distances, as shown in Figure 2.3, s approaches 1 and perturbative methods are no longer applicable. We enter the field of non-perturbative QCD, the regime of quark confinement where we can no longer speak of bare quarks. What appears at first sight to be a complication is actually a profound intellectual challenge which has stimulated rapid developments of different theoretical methods. Two approaches are being actively pursued: in one, full numerical solutions of Quantum Chromodynamics are obtained by simulating the theory on the most powerful existing computers. The quark and gluon fields are defined at discrete points 197

Section 2 in space and time forming a four-dimensional lattice, thus the name Lattice Quantum Chromodynamics (LQCD). The other approach uses effective field theories which retain the fundamental symmetries of QCD but - instead of quarks and gluons - use composite hadronic systems like baryons (made of three quarks) and mesons (made of a quark and an antiquark) as the relevant degrees of freedom. These effective theories provide a connection between QCD and the more phenomenological treatments developed over past years to account for observations in hadron and nuclear physics.

Figure 2.3. Coupling strength between two quarks as a function of their distance. For small distances ( 10-16 m) the strengths s is ~0.1, allowing a theoretical description by perturbative QCD . For distances comparable to the size of the nucleon, the strength becomes so large (strong QCD) that quarks can not be further separated: they remain confined within the nucleon. (The data are adapted from [1])

One of the fundamental symmetries of QCD is chiral symmetry. In the limit of vanishing quark masses this symmetry prevents a change of the chirality or handedness of quarks when they interact by gluon exchange (particles are called right (left) handed when their spin is parallel (antiparallel) to their direction of motion). However, the strong interactions of quarks and gluons result in a spontaneous break down of this chiral symmetry; in this process, almost massless "bare" quarks turn into massive "constituent" quarks. The mass of hadronic systems is thus intimately related to the spontaneous breaking of chiral symmetry. The program of understanding this symmetry breaking pattern is complementary to the compressed baryonic matter part of the proposal where the density dependence of chiral symmetry breaking is addressed by compressing nuclear matter to very high densities in heavy-ion collisions. In contrast, using antiprotons as projectiles, energy is transferred to the nuclear system without compressing it. The results obtained at normal nuclear matter density will thus provide the reference point for the heavy-ion studies. A full understanding requires investigations over the full density range. Although hitherto not used as probes at GSI, antiproton beams will thus become an 198

Section 2 integral part of the future GSI program devoted to unravelling the properties of the strong interaction. Apart from the up and down quarks which are the constituents of the nucleon, four additional quark flavors are known: charm, strange, top, and bottom. The bare masses of these quarks implement a hierarchy of scales with vastly different phenomena. The sector of the lightest (up and down) quarks, with masses smaller than about 1% of the nucleon mass, is governed by chiral symmetry and its spontaneous breaking. The heavy quark limit, on the other hand, involves different, heavy quark symmetries and the corresponding spectroscopy is characterized by the almost static behavior of the very massive bottom quark which is almost five times heavier than the nucleon. The top quark is too short lived to form any bound state. The dynamics of hadronic bottom states is dominated by confining properties of the QCD vacuum: two quarks interact more and more when they are pulled apart; gluonic flux tubes connect the static quarks and produce a confining potential. The masses of strange (s) and charmed (c) quarks, ms 0.1 mN and mc 1.4 mN , are intermediate between the chiral and heavy quark limits. What makes the physics of strange and charmed quarks so interesting is that they interpolate between the limiting scales of QCD. In particular, the very important domain of charmed quarks with its connection to gluon dynamics and the confinement problem is still not understood. The present proposal focuses in large part on the outstanding physics issues in this area. The situation calls for new experimental and theoretical efforts to study systems with c quarks and to explore QCD in its nonperturbative regime. As illustrated below, antiproton beams are an excellent tool in this field of research. The physics program with antiproton beams at the High Energy Storage Ring (HESR) offers a broad range of investigations from studies of Quantum Chromodynamics to the test of fundamental symmetries. In annihilation reactions between antiprotons and protons or nucleons in nuclei a whole wealth of particles is generated whose properties and interactions can be studied. The proposed charmonium ( cc ) spectroscopy program using p p annihilation is an extension of successful experiments performed recently at the Fermilab antiproton accumulator. Advanced p - cooling techniques and a more versatile detector set up will be employed allowing for the first time the measurement of both electromagnetic and hadronic decays. One goal is to make comprehensive measurements of the spectroscopy of the charmonium system and hence provide a detailed experimental study of the QCD confining forces in the charm region to complement theoretical investigations in the framework of LQCD. Recent experiments at LEAR/CERN have demonstrated that particles with gluonic degrees of freedom are produced with high probability in p p annihilation. One of the central parts of the program is the first search for gluonic excitations in the charmonium sector and the continuation of the experiments to tackle the problems of glueballs, including highly excited ones with exotic quantum numbers. Glueballs are 199

Section 2 massive hadrons predominantly built of massless gluons. Their mass is solely determined by the interaction among the constituents and is thus especially sensitive to the self-interaction among gluons which is the key for confinement. A study of the spectrum and subsequently deeper understanding of the gluonic modes would shed light on the confinement problem in QCD. Furthermore, the proposed experimental program at the HESR addresses open problems of in-medium modifications of hadrons with charmed quarks in nuclei and the interaction of these hadrons with nuclei. This is, on the one hand, an extension of the present chiral dynamics studies with partial restoration of chiral symmetry in hadronic environment, from the light quark to the open charm quark sector. On the other hand, this program is focused on the first experimental studies of the charmonium-nucleon and charmoniumnucleus interaction which is also of basic importance for ultra-relativistic heavy-ion collisions. Replacing an up- or down quark by a strange quark in a nucleon, which is bound in a nucleus, leads to the formation of a hypernucleus. A new quantum number, strangeness, is introduced into the nucleus, adding a third, almost unexplored dimension to the nuclear chart. This program opens new perspectives for nuclear structure studies and is thus a novel supplement to the proposal to study the structure of nuclei with exotic beams. The nucleon with an up- or down quark replaced by a strange quark (hyperon) is not restricted in the population of nuclear states as neutrons and protons are. Using antiproton beams, nuclei with even several strange quarks may be produced. These exotic nuclei offer a variety of new and exciting perspectives in nuclear spectroscopy and for studying the forces among hyperons and nucleons. CP symmetry implies that the laws of physics remain unchanged after replacing particles by their antiparticles and simultaneous reflection in a spatial mirror. If CP symmetry were to hold strictly, matter and antimatter in the universe would have completely annihilated each other and neither stars nor human beings would exist. The observed prevalence of matter over antimatter in the universe requires CP violation, an effect first observed directly in the decay of neutral kaons and very recently also in B mesons. With the HESR storage ring running at full luminosity, CP violation can be studied in the charm meson sector and in hyperon decays. The expectation is that an observation of significant CP violation in the charm sector would indicate physics beyond the Standard Model of particle physics. For the major part of the physics program a general purpose detector as shown in Figure 2.1 will be used. It allows the detection and identification of neutral and charged particles over the relevant angular and energy range. The inner part of the detector can be modified for the experiments with strange hypernuclei or the special needs of CP violation studies. With the realization of the HESR-project, GSI will play a pioneering role in the experimental exploration of strong (non-perturbative) QCD. 200

Section 2 In the following an experimental program using an antiproton beam of high quality and intensity is presented. Four experimental fields (2.1.2-2.1.5), which are foreseen for the first years of machine operation are highlighted. An overview of future options is given in chapter 2.1.6. A more detailed description of the physics motivations and of technical details can be found in the Letter of Intent [2] and in Ref. [3]. 2.1.2 Charmonium spectroscopy The fundamental understanding of strong interactions in terms of QCD was greatly stimulated by the 1974 discovery of J/, the vector state (JPC = 1) of charmonium, the system of a charm quark and a charm antiquark ( cc ). The charmonium system (Figure 2.4) has been considered a powerful tool for the understanding of the strong interaction and this is the reason why this system has often been called the positronium of QCD. It is charmonium spectroscopy where the potential models of quark-(anti) quark interactions are tuned. It is where the gluon condensate, which represents the energy density of the QCD vacuum, is determined. And it is the attenuation of charmonium in the presence of quark-gluon plasma (QGP) which may provide the unerring signal in heavy ion collisions for the existence of this unique state of primordial matter. The extraordinary versatility of charmonium spectroscopy derives from the fact that while in principle the qq interaction can be studied with quarks of any flavour, with charmonium cc a number of unique advantages are realized. As is well known, light quark (u, d, s) systems are of great complexity. Nearly 100 states, with widths ranging from 100 MeV to 400 MeV, are known to exist in the mass interval from 1 to 2 GeV. The strong coupling constant s is so large (> 0.7) as to rule out the use of perturbation theory, and relativistic effects are formidable. At the other extreme is the bottonium ( bb ) system. However, high precision spectroscopic measurements using antiprotons would require beams of such high energies (up to 58 GeV), such high energy resolution (< 100 keV), and such high intensity (several orders of magnitude larger than currently achievable), as to be essentially impossible in any near future. Charmonium is almost free from all of the above-mentioned problems. It has only eight bound states in a ~ 0.8 GeV mass interval. The states have small widths ( 20 MeV) and are well resolved. The coupling constant, s 0.3 is not too large and relativistic problems are considered manageable (< v2/c2 > 0.2). Of particular interest is the interplay of perturbative and non-perturbative effects: the charm quark mass is large enough to justify the application of perturbative QCD, but is not sufficiently large to suppress non-perturbative corrections.

201

Section 2Mass [MeV/c2] momentum p [GeV/c]

4100

(4040)

8.0* D*D

thresholds

39001

DD*

7.1

(3770) 3700 c (2 1S0) (2 S1)3

D2(2 )

+

3

D2(2 )DD

6.3 h c (1P1) 2 1 0

3500

e+e

3

PJ

5.5

3300

4.8

3100

c (1 1S0 )

3 J/ (1 S1 )

4.1

e+e

2900

0

+

1

1+

(0,1,2)

++

J

PC

3.4

Figure 2.4. Charmonium states and their decay modes. Undiscovered and poorly known states are marked by dashes.

Charmonium spectroscopy was extensively studied at e+e- colliders, notably at SLAC, Cornell, ORSAY and DESY during 19741980. While a number of important discoveries were made, the technique of studying charmonium via e+e- annihilations had important limitations. Only the vector states, JPC = 1 (J/, ') could be directly formed, all other states had to be produced by radiative transitions from J/ and/or ', with the consequent limitations in precision due to poor yield and poor energy resolution of the photon detector. The masses of several states were determined well, but the widths were generally poorly determined, if at all. Many decay channels of vector states were studied (e.g., 130 decay channels of J/), but less than 20 decay channels were studied for any other state, with far fewer (< 30 % of these) having branching factors determined with errors 30%. A good amount of discovery physics was done, but precision spectroscopy was not!

202

Section 2

Figure 2.5. Comparison of measurements of the c1 resonance obtained by e+e- annihilation (Crystal Ball, SLAC) and by pp annihilation (E835, Fermilab).

In order to surmount the intrinsic limitations of the e+e- annihilation experiments, a new technique was needed. It came in the form of the pioneering experiment, R704 at CERN, which demonstrated that p p annihilation was not only a viable tool for the study of charmonium spectroscopy, but had significant advantages in precision and versatility. The technique was further refined and exploited with great success by experiments E760/E835 at Fermilab. There are two primary reasons for the superiority of the p p annihilation technique for the study of charmonium spectroscopy. The first reason is that because pp annihilation must proceed via two or three gluons, it can lead to the direct formation of charmonium states of all JPC, unlike e+e- annihilation which is limited to the direct formation of only 1 states. A comparison between both techniques is shown in Figure 2.5 for measurements of the c1 state. The superiority of p p annihilation over e+e- annihilation is very obvious. This means that the precision achievable in the determination of masses and widths of all charmonium states depends only on the quality of the antiproton beam and target, and not on any detector properties. The second advantage comes about because antiproton beams can be cooled (stochastic and/or electron cooling) to obtain momentum resolution of less than 1 part in 105. Thus, the beam energy resolution can be translated directly to mass resolution when very thin targets of hydrogen gas jet or hydrogen pellets are used in the stored circulating beam of antiprotons. Fermilab experiments have successfully demonstrated these unique advantages of the pp annihilation technique for precision spectroscopy of charmonium. An example is shown in Figure 2.6 of how these technical advantages translate into precision in the determination of an important physics parameter, the two photon width of the 2 state of charmonium.

203

Section 2

Figure 2.6. Comparison of two-photon partial width measurements for the 2 state. The hatched areas represent theoretical predictions.

The Fermilab experiments E760/E835 have finished their last data taking in the year 2000, and do not plan to continue, being limited by the p energy limit ( 9 GeV), detector capability (inability to detect and identify charged hadrons), and lack of priority in Fermilabs plans for the future. This proposal not only has in mind to continue the Fermilab charmonium program, but to build and improve on it. It is intended to do so with higher energy p beams (15 GeV), higher luminosity (up to 21032 cm-2s-1), better cooling (stochastic + electron), and a state-of-the-art hermetic detector for both electromagnetic and charged particles. Although the Fermilab experiments improved the precision of many earlier e+emeasurements of charmonium states, and provided many new measurements of widths and branching fractions, many open questions remain which will be addressed by the GSI charmonium program. These open problems exist both below and above the DD threshold at 3.73 GeV. We list them below, and describe a few of them in some detail. The ground state of charmonium, c(11So) It is disappointing how little is known about the ground state of charmonium, c(1 1So). To this date its mass remains uncertain by more than 2 MeV, with the latest (post 2000) measurements disagreeing by more than 5 MeV [4]. The width of c is even more uncertain. Two recent experiments report a width (c) 24 MeV, nearly twice as large as was previously believed [4]. It is important to know the width of c, because a width as large as 25 MeV is essentially impossible to reconcile with simple quark models, and it has been suggested that instanton effects may be responsible [5]. Additional hints for possible instanton effects come from the unusually large 204

Section 2 branching ratios of c decays involving multiple kaons and pions. Clearly, a precision measurement of c mass, width, and branching ratios is of the utmost importance, and it can only be done by direct formation experiments with p p annihilation. It is worth noting that given the same luminosity in e+e- and p p annihilations, the direct formation yield of c in p p annihilation is nearly 50 times larger than the indirect production yield of c in e+e- annihilation, the relevant branching ratios being: Br ( pp c) = 1.2 (4) 103, and Br (e+e- ) Br ( c) = 2.5 (6) 10-5. It should also be noted that unlike E760/E835 experiments which were obliged to identify c formation in the extremely weak two photon channel (Br (c ) = 3 10-4) , at the new facility at GSI, with a hermetic detector capable of electromagnetic and charged particle identification, one would be able to study c in several channels which have hundred times larger branching ratios (c +-K+ K-, 2 (K+ K-), 2 (+-), KK , , etc.) The radial excitation of charmonium, 'c(21So) By comparing the hadronic decays of the J/ and ' [6], it has been shown that radial excitations of charmonium are far from being simple recursions of the ground states. On the one hand, they reach much further into the confinement region of the qq potential, and on the other hand, they are quite close to the cc continuum to which they may couple. It is therefore important to identify and study the first radial excitation of the charmonium ground states, c'(2S0). The only evidence for the c comes from an early experiment at SPEAR [7] which measured a mass difference of 90 MeV between the c and the . This large difference is hard to reconcile with theoretical predictions unless coupled channel effects are large. A search for the c was performed by the experiments E760 and E835 at Fermilab in the process p p c [8]. No signal was observed by either experiment. Based on a total integrated luminosity of 40 pb-1 for both experiments combined an upper limit (90% confidence level) to the product of the branching ratios in the mass region 35753660 MeV/c2 was determined. For example, for a resonance width = 10 MeV: Br( p p c ) Br( c ) < 12.0 10-8. The technique employed by E760/E835 is limited by the relatively high background from 00 and 0 compared to the small signal [9]. Further measurements using the same channel would require increased statistics and a substantial reduction of the background. Alternatively, hadronic decay modes such as c KK or c allow a clean identification of this state. A possible explanation of the non observation of the c at Fermilab might be that its mass (or that of the c) is shifted due to mixing with a nearby 0-+ glueball. A comparison of the ratios (c)/(c) and (c p p)/(c p p) could shed

205

Section 2 light on this question. It should be noted that in the absence of the c our knowledge of spin-spin forces in the qq interaction is based only on the J/-c splitting. The hc(1P1) resonance of charmonium The singlet P resonance of charmonium hc(1P1) is of extreme importance in determining the spin-dependent component of the qq confinement potential. Essentially nothing is known about it. In absence of any direct knowledge of the Lorentz character of the confinement potential, it is more-or-less assumed that it is scalar, not vector. It follows that it makes no contribution to the spin-spin part of the qq potential; it arises entirely from the single gluon-exchange Coulombic part of the potential. It is a contact potential, which is identically zero for all except s-wave states like J/ and c. If this is true, the hc(1P1) resonance should be degenerate with the centroid of the c(3P0,1,2) states. If not, one has to consider a possible spin-spin contribution arising from the confinement potential, which would have major impact on our current understanding of the confinement potential. It is therefore extremely important to identify, and study the characteristics of hc, the 1P1 state of charmonium. Furthermore, a measurement of the total width and of the partial width to c + will provide an estimate of the partial width to gluons. Since calculations of partial gluonic widths of 1,3P1 states suffer from infrared divergences to leading order, a comparison with measured values will be of considerable importance. Calculations of branching ratios for hadronic decays to lower cc states also give contradictory answers and the debate could profit from accurate experimental data [ 10]. Single photon transitions '(3S1) hc(1P1) are forbidden by charge conjugation. For this reason the hc(1P1) state could not be identified in e+e- experiments. A signal in the hc region was seen by E760 in the process p p hc J/+0 [11]. Due to the limited statistics the experiment was only able to determine the mass of this structure (3526.2 0.15 0.2 MeV/c2), and to put an upper limit of 1.1 MeV (90 % CL) on its width. The results need confirmation. E835 took data in the same energy region, and the relative analysis is under way. What can already be anticipated is that the signal in this channel is very low (at the pb level). The proposed experiment could improve substantially the study of this state by adding considerably in statistics and, above all, in the ability to detect other decay modes of the hc. Radiative transitions of the J(3P0,1,2) charmonium states The measurement of the angular distributions in the radiative decay of the 1 and 2 states formed in p p annihilations provides insight into the dynamics of the formation process, the multipole structure of the radiative decay and the properties of the cc bound state. The c radiative decay is dominated by the dipole term E1. The higher multipoles M2 (magnetic quadrupole) and E3 (electric octupole) arise in the relativistic treatment of the interaction between the electromagnetic field and the quarkonium system [10]. 206

Section 2 The study of the radiative decay angular distributions of both 1 and 2 states allows the measurement of the deviation from pure E1 transition, through the E1-M2 and E1-E3 interference terms. The ratio between the fractional M2 amplitudes measured for the decays of the two states can provide a check of the theory. A comparison of the E760 result at the 2 [12] with the Crystal Ball result at the 1 [13] is not consistent with theory, and may suggest the existence of additional contributions to the theoretical predictions of the M2 amplitude. The simultaneous measurement of both angular distributions has been recently performed by E835 [12]. They, too, observed a discrepancy with respect to theoretical predictions, which could indicate the presence of competing mechanism, leading to the cancellation of the M2 amplitude at the 1. The effect seen by E835 is at the 2.5 level, therefore further high-statistics measurements are clearly needed to increase the significance of this result. Charmonium states above the DD -threshold Above the DD breakup threshold at 3.73 GeV, essentially nothing is known with any certainty. The e+e- annihilation experiments have only measured R = (e+e- hadrons)/(e+e- + -) in large energy steps. Relatively washed out structures were observed by some detectors (Mark I, DELCO), while others (DASP, PLUTO) claimed well defined peaks assigned to higher vector states ((3S, 4S, 5S). The latest, much more accurate measurements from BES (see Figure 2.7) do not confirm the sharp states claimed by DASP and PLUTO. It is now very much an open question whether the presently assigned higher vectors at 4040, 4160, 4415 MeV are in fact real. It is of critical importance to confirm the existence of these states because they would be rich in DD decays, and are often called D-factories.

207

Section 2

R = had /

7 6 5 4 3 2 1 0 3.8 4 4.2 4.4

BES (2000) 4.04 3.77

4.16

4.42

4.6

4.8

Ec.m. (GeV)

R = had /

7 6 5 4 3 2 1 0

MARKI (1982) 3.77 4.04 4.16 4.42

3.8

4

4.2

4.4

4.6

4.8

Ec.m. (GeV)

Figure 2.7. Comparison of R = (e+e- hadrons)/(e+e- + -) data from the Bejing Electron Synchrotron (BES) and MARK1 (SLAC) (Ref [14]).

Because of their energy limitation, the Fermilab experiments E760/E835 were not able to study any physics above the DD threshold. Yet this is the region in which narrow 1D2, 3D2 states (which are narrow because they cannot decay to DD ), and the first radial excitations of the singlet and triplet P-states are expected to exist. It is important to identify the vectors and the P- and D-states. This will require high statistics, small-step scans of the entire region accessible at GSI. In addition to the topics discussed above, exclusive charmonium decays represent a very interesting test ground for QCD predictions, particularly for the study of higher Fock state contributions, which might produce quite sizeable effects in selected cases. In order to perform a proper comparison between prediction and experiment, more and better data on these decays are needed. Of particular interest are the following decays: Hadron helicity non-conserving processes: , c baryon-antibaryon, 0 baryon-antibaryon G-parity violating decays: +-, 0, Radiative decays: + , , ... J , , ...

208

Section 2 In recent years numerical construction of effective potentials for heavy mesons and quarkonia by lattice methods have been remarkably successful [15]. Spectra of cc and bb mesons are fairly well reproduced by a Schrdinger equation based description with the QCD derived interaction Hamiltonian which is formed from a scalar linear confining term and a single gluon exchange model for hyperfine interaction including spin and momentum dependent terms up to O(v2) [16]. It has also been shown that radiative decays of heavy quarkonia are a sensitive testing ground for the effective confining interaction and the associated two-quark exchange current contributions [17]. A larger and more precise experimental data base of the charmonium system is essential for a better understanding of the phenomenon of confinement and of the strong interaction on the quark level in general. 2.1.3 Hybrids and glueballs The QCD spectrum is much richer than that of the naive quark model, as the gluons, which mediate the strong force between quarks, can also act as principal components of entirely new types of hadrons. These gluonic hadrons fall into two general categories: glueballs and hybrids. Glueballs are excited states of pure glue, while hybrids are resonances consisting largely of a quark, an antiquark, and excited glue. The additional degrees of freedom carried by gluons allow glueballs and hybrids to have spin-exotic quantum numbers JPC that are forbidden for normal mesons and other fermion-antifermion systems. Exotic quantum numbers (e.g., JPC = 0 , 0+ , 1 +, 2+ ) are the easiest way to distinguish gluonic hadrons from qq states, but even nonexotic glueballs and hybrids can be identified by measuring an overpopulation of the experimental meson spectrum and by comparing their properties (masses, quantum numbers, decay channels, etc.) with predictions from models or Lattice Quantum Chromodynamics (LQCD). Since the properties of glueballs and hybrids are determined by the long-distance features of QCD, their study will yield fundamental insights into the structure of the QCD vacuum. The most promising results for gluonic hadrons in recent years come from antiproton annihilation experiments. Two particles with exotic JPC = 1+ quantum numbers, 1(1400) [18] and 1(1600) [19] are clearly seen in the data. In the search for the glueballs a narrow state at 1500 MeV/ c 2 , discovered in antiproton annihilations, is considered the best candidate for the glueball ground state (JPC = 0++). However, it mixes with nearby conventional 0++ qq states, which makes the unique interpretation as glueball somewhat more difficult. Until now, however, the search for glueballs and hybrids was mainly restricted to the mass region below 2.2 GeV/c2. Gluonic hadrons of any mass and any quantum numbers can be produced in antiproton annihilation, but the maximum p -momentum at LEAR (1.94 GeV/c) only allowed the production of states up to 2.2 GeV/c2. Another gluon-rich process, the central production of mesons in proton-proton collisions, has difficulties in generating higher masses because the production cross section of a particle is inversely proportional to the square of its mass. Radiative J/ decays could also produce gluonic hadrons up to 3 GeV/c2, but such measurements simply lack 209

Section 2 statistics. Therefore, to understand the nature of gluonic excitations, a careful study of the spectrum up to 5 GeV/c2 is an absolute necessity. Because of the unavoidable problems due to the high density of normal qq mesons below 2.5 GeV/c2 it would be experimentally very rewarding to go to higher masses, where the light quark states have melted into a structureless continuum, and heavy quark states, which are far fewer in number, can be easily resolved. This is particularly true of the charmonium region. With only eight states in the 0.8 GeV/c2 region below DD , and a relatively smooth continuum above, it may be expected that exotics which may exist in the 3-5 GeV/c2 mass region can be resolved and identified unambiguously. 2.1.3.1 Charmonium Hybrids Theoretical predictions Predictions for hybrids come mainly from calculations based on the bag model, flux tube model, and constituent gluon model and recently, with increasing precision, from Lattice QCD [20,21]. All of these calculations do quite well in predicting the properties of the already known QQ states (Q = heavy quark) and their free parameters are fixed according to those data. For hybrids, the theoretical results agree qualitatively, lending support to the premise that their predicted properties are not too far from reality. First of all, why can we expect to find charmonium hybrids? The effect of an extra gluonic degree of freedom in meson-like systems is evident in the confining potentials for the cc g system, e.g., as derived from LQCD calculations in the Born-Oppenheimer approximation (cf. [21]).

Figure 2.8. Potentials between static quarks at separation R, in units of r0 0.5fm, as derived from LQCD calculations (scaled from [21]). The Vcc curve represents the ground-state potential (no gluonic excitation) and corresponds to the conventional cc states J/, c, , etc. The VHybrid potential originates from the first excited state of gluonic flux, giving rise to cc g hybrid states. The cc g ground state has the spin-exotic quantum numbers JPC=1 + and is predicted to be a narrow state at about 4.2 GeV/c2.

210

Section 2 Figure 2.8 shows the potentials between static quarks as a function of their separation R. The ground-state potential Vcc is the 0th- order cc potential, i.e., with no gluonic excitation; it is equivalent to the usual cc confining force and thus applies to the conventional cc states (J/, c, ',etc.) that have been found experimentally. The lowest-order hybrid-type potential, V Hybrid , results from the first excited state of gluonic flux and gives rise to eight possible cc g hybrid states with different quantum numbers. For example, the cc g ground state has the spin-exotic quantum numbers JPC = 1 + and is expected to be a narrow state with a mass of approximately 4.2 GeV/c2 [22]. Compared to light hybrid candidates which have reported widths of 200 to 400 MeV [18,19,23,24] there are several reasons for expecting that charmonium hybrids are likely to be much narrower. Their properties and those of other states are discussed in more detail below. Mass spectrum All model predictions and LQCD calculations agree that the mass of the lowest-lying charmonium hybrids are between 3.9 and 4.5 GeV/c2 [25] and that the state with JPC = 1 + has the lowest mass. The mass splittings are caused by spin-dependent effects which are not well known; energy differences of some tens of MeV between states seem realistic, but an experimental determination would yield valuable insights into the nature of the spin-dependent effects that cause them. Quantum numbers Until now, discussions have centered only around the lowest-lying charmonium hybrids. Four of these states ( JPC = 2-+, 1-+, 1--, 0-+ ) correspond to a cc pair with JPC = 0-+ or 1--, coupled to a gluon in the lightest mode with JPC = 1+-. The other four states ( JPC = 2+-, 1+-, 1++, 0+-) with the gluon mode JPC = 1 + are probably a bit heavier. Three of these eight charmonium hybrids have spin-exotic quantum numbers (JPC = 1 +, 0+ , 2+ ), so mixing effects with nearby cc states are excluded for them, thus making their experimental identification especially easy. Partial decay widths Since the predicted hybrid masses are above the DD threshold of 3.8 GeV/c2, decays into members of the D family will dominate if they are not forbidden by quantumnumber conservation or dynamical selection rules. The former condition forbids the decay of the spin-exotic 0 + state into DD , D * D* , Ds D s, which would violate CP invariance. A dynamical selection rule forbidding cc g ( cq )L=0 + ( cq )L=0 transitions is part of various models and backed up by LQCD. According to this selection rule, the first allowed decay of a charmonium hybrid into open charm states would be D D ** ( D ** being the lightest L=1 meson with a mass of 2.42 GeV/c2). Charmonium hybrids below this D D ** threshold of 4.3 GeV/c2 cannot decay into D particles, so their widths would be several tens of MeV or less. A flux tube calculation finds the width for a 2+ hybrid to be as narrow as 4 MeV [22]. If no DD decay channels are possible, then the decay modes will involve cc states; for example, the 1 + charmonium hybrid could decay into J/ + X, where X is , , or . Charmonium hybrids above 4.3 GeV/c2 can certainly decay into open charm mesons. To estimate the contribution of the DD 211

Section 2 decay widths of the hybrids, we note that the known vector states, (3S), (4S), (5S), (6S), in the same mass region, which are known to decay dominantly into DD , have widths of ~ 25-40 MeV. Those hybrids which have quantum numbers which allow them to decay into DD may be expected to acquire similar widths. Mixing In principle, non-exotic hybrid states can mix with cc states if they have the same quantum numbers and their masses overlap. Such mixing can change their predicted masses and widths, and is believed to be important for light hybrids. This is much less of a problem for charmonium hybrids, though, because there are only very few cc states and they have much smaller widths than qq states. Cross sections From theoretical considerations, little is known about the production rate of hybrids in p p annihilations. However, the formation of cc states has been measured in p p cc processes, yielding cross sections of 5 b for J/ and 1 b for c particles. The cross section for production of a charmonium state together with another particle has also been measured, namely in the p p J/+ process; the cross section of 120 pb [26] confirms theoretical predictions [27]. From experiments at LEAR we know that production rates of such qq states are similar to those of states with exotic quantum numbers. We estimate that the cross sections for the formation and production of charmonium hybrids will be similar to those of normal charmonium states. Proposed experimental program Charmonium hybrids with masses below 4.3 GeV/c2 are predicted to have narrow widths, and they will be a prime target in the search for gluonic hadrons in the mass region from 3 to 4.3 GeV/c2. Two types of experiments can be done: formation experiments and production experiments. Formation experiments would generate non-exotic charmonium hybrids with high cross sections, while production experiments would yield a charmonium hybrid together with another particle, such as a or an . In p p annihilation, production experiments are the only way to obtain charmonium hybrids with exotic quantum numbers. This distinction is a very powerful tool from the experimental point of view: the detection of a state in production and its non-detection in formation is a clear, unique signature for exotic behavior. It is envisaged that the first step of exploring charmonium hybrids would consist of production measurements at the highest antiproton energy available (E p = 15 GeV, s = 5.46 GeV), and studying all possible production channels available. The next step would consist of formation measurements by scanning the antiproton energy in small steps in the regions in which promising hints of hybrids have been observed in the production measurements. The selection of final states is used as quantum-number filter for the decaying hybrid. Charmonium hybrids with JPC = 1 or 1+ would decay into J/ + , while those with JPC = 0 +, 1++, or 2 + would be selected by triggering on a J/ + final state. An exotic 1+ charmonium hybrid can be found in a production experiment by means of its decay into J/ + X (where X is , , or ). 212

Section 2 In a formation experiment - often also called a scanning experiment - the momentum of the antiproton beam is varied in steps. Typical steps are 50 MeV, 10 MeV, or even smaller if necessary. Forty days of measuring time would be sufficient for an initial mass scan between 3.9 and 4.3 GeV/c2. Given a machine luminosity of 10 pb1 per day, this translates into a total integrated luminosity of 300 pb1 and a peak detection rate of 10,000 charmonium hybrid decays per day. This number is based on the following assumptions: (i) a non-exotic hybrid with a mass of 4.3 GeV/c2 and a width of several MeV is formed at a similar rate as charmonium states; (ii) the decay branching ratio into the J/ + channel is 1%; (iii) the accelerator and detector efficiencies are 50%; and (iv) the J/ is detected in the leptonic e+e- and + decay modes, which constitute 12% of all J/-decays and provide a clean experimental trigger. Such formation experiments would run partly in parallel with charmonium spectroscopy experiments, as described in chapter 2.1.2. Production experiments, on the other hand, should reveal the 1+ hybrid and other such exotics. For a hybrid with a mass of 4.2 GeV/c2 and a width of 30 MeV/c2, a p momentum of 15 GeV/c in the lab can produce enough phase space for the reaction p p cc g + 0(). One hundred exotic hybrids can be detected per day under the following assumptions: (i) the hybrid is produced at a similar rate as the charmonium states in association with a neutral pion; (ii) the hybrid decays into J/ + X (where X is , , or ); (iii) the accelerator and detector efficiencies are 50%; and (iv) the J/ is detected in the leptonic e+e- and + decay modes. Such conditions would yield a high-statistics Dalitz plot after just several weeks of measurement. Although analysis of data from the energy scan requires only the registration of count rates as function of the antiproton momentum (known as bump hunting), the proposed detector also makes it possible to do a partial-wave analysis, which disentangles the contributing quantum numbers and thus facilitates the extraction of states together with their quantum numbers. This is important because a partialwave analysis of the data is mandatory for production experiments. The feasibility of such analyses has already been demonstrated with data from LEAR. At HESR energies there are, in principle, many partial waves that contribute to the p preactions; but only the lowest waves are expected to play a role near the p p threshold, as was the case in p p-reactions near the ss threshold [28]. As for presently existing data, only very few are relevant to the search for hybrids in the heavy-quark sector. They originate from measurements at e+e- colliders and the results clearly suffer from poor statistics. One example is the decay of B J/--K from e+e- reactions as measured by CLEO [29]. The few events that cluster around the J/- masses of 4.4 and 4.7 GeV/c2 may be the first hints for the existence of charmonium hybrids. Further evidence for heavy-quark hybrid states is a structure above the (4s)-state [30], which could be assigned to hybrids with open beauty. Finally, some irregularities in the decays of heavier 1( cc ) states have been explained via mixing with hybrid charmonium states [31].

213

Section 2 2.1.3.2 Light Hybrids Two spin-exotic 1+ states, 1 (1400) and 1 (1600), have been found in -p reactions and more recently in p p annihilations [23,24,18,19]. The most striking feature is that their production rate in p p annihilation is comparable to that of normal qq states (see Figure 2.9). This feature makes annihilation reactions a prime tool in the search for further exotic states. Several predictions put 1+ hybrids at masses around 2 GeV/c2 [22,32]. The discrepancy between these predictions and the experimentally measured 1 (1400) and 1 (1600) needs further clarification. This can be done by measuring an entire spectrum of light hybrids by means of formation and production experiments. As in the search for charmonium hybrids, experiments would take advantage of the dynamical selection rule forbidding cc g ( cq )L=0 + ( cq )L=0 reactions and thus enhancing light-hybrid decays into specific final states, e.g., f1(1285), b1(1235) and K1 K . Given the relatively large cross sections (on the order of b), conclusive results should be achievable within a few months of measuring time.

Figure 2.9. Square of the invariant mass as measured in the pd p reaction by the Crystal Barrel Collaboration at LEAR [18]. The peak at 1.7 GeV2/c4 corresponds to the well known 2++ q q meson a2 (1320), the structure at 2.0 GeV2/c4 to the exotic 1 (1400). The shoulder at 2.4 GeV2/c4 is a kinematical reflection of the (770). The 1 state has the quantum numbers JPC = 1-+ and thus cannot be a q q state. The a2 and the 1 signals are of similar size, thus demonstrating that q q and exotic states are produced in pN annihilations at similar rates.

0

2.1.3.3 Glueballs There have been many searches for the glueball ground state over the last twenty-five years, but the best candidate has emerged recently from p p annihilation experiments. This rather narrow state, called f0(1500), has the non-exotic quantum numbers JPC = 0++. Since it mixes with nearby conventional 0++ qq -states, the f0(1500) 214

Section 2 is not a pure glueball in the strict sense of the word. However, such mixing may not be a problem for excited glueball states at higher masses. This alone is enough to motivate an extended glueball search than has been done previously. In any case, in order to understand the nature of glueballs, it is clearly necessary to study a whole spectrum of glueballs and not just the ground state. LQCD calculations make rather detailed predictions for the glueball mass spectrum in the quenched approximation disregarding light quark loops [33]. For example, the calculated width of approximately 100 MeV/c2 [34] for the ground-state glueball matches the experimental results. In the mass range that is accessible to the HESR project, LQCD predicts the presence of about 15 glueballs, some with exotic quantum numbers (see Figure 2.10). Glueballs with exotic quantum numbers are called oddballs. Oddballs cannot mix with normal mesons; as a consequence, they are predicted to be rather narrow and easy to identify experimentally [35]. Since the spin structure of an oddball is different, it may very well be that comparing oddball properties with those of non-exotic glueballs will reveal deep insights into the so-far unknown glueball structure. The lightest oddball, with JPC = 2+ and a predicted mass of 4.3 GeV/c2, would be well within the range of the proposed experimental program.

Figure 2.10. The glueball spectrum as derived from recent LQCD calculations [33], with 15 states in the mass region between 1.5 and 5.0 GeV/c2. The relative mass uncertainties are indicated by the vertical extent of the boxes. For the JPC = 0++ ground state, a good experimental candidate has been found in pp -annihilations, namely the f0(1500). All other glueballs are awaiting experimental discovery. The 2+ state at 4.3 GeV/c2 is of special interest, because it has exotic quantum numbers and can thus be easily distinguished from qq states.

Like the charmonium hybrids, glueballs can either be formed directly in the p pannihilation process, or produced together with another particle. In both cases, the glueball decay into final states like or would be the most favorable reaction; it is easily distinguishable from other annihilation channels and should exhibit only low-l partial waves, thus facilitating the spin analyses. Finally, it is suppressed for 215

Section 2 ordinary qq mesons and should therefore provide a unique glueball signature. For the formation experiment, the cross sections are quite large (b), so a mass scan between 2 GeV/c2 and 4.5 GeV/c2 could give conclusive results after just a few months of running time. As with hybrids, the production experiment would select spin-exotic glueballs. Both types of glueball search experiments can be done partly in parallel with charmonium spectroscopy experiments or with the search for charmonium hybrids. It is worth stressing again that p p-annihilations present a unique possibility to search for heavier glueballs, since the alternative methods have severe limitations. The study of glueballs is a key to understanding long-distance QCD, so every effort should be made to find them. 2.1.4 Interactions of hidden and open charm particles with nucleons and nuclei The investigation of medium modifications of hadrons embedded in hadronic matter is one of the main research activities at GSI at present and in the near future. The main physics goal is to understand the origin of hadron masses in the context of spontaneous chiral symmetry breaking in QCD and their modification due to chiral dynamics and partial restoration of chiral symmetry in a hadronic environment. Because of the limited energy available these studies have so far focussed on the light quark sector. Evidence for mass changes in the medium for pions and kaons have been deduced from the observation of deeply bound pionic states (see Figure 2.11) and from a study of excitation functions for K+ and K production in heavy-ion collisions in comparison to elementary nucleon-nucleon reactions (see Figure 2.12).

Figure 2.11. Excitation energy spectrum of the [36].

206Pb(d,3He)

reaction. The figure is taken from

216

Section 2 From the binding energies of the 1s and 2p pionic states in Pb isotopes, measured in the Pb(d,3He) reaction, a pion-nucleus potential has been deduced which indicates an increase of the - mass by about 25 MeV in neutron-rich nuclear matter at normal density. Nearly equal cross sections for K+ and K production at the same available energy in heavy ion reactions imply a cross section enhancement for K production compared to the elementary production processes. This observation is consistent with the scenario that the K mass is lowered in compressed nuclear matter formed in heavy-ion collisions: for the same energy more K mesons can be produced. Transport calculations, which reproduce these observations, indicate a lowering of the K mass at normal nuclear matter density by about 100 MeV. In the near future, the HADES experiment, presently being commissioned at GSI, will study medium modifications of the light vector mesons (,,) for which substantial changes of spectral functions in the medium are predicted already at normal nuclear matter density [37]. The presently available information on mass modifications of pions and kaons is summarized in Figure 2.13.

Figure 2.12. Excitation functions for K+ and K- meson production in heavy-ion collisions and elementary nucleon-nucleon reactions. The figure is taken from [38].

217

Section 2

Figure 2.13. Schematic picture of hadron masses in vacuum and mass splitting in the nuclear medium at normal nuclear matter density as derived from [36,38,39].

The HESR [2,3] with high intensity p beams up to 15 GeV/c will allow an extension of this program to the charm sector. Whereas the low energy dynamics of u, d and s quarks is anchored in the approximate chiral SU(3) symmetry of QCD (though with substantial explicit symmetry breaking by the strange quark mass) the physics of the heavy charmed quarks works on a completely different scale governed by their large mass mc 1.3 - 1.5 GeV. The leading mechanism which drives the short-distance interaction of c quarks with color singlet hadrons is the exchange of two or more gluons. Investigating the interaction of charmed hadrons with nucleons and nuclei is therefore tantamount to exploring fundamental aspects of gluon dynamics in QCD. As meson masses in the charmonium sector are dominated by the large mass of the charm quark pair, only little sensitivity to changes in the quark condensate is expected for charmonium states. The in medium mass of these states would be affected primarily by a modification of the gluon condensate. Recent calculations [40] indicate, however, only small in-medium mass reductions of the order of 5 - 10 MeV for the J/ and c in nuclear matter. For D mesons, the situation is different. Built of a heavy c quark and a light antiquark, the D meson is the QCD analogue to the hydrogen atom. Thereby, D mesons provide the unique opportunity to study the inmedium dynamics of a system with a single light quark. Recent model calculations [39,41] predict a lowering of both the D+ and D- meson masses with a mass split of the order of 50 MeV as also shown in Figure 2.13. A measurable consequence would be a sub-threshold enhancement for D and D meson production in p annihilation on nuclei similar to previous observations in the 218

Section 2 strangeness sector in heavy-ion collisions. Corresponding predictions are shown in Figure 2.14. The D and D mesons can be detected and identified via their relatively large branching ratios (10 - 20 %) into semileptonic decay channels. Cross sections of typically 1 nb near threshold lead to about 100 events registered per day at a luminosity of 2 1031 cm-2 s-1. The envisaged D and D production rates at the HESR would allow a substantial D meson physics program. These experiments should also be extended to higher incident momenta in order to study the interactions of the various open charm mesons and charmonia with the nuclear medium which are expected to be quite different for D, D , Ds and D s or J/ and '.

Figure 2.14. Total D meson pair production cross section in p annihilation on Au (upper curves) and on nucleons (lower curve) as a function of the antiproton energy. The cross sections for annihlation on Au were calculated for free (dashed curve) and in-medium (upper solid curve) masses of the D mesons. The figure is taken from [42].

Moreover, a lowering of the DD threshold in the nuclear medium by about 100 MeV would allow the ' and c2 states of charmonium to decay into this channel as illustrated in Figure 2.15. While in vacuum these states have widths of 0.3 and 2.7 MeV, respectively, as they are below the free DD threshold, the widths would dramatically increase in the medium [41]. A sensitive probe for studying such phenomena would be the spectroscopy of charmonium states produced via p annihilation in nuclei. Furthermore, in vacuum the (3770) is 31 MeV above the DD threshold. Its decay is completely dominated by the DD channel giving rise to a width of 24 MeV. The branching ratio into electron pairs is of the order of 10-5. The free decay width is sufficiently large for the major fraction of decays to occur within a large nucleus like Pb. A reduction of the D mass should again lead to a significant

219

Section 2 broadening of the (3770) which could be detected via lepton pair spectroscopy with the proposed HESR detector system for suitable kinematic conditions.

Figure 2.15. Charmonium states in comparison to the D D threshold in vacuum and in the nuclear medium at normal and twice normal nuclear matter density. The figure is taken from [41].

The suppression of J/ production in ultra-relativistic heavy-ion reactions is one of the signatures believed to indicate the formation of the quark-gluon plasma. J/ suppression can, however, also occur by hadronic interactions with nucleons and mesons. The J/ - N absorption cross section needs to be accurately measured. The results of the existing measurements of (J/ - N) at high energies are highly questionable because of problems related to colour transparancy, feed down, coloroctet production and model dependencies in the analyses of the primary data. Photoproduction data suggest a value of 3 - 4 mb for (J/ - N) while charmonium absorption on nucleons in p + A and A + A collision is conventionally fitted by 6 - 7 mb at high relative momenta. A very clean and direct measurement of ( J/ - N) can be made by studying resonant J/ production in p annihilation on nucleons in nuclei [43]. Figure 2.16 shows a Monte-Carlo simulation of the kind of measurement which can be made with 3.4 - 3.6 GeV/c antiprotons interacting with nuclei. A complete program should, of course, include the study of J/, ' and J production on a series of nuclear targets. The elementary process of cc pair dissociation into open charm in the presence of nucleons can best be studied in p annihilation on a deuteron [44]. In the reaction p + d D + c+ antiproton annihilation occurs on the proton with the formation of a cc pair which subsequently interacts with the residual neutron to form the charmed D meson and the charmed hyperon in the final state. A detailed study of the energy dependence of this reaction should reveal important features of cc dissociation by gating on large transverse momentum events. 220

Section 2

(e+e-) (pb)

300

J/250

200

150

100

50

0

3.25

3.5

3.75

4

4.25

4.5

p momentum (GeV/c)

_

4.75

5

Figure 2.16. Monte -Carlo results for scans of Fermi momentum broadened J/ resonance in p annihilation on nuclear protons. The statistics corresponds to 250 h running at a luminosity of 2.51031cm-2s-1. The figure is taken from [43].

The possibility to measure the elastic J/ - N cross section down to lower relative momenta is provided by the reaction p + d J/ + + n [45], where a large fraction of the momentum can be taken up by the photon. As the final state involves only two hadrons their relative momentum can be tagged by the photon. Similarly, the reaction p p J/ + + meson will allow the determination of the J/ meson scattering cross section which is relevant for understanding the interaction of J/ with comoving mesons in ultra-relativistic heavy-ion collisions. The experimental realization of the above physics program requires high luminosity, high quality antiproton beams and a dedicated, multi-component detector system with a sophisticated trigger/data acquisition of high rate capability. Microvertex tracking and a narrow, well-defined interaction region is needed for D meson detection. Dilepton spectroscopy with mass resolution better than 1% is required for experiments studying modifications of charmonium states in the nuclear medium. At present, no suitable dedicated facility, combining both beam quality and detector capabilities, is available for the proposed research program. Summarizing, the High Energy Storage Ring with an intense beam of antiprotons up to 15 GeV/c would offer an optimal framework for exploring the interactions of charmed quarks with nucleons and nuclei. Such investigations are of key importance to the basic understanding of QCD in its non-perturbative regime.

221

Section 2 2.1.5 Strange Baryons in Nuclear Fields A hypernucleus contains one or more hyperons implanted as an tracer within the nuclear medium. This introduces a new quantum number, strangeness, into the atomic nucleus thereby adding a third dimension to the nuclear chart [46]. These exotic nuclei provide a large variety of new and exciting perspectives ranging from genuine hypernuclear states with new symmetries not available in ordinary nuclei, over non-mesonic weak decays which offer the unique chance to study the interplay of the quark-exchange and meson-exchange aspects of the baryon-baryon forces, up to the possibility to study basic properties of hyperons and strange exotic objects. The many facets of hypernuclei An important goal of measuring the level spectra and decay properties of multistrange hypernuclei is to test the energies and wave functions from microscopic structure models and to put constraints on baryon-baryon interaction models. The distinctive role played by the strangeness degree-of-freedom can be illustrated by the fact that, unlike the nucleon-nucleon interaction which is dominated by the one pionexchange, the -nucleon interaction has no one-pion component. Thus, the short range aspect of the nuclear force may no longer be hidden under the long range onepion exchange mechanism. It is however clear that a detailed and consistent understanding of the quark aspect of the baryon-baryon forces in the SU(3) space will not be possible as long as experimental information on the hyperon-hyperon channel is not at our disposal. Since direct scattering experiments between two hyperons are impractical, the spectroscopy of multi-strange hypernuclei provides a unique approach to explore the hyperon-hyperon interaction. Furthermore, non-mesonic weak decays of hypernuclei will offer exceptional information on the four-fermion, strangeness changing, baryon-baryon weak interaction [47]. Another interesting facet of hypernuclei arises from the fact that the hyperons are embedded in a nuclear medium. Comparing different nuclei with different structure the influence of the nuclear medium on strange baryons and on their mutual interaction can be explored under quasi-stationary conditions. High resolution spectroscopic studies of deeply bound hyperatoms will add further important information on the properties of hyperons in a nuclear environment. Thus these data will provide an important reference point for heavy ion collision experiments where the same aspect is studied in dynamically evolving systems. The possible existence of an S = -2 six quark (uuddss) H-dibaryon represents another challenging topic of hypernuclear physics [48]. Presently about thirty theoretical predictions of its mass are available, ranging from below twice the nucleon mass to about 2.8 GeV/c2 and concentrating around the mass. So far, however, the experimental searches for this lightest strangelet which seems to be inherent to QCD are inconclusive. In fact the short timescale of the order of 10-23 s available in a coalescence process may prevent the transition from a state to an H-particle in free space [49]. Here, double hypernuclei may serve as breeder for the H-particle: the long lifetime of the order of 10-10 s of two 's bound together in a nucleus may help 222

Section 2 to overcome a possible repulsive interaction at short distances. Furthermore the fact that the mass of the H-particle might drop inside a nucleus [50] opens the possibility to observe the H even if it is unbound in free space. Thus spectroscopy of S = -2 hypernuclei can give important information on this fascinating and long sought object. Finally, hyperatoms created during the capture process of the hyperon will supply new information on fundamental properties of the hyperons. For example, the is the only baryon who's static quadrupole moment can be determined experimentally. However, this measurement was not feasible until now. At the antiproton ring HESR the large number of produced atoms will enable us to determine the hyperfine splitting in an atom and thus for the first time provide information on the static deformation of a baryon. Experimental concept A drawback common to all theoretical investigations on the full baryon octet structure is the lack of high resolution and systematic data on and multi- hypernuclei. At present, only three candidates for double hypernuclei have been completely identified via their double pion decay. Furthermore, until the end of the last decade, the energy resolution on hypernuclear studies was limited to typical values of about 1 MeV. Only recently the door to high resolution -spectroscopy of hypernuclei has been opened with the advent of high-efficiency and high-rate Ge-arrays like the Hyperball [51] or the VEGA detector [52]. Combining such a device with the high luminosity antiproton storage ring and with a novel solid-state micro tracker, high resolution spectroscopy of double hypernuclei and atoms will become possible for the first time.Figure 2.17. Two-step process for the production of double hypernuclei: a hyperon produced in an antiproton-nucleus collisions is stopped in a secondary target and converted into two lambdas.

223

Section 2 Because of the low energy release of only 28 MeV in a conversion of a and a proton into two 's, attempts to produce -hypernuclei are generally based on the capture reaction (Figure 2.17). In order to minimize the background from associated particles, the production of hypernuclei and hyperatoms at HESR will use hyperon pair production close to threshold in antiproton nucleus collisions at momenta of 2.6 and 4.9 GeV/c for and respectively. Accordingly, single hypernuclei can be studied at p momenta above the threshold. The trigger will be based on the detection of a high momentum anti-hyperon at small angles or of positive kaons produced by the antihyperon absorbed in the primary target nucleus. The reconstruction of a high momentum is straight forward and will make use of the forward region of the detector described in the technical part. The 2K+ trigger will provide a significantly higher count rate (up to a factor of 103) but requires the detection of rather low momentum kaons of a few hundreds MeV/c. The second ingredient of the experiment is the deceleration of the inside the nucleus and subsequent absorption in a secondary active target. The geometry of this secondary target is determined by the short mean life of the of only 0.164 ns. In order to optimize the survival probability of the during the stopping process the separation between the primary target and the secondary absorber should be minimized. If this distance is too large, low momentum will decay prior to full stopping. On the other hand, energetic with momenta beyond approximately 500 MeV/c will anyhow not be stopped prior to their decay. For a target-absorber distance of 5 mm and momenta between 100 and 500 MeV/c the stopping probability reaches a peak value of about 40%.

Figure 2.18. Schematic view of the central detector components for hypernucleus experiments .

224

Section 2 This limits the required thickness of the active secondary target to about 30 mm. In order to track the stopped and the charged fragments resulting from the decay of the produced hypernuclei, this device has to be a very compact, high-resolution solidstate micro-tracking detector with high rate capability. A compact stack of diamond or silicon micro strip detectors can meet these conditions. Alternatively one may envisage the use of a capillary fiber detector [53] read out by a hybrid phototube combined with the ALICE pixel readout chip [54] or by Visible Light Photon Counters (VLPC) based on low-temperature avalanche photodiode arrays [55]. Finally, an efficient, position sensitive germanium -array is required for the highresolution spectroscopy of excited hypernuclear states. To maximize the detection efficiency the -detectors must be arranged as close as possible to the target. Hereby, the main limitation is the load of particles from reactions. Fully digital electronics currently developed for the new generation of highly segmented Ge-arrays like VEGA [52] or AGATA [56] will - in connection with fast preamplifiers - allow a load of background events of more than 100 kHz for each detector element. Since most of the produced particles are emitted in the forward region not covered by the Ge-array an interaction rate of a few 107/s seems to be manageable. The experimental set up is sufficiently small to fit inside the calorimeter of the general-purpose detector presented in the technical part (Figure 2.18). After removing the end cap of the calorimeter upstream of the target the detectors will be placed close to the straw chambers thus making use of the tracking capability of the main detector in forward direction. In order to trigger on low momentum kaons a thin start detector (e.g. diamond strip) will be placed close to the target. The stop signal will be provided by scintillation or resistive plate counters, which replace or - if space permits complement the DIRC and MDC detectors. All hypernuclei experiments performed to date have collected of the order of 10000 stopped in total. At HESR we will be able to reconstruct approximately 3000 stopped with the unique trigger per day. This rate is comparable to the total number of stopped envisaged at the Japan Hadron Facility in (K-, K+) reactions. However, by applying the kaon trigger at HESR we will exceed these numbers by up to 2 orders of magnitude thus providing us with approximately 300000 stopped per day. Taking into account typical -branching ratios and -detection efficiencies we will observe several hundred -transitions for specific levels per day. 2.1.6 Further Possibilities The physics program as discussed above comprises topics we presently think of being highly relevant for the advancement of hadron physics. It encompasses experiments which are in reach by modern detector technology as well as realistic anticipations for quality and intensity of antiproton beams. The field of physics with antiprotons is of course much wider. Some of the possibilities are discussed more in detail in the following three sections.

225

Section 2 2.1.6.1 CP violation in the charmed meson and hyperon decay Charm sector CP violation [57] has been observed in neutral kaon and in neutral B meson decay[58,59]. In the standard model, CP violation arises from a single phase entering the CKM matrix. As a result, two elements of this matrix, i.e. Vub and Vtd have large phases. Because these elements have small magnitudes and involve the third generation, CP violation is small in the K0 system and is predicted to be even smaller in the D0 system[60]. Thus, a deviation from the small standard model effect indicating new physics can be more easily distinguished in experiments in the Dmeson system. CP-violation in neutral meson decays can be detected by observing final states which can be reached by one of two processes involving different weak amplitudes and phases. The first kind of CP violation is called indirect CP violation, and is due to a mixing of D0 and D 0, since both can decay to the same final state. It will be extremely small for the D system owing to the small mixing parameter expected in the standard model (of the order of rD 10-8 ). In so far, the observation of a relatively large flavor mixing rate would already indicate physics beyond the standard model. The second kind is called direct CP violation and occurs at the level of the decay amplitudes. Possible and relevant decay amplitudes for the D system involve singly Cabibbo-suppressed (SCS) amplitudes (branching fractions of 0.1%) and penguindiagrams. These amplitudes involve different elements of the weak quark mixing matrix with different phases so that an interference of these amplitudes can give rise to CP-violation. The relevant question is: will there be strong final state effects which can enhance the decay amplitudes to give an observable effect on the asymmetries or are there other physics processes not yet accounted for in the standard model also giving rise to an enhancement of such amplitudes? In the standard model using optimistic estimates on the strong amplitudes we may expect asymmetries of the order of CP~10-3. On the other hand Cabibbo-favored decays as well as doubly Cabibbo-suppressed decays (DCS) should not show any CP violation according to the standard model, whereas new physics can produce measurable CP violation effects. Working with D mesons produced at the DD threshold has also some advantages arising from the strong correlation of the DD pair, which is kept in the hadronization process. Formed near threshold, no asymmetries are expected in the production process. and the observation of one D meson tells about the quantum numbers of the other one at production in a charge symmetric environment (flavor tagging). Thus flavor ( DD ) mixing and CP violation can be searched for analog to methods in the Bsystem produced on the Y(4S). The drawback of the charm system is the very large number of decays which have to be observed in order to access this physics (translated into typically about 108 observed decays in the dominant channel), which requires a large sensitivity and high rejection power to possible sources of background. In order to obtain this count rate, HESR has to run at the peak luminosity, using about 510 7 p /s which equals the 226

Section 2 HERA-B interaction rate and corresponds to about twice the currently foreseen p production rate. Assuming a partial cross section of 510-6 for DD production around the (2S), about 250/s D meson pairs are accumulated thus producing 2.5109 D meson pairs within one effective year (107 seconds). With a trigger efficiency of 10%, useful branching ratios for the dominant decay channels of 25% and a reconstruction efficiency of 40% we may collect 2.5107 /year reconstructed charmed mesons, thus 5107 in 2 years of projected running (as compared to about 106 accumulated by E831 at Fermilab). For the CP asymmetry measurement, only SCS-modes are relevant. For the D 0D0 mixing we estimate for all lepton tagged events, using all decay modes, a limit of 10-4 (given by DCS decays) or a limit of 10-6, assuming double lepton tagged decays and a 1/N behavior of the upper limit. In addition to sufficient count rate, it is also required to have a successful strategy to deal with the systematic errors that result e.g. from the low signal to background ratio. Investigations of the suitability of the proposed detector system to measure these processes are under way (see chapter 2.2.5). Hyperon sector In self-analyzing two-body decays of hyperons the polarization of the mother particle can be obtained directly from the daughters. In these decays the orbital angular momentum of the final state can be L = 0, 1, in other words the decay amplitude can be an S-wave or a P-wave. Having two amplitudes, interference can occur and CPviolating phases can enter. There are two characteristic parameters, which govern the decay dynamics: The quantity denotes the asymmetry of the decay angular distribution, the quantity gives the decay-baryon polarization. With these quantities and the decay width , CP asymmetries can be formed [61]:

A=

+ +

B=

+ +

D=

+

For the asymmetry A, standard model predictions are of the order 2105. Some models beyond the standard model predict CP asymmetries of the order of several 10 4. To reach the standard model limit about 1010 hyperon decays have to be reconstructed, which could be done within one year under ideal conditions. The detector for this experiment requires good vertex reconstruction, excellent particle identification both in forward direction and at large angles and reliable long term stability. So the proposed detector system would largely comply with the needs of the experimentalist.

227

Section 2 2.1.6.2 D meson spectroscopy HESR running with full luminosity (limited by the production rate of 2107 p /s) at momenta larger than 6.4 GeV/c would produce large numbers of D mesons in pairs. Such an installation can be considered as a hadronic factory for tagged open charm, with about 100/s charmed pairs around (3S). Despite the small ratio (510-6) of charm production to total cross section, the background conditions are expected to be favorable because the hadrons are produced at threshold with no room for additional hadrons in the same process. The high yield and the well defined production kinematics of D meson pairs allows to look for: Rare decays The study of rare decays can open a window to non-standard model physics since it probes symmetry violation. Lepton flavor number violating decays e.g. D0 e or D e are interesting to search for. Flavor changing neutral currents like in the decay D0 +- can occur in the standard model through box graphs or weak penguin graphs with branching fractions smaller than 10-15. However the signatures of these decays are clean leaving hope for their observation, if processes exist boosting the decay branch. Leptonic decays Leptonic decays of D mesons like D+ , open the road to the understanding of hadron structure. The decay width is proportional to the wave function overlap of heavy and light quark. This quantity can be calculated in many quark models and is an ideal testing ground for LQCD [62]. The drawback of the charm system is the very large number of decays which have to be observed in order to access this physics (108 observed decays in the dominant hadronic and semileptonic channels) which requires a large sensitivity of the detector combined with high rejection power to possible sources of background. The last round of fixed target experiments (mostly Fermilab) have accumulated about 106 reconstructed D mesons. Today, only about 200 events for Ds + have been observed at CLEO quoting an 15% error on the decay constant both statistically and systematically. No events could yet be detected for D + with an expected branching fraction of about 10-4. Limits on rare D decays are between 10-6 - 10-4 about 6 orders of magnitude above the largest predicted numbers. The field of rare processes in D decays has also beenput forward in the BTEV proposal at FNAL. Current B-factories also produce large numbers of charmed hadrons. The competition, however, lies mostly in the field of leptonic decays. COMPASS envisages an accuracy of 10% in the measurement of the D and Ds decay constants fD , fDs . At this point it is difficult to estimate the precision to which the D meson decay constants can be measured at the HESR. If excellent vertex reconstruction with the detector can be achieved, and if sources of systematic errors to the events can be controlled, then there is a great prospect for interesting physics to do. 228

Section 2 2.1.6.3 Inverted deeply virtual and wide angle Compton scattering One of the prime goals of future experiments at HERMES and JLAB is the in-depth study of generalized parton distributions (GPD). Their large information content offers the possibility to explore the dynamics of quarks and gluons in the nucleon and in other hadrons. A key process to study GPDs is deeply virtual Compton scattering (DVCS): e + p e + * + p e + + p, which is described in terms of the so-called handbag diagram A similar mechanism (Figure 2.19) is believed to dominate the reaction p + p * + - at large s, where one tests, however, a completely different kinematical domain. At large Q2 the process is the timelike analogon of DVCS whereas at very small Q2 the process corresponds to the inverse of wide angle Compton scattering. Within the theoretical framework of the related double distributions the information obtained from both reactions is indeed very nicely complementary. From the experimental point of view, pp annihilation is somewhat easier. In DVCS one has to ensure that the outgoing