STUDY OF PHOTOPRODUCTION OF NEUTRAL MESONS USING
THE CRYSTAL BARREL AT ELSA
PNPI participants of the CBELSA/TAPS Collaboration:
D.E. Bayadilov, Yu.A. Beloglazov, A.B. Gridnev, I.V. Lopatin, D.V. Novinsky,
A.K. Radkov, V.V. Sumachev
The problem of baryon spectroscopy which has not been solved fully yet is the nature and origin of the so
called “missing resonances”. As a matter of fact, till now the number of the excited states of baryons
observed experimentally is significantly less than that predicted by different theoretical models. The reason
of such non-consistence may be two-fold. One possibility – such resonance states non-observed in
experiments do not exist in reality as a consequence of a “hidden” symmetry. The quark-diquark structure of
baryons might be considered as an example of such a “hidden” symmetry. Such a structure could freeze one
of the internal degrees of freedom and, hence, would decrease the number of expected resonances. Another
possibility with less dramatic consequences – some of observables simply have not been found in
experiments. Till recently, practically all information on characteristics of baryon resonances were extracted
from experimental data on elastic scattering using a partial-wave analysis. However, if some of
resonances have anomaly weak coupling with the  channelit is impossible to find such resonances in
these experiments. On the other hand, the information on resonances coupling weakly with the channel
may be obtained from experiments on photoproduction of neutral mesons. Thus, the photoproduction of
baryon resonances provides an alternative tool to study nucleon states.
The main component of the experimental set-up which is used for studying photoproduction of neutral
mesons at the tagged photon beam of the electron accelerator ELSA in Bonn is the multi-crystal photon
detector Crystal Barrel (CB). The CB at ELSA is an ideal tool to study nucleon resonances. In particular, it
provides a wide angular coverage, which is needed to separate the resonant from non-resonant contributions.
Fig. 1. Schematic view of the multi-crystal photon spectrometer Crystal Barrel
The experiment was carried out at the electron accelerator ELSA with the maximal available electron
energy of 3.2 GeV. Electrons hit a radiator target, where they produced Bremsstrahlung. The energy of
photons, in the range between 25% and 85% of the primary electron energy, was measured by detection of
the corresponding scattered electrons in the tagging system (“tagger”). It consisted of a dipole bending
1
magnet and two detecting parts – an array of fourteen scintillation counters providing fast timing information
and two multiwire proportional chambers (MWPCs) with a total of 348 channels to provide information
about trajectories of the scattered electrons. Each scintillator of the array was observed by two
photomultiplier tubes attached to both sides. A logical OR of the left-right coincidence from all the
scintillators was required in the first-level trigger.
The tagger was calibrated by direct injection of a very low intensity e-beam of 600 or 800 MeV, after
removing the radiator. Variations of the magnetic field of the tagging dipole magnet enabled us to execute
a scan of several spatial positions over the MWPCs. For a given wire, the electron momentum is proportional
to the magnetic field strength. The calibration was checked by a Monte Carlo simulation of electron
trajectories through the tagger. The set-up geometry, dimensions of the electron beam, angular divergences,
multiple and MØller scattering in the radiator foil and air were taken into account. From these simulations,
the energy corresponding to each MWPC wire was obtained by a polynomial fit. The uncertainty in the
simulation was estimated to be of the same order of magnitude as the “energy width” of the respective wire.
The CB is the central component of the experiment. It consists of 1230 CsI(Tl) crystals of 30 cm length
(16 X0) and has an excellent photon-detection efficiency. Large solid-angle coverage (excluding the 12
openings on both sides of the barrel necessary for technical reasons) and high granularity allow to
reconstruct multi-photon final states. The spatial resolution determined by the size of individual crystals is
about 20 mrad (1.1); it allows separation of two photons stemming from a decay of 0 mesons with the
maximum momentum 1 GeV/c corresponding to a minimum opening angle of 16.6. The energy resolution
of the calorimeter was empirically described as σE /E = 2.5% / [E(GeV)]0.25.
The three-layer scintillating fiber (scifi) detector was inserted into the inner cavity of the CB to identify
charged particles leaving the target and to determine their intersection point with the detector. The fibers
were 2 mm in diameter; one of the layers was flat, the fibers of other two layers encircled the target over
25 with respect to the first layer. The innermost layer covered the solid angle of 92.6% of 4.
A coincidence between the tagger and the scifi detector provided the first-level trigger of the experiment.
From the hit pattern in the CB, the fast cluster logic determined the number of “particles” defined by the
clusters of contiguous crystals with individual energy deposits above 15 MeV. The second-level trigger was
generated for events with two or more particles determined by the cluster logic. In the data analysis, the
clusters with two local maxima were split into two particles sharing the total energy deposit. The offline
threshold for the accepted particles was set to 20 MeV. The proton kinetic energy had to exceed 35 MeV to
traverse two inner scifi layers and to produce a trigger. A proton of 90 MeV was needed to reach the barrel
calorimeter and to deposit the minimum cluster energy of 20 MeV. The CB acceptance was determined by a
Monte Carlo simulation. It vanished for forward protons leaving the CB through the forward outlet and for
protons with very low energies.
At the first stage of experiments, the CB was used for measuring the total cross sections and the
differential cross sections (angular distributions) for photoproduction reactions  p  p0,  p  p and
 p  p00. The experiments were carried out with primary electron energies 1.4, 2.6 and 3.2 GeV,
providing a total range of photon energies from 0.3 to 3 GeV. Introducing into the partial-wave analysis new
cross sections, measured for the above mentioned photoproduction reactions, one can search baryon
resonances which were not known earlier and improve our knowledge of properties of the known
resonances. Just this way was chosen by the group of physicists from Bonn and Gatchina (group Bn-Ga) for
processing new data on the cross sections for photoproduction of neutral mesons.
The next logical step of experiments which are carried out with the СB spectrometer (in combination with
the TAPS) is to use beams of linearly (or circularly) polarized tagged photons produced at the electron
accelerator ELSA. Linearly polarized photons are produced by coherent bremsstrahlung from a thin
(100 μm) diamond radiator which is mounted on a goniometer where its orientation with respect to the beam
can be adjusted precisely. The degree of polarization is then deduced from the coherent peak which is
measured in the tagging system. The maximal value of the linearly beam polarization is about 50%.
Circularly polarized photons are produced via incoherent bremsstrahlung of longitudinally polarized
electrons off an amorphous radiator.
In order to have a possibility to detect photons and charged particles emitted from the target with high
angular and energy resolution, the main detector of the set-up – the multi-crystal photon spectrometer Crystal
2
Barrel – was equipped in addition with two forward detectors (see Figs. 2, 3) built specially for this purpose.
One of them (the so called Forward Plug) assembled of 90 CsI(Tl) crystals was inserted into the open
forward cone of the Crystal Barrel for registering particles emitted at angles from 10 to 30 and, thus, for
increasing the total covered solid angle. 180 overlapping veto-counters made of thin scintillators were placed
in front of the CsI(Tl) crystals in order to determine the kind of a particle (photon or proton) entering each
crystal. Another forward detector (see the right part of Fig. 3) is a “wall” consisting of 216 BaF2 crystals and
216 veto-counters located in front of the crystals. The central crystal was removed to make a hole for the
photon beam. The Mini-TAPS registers particles (photons and protons) emitted at angles from 1 to 10.
Fig. 2. Forward Plug
Fig. 3. Schematic view of the experimental set-up for measuring the beam
asymmetry
With the use of the linearly-polarized photon beam and the experimental set-up CB (Crystal Barrel with
the Forward Plug + TAPS), measurements of the beam asymmetry Σ were performed for 0, , and 0
photoproduction [1–6]. A combination of the polarized photon beam and the above mentioned set-up enabled
high-resolution detection of multiple photons, important for clean detection of 2γ and 30 decays of 0 and 
mesons. As an illustration, we present in Fig. 4 (taken from Ref. [1]) the angular dependence of the beam
asymmetry Σ measured for the reaction  p  p. A continuous 3.2 GeV ELSA electron beam was used to
produce a linearly polarized tagged photon beam by means of coherent bremsstrahlung off a diamond
crystal, covering a photon energy range E = 800 –1400 MeV with the maximal achievable polarization up to
49% at the energy of 1305 MeV. Both dominant decay modes of  meson  2 and   3 0 were used to
extract the beam asymmetries from the azimuthal modulation of the cross section. The measurements cover
the angular range from 50° to 150°.
In Fig. 4, our new data are compared to two standard calculations, the Mainz isobar model eta-MAID and
the Bn-Ga partial-wave analysis. In contrast to eta-MAID, the Bn-Ga analysis in addition to N also takes
the N, K and KΣ coupled channels into account. To calculate the photon asymmetry, the preliminary high
energy GRAAL data have already been used in the Bn-Ga fit. This might be the reason for a slightly better
description of our data. The overall agreement between the data and both models seems quite satisfactory at
first glance. However, a closer examination reveals distinct inconsistencies. While the full model results
agree with each other, the individual resonance contributions differ substantially, as is illustrated in Fig. 4.
This unsatisfactory situation cannot be resolved using the measurements of the photon asymmetry alone.
Yet, such data provide the necessary basis to be extended with double polarization observables in order to
get closer to, or even to accomplish the complete experiment in terms of an introductory discussion. Such an
extended program is underway now at ELSA within the Collaborative Research Project SFB/TR-16.
3
Fig. 4. Angular dependence of the beam asymmetry Σ for the reaction p  pη at E = (1250 ± 50) MeV. Comparison of
calculations with the experimental data. Left: the eta-MAID result is shown. Right: the result of the BnGa analysis. The full lines
represent the respective full calculations. The broken curves illustrate the impact of “turning-off” individual resonances: Long
dashed – without P13 (1720), long dash-dotted – without P11 (1710) (no difference to full calculation in the BnGa analysis), short
dashed – without D31 (1520), and short dash-dotted – without D15 (1675)
After completing the above described stage of the experiments on photoproduction of neutral mesons,
measurements of the double polarization observables were started. Such kind of measurements provide
physicists with a unique information for identification of different baryonic states, especially in the region of
overlapping resonances where “missing” baryonic states could be found.
These complicated experiments need both a beam of polarized photons and a polarized proton target. To
realize this, an upgrade of the experimental set-up was performed. First of all, the Crystal Barrel was moved
to another beam-line – Fig. 5. The tagging system installed earlier at this line was equipped with 96 new
scintillation counters (with overlapping scintillators of varying width) and with an array of 480 scintillating
fibers (two layers of 2 mm diameter fibers). This allowed to tag photons with 18% to 95% of the incoming
electron energy with an energy resolution between 0.2% and 2.2%.
Fig. 5. Disposition of the experimental set-up after its upgrading and replacing to another beam-line
4
The displacement of the Crystal Barrel to another beam-line allowed to insert a longitudinally polarized
proton target into the Crystal Barrel inner cavity and to arrange its operation in the “frozen spin” mode by
locating a system of polarization “pumping” outside the Crystal Barrel. This system consists of a horizontal
dilution refrigerator and a 5 T superconducting polarization magnet, which will be used in the polarization
phase together with a microwave system for dynamical nuclear polarization. The polarization will be
preserved in the “frozen spin” mode at the temperature of about 50 mK by a very thin superconducting
solenoid located within the target cryostat, which will produce a 0.5 T longitudinal holding magnetic field.
In general, the maximal polarization of 90% is possible. The relaxation time of the target is about 200 h.
Firstly, it was proposed to measure the double polarization observables for 0,  and ω photoproduction
on the longitudinally polarized proton target. When using linearly polarized photons, we can measure the
double polarization asymmetry G. This asymmetry, for which scarce data exist, plays an important role to
disentangle contributions of various nucleon resonances by the imaginary part of the interference between
different partial waves. The asymmetry G has high sensitivity to the Roper resonance P11(1440) in 0
production at photon excitation energies around 600 MeV, to the resonances P11(1710) and P13(1720) in 
photoproduction around 1100 MeV, and to the recently found resonance D15(2070) at 1600 MeV. In an
experiment with circularly polarized photons, we were specially interested in measuring the helicity
difference of the cгoss sections d/d1/2 and d/d3/2 (and, correspondingly, the helicity asymmetry
E = [d/d1/2  d/d3/2] / [d/d1/2 + d/d3/2]) for 0 and  photoproduction off the proton. These
measurements give possibility to distinguish different model predictions of double polarization observables.
Furthermore, they allow to fix the quantum numbers of the newly discovered nucleon resonance at
W = 2070 MeV.
It should be stressed that first measurements of the polarization observable G have been performed
already for the reaction  p  p0 with linearly polarized photons at the energy from 800 to 1100 MeV using
the CB with the longitudinally polarized proton target. The first obtained results are published in Ref. [7].
To complete the series of experiments on measurements of the double polarization observables it was
necessary to make one more modernization of the set-up by changing the construction of the polarized target.
The cryostat of the Bonn frozen spin target was equipped with a saddle type superconducting holding coil
producing a transverse magnetic field. The coil is very thin in order to allow particles to pass the coil
virtually unaffected. The target material is butanol C4H9OH, which has an effective density of 0.57 g/cm3
and dilution factor D = 0.1335. In general, in this target modification the vector of its polarization is directed
transversally in respect to the direction of the incident photon beam. The maximal value of the target
polarization can reach 90%. The relaxation time of the target is about 250 hours.
A possibility of inserting the transversally polarized target into the Crystal Barrel allows to measure the
double polarization observables for 0,  and 0  photoproduction using both the circularly polarized and
the linearly polarized photon beams. In particular, the objects of further measurements are the double
polarization observables H, P, and T.
References
1. D. Elsner, …, D.E. Bayadilov, Yu.A. Beloglazov et al., Eur. Phys. J. A 33, 147 (2007).
2. E. Gutz, …, D.E. Bayadilov, Yu.A. Beloglazov et al., Eur. Phys. J. A 35, 291 (2008).
3. Frank Klein, ..., D.E. Bayadilov, Yu.A. Beloglazov et al., Phys. Rev. D 78, 117101 (2008).
4. D. Elsner, …, D.E. Bayadilov, Yu.A. Beloglazov et al., Eur. Phys. J. A 39, 373 (2009).
5. M. Sparks, …, D.E. Bayadilov, Yu.A. Beloglazov et al., Phys. Rev. C 81, 065210 (2010).
6. E. Gutz, …, D.E. Bayadilov, Yu.A. Beloglazov et al., Phys. Lett. B 687, 11 (2010).
7. A. Thiel, …, D.E. Bayadilov, Yu.A. Beloglazov et al., Phys. Rev. Lett. 109, 102001 (2012).
5