Double pi(0) Photoproduction on the Proton at GRAAL.

Abstract

The double pi(0) photoproduction off the proton has been measured in the beam energy range of 0.65-1.5 GeV. The total and differential cross sections and the Sigma beam asymmetry were extracted. The total cross section measured for the first time in the third resonance region of the nucleon shows a prominent peak. The interpretation of these results by two independent theoretical models infers mostly the selective excitation of P11- and D13-nucleon resonances.

Full text

Page 1

Double ?0Photoproduction on the Proton at GRAAL
Y. Assafiri,1,*O. Bartalini,2V. Bellini,3J. P. Bocquet,4S. Bouchigny,1M. Capogni,1,2M. Castoldi,5A. D’Angelo,2
J. P. Didelez,1R. Di Salvo,1,2A. Fantini,2L. Fichen,1G. Gervino,6F. Ghio,7B. Girolami,7A. Giusa,3M. Guidal,1
E. Hourany,1,†V. Kouznetsov,4,8R. Kunne,1J-M. Laget,11A. Lapik,8P. Levi Sandri,9A. Lleres,4D. Moricciani,2
V. Nedorezov,8D. Rebreyend,4C. Randieri,3F. Renard,4N. Rudnev,8C. Schaerf,2M. Sperduto,3M. Sutera,3
A. Turinge,10and A. Zucchiatti5
1IN2P3, Institut de Physique Nucle ´aire, 91406 Orsay, France
2INFN, Sezione di Roma II and Universita ` di Roma ‘‘Tor Vergata,’’ 00133 Roma, Italy
3INFN, Laboratori Nazionali del Sud and Universita ` di Catania, 95123 Catania, Italy
4IN2P3, Institut des Sciences Nucle ´aires, 38026 Grenoble, France
5INFN, Sezione di Genova, 16146 Genova, Italy
6INFN, Sezione di Torino and Universita ` di Torino, 10125 Torino, Italy
7INFN, Sezione Sanita ` and Istituto Superiore di Sanita `, 00191 Roma, Italy
8Institute for Nuclear Research, 117 312 Moscow, Russia
9INFN, Laboratori Nazionali di Frascati, 00044 Frascati, Italy
10I. Kurchatov Institute of Atomic Energy, Moscow, Russia
11DAPNIA/SPhN, CEA-Saclay, 91191 Gif-Sur-Yvette, France
(Received 17 January 2003; published 3 June 2003)
The double ?0photoproduction off the proton has been measured in the beam energy range of 0.65–
1.5 GeV. The total and differential cross sections and the ? beam asymmetry were extracted. The total
cross section measured for the first time in the third resonance region of the nucleon shows a prominent
peak. The interpretation of these results by two independent theoretical models infers mostly the
selective excitation of P11- and D13-nucleon resonances.
DOI: 10.1103/PhysRevLett.90.222001PACS numbers: 13.60.Le, 13.88.+e, 14.40.Aq
The study of nucleon resonances is a long-standing
challenge of hadronic physics. Up to now, they have
been mainly observed through ?N scattering and ?
photoproduction on the nucleon [1]. A more recent
method is the double pion photoproduction which gives
complementary information, particularly on resonances
that couple weakly to a single pion.
Among the three channels of the 2? photoproduction
on the proton (?p ! p?0?0, ?p ! p????, ?p !
n???0) the 2?0channel is the most selective one.
Because of the vanishing charge of the ?0, Born terms
associated with pion exchange are suppressed and since
the ? meson cannot decay into ?0?0, its production is
forbidden. Compared to the ? photoproduction channel,
which is almost exclusively sensitive to the S11?1535?
resonance [2], the 2?0channel appeared to be mainly
sensitive to P11?1440? and D13?1520? resonances, below
E?? 0:8 GeV [3].
The precise study of the 2?0photoproduction on the
proton has become possible with the advent of the new
generation of continuous wave electron accelerators and
large acceptance detectors. Three successive sets of data
from MAMI (Mainz Microtron) have covered the
cross section measurements up to 800 MeV [3–5]. The
present data measured at GRAAL (Grenoble Anneau
Acce ´le ´rateur Laser) extend the cross section results up
to E?? 1:5 GeV and provide invariant mass spectra and
beam asymmetry observables which strongly constrain
the theoretical models.
A theoretical model taking into account a largenumber
of diagrams has been developed by the Valencia group
[6–8], to study the general problem of the photoproduc-
tion of two pions, neutral or charged, on the proton and
the neutron. They satisfactorily explained the data ob-
tained at the MAMI accelerator [3–5] by allowing the
excitation not only of P11?1440? and D13?1520? but also of
D33?1700?.
Another model [9], which we refer to as the Laget
model, is the extension to higher energies of the Laget-
Murphy model [10]. Besides Born terms, it involves the
excitation of P11, D13, and D33resonances, allowing the
P11resonance to decay via the N? channel as well as the
?? channel. Indeed, the pair of ?0’s has the quantum
numbers of the ? meson and could be considered as
originating from the photoproduction of this meson and
its subsequent decay into two ?0’s.
The present data have been obtained with the GRAAL
setup, using a tagged and linearly polarized photon beam
and a large acceptance detector [2,11–13]. The photon
beam of 0.6 to 1.5 GeV is produced by backscattering a
laser beam on the electron beam of 6.04 GeVof the ESRF
(European Synchrotron Radiation Facility). The detector
consists of three layers: wire chambers, scintillators, and
calorimeters. In the central part, a bismuth germanate
(BGO) calorimeter of 90% of 4?, covered from the target
side with a barrel of scintillators, measures with a good
resolution the neutral mesons, ?0and ?, through their
decays into ?’s [14]. In the forward direction (? ? 25?) a
PH YSICA LR EVI EW L ET T ERS
week ending
6 JUNE 2003
VOLUME 90, NUMBER 22
222001-10031-9007=03=90(22)=222001(4)$20.00
 2003 The American Physical Society222001-1

Page 2

double wall of scintillators and a shower wall [15] mea-
sure the time of flight of the proton and of the neutron. On
the backward side, a large part of the hole is closed by
a disk of scintillators. A liquid H2target, 6 cm thick,
was used.
The energy spectrum of the photon beam is flat and has
a high degree of polarization in its upper part. For a given
measurement, there are choices of either the green line of
the laser giving an energy range from 0.6 up to1.1GeVor
theUV linegivinganenergy rangefrom0.8upto1.5GeV.
We made two experimental runs with both laser lines in
order to cover thefull energy range (0.65–1.5GeV) witha
degree of polarization between 0.6 and 0.98, and, con-
sequently, we could compare the results in the overlap
region. The GRAAL experiment is well adapted to mea-
suring the beam asymmetry ?. We applied the method
which was previously used for the photoproduction of a
single pseudoscalar meson [2,11–13].
The measurement and monitoring of the photon
beam flux ( ? 1:0 ? 106?=s) were carried out with
two independent detectors: one thin scintillator sensi-
tive to 2.7% of the intensity of the beam and a full
absorption lead-scintillator sandwich detector. The inci-
dent energy is measured by the tagging system with
an energy resolution of 16 MeV (FWHM) (full width at
half maximum).
For the reaction ?p ! p?0?0, the ? and ? angles of
the proton were measured at large ? angles by the BGO-
barrel assembly and in the forward direction by the
double wall of scintillators. The energy and the angles
of the four ?’s are measured by the BGO calorimeter or
by the forward shower wall.The energy loss and the time
spectra given by the barrel and the forward doublewall of
scintillators serve for the identification of the proton and
the monitoring of random events.
Two types of events were selected, one with four ?’s
detected in the BGOcalorimeter ( ? 440000 events, with
UV line), and the other allowing for only one among
the four ?’s to be detected in the forward shower wall
( ? 220000 events, with UV line). In the latter case, only
the angles could be measured and the energy is deduced
from the momentum and energy balance of the reaction.
The resolution (FWHM) of the reconstructed invariant
mass of ?0is 23 MeV when the two ?’s are seen in the
BGO and 30 MeVwhen only one ? is in the BGO and the
other in the shower wall.
Since a fivefold coincidence was required between the
four?’sandtheproton,andsincetherandom ratewasless
than 1.5%, the background was less than 1%. The con-
tamination from the ?p ! p? where the ? can decay
into three ?0’s is less than 1% in any bin of the beam
energy.
The acceptance correction, which was deduced with
the simulation code LAGGEN built on the GEANT3 code
from the CERN library, uses three kinematical variables:
thephotonbeam energy,andthe momentum andtheangle
of the 2?0system.The acceptance correction involved an
average efficiency of 30% and an extrapolation for the
uncovered phase space of less than 3% at any photon
energy. Two types of event generators for the simulation
wereused,basedonphasespace distributionsofeither the
?p ! ???0reaction or the ?p ! p?0?0reaction. Both
distributions produced compatible outputs for the total
cross section, the invariant mass spectra, and the beam
asymmetry.
The total cross section for the reaction ?p ! p?0?0is
shown in Fig. 1 for an incident energy going from the
threshold to 1.5 GeV. At 0.7 GeV there is a peak confirm-
ing the previous data sets from MAMI [3,5].This peak is
mainly attributed to the excitation of P11?1440? and
D13?1520? resonances. However, the GRAAL peak is
narrower than the one in the recent set of MAMI which
shows an excess at the high energy side. At high energy,
the GRAAL data show a new feature not seen before, a
peak located at 1.1 GeV.
We show in the same figure the calculations of Oset
(Valencia group) [6,7] covering the low energy part
and the calculation of Laget covering the whole range
0
300
4
8
12
500700 9001100 1300
Eγ (MeV)
1500
σ (µb)
1300 1400 15001600 1700 1800 1900
W (MeV)
GRAAL data
TAPS(1996)
TAPS(2000)
Lines − and -- Oset
Line ... Laget
12
3
4
a
c
b
0
300
2
4
6
8
500 700900
FIG. 1(color online).
p?0?0. GRAAL data together with previous data from
MAMI, TAPS(1996) and TAPS(2000) and theoretical calcula-
tion from Oset and Laget models. The GRAAL data are shown
with statistical error bars only and the systematical error is
estimated to be 3%. In the lower part the partial contributions
of the different diagrams of Laget model are given, labeled
with numbers: label 1 associated with the dash-dotted line
corresponds to ?p ! P11?1440? ! ??; label 2, dashed lines,
?p ! D13?1520?;D13?1700? ! ??; label 3, continuous line,
?p ! P11?1440?;P11?1710? ! ?p; label 4, dotted line ?p !
?p. The contribution of D33?1700? is too small to be seen. In
the inset, the main contributions of the Oset calculation
are given [6]: thick solid line, full calculation; lines (a), (b),
and (c) for D13?1520?, ?, and P11?1440? intermediate states,
respectively.
Total cross section of the reaction ?p !
PH YSICA LR EVI EWL ET T ERS
week ending
6 JUNE 2003
VOLUME 90, NUMBER 22
222001-2 222001-2

Page 3

of energy. The calculations of Oset for two different
parametrizations, drawn in continuous and dashed
lines, include N, P33?1232?, P11?1440?, D13?1520?, and
D33?1700? as intermediate baryonic states. They give a
peak at 700 MeV with a width compatible with GRAAL
data. The Laget model takes into account three mecha-
nisms in the 2?0channel: (i) the excitation of P11?1440?,
D13?1520?,P11?1710?,D13?1700?,and D33?1700? resonan-
ces which decay into a ??and ?0, followed by the decay
of the ??into a proton and ?0, (ii) the excitation of
P11?1440? and P11?1710? followed by a direct decay to
the proton with the emission of a ? meson which decays
into?0?0,and(iii) the direct emissionof the?mesonvia
? exchange between the incoming photon and the target
nucleon.The parameters of the resonances taken by Laget
aregiven in Table I.Theyhave been adjusted in order to fit
the total cross section and lie in the range of previously
determinedvalues[1].ThemassofP11?1440? resonanceis
at theedge,but it issowide that a precise determination is
meaningless.Theresult isdrawnasthe dotted lineandthe
contributions of the different diagrams are plotted in the
lower part of the figure. The model reproduces quite
satisfactorily the two peaks which dominate the cross
section. In both cases, the key to the success is the
interference between the direct emission of the ? meson
and the decay of the relevant P11resonance into the ?N
channel: this is the only way to get the magnitude of the
cross section with reasonable resonance parameters (see
Table I), compatible with Particle Data Group data [1]. It
is worth noting that both models reproduce the total cross
section at low energy with different combinations of
resonances: the D13?1520? dominates theValencia model,
while the P11?1440? dominates the Laget
Comparisons with experimental differential cross sec-
tions and spin observables are clearly needed.
The differential cross sections were evaluated as a
function of the invariant mass of any group of two par-
ticles in the final state. In Fig. 2 the experimental results
of GRAAL are presented with dots, on the left-hand side
in terms of the invariant mass of the two ?0’s and on the
right-hand side in terms of the invariant mass (IM) of the
model.
proton and any one of the two ?0’s, for four narrow bins
(40 MeV) of beam energy centered at 720, 850, 1100, and
1300 MeV. The corresponding differential cross sections
predicted by the Laget model are plotted as a continuous
line, while the empty circles on the 720 MeVspectra are
experimental data from MAMI. The phase space results
are also shown as a dashed line for the ?p ! ???0
reaction and as a dotted line for the three-body final state
reaction ?p ! p?0?0. The overall trends are similar for
the GRAAL data and for the Laget predictions. At the
two lower energies the structures are better reproduced
than at the higher ones. At high energy, the experimental
data show a stronger peaking in the IM?p?0? plot at the ?
mass position with a complementary bump on its right.
Looking at the phase space results, the shape of GRAAL
data is roughly intermediate between the shapes of ?p !
???0and ?p ! p?0?0.
In Fig. 3, the beam asymmetry ? results are presented
in terms of the invariant mass in a way similar to that in
Fig. 2. The error bars originate from the statistics and the
fit to the ? distribution. The systematical error is 3% and
is mainly due to the determination of the polarization.
The asymmetries calculated by the Laget model are
plotted as continuous lines. At the lowest beam energy
[part (a)], the experimental asymmetries are small and
hardly reproduced by the Laget calculation. At the other
energies [parts (b), (c), and (d)], the asymmetries are
sizable and the calculation reproduces the sign and the
general trend of the data. Another presentation of the
beam asymmetries, in terms of ?CM, is reported in
Fig. 4, where at the lowest energy bin (a) the results of
the Valencia group [8] are also plotted and show poor
agreement with the data.
TABLE I.
model. The decay widths for the ? channel and ‘‘others’’ are
given for completeness but do not contribute to the ?0?0
channel.
Parameters of the resonances used in the Laget
Widths in MeV
Partial
??
??N
0.324
0.055
0.79
0.30
0.010
Mass Full
?
Reson. in MeV
???
24
85
50
158
76
??N
0
0
0
25
25
??N
72
10
81
167
34
?others
0
0
110
0
0
D13
D13
D33
P11
P11
1520
1650
1700
1500
1720
120
100
250
350
135
0.00
0.05
0.10
γp→p(πoπo)
720 MeV
γp→(pπo)πo
720 MeV
0.00
0.025
dσ/d(IM) (µb/MeV)
850 MeV850 MeV
0.00
0.02
1100 MeV1100 MeV
0.00
0.01
400600800
IM(πoπo) (MeV)
1300 MeV
1200 1400 1600
IM(pπo) (MeV)
1300 MeV
FIG. 2.
at four beam energies. See explanations in the text. At E??
720 MeV the insets compare the Laget (solid line) and Oset [6]
(dashed lines) models with the GRAAL data (dots).
Invariant mass spectra for the reaction ?p ! p?0?0
PH YSICA LR EVI EWL ET T ERS
week ending
6 JUNE 2003
VOLUME 90, NUMBER 22
222001-3222001-3

Page 4

To summarize, the 2?0photoproduction off the proton
has been measured from 650 up to 1500 MeV , extending
previous data. The total cross section shows evidence of
two prominent peaks: one already seen at 700 MeV and
corresponding to the excitation of the P11?1440? and
D13?1520? resonances and the other at 1100 MeV , seen
for the first time and corresponding to the excitation of
the P11?1710? and D13?1700? states. The sign of the beam
asymmetry and the general trend of the differential cross
section and the beam asymmetry were reproduced satis-
factorily by the Laget model, where the ? meson was
involved in two dominant mechanisms, namely, the ex-
citation of P11resonances which decay by ? emission and
the direct emission of the ? meson during the initial ?
exchange. Our interpretation with the Laget model is
based on a minimum set of resonances whose parameters
lie in the range of accepted values. It has to be confirmed
by a full partial wave analysis requiring a large database
for whichourdatawould represent a significant input.The
2?0channel not only would provide a powerful filter to
study P11resonances but also would be an important
source of the ? meson. Finally, our conclusions support
a dedicated program of 2?0photoproduction on heavy
targets, in order to study the propagation of the ? meson
in hadronic matter [16–19].
It is a pleasure to thank the ESRF for a reliable and
stable operation of the storage ring and the technical staff
of the contributing institutions for essential help in the
realization and maintenance of the apparatus.
*On sabbatical leave from the Faculty of Sciences of the
Lebanese University, Tripoli, Lebanon.
†Electronic address: hourany@ipno.in2p3.fr
[1] D. E. Groom et al., Eur. Phys. J. C 15, 1 (2000).
[2] F. Renard et al., Phys. Lett. B 528, 215 (2002).
[3] M. Wolf et al., Eur. Phys. J. A 9, 5–8 (2000).
[4] A. Braghieri et al., Phys. Lett. B 363, 46 (1995).
[5] F. Harter et al., Phys. Lett. B 401, 229–233 (1997).
[6] J. A. Gomez-Tejedor and E. Oset, Nucl. Phys. A600, 413
(1996).
[7] J.C. Nacher et al., Nucl. Phys. A695, 295–327 (2001).
[8] J.C. Nacher and E. Oset, Nucl. Phys. A697, 372–387
(2002).
[9] J.-M. Laget and L. Murphy (private communication).
[10] L. Murphy and J. M. Laget, CEA Report No. DAPNIA
96-10, Saclay, 1996.
[11] J. Ajaka et al., Phys. Rev. Lett. 81, 1797 (1998).
[12] J. Ajaka et al., Phys. Lett. B 475, 372 (2000).
[13] O. Bartalini et al., Phys. Lett. B 544, 113 (2002).
[14] F. Ghio et al., Nucl. Instrum. Methods Phys. Res., Sect. A
404, 71 (1998).
[15] V. Kouznetsov et al., Nucl. Instrum. Methods Phys. Res.,
Sect. A 487, 396 (2002).
[16] P. Schuck (private communication).
[17] P. Schuck et al., Z. Phys. A 330, 119 (1988).
[18] T. Hatsuda et al., Phys. Rev. Lett. 82, 2840 (1999).
[19] J.G. Messchendorp et al., Phys. Rev. Lett. 89, 222302
(2002).
-0.4
0.2
-0.2
0.0
0.2
γp→p(πoπo)
(a)
γp→(pπo)πo
(a)
-0.4
0.2
-0.2
0.0
Beam Asymmetry Σ
(b) (b)
-0.4
0.2
-0.2
0.0
(c)(c)
-0.4
-0.2
0.0
200400 600800
(d)
IM(πoπo) (MeV)
1000 1200 1400 1600 1800
IM(pπo) (MeV)
(d)
FIG. 3 (color online).
tion ?p ! p?0?0at four beam energies: (a) 650–780 MeV ,
(b) 780–970 MeV , (c) 970–1200 MeV , and (d) 1200–1450 MeV.
On the left, ? as a function of the IM of the system (?0?0) and
on the right, of the IM of the system (p?0). The continuous line
shows the Laget model results. The reaction plane used to
define the ? angle in order to extract ? was, on the left, the
plane of the incident photon and the outgoing proton, and on
the right, of the incident photon and any one of the two
outgoing ?0’s .
The beam asymmetry ? for the reac-
-0.3
0.0
γp→p(πoπo)
(a)
γp→(pπo)πo
(a)
-0.3
0.0
Beam Asymmetry Σ
(b)(b)
-0.3
0.0
(c)(c)
-0.3
0.0
090
(d)
ΘCM(πoπo) (Deg.)
0 90180
(d)
ΘCM(πo) (Deg.)
FIG. 4 (color online).
tion ?p ! p?0?0at for bins of beam energy in terms of
?CM??0?0? (on the left) and of ?CM??0? (on the right). The
calculations of Laget (solid line) and the Valencia group
(dashed line) are shown. Notations as in Fig. 3.
The beam asymmetry ? for the reac-
PH YSICA LR EVI EWL ET T ERS
week ending
6 JUNE 2003
VOLUME 90, NUMBER 22
222001-4222001-4