A narrow resonance in the system K0p at COSY-TOF

Forschungszentrum Kilichin der Helmholtz-Gemeinschaft

Possible evidence for the O+pentaquark state

�������������������������������� ����

i

Annual Report 2003

Institut für Kernphysik / COSY Managing Director: Prof. Dr. R. Maier Experimental Nuclear Physics I, Director: Prof. Dr. K. Kilian Experimental Nuclear Physics II, Director: Prof. Dr. H. Ströher Accelerator Division, Director: Prof. Dr. R. Maier Theoretical Physics, Director: Prof. Dr. J. Speth/ Prof Dr. U.-G. Meißner EDITORIAL BOARD: Prof. Dr. Gerhard Baur Dr. Markus Büscher Prof. Dr. Detlef Filges Prof. Dr. Kurt Kilian Prof. Dr. Rudolf Maier Dr. Peter von Rossen Prof. Dr. Josef Speth/

Prof. Dr. Ulf-Gerrit Meißner Prof. Dr. Hans Ströher and Prof. Dr. Hartwig Freiesleben, TU Dresden Cover Picture In 2003 strongly increasing experimental evidences have been published for a new particle Θ+, which may represent a new form of matter. Strange nucleon (hyperon) production was intensely studied at COSY- TOF. One of these hyperons is the Σ+ particle, which is produced in proton proton collisions together with a neutral kaon and a proton. This reaction enables us to study the proton - kaon subsystem, where the five-quark state Θ+ may be visible as a resonance according to the following scheme: pp → (Θ+) + Σ+ → (p+ K0) + Σ+ . A dedicated detector system very close to the target was used to register decays of particles as close as a few millimeters to the target. With this system the K0 decay into two pions and the Σ+ decay into a nucleon and a pion are detected in coincidence. By including the track of the proton the invariant mass between the proton and the reconstructed K0 is determined. The results of this distribution - measured at a beam momentum of 2.95 GeV/c - are shown. A clear signal is observed near 1.53 GeV/c2, which may be the first indication for the pentaquark state Θ+ produced in proton proton collisions. More details are found in the report.

ii

iii

PREFACE We have been very pleased by the positive response we received in connection with the new format of the annual report that was inaugurated last year and present for this reason this year’s activities in the same manner. The printed part concentrates on highlights and gives background information together with information about the institute. The accompanying CD-ROM contains the scientific articles in PDF-format and can easily be looked at with any PC or notebook computer.

Firstly, let me mention a most notable event for our institute, the retirement of Professor J. Speth, who had been in charge as director of the Institute of Theoretical Physics for more than 20 years. With deep gratitude we remember his many outstanding contributions and scientific visions, that have been instrumental to obtain with COSY a new powerful research instrument. We are fortunate that he will be still at our institute, which allows us to benefit also in the future from his distinguished experience. At the same time it is a great pleasure for me to welcome Professor Ulf-G. Meißner, who took over this position at a time that has great significance for the future of the institute. We highly esteem his decision to join us despite his many obligations that he has as chair in theoretical physics at the University of Bonn. His dedication to Quantumchromodynamics, symmetries, and effective field theories will be of great importance for the future physics program.

The year 2003 had been special in many ways as the institute is preparing itself for a paradigmatic change in its future funding. This initiative had been started by the federal government under the term “Program oriented Funding” (PoF) and aims to set long term perspectives for the various research fields. Our institute jointly with GSI has outlined mid term and long-term perspectives in the general field “Structure of Matter” and specifically in our case “Physics of Hadrons”. The objectives are summarized in a Helmholtz Research Programme Report under the title “Physics of Hadrons and Nuclei”

Our institute also contributed its expertise to advance accelerator development within the European framework. HP-NIS is aimed to extend the performance limits of High Performance Negative Ion Sources while our contribution in the framework of CARE (Coordinated Accelerator Research in Europe) concentrates on the improvement of new acceleration structures for high intensity pulsed machines.

The endorsement of the federal government for the GSI proposal to build a large and universal accelerator facility brings into reach the exciting opportunity to extend the future physics program into the charmed regime. We have, therefore, committed ourselves to actively participate in this promising project by taking a leading role in the construction of the HESR ring, which will provide anti-protons with unprecedented intensity and beam quality.

Another important milestone for our institute has been the decision to become host for the WASA detector, which presently is taking data at the CESIUS ring. The decision of the WASA collaboration to transfer the detector to COSY will have a major impact on our physics program as it dramatically extends our detection capabilities.

There has been a general excitement in hadron physics because of experimental indications for a new exotic quark matter, which seemingly may also have appeared in data taken by the TOF collaboration. Naturally, the whole issue is still in a state of intense and controversial debate. But unanimity exists that more experimental investigations are required to come to an all-convincing conclusion and that if this new state of matter exists it will turn a new page in strong interaction hadron physics.

CANU has always been at the heart of COSY. As our user community has meanwhile widened and gotten much more international it was felt necessary to change the meaning of the acronym to reflect this. The CANU assembly has unanimously adopted the new definition “COSY Association of Networking Universities” in its December meeting.

The work presented in this report would not have been possible without the dedication of our technicians and engineers. Equally important has been the support of the IKP service groups and of the infrastructure of the Forschungszentrum Jülich, especially those from central divisions for electronics (ZEL) and advanced technology (ZAT). As always, the students working on their diploma and on their PhD-theses had a substantial impact on advancing the research at COSY. We gladly mention with gratitude, too, the fruitful collaboration with CANU and the many outside users, who have been the foundation of the common success.

Our gratitude extends to the advisory committees whose help was crucial in the often-difficult decision-making process. Indispensable as well was the support that many outside users obtained by the BMBF and through the FFE-program of the Forschungszentrum Jülich. Last but not least we are deeply indebted to the board of directors of the Forschungszentrum Jülich for their sustained commitment to the COSY research program.

R. Maier

iv

v

CONTENTS

I. Experiments at COSY and other Facilities 1 a. Highlights from IKP 1 3 b. Highlights from IKP 2 11

II. WASA 17 III. Accelerator Physics 21 IV. Theoretical Investigations 31 APPENDIX A. Scientific Council IKP 41 B. Program Advisory Committee (of COSY) 41 C. Collaborations 42 D. Personnel 53

vi

1

I. Experiments at COSY and other Facilities

2

3

HIGHLIGHTS FROM THE IKP1

COSY-TOF Spectrometer The COSY TOF (Time-Of-Flight) spectrometer (Fig. 1) detects charged particles with huge angular acceptance of nearly 2π by reconstructing the four vectors from measured directions and time of flight. This spectrometer is especially capable to examine multi-particle reactions like meson productions from threshold up to excess energies of several 100 MeV. Due to the very high acceptance (in the cm system) and the kinematical completeness of all events, all (!) differential observables one might think of, (like differential cross sections and analyzing powers, invariant mass distributions of any subsystem, multi differential distributions, spatial correlations and so on). can be shown Limitations come from statistics and resolution.

Fig. 1: Experimental Setup of the Time of Flight

Spectrometer. Quirl, Ring, and Barrel counter hodoscopes are in the big vacuum tank shown. Beam comes from the right, going first through the very small cryo-target. Reaction particles pass the following Erlangen multiplicity and tracking detector system which starts right after the target.

Unique at TOF is the detection of track multiplicity both very close to the nearly point like target (few mm3 overlap with the beam) and very far away. An increase of charged tracks is due to delayed decays (e.g. Λ→pπ- , K0 →π+ π- ) and provides a very clean on line trigger for reactions with strangeness. Strangeness production with its overall branching ratio of

4

Measurement with a Deuteron Beam A deuteron beam was for the first time extracted to the TOF spectrometer. We looked for the reaction dp→pppπ- in order to investigate the reaction np→ppπ- where the protons from the deuterons were detected as spectators. (close to beam velocity) The beam momentum of 1.85 GeV/c corresponds to an excess energy of the free np reaction of 40 MeV in the ppπ- exit channel. For calibration purposes we used the quasi-free pp elastic scattering reaction, where the neutron is detected as spectator. Fig. 3 shows a typical event. (Nearly planar and close to 900 between the proton tracks)

Fig. 3: Detector response to a quasi-elastic scattering event dp→nSpectpp. The neutron is detected as spectator with the calorimeter (small yellow hexagon). The thin lines represent the fiber hodoscopes, the narrow, filled wedges the barrel bars, and the wide wedges the start counters.

The GEM Experiment (BIG KARL)The η mass arouses recently great interest. Measurements taken before the sixties report a mass between 548.0 MeV/c2 to 549.0 MeV/c2. In contradiction measurements taken in the seventies up to the nineties, report values, smaller by typically 1 MeV/c2. These are compiled in Fig.4. The large discrepancy led the particle data group (PDG) to omit the earlier data and come to a mean value of 547.30±0.12 MeV/c2. Recently, a new measurement giving 547.843±0.051 MeV/c2was published by the NA48 collaboration, which is in contrast to the mean value of the PDG. The small uncertainty reported moves up the mean value. The NA48 value differs from the earlier mean by 5σ!

546.8

547.0

547.2

547.4

547.6

547.8

548.0

548.2

m(η

) (M

eV/ c

2 )

1970 1975 1980 1985 1990 1995 2000 2005year

Douane

Plouin

Krusche

Lai

this work

Fig. 4: Published values of the η mass over the last 30

years. The most recent NA48 result, indicated by Lai is in contradiction to our result.

If one wants to resolve this discrepancy then one has to improve control of the systematic errors. This was possible by measuring the 3He momentum distribution in the two-body η production reaction

pd→3Hη

reconstructing then from a big enough sample the missing η mass.

What one needs however is an absolute calibration both for the beam momentum and the spectrometer setting. For that purpose we could take advantage of a “magic kinematical situation”. At about 1620 MeV/c beam momentum, one gets in the reaction

pd→3Hπ+ both forward (in cm) π+ and the coincident backward (in cm) 3H at exactly the same lab momentum – namely half beam momentum - when they are at 00. This situation depends only on the masses of the particles, which are involved, and not on a specific spectrometer or accelerator! This calibration trick can be extended to finite scattering angles of the coincident reaction partners and also to slightly different beam momenta.

At a beam momentum of about 1640 MeV/c we took advantage of a second “lucky accident”. Here the doubly charged 3He from the η production reaction on deuterium comes with close to beam momentum and just also fits into the acceptance of BIG KARL. No change of any parameter is needed. Calibration of beam and spectrometer momentum and the η mass determination run simultaneously.

For the first time a beam which was electron cooled at injection was used in that experiment! Together with a thin (

5

p+p→ d+π+ at different momenta). The effective target thickness was monitored. The momenta of recoiling 3He were corrected for degrading in the target. The systematical error was found from simulation calculations where the uncertainties of the fitted parameters were included. This leads to a systematical error of 0.021 MeV/c2. In addition an asymmetric component of the beam was found giving another uncertainty of -0.015 MeV/c2. As result we get

m(η) = (547.356 ± 0.027 ± 0.021 ± 0.015) ΜeV/c2 This value, with all errors added in quadrature, is shown in Fig. 4 The total error is smaller than those of all previous measurements. It is in agreement with the measurements before 2002 but in disagreement with the result from NA48.

In one of the calibration runs the reaction pp→π+X

with X=d or X=pn at a pion emission angle of zero degree was measured. We have used these data to study the pn interaction. In previous studies of this type was the missing mass resolution found to be rather poor. The deuteron yield punches through to the break up region. Hence it was impossible to study the threshold region of the break up, where one expects a narrow quasi-bound spin singlet state. The high resolution data obtained in the present experiment solve this problem. They show the 2.22 MeV wide gap between deuteron production and the pn continuum. (Fig 5).

102

103

104

coun

ts

-10 0 10 20 30 40

Q (MeV)

Uppsala ≈30 degreeGEM 0 degree

Fig. 5: Comparison of pp→π+pn data taken earlier at

CELSIUS with the present data.

A missing mass resolution of 92 keV/c2 was obtained. This is about the same relative resolution as found for the cooled beam. For such high missing mass resolution not only the value of the momentum of the pion but also the emission angles have to be measured with similar precision.

Fäldt and Wilkin had derived an analytical continuation of the pole to the continuum. Though this relation is exact only at the pole it is a good approximation for small momenta and also for quasi-bound states, if the production operator is of short range. This is the case for pion production. When one uses this relation for the spin triplet cross section, the break up part is fixed by the deuteron yield. The singlet cross section is then given by

Ns → Ntot –Nt This yield is compatible with zero if a constant background as strong as the one below the deuteron peak is assumed. The same finding holds when the form of the singlet continuum plus a polynomial representing the background is fitted to Ns. From these procedures upper limits for the quasi-bound pole and the singlet continuum are found. The latter is Ns < 0.03Ntot where a statistical weight of 1/4 to 3/4 for singlet and triplet contribution would lead to Ns=0.40Ntot. This upper limit is by a factor of three smaller than in previous studies.

The Germanium Wall of the GEM set up was for the first time complete, i. e. the quirl micro strip detector was followed by three thick pizza detectors (50mm thick in total). It was used to measure cross sections and vector analysing powers of the reaction

pd→3 Heη in the S11 resonance region. In another experiment the same set up was used to study the reaction

dd→αη For the first time a vector and tensor polarised deuteron beam was employed in an experiment.

A search for η-mesic nuclei is planned. The experiment will make use of recoil free kinematics in the reaction

( )**

32 2 .A Ap He

d η

− −+ → +

at 1736 MeV/c “magic” beam momentum. The η can be produced at rest in the laboratory system. Therefore, one has the best chance to find binding effects. The residual nuclear part as well as the mesonic bound state can be excited.

Fig. 6: View into the ENSTAR detector. Different

scintillators and the woven tube for the target holder can be seen. Two pieces from the outer layer were removed to make the axial segmentation in the target region visible.

Measurement of the recoiling 3He will allow to observe binding effects in the missing mass. However, that spectrum will also contain contributions from multi pion

6

production. In order to reduce this physical background the ENSTAR detector was developed. It will detect characteristic decay products from the mesic system. The ENSTAR detector has barrel geometry with three cylindrical layers of scintillators. The nearly completely assembled detector is shown in Fig. 6. One can see an

additional sandwich structure in axial direction in the target region. A vertical pipe connects to the horizontal beam tube and allows coming in with the target. These vacuum pipes are manufactured from epoxy- carbon fibre compound.

The COSY-11 ExperimentThe internal COSY-11 installation uses an accelerator dipole as a magnetic spectrometer with a detection system attached. Fig. 7 shows a photo and fig. 8 a sketch of the detectors.

Fig. 7: The COSY-11 installation arranged at a COSY

machine dipole.

Fig. 8: The COSY-11 detection setup in the bending

section of the COSY ring. Depending on the studied reaction an additional drift chamber and a start scintillator hodoscope is positioned directly in front of the exit window.

The main detection system consists of drift chambers and scintillator hodoscopes for a precise 4-momentum determination of positively charged ejectiles. A silicon pad detector inside the dipole allows detecting negatively charged ejectiles and the forward going neutrons are registered in an arrangement of alternating lead and scintillator layers. To study pn induced reactions at a deuteron target a spectator proton detector with 18% of 4π solid angle coverage is installed close to the target. It has been provided by the WASA/PROMICE collaboration. The setup is extended for certain reaction channels by a start scintillator hodoscope and an additional drift chamber in front of the exit window of the vacuum chamber.

Reaction channels investigated at COSY-11 are meson (η ,ω and η’), two meson (π+π- , K+K- ) and hyperon production in pp as well as pd and dp induced reactions at a H2/D2 cluster target.

The hyperon production studies have been continued. The final analysis of the comparative production of Λ and Σ0 hyperons via pp→p K+Y results in data shown in fig. 9.

σ (p

p →

pK

+Λ

)

excess energy [MeV]40 60200

σ (p

p →

pK

+ Σ

0 )

> 300

10

20

40

30

0

Fig. 9.: Cross section ratio of Λ and Σ0 production via

pp→pK+Y. The solid line corresponds to the ratio of the fitted parameterisation including p-Y FSI.

The large cross section ratio σ(Λ)/σ(Σ0) observed close to threshold decreases strongly between 10 and 20 MeV excess energy. The solid line in fig. 9 results from parameterisations of the individual Λ and Σ0 excitation functions by an s-wave phase space distributions modified by the p-Y final state interactions. The p-Λ FSI seems to be much stronger than FSI in the p-Σ system.

This simple description accounts for the gross features of the energy dependence of the cross section ratio. In detail some more physics must be included. More about the existing model descriptions can be found in the individual contributions of this annual report.

7

In 2003 data have been taken for Σ+ hyperon production via pp→nK+Σ+, which is presently under analysis. This channel gives information complementary to the earlier σ(Λ)/σ(Σ0) data. It may help to disentangle the importance and phase relations of different boson exchange contributions in threshold hyperon production.

Furthermore data on η production in pp→ppη with polarized beam, pd→psppnη and dp →3Heη have been taken. The preliminary analysis of the η production on a deuteron target clearly demonstrates that pn reactions with spectator proton tagging can be done at COSY-11. The quasi-free pn→pnη reaction at a deuteron target runs via pd→psppnη. To identify the η meson via the missing mass requires measuring the four vectors of the fast p and n and the slow spectator psp. The WASA/PROMICE tagger does this perfectly with 18% solid angle coverage in an angular range 1000

8

isotopes are not separated, and 14% for Be where the isotopical structure is visible has been found. While the low energy part of measured kinetic energy spectra can be nicely reproduced by statistical models also for composite particles like He, Li, Be and B, intra-nuclear cascade codes developed for modelling the first phase of the reaction fail completely to describe the pre-equilibrium component even for He. Very first approaches with the relativistic quantum molecular dynamics model (RQMD) show a much better agreement with the experimental data for light composite particles as He and Li. Qualitatively the RQMD reproduced both, the low and the high energy part of the distributions. To derive conclusions concerning the reaction mechanism the calculation needs to be extended on particles from excited remnants of the fast stage of the reaction

dEdx

chart ofthe Bragg curve

Penetration depth

Particle

Cathode Anode

Frish Grid

Resistor Chain

electrons

field shaping rings Fig 12.: BCD as used for spallation studies at the COSY.

Representative the mass identification for isotopes up to 14N is shown in Fig. 13 for 15o in respect to the beam direction. Yield ratios of 7,9,10Be and 8,10,11B given in Fig. 13 reproduce roughly the ones published in literature The analysis on absolute cross sections and comparison to published data is currently in progress.

Distance from COG (a.u.)-150 -100 -50 0 50 100

Co

un

ts

0

10

20

30

40

50

606Li

7Li

8Li

Distance from COG (a.u.)-200 -100 0 100

Co

un

ts

0

5

1015

20

25

3035

40

45 7Be9Be

10Be

Distance from COG (a.u.)-100 -50 0 50

Co

un

ts

0

20

40

60

80

100

120

140

10B

11B

Distance from COG (a.u.)-100 -50 0 50 100

Co

un

ts

0

20

40

60

80

100

120

140

160

11C

12C

13C

14C

Distance from COG (a.u.)-100 -50 0 50

Co

un

ts

0

20

40

60

80

100

120

140 13N14N

Fig. 13: Mass identification spectra of the 1.9 GeV p+Ni

reaction products measured at 15o

Acknowledgement This work is supported by the WTZ project POL01/90, the EU-LIFE program, the EU HINDAS project FIS5-1999-00150, and the EU TMR project ERB-FMRX-CT98-0244.

Investigation of Cold Neutron Spectra (JESSICA)The JESSICA (Juelich Experimental Spallation Target Set-up In COSY Area) experiment is used for the characterization of the energy spectra of advanced cold moderators. The knowledge of the moderator spectra of such moderators is necessary when designing new, high power spallation sources. In order to reduce the necessary time for measurements, especially with cold neutrons, the increase of the neutron flux is one main goal when designing a new facility. With the JESSICA experiment (see Figure 14) we are able to characterize the neutron spectrum and can validate existing neutron scattering kernels.

The experimentally determined spectra are compared with Monte-Carlo simulations performed with MCNPX in order to validate new developed neutron scattering kernels and the particle transport through matter in the low energy region.

Fig. 14: The JESSICA led reflector. The mercury target is

partly pulled back.

9

Fig. 15: Comparison of the energy spectra between

JESSICA (solid line) and MCNPX (circles) for an ice moderator operated at 20 K and 70 K and a water moderator at room temperature.

Progress was done with the simulation of the energy spectra of cold ice moderators at 20 K and 70 K. Figure 15 illustrates the comparison of the simulated energy spectra and the experimentally determined data from the JESSICA experiment. Whereas the simulated and measured spectra for an ambient temperature water moderator –a material with well known properties- are in a very good accordance, difficulties in the simulation of the energy spectra of cold ice moderators are observed. As can be seen in Figure 15 the position of the maximum of the simulation is shifted towards lower kinetic energies. One material supposed to have advantageous properties for the moderation of neutrons is methane-hydrate, where a single methane molecule is encaged in an ice cage formed by six water molecules. It was assumed that methane-hydrate combines the properties of an ice moderator and those of a solid methane moderator. Figure 16 shows the energy spectra of an ice, methane, and methane-hydrate moderator at a temperature of 20 K. The methane-hydrate moderator matches the energy spectrum of the methane moderator up to 4 meV exactly. The maximum at 2 meV is located at the same position for methane and methane-hydrate.

Fig. 16: Energy spectra of solid methane, ice, and

methane-hydrate at T=20 K.

Whereas the methane moderator shows a strong decrease of the neutron flux for higher energetic neutrons, the methane-hydrate moderator yields up to a factor of three higher flux in the energy range between 5 meV and 400 meV. The ice moderator is superior to methane-hydrate in the energy range above 40 meV. This effect is supposed to be due to the different moderator properties of normal ice Ih and the ice cage of methane-hydrate. On the one hand it can be seen, that methane-hydrate is a material which combines the advantageous properties of both materials -methane and ice- into one single material, but it is not a superposition of both single spectra. A further moderator material which was investigated was mesitylene. The observed energy spectrum is similar to the spectrum of methane-hydrate. The charme of mesitylene is, that it does not show irradiation damage effects, which is advantageous for moderators to be operated in high radiation fields.

Production of Cold Anti-hydrogenThe ATRAP experiment at the CERN Anti-proton Decelerator AD aims for a test of the CPT invariance by a comparison of the spectroscopy of hydrogen- to anti-hydrogen atoms. To achieve the required high precision in the measurements of atomic transitions cold atoms of anti-hydrogen are essential and consequently trapped neutral anti-hydrogen atoms at rest, up to now not available, have to be used.

The ATRAP-collaboration performed their anti-hydrogen studies at two different installations: ATRAP-I, using an already existing magnet and being now in operation since 3 years with increasing progress and success for preliminary studies, and ATRAP-II, where the installation started in 2003.

In both experiments the Penning traps for anti-protons and positrons are positioned in superconducting solenoids

surrounded by detectors for the annihilation products. ATRAP-I is strongly limited in space which allows only a very restricted detector equipment of scintillating fibre layers in straight and bended orientation, see fig. 17. At ATRAP-II a dedicated solenoid will be used with a sufficient large bore diameter of 0.5 meters for a charged particle tracking system as well as a neutral particle trap.

The anti-hydrogen production via 3-body recombination routinely operated at ATRAP-I was studied in more detail in order to characterize this production mechanism in view of future trapping of neutral anti-hydrogen atoms. Shape parameters of the anti-proton and positron clouds and the anti-hydrogen production probability as a function of relevant parameters have been measured. The analysis of the N-state distribution of the produced Rydberg anti-hydrogen was extended to lower N-states down to N > 20.

10

In fig. 18 the typical potential distribution in the trap used for these measurements is shown.

Fig.17: Configuration of scintillating fibre detector and

Penning traps installed in the solenoid of ATRAP-I. The Penning trap is moved into the detector for the experiments. To illustrate the fibre orientation a few fibres are illuminated by coloured light.

First measurements of the anti-hydrogen velocity have been performed by applying an additional sinusoidal potential to the electrode T8, see fig. 18. The maximum field strength is sufficient to ionize most of the produced anti-hydrogen atoms. With increasing frequency of the potential the time window in which the field is below the ionization for a certain N-state is reduced. Therefore the measurement of the number of anti-hydrogen traversing the electrode region as a function of frequency allows to estimating the velocity distribution. The analysis of the data is still in progress. However, it became apparent by these investigations that the anti-hydrogen atoms are still

hotter than needed for being trapped in a neutral particle trap.

For this and among other reason an alternative production mechanism via double charge exchange has been studied. This method allows producing anti-hydrogen in well defined Rydberg states. Via laser excitation with two subsequent frequencies a beam of Rydberg Cs atoms (with a main quantum number of N ~ 37) is produced which traverses a cloud of positrons. The outer electron of the Cs is caught by a positron and positronium is built in a well defined state. This neutral positronium is moving isotropic and a small fraction also traverses a prepared anti-proton cloud at 4 K temperature where in a second charge exchange process anti-hydrogen atoms can be produced again in a well defined Rydberg state. The production of anti-hydrogen atoms via this double charge exchange process was clearly observed. Although the production rate is smaller than in the 3-body recombination the atoms are clearly produced at a temperature of 4 K.

Presently possible solutions for construction of a magnetic trap for the neutral anti-hydrogen atoms are discussed and first tests of a combined Joffe/Penning trap will start next year.

Fig. 18: Potential distribution in the trap for typical anti-

hydrogen experiments. At the stripping well the field increases strongly (high potential difference to the following electrode) and the anti-hydrogen atoms produced down to low N-states are ionized. For the velocity measurements a sinusoidal potential is applied to the electrode marked by the arrow.

Penning trap for positrons

Penning trap for antiprotons

Scintillating fibres one straight and two helical layers

Fiber ends coupled to 16-fold photomultipliers

11

HIGHLIGHTS FROM THE IKP 2

The ANKE ExperimentOverview ANKE, a magnetic spectrometer and detection system at an internal target position of COSY, allows one to separate and momentum analyse ejectiles from the circulating COSY beam with forward emission angles around 0˚. Due to its large solid angle and wide momentum acceptance ANKE can be used for a variety of experimental studies. Outstanding results from ANKE in 2003 are:

1. First measurement of the K+-production cross section in proton-neutron interactions,

2. Study of a0+-resonance production in pp reactions, 3. Measurement of ω and η-meson production in

proton-neutron collisions, 4. Measurement of the proton-induced deuteron

breakup Further major technical developments for ANKE comprise:

a) The development of the atomic beam source for the polarized internal target has been finalized. This target will be used for double polarization experiments at ANKE.

b) The ANKE pellet target has produced droplets of frozen hydrogen and liquid nitrogen.

Fig.1: Photo of ANKE and the detection systems.

Strangeness Production in pp and pn interactions The layout of ANKE, including detectors and the DAQ system, has been optimised and used to study K+-spectra from proton-nucleus collisions at beam energies down to Tp = 1.0 GeV (p = 1.70 GeV/c), thus far below the free nucleon-nucleon threshold at TNN = 1.58 GeV. This is a very challenging task because of the small K+-production cross sections, e.g. 39 nb for pC collisions at 1.0 GeV. In subsequent measurements ANKE has been used to measure K+-mesons in coincidence with protons and deuterons from pn → pK+X, pp → d K+ K 0, and p12C → p/d K+X reactions.

K+-production in pn interactions Data on the K+-production cross section from pn interactions in the close-to-threshold regime are not available yet. This quantity is, for example, crucial for the theoretical description of proton-nucleus and nucleus-nucleus data since it has to be used as an input parameter for corresponding model calculations. Predictions for the ratio σn/σp still have large uncertainties, they range from one to six, depending on the underlying model assumptions.

K+-production in proton-deuteron collisions has been measured with ANKE at two beam momenta, p = 2.60 and 2.81 GeV/c (see Fig. 2). Based on the assumption that the K+-production cross section on the deuteron is governed by the sum of the elementary pp and the pn cross sections, i.e. σD = σp + σn, the experimental distributions have been simulated, treating the ratio σn/σp as a free parameter. Figure 2 shows the resulting momentum spectra for σn = σp, 2σp, 3σp and 4σp. The best agreement between data and calculations is obtained for σn/σp ~ 3 at 2.60 GeV/c and σn/σp ~ 4 at 2.81 GeV/c.

Fig. 2: Upper: Preliminary double differential pD → K+X

cross section at 2.60 and 2.81 GeV/c in comparison with model calculations using different ratios σn/σp (lines). Lower: Missing mass mX for pD → K+pX(psp) events at p=2.81 GeV/c (preliminary) in comparison with model calculations.

The resulting large value of σn/σp from the inclusive spectra is supported by the analysis of missing-mass spectra from pD → K+pX(psp) events (where psp denotes the unobserved spectator proton) from the same beam time. The spectrum for 2.81 GeV/c is also shown in Fig. 2 and is compared with the result of the Monte-Carlo simulations, again for different ratios σn/σp. The best

12

agreement between data and simulations is obtained for σn/σp ~(4–5). Study of light scalar resonances A primary goal of hadronic physics is to understand the structure of mesons and baryons, their production and decays, in terms of quarks and gluons. The non-perturbative character of the underlying theory – Quantum Chromo Dynamics (QCD)– hinders straightforward calculations. QCD can be treated explicitly in the low momentum-transfer regime using lattice techniques, which are, however, not yet in the position to make quantitative statements about the light scalar (JP = 0+) states. Alternatively, QCD inspired models, which use effective degrees of freedom, are to be used. The constituent quark model is one of the most successful in this respect. This approach treats the lightest scalar resonances a0/f0(980) as conventional q q states. However, they have also been identified with KK molecules or compact qq- qq states. It has even been suggested that at masses below 1.0 GeV a full nonet of 4-quark states might exist. Such possible deviations from the minimal quark model have a parallel in the baryon sector, where the recently discovered Θ+ state requires at least five quarks.

The existing data are insufficient to conclude on the structure of the light scalars and additional observables are urgently called for. In this context the charge-symmetry breaking (CSB) a0-f0 mixing plays an exceptional role since it is sensitive to the overlap of the two wave functions. It should be stressed that, although predicted to be large long ago, this mixing has not unambiguously been identified yet in corresponding experiments.

An experimental program has been started at COSY, which aims at exclusive data on a0/f0 production close to the KK threshold from pp, pn, pd and dd interactions - i.e. different isospin combinations in the initial state. The reactions pp → ppK+K- and pd → 3He K+K- have been measured at COSY-11 and MOMO, respectively, at excitation energies up to Q = 56 MeV above the KK threshold. However, mainly due to the lack of precise angular distributions, the contribution of the a0/f0 resonances to KK production remains unclear for these reactions.

During the first experiment which has been made at ANKE, the reaction pp → dK+ 0K has been measured exclusively (by reconstructing the 0K from the measured dK+ missing mass) at beam momenta of p = 3.46 and 3.65 GeV/c (Q = 46 and 103 MeV). These measurements crucially depend on the high luminosities achievable with internal targets, the large acceptance of ANKE for close-to-threshold reactions, and the excellent kaon identification with the ANKE detectors. The obtained differential spectra for the lower beam momentum are shown in Fig. 3.

The background of misidentified events in the spectra of Fig. 3 is less than 10%, which is crucial for the partial-

wave analysis. This analysis reveals that the 0KK + pairs are mainly (83%) produced in a relative S-wave (dashed line in Fig.3), which has been interpreted in terms of dominant a0+-resonance production.

Fig. 3: ANKE data for the reaction p(3.46 GeV/c}p →

dK+ 0K . The shaded areas correspond to the systematic uncertainties of the acceptance correction. The dashed (dotted) line corresponds to K+ 0K -production in a relative S-(P-) wave and the solid line is the sum of both contributions.

Data on the reaction pp → dπ+X have been obtained at ANKE in parallel to the kaon data. In contrast to the latter, where the spectra are almost background free, the π+η signal is on top of a huge multi-pion background. This makes the analysis of this channel more demanding and at present even model dependent. A total cross section of σ(pp → dπ+η) ~ 4.6 µb has been extracted from the data with a resonant contribution of σ(pp → da0+ → dπ+η) ~ 1.1 µb. Together with the cross section for the a0+ → K+ 0K channel this yields a branching ratio of B.R.( KK /πη) ~ (0.029 ± 0.005stat ± 0.02syst) which is in reasonable agreement with model calculations B.R.(

KK /πη) ~ 0.04 for this beam momentum, thus confirming the interpretation of dominant resonant K+ 0K production via the a0+ .

The experimental results on a0+ production in pp interactions can also be regarded as a successful feasibility test for a longer experimental program with the final goal to determine the charge-symmetry breaking a0-f0 mixing amplitude. These measurements will require the use of a photon detector, which is not yet available at COSY. However, it is planned to relocate the WASA detector from its current location at CELSIUS/TSL to COSY in summer 2005, which will then make these experiments feasible.

13

Heavy hyperon production Compared to the spectrum of nucleon resonances, the excitation spectrum of hyperons (Λ, Σ) is not well known. Of particular interest is the Λ(1405) where quark models have difficulties to explain its low mass, and which alternatively has been interpreted as a NK bound state. Precise data from hadronic interactions, in particular on the Λ(1405) mass distribution, are still lacking. The Σ(1480) is even less known, the Particle Data Group cites it as a “bump” with unknown quantum numbers.

The reaction pp → K+pY has been measured with ANKE at maximum COSY momentum of 3.65 GeV/c. At this momentum six hyperons can be produced: Y = Λ(1116), Σ(1192), Σ(1385), Λ(1405), Σ(1480), and Λ(1520). They can be identified at ANKE by detecting events of the type pp → K+pπ±X, i.e. charged pions from the heavy hyperon decays (like Σ(1480) → mΣ±π ), in coincidence with K+p pairs. Figure 4 shows the K+p missing mass distribution for selected events with mX = mΣ .

Fig. 4: Missing mass mY spectrum for the reaction pp →

K+pY. The solid line shows the result of a Monte Carlo simulation with contributions of the indicated heavy hyperons.

Also shown in the Figure is the result of a Monte-Carlo simulation. Significant contributions of all four heavy hyperons are needed to reproduce the shape of the measured distribution. The best agreement between data and simulation is obtained for the Σ(1480) mass being equal to 1470 MeV.

Planned measurements of the Θ+ pentaquark state In Feb. 2004 a measurement will be performed at ANKE aiming at the detection of the recently discovered Θ+ resonance with a mass of about 1540 MeV. It will be tested whether the Θ+ can be detected in the reactions pp → K0 pΛπ+, pp → K0 pΣ+, and pn → K0 pΛ at maximum COSY energy.

Meson Production in pn and dd interactions One advantage of meson production in nucleon–nucleon collisions is that, besides the spin of the particles, the isospin is an internal degree of freedom to be manipulated. A two-nucleon initial state can either be in an isotriplet configuration (T = 1 with T3 = +1, 0, –1 for pp, pn or nn states, respectively) or in an isosinglet state (T = T3 = 0 for pn). Variation of the initial-state isospin in pN → pNx reactions (where x denotes an isoscalar meson) may provide information about the production operator. In case of η production the observed ratio R = σtot(pn → pnη)/σtot(pp → ppη) ~ 6.5 is generally attributed to isovector dominance in model calculations based on meson exchanges. Data on the production of heavier isoscalar mesons (ω, φ) in pn interactions are not available yet.

At ANKE pn interactions can be studied using a deuterium cluster-jet target as an effective neutron target and detecting low momentum recoil protons (psp) in a silicon telescope placed inside the COSY vacuum close to the target. These recoil protons can be treated as ”spectators” that influence the reaction only through their modification of the kinematics.

Fig. 5: Measured total cross section for ω-meson production in pp (green) and pn (red triangles, preliminary) interactions in comparison with model calculations (blue lines).

Figure 5 shows the measured pn → dω cross section from ANKE in comparison with pp data from literature and model calculations as a function of the excess energy Q. The fact that the pn cross sections – although lying above the corresponding pp data – are significantly smaller than the theoretical predictions, suggests that the reaction mechanism for ω production differs from that for the η, possibly implying a relatively larger contribution from isoscalar meson exchange.

Figure 6 shows the result of a first beam time on the reaction pn → dη. The data have also been obtained with

14

the spectator-tagging technique and a clear η signal has been observed for Q values in the range 0…45 MeV. The data analysis, in particular the extraction of absolute cross sections close to threshold, is still in progress.

Fig. 6: Missing mass distribution for the pD → psp dX

reaction.

The ANKE data on η production in pn interactions are not only interesting in view of production mechanisms, but one may also expect information about the low-energy η-nucleon interaction. The strength of this interaction has led to speculations on the existence of quasibound η-nuclear states, even in the two-nucleon system. Such a state should show up as a threshold enhancement of η-production in nuclear reactions. Therefore, the data analysis and future measurements of pn → dη should aim at cross sections with high Q resolution, in particular for excitation energies smaller than 10 MeV.

An enhancement of the threshold production amplitudes has also been reported for the 3Heη and 4Heη final states. However, the data from literature do not provide precise information on differential cross sections, in both cases it has to be assumed that close to threshold only S-waves contribute. For dd → 4Heη only the total cross sections are available, and data for higher Q values are missing which could provide some information about the onset of P-waves. Measurements of the dd → 4Heη reaction have been started at ANKE, total cross sections together with angular distributions will soon be available for Q ≤ 43 MeV.

During a beam time in Feb. 2004, aiming at the study of a0/f0(980) production in pn interactions, also data on the reaction pn → dφ → dK+K- will be taken. This will allow one to determine the ratio R = σtot(pn → dφ)/σtot(pp → ppφ), using pp data from ANKE which currently are being analysed.

Deuteron-Breakup Studies The deuteron breakup reaction pd → ppn at GeV beam energies and in kinematics similar to backward elastic scattering pd → dp provides a new tool to investigate the pd dynamics at high-momentum transfer. This process has been investigated at ANKE for beam energies Tp = 0.6…1.9 GeV and forward emission of fast proton pairs with small relative energy Epp < 3 MeV. The results have

been compared with calculations based on a theoretical model previously applied to the pd → dp process. Using the the Reid soft core (RSC) and Paris potentials, the model only reproduces the measured cross sections at the lowest energy (Fig. 7). Recently, it has been shown that the use of the Bonn potential in the framework of the same model yields a much better agreement for all beam energies.

Fig. 7: Measured (•) and calculated (lines) cross sections

for the breakup reaction pd → ppn and of pd → dp backward elastic scattering (ο).

For further insight, additional data, in particular polarization measurements, are needed to provide a complete set of observables. A first measurement of the spin up/spin down asymmetry for the breakup process has been carried out, during which the EDDA detector has been used to obtain a reference value for the beam polarization. The analyzing power Ayp predicted by our model is expected to follow an approximately linear function of the neutron emission angle in the range from 180° to 166°. The preliminary measured slope differs by about two standard deviations from the value predicted by the RSC potential.

For the few-nucleon interaction studies with polarized beams and targets at COSY a polarized gas target of internal storage cell type is currently being developed.

The Charge-Exchange Reaction dr

p → (pp)n The complete description of the NN interaction requires precise data, in particular from double polarization experiments, for phase-shift analyses (PSA), from which the scattering amplitudes can be reconstructed. For the pp system such experiments have been carried out to beam energies of about 3.0 GeV, whereas much less information is available on spin observables in elastic np scattering, especially above 0.8 GeV. The PSA results for the isospin I = 0 NN system are poorly tested and measurements of any observable at small angles is highly desirable in order to improve the PSA solutions below 40° (c.m.s.). ANKE covers the angular range ϑc.m.s. ~ 0…30°.

15

Information on the spin-dependent np elastic amplitudes in backward direction (i.e. the Charge-Exchange (CE) region) can be obtained by measuring the CE breakup of polarised deuterons scattered on a hydrogen target.

The ANKE experimental program will first utilise unpolarised and tensor-polarised deuteron beams and an unpolarised hydrogen cluster target. The differential cross section will give the overall intensity of the spin-dependent parts of the CE process. The tensor-polarised deuteron beam allows one to separate the absolute values of three spin-dependent amplitudes. The use of transversely polarised deuterons impinging on a polarised hydrogen target and measuring the spin-correlation coefficient opens the possibility of obtaining the relative phases between the amplitudes. A first test measurement has been carried out with the polarised deuteron beam at an of energy Td = 1.2 GeV. Its aim was to check the feasibility of the experiment and to develop polarimetry of the COSY deuteron beam with ANKE.

For the dr

p → 3Heπ0 reaction the tensor analysing power T20 has been measured at SATURNE. These data can be used to determine the tensor polarisation of the deuteron beam. A preliminary result for the identification of this reaction is shown in Fig. 8.

Fig. 8: Missing mass squared for the reaction dr

p → 3He X. The results of the peak fit are indicated.

The Polarized Internal Target (PIT) for ANKE The development of the polarized atomic beam source (ABS) for ANKE has been finalized. Beam intensities of 7.8×1016 polarized hydrogen atoms per second in two hyperfine states have been reached, with polarizations of about 90%. In order to increase the target density for experiments at ANKE, the polarized beams from the ABS will be fed into a storage cell which will be located ~50 cm in front of the spectrometer dipole D2. Figure 9 shows the future location of the PIT at ANKE.

Fig. 9: Future position of the atomic beam source (vertical grey) and the Lamb-shift polarimeter at ANKE between the dipoles D1 and D2 (blue).

Installed at the PIT of the ANKE spectrometer, a Lamb-shift polarimeter (LSP) will be used to measure the nuclear polarization of the hydrogen or deuterium cell gas. A small fraction of the cell gas is deduced by a polarization-sample tube and fed into the LSP ionizer. Compared to the intensity of the directed beam from the ABS, the intensity of the incomming atoms drops by about 4 orders of magnitude. To study the feasibility of these measurements, a test setup of feeding, storage, and sample tubes was fed by the polarized H beam from the ABS. From the new ionizer the expected number of polarized protons (109 p/s ~ 0.1 nA) could be extracted. However, the background by unpolarized protons, stemming from residual gas, was higher than expected. Further measurements are underway.

The ANKE Frozen Pellet Target Preparation of the pellet target for operation at an internal target position of COSY has made significant progress in 2003. Droplets of frozen hydrogen (“pellets”) have been observed for the first time behind the exit sluice from the triple point chamber (TPC) and a continuous flow of liquid nitrogen droplets could be generated in the TPC and has been injected into the first vacuum chamber.

Figure 10 shows the flow of hydrogen pellets which could be generated during test runs of several days. The pellets have been produced from a liquid hydrogen jet with 60

16

µm diameter and a velocity of ~3 m/s. The frequency of the vibrating nozzle has been varied between 3–4 kHz.

Fig. 10: Frozen hydrogen droplets leaving the triple-point

chamber (TPC) of the ANKE pellet target into vacuum.

In another test run new metal nozzles with channels of 38 µm diameter have been used and, for the first time, nitrogen droplets have been observed in the TPC, see Fig. 11. For these tests the target was operated at nitrogen triple point conditions and a droplet generator frequency of 3 kHz. The temperature in the gas condenser was stabilized and controlled with the standard helium cooling system.

Fig. 11: Continuous flow of a liquid nitrogen jet in the

TPC breaking into droplets by acoustic excitation. Acknowledgements The work at the ANKE spectrometer has partially been supported by: BMBF (grants WTZ-RUS-211-00, WTZ-RUS-691-01, WTZ-POL-007-99, WTZ-POL-015-01, WTZ-POL-041-01), DFG (436 RUS 113/444, 436 RUS 113/561, 436 RUS 113/561); Polish State Committee for Scientific Research (2 P03B 101 19); Russian Academy of Science (RFBR99-02-04034, RFBR99-02-18179a, RFBR99-02-06518, RFBR02-02-16349), Russian Academy of Science; ISTC (1861, 1966).

System development for COSY experiments is done in close cooperation with the Central Laboratory for Electronics (ZEL). The main goal is to improve the efficiency, flexibility and standardization including state of the art technologies.

17

II. WASA

18

19

WASA – NEW OPPORTUNITIES FOR HADRON PHYSICS AT COSY

Introduction WASA, the ”Wide Angle Shower Apparatus“ will be relocated from the CELSIUS accelerator of The Svedberg Laboratory (TSL, Uppsala, Sweden) to COSY at the Research Center Jülich (Forschungszentrum Jülich, FZJ) in 2005.

The decision of the Swedish Research Council (SRC) in 2002 to terminate the status of TSL as a Swedish National Laboratory and to hand it back to Uppsala University has a bearing on the future of CELSIUS: it will have to stop operation latest in the middle of 2005. The major detector system at CELSIUS, WASA, will then be available for further use. The Institute for Nuclear Physics (Institut für Kernphysik, IKP) of FZJ has put forward the proposal of a transfer of WASA to COSY, which has been enthusiastically accepted by the WASA collaboration as well as the COSY users. Subsequently, a Letter of Intent for “WASA at COSY” has been submitted to the Pro-gram Advisory Committee (PAC) of COSY [see e.g. www.ankecvs.ikp.kfa-juelich.de/emc], which responded, saying it “very much welcomes this initiative from the whole WASA collaboration, the IKP, and the COSY users and awaits the submission of a full proposal”.

Forging the Future Meanwhile, a task force “WASA at COSY” has been installed to take care of the physics, technical and organizational issues. A seminar series at IKP, the CANU meeting in December 2003 and a special workshop in January 2004 to a large part have been devoted to a discussion of the physics program of “WASA at COSY”. It is foreseen that later during 2004, a new WASA collaboration will be founded, which will work out the requested proposal: everyone interested to join in is most welcome! For the time being as one of the most important and urgent tasks one has to come to a decision about where to install WASA at COSY – at the internal beam or at an external target position.

Detector Capabilities WASA is a unique detection system which has been built by a large international collaboration since about 1990 for use at the internal beam of CELSIUS. It comprises as the major components (see Fig.1):

• a Central Detector, comprising tracking de-tectors, a calorimeter and a superconducting solenoid,

• a Forward Detector • a Pellet Target, and • a Zero Degree Tagging Spectrometer.

The large solid angle electromagnetic calorimeter, made of 1012 CsI detectors, constitutes the main part of the central detector (see Fig. 2). Since the detectors that are currently operated at COSY do not have photon detection capability, this feature will in

particular be a new asset for the intended COSY physics program:

• Symmetry tests, rare decays • Baryon resonances, exotic hadronic states • Scalar mesons • Baryon-baryon interaction • Hadrons in the nuclear medium

We – IKP, the WASA collaboration, and the COSY user community – are all looking forward to this new exciting possibility for hadron physics with hadronic probes that will open up with “WASA at COSY”.

Fig.1: Schematic drawings of the WASA detector.

Fig.2: Photo of WASA showing the CsI electro-magnetic

calorimeter and the forward detectors .

20

21

III. Accelerator Physics

22

23

COSY - OPERATION AND DEVELOPMENTS

Introduction The sustained efforts of the COSY crew have been successful in further enhancing the environment for performing precision experiments in the field of hadron physics. This concerns besides others the available beam intensities for polarized and unpolarized proton beams, the provision of external polarized deuteron beams, an electronic feedback system capable to dampen the destructive oscillation of high intensity cooled beams, and the spin preparation and manipulation inside the COSY racetrack.

As in the years before vital support was given by the COSY team to collaborations setting up their experiments and in fine-tuning the beam properties to match difficult requirements. Additionally, irradiations were performed for external institutes.

Our institute also shared its expertise to advance accelerator development within the European framework. The HP-NIS project is targeted to expand the performance limits of High Performance Negative Ion Sources while our efforts in the framework of CARE (Coordinated Accelerator Research in Europe) is focused on the advancement of acceleration structures for high intensity pulsed machines.

Intensity Upgrade The suspension of the super-conducting linac project caused by the unprecedented budget cuts of the federal government did jeopardize a whole range of planned future experiments depending on very high intensity beams for polarized particles. To ameliorate the impact of this unforeseen set back for the experimental program a task group has been formed with the specific aim to exploit all conceivable avenues holding promise for higher intensities for polarized beams. From the present analysis it seems clear that not one single measure will result in a big leap forward instead the desired increase in intensity will be the accumulative effect of a great number of small steps. Nevertheless it still holds that the projected result of all measures combined will still fall short by a sizable factor to what one could have easily reached with the aid of a new injector.

To study in a first step the ion optical conditions tests had been performed with unpolarized beams. The results of the task group in this first phase are remarkable and encouraging. After several test runs one was able to raise the intensity for unpolarized protons inside the ring to 1.5⋅1011 for the PISA collaboration which is taking data at an internal experiment. To achieve such an intensity in the flat top it had been necessary to inject 3⋅1011 particles at injection energy. This value is highly notable because it is substantially above the original design limit of 2⋅1011, which was deduced from calculated space charge effects.

Allocation of Beams The diagram in figure 1 gives an overview of the usage pattern for different types of beams. Clearly the anticipated trend for an increased demand of polarized beams has continued and the significance of deuterons, polarized and unpolarized, has surged as they offer new scientific opportunities.

With about 94 % the reliability of the machine proved to be as high as in the years before, which has been a result of the high standards in machine maintenance.

Shutdowns have been used to install new equipment necessary for the operation of the machine or the improvement and extension of internal experiments. A statistical breakdown how time had been shared for the various activities is depicted in figure 2.

Great efforts have been invested in improving the cooling techniques and to overcome instabilities that limited the intensities of electron cooled beams with the aid of an electronic feedback system. This method offers now the possibility of cooling and stacking of polarized beams and boost by this the intensity. Another highlight has been the successful manipulation of the spin of circulating ions by using rf-dipoles. Such techniques will be of great importance in reducing systematic errors in future precision spin experiments.

u n p o la r iz e d D e u te ro n s

1 3 %

p o la riz e d D e u te ro n s

1 1 %

u n p o la r iz e d P ro to n s

5 4 %

p o la r iz e d P ro to n s

2 2 %

Fig. 1: Distribution according to ion species

External Exp.18%

Internal Exp.46%

MD / Exp. Setup22%

Maintenance & Shutdown

14%

Fig. 2: Sharing the available time for the different tasks, MD means machine development

24

BEAM FEEDBACK SYSTEM FOR COOLED IONS

Introduction For precision experiments in hadron physics cooled beams are indispensable to ensure a narrowly defined phase space. This holds for the transversal as well as for the longitudinal dimension. At the same time the experiments demand also highest intensities to be able to investigate reactions with small cross sections.

Principle of Electron Cooling The basic principle of electron cooling is sketched in figure 1.

The coasting beam inside COSY after injection occupies a rather large phase space. To shrink it the beam is brought into close contact with a highly ordered electron beam having the same speed but whose intensity is several order of magnitude higher. To increase the cooling force the electrons spiral in narrow circles along the field lines of a longitudinal field which has been carefully shaped to high uniformity. Due to the electron-proton mass ratio the energy of those electrons needs only to be about 1/1836 of the proton energy, which technically eases the task to produce electrons having a minute phase space. Although the cooling length is only about 1% of the rings circumference the cooling process being applied millions of time is able to shrink the proton phase space within a couple of seconds to a small fraction of the original size. The technical realization is depicted in figure 2 showing the two toroidal magnets connected with a large solenoid in which the cooling takes place. The high voltage terminals extent beyond the concrete shielding and are not visible.

In order to make the cooling process work as intended a very careful and delicate tuning of the accelerator's ion optics is necessary. An important monitoring device for the tuning procedure are wire chambers and scintillators that measure the trajectories and number of neutral hydrogen atoms formed inside the electron cooler that are positioned after de dipoles following the straight section that contains the electron cooler. Although the neutralization of protons by catching an electron is a rare

process the number of hydrogen atoms formed is entirely sufficient to look with great precision to changes made to the characteristics of the coasting beam.

Stability Problems Unfortunately, a significant limitation turned up for electron-cooled beams because of instabilities resulting in beam loss if proton intensities went beyond a certain limit. An investigation revealed that there are two different particle loss mechanisms. One is an incoherent loss just after injection that is due to the interplay of the electric field of the electron beam and ions with large betatron amplitudes that pass the cooler section without overlap with the electron beam. Another loss mechanism apearing after several seconds is a result of coherent betatron oscillations, which are present in both planes. These instabilities can be understood in the framework of criteria that need to be met for high-density ion beams to ensure transverse and longitudinal stability. Oscillations are fanned both by the increase of the space charge impedance, a consequence of the reduction in transverse beam extension, and the weakening of the Landau damping because of the shrinking momentum spread during the cooling process..

Fig. 1: Principle of electron cooling

Fig. 2: View of the COSY electron cooler

25

Transverse Beam Feedback To counter these detrimental effects a feedback system has been conceived to actively dampen those oscillations. Fig. 3 shows the schematic for the electronic circuit used at COSY.

The effect of such beam oscillations can be seen in figure 4. Plotted are the signals from the beam position monitors for the vertical and the horizontal directions and the intensity of the beam as measured by the beam current transformer (BCT).

Just after injection incoherent losses occur. Coherent losses start after 20 s and the beam position monitor (BPM) ∆-signals indicate strong horizontal and vertical oscillations. The dominating particle loss is due to the vertical oscillation.

The transverse feedback (FB) system allows to dampen coherent beam oscillations without any change of the machine optics. A so called “waterfall” frequency spectrum without activated feedback system and the damping after it has been switched on can be seen in figure 5. The time progresses from top to bottom. The frequency extends up to 8 MHz. The effectiveness of the active beam damping is clearly demonstrated. The important advantage of such a system is that it allows to boost intensities for cooled beams by significant factors.

The process of injecting and cooling the beam with and without a feedback is graphed in figure 6.

The upper trace represents the intensity of the cooled beam as measured by the Ho-count rate. By avoiding oscillations the current is boosted 6 fold.

Another important application of the feedback system is the stacking of cooled ion beams by repeated injections as shown in figure 7.

The vertical FB system allowed to stabilize the cooled proton beam at a level of 2⋅1010 particles (1,8 mA) after a single injection. After about 20 successive injections the intensity levels off. The stacking technique enabled the accumulation of 1,2⋅1011 cooled protons (9,2 mA) at injection energy. This method is especially valuable for internal experiments that take data for long periods before they need a refill of the ring.

Fig. 3: Schematic of the feedback system 1 pick-up, 2 pre-amp, 3 analog processing, 4 delay, 5 power amp, 6 splitter, 7 kicker

I o n B e a m

Fig. 4: Beam behavior during electron cooling. The beam current transformer (BCT) signal (black) is plotted together with horizontal (red) und vertical (blue) BPM ∆-signals. Time scale 5 s/div

8 MHz 0 Hz Frequency

Fig. 5: Frequency spectra of the vertical beam position monitor signal with and without feedback

Fig. 6: Countrate of Ho-atoms formed inside the cooler for

different settings of the feedback

Fig. 7: BCT signal during stacking of the cooled proton beam with active FB. Time scale 50 s/div. Intensity scale 2 mA/div

26

SPIN@COSY: SPIN MANIPULATION OF POLARIZED PROTONS AND DEUTERONS STORED IN COSY

Introduction During the past decade, polarized proton and electron beam experiments have become an important part of the programs in storage rings such as the IUCF Cooler Ring, AmPS at NIKHEF, the MIT-Bates Storage Ring, COSY, LEP at CERN, RHIC at BNL and HERA at DESY. Many polarized scattering experiments require frequent spin-direction reversals (spin-flips), while the polarized beam is stored, to reduce their systematic errors. Spin resonances induced by either an rf-solenoid or rf-dipole can produce such spin-flips in a well-controlled way. At high energy, the spin-flipping efficiency with an rf-dipole should be essentially independent of energy due to the Lorentz invariance of a dipole magnet's ∫B dl; this is important for very high energy rings and colliders with polarized beams.

Investigating the spin-manipulation of polarized deuterons is an important step towards polarized deuteron beams, and thus polarized neutron scattering experiments. Spin flipping and spin manipulation of a simultaneously vector and tensor polarized stored deuteron beam was first studied in the IUCF Cooler Ring at 270 MeV.

In 2002, the SPIN@COSY collaboration was founded to continue these unique polarized beam studies in the GeV-regime at COSY; the collaboration includes: University of Michigan, IKP of Forschungszentrum Jülich, Universität Bonn, Universität Hamburg, KEK and Brookhaven. In February 2003, a new air core rf-dipole was used to spin- flip 1.85 GeV/c vertically polarized deuterons stored in COSY, by ramping the rf dipole's frequency through an rf-induced spin resonance. A similar experiment was done in April 2003, with 1.94 GeV/c polarized protons stored in COSY. This paper summarizes the results of these two runs of SPIN@COSY, and describes the new ferrite rf dipole used in the December 2003 polarized deuteron run.

Spin manipulation of polarized beams In any flat circular accelerator or storage ring, each particles’ spin precesses around the Stable Spin Direction (SSD). With no horizontal magnetic fields in the ring, the SSD points along the vertical fields of the ring's dipole magnets. The spin tune νs, which is the number of spin precessions during one turn around the ring, is given by :

,γν Gs = (1) where G = (g-2)/2 = 1.792847 and 0.1426 are the proton and deuteron gyromagnetic anomaly, respectively, and γ is its Lorentz energy factor.

The polarization can be perturbed by the horizontal rf magnetic field from an rf-dipole. This perturbation can induce an rf spin resonance, which can flip the spin of the stored polarized beam; the resonance's frequency is

),( scr kff ν±= (2) where fc is the particles’ circulation frequency and k is an integer. Sweeping the rf magnet's frequency through fr

can cause a spin-flip. A modified Froissart-Stora formula can relate the beam's initial vector polarization Pi, to its final polarization Pf, after crossing the resonance:

2( )(1 ) exp/

f c

i

P fP f t

π εη η η ≡ − = + − − ∆ ∆

) , (3)

where η is defined as the upper limit of η), which is the measured spin-flip efficiency; when the exponential is zero, then η) = η. The ∆f /∆t is the resonance crossing rate, where ∆f is the ramp's frequency range during the ramp time ∆t. The resonance strength ε is given by:

(1 )1 ,

2 rmse G B dl

pγε

π+

= ∫ (4) where e is the particle's charge, p is its momentum, and ∫Brms dl is the rf-dipole's rms magnetic field integral.

In addition to the ordinary vector polarization components of spin-1/2 particles, such as protons and electrons, the polarization of spin-1 particles, such as deuterons, have both a vector and tensor polarization.

Experimental setup The layout for SPIN@COSY is shown in Fig. 1. The 6-turn air-core copper-coil rf dipole was installed around a ceramic chamber; it was part of an LC resonant circuit, which ran near 4.3 kV rms and provided an ∫Brmsdl of about 0.15 T.mm rms. The polarized beam emerging from the H-/D- ion source was accelerated by the Cyclotron to COSY's injection energy. The Low Energy Polarimeter (LEP), between the Cyclotron and COSY, monitored the beam's polarization before injection into COSY. The EDDA detector measured the beam polarization before or after the spin-manipulation.

Fig. 1: Layout of the Cooler Synchrotron Storage Ring

COSY, with its injector Cyclotron JULIC and polarized ion source. Also shown are, the rf dipole, the EDDA detector and the Low Energy Polarimeter.

27

Spin manipulation of stored polarized deuterons The February 2003 run at 1.85 GeV/c used COSY’s first stored polarized deuteron beam. We spin-flipped it by ramping the rf dipole's frequency through an rf-induced spin resonance to manipulate the polarization direction of the deuteron beam. We reduced the systematic errors by cycling the deuteron source through five states with nominal vector polarization values of 0, -2/3, -1/3, -1, +1. Since no 1.85 GeV/c d-p analyzing power data was available, we could only measure the asymmetries of deuteron scattering into the EDDA detector’s six regions, which are related to the vector and tensor polarizations. The following procedure was used to optimize the rf-dipole’s parameters for the highest spin-flip efficiency: we experimentally determined the resonance’s frequency; then set the rf dipole's ∫Brmsdl to its maximum; and then varied its frequency ramp time ∆t and frequency range ∆f.

The rf dipole frequency was first centered at the spin resonance’s approximate location of fr = fc (1−|νs|). We then ramped its frequency through ±100 Hz around the calculated fr = 917.4 kHz; we next made ±100 Hz ramps on each side of the previous frequency range until the beam was depolarized, as shown in Fig. 2. For these time-consuming studies, we cycled the deuteron beam through only the +1 and –1 vector polarized states. After each rf dipole ramp we measured the left-right deuteron scattering asymmetry, which was linearly proportional to the beam’s vector polarization. We then plotted in Fig. 2 the ratio of the average asymmetries measured before and after ramping the rf dipole.

Fig. 2: The measured deuteron’s final-to-initial vector polarization ratio at 1.85 GeV/c is plotted against each ramp’s frequency range ∆f shown by a horizontal bar.

These data show that the resonance was near 916.85 kHz, which is close to the calculated fr = 917.4 kHz. We then mapped the resonance by measuring the asymmetry, with the rf dipole at some fixed frequencies near 916.85 kHz. This mapping data showed a wide and shallow dip also

centered at 916.85 kHz; its wide shape may be due to the rather weak rf dipole and the non-zero spin tune width.

After setting the rf dipole’s ∫Brmsdl at its maximum and its frequency range at ∆f = ±50 Hz, we spin-flipped the beam by varying its frequency sweep time ∆t. The measured left-right scattering asymmetries for all five deuteron spin states are plotted in Fig. 3; the curves show the fit to Eq. (3) for each polarization state. Notice that all five curves cross at the same point near 100 s. This indicates that the left-right scattering asymmetry was indeed linearly proportional to the state’s vector polarization. Fig. 3 shows that the vector polarization was partially spin-flipped for the two longest ramp times of 200 and 400 s.

Fig. 3: The measured left-right scattering asymmetry for

the 5 deuteron spin states at 1.85 GeV/c is plotted vs. the rf-dipole’s ramp time ∆t, with ∆f/2 = 50 Hz. The curves are fits using Eq. (3).

The data in Fig. 3 suggest that increasing the ramp time even further would probably not significantly increase the spin-flip efficiency. This could be because the ±50 Hz frequency range did not fully overlap the resonance, which limited the maximum spin-flip efficiency. To analyze the data, we first subtracted from each asymmetry data point the unpolarized offset, shown by a solid line in Fig. 3, and then divided by its initial asymmetry. We then averaged these ratios for all four polarization states; this gave a spin-flip efficiency η = 48 ± 1% for the ramp time of 400 s.

Spin manipulations of stored polarized protons In the April 2003 Run, 1.941 GeV/c protons were stored in COSY; their circulation frequency was fc = 1.471 MHz, giving a nominal Lorentz energy factor of γ = 2.2977. Thus, Eq. (1) gave a spin tune νs = G γ of 4.1195; the resulting k = 5 depolarizing resonance frequency fr was:

(5 ) 1295.4 ,r cf f G kHzγ= − = (5) We determined fr by linearly ramping the rf-dipole's frequency by ∆f/2 = ±2 kHz around the calculated fr; we then made ±2 kHz ramps next to each side of the previous frequency range until the beam was either spin-flipped or

28

depolarized, as in Fig. 2. The data suggested that the resonance width was comparable to the ±2 kHz frequency ramps. Thus, we next studied ∆f/2 = ±1kHz frequency ramps with different central frequencies. These ±1 kHz data gave fr = 1306.0±0.5 kHz and an upper limit width of w = 2.3±0.9 kHz FWHM, when fit to a first-order Lorentzian. The width due to the resonance's ε, given by w = 2 ε fr, is only 59±3 Hz. However, the beam had a measured momentum spread ∆p/p of about 6*10-4 FWHM; this adds a width of about 3 kHz in Eq. (5), which may dominate the measured upper limit of w = 2.3±0.9 kHz.

We then spin-flipped the proton beam by ramping the rf-dipole's frequency, with various ramp times ∆t, through the frequency range ∆f/2 = ±5 kHz, which seemed to safely cover the whole resonance; we measured the polarizations after each ramp. The measured data are plotted against the ramp time in Fig. 4; this suggested setting ∆t near 7 s. Since the spin-flip efficiency is never exactly 100%, the Modified Froissart-Stora Equation, Eq. (3), was earlier introduced to describe non-ideal single-flip data. Fitting the Fig. 4 data to Eq. (3) gave a spin-flip efficiency η of 100.8 ± 1.2% and a resonance strength of (20.7 ± 0.2)10-6; this ε is consistent with the ε of (20±1)10-6 obtained from Eq. (4) with ∫Brms dl = 0.11 ± 0.005 T·mm.

Fig. 4: The measured proton polarization at 1.941 GeV/c

is plotted against the rf-dipole ramp time ∆t. The rf-dipole's frequency half-range ∆f/2 was 5 kHz, and its ∫Brms dl was 0.11 Tmm. The curve is a fit using Eq. (3).

To more precisely determine the spin-flip efficiency, we then measured the polarization after 11 spin-flips, while varying the rf-dipole's rms ∫Brmsdl, its ramp time ∆t, and its frequency half-range ∆f/2. This technique enhanced small changes in the spin-flip efficiency's dependence on the rf-dipole's parameters, because the 11th power, of even a small single-spin-flip depolarization, is large.

Fig. 5: The measured proton polarization at 1.941 GeV/c

is plotted against the number of spin-flips. The rf-dipole's frequency ramp time ∆t was 10 s; its frequency half-range ∆f/2 was 4 kHz, and its ∫Brmsdl was 0.11 T·mm. The line is a fit using Eq. (6).

After setting ∆t, ∆f and ∫Brmsdl to maximize the spin-flip efficiency, we then measured it precisely by varying the number of spin-flips. The vertical polarization after 0, 1, 11 and 30 spin-flips, was measured while keeping ∆t, ∆f and ∫B dl all fixed; these data are plotted against the number of spin-flips in Fig. 5. We fit these data using the measured η) defined by:

( ) ,nn iP P η≡ −)

(6) where Pn is the measured polarization after n flips. The fit gives a measured spin-flip efficiency of η) = 99.3±0.1%. Note that, when the exponential in Eq. (3) goes to zero, the equation yields η = η) for one spin-flip.

Summary and Conclusion In February 2003, by adiabatically ramping a new air-core rf-dipole's frequency through an rf-induced spin resonance, we were able to spin-flip the polarization of COSY’s first stored deuteron beam. The vector spin-flip efficiency at 1.85 GeV/c was η = 48±1% with a ±50 Hz frequency ramp and a ramp time of 400 s. Since no tensor analyzing power data was available at 1.85 GeV/c, we could not obtain absolute values of the deuteron's tensor polarization from the measured data.

In April 2003, with stored 1.941 GeV/c polarized protons, we obtained a 99.3±0.1% measured spin-flip efficiency by ramping the air-core rf dipole though an rf spin resonance with a ∆f/2 = 4 kHz frequency ramp and 10 s ramp time.

29

We recently increased the rf-dipole's strength by modifying its coil design, enclosing it in a ferrite box, and adding water-cooled coils to allow running at a higher current. The new ferrite rf dipole, shown in Fig. 6, had about four times more ∫Brmsdl (0.58 T·mm). Thus, when it was used in a December 2003 run with 1.85 GeV/c polarized deuterons, their vector spin-flip efficiency was increased to well above 90%, their vector and tensor analyzing powers were calibrated, and their tensor spin manipulation was studied in considerable detail. Acknowledgements: This research was supported by grants from the U.S. Dept. of Energy and the German BMBF Ministry.

Fig. 6: New ferrite rf dipole with an ∫Brmsdl of 0.58 T·mm. When it was used in the December 2003 polarized deuteron run, it produced a deuteron spin-flip efficiency of well over 90%.

30

31

IV. Theoretical Investigations

32

33

THEORETICAL INVESTIGATIONS Introduction The least understood part of the Standard Model of the strong and electroweak forces are the strong interactions at low energies. Through the study of hadrons and their interactions one aims at unraveling the mechanisms behind color confinement and chiral symmetry breaking. A precise knowledge of nuclear interactions and structure is also of importance in other fields like astrophysics and cosmology. Open questions in hadron physics The basic theory of the strong interaction, Quantum Chromodynamics (QCD), has two special symmetries, color and chirality, that determine, to a large extent, the interaction between the fundamental constituents of hadronic matter, quarks and gluons. For processes involving large momentum transfers (the ultraviolet regime), QCD can be treated accurately in perturbation theory, in close analogy to Quantum Electrodynamics. This is due to an important property of QCD, asymptotic freedom, which implies a decrease of the coupling strength between quarks and gluons as the momentum transfer increases. In the infrared regime (energies up to a few GeV), the so-called non-perturbative regime, the coupling becomes too strong to allow for a perturbative treatment of QCD. To make predictions for non-perturbative phenomena is a great challenge for hadron theory. One of the most exciting aspects of non-perturbative hadron physics is known as confinement: Quarks and gluons, the fundamental particles of QCD, have never been observed in isolation. The only strongly interacting particles that can be counted in detectors are color-neutral composites of quarks and gluons, i.e. baryons and mesons. An explanation of confinement is a fundamental goal of hadron physics closely linked to an understanding of the spectrum of mesons, baryons as well as exotica. Another basic question concerns the origin of the spontaneous breaking of chiral symmetry, the hadron masses and the values of the quark condensates. Strategies in Hadron Physics The hadron physics community focuses on three major activities in order to improve our understanding of the strong interaction in the non-perturbative regime: 1. Reactions of few-hadron systems allow to study

hadron dynamics in the non-perturbative regime of QCD in detail. Experimental research projects are performed at electron and hadron facilities, where one investigates reactions at low to moderate momentum transfers. One of the open problems addressed here is the question whether hadrons are limited to quark-antiquark and three-quark bound states or whether other, so-called exotic combinations of quarks, antiquarks and gluons, such as glueballs and dibaryons, which are allowed from a theoretical point of view, are also realized in nature.

2. Heavy ion scattering and proton-nucleus reactions introduce temperature and density as external, tunable variables of hadronic matter. This opens up a qualitatively new alley for investigating, for example,

non-perturbative hadron condensate quantities through their functional dependence on the external variables.

3. Deep inelastic lepton scattering probes hadrons at large momentum transfers. This is the kinematical region where perturbative QCD applies. It has turned out, however, that reactions involving spin degrees of freedom and semi-inclusive reactions as well as deep inelastic scattering off heavy nuclei are sensitive to non-perturbative physics.

Theory Projects The Jülich theory group has shifted the theoretical methods from the application of specific models of hadron structure and dynamics to the development of effective field theories. These are field theories which are based on the relevant degrees of freedom in accordance with the symmetries of the underlying theory and which allow systematic expansions with respect to a set of small parameters, like e.g. external momenta and quark masses. The most studied example is chiral perturbation theory which utilizes the spontaneously broken chiral symmetry of QCD to define an expansion scheme of the strong interaction amplitudes in external momenta and masses of the Goldstone bosons, i.e. the pions, kaons and eta. The study of the symmetry breaking mechanisms allows insight into the role played by quark condensates in the generation of the non-perturbative part of the hadron masses. At IKP, EFTs for pion-nucleon processes and few nucleon systems have been successfully developed. The next step will be the development of an EFT for meson production, which will allow for a model independent analysis of the many precise data from COSY, IUCF and TSL. This is a long term (~ 5 year) project that can only be undertaken within the Helmholtz infrastructure.

The breaking of the isospin symmetry is a way to extract quark mass ratios from the hadronic observables, provided one has a systematic framework to disentangle electromagnetic (virtual photon effects) and strong (quark mass effects) isospin violation. This is only possible within chiral EFT. The Jülich theory group is involved in the theoretical analysis of the reactions dd → απ0 as well as the forward-backward asymmetry in pn → dπ0 recently measured at IUCF and TRIUMF. Moreover, it supports the experimental program at COSY by studying the possible mixing of the lightest scalar mesons: the isovector a0(980) and the isoscalar f0(980) and investigating how to extract the charge symmetry breaking a0-f0 transition amplitude - a quantity believed to shed light on the nature of these controversial resonances.

The study of strangeness in hadrons and nuclei is of utmost importance to understand the strong interaction vacuum (the size of the 3-flavor quark condensate) and its excitations. Since particles that contain strange quarks decay due to the weak interaction, direct scattering experiments are very difficult and consequently the low energy scattering parameters like e.g. the hyperon--nucleon scattering lengths are badly determined. Production reactions with a hyperon-nucleon system in

34

the final state opens a promising alternative. Therefore the theory focuses on the theoretical study of hyperon production with particular emphasis on polarization observables. It was recently demonstrated that for example the ΛΝ scattering lengths can be extracted from the reaction pp → Λ pK+ with an uncertainty of 0.5 fm for an experiment with resolution of 2 MeV, a number that can be reached at COSY. It should be stressed that the spin dependence of the hyperon-nucleon interaction is not only interesting in itself, but also is relevant for the formation of hypernuclei as well as the matter decomposition of neutron stars. In a more general framework, the TH group also addresses the analysis of flavor symmetry breaking in various observables to get further bounds on the size of the strange quark mass.

Excited baryons promise insight into non-perturbative hadron dynamics, in particular because of possible non-q3 candidates, such as the Roper resonance or the recently found pentaquark. Theoretical investigations connect electromagnetically-induced meson production on the nucleon with the meson production via the hadronic reactions. An important aspect of the theoretical approach is to pro